sign in sign up
<<
Home , Coppery brushtail possum, Eastern barred bandicoot, Ekaltadeta, Gerard Krefft, Golden bandicoot, Julia Creek dunnart

LIFE OF

HUGH TYNDALE-BISCOE © Hugh Tyndale-Biscoe 2005

All rights reserved. Except under the conditions described in the Australian Copyright Act 1968 and subsequent amendments, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocop- ying, recording, duplicating or otherwise, without the prior permission of the copyright owner. Contact CSIRO PUBLISHING for all permission requests.

National Library of Cataloguing-in-Publication entry Tyndale-Biscoe, Hugh. Life of marsupials.

[New ed.]. Bibliography. Includes index. ISBN 0 643 06257 2 (Hardback). ISBN 0 643 09199 8 (Paperback). ISBN 0 643 09220 X (netLibrary eBook).

1. Marsupials. I. CSIRO Publishing. II. Title.

599.2

Available from CSIRO PUBLISHING PO Box 1139 (150 Oxford St) Collingwood VIC 3066 Australia

Telephone: +61 3 9662 7666 Local call: 1300 788 000 (Australia only) Fax: +61 3 9662 7555 Email: [email protected] Web site: www.publish.csiro.au

Front cover (clockwise from top left): Gray four-eyed opposum (Hugh Tyndale-Biscoe); Stages in tammar development: the unattached vesicle (Ivan ) and a newborn tammar (LA Hinds); Chromosome painting (JAM Graves); (Ederic Slater); Julia Creek with 60-day-old litter (PA Woolley and D Walsh).

Spine: Yellow-footed rock wallaby (Esther Beaton).

Back cover: Male on Banksia inflorescence (PA Woolley and D Walsh).

Set in Minion 10/12 Cover and text design by James Kelly Typeset by J&M Typesetting Printed in Australia by Ligare Contents

Preface v

1 What is a ? 1

2 Reproduction and development 37

3 Opossums of the Americas: cousins from a distant time 103

4 Predatory marsupials of Australasia: bright-eyed killers of the night 139

5 : fast-living opportunists 165

6 Pygmy possums and sugar gliders: pollen eaters and sap suckers 183

7 Life in the trees: , and possum 219

8 : vegetarians of the underworld 267

9 Consummate 287

10 Marsupials and people: past and present 365

References 385

Index 421 This page intentionally left blank Preface

When the first edition of this book was written in 1970, the old debate about the inferior status of marsupials, compared to other , was still active. The work reviewed then on a few of marsupial in Australia and North America was beginning to dispel this idea but it still prevailed in other countries, particularly in the Northern Hemisphere. Thirty-five years later it is no longer an issue of importance. Now, much more is known about the past history and the present relationships of marsupials in Australia, and South and Central America, so that the long evolution of this separate line of mammals is becoming much clearer. As well as this, there are now detailed studies on the physiology, reproduction, ecology and behaviour of representatives of all the main families of marsupials, so that comparisons and generalisations can be made with much more confidence. In the first edition of this book it was also still possible to cover the whole literature on marsupials. That is no longer possible and a small book now must be selective in its coverage and its acknowledgement of sources. However, for most topics and for most groups of marsupials there are now excellent monographs or reviews that enable the interested reader to follow any topic further. Several themes about marsupials have developed among the people who have studied them during the past 35 years and these resonate through all the work that is described here. The oldest of these themes is the remarkable convergence of adaptations seen in Australasian marsu- pials and mammals on other continents. When first seen by European explorers these similari- ties were thought to indicate close relationship but deeper understanding soon showed that these were two independent evolutionary lines responding to similar external imperatives. As well as these large convergences we can now recognise convergences between marsupials from Australasia and those from South America. Within the Australasian marsupials there are also convergences to similar food sources, such as the leaf eating koala, possums and ringtails, or the nectar-eating marsupials from four separate families. However, the most interesting outcome of the new work on marsupials has been a much greater appreciation of how marsupials have adapted to the special conditions of the Australian environment, its unpredictable climate, low fertility soils and unpalatable plants. It is an important and interesting aspect of the adaptive radiation of marsupials in Australia and raises the question how marsupials came to prevail in Australia but shared South America with other kinds of mammals: it also tells us how we must adapt to the land if we wish to live here in the long term, and what we must do to let these long time residents continue to live here also. Because it is not possible for one person to command a knowledge of so many fields as this book covers, I have depended on the expert advice of colleagues in several fields: while taking full responsibility for what is written, I am deeply grateful for the generous help of Ken Aplin, Bill Foley, Jennifer Graves, Brian Green, Stephen Ho, Peter Janssens, John Kirsch, the late Richard Mark, Lauren Marotte, Jim Merchant, David Ride, Phil Waite, Mike Westerman and Patricia Woolley. For each reading several chapters as a non-expert and thereby helping me to express things more clearly than I otherwise would have, I sincerely thank Meredith McKinney and Nicola Tyndale-Biscoe. I am also very pleased to acknowledge CSIRO: this great organisation has supported research on Australasian marsupials since 1950 and my own research for more than 40 years, so that much that is discussed in this book stems directly from that support. Then, when I began this book Brian Walker, Chief of CSIRO Wildlife and Ecology, offered me generous and stimulating hospitality to prepare it, and his successors in CSIRO Sustainable Ecosystems have graciously continued to do so, to its completion. At CSIRO Sustainable Ecosystems I have been wonderfully well supported by many people and I especially thank Alice Kenney for preparing the figures; Margaret Hindley, Trish Kelly, Megan Edwards and Inge Newman for tracking down difficult or unusual references with speed and efficiency; and Andrew Bishop, Brian Davis and Yechiam Marks for leading me courteously through the complexities of information technology. At CSIRO Publishing I thank Paul Reekie for great patience as deadlines passed and Nick Alexander and Briana Elwood for producing the finished work with diligence and despatch. I also thank Alexa Cloud for superb copy editing. Finally, I thank Marina, who read and commented on every chapter in draft and then read the proofs, and has sustained me throughout the whole saga as one year passed into another and the end remained a mirage too far away: thank you for everything.

Hugh Tyndale-Biscoe January 2005 Chapter 1

What is a marsupial? South American opossum; steel engraving from Buffon (1749). What is a marsupial? 3

his is the story of a group of mammals that were isolated from the rest of the world for many millions of years. It is set on the great southern continent of Gondwana that Tstretched from the Caribbean to the islands of New Guinea and included the three present-day continents of South America, Antarctica and Australasia. The characters are the marsupial mammals and the plot is how they came to be there and how they adapted to the special conditions of their vast homeland. Today marsupials only occur in Australasia and the Americas, although fossil marsupials have been discovered on every continent of the world. If they occurred on all the continents in the past, why are they not more widely distributed today? We first need to ask whether marsupi- als really share a common relationship closer than that to any other group of mammals. If they do, where did marsupials originate and how did they come to be where they are today? The features that were described first by the Europeans were the pouch of the female and the extraordinarily small size of the young at birth. These are their two most distinctive features and reproduction is what sets marsupials apart from other mammals and permeates the life history of every species. But what confused the early European explorers was that many marsupials closely resembled mammals more familiar to them that follow similar life styles.

First European encounter with American marsupials The first marsupial brought to Europe from America was a common opossum collected by Vincente Yañez Pinzón on his first voyage to the New World in 1500. He collected a female, which had young in its pouch, and later he presented it to Queen Isabella and King Ferdinand II in Grenada. By the end of the voyage to Spain the young were gone and the mother opossum dead but the Queen inserted her fingers into the ‘second belly’ of this strange creature from the New World. This extraordinary organ and the young it enclosed caused astonishment in scientific circles throughout Europe and led to speculation about how the tiny young reached the pouch – did they grow from the teats as buds, or were they blown into the pouch from the mother’s nostrils? Both ideas had a long currency but the equally astonishing fact that they crawl to the pouch unaided by their mother was not discovered for another 420 years. The strange appearance of this New World , with a fox-like head and a monkey’s hands, also puzzled European scientists, who described it as the monkey-fox or ‘simivulpa’.

First European contact with Australasian marsupials In 1493 Pope Alexander IV divided the world between the two major European powers – Spain to the west and Portugal to the east – so it was Spanish explorers who discovered marsupials in the Americas, while Portuguese traders made the first observations on Australasian marsupials. The earliest description was by Antonio Galvao, Station Captain of the Portuguese settlement on Ternate in the Moluccas from 1536 to 1540. He brought back to Lisbon extensive notes from which he intended to write a treatise on the Moluccas. He wrote (Jacobs 1971):

Some resemble ferrets, only a little bigger. They are called Kusus. They have a long tail with which they hang from the trees in which they live continuously, winding it once or twice around a branch. On their belly they have a pocket like an intermediate balcony; as soon as they give birth to a young one they grow it inside there at a nipple until it does not need nursing any more. As soon as she has borne and nourished it, the mother becomes pregnant again.

This is a good description of the common , orientalis, which still lives on Ternate. Galvao’s manuscript lay forgotten in the Jesuit Library at Seville for 400 years, until discovered and published by Father Hubert Jacobs in 1971. However, the manuscript may well 4 Life of Marsupials

have circulated in Europe because there are other references in the 17th century to animals similar to the opossum being found in the Moluccas (eg Piso 1648). As the rivalry of Portugal and Holland for the rich takings of the Spice Islands increased, travellers also reported on the strange animals of New Holland and New Guinea. In 1606 Captain Don Diego de Prado y Tovar landed at San Millaus Bay, on the southern coast of New Guinea and wrote (quoted by Stevens 1930):

Here we killed an animal which is in the shape of a dog, smaller than a greyhound, with a bare scaly tail like that of a snake, and his testicles hang from a nerve like a thin cord; they say it was the castor, we ate it and it was like venison, its stomach full of ginger leaves and for that reason we ate it.

The species that most nearly fits this description (Calaby 1965) is the Thylogale brunii, which makes this the earliest European record of a member of the family. The Dutch Captain Pelsaert, wrecked on the inhospitable Abrolhos Islands off the west coast of Australia in 1625, described another animal with the same remarkable pouch and tiny offspring within it. This was the tammar wallaby, eugenii, the second member of the kangaroo family to be described by Europeans. Then on Dirk Hartog Island, the buccaneer, William Dampier in 1699, described the banded hare wallaby, Lagostrophus fasciatus:

A sort of raccoons, different from those of the West Indies chiefl y as to their legs; for these have very short fore-legs; but go jumping upon them as the others do, and like them are very good meat.

Other navigators were also encountering similarly bizarre animals when they made landfalls on the coast of and, from their reports, scientists in Europe recognised the similarities to the American animals. Pallas (1766) named the cuscus from the Moluccas Didelphis orientalis but Storr (1780) 14 years later noticed that the 2nd and 3rd toes or phalanges of their hind feet are partly fused and changed it to Phalanger orientalis, the name it has today. It has also given its name to the family of Australasian possums and , the .

Adaptive radiations on different continents While all the early European observers were much struck by the pouch of female marsupials, they were also struck by the astonishing similarities of Australian marsupials to familiar mammals from Europe. In the scientific names they gave them they often used Greek prefixes that meant pouch or pocket and the scientific name of the familiar that the marsupial resembled. So we find Pera-meles the pouched badger, Pera-dorcas the pouched antelope, Phascol-arctos, the pouched bear, Thylo-gale, the pouched hare and Thyla-cinus, the pouched dog. As the Australian continent was explored yet more remarkable similarities were discov- ered, the most extraordinary being the , Notoryctes typhlops, which looks and behaves like the golden mole of Namibia, Eremitalpa granti, especially in that both species ‘swim’ through dry sand, which collapses behind them leaving no burrow; and the exquisite , Myrmecobius fasciatus, adapted to living on termites and ants. The tree-living marsupials of Australia and New Guinea show remarkable yet superficial likenesses to the various species of lemur of Madagascar (ie Daubentonia madagascarensis) and the tree sloths (ie Bradypus tridac- tylus) of South America. More remarkable yet are the many physiological similarities between marsupials and other mammals in such functions as fermentation of grass in the forestomach of kangaroos and ruminants. Apart from a pouch, what do all these species that are called What is a marsupial? 5

marsupials have in common that tells us that they are uniquely related to one another and separate from all the other mammals that they variously resemble?

Distinctive features of marsupials The person who first looked beneath the superficial similarities to find the fundamental criteria for determining relationships between mammals was de Blainville in 1816. He took as the defin- ing character the anatomy of the female reproductive tract (Fig. 1.1). In marsupials there are two vaginae, two uteri and two oviducts, whereas in other mammals there is a single vagina, cervix and uterus and only the oviducts are paired. He named the marsupials the Di-delphia from the Greek words for two uteri and other mammals he called the Mono-delphia. He later realised that the platypus and echidna did not fit in either group, having a reproductive tract like that of birds and reptiles with a single opening for discharging products from the gut, the bladder and the gonads, and he termed them the Ornitho-delphia, or bird-uterus. His division of mammals has stood to this day, although his terms have not been retained. Instead, the Ornithodelphia are known as the Monotremata (Greek for one hole), the Didelphia are generally called the Marsupialia (Latin for pocket or pouch) and the Monodelphia are called the Placentalia, because of the great development of the placenta as an organ of exchange during pregnancy. These terms are also unsatisfactory because they are not exclusive. Thus, the marsupials share with the monotremes a single opening, the cloaca, for the discharge of all products. Second, not all female marsupials possess a well-developed pouch, whereas the female echidna, which is a monotreme, develops a pouch during lactation. And third, all marsupials have a placenta during intra-uterine development and in some species it is a complex structure with an intimate connection to the uterus. To avoid this confusion of terms some biologists favour Huxley’s (1880) terms for the living mammals. He saw the three groups of mammals as evolutionary stages on the way to ‘true’ mammals, by which he meant the group of mammals to which we human beings belong. So he coined the terms Proto-theria, or first mammals (the monotremes), Meta-theria or halfway mammals (the marsupials) and Eu-theria or true mammals (the placentals). These terms themselves imply progress, and so are also unsatisfactory. It is probably best now to ignore their original meanings and accept the terms as neutral descriptors of the groups they have been assigned to: in this book I will refer to monotremes, marsupials and placentals.

Anatomy of the reproductive organs of marsupials De Blainville was right to choose the anatomy of the reproductive organs as the central criterion of his classification because they are unequivocally distinct between the three groups of living mammals. To appreciate this we must consider the development of the kidney and the ducts that convey urine, and the genital ducts that convey gametes and embryos to and from the body. If we compare a fetus from a tammar wallaby four days before birth and a human embryo at five weeks of gestation, for example, both have the same arrangement of kidney ducts, genital ducts and gonads (Fig. 1.1). Both the genital ducts and the kidney ducts (ureters) enter a common tube, the urogenital sinus, on its dorsal side and the future bladder is on the opposite, ventral side. In later development of both groups the ureters migrate to the ventral side to enter the bladder, while the genital ducts remain dorsal. This migration of the ureters only occurs in placentals and marsupials and probably arose as an adaptation for the more effective storage of urine in the bladder. In the monotremes, the ureters still discharge into the top of the sinus and the urine must pass across the sinus to enter the bladder, by a process still not understood. 6 Life of Marsupials

Figure 1.1: The primary difference between marsupial and placental mammals. The relative position of the ureters and genital ducts and how it is derived from a common pattern of kidney and genital ducts at the indifferent stage in the embryo.

In the human, and all placental mammals, the ureters migrate outside and below the genital ducts, while in the tammar, as in all other marsupials, the ureters migrate inside and above the genital ducts. While the initial adaptation may have had more to do with excretion in the ancestral mammals, the route that the ureters take to reach the bladder unequivocally distinguishes living marsupials from living placentals. And it has had profound consequences for reproduction. Its most obvious effect is seen in the female reproductive tract. Each oviduct transforms into the specialised regions of Fallopian tube, uterus, cervix and vagina (Fig. 1.1). In placental mammals the left and right oviducts join in the midline to form a single vagina and, in most species, also join to form a single cervix and a single uterus. The only portions that remain separate are the two Fallopian tubes, which each receive an egg from their respective ovary. What is a marsupial? 7

In marsupials this joining of the oviducts cannot occur because the ureters pass between them, so there are two lateral vaginae, each arising from the common urogenital sinus posteriorly. Above the ureters the lateral vaginae loop back to the midline and become partially fused (Fig. 1.1). Then there are two cervices, two uteri and two separate Fallopian tubes. At copulation the semen is deposited in the lateral vaginae and the sperm pass through the two cervices and uteri to the Fallopian tubes, where fertilisation occurs. In the males of many species the glans of the penis is also divided and it is supposed that the semen is directed separately to each lateral vagina, but this has not been established. The same relationship of the ureters and the genital ducts in males means that the vas deferens in placental mammals loops over the ureter, whereas in marsupial males the arrangement is simpler (Fig. 1.1). At birth the young marsupial passes through a new-formed canal in the tissues between the ureters, direct from the lateral vaginae to the urogenital sinus. In almost all marsupials this extraordinary arrangement, the pseudo-vaginal or birth canal, is re-formed at each birth: only in some kangaroos, Macropus, and the honey possum, Tarsipes rostratus, does it remain open after the first birth and in these species it is called the median vagina. So, female kangaroos have three vaginas, two for sperm and one for the young at birth. It has long been held that the small size of the divided uteri and the inadequacy of the birth canal are the reasons that marsupial young at birth are so very small. Whether or not this was the cause, the young of all marsupials at birth are much smaller than the most immature of placental young.

Figure 1.2: Relationship between maternal body weight and weight of the newborn marsupial and its stage of development in different species. 8 Life of Marsupials

Small size at birth Females of the largest living marsupials, the , Macropus giganteus, and the red kangaroo, Macropus rufus, weigh 28 kg and deliver one young that weighs 830 mg, half the size of a newborn mouse. The newborn kangaroo is 0.003% of its mother’s weight, compared to a newborn mouse or human baby that are about 5%: almost a two thousand-fold difference. Most marsupials at birth are between 200 mg and 400 mg (Fig. 1.2) but those of small dasyurid marsupials are under 10 mg and the very smallest newborn marsupial is that of the honey possum at 4 mg. It is little wonder that the first European explorers to see these tiny creatures were unable to believe that they were born in the normal manner or could travel unaided to the pouch and there find and attach to a teat. It seemed impossible that the young marsupials could be active and possess sensory organs or have sufficient neuromuscular coordination to move independently. But they do and, even more remarkably, they control the onset of their own birth like the much more advanced placental young do (see Chapter 2). One reason often cited for the small size of marsupials at birth is that pregnancy is very short, compared to placental counterparts. Although it is true that in some species pregnancy is very short (less than two weeks), in other species it is longer than in a comparable sized placental species. The real point is not the duration of pregnancy but the mass of young produced at the end of pregnancy. All placental mammals bring forth very much larger young, at a more advanced stage of development, because the major growth phase of the young occurs during pregnancy, via a well-developed placenta. By contrast, in marsupials almost all growth and development occurs after birth during the long and complex period of lactation, when large changes in the volume and composition of the milk occur that support the changing needs of the young. Thus, the placental female makes her major investment in reproduction during pregnancy, the female marsupial makes hers during lactation. This difference has important consequences in the ecology of the species.

Control of sexual differentiation in marsupials Superficially the external genitalia of marsupials and placentals are similar but the control of devel- opment to the adult form is different in the two groups of mammals. In placental mammals sexual differentiation of the fetus takes place during gestation when the external appearance of both sexes is the same, the so-called indifferent stage. Later under the influence of testosterone secreted by the developing testes of male fetuses the genital tubercle develops into a penis and the scrotum forms behind it; in the absence of testes the same structures develop in female fetuses into the clitoris and the outer lips of the vulva, respectively; nipples and mammary glands are formed in both sexes and retained through life. The whole cascade of change from the indifferent stage to the adult form is controlled by the expression of one gene on the tiny Y chromosome, called the ‘sex determining region’ of the Y chromosome, or SRY gene. Possession of this gene directs the gonads to differenti- ate into testes, which then secrete testosterone and transform the other genital structures to the male pattern: in the absence of the SRY gene the fetus becomes female. The normal complement of sex chromosomes in placentals and in marsupials is two X chromosomes in females and one X and one Y chromosome in males, but an individual with one X and no Y chromosome (XO) is female and an individual with two X and one Y chromosome (XXY) is male. By contrast in marsupials the scrotum forms as two bulges in front of the genital tubercle, there is nothing equivalent to the outer lips of the vulva in female marsupials, and the pouch and mammary glands only differentiate in females. Male marsupials have a Y chromosome and the SRY gene, which directs the differentiation of the gonads to become testes and secrete testosterone leading to the differentiation of the internal genitalia and the genital tubercle to the male form: absence of the SRY gene results in the female form of the internal genitalia and genital tubercle. What is a marsupial? 9

However, the developing testes do not control the differentiation of mammary glands, pouch or scrotum. Indeed, scrotal bulges develop in genetic males and mammary glands and pouch in genetic females many days before the gonads can be distinguished as ovary or testis, and these organs are not affected in their later development by sex hormones. The present thinking is that these external organs in marsupials (mammary glands, pouch or scrotum) are controlled directly by the sex chromosome constitution of the tissues themselves, particularly the X chromosomes (Cooper 1993). Thus, possession of one X chromosome, as in a normal male marsupial, leads to differentiation of scrotal bulges; and possession of two X chromosomes, as in a normal female marsupial, leads to differentiation of mammary glands and a pouch (Renfree et al 1996a). In genetically abnormal tammars, XO individuals have female organs internally, as in placental species, but externally they do not have mammary glands or a pouch but do have a well-developed, empty scrotum. Conversely, XXY tammars have male organs internally and a well-developed penis, in accordance with their possession of a Y chromosome, but instead of a scrotum, they have a small pouch and mammary glands, in accordance with their possession of two X chromosomes (Sharman et al 1990). Thus, the control of sexual differentiation in marsupials has followed a different path from that followed by placentals, although the end result is deceptively similar.

Physiological differences between marsupials and other mammals Marsupials, like other mammals and birds, maintain their body temperature at a fairly constant level. However, the normal body temperature of marsupials is about 35.5°C, which is 2.5°C lower than that of most placentals, which in turn are lower than birds (Table 1.1). It is not clear why this should be so but it does appear to be something that is genetically determined and affects the lives of marsupials as profoundly as their mode of reproduction. To appreciate this we need to understand the underlying physiological process of which body temperature is an outward manifestation. Since the rate of chemical reactions doubles for every rise of 10°C, the rate in marsupials must be about 25% lower than in placentals, which in turn must be about 25% lower than passerine (song) birds. This is clearly seen in the cost of maintaining a constant body temperature, which rises with increasing basal body temperature (BBT) (Table 1.1).

Table 1.1: A comparison of the standard metabolic rate of terrestrial vertebrates Standard metabolic rate (SMR; kJ/kg0.75 per day); basal body temperature (BBT; °C) (after Dawson and Hulbert 1970).

Reptiles Mammals Birds Lizards Monotremes Marsupials Placentals Non-passerine Passerine BBT 30 30 35.5 38.0 39.5 40.5 SMR 31 142 204 289 347 598 SMR at 38°C 82 260 260 289 301 477

To maintain their body temperature mammals and birds expend the least amount of energy when the surrounding temperature is nearly the same as their body temperature. When the surrounding, or ambient, temperature is lower more energy is required to generate heat and when it is higher than the body temperature more energy is expended in cooling devices, such as panting and sweating. The ambient temperature where minimum energy is used by the non-feeding mammal at rest is called its thermo-neutral zone. This minimum value is termed the standard metabolic rate (SMR) and it represents the energy required to maintain essential 10 Life of Marsupials

functions of the living body at a constant body temperature. It is usually determined as the

oxygen (O2) consumed or carbon dioxide (CO2) produced under these conditions in a given period of time (Table 1.2).

Table 1.2: Resting body temperature (TBody), oxygen consumption and standard metabolic rate (SMR) of seven marsupials, arranged according to body mass (Hume 1999).

Species Mass TBody Oxygen consumption SMR per hour

0.75 (kg) (°C) (mL O2/g) (kJ/kg per day) ingrami 0.007 34.8 2.130 310 Monodelphis brevicaudata 0.076 0.800 211 nasuta 0.667 36.1 0.479 209 Didelphis virginiana 2.403 0.380 238 Macropus eugenii 4.878 36.4 0.283 212 Vombatus ursinusA 28.000 0.202 210 Macropus robustus 30.000 36.0 0.178 210

AGowland (1973).

The SMR values can be converted to energy used (Joules) if the composition of the food being

eaten is known. For carbohydrate one molecule of sugar and 6 molecules of O2 are converted into 6 molecules of water and 6 molecules of CO2. This is a respiratory quotient (RQ) of 1 and the water formed is called metabolic water. For fat the RQ is 0.7 and more metabolic water is formed, while for protein the RQ is 0.8 and less metabolic water is produced. If it is assumed that the animal’s food is a mixture of carbohydrate, fat and protein, with an average RQ of 0.8,

then 1 mL of O2 consumed is equivalent to 21 Joules of energy and 1 mL of CO2 produced is equivalent to 26 Joules. The values for O2 consumed by a range of marsupials of increasing body size are given in Table 1.2. Although the body temperature of all the species is much the same,

the smallest species consumed 12 times more O2 per gram of body tissue than the largest species. Why is the cost of living for the small species greater than for the large species?

Body size in relation to metabolic processes For all animals there is an important relationship between body mass and standard metabolic rate, which is much more pronounced for birds and mammals than for other animals. In its simplest terms the mass of the body increases by the cube power, whereas the surface area increases by the square power, so the small species has a relatively larger surface area than the larger species. Since all metabolic functions occur at surfaces the smaller species has a relatively higher metabolic rate than the larger one. This affects all sorts of physiological functions. For instance, the heart rate of the smallest mammals are about 1000 beats per minute, compared to about 70 beats per minute for humans and fewer than 10 beats per minute for large whales. Again, the strength of muscle and bone depends on the cross-sectional area, so the strengths of these tissues increase by the square power also, so that a comparison between a small and a large mammal shows the small one to be proportionately much stronger than the larger one. For instance, a female , weighing 30 g, can carry a litter of young that weighs more than she does, whereas the female kangaroo ejects her single young from the pouch when it weighs about one-tenth of her own weight. What is a marsupial? 11

Conversely, because the small mammal has a proportionately greater surface area than the larger species, it loses heat and water across its skin and lungs more readily. This greater energy and water loss must be made up from the food and water ingested, so that less is available for synthesis into stored material. Hence, small animals cannot survive starvation for as long as large ones. For example, a small marsupial, such as a dunnart, Sminthopsis, consumes the equivalent of its body weight each day, whereas a person consumes about 1%, so a dunnart cannot survive more than a few days without food, whereas a person can survive for several weeks. For the same reason, one large mammal takes much longer to exhaust its food supply than an equivalent mass of many small mammals. Hence the argument: 10 eat as much as one sheep. There is, thus, in body mass a nice balance of advantages and disadvantages. Under favourable conditions the small mammal converts food more rapidly and the population proliferates faster than the large species. In adverse times, however, the population of a small species will decline rapidly, as its members succumb to the lack of food or the adverse environment, whereas the members of a large species can withstand adversity for much longer. In the later chapters of this book the importance of body size in the economy of different marsupials will recur often.

Comparing mammals of different size In order to compare the performances of animals of different diets, life styles and ancestry it is necessary to agree on mathematical functions that reduce the variability due to body mass. The formula that has generally been adopted is: y = bxk where y is the size-dependent variable (eg O2 consumption, heart rate or food consumption), x is body mass in kilograms, b is the intercept constant and k is the slope (Fig. 1.3). For metabolic rate the exponent 0.75 is still the best approximation for k, despite considerable variability among species. Thus, to compare the SMR of a range of mammals of differing size, the formula used is kJ/kg0.75 per day. Using this formula the average SMR of 56 marsupials from small 7 g dasyurids to large 29 kg kangaroos is 204 kJ/kg0.75per day (Table 1.1). The SMR values for individual species vary from 140 for some desert dasyurids (eg Dasycercus cristicauda, 160 for

Figure 1.3: The relationship between body mass (on a log scale) and daily energy consumption for marsupials and placentals. Closed lines show standard metabolic rate and dashed lines show field metabolic rate. Note that the slope of the field metabolic rate for marsupials is not parallel to that for placentals, the smallest species having a much higher metabolic scope. 12 Life of Marsupials

the koala, cinereus, to 310 for the tiny planigale, Planigale ingrami. This compares with the average SMR for a group of 272 placental mammals, 289 kJ/kg0.75 per day, ranging from the house mouse to the elephant. Thus, the average value for marsupials is about 70% of the average value for placental mammals. In Table 1.1 these values are compared to other terrestrial vertebrates, measured at their thermo-neutral zone and also at 38°C, the body temperature of placental mammals. What is evident from this is that the reptiles, which do not control their body temperature, have a much lower metabolic rate than the mammals and birds under both conditions. More interestingly, each group of mammals and birds has a characteristic level and each is positively correlated with body temperature. The monotremes have the lowest body temperature and the lowest metabolism, while the passerine birds have the highest. The high values of the latter are probably associated with the special requirements of flight. During the past 30 years there has been much discussion about the significance of the apparent difference between marsupials and placentals, some people arguing that life style and diet may be more significant factors in determining SMR than ancestry (see especially Lee and Cockburn 1985, McNab 1986, 1988). Thus, some placentals, such as sloths, Bradypus, have a SMR below the marsupial average and some marsupials, such as planigale and the honey possum have a SMR of, respectively, 106 and 158% of the placental average. Nevertheless, the mean value for marsupials is 30% below the mean for placental mammals, which suggests that there is a basic underlying difference in SMR, although food habits and activity are sometimes strong enough to mask phylogeny. It is not clear why the set point should vary between different kinds of animals, nor what controls it. Hulbert and Else (1999) have observed that the plasma membranes of the mitochondria and cells of vertebrates with high metabolic activity have a relatively high proportion of polyunsaturated lipids, whereas those with lower metabolic activity have membranes that are relatively mono-unsaturated. They have proposed that these differences may be the basis of the pacemaker for metabolism. This suggests that there may be a fundamental difference between the cell membranes of marsupials and placentals, which affects all metabolic processes, including respiration, excretion, water balance, nerve conduction and growth rate. The lower setting of the rate in marsupials is reflected in several other physiological functions: their heart rates, adjusted for body mass, are lower than for placentals, as are nitrogen requirements in some species. The lower SMR of marsupials also means that they have lower food requirements and water turnover rates, which may confer special advantages in adverse conditions or arid environments. Conversely, the advantage for placentals of a higher body temperature and SMR is faster nerve conduction and smooth muscle contraction, faster growth rates and faster reproduction, with the trade off in higher food requirements. In those environments where soils are rich, the climate is equable and food predictable the higher metabolic and reproductive rate of placentals is advantageous. However, in less benign environments, where the soils are infertile and climate unpredictable, the lower metabolism of marsupials may confer an advantage. The spectacular radiations of marsupials in South America and Australia alongside placentals may in part have been due to environmental constraints in these continents that favoured species that conserved limited resources.

Field metabolic rates A major development in the last 20 years has been the measurement of field metabolic rate (FMR) in 28 species of marsupial, using isotope dilution techniques (see Box 1.1). This is a much more informative measure of a species’ actual metabolic needs than SMR and the difference between SMR and a species’ maximum metabolic rate provides a measure of its metabolic scope. While most marsupials have a low SMR, compared to most placentals, the FMRs of species under 100 g What is a marsupial? 13

Box 1.1: Measuring fi eld metabolism To compare the metabolic strategies that different species have evolved we need to understand the various components of metabolism. These are metabolic rate, water turnover, and the nutrients (eg nitrogen) required for an animal to maintain itself in its natural environment and reproduce. Each of these three components can be measured in captive animals in feeding trials, and by the use of respirometers. The use of water labelled with radioactive or other identifi able isotopes of hydrogen and oxygen give more realistic measures from free-living animals. Doubly 3 18 3 2 labelled water ( H2O, H2 O), composed of tritium ( H2) or deuterium ( H2), and a non-radioactive 18 isotope of oxygen ( O), can provide measures of respiration (CO2 production), water turnover, fat deposition and food consumption. If the sodium content of the food is known, food intake can also be estimated in the same way, using an isotope of sodium (22Na). The technique is to capture the animal, inject it with a known quantity of the particular isotope and then let the isotope equilibrate with the animal’s body tissues, a process that usually takes 2 to 6 hours. Then a blood sample is taken to measure the concentration of the isotope distributed in the body. This initial ratio of labelled to unlabelled isotope at equilibration gives a measure of total body content, or pool size, of that element:

Pool size = concentration of isotope injected × sample volume concentration of isotope in sample

The animal is then released and, after some days, when it is recaptured, a second sample of blood is taken and the concentration of the isotope measured again. The difference between the initial and subsequent concentrations is a measure of the dilution that has taken place in the elapsed time by respiration, and by the ingestion and excretion of water and food by the animal. These measures can be converted into fi eld metabolic rate, water turnover rate and food consumption, respectively. For instance, oxygen turnover is measured by the dilution of the 18O isotope compared to the common isotope 16O, using a mass spectrometer. Because each of these rates is substantially affected by the body mass, different species can be compared only if the values are expressed by an allometric exponent. For metabolic rate and food intake the exponent is usually taken to be the three-quarter power of body mass (kg0.75), for water turnover it is kg0.8. However, both of these are approximations based on placental mammals and standard conditions. Green (1997) has shown that there are wide differences between species of marsupial and within a species at different times of the year, and that other exponents may more accurately refl ect reality. However, we will follow convention here when comparing species in each Chapter. In order to estimate the fi eld metabolic rate of an animal we need to measure the volume of

O2 consumed or CO2 produced, and to know the available energy content of the food. The mean energy content of the food varies according to the relative composition of carbohydrate, fat and protein. For example, it is 3–4 kJ/g for and 6 kJ/g for mammalian fl esh. However, the net energy that can be used by the animal is less than this because about 10% is lost in faeces and 8% in urinary excretion. The balance is the net metabolically available energy of the food and, with this information and the SMR (calculated as kJ/kg0.75 per day), the daily food requirements of the animal can be estimated. By comparing these several measures in different species of marsupial in later Chapters we can understand how each species is using the available resources to meet its basic requirements at different periods of its life cycle or in different environments. 14 Life of Marsupials

body mass are much the same, which means that their metabolic scopes are greater (Fig.1.3). For instance, the fat-tailed dunnart, Sminthopsis crassicaudata, has a FMR up to seven times higher than its SMR (Nagy et al 1988). It is unusual for placental mammals to have FMR more than three times SMR, which means that the dunnart’s metabolic scope is considerably greater than equivalent placental species. This may be a definite advantage in extreme environments, since the low SMR conserves resources of the animal at rest, without limiting its metabolism when active. These several attributes of marsupial metabolism will be discussed when considering the adaptations of particular species in later chapters.

Relationships within marsupials If we accept that marsupials are a distinct group of mammals, what are the relationships within the group, and how have they evolved? For 100 years marsupials have been classified on the basis of the number and kinds of teeth and on the number of digits on the feet. These characters are undoubtedly affected by the mode of life of the possessors but have the advantage of being available in fossil remains and can be accurately dated and used to calibrate rates of change in other criteria, based on living animals. Most information comes from the structure of the molar teeth, whereas the structures of the simpler canines and incisors have only been used to determine total dentition. There may also be subtle differences in structure of the enamel and other features of these teeth that have not been disclosed so far. In classifying living marsupials the anatomy of the soft tissues can also be used to distinguish between related species or groups, as can the form and number of chromosomes. In the past 30 years new techniques in biochemistry and molecular biology have greatly extended our understanding of the relationships of modern day marsupials: amino acid analyses and immunological techniques opened up new ways to compare relationships of living species by comparing proteins, and analyses of base sequences in nuclear and mitochondrial DNA has provided an even more powerful means to determine relationships of living marsupials, and to provide increasingly precise measures of the time since related groups diverged in the past. We will begin by considering anatomical characters and then see how relationships based on the newer techniques corroborate or refute the classification of marsupials based on anatomy.

Relationships based on anatomy Teeth Teeth develop in the jaws of young mammals as a coalescence of cells derived from the base of a deep groove of surface epithelium and the underlying tissues in the jaw. The initial tooth bud develops as a central core and an overlying cap of epithelium in which dentine and enamel is later laid down. The tooth bud rises to the surface of the jaw as it develops, eventually piercing the groove and erupting; other tooth buds follow successively behind it in the jaw. In the front teeth the structure remains simple with a single cusp and a single root, but the back teeth develop more cusps and roots. These secondary cusps develop, like the primary one, as subsidiary buds and their position relative to the primary cusp is characteristic of the particular tooth and for the particular species. This pattern of development has been understood for a long time but in the last 10 years the way in which the pattern is controlled genetically has become clearer through the expression of developmental genes in the tooth bud (for review see Thesleff and Sharpe 1997). One group of genes determine that the tooth bud will develop into a complex molar instead of a simple front What is a marsupial? 15

tooth, while other genes determine the position of the secondary buds relative to the primary bud. What is interesting is that the expression of the genes precedes by one or two days the first morphological changes in the cells of the jaw, indicating that the information about pattern is set up before the buds begin to form and not as a result of interaction between adjacent tooth buds in the jaw. Thus, gene expression predicts future cusp pattern and affects the very early stages of tooth development. In evolutionary terms small changes in gene expression could have profound effects on tooth morphology. As an example of what may become possible, Jernvall et al (2000) compared the expres- sion of specific genes in the formation of molar teeth in mouse, Mus musculus, and vole, Microtis rossiameridionalis, which evolved from a common ancestor in the early epoch 20 million years ago. They showed that the big differences in the molar pattern between mice and voles is the result of small changes in the expression of the genes that determine the relative position of the secondary cusp, whether it is parallel to the primary cusp (in mice) or diagonal to the primary cusp (in voles). What is the relevance of this to marsupials? Marsupial teeth develop in the same manner as the teeth of placental mammals, reflecting their common ancestry more than 120 million years ago. It is, therefore, reasonable to assume that the same or similar genes control the pattern of molar cusps in marsupials. More than this, the discovery that the cusp pattern of molar teeth is controlled by the expression of particular genes brings closer the day when the evidence of palaeontology, based largely on the morphology of molar teeth, and the evidence of relation- ships based on molecular genetics, can be integrated more precisely. Teeth in mammals have two primary functions: the front teeth are used to bring food into the mouth, while the back teeth are used to process it for digestion, either by cutting or grinding the food into fine pieces. Teeth may also be used in aggressive or defensive displays, for groom- ing, or for grasping the young. The front teeth are single rooted and comprise a variable number of chisel-shaped incisors and a single set of pointed canines. Behind the canines are up to four larger teeth, each with two roots, called premolars and behind the premolars are up to four sets of molars, each with three or four roots. The structure of teeth reflects the uses to which they are put in different species and they can therefore tell us a considerable amount about the life of the possessor. This is especially useful when examining fossil specimens, which often consist only of a few teeth, because these are the most durable part of the body. The size and shape of the front teeth can tell us whether the mammal catches moving prey and what sort of prey, or whether its diet is largely of plant material. Carnivorous species have many sharp incisors, prominent canines and many sharp points on the premolars and molars. Their molars intersect in such a way as to provide many shear surfaces, like scissor blades, and less emphasis on flat, opposed surfaces where food can be crushed or ground small. In herbivorous species, by contrast, the canines are small or absent, the incisors form two opposing rows of chisel-like teeth with which herbage can be cut, the molars and premolars have fewer sharp points and shear surfaces but have a much larger area for grinding. Sanson (1985) has suggested that the different types of dentition reflect the forces required to penetrate the bodies of prey species. For instance, the impact strength of bone is about 2 kg/cm2 while that of cuticle is 23 kg/cm2, so puncturing the cuticle with sharp points is an easier option than crushing it. The impact strength of plant cell walls, however, is about 76 kg/cm2, which explains why the teeth of have such highly developed crushing surfaces. It also accounts for the different anatomy of the jaws of carnivores and herbivores: carnivores have the jaw muscles grouped at the back of the jaw, which allows for a wide gape but delivers less force than the forward disposition of the jaw muscles of herbivores, such as kangaroos (see Chapter 9). 16 Life of Marsupials

Since the earliest mammals and all the ancestral marsupials and placentals were small insectivores, the dentition of small carnivores resembles more closely the ancestral dentition of mammals from which the several types of herbivorous dentition have evolved. In the study of teeth of mammals the molars provide the greatest amount of information about diet, about relationships between species, and about the evolution of the main lines of descent from the period ancestors. Because of their importance in the subsequent discussion on the evolution of marsupials, some knowledge of the anatomy of molars is necessary.

Diprododont Marsupial Polyprododont Marsupial

Phascolarctos stE stA pa

me prd me Perameles

mcl mcl pad med med Tertiary

Didelphis, Dasyurus

Placental

Didelphodus Alphadon

Upper Cretaceous

Pappotherium stylar shelf Clemensia

Lower prd Cretaceous hyd (Albian) end pad med

Eupantothere me pa Upper pr

Pantothere Upper outer (Rhaetic) posterior anterior

inner Figure 1.4: Probable evolution of molar teeth in marsupials and placentals from pantothere ancestors 250 million years ago. The lower left molar is shown in occlusal view, stippled, and the upper molar is superimposed in outline. In the upper molar the primary cusps are the (pr) protocone, (pa) paracone and (me) metacone, to which were later added posteriorly the (mcl) metaconule and laterally the (st) stylar shelf. In the lower molars the primary cusps are the (prd) protoconid, (pad) paraconid and (med) metaconid, to which were later added posteriorly the (end) entoconid and (hyd) hypoconid. Data from Romer (1966) and Archer (1976). What is a marsupial? 17

The structure of molar teeth can be traced from the earliest mammals, which had triangular, three-cusped teeth in the upper and lower jaws arranged so that, seen in surface view, the apex of each upper tooth (protocone) pointed inwards while the apex of each lower molar (protoconid) pointed outwards (Fig. 1.4). When the mouth closed the upper and lower molars fitted closely between each other providing a zigzag shear surface. This can still be seen in the skulls of small carnivorous marsupials (see Fig. 4.2) and placentals. At an early stage of mammalian evolution, before the separation of marsupials and placentals, the simple arrangement of three cusps was extended. In the lower jaw two additional cusps developed on the posterior face of the molars, the entoconid and hypoconid, thereby providing a basin into which the apex, or protocone, of the upper molar fitted. This provides a grinding surface in addition to the shearing component of the molar teeth. In the upper molars additional cusps developed, both between the primary three cusps and also on the outside of them. The outer series of five small cusps, called the stylar shelf, do not meet complementary parts of the lower molars so that their function in mastication is unclear. Although the earliest placentals had two stylar cusps, later placentals do not, whereas all early marsupials had the full complement of five cusps, most of which are retained in the present day American opossums (Fig. 1.4) and dasyurids (see Fig. 4.2). The presence of the stylar shelf is a diagnostic tool in differentiating between fossil placentals and fossil marsupials. Ridges may form between these several cusps and the pattern of the ridges and the relative sizes of the cusps are used to determine phylogenetic relationships between extinct and living mammals. In addition the fine surface structure of the teeth and surface scratch marks can provide further information about relationships and the type of food that was processed by the living animal. All the living American marsupials and the Australian carnivorous species have long snouts bearing a battery of simple, sharp-pointed teeth. In each jaw there are four molars, three premo- lars and one prominent canine, as well as four or five incisors in the upper jaw and three in the lower jaw (Fig. 1.5). On this criterion they are grouped together as the Polyprotodontia

Figure 1.5: Marsupial relationships based on teeth and feet. Upper panel to show representatives of didactyl and syndactyl feet, the latter differentiated by the small, paired digits 2 and 3 on the pes; also note the large digit 4 in kangaroos; lower panel shows representative skulls of polyprotodont and diprotodont species, differentiated by the presence of 4 or 5 incisor teeth (i) in each jaw in front of the canines (c) in all polyprotodonts and only one in each lower jaw of diprotodonts; all have four molars (m) and up to three premolars (pm). After Jones (1924) and Tyndale-Biscoe (1973). 18 Life of Marsupials

(meaning many front teeth). The bandicoots are included in this group on this criterion but they are more omnivorous than the dasyurids and their molar teeth are squared up with an extra cusp at the back, called the metaconule (Fig. 1.4 and 5.2). The herbivorous species of Australasia have fewer premolars, the canines are small or absent and there are only one to three incisors in the upper jaw and a single large pair of incisors in the lower jaw. With fewer teeth, there is often a gap between the front teeth and the cheek teeth, called the diastema, which enables the herbage to be presented by the tongue to the battery of grinding molars. The single pair of incisors in each lower jaw is the defining character for this diverse group, which is called the (meaning two front teeth). Again, there is a superficial link to one group of South American marsupials, the Caenolestidae, in which the first incisors are large and procumbent, but in them the other incisors are present. Furthermore, in the embryologi- cal development of the teeth of the Diprotodontia, the large incisor is the second, not the first, incisor so this is another case of convergence rather than indicating a close relationship. Within the Diprotodontia further distinctions can be made on the basis of teeth. As in bandi- coots, they have the metaconule in the upper molars, but have lost the stylar shelf, and their lower molars are also squared up by the loss of the paraconid at the front (Fig. 1.4). Also the grinding surface of the molars is more rounded than the sharp pointed cusps of the Polyprotodontia and are variously shaped into crescents or ridges (see Fig. 7.7). This development of a grinding battery of molars reaches its greatest development in two groups of grass eaters: in the wombats the incisors in the upper jaw are reduced to one pair and all the teeth grow continuously through life in a similar manner to the teeth of rodents and rabbits; in the large kangaroos the molar teeth are high crowned, like those of sheep, and the whole battery moves forward in the jaw, the molars being used successively as the more anterior ones are worn down and shed (see Fig. 9.3). Elephants do this too, but on a grander scale.

Foot structure When foot structure is used as the criterion for grouping marsupials, all the American marsupials and the dasyurid marsupials of Australia share a common feature of hind feet with five separate, subequal digits (Fig. 1.5). This is termed didactyly, or separated digits. The remaining Australian marsupials have digit 1 of the hind foot reduced to a nubbin and digits 2 and 3 partly fused and together equal in size to digit 5, and this is termed syndactyly, or fused digits. In some species, especially the large kangaroos, digit 4 is much larger and longer than the other digits (Fig. 1.5) and takes the main thrust during jumping. The semi-fused digits 2 and 3 are used in grooming by some species (see Fig. 9.2e), but whether they evolved primarily for this function or represent a progressive reduction for speedier locomotion, as in the evolution of horses and ruminants, is not clear. A long-standing paradox in understanding marsupial relationships is that bandicoots (Peramelomorphia) have a dentition like the American and Australian carnivorous marsupi- als but have a foot structure that is apparently identical with the Diprotodontia. This paradox is slowly being resolved in favour of bandicoots and the Diprotodontia having acquired fused toes independently. Although this is a very remarkable convergence, it is under a simple genetic control and so could have arisen more than once.

Ankle bones Szalay (1982) proposed that the anatomy of the ankle bones could differentiate between American and Australasian marsupials: in all the Australian species that he examined there was a single articular facet where the ends of the long bones attach to the bones of the ankle joint, whereas in the two main families of American marsupials he found two facets. The only American species What is a marsupial? 19

that had undivided facets like the Australian species was a species that lives in Chile, the monito del monte, Dromiciops gliroides; this alerted biologists to look for closer links between American and Australian marsupials, which has gained strong support from the molecular studies (see Marsupial relationships based on protein analysis). While subsequent work has confirmed a closer relationship of Dromiciops with Australasian marsupials than with American species, the crite- rion itself has not proved to be as consistent as Szalay thought. Hershkovitz (1992) examined a much larger series of species than Szalay and found that both types of ankle joint occur among the American marsupials and among the Australasian species, so it is not an exclusive character.

Brain anatomy Links between the two halves of the forebrain provide another way to distinguish relationships between seemingly similar marsupial groups. Two nerve tracts, or commissures, link the cerebral hemispheres of the forebrain: the large anterior commissure, which also links the two olfactory lobes of the forebrain, and the smaller hippocampal commissure (Johnson 1977). In placen- tals a third commissure, the corpus callosum, links the cerebral cortex of each side, but it is absent from monotremes and marsupials. Within the marsupials a clear distinction can be made between the Diprotodontia and the Polyprotodontia: the former group has an additional tract of fibres, called the fasciculus aberrans, that extends the links of the anterior commissure between the two sides of the cerebral cortex. The absence of this tract in the Caenolestidae supports the conclusion that their diprotodont dentition was independently acquired, and the absence of this tract in the bandicoots also supports the conclusion that syndactyly was independently evolved in them. The brain of Dromiciops has not been described but it would be pertinent to the argument about its relationship to Australian marsupials to know whether it has the fasciculus aberrans.

Sperm morphology and other anatomy The morphology and fine structure of the spermatozoa of marsupials can also disclose relation- ships (Fig.1.6). In the two main families of American marsupials, the Didelphidae and the Caenolestidae, the sperm occur as conjoined pairs. The head of each sperm is intimately adpressed to the head of the other and they remain like this from their passage through the epididymis of the male until they reach the vicinity of the egg in the oviducts of the female. Conjugation of sperm is unknown in any placental mammal, or in any species of marsupial from Australia. Until 1982 this character was thought to separate the American and Australian marsupials from each other and represent a very old divergence. However, Dromiciops has unconjugated sperm, like Australian marsupials. This was further support for closer links between the two geographical groups. Within each family the sperm head has a characteristic shape, which distinguishes the three families of American marsupials from each other, and within the Australian marsupials, distin- guishes between the , the and the (Fig. 1.6). The wombats (Vombatidae) and the koala have similarly hook-shaped sperm heads distinct from any other marsupial, which supports their close relationship to each other and differences from possums. And the minute marsupial, Tarsipes, has the largest sperm of any mammal so far described. It is 360 µm in length or 4.5 times longer than a human sperm. As mentioned earlier, the head of the penis is more or less divided into two lobes in all species of South American marsupials and most Australian species. It is only partly divided in the koala and wombats, while in the Macropodidae it is a single structure and in Tarsipes there is no glans at all. Among some dasyurids the anatomy of the penis is a useful character for classification (Woolley 1982). 20 Life of Marsupials

Figure 1.6: Marsupial relationships based on head and midpiece of spermatozoa. Note three forms of pairing at the head in (a) didelphid, (b) caenolestid, (c) caluromid marsupials of America and (d) unpaired sperm of the American Dromiciops and Australian phalangerid, petaurid, burramyid and macropodid species. Other family specific forms of the sperm head are shown for (e) peramelids, (f) dasyurids, (g) honey possum, Tarsipes and (h) koala, Phascolarctos cinereus, and wombats, . From Temple-Smith (1994) Reproduction, Fertility and Development 6, 423, with permission of CSIRO Publishing.

On the anatomical characteristics of the teeth, feet, brain, sperm and penis shape the 16 families of living marsupials separate into five groups, one exclusively in South America, three exclusively in Australasia and one with representatives in both regions (Table 1.3). These groups will now be used as the basis for comparing the other characters of cytology, proteins and genes.

Relationships based on chromosomes The double-stranded molecules of DNA, which carry the genetic information of the individual animal in all its cells, is packaged in a variable number of pairs of chromosomes, characteristic for each species. At cell division each chromosome divides along its length by a separation of the two strands of the DNA molecule and one strand goes to each daughter cell where it synthesises the complementary strand to reconstitute the double-stranded molecule. The separation of the strands begins from a body called the centromere, which first divides and then draws the dividing DNA to opposite poles of the cell. The centromere may be at one end of the chromosome, in which case the chromosome has one arm and is called acrocentric, or it is at some point along the length of the chromosome, in which case the chromosome has two arms and is called metacentric. In placental mammals the commonest number of chromosomes is 48, that is, 23 pairs of autosomes and two sex-determining chromosomes, but between species the number ranges from 6 to 92. Marsupials have about the same amount of DNA as placentals but it is packaged in fewer, larger chromosomes: the number of chromosomes ranges from 10 in the , What is a marsupial? 21

Table 1.3: Distribution of anatomical characters among American and Australasian marsupial families

Chapter Family RegionA Sperm Digits Incisor teeth Brain Glans 2 and 3 commissure penis on hind feet 3 Didelphidae Am paired separate 10/8 single divided 3 Caenolestidae S Am 3 Microbiotheriidae S Am single 4 Dasyuridae ANG 4 ANG 4 Myrmecobiidae Aust 4 Aust 5 Peramelidae ANG united 6 Tarsipedidae Aust 6/2 double 6 ANG 6 Burramyidae ANG 7 Phalangeridae ANG 7 ANG 7 Aust 8 Vombatidae Aust 9 Macropodidae ANG single AAm, North and South America; S Am, South America; ANG, Australia and New Guinea; Aust, Australia.

Wallabia bicolor, to 32 in the rufous , Aepyprymnus rufescens, with 111 species having 14 chromosomes and 37 species having 22 (Table 1.4). In both America and in Australasia there are marsupials with 14 and with 22 chromosomes, so it was for long uncertain which was the ancestral number. However, with techniques that can identify parts of the chromosomes, it is now clear that the chromosomes of all marsupials can be derived from an ancestral number of 14, composed of six pairs of autosomes and the two sex chromosomes. Using specific Giemsa dyes that disclose a pattern along the chromosome, called G-banding, rather like a bar code, it is possible to recognise portions of chromosomes of different species that have the same pattern (Rofe 1978). However, a much more precise technique, called chromosome painting (Fig. 1.7, Plate 1), was developed in 1995, which is transforming understanding of the relationships between parts of chromosomes in different species. In this technique individual chromosomes of one species are isolated and the DNA amplified by a process called polymerase chain reaction and then labelled with a fluorescent dye of a particular colour. This labelled DNA is then mixed with chromosomes from the other species to be compared. It attaches to the complementary DNA in them, and can be visualised by fluorescence microscopy. Only those parts of the chromosomes that are the same in both species will be coloured. By labelling different chromosomes with different coloured dyes it is possible to show that the chromosomes in one species are composed of parts of several chromosomes from the other. This beautiful technique now makes it possible to resolve the paradoxes of closely related species having very different numbers of chromosomes. For marsupials (De Leo et al 1999, Rens et al 2003) it is also enabling the relationships of the different orders and families to be better understood (Fig. 1.8, Plate 1), even those between Australasian and American marsupials (Rens et al 2001). 22 Life of Marsupials

Table 1.4: Distribution by family of chromosome numbers (karyotype) among 211 species of living marsupials Hayman (1990), with 31 additions since 1990.

Family 10/11 12/13 14 15/16 16 18 18/19 20 22 24 32 Total Didelphidae 22 6 10 38 Microbiotheriidae 1 1 Caenolestidae 5 5 Thylacinidae 1 1 Dasyuridae 55 55 Myrmecobiidae 1 1 Notoryctidae 1 1 Peramelidae 8 8 Thylacomyidae 1 1 Vombatidae 2 2 Phascolarctidae 1 1 Burramyidae 5 5 2 2 Phalangeridae 5 3 8 Petauridae 1 1 4 3 8 5 22 Tarsipedidae 1 1 1 1 Potoroinae 2 4 1 1 8 1 2 4 1 12 5 7 17 49 1 1 Total 2 5 111 1 17 14 1 19 37 3 1 211

By G-banding and chromosome painting the 14 two-arm chromosomes of South American species are seen to be very similar to the 14 two-arm chromosomes of Australian species, with differences being due to inversions and translocations of bits within individual chromosomes. In species that have more than 14 chromosomes the individual chromosomes can be identified as equivalent to separate parts of these 14. Thus in the koala, with 16 chromosomes, one two-arm chromosome has divided to produce two one-arm chromosomes. Similarly, the 18 chromosomes of the South American species of Monodelphis can be derived by the division of four of the two-arm chromosomes of the related Marmosa species group. In the Australasian family, Phalangeridae, the cuscuses have 14, while the related Australian brushtail possums (Trichosurus) have 20 chromosomes, the latter being derived by division of eight two-arm chromosomes followed by fusion of 4 one-armed chromosomes (Rens et al 2003). The Petauridae and Macropodidae show the greatest variation in chromosome number within each family. For the latter family, this variation has been interpreted as due to several fissions of the ancestral 14 chromosomes to give 22, and then later fusions of some one-armed chromosomes, in several different arrangements, to give the range of numbers from 10 to 32 (Glas et al 1999, Rens et al 2003). This variability reaches its most extreme expression among the rock (see Chapter 9). It is not clear what the selective advantage of arranging the DNA in different numbers of chromosomes is, although there is some evidence that the upper size limit for a single chromo- some is half the length of the spindle formed at each mitotic cell division (Schubert and Oud 1997). Nor is there any clear reason for the bimodal distribution of chromosome number in marsupials from what is now agreed to be the ancestral number of 14, to 22, a shift that has What is a marsupial? 23

occurred in at least four families of marsupials independently. This illustrates the point, made earlier in regard to teeth and feet, that genetic changes may occur independently and yet appear to be superficially similar (Rens et al 2003). It may be that there is some fundamental advantage in arranging the DNA into 14 or 22 chromosomes, rather than into any other combination.

Sex chromosomes The genes that determine the sexual differentiation of an individual mammal are usually carried on one pair of chromosomes, known as the sex chromosomes. The X chromosome is usually of normal size and contains many genes that are not directly concerned with sexual differentia- tion, whereas the Y chromosome is usually very small and contains few other genes than those concerned with sex determination. Most marsupials have small X chromosomes, representing less than 3% of the total DNA, and minute Y chromosomes, which do not even pair up with the corresponding X. By contrast in placental mammals the X chromosome represents more than 5% of the total DNA and the Y chromosome has some common genes with the X and does pair with it at meiosis. Jennifer Graves (1996) thinks the marsupial pattern is the original or primitive one and that at some early stage of evolution an additional part of an autosome became attached to the original placental X and Y chromosome. A similar fusion of an autosomal chromosome to the original X chromosome has occurred independently in four species of marsupials, which have large X chromosomes. For instance in the long-nosed , Potorous tridactylus, females have 10 autosomes and two large X-chromosomes, while males have 10 autosomes, one large X chromosome, one large Y and one small Y chromosome. The large Y chromosome is actually the other half of the autosome that has fused with the original X and at meiosis it is paired with one arm of the large X and the small Y with the other arm. A similar thing must have occurred independently in the swamp wallaby, which has 8 autosomes, a large X chromosome and one large and one small Y chromosome. Chromosome painting with antibodies to tammar chromosomes has shown how this came about. The long arm of the X and the whole of the large Y chromosome are homologous with chromosomes 2 and 7 of the tammar and only the short arm of the X is homologous with the tammar X chromosome (Fig. 1.7, Plate 1) (Toder et al 1997). These unusual arrangements appear to have no significance in the life of the species concerned, since closely related species have the normal XX/XY sex chromosome arrangement.

Dosage compensation in female mammals Because female mammals have two X chromosomes, one from their father and one from their mother, they have a double dose of genes that reside on the X chromosome, compared to males, which only have one X chromosome from their mother. In both placentals and marsupials one of the X chromosomes in females is inactive and replicates later than the other one. In placen- tals the X chromosome that is inactive in any cell may be the one that came from the father or from the mother, so it is called random X inactivation. In marsupials, by contrast, it is almost invariably the X chromosome that came from the father that is inactivated, and this is termed paternal X inactivation. It has been most thoroughly studied in kangaroos but there is evidence for similar processes in other families. In the bandicoots most tissues of the body possess only one X chromosome, the second X in females and the Y chromosome in males being lost during development. Only the tissues of the gonads retain the full sex chromosome complement (see Hayman 1990), a condition termed sex chromosome mosaicism. It is clear from the above that chromosome number and arrangement reflects the relation- ships between marsupials derived from other criteria, but there have been many changes in the order of the chromosomes and their number that make it hard to interpret. As a primary crite- rion for understanding the relationships of marsupials, chromosome number is not satisfactory, 24 Life of Marsupials

although it can add secondary support to more comprehensive criteria, such as protein and DNA sequences.

Marsupial relationships based on protein analysis With the development in the 1960s of amino acid analysers that could determine the structure of large protein molecules, it became possible to use this as a criterion for examining relation- ships between mammals. The first proteins to be examined were haemoglobins, albumins and some enzymes. These proteins were used to estimate the time of separation of placentals and marsupials on the basis of the number of differences in the amino acids in the comparable molecules. It was assumed that changes in the sequences and composition had changed over time by random mutation and that the rate had been constant. Neither of these assumptions could be sustained adequately: any change in the amino acid composition of the protein would be likely to affect its function and hence not be random, and the rate was determined by the known time of separation determined by fossils and so was a circular argument. The approach with albumins and enzymes of intermediary metabolism were more successful at determining relationships within closely related marsupials and between families but suffered from the same limitation that change was unlikely to be random. The most important development was John Kirsch’s (1977) comparisons between different marsupials of whole serum. With its great number of individual proteins, serology redefined the classification of marsupials for 20 years. Kirsch exploited the then new techniques of immunol- ogy to raise antibodies to the serum of one species in a and challenge other related species with the antiserum from the rabbit. If the challenged species were closely related to the original species, they would react to many antibodies in the rabbit serum; if more distantly related, they would react to fewer. By making numerous such challenges, using many representative species, Kirsch was able to build up a picture of relationships of all the main families of marsupials. His results led to a revision of the earlier classification in several ways. First, the Polyprotodontia are not a single group, the South American didelphids being distinct from the caenolestids and from the Australasian dasyurids and peramelids. Second, the two groups with fused toes, the peramelids and Diprotodontia, were not closely related. But the most surprising result was that Dromiciops of southern Chile appeared to be more closely related to the Australian dasyurid and diprotodont marsupials than to the other South American families. In subsequent years the spermatozoa of Dromiciops were found to be unconjugated, unlike those of all other South American marsupials, and the cytology (Sharman 1982) and anatomy of the ankle joints (Szalay 1982) all supported a closer relationship of this species to Australian marsu- pials than to South American species. Serology gave no support to a close relationship between ringtail possums (Pseudocheiridae) and the koala, despite their similar type of molar teeth, but rather confirmed a close relationship between koala and wombats, previously suggested from penis and sperm morphology. Why is serology a more powerful criterion for determining relationships than anatomy? One reason is because it deals with interactions between many different proteins, each of which may act as a foreign antigen, so that many different genes are involved. Conversely, using individual anatomical characters, fewer genes are involved and it is difficult to distinguish characters held in common from characters acquired independently by parallel evolution. For instance, the failure of adjacent toes to separate (syndactyly) may be controlled by a single gene. The same holds for comparisons between the amino acid sequences of individual proteins, such as albumin, haemo- globin or myoglobin, which have all been used as criteria for determining relationships. What is a marsupial? 25

Marsupial relationships based on DNA The most powerful technique available for determining relationships is the comparison of whole genomes of related species. This can be achieved in two ways. The sequence of base pairs along part of the genome can be examined by comparing the sequence with that from other species. Most of the analyses have been done using four mitochondrial genes (12S rRNA, valine tRNA, 16S rRNA and cytochrome b DNA) and two nuclear genes (exon and intron of the protamine P1 gene and the exon of inter photo receptor retinoid binding protein or IRBP). These provide information about short lengths of the total genome but, unless many sequences are used together, they suffer from the same shortcomings as amino acid sequences in proteins. The other technique, called DNA/DNA hybridisation compares differences in the entire genome, including the so-called ‘nonsense,’ or non-coding DNA, which makes up 95% of the genome. Both sequence analyses and DNA/DNA hybridisation are based on the assumption that small changes in the DNA occur at a fairly constant rate over time, so the greater the difference in the DNA between two species the longer the time since they had a common ancestor. By analys- ing several different types of DNA, including mitochondrial DNA, with both techniques, the separate values provide internal checks on the accuracy of the estimates. In addition, independ- ent checks can be made when the time of separation of two related species is known from the fossil record. However, a much more cogent argument for the accuracy of this technique is that most of the DNA in the genome of all mammals does not encode for specific proteins and its function is unknown. There are three kinds of DNA in the nucleus: unique or single copy DNA, which codes for proteins and makes up 10–20% of the genome; medium repeat DNA, comprising one thousand to one hundred thousand copies, making up another 20%; highly repeated DNA with more than one million copies, which until recently was called nonsense, redundant or junk DNA because it does not code for proteins. One view is that the latter DNA may be the residue of viruses that infected the species in the distant past, were inactivated in the nucleus but were then trapped in the genome. If highly repeated DNA does not code for protein or have an important function in the life of the mammal, it is unlikely that natural selection will act against changes that occur in it. Hence, mutations in repeat DNA can be assumed to occur at random and to increase with time, so that differences between compared species are more likely to represent real differences in the time since they shared a common ancestor: such differences are time dependent. Time dependency has been tested by comparing the distances between closely related species to another that is much more distantly related. For instance, species representing several differ- ent families of marsupial can be compared with a placental species. If rates of mutation since the separation of placentals and marsupials have been constant, two things follow: the genetic distances between each of the marsupials and the placental should be approximately the same; and in every case the distance between any two marsupials should be less than the distance between them and the placental. In several independent trials using DNA/DNA hybridisation this has been found to hold. As Sibley and Ahlquist (1986) said:

DNA and the morphological characters traditionally used to reconstruct phylogeny serve to provide different kinds of information. Morphology shows how natural selection has modifi ed structure to adapt organisms to the environments, whereas DNA comparisons give a direct indication of the branching pattern and the approximate branching dates among living lineages. Morphology is functional; the DNA clock keeps time. 26 Life of Marsupials

Box 1.2: DNA/DNA hybridisation Each strand of the deoxyribonucleic acid or DNA molecule is made up of a sequence of four kinds of nucleotides, each composed of a fi ve-carbon sugar (ribose) and phosphate group and one of four nucleic acids. These nucleic acids are adenine (A), thymine (T), cytosine (C) and guanine (G), which form two complementary pairs A–T and G–C, held together by hydrogen bonds. There are about three billion nucleotides in the nucleus of a mammal and a triplet of nucleotides codes for an amino acid, so the order of the nucleotides on the strand determines the order in which amino acids line up to form protein chains. Because the nucleotides are complementary, each strand of the DNA is the complement of the other and, when the cell divides, each divided strand synthesises the complementary strand in the new cell. The sugar–phosphate bonds that maintain each strand are very strong but the hydrogen bonds between each nucleotide pair, which keep the two strands together, are weak. When DNA is isolated from a cell and heated to boiling point the weak hydrogen bonds are broken and the two strands separate, without affecting the sugar–phosphate bonds of the separated strands themselves. As the strands cool they may collide with other single strands and, if the nucleotides match each other, the strands will combine. If the separate strands come from the same species, the recombination will be strong and will occur at a high temperature. However, if separated strands from two related species are mixed, not all the nucleotides will encounter a match and the combination will be weaker (Fig.1.9, Plate 2). Being weaker they will only stay together at lower temperatures. This technique of introducing single strands of DNA from different species and testing their ability to combine is called DNA/ DNA hybridisation and allows the closeness of a relationship to be measured directly by the temperature at which the strands hold together. An important assumption is that changes in the nucleotide sequence along the DNA strands occur at a fairly uniform rate. Hence, the longer two species have been separated in time, the greater the number of differences between their nucleotide sequences and the lower the temperature at which they will associate. This technique was fi rst developed in the 1970s and has been used with increasing precision to explore the relationships of living marsupials and their origins. It does two things: it provides values for the closeness of relationships between living species when compared one with another and, by iteration of many individual comparisons, it can allow family trees to be constructed; it also indicates how long ago the separation between each group occurred. These two functions have transformed our understanding of marsupial systematics and phylogeny.

All the families of living marsupials, represented by more than 100 species, have been analysed by DNA/DNA hybridisation and the results compared to other analyses using DNA sequencing data from less complete series (Kirsch et al 1997, Springer et al 1997). These results enlarge and to a substantial degree confirm the earlier results from serology. The major division is not between South American and Australasian marsupials, as previously supposed, but between the Didelphidae, Caenolestidae and Peramelidae on the one hand and on the other the Dasyuridae, with Notoryctidae, the Microbiotheriidae (Dromiciops) and the large multifamily Diprotodontia, comprising most of the Australasian species (as summarised in Fig. 1.10). Using this outline as the best representation of marsupial relationships, it is apparent that many of the anatomical features that have previously been used to classify marsupials have arisen more than once and, therefore, do not represent special affinities (Fig. 1.11). Thus, species with 14 chromosomes are found in every major group, procumbent lower incisors are found in the caenolestids of South America and in all the diprotontids of Australia; syndactyly was almost certainly acquired independently in the peramelids and the diprotodontids, a conclusion supported by brain What is a marsupial? 27

commissures and the markedly different mode of reproduction in the two groups; conjugated sperm occur in only two families of South American marsupials. Other features can now be seen to have been independently acquired: the burrowing habit has arisen independently in the bilby ( lagotis), the marsupial mole (Notoryctes typhlops), the wombats, the rufous hare wallaby ( hirsutus), and the burrowing bettong (Bettongia lesueur); arboreal leaf eating has also evolved independently in the koala, in ringtail possums and the greater glider (Petauroides volans), and in brushtail possums and cuscuses (Phalanger); gliding membranes have evolved independently at least three times; and the pouch may have arisen independently several times also. We will return to the details of this classification as we deal with each group of families in the later chapters.

Marsupial distribution in space and time The earliest mammals appeared about 220 million years ago, before the age of dinosaurs, but they remained a small and insignificant part of the world fauna for the next 100 million years. During the early part of this long span of time the continents of the world were contiguous, so that the different kinds of mammals could spread to all of them. In South America fossils of two extinct groups of early mammals, the dryolestids and symmetrodonts, have been found, and in Australia two monotremes, Steropodon and Kollikodon, have been described from the opal fields of Lightning Ridge, NSW, dated at 110 million years ago (Archer et al 1985, Flannery et al 1995) and another mammal, Ausktribosphenus has been described from Flat Rocks, , dated to 115 million years ago (Rich et al 1997). The latter was first thought to be a placental mammal but is now considered by Kielan-Jaworowska et al (1998) to be either a symmetrodont or a multituberculate, both of which groups had a long separate evolution from the placentals and marsupials. Ausktribosphenus has left no living representatives in Australia or elsewhere but the monotremes have persisted to the present day and at least one species was present in South America in the early Tertiary period, 60 million years ago. Beginning in the Jurassic period (200–140 million years ago) the landmass of the world split into two super continents: Laurasia comprised Europe, Asia and North America; Gondwana comprised South America, Africa and Madagascar, India, Australia and Antarctica. Since the common ancestor of marsupials and placentals arose during the early Cretaceous in the northern hemisphere, after South America, Antarctica and Australia were isolated from Africa and the northern continents, the first half of their evolutionary history occurred in Asia, Europe and North America. The precise time when the two groups of mammals separated is still conjectural because the critical anatomical features that distinguish them leave no direct evidence in the hard parts that become fossils. Now that protein and DNA criteria can be used with increasing precision to determine the time that living mammals have had a separate history, the time of separation is being narrowed down to between 120 and 100 million years ago. Luo et al (2003) described the whole skeleton and skull and even impressions of fur of a small mammal that lived in China 120 million years ago. They consider that this is the earliest known marsupial on the basis of several features of its dentition, the cusp pattern of its molar teeth, the structure of its limbs and feet, and the possession of epipubic bones, found only in marsupials and monotremes. Equally interesting is the discovery in the same fossil beds of another small mammal that they consider to be a placental mammal. The next oldest fossils of both kinds of mammal occur in Asia and in North America at about 110 million years ago and then with increasing abundance up to the end of the Cretaceous at 65 million years ago. 28 Life of Marsupials

Figure 1.10: A family tree of all living marsupial families, based on DNA/DNA hybridisation and DNA sequence data. The column on the left refers to the genera of marsupials used to construct the tree and that on the right the families to which they belong. The length of the horizontal connecting lines represent the degree of relatedness and probable time since a common ancestor. Time scale at bottom is only approximate. After Kirsch et al (1997) and Springer et al (1997). What is a marsupial? 29

Figure 1.11: How the molecular family tree can be used to test other criteria, such as chromosome number, diprotodont incisors, syndactyl toes and sperm morphology between American and Australian marsupials. After Springer et al (1997). 30 Life of Marsupials

In Asia the fossils are predominantly of placental mammals and in North America, predominantly marsupial (Cifelli and Davis 2003, Kielan-Jaworowska et al 2004): there is no evidence of fossil marsupials or placental mammals in the three southern continents before the beginning of the Tertiary. Since North and South America were separated in the Cretaceous, placentals and marsupials must have reached South America across an island chain but being a minor element of the fauna left few fossils. Then the world changed.

Marsupial history through the Tertiary Sixty-five million years ago an asteroid of about 10 km diameter struck the earth at a point that is today the Yucatan Peninsula of Mexico (Alvarez 1998). This resulted in the of 70% of all living things in the sea and on the land. It included the demise of the large dinosaurs that had been the dominant vertebrates for 100 million years. In addition, the impact area must have closed off any connection between North and South America, so that the surviving southern fauna became isolated from the rest of the world except Antarctica and Australia. In the aftermath of the asteroid, some small mammals survived and prospered: in North America small marsupials diversified into several kinds during the early Tertiary but then died out and there was no later connection to the marsupials that evolved in South America. Thus, the great southern continent, extending from the equator to the pole and round the other side, was swept clear of most of its fauna from the Cretaceous and isolated from further input until the end of the Tertiary (Fig.1.12). The main marsupial story now takes place on the three southern continents and is intimately associated with the movements of the continents through this time.

Figure 1.12: Disposition of South America, Antarctica and Australasia at the close of the Cretaceous, showing the site of the asteroid impact in the Yucutan peninsula that isolated the continents from North America. Sites V in all three continents where fossil marsupials of the early Tertiary period (65–25 million years ago) have been found. The lower figure shows the disposition of the continents at about 25 million years ago, after their separation and when the circumpolar current had begun to flow. After Archer et al (1993) and Woodburne and Case (1996).

Continental movements through the Tertiary During the past 30 years the precise positions of all the continents and larger islands have been plotted for the whole of the Tertiary, so that their relative movements and times of separation are now understood in detail. The crucial evidence was palaeomagnetism: igneous rocks that What is a marsupial? 31

contain iron particles retain the magnetic field that prevailed on the earth when the rock was molten. Because the polarity of the Earth has switched from one direction to the other on many occasions through geological time, it is possible to estimate when a particular rock was molten. This technique disclosed that the youngest rocks occur in the mid-ocean ridges with progressively older rocks of complementary ages outwards on either side of the ridge. This observation can be explained only by spreading of the sea floor with new rock emerging in the mid-ocean ridge and the continents on each side moving apart. The exciting thing is how the findings of palaeontology and the position of the continents through the Tertiary complement the conclusions, based on the findings from DNA analyses, for the times of divergence of the various groups of marsupials in South America and Australasia. At the close of the Cretaceous, 65 million years ago, Gondwana had already begun to break up. Africa and India had separated from South America and Australia, respectively, but South America and Antarctica were still united at what is now the Antarctic (formerly Palmer) Peninsula, and the southern coast of Australia was contiguous with eastern Antarctica, so that the three continents still formed one super continent (Fig. 1.12). No icesheet covered Antarctica and this vast continent in high latitudes was probably rather like Siberia is today, except that it was much warmer and was heavily forested. For the next 20 to 25 million years mammals could have moved each way between South America and Australia via Antarctica. However, since its position over the South Pole has not changed substantially, any species of mammal or bird would have had to survive long polar nights on that continent. As the Australian plate moved north, shallow seas in the west gradually separated it from eastern Antarctica, the last connec- tion being south of . When this submerged, Australia became wholly separated from Antarctica between 45 and 38 million years ago, closing off further opportunity for migration of land animals from South America and Antarctica. The land connection between South America and the Antarctic Peninsula of Western Antarctica remained for another 15 million years, to the end of the epoch (23 million years ago). After the final separation and the opening of Drake Passage the circumpolar current began to run as it does to this day (Fig. 1.12), bringing profound changes to the climate of Australia and Antarctica. Antarctica became progressively colder and developed a thick ice sheet that obliterated almost all life, and Australia became drier. Since the start of the Tertiary the Australian plate has been moving northwards at a fairly constant rate of 1 degree of latitude every 2.1 million years, or 50 mm per year, and this also contributed to profound changes in its climate and vegetation. While the south- ern half of New Guinea was a part of the Australian plate and moved north into the tropics ahead of it, for all of the early and mid Tertiary it was either submerged or separated by shallow seas from Australia. It is only in the last 5 million years that it has become a prominent landmass with high mountains and an intermittent connection with Australia (see Fig. 10.2). The only fossil marsupi- als from New Guinea are of this time: nothing earlier has been discovered (Flannery 1995).

Climate changes through the Tertiary Estimates of the sea temperature through the Tertiary suggest that the climate of the early super continent in the south was initially very warm (20oC) but cooled to 1oC during the Oligocene, after Australia had separated from Antarctica (Kemp 1981, Galloway and Kemp 1981). A polar ice cap began to form in the late Eocene epoch (Fig. 1.13) and increased through the Oligocene, when sea ice first occurred. Sea temperatures rose again to 10oC in the early Miocene, 20 million years ago, which was associated with warm, moist climates in Australia. Then, through the second half of the Miocene the sea temperatures were again low, a thick ice sheet covered Antarctica, and much of Australia was dry. There was a return to warm moist climates in the , 5–2 million years ago, followed by cold, dry conditions in the , which have persisted to the present time. 32 Life of Marsupials

South America in the Early Tertiary era After its isolation there followed a large radiation of placental mammals in South America, which included the ancestors of today’s sloths and armadillos (Edentata) and a wealth of other groups now extinct, such as horse-like litopterns, tapir- and camel-like forms and rhinoceros- like notoungulates (Patterson and Pascual 1972). Most of these placental mammals died out when connection with North America was resumed in the late Pliocene, 3 million years ago, being displaced by the more familiar mammals of the northern continents, such as llamas, bears and large . The only surviving descendants of the old Tertiary radiation are the sloths and armadillos. The marsupials in South America fared little better, with the large marsupial carnivores becoming extinct at the same time as the old placental mammals. The three surviving lines can be traced to the Early Tertiary (Fig. 1.13): the omnivorous Didelphidae, which gave rise to the large borhyaenid top carnivores (eg Thylacosmilus); the Caenolestidae, now represented by a few small insectivorous species in the high Andes and southern Chile; and the Microbiotheriidae, represented today by a single species, Dromiciops gliroides, in southern Chile (see Chapter 3).

Australasia in the Early Tertiary Vegetation change At the beginning of the Tertiary, the vegetation across the great southern continent was composed of the ancient conifer genera Araucaria, Podocarpus, Phyllocladus and Dacridium, as well as the cool temperate species of flowering plants, such as Nothofagus, Casuarina and the families Myrtaceae, Proteaceae and Banksiae (Martin 1981). All of these persist to the present day in Chile, New Zealand, Tasmania, eastern Australia and New Guinea. Other early elements of the Australian flora that first appear in the Eocene were members of the families Loranthaceae (mistletoe), Cupressaceae and Graminae (grasses). Closed forests covered most of Australia through this immense span of time, while the climate was warm and moist. With the onset of the great arid period through the Oligocene, the evolution of hard-leafed (xerophilous) species began, possibly in response to the leached soils and arid conditions. This is when the genera Eucalyptus, Acacia, Melaleuca and Eremophila first appear. In the Miocene these species largely replaced the Nothofagus and other closed forests. The eucalypts were astonishingly successful in the Miocene, especially in the western, drier half of the continent. Grasses and species of Compositae also become predominant in mid Miocene and remained so thereafter.

Mammal history of Australasia Until the 1990s, the mammal history of the first half of the Tertiary of Australia was unknown, the earliest fossil mammals being from the Miocene (about 25 million years ago) and recognis- ably modern (Fig. 1.13). What had occurred in the first 40 million years of the Tertiary? Now, at a site in southern , near Murgon, the fossil record reaches back to 55 million years ago (Godthelp et al 1992). There is some dispute about the precise age of the Murgon site, but volcanic rock overlaying it is 40 million years old, so the site must be older than this. At Murgon there are fossil marsupials, representing Australian bandicoots and dasyurids, as well as other fossils that resemble contemporaneous fossil didelphids from Peru (Djiarthia murgon- ensis, Godthelp et al 1999; Thylacotinga bartholomaii, Archer et al 1993) and a microbiotheriid (Chulpasia, M. Archer pers. comm. 2001). The latter discovery is of particular interest because the living microbiotheriid, Dromiciops in Chile, is more closely related to Australian possums and kangaroos than to either of the other South American families (Springer et al 1994, 1997). Conversely, the Australian bandicoots appear to have diverged from other Australian marsupials 60 million years ago (Fig. 1.10) and are more closely related to the living caenolestids of South America than to any Australian family (Palma and Spotorno 1999). These astonishing results What is a marsupial? 33

Figure 1.13: A summary of marsupial history in Asia, Africa, Europe and North America in the Cretaceous and Tertiary periods, and in South America, Antarctica and Australia in the Tertiary and Quarternary. The dotted lines indicate approximate times of migration between continents. The living families are listed in the same order as in Figure 1.10. After Archer et al (1999) and Marshall et al (1990). 34 Life of Marsupials

have led to the idea that there must have been considerable exchange both ways between South America and Australia while the connection through Antarctica persisted (Kirsch et al 1991). If there was such an exchange during the first half of the Tertiary, why did none of the placen- tal mammals of South America reach Australia at the same time as the marsupials and undergo adaptive radiation in Australia, as they did in South America? Maybe some did reach Australia but failed to survive: two lower molar teeth and a petrosal (the bone in the skull that encloses the inner ear) discovered at Murgon have been provisionally identified as placental (Tingamara porterorum, Godthelp et al 1992), and palaeontologists from South America consider the teeth to resemble those of the early Tertiary condylarths, which gave rise to the large herbivores in South America. If these claims of relationship are sustained and the age of 55 million years for the Murgon site is confirmed, this is the best evidence yet for a direct link between the earliest South American fauna and the earliest Australian marsupial fauna. It allows for the possibility that for the first 15 million years of the Tertiary there was interchange of mammals across the super continent of South America–Antarctica–Australasia, with a number of separate lineages of marsupials evolv- ing on one or the other part. This could account for the closer relationships of Dromiciops to the Australasian families than to the other living South American families, and for the apparently closer relationship of the Peramelidae to South American Caenolestidae than to any Australian family.

Miocene fossil history in Australasia After Murgon nothing is known of mammals in Australia for another 30 million years. What took place in that long span of time remains a mystery but we can infer that much did because, when the Miocene opened about 25 million years ago, an incredible variety of marsupials appear in several sites across Australia. By far the richest of these sites is at Riversleigh, northern Queensland. Here are found representatives of almost all the present day families of marsupials, some very similar to existing species, as well as a host of other kinds that have no living descend- ants. In the earliest deposits at Riversleigh the climate and vegetation of the site was still rainfor- est but as time passed the site became progressively drier and the Nothofagus-type forests were replaced with Eucalyptus dominated forest. The new animals reflect these changes: apart from a variety of carnivorous and insectivorous species, similar to the Murgon fossils, there are now present herbivores of various kinds, which were not present 30 million years before. There are koala-like animals adapted to feeding on Eucalyptus, the ancestors of ringtail possums, striped possums and -like possums, as well as the ancestors of rat kangaroos, the forerunners of the grass eaters of the late Miocene. Whether these various groups evolved in response to the new food resource of grasses, eucalypts and other nectar rich plant species cannot be resolved until more is known about the 30 million year period after Murgon. No longer present are any placental mammals, except bats, which could have reached Australia from the north. The brief appearance at Murgon of a possible placental has not persisted. Was Tingamara really a placen- tal, which left no descendants, or was it an aberrant marsupial? If it were indeed a placental, this raises the interesting question of why placental mammals failed to become established in Australia when they flourished through the Tertiary in South America.

Darwin’s ideas and alternatives Darwin (1859) and others before and since him have assumed that the adaptive radiation of marsupials occurred in Australia because they were protected from competition with placen- tals by the isolation of Australia through the early Tertiary. If Tingamara is indeed a placental mammal, this argument loses much of its force and we must ask why the placentals did not survive in Australia after its separation from Antarctica, while marsupials did? One possibil- ity is that marsupials prevailed in Australia because they were pre-adapted to the Australian What is a marsupial? 35

environment that developed after the separation from Antarctica about 45–38 million years ago (Woodburne and Case 1996). These climatic changes could have presented a critical challenge to the mammals isolated on the Australian continent, which were still all very small creatures. As already discussed, two features distinguish living marsupials from living placentals and may have pre-adapted them for survival in Australia: their lower metabolic rate and their manner of reproduction. If either or both of these features were important in the survival of marsupials in Australia, it suggests that Darwin was not right on this matter and that marsupials were pre- adapted for the special conditions of low fertility soils and uncertain climates. In subsequent chapters we will examine the evidence for this idea. This page intentionally left blank Chapter 2

Reproduction and development A neonatal ; pen and ink sketch by AG Lyne. Reproduction and development 39

eproduction is a costly business for any mammal, but essential for survival of the species. For females it demands the transfer of nutrients to the young throughout the dependent Rperiod of its development and growth, whether this be by way of yolk, uterine secretions or milk; for males it demands the expenditure of much energy in establishing territory, repulsing rival males and seeking oestrous females; and for the young the stage when it begins to leave its mother is the most vulnerable time of its life. There is, therefore, a high premium on synchronising the most demanding stages of reproduc- tion with the most favourable times of the year. But most reproductive functions have a long lead- time: it takes from 20 to 60 days to mature an egg for ovulation, 80 days to fashion a sperm capable

Figure 2.1: Distribution of births and time of pouch exit (arrow) for brushtail possum, Trichosurus vulpecula, tammar wallaby, Macropus eugenii, and the western grey kangaroo, Macropus fuliginosus, and pattern of rainfall on , to show that the young of all three species leave the pouch at the most favourable time of the year. For the tammar this is when the daily consumption of milk is highest. Milk consumption after Cork and Dove (1989). 40 Life of Marsupials

of fertilising the egg, and much longer from fertilisation to independence of the offspring. So to match the most vulnerable or costly part of the reproductive process to the most favourable time of the year the first steps in the process must take place many months before the critical event. For mammals the most vulnerable period is when the young are weaned and this usually coincides with the most favourable period of the year. Because lactation in marsupials is long, mating and birth may occur many months before, at an unfavourable season. For instance, the , Antechinus stuartii, gives birth in winter but the young are weaned three months later in late spring when there is abundant insect prey; the brushtail possum, Trichosurus vulpecula, with a five month pouch life gives birth in autumn and the young leaves the pouch in late spring, while tammar wallabies, Macropus eugenii, give birth in mid summer at the hottest, driest time of the year when they are in a poor state of nutrition, but again the young one emerges from the pouch eight months later in spring when its demands for milk are highest and there is abundant new feed after winter rains (Fig. 2.1). How is this timing accomplished? In most habitats the abundance of food varies through the year, whether this be new plant growth in spring after winter rains, the seasonal abundance of insects in a forest or fruiting times in tropical rainforest. If these episodes of food abundance occur regularly each year, mammals that time their reproductive cycle to coincide with it will leave more offspring than competitors that do not. The most widely used predictor in temperate regions of the world is the changing length of the day, which is longest at mid summer and shortest at mid winter. Some species respond to a shortening day length and are termed ‘short-day breeders’ while others respond to increasing day length and are termed ‘long-day breeders’. In the examples given above, male and female brown antechinus come into breeding condition after mid winter and presumably respond to increasing day length (see Chapter 4), whereas brushtail possums and tammars are responding to declining day length after mid summer (see Chapters 7 and 9). In very few marsupials has the response to changing day length been experimentally tested, so inferences about its importance in synchronising breeding must be made with caution. For instance, where the brown antechinus and the , Antechinus swainsonii, occur together both breed after the winter solstice but the latter species begins five weeks earlier than the former: where the two species live in separate places, both breed at the same time (Dickmann 1982). Clearly, other factors as well as photoperiod are involved (see Chapter 4). Brushtail possums came into oestrus two months early when the normal changes in day length were experimentally advanced (Gemmell and Sernia 1992), but in New Zealand the actual onset of breeding in the autumn varies from year to year in response to other factors, such as an abundant set in southern beech trees (Brockie 1992) (see Chapter 7). Similarly, Virginia opossums, Didelphis virginiana, kept in Philadelphia by Farris (1950) were induced to breed in mid winter, two months earlier than normal, by increasing the hours of light in October to mimic the increasing day length that naturally occurs after the winter solstice on 21 December. At this latitude the Virginia opossum is a seasonal breeder but further south in the tropics, where the change in day length is slight, its close relative, the common opossum, Didelphis marsupialis, breeds for most of the year and it is unlikely that change in day length is used to predict favourable times to come. Many tropical species, such as the woolly opossum, Caluromys philander, begin to breed at the end of the annual dry season and their first young leave the pouch when fruit is abundant, while subsequent litters born after the main fruiting season do not survive (Atramentowicz 1982) (see Chapter 3). Conversely, in central Australia, where the rainfall is unpredictable, changes in day length cannot predict favourable times to come and the desert kangaroos have evolved an opportunistic breeding strategy, which is responsive to breaking rains (see Chapter 9). Details of the various breeding strategies that marsupials display will be considered in the later chapters. In this chapter we consider the common features of marsupial reproduction: the Reproduction and development 41

production of mature gametes by males and females at the same time so that fertilisation can occur; the development of the embryo during pregnancy and the extraordinary process of marsu- pial birth; the changing composition of milk during the lengthy lactation; and the development of the young marsupial to independence. The tammar is the main example against which other species are compared: this is because more is known about reproduction in this species than in any other marsupial.

From gametogenesis to fertilisation Marsupials have between 10 and 32 chromosomes, comprising between 4 and 15 pairs of autosomes and a single pair of sex chromosomes (see Table 1.4). At each division of a body cell, the double strand of DNA of each chromosome separates into its component strands and each reconstitutes the complementary half, so that the two new cells each have a full complement of chromosomes (mitosis). In the transformation of a cell into a gamete (gametogenesis to produce a gamete, or sex cell), however, a two-step division takes place (meiosis): at the first division each chromosome lines up with its partner, the DNA strands of both separate, entwine with comple- mentary strands of the other pair, then separate and go to opposite poles of the dividing cell; at the second division each pair of chromosomes separates, so that each daughter cell receives half the autosomes and one sex chromosome. Since female mammals generally have two X chromo- somes, each egg has one X chromosome and since each male mammal has an X and a Y chromo- some, half the sperm carry an X chromosome and half carry a Y chromosome. At fertilisation each gamete brings half the autosomes and one sex chromosome to the union, so the sex of the new individual is determined by the sex chromosome carried by the fertilising sperm.

Oogenesis: production of the egg from stem cell to ovulation The stem cells that will ultimately give rise to eggs, called primordial germ cells, are set aside at a very early stage in the formation of an embryo and subsequently migrate to the site of the future gonad: in the tammar the primordial germ cells have been identified in the day 17 embryo (Ullmann et al 1997) and have reached the ovary by the time the young female is born on day 26. During their migration and after they reach the ovaries they undergo many mitotic cell divisions to give rise to half a million germ cells by two months after birth; and this is followed by the commencement of meiosis (Alcorn and Robinson 1983). This is typical of all mammals: for instance, in the human female fetus the peak of three million germ cells is reached in the second trimester of pregnancy and has declined to one million by birth; no more are formed after this time and so the number steadily declines through life. The other remarkable thing about the germ cells in the mammalian ovary is that after they enter the first phase of meiosis, when the DNA strands have unravelled, they remain in this state, called the primary oocyte, until after leaving the ovary at ovulation, which may be years later. For long-lived species, such as humans, this can be hazardous because the unravelled chromosomes are more susceptible to damage from radiation and so the incidence of genetic abnormalities increases with increasing age of the mother. In the tammar each oocyte grows to a maximum diameter of 0.12 mm and secretes an outer coat of special proteins around itself, called the zona pellucida. Four or five layers of cells surround the zona pellucida, growing on a basement membrane that isolates the oocyte from the blood circulation of the female; the whole structure is called a primary follicle. Further growth of the follicle, leading to ovulation, depends on hormones secreted by the pituitary gland during the oestrous cycle of the adult female (see Ovulation). The cells of the follicle wall secrete a fluid so that the oocyte lies in a central fluid-filled cavity, the antrum, and 42 Life of Marsupials

Figure 2.2: Changes during the oestrous cycle of the brushtail possum, Trichosurus vulpecula, in the corpus luteum and uterine glands and their correlation with progesterone concentration and relaxin activity in the peripheral circulation. Lipid vesicles in the corpus luteal cells contain precursors of progesterone, which are converted to progesterone at day 8–12 before being secreted into the adjacent capillaries. The small dense bodies, which appear later and are associated with the Golgi, probably release relaxin between day 12 and 16. The uterine glands begin to synthesise material by day 4 and maximum release into the lumen occurs between day 8 and 12, after which the exhausted cells are replaced by an underlying epithelium, newly formed from stromal cells. Redrawn from Shorey and Hughes (1973a, b). Reproduction and development 43

in the tammar the whole structure expands to a maximum diameter of 4 mm: it is now called a Graafian follicle, after the Dutch anatomist, Reinier de Graaf, who first described it in the ovary of the rabbit. It is not clear how the selection of follicles that will grow in each cycle is made but the number varies from species to species. In the Virginian opossum and in the , harrisii, up to 50 follicles may grow and ovulate at once, whereas in the kangaroos and brushtail possum very few grow and it is rare for more than one of these to ovulate at one time. In the latter species ovulation occurs in the opposite ovary to the last ovulation, which suggests that there is some local influence in each ovary that stops other follicles from growing when a Graafian follicle is present. This particular influence does not apply in species that ovulate many eggs each time from both ovaries.

Ovulation At ovulation the outer wall of the Graafian follicle ruptures, the fluid and the contained oocyte are extruded into the oviduct, where cilia on the surface cells and contractions of the wall move it on to meet the awaiting sperm. Up to the moment of ovulation the oocyte nucleus still contains four sets of chromosomes but ovulation triggers the final process of maturation: at the first division two sets of chromosomes remain in the egg cell and two sets are discarded with a little cytoplasm as the first polar body; the second division only begins when a sperm penetrates the egg membrane (see Fertilisation) and the chromosomes then reduce to a single set ready for fusion with the sperm nucleus, while the other set are discarded as the second polar body.

The corpus luteum After ovulation the remains of the follicle left behind in the ovary collapse, the cells enlarge and change their function from nourishing the oocyte and secreting oestrogen, to secreting proges- terone, the hormone of pregnancy. A network of fine capillaries and lymph ducts grow among them so that a highly vascular, compact sphere is formed, which is called the corpus luteum, or yellow body, from the colour it acquires as it ages; and its cells are now known as luteal cells (Fig. 2.2). Although the corpus luteum is an inconspicuous little object on the surface of the ovary, it is profoundly important in marsupial reproduction. In the brushtail possum and the tammar it is essential at the beginning and at the end of pregnancy and at the onset of lactation; also in its presence further ovulation is inhibited and in kangaroos it is involved in the phenomenon of embryonic diapause (see Embryonic diapause). It synthesises and secretes at least two hormones: progesterone and relaxin. Progesterone is synthesised on membrane stacks inside the cell and passes easily through the outer cell membrane into the blood circulation, where it can be measured in the blood leaving the ovarian vein and in the general circulation (Fig. 2.2). In the blood it is bound to a protein, called sex hormone binding globulin, which selectively binds to receptors on the cell membranes of target tissues, such as the uterine glands and later the mammary glands, where it releases the progesterone into the cell. The protein hormone relaxin also occurs in the luteal cells of the brushtail possum and the tammar and its concentration rises to a maximum in the middle of the oestrous cycle of the brushtail possum (Fig. 2.2) and at the end of pregnancy in the tammar. The luteal cells become crammed with exceedingly small bodies 200 µm in diameter, budded off the Golgi apparatus, which contain the relaxin (Parry et al 1997b). This was shown by an elegant technique in which luteal cells were incubated with an antibody to relaxin that had been labelled with colloidal gold. Under the the electron microscope it could be seen that gold particles were concentrated in the tiny granules and nowhere else in the cells. By two days after birth both the granules and 44 Life of Marsupials

the associated gold particles had almost all gone, presumably released into the blood circulation at birth.

The oestrous cycle in marsupials To breed successfully a female marsupial has to resolve two different functions: to prepare the genital tract to receive and convey sperm to the vicinity of ripe eggs; and to prepare the genital tract to secrete egg coats and nourish the developing embryo to full term. For all species of Antechinus and a few other species there is a single protracted phase of oestrus lasting for one or two weeks before ovulation occurs; and this is followed by a luteal phase controlled by proges- terone from the corpus luteum. During the long oestrus the female mates with several males and the accumulated sperm survive in special crypts in the oviducts, awaiting the arrival of unferti- lised eggs: in the luteal phase the fertilised eggs complete their development. If the female fails to conceive during the prolonged oestrus phase, she fails to become pregnant for that year. Most female marsupials, however, have more than one oestrous event in a year and the inter- val between one oestrus and the next is called an oestrous cycle. In the tammar this interval is 28 days; in the brushtail possum, 25 days. In other species the length of the oestrous cycle ranges from 21 days in bandicoots (ie Isoodon macrourus), and some rat kangaroos (ie Bettongia), to 25 days in opossums (ie Marmosa robinsoni and Didelphis virginiana) to 45 days in the eastern grey kangaroo (Macropus giganteus). Two phases can be distinguished in the marsupial oestrous cycle: the pro-oestrous phase during the growth of the Graafian follicle leading to oestrus and ovulation, and the luteal phase controlled by the newly formed corpus luteum. In the Virginia opossum, the brushtail possum and most marsupials, pregnancy is accommodated in the luteal phase, which is about two-thirds of the length of the oestrous cycle. After the young are born their sucking suppresses the next pro-oestrous phase and the subsequent oestrus and ovulation. By contrast, in kangaroos and rat kangaroos the length of pregnancy is about the same length as the oestrous cycle, so that oestrus occurs a few hours after the birth of the young, when the female is able to become pregnant again. However, if the newborn young finds a teat and suckles, development of the newly formed corpus luteum and the new embryo are suspended until the end of lactation: this phenomenon is called embryonic diapause (see Embryonic diapause). In all marsupials that have been studied the events in the ovaries and the reproductive tract during the oestrous cycle follow a similar sequence, albeit at different rates. What follows is the sequence in the tammar with some reference to the brushtail possum.

Hormonal control of ovulation and pregnancy Four hormones are involved in the process of follicle growth and ovulation and two more in pregnancy. The pituitary gland, which lies directly underneath the midbrain, secretes two hormones, each from a separate cell type. Follicle stimulating hormone (FSH) stimulates follicle growth, expan- sion of the fluid-filled antrum and secretion of oestrogen by the follicle cells before ovulation. Luteinising hormone (LH) is also important in follicle maturation but its main role is to induce ovulation and the transformation of the cells of the follicle wall into luteal cells that secrete proges- terone. Oestrogen, in turn, affects the concentration of pituitary hormones in the circulation. At low concentrations oestrogen inhibits the secretion of LH and FSH – a negative feedback effect – but at high concentration it induces a massive outpouring of LH – a positive feedback effect. The secretion of the two pituitary hormones is also controlled by another hormone, called Gonadotrophin releasing hormone (GnRH), which is synthesised by nerve cells in the brain. It is secreted in regular pulses of about three per hour, which causes the pituitary hormones to likewise be secreted in pulses, especially LH. Both the pulse rate and the amount of hormone Reproduction and development 45

released into the blood stream are important factors in stimulating the ovary to grow follicles and secrete oestrogen. GnRH has not been measured directly in any marsupial but its role can be inferred from three experiments in tammars. When female tammars were injected with GnRH a pulse of LH was detected in the blood within one hour. Furthermore, after removal of the ovaries of female tammars the levels of both FSH and LH rose, indicating the removal of oestrogen negative feedback. The pattern of LH in these tammars was pulsatile with a frequency of 1 pulse every two hours. When injected with low doses of oestrogen, equivalent to levels in intact females, there was a rapid reduction in both hormones – the negative feedback, while higher doses equivalent to oestrous levels caused a very large pulse of LH 20 h after the injection – the positive feedback effect (Horn et al 1985). If the pituitary gland is surgically removed from a tammar, no follicles grow in the ovaries and those that have already begun, go no further (Hearn 1975). A similar effect has been induced in female tammars that were immunised against GnRH (Short et al 1985), so that the pituitary is unable to secrete FSH and LH. The appearance of the ovary under both these experimental conditions looks very like the anoestrous ovaries of species that have a short breeding season, such as the greater glider, Petauroides volans, and the brushtail possum, or red kangaroos during a severe drought, so it is likely that their pituitary hormones are at low levels at these times. Presumably the onset of the breeding season in these species begins when the pulse rate of GnRH secretion from the brain increases and the pituitary gland secretes FSH, which stimulates some follicles to begin their final growth. In the anoestrous red kangaroo the response to break- ing rains after a drought is dramatic: in three studies the females had growing follicles and some had come into oestrus 14 days after the rains began (see Chapter 9). The process is slower in the greater glider, taking one month from the start of changes in the ovary to the time of oestrus and ovulation (see Chapter 7). This complex interaction of hormones secreted by the brain, the pituitary gland and the ovary provide the means for each species to adapt its investment in reproduction to the most favourable times of the year.

Hormone changes during the oestrous cycle in the tammar wallaby Changes in progesterone and oestrogen in the tammar have been measured in blood samples taken at daily or shorter intervals during pregnancy or the oestrous cycle (Fig. 2.3). Four or five days before birth and oestrus the concentration of progesterone in circulation is high, oestrogen and LH undetectable and FSH rising. At birth, a few hours before oestrus, the progesterone level falls to one-quarter of its former level while the oestrogen level rises to reach its peak level of 15–20 pg/mL at oestrus (see Fig. 2.12). Coincident with, or eight hours after the oestrogen peak, there is a large, transient, pulse of LH, and this is followed 24 h later by ovulation. By sampling blood from the veins draining each ovary (Fig. 2.3) it is clear that the progesterone comes from the ovary bearing the corpus luteum of pregnancy and the oestrogen comes from the other ovary bearing the Graafian follicle destined to ovulate imminently (Harder et al 1984). For the first five days after ovulation the concentration of progesterone in the circulation is low, less than 0.2 ng/mL, and then on day 6 or 7 there is a brief rise in concentration to 0.5 ng/mL (Fig. 2.3). On about day 10 a gradual rise begins to reach a relatively high concentration (about 0.5 ng/mL), which persists to the end of the cycle, when there is a return to the basal level. The first rise results from an increased rate of secretion of progesterone by luteal cells and the second, prolonged rise is due to the increase in the number and size of the luteal cells themselves, not to a change in the rate of progesterone secretion. Neither the secretion rate nor the subsequent growth was affected by removing the pituitary gland, which shows that the corpus luteum, once formed after ovulation, has an independent 46 Life of Marsupials

Brushtail possum

Post-partum uterus Non-pregnant uterus

Figure 2.3: Central panel, changing levels of progesterone in the general circulation of the tammar wallaby, Macropus eugenii, through pregnancy, showing the early pulse and later substantial rise in late pregnancy and rapid fall at birth. Side panels, showing the large output of progesterone from the corpus luteum on one ovary through late pregnancy, and the large output of oestrogen from the Graafian follicle on the other in the post partum period. Data from Harder et al (1985), Tyndale- Biscoe et al (1986).

life. How the small luteal cells can independently increase and then decrease their rate of progesterone synthesis and secretion remains unknown. And more remarkable is that they do this while simultaneously dividing and beginning to enlarge. No other endocrine cell is known to do this. At the end of the oestrous cycle the luteal cells cease to secrete progesterone and shrink in size and the corpus luteum becomes a functionless scar in the ovary.

Changes in the vaginal complex during the oestrous cycle During the pro-oestrous phase, while the follicles are maturing, the vaginal complex enlarges, the lining thickens and the two lateral vaginae open. At oestrus the vaginal complex becomes several times larger than in the quiescent state and fluid is secreted into the lumen by the glandu- lar lining of the median vagina and by the shedding of lining cells: the appearance of these cells in a swab taken from the genital tract provides evidence of oestrus. The great size of the vaginal complex in oestrous kangaroos led early observers to think that it is the site of sperm storage: but this not so, since sperm do not remain for more than a few days in the female tract of tammars and other kangaroos. However, in tammars that mate with several males during their brief oestrus, the vaginal complex becomes grossly distended with more than 100 mL of seminal fluid, some of which may coagulate to form a plug. The significance of this is not clear, but the fact that the prostate glands of males enlarge greatly during the main breed- ing period (see Breeding strategies of male marsupials) when the vaginal apparatus of females is enlarged, indicates some complementary function, either in sperm competition or to aid the transport of spermatozoa by stimulating the uterus and oviduct to undergo regular peristaltic contractions during this time. After oestrus the vaginal complex shrinks, while the two uteri enlarge and become more vascular and swollen. Reproduction and development 47

Uterine changes during the oestrous cycle The enlargement of the uteri is due to an increase in the number of cells and their transforma- tion to active secretion. The cells divide during and after oestrus and, in the brushtail possum, this appears to be at a greater rate in the uterus contiguous to the ovary bearing the Graafian follicle or new corpus luteum. Subsequently, when the cells enlarge, this uterus is larger than the other because of the greater number of cells. In the brushtail possum and the tammar each ovarian vein forms a network of small branches that are intimately associated with branches of the uterine artery of the same side (Lee and O’Shea 1977, Towers et al 1986), so that oestrogen, arising in the Graafian follicle, can probably be conveyed directly from the ovary to the contigu- ous uterus, without being diluted in the general circulation. Later in the cycle in the tammar secretions from the uterus associated with the new corpus luteum consistently contain greater amounts of protein than from the opposite uterus. This may be important for nourishing the developing embryo, which is carried in this uterus. As the corpus luteum grows and progesterone in circulation increases, the gland cells of the uterus are transformed from cuboidal cells with large central nuclei to elongate columnar

Figure 2.4: Protein components, separated by electrophoresis, of maternal blood serum and uterine secretions, and fetal yolk sac, allantoic and amniotic fluid during pregnancy in the tammar wallaby, Macropus eugenii. After Renfree (1973a). 48 Life of Marsupials

cells with small basal nuclei. This characteristic form has been termed the luteal phase because progesterone induces it. In the brushtail possum it extends from day 4 (when progesterone begins to increase in the blood) to day 18. Shorey and Hughes (1973a) showed by using the electron microscope that the region around the nucleus of each cell is undergoing active protein synthesis (Fig. 2.2). After day 8 synthetic activity declines and the accumulated secretion pours from the apices of the cells and may even include cell components. This material continues to flow into the lumen of the uterus for several days, reaching a peak at day 12 and declining thereafter. This secretion is essential for the early development of the embryo. Uterine gland cells of the tammar in the luteal phase show a similar appearance under the electron microscope. In this species the secretion has been analysed (Renfree 1973a). It contains glucose and protein, the latter being composed almost entirely of albumin and pre-albumin fractions (Fig. 2.4), whereas the slower moving proteins, characteristic of blood serum, are not detectable. This supports the evidence from electron microscopy that uterine fluid is a product of active secretion rather than an exudate of lymph or serum, as was formerly thought. By day 18 in the brushtail possum, when birth occurs, the gland and epithelial cells are exhausted and dying. Meanwhile undifferentiated cells in the uterus, having assembled under the basement membrane, establish contact with each other and form a new basement membrane beneath the old one. When this is complete the old epithelium sloughs away into the lumen of the uterus where it is absorbed and the new layer becomes a new epithelium in readiness for the next cycle (Fig. 2.2) (Shorey and Hughes 1973a). A similar regeneration occurs in other species; and in some dasyurids, such as the , Dasycercus byrnei (Fletcher 1989), the breakdown at the end of the luteal phase leads to bleeding at the time when birth occurs in the pregnant animal.

The saga of the sperm As in the female tammar, the primordial germ cells of the male have migrated to the future testis by the day of birth, where they undergo many mitotic divisions, and become arranged in the wall of the seminiferous tubules of the testis. But their transformation into sperm does not begin until the young male reaches sexual maturity between one and two years later. At sexual maturity these cells undergo waves of mitotic divisions at intervals of 16 days for the rest of the animal’s life (Jones 1989). The wall of the seminiferous tubule comprises two kinds of cell: the spermatogonia that give rise to sperm, and the Sertoli cells that support them. The first divisions increase the number of cells destined to become sperm from the spermatogo- nia that remain on the wall of the tubule and give rise to the next wave of sperm. This process in the tammar takes 26 days and is followed by meiosis, which takes a further 21 days. Unlike in the female, where each primary oocyte gives rise to one egg and two polar bodies, each spermatocyte gives rise to four spermatozoa, two bearing 7 autosomes and an X chromosome and two bearing 7 autosomes and a Y chromosome. Supported by the surrounding Sertoli cells, these cells now transform into spermatozoa: the nucleus condenses and changes shape; the Golgi apparatus migrates to a position in front of the nucleus and becomes the acrosome; the mitochondria are rearranged into a spiral around the flagellum to provide the energy for swimming; and the Sertoli cells absorb the surplus cytoplasm. In the tammar and brushtail possum this process takes a further 25 days, very similar to placen- tal mammals. At its completion the young sperm detaches from its supporting Sertoli cell, and begins its journey along the length of the seminiferous tubule to the epididymis. The whole process to this point has taken 72 days. The epididymis is a highly convoluted tube, 35 m long in the tammar, and it is while it passes along this tube that the sperm acquires its final form, characteristic of the particular species: in the American marsupials it is in the epididymis that pairs of sperm become attached Reproduction and development 49

to each other by their respective acrosomes (see Fig. 1.6). In the tammar and brushtail possum the sperm take 13 days to reach the end of the epididymis, travelling at about 11 cm/h. Although the spermatozoa can now move and are capable of fertilisation, they can remain here immotile for up to 28 days or until ejaculated at copulation, or shed in the urine. Thus, it takes 85 days to fashion a sperm capable of fertilisation and it can remain in storage for one month. The caudal region of the epididymis lies at the lower pole of the testis and is, therefore, nearest to the bottom of the scrotum. For some reason, not understood, the core body tempera- ture in most mammals is too high for survival of sperm and, if the testes are experimentally returned to the body cavity, or the scrotum is heated, spermatogenesis is disrupted and sperm in the epididymis lose their potential for fertilisation. In a variety of marsupials that have been examined the temperature of the testis and epididymis is about 5oC lower than the core body temperature. In the tammar this is achieved by a counter current heat exchange system in the neck of the scrotum (Setchell and Waites 1969). The testicular artery of the tammar divides into about 150 branches that are intimately associated with the similarly subdivided testicular vein of 50 branches, so that heat is transferred from the warm arterial blood, entering the scrotum, to the cooler venous blood returning to the body. In other marsupials the number of branches is highest in large species, less in small species and there is none in the marsupial mole, Notoryctes

Figure 2.5: Changes in weight of prostate glands of the brushtail possum, Trichosurus vulpecula, and the tammar wallaby, Macropus eugenii, associated with the main and subsidiary periods of the year when females are in oestrus. After Gilmore (1969) and Inns (1982). 50 Life of Marsupials

typhlops, which carries the testes in the abdominal wall. This arrangement for cooling the testes is quite different from the arrangement in placental mammals and is further evidence that the evolution of the scrotum in marsupials was independent of that in placentals.

Prostate gland The prostate in marsupial males is a glandular outgrowth of the wall of the urethra through which sperm and urine pass to the exterior. Secretions from the prostate glands provide the carbohydrate fuel for the sperm in the female reproductive tract. The size of the prostate varies seasonally in some species, such as the brown antechinus, and in others, such as the brushtail possum and the tammar the size is maximal when females are in oestrus during the breeding season (Fig. 2.5). As in other mammals the prostate is stimulated by testosterone secreted by the testes and this hormone is elevated during the breeding season.

Hormonal control of reproduction in the male marsupial Only in the tammar has the role of the pituitary in reproduction been examined. After its removal spermatogenesis ceased, the testosterone-secreting cells shrank and all sperm were lost from the epididymis after two months (Hearn 1975). Likewise, the prostate gland shrank in the same way as it does after castration, indicating that the testes were unable to secrete testosterone in the absence of the pituitary. Conversely when intact male tammars were injected with the brain hormone GnRH the concentration of the pituitary hormone LH rose in the circulation and testosterone was also elevated (Lincoln 1978). From the appearance of the tissues in other species we can infer that pituitary hormones, likewise, stimulate them leading up to the breeding season.

Breeding strategies of male marsupials Because of the long period required to produce sperm, male marsupials have three possible strategies for ensuring that they are able to deliver sperm capable of fertilisation when females are in oestrus. First, they can produce sperm continuously and be capable of fertilisation at all times of the year. Second, they can respond directly to the changes in the females; or third, they can independently respond to seasonal changes and only produce sperm during a brief breeding season, synchronised to the breeding season of females.

Continuous breeders The adult males of all species of kangaroo, wallaby and rat kangaroo are continuously fertile, as are many species of opossum and the larger arboreal marsupials, such as the brushtail possum and koala, Phascolarctos cinereus. Since male marsupials produce about 25 million sperm each day, there is some cost in this strategy and, under extremely arid conditions, sperm production in kangaroos may be impaired (see Chapter 9). However, a larger cost may be the production of sufficient seminal plasma, mainly from the prostate. In tammars and brushtail possums the males are continuously fertile but the size of the prostate gland is enlarged only during peak breeding, which for the possum is April to May (Gilmore 1969) and for the tammar is January to March (Inns 1982). In both species there is a second, smaller increase in prostate weight in September to October, which coincides with the time when young females of the year come into first oestrus (Fig. 2.5). Both peaks in prostate weight of the tammar were associated with elevated testosterone in the circulation (Inns 1982), which only occurs if the males are associating with females (see Fig. 9.21). Thus, the males are responding to changes taking place in the females as they come into breeding condition. Reproduction and development 51

Seasonal breeders A second strategy, seen in highly seasonal breeders, such as the greater glider, is for the testes to be shut down for most of the year and for sperm to be produced only during the short breeding season of the female. Because of the long lead time needed to produce mature sperm, the males must respond to some external signal, rather than to the immediate condition of the females. In greater gliders the testes begin to enlarge in mid January and spermatogenesis takes place from then until the end of February, with mature sperm being produced only between mid March and mid May (Smith 1969). The ovaries of the females increase in mass from mid March, when pair formation and copulation occurs, and young are born in April to May. After this time the testes of the males shrink and they cease to produce sperm, so females that fail to conceive at the first oestrus of the year do not have another opportunity to become pregnant. This highly synchro- nised breeding season results in a single, even-aged cohort of young emerging from the pouch in September and becoming independent of their mothers by January (see Chapter 7). The most likely signal that the males respond to is change in day length after the summer solstice, but this has not been investigated in this species.

Single cycle breeders The ultimate strategy is seen in all the species of Antechinus and in which there is only one wave of spermatogenesis, resulting in a single cohort of sperm (Kerr and Hedger 1983). The onset of spermatogenesis is probably triggered by a photoperiod signal (although this has not been investigated critically) so that the sperm complete their development in the epididymis, and the prostate reaches maximum size, at precisely the same time as females come into oestrus. Within days of the females becoming pregnant all the males die. Males taken into captivity and nursed through the post mating period lived for another year but never produced any more sperm because their seminiferous tubules were totally degenerate (Woolley 1966). Some of the small didelphids of South America have a similarly brief period of sperm production and males only contribute to one breeding season (Pine et al 1985, Lorini et al 1994) but it is not clear that the males are as strictly programmed for one wave of spermatogenesis as the dasyurids are.

Copulation and the fate of sperm in the female tract The tammar female will accept the male for a brief period of about eight hours at the time when a mature Graafian follicle is present in one ovary, oestradiol is at a peak level and progesterone is in decline. The essential hormonal conditions that elicit oestrus may be the change in the ratio of the two hormones (Figs 2.3, 2.12). As the female comes into oestrus several males follow her persistently, with the dominant male keeping a close guard and repulsing other males. This male is the first to copulate but other males will do so over the next several hours. A similar pattern has been described in the Virginia opossum, where the largest male mates first and guards the female for several hours after (see Chapter 3). Since ovulation in the tammar does not occur until about 30 h after the end of oestrus the sperm of all the males that mate will be available to fertilise the egg. So, what is the fate of sperm in the female reproductive tract after copulation? In the tammar the sperm enter the cervix within one hour of the first mating and some have reached the oviducts one to four hours later, others arriving over the next 24 h (Tyndale-Biscoe and Rodger 1978). The question of whether sperm travel the distance by their own efforts or are carried by peristaltic contractions of the tract is still debated. In a dilute medium the sperm of bandicoots and mouse opossums do not move forward. However, when placed in medium of the viscosity of the genital tract secretions, the sperm of both species showed directional movement. In the case of bandicoots the rate of 52 Life of Marsupials

movement was 0.45 mm/s (Taggart 1994), which is quite fast enough to cover the length of the tract in the time available. Nevertheless, many sperm do not make the distance because they go into the blind ends of uterine glands. Since ovulation occurs about 30 h after the end of oestrus, many sperm will already be at the top of the oviduct when the egg enters it. This is important because the oviduct secretes a second protein coat around the egg soon after it enters the oviduct, which is impenetrable to sperm. Thus, there is only a very brief opportunity for the sperm to penetrate the zona pellucida after ovulation and only those sperm that are present will be able to engage with the egg surface and fertilise it. In the gray short-tailed opossum, Monodelphis domestica, Moore (1996) showed that there is only about 15 minutes after the egg reaches the oviduct when sperm can attach to the zona pellucida, before the new layer to which the sperm cannot attach covers the egg. The same is true of the Virginia opossum (Rodger and Bedford 1982a) and probably all marsupials. Therefore, fertilisation must always occur at the top of the Fallopian tube.

Whose sperm fertilises the egg? Because male tammars are larger than females and compete vigorously for access to the female it has been supposed that sperm of the dominant male, which mates first, will reach the egg first. Peter Temple-Smith and colleagues investigated this question in captive tammars (Ewen et al 1993) by comparing minisatellite DNA profiles from the competing males and the subsequent progeny, to identify the father of each one. The results did not support the supposition: there was no correlation between the order of mating and the sire of the young produced. Similar results have subsequently been reported in other captive tammars but no one has repeated their work in wild animals, which would be necessary to be certain that the most dominant male has no selective advantage in the paternity stakes. In antechinus oestrus lasts for several days and ovulation does not occur until the end of that time. During oestrus a female will copulate with up to four different males, so sperm must survive for much longer than in species with a brief period of oestrus. Large numbers of sperm accumulate in the oviducts, where they are packed together in ordered ranks in special crypts (Breed 1994). Unlike in the tammar, many eggs are shed at ovulation in antechinus, which again raises the question of the paternity of the resulting litter. Because they only have this one time in their brief lives to leave offspring there is strong competition among males to mate with oestrous females. Unlike other marsupials and placentals, the number of sperm produced by male antechinus are surprisingly low – less than half a million per ejaculate – and the number declines through the brief breeding season as the limited supply is used up. Nevertheless, the transport of sperm through the female tract is much more efficient than in other species, so that the number of sperm found in the oviduct of a mated female are similar to those of other species (Taggart and Temple-Smith 1991). When the paternity of the resulting litters was determined by genomic DNA profiles it showed that several males share paternity of the litter, with the last to mate leaving the largest number of offspring (Shimmin et al 2000, Kraaijeveld-Smit et al 2002).

Life before birth – fertilisation to parturition Fertilisation The process of fertilisation has now been described in five species of marsupial: the Virginia opossum, the gray short-tailed opossum, the brushtail possum, the brushtailed bettong, Bettongia penicillata, and the fat-tailed dunnart, Sminthopsis crassicaudata, and the details are similar in all of them (Rodger and Bedford 1982b, Baggott and Moore 1990, Breed 1996). Reproduction and development 53

The sperm attaches to the zona pellucida by its acrosomal face and this causes the acrosome to release four enzymes, which dissolve the proteins of the zona pellucida and enable the sperm to push through (Fig. 2.6) (Breed 1996). As soon as the head of the first sperm is through the zona pellucida it attaches to the outer membrane of the egg, setting off a reaction in the outer part of the egg that prevents any other sperm from attaching. It also provokes the egg to complete its second maturation division, extruding the second polar body and reducing the egg nucleus to a single set of chromosomes, ready to meet the complementary set from the sperm nucleus.

Figure 2.6: The moment of fertilisation in the short-tailed opossum, Monodelphis domestica. The head of the sperm is burrowing into the outer matrix of the zona pellucida of the oocyte. From Breed (1996) Reproduction, Fertility and Development 6, 627, fig 16d, with permission of CSIRO Publishing.

Only the head and midpiece of the sperm pass into the egg, the outer membrane and the tail being discarded. In the sperm nucleus, which until now has been very compact, the chromo- somes separate in preparation for their pas de deux with the chromosomes of the egg nucleus. Each chromosome meets its pair from the other parent and becomes intertwined with it on the spindle. Each then replicates itself and one full set of the new chromosomes separates to each 54 Life of Marsupials

pole of the spindle. The egg cytoplasm divides between the two sets and the first division of the newly fertilised egg is completed. All of this activity takes place in the first 24 h after ovulation, as the egg is carried along the oviduct and enters the uterus. This is much faster than in placental mammals, where the first four or five divisions occur in the oviduct, and the embryo enters the uterus three to five days after ovulation as a ball of about 60 cells.

Figure 2.7: Early development of the tammar wallaby, Macropus eugenii: (a) the mature ovarian oocyte, surrounded by the zona pellucida; (b) the fertilised egg in the oviduct, surrounded by the mucoid layer and engulfed sperm; (c) the 4-cell stage; (d) the12-cell stage in the uterus, surrounded by the shell membrane; (e) an early formed blastocyst on day 6; and (f) a blastocyst in diapause. From own collection. Reproduction and development 55

As the fertilised egg passes along the oviduct it is covered by the second egg coat, in which many sperm become trapped. Then, upon entering the uterus on the second day after ovulation, the egg is covered by a third coat of keratin, secreted by the cells of the uterus (Fig. 2.7). These two egg coats are homologous to the albumin and soft shell membrane of a snake’s egg and of the eggs of platypus and echidna. No placental mammal has a shell membrane but a few species, such as the rabbit and hare, do have the middle mucoid coat.

Early development of the marsupial embryo The early development of five species of marsupial is known in detail and of several others less completely (see Tyndale-Biscoe and Renfree 1987, Selwood 1992, Renfree and Lewis 1996).

Cleavage and blastocyst formation The marsupial egg contains material, which has been called yolk, although it is not clear that it is the same as the yolk of a hen’s egg. During the first cleavage division it is extruded from the cells into the surrounding space where it forms either a discrete body (in dasyurids and peramelids) or a diffuse mass (in didelphids). In the diprotodonts (eg honey possum, Tarsipes rostratus, greater glider, brushtail possum, tammar and eastern grey kangaroo) there is no evidence of a separate yolk mass and the first two cleavage divisions result in a tetrad of cells. As further cleav- age divisions occur the cells become flattened against the inner surface of the zona, which acts as a scaffold, to form a hollow sphere of 60–80 cells, the blastocyst (Fig. 2.7). With the electron microscope it can be seen that these flattened cells are joined at their margins to each other by junctional complexes, which effectively close off the space inside, containing the yolk sac fluid, from the surrounding environment (Renfree and Lewis 1996). Thus, all subsequent movement of substances must be through the cells themselves. In the tammar all the blastocyst cells appear to be identical, with no hint of where the embryo proper will form. However, in dasyurids, the position of the yolk body gives the blastocyst a polarity, with smaller cells near it and larger cells further away. Yousef and Selwood (1996) traced these individual cells through later development in Sminthopsis and Antechinus and showed that the small cells eventually give rise to the embryo proper and the remaining cells become the fetal membranes. This confirmed the views of Hill (1910), who described the development of the eastern , Dasyurus viverrinus, and surmised that the unequal sized cells might have different developmental fates. This finding brings marsupial development closer to the placental pattern (Johnson and Selwood 1996). In most placentals the first cleavage divisions give rise to a solid ball of cells called a morula (Latin for mulberry) without a central cavity containing yolk, and junctional complexes form only between the outside cells. The cells trapped inside become the embryonic stem cells that give rise to the embryo proper and the outside cells give rise to the placenta. Thus, the true embryo is set apart from the supporting tissues from a very early stage of development, whereas in marsupials it had been supposed that the embryo is differentiated much later. Now it seems that early development may not be so different. Indeed, there are some placentals, such as the elephant shrew, Elephantulus myurus, (van der Horst 1942), in which the fertilised egg enters the uterus in early cleavage and becomes a hollow sphere without an apparent presumptive embryo, just like diprotodont marsupials do. The blastocyst is the stage of development that the embryo can reach in both groups of mammals on the resources carried in the egg at fertilisation: further development depends absolutely on nourishment provided by the enclosing uterus. In placentals the embryo hatches from the zona pellucida at this stage and the outer trophoblast cells attach to, or actually invade, the uterus by a process called implantation. By contrast, in marsupials the blastocyst remains 56 Life of Marsupials

enclosed by the three egg coats and it grows by absorption of uterine secretions across the blasto- cyst wall. If the uterine glands are not in a secretory condition, further development cannot take place.

Embryonic diapause In a wide variety of marsupials and placentals the blastocyst has the ability to enter a state of dormancy, called embryonic diapause, during which cell division and growth either cease or continue at a very slow pace until an appropriate signal from the mother is received. The signal may be a rise in an ovarian hormone or a specific growth-promoting substance secreted by the uterus. Embryonic diapause occurs in all kangaroos, wallabies and rat kangaroos, except the western grey kangaroo, Macropus fuliginosus, Lumholtz’s tree kangaroo, Dendrolagus lumholtzi, and the musky rat kangaroo, moschatus (see Chapter 9), and it occurs in pygmy possums, the feathertail glider, Acrobates pygmaeus, and the honey possum (see Chapter 6). Even in the dasyurids, which do not undergo diapause, there is an important change during the formation of the blastocyst: whereas it is possible to grow embryos through the early cleavage divisions in culture, it is not yet possible to carry development through to the blastocyst stage, although blastocysts recovered after this critical period will readily grow on in culture (Selwood and Young 1983). A similar block has been observed in the mouse and it is possible that in all mammals there is a change in the metabolic pathway on the completion of cleavage divisions and the beginning of cell growth. The way in which diapause is controlled in marsupials has been thoroughly investigated only in the tammar wallaby and the results are examined in the following two sections.

Initiation and maintenance of embryonic diapause Tammar females come into oestrus within a few hours of giving birth and, if the newborn attaches to a teat and is suckled, the new pregnancy is delayed during lactation. If the young in the pouch dies, or is experimentally removed, the corpus luteum begins to grow, the delayed pregnancy resumes and birth occurs 26–27 days later. So to address the first question of when diapause begins, we can compare the progress of early pregnancy in females that are carrying a new young in the pouch with females that are not. There is no discernable difference in the rate of cleavage up to day 8, at which time the blasto- cyst has formed and consists of 80–100 cells (Fig. 2.7f). In lactating females no further change occurs in the blastocyst, the corpus luteum does not grow, the early pulse of progesterone does not occur and the uterine glands remain small. By contrast, in the non-lactating female the new corpus luteum grows, the transient early pulse of progesterone in the circulation occurs on day 7, the uterine glands enlarge and become secretory, and cell division continues in the blastocyst, which increases in size, so that by day 12 it is 1 mm in diameter and growing fast. To test the role of the corpus luteum, Sharman and Berger (1969) surgically removed it from several non-lactating tammars on day 2 after oestrus, when the fertilised egg would just have reached the uterus and begun to divide. When they examined the tammars 10 days later all had quiescent blastocysts, whereas unoperated females on the same day had enlarged blastocysts. From this they concluded that the fertilised egg can reach the blastocyst stage without the corpus luteum but it can go no further. How long can a blastocyst remain dormant and not die? In the tammar the blastocyst remains in diapause throughout lactation and then for some months after, only resuming its interrupted development after the summer solstice, when the corpus luteum grows and progesterone levels increase in the circulation (see Chapter 9). Indeed, blastocysts can survive for several months in Reproduction and development 57

ovariectomised females, which suggests that diapause is a passive state from which the blastocyst must be awakened by a special signal.

The end of diapause, first steps in reactivation Removing the pouch young starts a train of events in the ovaries, the uterus and the dormant embryo that culminates in birth 26–27 days later (Fig. 2.8). During days 1–3 the process can be reversed if the young is returned to the pouch and resumes sucking: after this day the blasto- cyst reactivates irreversibly and either develops to full term or expands briefly and then dies (K Gordon et al 1988). Days 1–3 encompass the time it takes for the corpus luteum to resume its own development, which is evident on day 4 when the luteal cells begin to divide and enlarge. Progesterone secretion increases and the transient pulse occurs on day 5, 6 or 7.

Diapause ----- Reactivation Embryo anaerobic glucose glucose metabolism glycolosis oxidation uptake increases Embryo none cells resume increased blastocyst shell fetus begins birth development division protein expansion membrane to secrete synthesis begins breaks down cortisol Endometrium no change increased luteal phase release of protein in glands uterine synthesis secretion Corpus reversible cells divide, large pulse of low level of elevated elevated sharp fall in luteum if sucking progesterone progesterone progesterone progesterone progesterone progesterone resumes increases Effect of blocks embryo embryo goes blocks blocks birth removing reactivation reactivates, to term mammary and lactation corpus then dies gland luteum development Days after removing 1 23 4 5 6 7 8 9 10 18 19 23 24 25 26 pouch young

Figure 2.8: Summary of changes after diapause in the tammar wallaby, Macropus eugenii: the series of changes in the corpus luteum, the endometrium of the uterus, and in the metabolism and growth of the blastocyst after the sucking inhibition is lifted by removing the pouch young (RPY), and the effects of removing the corpus luteum on successive days after RPY.

When the corpus luteum was removed on day 4 the blastocyst resumed its development but then died; but when the corpus luteum was removed on or after day 6, development went to full term but birth did not occur (Sharman and Berger 1969). Thus, the corpus luteum is necessary for the first six days of active pregnancy and for parturition but is not required for the whole process of growth and differentiation of the embryo. Very similar results have been found in the , Setonix brachyurus, the Virginia opossum, the brushtail possum and the long-nosed potoroo, Potorous tridactylus (Tyndale-Biscoe and Renfree 1987). In all these species the essential factor for gestation to continue is that the uterus is in the luteal phase at the time of operation: once established, its secretions are adequate to initiate and maintain development of the embryo to full term. If the blastocyst awaits a signal from the stimulated uterus to reactivate, what is the nature of that signal? It cannot be the abundant secretions of the luteal uterus because they do not occur until later, so it must be something more subtle coming from the corpus luteum. The first 58 Life of Marsupials

changes in the reawakened blastocyst have been the subject of much research and point to what the signal may be. The first sign that the blastocyst is reactivating occurs on day 4 when the first mitotic figures are seen in its cells as they resume division; and on day 5 the first changes in the synthesis of nuclear and cytoplasmic RNA occur (Moore 1978, Shaw and Renfree 1986), the precursor to new protein synthesis (Fig. 2.8). Cell division and protein synthesis require energy, which comes from carbohydrate metabolism in the blastocyst. During diapause the main pathway of glucose metabolism is anaerobic glycolysis (Spindler et al 1998) with a high ratio of adenosine triphosphate to adenosine diphosphate (ATP:ADP), which may inhibit the enzymes of oxidative metabolism. Between day 3 and day 4 after RPY this changes dramatically: ATP stores become depleted so that the ATP:ADP ratio falls significantly by day 3 and the contribution of glucose oxidation to ATP production rises significantly by day 4. This means that the capacity of the blastocyst to generate energy from each molecule of glucose increases nearly 20-fold between day 0 and day 5. This is followed by a 10-fold increase in uptake of glucose by day 10, reflecting the profound changes in the metabolism of the growing blastocyst (Pike 1981) and the increasing demands for energy to fuel the active transport of uterine secretion across the blastocyst wall as it expands. Without the increased resources of uterine secretions, including glucose, expansion must fail, as it does when the corpus luteum is removed before day 6. The findings of Spindler et al (1998) show that the first signal from the corpus luteum to the blastocyst must occur on day 3, coinciding with its own release from the quiescent state. Is this signal a change in the concentration of progesterone coming directly to the uterus via the network of small branches of the ovarian vein and the uterine artery, mentioned earlier (Fig. 2.3), or is it a specific growth factor secreted by the uterus? Geoff Shaw (1996) has described the increase in the number of proteins in uterine secretion when it is stimulated with oestrogen or progesterone and one such factor, platelet activating factor (PAF) was found to be elevated on day 3 in tammars (Kojima et al 1993), but it is not known whether this factor is involved in blastocyst reactivation. Several other polypeptide growth factors could also be candidates but at present none has been identified. So the question as to the nature of the early signal remains unanswered, but the critical time for it has been narrowed down to early on day 3. Once reactivated and with an adequate supply of nutrients from uterine secretions, the embryo continues without further delay to completion of pregnancy: indeed, in the tammar the time from the early pulse of progesterone to birth is constant at 22 days.

Development of the fetus The blastocyst increases in size by absorption of fluid and by a rapid increase in the number of cells: by day 10 it is a translucent vesicle 5–10 mm in diameter. It then develops a polarity that shapes its future. In one hemisphere some cells become detached from the outer layer and lie free inside and spread around the inside of the vesicle forming a rather loose network held together by thin strands. This becomes the inner layer of the two-layered blastocyst and its cells are destined to become the linings of the gut, lungs and bladder – the so-called endoderm. The first formed endoderm cells induce the overlying cells to differentiate into an oval plate of thick cuboidal cells, which will become the outer layer of the embryo – the ectoderm, which gives rise to the skin, brain and eyes. As the blastocyst continues to expand the outer cells of the yolk sac become very thin but still retain complete attachment with each other. The cells of the future embryo, however, do not expand so that this area becomes more distinct and is clearly visible in living vesicles as an oval plate (Fig. 2.9a, Plate 3). The embryo now develops a bilateral symmetry with the appearance of a midline groove and knot of cells at one end, in the same manner as does the chicken egg at 18 h of incubation. As in Reproduction and development 59

the chicken, these are the site of a remarkable migration of cells from the outside between the two existing layers to form the third or mesoderm layer, which will give rise to blood, muscle and bone. At the same time the notochord is formed and over it develops the neural tube from which develops the brain vesicles and spinal cord. From the lateral mesoderm differentiate the somites, which will later form the trunk muscles; these first appear on day 16 in the tammar embryo (Fig. 2.9a, Plate 3) and on day 17 the primordial germ cells can be seen at the periphery of the somites. On day 19 the great vessels of the heart appear and soon after they begin to pulse; and the limb buds appear. Between days 19 and 26 the embryo completes the development of all the organ systems necessary for life outside the uterus. This development is entirely dependent on the functions of the fetal membranes that surround it and form the placental connection to the uterus that provides its nourishment, respiration and excretion.

Egg coats, fetal membranes and placental transport Possible functions of the egg coats The three egg coats consist of the inner zona pellucida, laid down by the oocyte while in the ovary, the mucoid coat laid down in the oviduct and the keratinous shell membrane secreted by the uterine cells. As already mentioned, an important role for the zona pellucida is to act as a scaffold for the early development of the blastocyst before its cells have made strong connections to each other, but it is not clear what function the other two coats have in marsupials: the zona pellucida and mucoid coat disappear as the embryo enlarges and they may provide nutrients at this stage. By contrast, the shell membrane remains intact for more than three-quarters of the length of pregnancy, as the fertilised egg transforms into a fluid-filled vesicle and developing embryo. During the enormous expansion of the embryonic vesicle the shell membrane becomes extremely thin but remains intact, so much so that the vesicle can be rolled out of the uterus and grown in culture without damage. Since the shell membrane is intact uterine secretions must be able to pass through it for synthesis into new tissues. In the last few days before birth, proteolytic enzymes secreted by the fetal membranes of the embryo break down the shell membrane (Denker and Tyndale-Biscoe 1986), and an intimate connection is made between the membranes and the uterine lining to form the yolk sac placenta. This pattern holds for all marsupials: the period before breakdown of the shell membrane ranges from 8 to 23 days, and much longer when embryonic diapause occurs, whereas after the shell membrane breaks down the rest of gestation is brief, varying from four to five days in bandicoots and opossums to 8–10 days in macropods. The reason for the persistence of the shell membrane has long been a puzzle. One idea is that it protects the developing embryo from exposure to maternal antibodies that would otherwise be directed at the embryo’s proteins, which are genetically different from the mother’s. The amount of time after the shell membrane breaks down and the fetal tissues are exposed is sufficiently short to avoid such a reaction by the mother’s immune system and consequent rejection of the fetus. Evolutionary biologists have suggested that this was the key difference between marsu- pials and placental mammals: placental mammals evolved a way to avoid maternal rejection of the fetus, which opened the way for prolonged gestation whereas marsupials did not and were constrained to have a very brief gestation and deliver very immature young (Amoroso and Perry 1975). One way to test this idea is to sensitise the female marsupial to the antigens of the prospec- tive father, by giving her two successive skin grafts from him: foreign skin is very antigenic, and since the fetus will inherit paternal antigens, this should provoke a strong reaction at the end of gestation. When we did this with tammars the females developed very strong immune responses to the male, rejecting the second skin graft within a few days, as was to be expected 60 Life of Marsupials

(Laplante et al 1969). Nevertheless, they subsequently conceived a succession of offspring with the male to whom they were sensitised, and some females produced young to the same male for five years without any reduction in their fertility (Walker and Tyndale-Biscoe 1978, Rodger et al 1985). These experiments therefore do not support the idea that the shell membrane protects the female from male transplantation antigens and leaves open the question of what the function of the shell membrane is. Another possible function is that it is a resilient container during the formation of the fetal membranes, rather as the zona pellucida is for the early development of the blastocyst. While

Figure 2.10: Fetal membranes in four marsupials to show the different proportions of the yolk sac, allantois and amnion in late pregnancy. From the top they are the quokka, Setonix brachyurus, , Dasyurus viverrinus, koala, Phascolarctos cinereus, and long-nosed , Perameles nasuta. Thin lines represent ectoderm, dashed lines, endoderm, and thick lines, mesoderm. After Tyndale-Biscoe and Renfree (1987). Reproduction and development 61

studying tammar embryonic vesicles growing in a test tube, from which the shell membrane had been removed, we made a curious discovery: instead of the amnion growing over and enclosing the embryo, as occurs in the uterus, the folds of amnion remained on each side of the embryo like two large cushions. Without the constraint of the shell membrane the folds of the amnion were apparently unable to press the embryo down into the yolk sac and close over it and so the embryo remained exposed. It is interesting that in all marsupials the amnion envelops the fetus before the shell membrane breaks down.

Development of the yolk sac, amnion and allantois While the main form of the embryo is being laid down, mesoderm is spreading out beyond the limits of the embryo plate, until it extends nearly half way around the yolk sac. However, it never reaches the whole way, so that the marsupial yolk sac wall is two layered in its lower half and three layered in its upper pole (Fig. 2.10). As blood vessels develop only from the mesoderm, the three-layered part becomes the vascular yolk sac (Fig. 2.9b, Plate 3) and the main respiratory organ of the fetus, while the two-layered, non-vascular yolk sac is the main route of absorption of uterine secretions: together these comprise the yolk sac placenta of marsupials. Within the limits of the embryo and for a short distance beyond, the mesoderm splits into two layers, one being applied to the ectoderm and the other to the endoderm. The space between these layers is the coelom or body cavity. Folds of the outer layer now rise up at the head end and sides of the embryo and enshroud it within two membranes, an outer chorion and an inner amnion, which encloses a fluid-filled space, the amniotic cavity in which the developing fetus now lies (Fig. 2.9c, Plate 3, Fig. 2.10). Once again, this manner of forming the amnion and chorion closely resembles birds, reptiles and monotremes and in all of them the embryo is contained within a shell membrane when the amnion is formed. In all reptiles, birds and mammals (amniote animals) another sac grows behind the yolk sac: this is the allantois (Fig. 2.9c, Plate 3), which in birds, reptiles and monotremes becomes the main respiratory surface for the embryo in the egg. In placental mammals the allantois retains this function and also assumes the major nutritive role as the fetal component of the definitive, allantoic placenta. In most marsupials, however, the allantois is an inconspicuous sac with a modest vascular supply buried in the enfolding yolk sac (Fig. 2.10); and its main function is to store urine excreted by the fetal kidneys in the last days of gestation. In a few species, such as the koala, it is larger and in the bandicoots it becomes highly vascular, its cells fuse with cells of the uterus, and it forms an allantoic placenta, which functions for the last three days of gestation (Fig. 2.10). In summary, marsupials have developed the yolk sac as the main organ of exchange between the fetus and the mother (a yolk sac placenta) and the placental mammals have developed the allantois as the main organ of exchange (the allantoic placenta). Notwithstanding this main distinction, some placental mammals, such as the rabbit, have a yolk sac placenta in the early part of gestation and some marsupials have an allantoic placenta in the last days of gestation.

Placental transport in the tammar In the tammar, as we have seen, the protein constituents of the uterine secretion differ from serum in several particulars, which is further evidence that it is not a simple exudate but is the product of active secretion (Fig. 2.4). The outer surface of the yolk sac wall of the tammar is seen under the electron microscope to carry a thick weft of microvilli with pinocytotic vesicles at their bases (Tyndale-Biscoe and Renfree 1987). Mitochondria are numerous, as also is rough endoplasmic reticulum, all of which indicate transport and metabolism of uterine secretions. Later, when the shell membrane breaks down, the yolk sac membrane has direct contact with the uterine epithelium and the microvilli of the two tissues interdigitate and establish an intimate 62 Life of Marsupials

contact: although younger embryos can be rolled unharmed from the opened uterus, this is not possible after the shell membrane has gone. This intimate contact of the tissues brings the fetal and maternal blood vessels very close together, thereby facilitating exchange between them. These histological changes are reflected in changes in the constituents of the yolk sac fluid.

Functions of the yolk sac placenta During the pre-attachment phase, nutrient and respiratory requirements must be met by exchanges with uterine secretions across the yolk sac membrane and at this stage the concentra- tions of glucose and urea in yolk sac fluid resemble those in maternal serum (Fig. 2.11) (Renfree 1973b). The protein concentration in yolk sac fluid is much lower than in serum and is so similar to uterine secretion, which comprises just albumin and pre-albumin (Fig. 2.4), so it is very likely that the great increase in the volume of the yolk sac at this stage is caused by direct uptake of uterine secretion across the yolk sac membrane. After attachment there are big changes in the composition of yolk sac fluid. The protein contents now resemble those of serum, although there has been no change in the composition of the uterine secretion, the concentration of glucose increases steadily (Fig. 2.11) and the colour of yolk sac fluid changes from clear, through straw to yellow by full term. The yellow colour results from breakdown products of haemoglobin and reflects the development of the fetal liver in which this takes place. By contrast allantoic and amniotic fluids remain clear and contain the same limited number of

Figure 2.11: Glucose and urea content of yolk sac fluid (O) and allantoic fluid (O) through pregnancy in the tammar wallaby, Macropus eugenii. Shaded regions represent the range of concentrations in maternal serum. After Renfree (1973b). Reproduction and development 63

proteins as formerly (Fig. 2.4). Glucose concentration in the allantois is low but the urea concentra- tion increases progressively to four times that of the other compartments and of fetal serum: this coincides with the development of the fetal kidneys as excretory organs. The marked changes in the yolk sac fluid after attachment could result from greater ease of transport of maternal proteins across the contiguous circulations, or they may reflect the growing maturity of the fetal systems, such as the liver and kidneys. The evidence supports the second conclusion: when maternal proteins, labeled with an isotope of iodine, were injected into the maternal circulation in late pregnancy almost none of the label was later detected in the yolk sac; and iron-binding proteins, called transferrins in yolk sac fluid have a different mobility to transferrins in maternal serum, indicating that they have been synthesised by the fetus itself. Thus, fetal membranes in the final days of gestation are actively controlling the transfer of substances found in the fluid compartments, reflecting the growing autonomy of fetus as it reaches the time to be born.

Parturition The birth of a marsupial is an extraordinary phenomenon. The newborn is so small and so undeveloped that it belies belief that it could be anything but passive in the events that take it from the uterus, through the temporary birth canal and thence from the cloaca to the pouch and attachment to a teat. For as long as it has been known many explanations have been entertained for how marsupial parturition is achieved. Even after Hartman (1920) established that the young make the journey to the pouch unaided by the mother, direct involvement of the young in the process of parturition was still discounted. About 30 years ago we were still convinced that no special endocrine changes were associated with parturition: now we know that the tiny fetus is as much involved in the events leading to its birth as is the lamb or the human baby.

Role of the corpus luteum in parturition The single most important organ in the preparation of the female tammar for birth is the corpus luteum of pregnancy. If the corpus luteum is surgically removed after day 6, the pregnancy will continue through to full term but the fetus is not born: the fetus leaves the uterus but is then impounded in the median vagina, where it dies. If the corpus luteum is removed on or after day 23 (of the 26-day gestation), however, the fetus is born and reaches the pouch but dies one day later, presumably because the mammary gland is unprepared for lactation (see Preparation of the mammary glands, or mammogenesis) (Young and Renfree 1979, Harder et al 1984). Progesterone is the main hormone that prepares the genital tract and the mammary gland, since injections of progesterone will substitute for the absent corpus luteum. Without proges- terone the temporary canal through which the young must pass after it leaves the uterus is small or absent, whereas progesterone softens the tissues, making it possible for the young to pass through, still enclosed in its fetal membranes. This is an important point because it was earlier thought that the young might force its way through, using its well-developed arms. Clearly it is unable to do this while still in the amnion. The other hormone secreted by the corpus luteum is relaxin, but its role in preparing the genital tract is still unclear. The peak production of relaxin is three days before birth and it declines rapidly after birth, so the role is inferred (Tyndale-Biscoe 1981, Parry et al 1997b). Relaxin may act with progesterone to soften and dilate the cervix, as it does in other mammals, and it may be involved in preparing the birth canal, but critical experi- ments to test this have not been done. Because in the tammar and other macropods, birth is followed after a few hours by oestrus, oestrogen might also play a role in birth. However, removal of the Graafian follicle, the source 64 Life of Marsupials

of oestrogen at this time, during late pregnancy did not prevent parturition at the normal time, although it did abolish post-partum oestrus (Harder et al 1985). This result confirms the earlier work of Newsome (1964), who observed that red kangaroos could give birth in the absence of growing follicles, and consolidates Poole’s (1975) study of grey kangaroos that can give birth without ever having a post-partum oestrus.

Release of the young from the uterus The first hint that the marsupial fetus might be involved in determining the time of its birth came from a study of hybrids between the two species of grey kangaroos by Kirsch and Poole (1972). The pregnancy of the eastern grey kangaroo, Macropus giganteus, is 36 days while that of the western grey kangaroo, Macropus fuliginosus, is 31 days. However, when eastern grey females were crossed with western grey males, pregnancy was 34 days, indicating that the genetic status of the hybrid fetus had shortened it. When the female hybrids grew up and were themselves mated to western grey males, their pregnancy was further shortened to 31.6 days. In tammars Merchant (1979) discovered that the interval from oestrus to post-partum oestrus in pregnant tammars was significantly shorter than the interval from oestrus to oestrus when the same tammars were not pregnant, which further implicated the fetus and/or the placenta in controlling the events around birth. More direct evidence of the fetal role came later when it became possible, simultaneously, to measure six hormones in circulation around the time of birth in tammars, compared to the same hormones circulating at the equivalent time in non-pregnant tammars. These results clearly showed that the pattern and occurrence of all the hormones differed between the two states. Progesterone falls rapidly at the time of birth but gradually over several more days in the non-pregnant cycle. Large but highly transient peaks of the hormones prolactin, prostaglandin and mesotocin occur close to the time of birth (Fig. 2.12), whereas none of these hormones is detectable in the non-pregnant tammar at the equivalent time. Since these hormones are associated with parturition in placental mammals it was interesting to find that they are also associated with parturition in the tammar. Within one hour of giving birth the female tammar comes into oestrus, which is associated with high levels of oestrogen and is followed 24 hours later by a sharp peak of LH, followed by ovulation. In the non-pregnant female the same sequence is followed but two to three days later. The questions are the respective roles of each hormone and how the fetus is involved in their secretion. One way to test the importance of the fetus is experimentally to advance its development ahead of its mother’s. Clark (1968) did this in red kangaroos by injecting lactating females with progesterone, which mimicked the action of the corpus luteum and reactivated their diapausing blastocysts (Fig. 2.8); 10 days later she removed the young from the pouches of their mothers, which initiated development of the corpus luteum in their ovaries. If the mother determines the time of birth, the young should not have been born until 31 days after remov- ing the pouch young, whereas if the fetus determines the time it should have been born 10 days sooner: the latter result occurred. This design was repeated some years later in tammars when the several hormones were also measured (Tyndale-Biscoe et al 1988). As with red kangaroos, the tammars gave birth early and the pre-partum pulse of prolactin occurred at the time of birth, three days earlier than in control animals, again showing that the fetus is influencing the hormonal events. There is now good evidence on how the tammar fetus influences the cascade of events around the time of birth (Fig. 2.13). Briefly, the pituitary gland in the head of the fetus stimu- lates the fetal adrenal gland to produce the hormone cortisol, which in turn causes a massive release of prostaglandin from the yolk sac placenta and endometrium of the uterus. This in turn Reproduction and development 65

Figure 2.12: Profiles of six hormones in circulation around the time of parturition in the tammar wallaby, Macropus eugenii. The precipitate fall in progesterone and the very brief pulses of prostaglandin, mesotocin and prolactin only occur in the presence of a fetus and not in non- pregnant tammars at the same time after oestrus. From Shaw and Renfree (2001) Reproduction, Fertility and Development 13, 657, with permission of CSIRO Publishing. 66 Life of Marsupials

provokes the release of prolactin and mesotocin from the mother’s pituitary gland. Together the prostaglandin and the mesotocin stimulate contractions in the uterine muscle, and the fetus is expelled through the cervix into the median vagina and thence to the outside. The prostaglandin also provokes the whole repertoire of the birth position and behaviour by the mother and the mesotocin and prolactin prepare the mammary glands for the start of lactation (for details see Boxes 2.1–2.4).

Figure 2.13: Diagram to show how the hormones in Figure 2.12 interact during birth. While the corpus luteum prepares the cervix and birth canal for passage of the young, and the mammary glands for lactation, it is the fetus that starts the whole cascade of events, through four steps of amplification: fetal pituitary; fetal adrenal; placenta and endometrium; and maternal pituitary and brain. Reproduction and development 67

Box 2.1: Role of prostaglandins during birth Shaw (1983) showed that prostaglandins stimulate contractions in the uterus of the tammar, these contractions become more marked as the time of birth is approached and are larger in the pregnant than in the non-pregnant uterus. Shaw et al (1999) showed that two kinds of prostaglandins are synthesised in the placenta and endometrium in late pregnancy. They cultured pieces of each tissue in vitro and measured the production of the active forms and their metabolite, , PGFM. The active form of prostaglandin, PGF2_ is produced in the endometrium of the pregnant uterus from day 18 and the rate increases three-fold by the day of birth. It is also produced in the non-pregnant uterus but does not increase in a consistent way and its contribution to the total output of prostaglandins is probably negligible. However, the production of PGF2_ by the fetal membranes is considerably higher and increases at a faster rate. The production rate in none of these tissues is affected by mesotocin or the cortisol-like hormone, dexamethasone, so the inference is that neither the posterior pituitary nor the fetal adrenal gland are involved in the increased rate of production. PGF2_ is metabolised to another form called PGFM, which is the form that is measured in the circulation around the time of birth. PGFM is not produced by the fetal membranes and only to a very small extent by the endometrium, so the origin of the

PGFM in the blood is not clear. It is possible that the PGF2_ is released into the circulation and is metabolised to PGFM in the lungs. Whatever the process, PGF2_ is essential for parturition to occur and for the female to adopt the birth position, because both events are prevented when pregnant tammars are injected with an antagonist of PGF2_, indomethacin, from day 24 to 28

(Renfree et al 1994). PGF2_ also induces the release of prolactin from the maternal pituitary (Hinds et al 1990) and causes the precipitous fall in progesterone from the corpus luteum (Renfree et al 1994).

Box 2.2: Role of mesotocin during birth Mesotocin is a peptide molecule of nine amino acids, very similar to the pituitary hormone oxytocin secreted by the posterior pituitary of placental mammals. Indeed, it only differs by one amino acid and has the same functions in kangaroos that oxytocin does in placentals, namely to stimulate contractions in the uterine muscle and stimulate milk let down. This is supported by the results of surgically removing the posterior pituitary from pregnant tammars: they failed to give birth and the fetuses were found dead in the uterus (Hearn 1974). Mesotocin is found in the posterior pituitary of marsupials (Bathgate et al 1995) and there is a marked pulse of mesotocin at parturition (Parry et al 1996), which almost certainly comes from the maternal pituitary. The link to the uterus at parturition is supported by the presence, in the muscle tissue surrounding the pregnant uterus, of specifi c receptors that bind mesotocin (Parry et al 1997a). The concentration of these receptors was higher in the pregnant than in the non-pregnant uterus and reached a maximum just before the time of birth, suggesting that the presence of the fetus may stimulate synthesis of the receptor protein. Renfree et al (1996b) tested the role of mesotocin by infusing an antagonist of its receptors into tammars on the last three days of pregnancy. The effect was to delay birth but not prevent it taking place. These separate pieces of information all support a role for mesotocin at the time of birth by stimulating contractions of the gravid uterus and so releasing the fetus for the start of its journey to the pouch. As the peak of mesotocin in circulation occurs after the young has been born, it may also be important in the initiation of lactation. 68 Life of Marsupials

Box 2.3: Role of prolactin during birth The maternal pituitary also secretes the transient pulse of prolactin but its timing is clearly infl uenced by the fetus, since it only occurs in pregnant tammars and was advanced when the fetuses were experimentally advanced. However, its role is still unclear: when fi rst discovered it was thought to be the signal that initiated birth (Hinds and Tyndale-Biscoe 1985) but when Fletcher et al (1990) tested this by treating pregnant tammars with the prolactin antagonist, bromocriptine, which suppressed the pulse, birth was not delayed but the young died soon after birth. Like mesotocin, the main role of the prolactin pulse may be after birth, preparing the mammary gland for imminent lactation.

Box 2.4: Role of the fetus during birth Despite its tiny size and immaturity, the tammar fetus has surprisingly well-differentiated pituitary and adrenal glands. Leatherland and Renfree (1983) described cells in the pituitary four days before birth, that appeared to be capable of secreting hormones, such as adrenocorticotrophic hormone (ACTH), which controls the activity of the adrenal cortex. The adrenal cortex of the tammar fetus was suspected to be active many years ago when Catling and Vinson (1976) measured cortisol in the blood of two full-term fetuses. Shaw et al (1996) tested the importance of cortisol in initiating parturition by treating pregnant tammars with a related steroid, dexamethazone, which caused them to deliver their young a day earlier than control tammars treated with saline. While the young were signifi cantly lighter than the controls the interesting point was that the other hormones, PGFM, prolactin and progesterone, underwent their normal sequence of changes a day earlier than the control tammars. This strongly suggested that cortisol from the fetus was acting as the signal to start the cascade of events leading to parturition. Now Ingram et al (1999) have demonstrated that the source of cortisol at parturition is indeed the fetus: from day 24 to day 26 the adrenal cortex of the fetus increases in size and the content of cortisol to a peak just before birth, as does the concentration of cortisol in yolk sac, allantoic fl uid and fetal blood plasma. More importantly, the production of cortisol by fetal adrenal tissue in culture was increased three-fold

in the presence of ACTH or prostaglandin (PGE2). This was the evidence needed to show that the fetus is capable of timing its own delivery.

We can now try to put these several pieces together in the order in which they occur in the living tammar (Fig. 2.13). While the cervix and birth canal are being softened by elevated proges- terone and relaxin from the maternal corpus luteum, the fetal pituitary is maturing sufficiently to secrete ACTH, the fetal adrenal is enlarging and beginning to synthesise cortisol, and the yolk

sac placenta is synthesising prostaglandin PGF2_. Cortisol secretion accelerates towards the day of birth under the influence of fetal ACTH and prostaglandin. Fetal cortisol may influence the release of prostaglandins from the yolk sac and endometrium, which starts muscle contractions of the uterine wall and the evacuation of the fetus. The rapid rise of PGFM in maternal circula- tion as birth occurs may cause the immediate release of mesotocin from the mother’s posterior pituitary gland, which enhances those contractions but is not essential for birth. PGFM also induces the prolactin pulse from the mother’s anterior pituitary gland and the fall in progester- one from the declining corpus luteum. This sequence and the several interactions are very similar to the events that occur at parturition in the best-studied placental mammals – sheep and humans. In both these placental species the fetal pituitary and adrenal cortex are central in the onset of parturition, through their interplay with placental and uterine prostaglandins and oxytocin. But in both species Reproduction and development 69

the fetus is far more advanced in its development at birth than a tammar is, and it is easier to comprehend that it could have well-developed hormonal controls in place. It has been much harder to appreciate that a fetus of 200–400 mg is equally capable of controlling its own birth, and in the same way. Now consider that the tammar at birth is one of the largest marsupial neonates (see Fig. 1.2). Do all marsupial neonates have the same capacity to control the time of their birth? The young of dasyurids weigh less than 20 mg at birth and those of the honey possum weigh 4 mg! What little has been found out about these matters in other species of marsupial seems to be in accord with the tammar: bandicoots, brushtail possums and northern , Dasyurus hallucatus, have well-developed pituitaries and adrenals at birth (Gemmell and Nelson 1988), the cells of which contain secretory granules similar to those in hormone-secreting glands. In the bandicoot a large pulse of PGFM occurs around the time of birth (Gemmell et al 1980) and injections of prostaglandin induce birth behaviour in several species. This points to the likelihood that fetuses of other marsupials are similar to the tammar fetus in being able to control events at birth. But if one thinks the achievements of the prenatal marsupial are remarkable, consider what it does after it is born.

Maternal behaviour during parturition Birth has now been observed and described for many species of marsupial and the story of its discovery often told. In most species the young makes its way from the cloaca to the pouch by its own efforts within minutes and the mother does not directly help it. However, she does adopt a particular posture and pattern of behaviour that aid the young in its journey. Detailed observa- tions of the entire behaviour leading up to birth in the red kangaroo and the tammar wallaby have been made and both recorded on film. In the tammar the sequence of events can be related to the hormonal changes outlined (see Parturition) (see Renfree et al 1989 for an illustrated account). The pouches of non-lactating kangaroos and tammars contain a brown, dry scale and in the last week of pregnancy this is removed by the female putting her muzzle into the pouch, while holding the sides open with her forepaws. The pouch appears clean and moist and a small bud develops on the end of each teat. Non-pregnant kangaroos at the same stage of the cycle also clean the pouch in the same way. Since in both states this is the stage of the oestrous cycle when the progesterone level is elevated and before the levels of other hormones have risen, it is likely that increased progesterone provokes this behaviour. While cleaning the pouch the female may adopt the so-called birth posture, in which she sits on the butt of the tail with the tail and hind legs extended forward, so that the cloaca is directed upwards. In the hour or so preceding birth, pregnant animals adopt this posture much more frequently than non-pregnant animals and the intensity of pouch cleaning increases greatly. At the same time the animal begins to lick the cloaca as well. The physiological state of the pregnant female also changes: alertness is lost, shivering and whole body spasms occur and she is undisturbed by outside events; she can be picked up without showing alarm and when released will immediately resume the birth posture and persistently lick between the cloaca and the pouch.

The discovery that PGF2_ would induce this intense behaviour was made by chance while trying to work out the respective roles of prolactin and PGF2_ for the rapid decline of proges- terone at the end of pregnancy (Hinds et al 1990). Non-pregnant females at the end of the oestrous cycle were injected with PGF2_ and almost immediately went into the full repertoire of birth position and behaviour and remained like that until the decline of hormone in the blood (Fig. 2.14). Other females injected with prolactin showed no change in behaviour. Subsequently Shaw (1990) showed that the intensity and duration of the behaviour was directly related to the level of prostaglandin in circulation and that it could be induced in females that had never been 70 Life of Marsupials

Figure 2.14: Two non-pregnant female tammar wallabies, Macropus eugenii, that adopted the birth position a few minutes after receiving an injection of prostaglandin into a tail vein.

pregnant and, more remarkably, in adult male tammars. The males adopted the birth position and intently licked the scrotum and genital region! These results show that birth behaviour is

not learned but is ‘hard wired’ in the brain. In other marsupials that have since been tested PGF2_ has the same effect on birth behaviour. Indeed, this may be a very ancient function of PGF2_, since it also induces spawning behaviour in fish and sexual behaviour in a wide range of other vertebrates. From the less complete observations on other marsupials birth appears to be much the same as for the red kangaroo and tammar wallabies. The quokka, brushtail possum, koala and Virginia opossum adopt the same posture as the red kangaroo and tammar but grey kangaroos give birth in a standing position. In the small dasyurids, Antechinus swainsonii and Dasycercus byrnei, and in the , the female stands with its hips raised so that the young travel down from the cloaca to the teats: in the latter species two gelatinous strands are ejected from the uteri in which the emerging young pass towards the pouch area, where each one grasps a hair and wriggles to a teat (Nelson and Gemmell 2003). In the , Isoodon macrourus, the female lays on her side with the upper leg raised and the cloaca brought close to the backward opening pouch in such a manner that when the young appear they almost fall into the pouch, still attached by the elongated umbilical chord (Gemmell et al 1999). In the kangaroo and tammar, the onset of birth is heralded by a flow of fluid at the cloaca from the ruptured yolk sac. This is immediately licked up and is followed, in the red kangaroo, by the appearance of the intact allantois, which may fall to the ground. Then the young emerges head first, still enclosed in the fluid-filled amnion, from which it now frees itself with its swing- ing, clawed arms. In the tammar the young also emerges in the amnion but it is followed by the allantois and the remaining yolk sac membranes, trailing away from the umbilicus, from which it breaks free, partly aided by the mother’s persistent licking. If the mother is undisturbed, the young orientates itself towards the pouch and rapidly moves there by grasping the fur in its claws Reproduction and development 71

and by alternate movements of the forelimbs. At the same time the head turns from side to side, so that the muzzle describes an arc. Within a few minutes, usually less than five, it has gained the lower lip of the pouch and thence passes from view. If the mother has not been suckling a previous offspring, all four teats are available, otherwise three are. These have developed a small bud at the tip, which the young now draws into its mouth. For two days after birth the young can be removed from one teat and placed on another and it will suckle successfully, but as time passes this fails because only the suckled gland secretes milk (see Journey of a lifetime – adapta- tions for reaching and attaching to a teat). Similarly, young transferred to the budded teat of a non-pregnant female will attach and be suckled successfully until weaned. In tammars it is possi- ble, experimentally, to place a newborn young on each of the four teats and all four young will grow for a few weeks, showing that all the teats have the potential to support a young at the time of birth. Of course, in other species that have multiple young at birth, all the teats will usually be occupied. In some species, such as the Virginia opossum, the number of young born often exceeds the number of teats and those that fail to grasp a teat die.

Journey of a lifetime – adaptations for reaching and attaching to a teat The marsupial at birth is a marvelous composite of embryonic structures and precociously developed functional organs (Figs 2.9d, Plate 3, Fig. 2.15). The head, shoulders and forelimbs are relatively large but the hind legs are mere paddles, and the hips and tail are small. The ears and eyelids are shut but the mouth and tongue are large, the nostrils are open and the blood vessels beneath the skin are prominent. These precociously developed organs enable it to reach the pouch unaided, to respire, to attach to a teat and gain nourishment from the mammary gland.

Locomotion – forelimb skeleton, musculature and innervation The forelimbs and shoulders are well developed and the digits are armed with sharp recurved claws, with which the newborn grasps the mother’s fur. These claws are extensions of the outer keratin layer of the skin and are shed a few days later: they are not the definitive fingernails, which appear later in development. The skeleton of the shoulder region is a single cartilage, comprising the future scapula, glenoid, coracoid and sternum, which provide firm support for the forearm during the journey to the pouch (Fig. 2.15b) (Klima 1987). This is similar to the arrangement in adult monotremes and reptiles. A week after birth the coracoid separates from the sternum and is reduced to a small nubbin, so that the scapula is only loosely joined to the sternum by the clavicle, as in adult placentals and marsupials. One hundred years ago, before it was appreciated that the young travels unaided to the pouch, this anatomy was seen as a vestige from the reptilian ancestry of marsupials, rather than a functional adaptation. However, the relatively large coracoids do not support large coraco- brachial muscles, as these bones do in monotremes and reptiles: the muscles of the shoulder and forelimb of the newborn marsupial have the adult mammalian form and are innervated by a large brachial nerve (Fig. 2.15a). The muscles of the neck and thorax are also well developed, and the young animal carries out a series of alternate lateral contractions, starting at the head and ending with the forelimbs. What controls these movements? Langworthy (1925) showed that the movements continued normally after he removed the cerebral hemispheres of newborn Virginia opossums, so the control centre must be in the brain stem or spinal cord. The same probably holds for other marsupials, since few of the neurons that will constitute the cerebral cortex have migrated there at the time of birth (Reynolds and Saunders 1988). Furthermore, the special tract of fibres that carries signals from the cerebral cortex to the spinal cord, called the pyramidal tract, does not develop until the week 5 after birth in the opossum (Ward 1954). So how is this complex and highly coordinated series of 72 Life of Marsupials

Figure 2.15: Anatomy of the newborn marsupial: (a) diagram of the head and forearm of a newborn tammar wallaby, Macropus eugenii, to show the innervation of the nose by cranial nerve I; the innervation of the lips, jaws and tongue by the trigeminal ganglion (cranial nerves V and VII); and the innervation of the forearm by the brachial plexus (redrawn after Hughes et al 1989); (b) the shoulder girdle of the newborn brushtail possum, Trichosurus vulpecula, to show how it is united with the sternum on the day of birth but has separated from it 14 days later (redrawn after Klima 1987).

movements controlled? Is it entirely controlled at the level of the brachial plexus in the spinal cord, or are higher centres in the brain stem involved? Ho (1997) and Ho and Stirling (1998) have investigated the neural control of the alternat- ing opening and closing of the hands and movement of the forelimbs in the newborn tammar, which they call clock-like and which is retained in the later stages of pouch life. While the young tammar was attached to a teat and warm the movements ceased but if it was cooled or removed from the teat, the clock-like alternating movement resumed. When one limb of the newborn was Reproduction and development 73

restrained the movement of the other became irregular, which indicates that there is coordina- tion of the two sides in the spinal cord. The information about this is probably conveyed to the brachial plexus via sensory nerves from stretch receptors in the forelimbs. However, information about the limbs cannot go further because sensory nerve connections from the brachial plexus to the brain stem have not developed at this stage. When Ho isolated the brain stem and spinal cord of the newborn tammar in a water bath the alternate firing of the left and right brachial motor nerves continued, thereby showing that control of rhythmic limb movement is located in the spinal cord. In addition, motor nerve connections from the brain stem to the brachial plexus can modulate the movements of the limbs. This means that the information needed to find the pouch and attach to a teat can be provided from sensory input from the mouth. Hughes et al (1989) strikingly demonstrated this with freely held newborn tammars: when their lips were gently stroked the clock-like movements immediately ceased.

Sense organs that function at birth – smell, touch and balance The behaviour of the newborn marsupial suggests that three sensory systems are functional at birth, namely smell, touch and balance. The evidence for a sense of smell or touch is the way marsupial young redirect their movements when they reach the edge of the pouch and enter it. This indicates a response to the texture of the naked, moist skin in the pouch or to a particular smell emanating from the pouch. Sensory cells, called olfactory knobs, have been described in the nasal epithelium of three species (Gemmell and Rose 1989) and well-developed nerve fibres connect to the relatively large olfactory lobes of the brain (Hill and Hill 1955). Touch receptors, called Merkel cells, with associated nerve cells, have also been described in the skin surrounding the mouth of the newborn of six species, representing five families (Gemmell et al 1988, Hughes and Hall 1988). The trigeminal nerve plexus (cranial nerve V), a prominent feature of the newborn (Fig. 2.15a), conveys sensory information from the mouth, lips and tongue to the brain stem. Its motor fibres, together with those from cranial nerve VII, control the movements involved in sucking and breathing (Hughes et al 1989). The trigeminal plexus enters the brain stem close to the motor fibres that pass down to the brachial plexus, so it is highly likely that it is involved in the coordination of the sensory input from the nose and lips and the movements of the trunk and forelimbs. Hughes and Hall (1988) and Hughes et al (1989) called this complex system the ‘locomo- tor generator’, which is most highly developed in the newborn tammar, less developed in the newborn brushtail possum and least developed in the newborn quoll. Since much less is known about newborn marsupials other than the tammar, it is not possible to correlate the anatomical level of development with neonatal behaviour. For instance, is the less developed condition of the northern quoll associated with Nelson and Gemmell’s (2003) observations that the newborn, still enclosed in the amnion, are aided in their journey to the pouch by the gelatinous strands from the mother, rather than making the journey to the teat by their own efforts? They have noted that, unlike tammar newborn, these tiny creatures cannot repeat the journey if returned to the cloaca.

Does the newborn have a sense of gravity? The question of whether the newborn marsupial has a sense of gravity has been more conten- tious. Because the young one usually climbs upwards to enter the pouch it was thought that it must orientate according to gravity. However, the first studies of the Virginia opossum led McCrady (1938) to conclude that neither the utricle nor semicircular canals of the inner ear were sufficiently developed to provide any information on gravity to the newborn. He suggested that the young orientates itself by the peculiar geometry of its body at birth, the passive hind 74 Life of Marsupials

part of the body hanging down between the two large forelimbs. Cannon et al (1976) supported McCrady’s observations in experiments with newborn . They found that the young invariably orientated themselves upwards and, when the mother was laid on her side, so that the opening of the pouch was not directly above the young, it passed within 10 mm of the pouch and went on towards the mother’s chest. They concluded from this that olfaction is not important, and that gravity is the sole influence directing the young to the pouch, although they did not know how gravity was being sensed. Since then the structure of the inner ear of the newborn brushtail possum, brushtailed bettong and northern quoll have been studied with the electron microscope (Gemmell and Nelson 1989b, Gemmell and Rose 1989) and all three species have a well-differentiated sensory, ciliated epithelium on the inner surface of the vesicle that will become the inner ear, and lying on the epithelium is a cluster of crystalline granules or otoliths. This structure is probably capable of providing information to the newborn marsu- pial about gravity and thus position, although the neural connection to the brain has not been established.

Adaptations for sucking – lips and hyoid musculature The buccal cavity is large and the tongue is large and muscular. After the young becomes attached to the apical bud the teat expands inside the mouth and the lips and tongue grow around it, so that the young becomes firmly attached and can be removed only with difficulty. This arrange- ment also allows the young to develop a negative pressure in the buccal cavity by depressing the floor of the chamber with the hyoid muscles and so suck milk from the teat. The large trigeminal nerve supplies these muscles also (Fig. 2.15a). Griffiths and Slater (1988) demonstrated that newborn tammars and red kangaroos are able to suck, by offering them a fine pipette filled with warm water or milk and held so that the pipette was below the mouth of the young: all these young animals sucked up 20–30 mg of fluid in about two to three minutes and the fluid in the pipette exhibited motions that indicated that the young were sucking. This disposed of an old belief that the marsupial young is unable to suck and that the milk is forced into it by contrac- tions of a special muscle, the ilio-marsupialis, which penetrates the mammary gland. However, when this muscle was electrically stimulated, no milk was expressed from the teat (Enders 1966, Griffiths and Slater 1988). This muscle has a different function, which will be discussed later.

Swallowing milk – epiglottis and stomach Because the young is permanently attached to the teat for several weeks it might be thought that the flow of milk to the gullet would interfere with breathing, especially as the air from the nostrils must cross the gullet to reach the trachea. However, newborn marsupials have a special arrangement that avoids this: at the back of the mouth the epiglottis is large and extends through the soft palate so that the glottis opens into the nasopharynx. The buccal cavity extends around each side of it and communicates with the oesophagus, so that the young can feed and breathe simultaneously. The stomach and duodenum are in an advanced stage of development at birth and so is the pancreas. By contrast, the small intestine and colon are not.

Respiration via skin and lungs: surfactant and the role of cortisol Baudinette et al (1988b) have studied the metabolism of the newborn tammar as it makes its first journey. The effort required to reach the pouch is considerable and the young has energy reserves sufficient for little more than one journey. As in any newborn mammal the first require- ment for respiration is the switch from placental exchange to air, and the rearrangement of the great vessels entering and leaving the heart to convey oxygenated blood from the lungs instead of from the placenta. This involves the separation of the right and left halves of the heart by the Reproduction and development 75

formation of a wall between the two atria and between the two ventricles. It also involves closure of the connecting artery (ductus arteriosus) between the pulmonary artery, taking blood to the lungs, and the dorsal aorta taking blood to the rest of the body. In placentals, such as a baby, these three events occur immediately after birth, at the first breath; if they do not, the baby goes blue through lack of oxygenated blood. In marsupials the process is slower and in the tammar closure is not complete until day 3 (Runciman et al 1995). As the young tammar is born, blood flow from the yolk sac placenta ceases and blood flow to the lungs increases, as they fill with air. Since the circulation to the lungs is not separate from that to the rest of the body during the first hours after birth, mixing of oxygenated blood from the lungs with blood from the rest of the body must occur until day 3. Is this difference from placental mammals because the young of marsupials are so small that respiratory efficiency is less important after birth, or is it because there is another route for gas exchange – the skin? The newly born brushtail possum, kangaroo and tammar are a fiery red colour on the first day after birth and they have very conspicuous subcutaneous blood vessels (Fig. 2.9d, Plate 3) and a moist skin. John Shield (pers. comm. in Richardson and Russell 1969) noted that newborn quokkas paled in an atmosphere low in oxygen and recovered their bright red colour when returned to an oxygen-rich atmosphere. If respiration can take place across the skin as well as the lungs, there is clearly an advantage in allowing the blood to be mixed in the heart. During the journey to the pouch, when the demands for oxygen are great and the lungs are filling with air for the first time, gas exchange across the skin may well be important, especially in the smallest marsupials. Baudinette et al (1988a) discount the importance of this in the tammar for two reasons: by day 3 after birth the distance between the air and the blood in vessels under the skin was 100 times greater than the distance across the lung epithelium (<1 µm) (Randall et al 1984), and the relative uptake of oxygen through the skin was found to be only 2% of uptake across the lung. They acknowledge that the skin route may be important for release of carbon dioxide, which diffuses through tissues 20 times more readily than oxygen. Since all their obser- vations were done at day 3, when the circulation has become separate, the importance of cutane- ous respiration on day 1 remains an open question in the tammar. In the , Sminthopsis douglasi, which at 13–15 mg is far smaller than the newborn tammar (360 mg), the main source of oxygen after birth is indeed across the skin (Mortola et al 1999). For the first three days after birth gas exchange across the skin was 10 to 90 times higher than across the lungs, dropping to equality by day 10. At day 20 the cutaneous route still accounted for one-third of the oxygen consumed, and by then the young weighed 290 mg, which is still smaller than the newborn tammar. Thus, cutaneous respiration may be far more important in the very small newborn marsupials, in which the movement of air into and out of the lungs is constrained by their minute size, than in the larger species: even in the larger species, like the tammar, cutaneous respiration may be important during that first essential journey to the teat. In bandicoots a possible third route for the supply of oxygen during the journey to the pouch is the well-developed allantoic placenta. The umbilicus remains intact between the placenta in the uterus and the young as it travels to the pouch and for up to 50 minutes after it attaches to a teat (Stodart 1966, Lyne 1974). In these species the placenta is not shed from the uterus and it may continue to supply the young with oxygenated blood while it remains connected.

Carrying oxygen to the tissues of the body An important aspect of pulmonary respiration is getting the oxygen into solution, so that it can then pass across the lung surface into the blood and thence combine with haemoglobin in the red cells. The substance responsible for this, called surfactant, is a complex mixture of lipids, proteins and carbohydrates, which reduces surface tension and facilitates oxygen uptake. It is 76 Life of Marsupials

secreted by the lung epithelium under the stimulus of cortisol. Newborn marsupials, as already noted, have a well-developed adrenal cortex, which produces cortisol before and after birth, and they also have abundant surfactant in their lungs at this time (Ribbons et al 1989). This applies even to the very small young of the northern quoll (Gemmell and Nelson 1989a). Once the dissolved oxygen reaches the blood it must be taken up by haemoglobin and much has been learned about the properties of haemoglobins in fetuses and newborn mammals. In placental mammals fetal haemoglobin has a greater affinity for oxygen than adult haemoglobin, so that the oxygen readily passes across the placental barrier to supply the fetus, and then the adult haemoglobin takes up more oxygen when it reaches the oxygen-rich lungs. Haemoglobins in the blood of prenatal and newborn tammars, and newborn brushtail possums and fat-tailed , have different properties from the fetal haemoglobin of placen- tals (Holland et al 1988, 1994, Calvert et al 1994, Tibben et al 1991). In tammar fetuses during the last five days of pregnancy there are four distinct embryonic haemoglobins, each different from the haemoglobin in the red blood cells of the adult tammar, or the pouch young two to three days after birth. The embryonic haemoglobins have a lower affinity for oxygen than adult haemoglobin at the same partial pressure of oxygen, which means that very little oxygen can be unloaded at the placenta. However, if there is a very high blood flow to the uterine epithelium from the maternal side, sufficient oxygen could be taken up by the embryonic haemoglobin, and its low affinity for oxygen then facilitates delivery of the oxygen to the tissues of the fetus. At present these are only suppositions since the blood flow measurements have not been attempted in these tiny fetuses. Another speculation is that the embryonic haemoglobins, each with differ- ent affinities for oxygen, may play a role in the transition from placental respiration, through mixed cutaneous–pulmonary respiration, to solely pulmonary respiration. They disappear from the blood a few days after birth, as the nucleated red cells are replaced by non-nucleated red cells carrying adult haemoglobin.

How the young marsupial is succoured from birth to independence

The pouch and mammary glands For its continued survival after birth the newborn marsupial must attach itself to a teat and remain there continuously for several weeks, or even months. In all the larger species the teats are accommodated in a commodious pouch, which supports the young as it grows and provides a moist atmosphere and a constant temperature. In many of the smaller species, however, there is either no pouch or folds of skin develop during pregnancy and grow around the young after attachment, partially enclosing them. As they grow the young become exposed and so the pouch does not act as a carrying bag, as it does in the larger species. In these species the ilio- marsupialis muscle penetrates the mammary glands and slips of muscle pass up each teat. Formerly this muscle was thought to be responsible for pumping milk from the mammary glands, but Griffiths and Slater (1988) demonstrated that its function is to support the weight of the young while attached to the teats: when they anaesthetised female brown antechinus and female eastern quolls carrying litters of young and suspended them by their chest and tail the young hung down beneath on the end of flaccid teats. As the female regained consciousness the muscle contracted and each young was drawn up and held close to the body. This muscle in the female is the same as the cremaster muscle in the male marsupial, which draws the testes up to the body wall inside the scrotum, a somewhat similar function. Reproduction and development 77

Figure 2.16: Six different arrangements of the teats in the mammary area or pouch in marsupials, B indicates anterior. 1, caenolestids, small species of didelphids, most small dasyurids; 2, , spotted-tailed quoll, Dasyurus maculatus, and Tasmanian devil, Sarcophilus harrisii; 3, dunnarts; 4, yellow-bellied glider, australis, musky rat kangaroo; 5, large opossums, pygmy possums, possums and gliders, all kangaroos and rat kangaroos; 6, marsupial mole, water opossum, bandicoots, wombats, koala, Phascolarctos cinereus. After Tyndale-Biscoe and Renfree (1987).

The ilio-marsupialis muscle may have an additional function in the second half of lacta- tion. Woolley noticed that when a Julia Creek dunnart is about to leave her young in a nest and go foraging without them, she appears to contract the ilio-marsupialis voluntarily and shake her body, so that the young become detached from the teats. This led to an investigation of the muscle fibres during lactation (Woolley et al 2002). In unsucked teats the muscle fibres are undifferentiated or composite fibres. In sucked teats during the first half of lactation, however, when the young are permanently attached and drawn up close to the mother’s body, the muscle is composed of fast twitch fibres; after the young are left in a nest but the mammary glands are still enlarged, half the fibres are slow twitch, fatigue resistant, fibres. Fast twitch fibres again predominate in the last stage of lactation and may provide the mother with the capacity to shake off the sucking young. 78 Life of Marsupials

The shape of the pouch and the arrangement of the teats within it vary between marsupials and have been grouped into six types for convenience (Fig. 2.16). Among the small dasyurids and didelphids folds of abdominal skin surround the teats, or they are enclosed in pockets at the anterior end or the sides. The number of teats in these species is inversely proportional to the body size of the female, the smallest species having the largest number of teats, and the larger dasyurids, such as the Tasmanian devil and the thylacine having four teats. A curious difference between these two groups is that all female didelphids have an uneven number of teats, whereas all other marsupials have an even number. In the large opossums, the bandicoots, possums, koala, wombats, and kangaroos the pouch is a large bag, which the female can close by contracting part of the underlying muscle, panniculus carnosus. The number of teats in these species varies from 13 in the Virginia opossum, eight in bandicoots, four in kangaroos, to two in brushtail possums, koala and wombats. Because each young must have exclusive use of a teat for the first half of pouch life, in each species the number of teats determines the maximum litter size. Nevertheless, the normal litter size is usually less than the number of teats, either because the number of eggs shed at ovulation is less or the mammary glands are not equally functional (see Chapter 3). For instance, kangaroos have four teats but very seldom give birth to more than one young. Likewise, bandicoots usually give birth to four or fewer young, although there are eight teats in the pouch. In both these groups of marsupials successive young use one of the teats that was not used by the last offspring.

Marsupial lactation In placental mammals lactation is divided into two phases: preparation of the mammary gland before birth of the young; and milk synthesis and secretion after birth. In marsupials the second

Figure 2.17: Summary of events in the development and growth of the tammar wallaby, Macropus eugenii, from birth to weaning, and the corresponding phases of lactation. Phase 1 is the period of preparation of the mammary glands for lactation; in the early stage of Phase 2 the young is continuously attached to one teat, and in the later stage it begins to relinquish the teat but is still wholly dependent on milk for its nourishment; at the end of Phase 2 it is physiologically independent and begins to venture out of the pouch; during Phase 3 it gradually shifts its diet from milk to herbage. After Tyndale-Biscoe and Janssens (1988). Reproduction and development 79

phase is relatively longer and is divided into the period from birth until approximately the time when the young voluntarily releases the teat, which has no equivalence in placental mammals, and the later phase, when the young are growing fast, which is equivalent to the whole of lacta- tion in placentals. So we recognise three phases of lactation in marsupials (Fig. 2.17): Phase 1 is the phase of preparation before birth, or mammogenesis; Phase 2 is the phase of permanent attachment, unique to marsupials; and Phase 3, the growing phase of the young. Each of these phases is under a particular hormonal control.

Preparation of the mammary glands, or mammogenesis Like other mammals, the mammary gland of the marsupial begins as a few simple ducts each opening at the end of a teat. Each duct branches many times and the blind ends of each branch, called an alveolus (Latin, a little hollow), is lined with cells that will eventually synthesise milk, intermixed with special contractile cells, which respond to the hormone mesotocin by strong contractions. These contractions squeeze the milk into the ducts and reservoirs, from where the young can suck it out. During pregnancy, the mammary glands grow in size by the increased branching of the ducts and enlargement of the milk-secreting cells. They are stimulated by progesterone from the corpus luteum and are equally developed in non-pregnant female marsu- pials at the equivalent stage of the oestrous cycle. Sharman (1962) was the first to show this conclusively: he transferred newborn brushtail possums to the pouches of females that had never been pregnant, but were at the equivalent stage of the oestrous cycle, and the young grew to independence in their foster mothers’ pouches. Since then the procedure has been done in other species (Merchant and Sharman 1966) and is now being developed as a means of raising the young of rare species in the pouches of common species. Surgical removal of the corpus luteum late in pregnancy prevents the growth of the mammary glands. In tammars, if the corpus luteum was removed before day 23 of the 26-day pregnancy, the mammary gland could not produce milk and the young died after birth. In these tammars the mammary glands did not grow to their normal size but, more importantly, special proteins that selectively bind the hormone prolactin – the so-called prolactin receptor molecules – were not produced (Stewart 1984). Prolactin, secreted by the mother’s pituitary gland, is essential for milk to be synthesised by the mammary gland cells of the tammar. This has been demonstrated in two ways. If the pituitary is removed during pregnancy or at any time during lactation, milk production ceases, as it does in placentals (Hearn 1974). In placentals prolactin is essential for lactation and so is insulin from the pancreatic islets, thyroxin from the thyroid gland and cortisol from the adrenal glands. The thyroid and adrenal glands will only secrete their hormones if the pituitary is present. However, in the tammar, the only hormone needed to start milk synthesis is prolactin, at the same concen- tration as it occurs in the blood of the pregnant tammar (Nicholas and Tyndale-Biscoe 1985). So, at the end of the first pregnancy of a tammar, or the equivalent time in a non-pregnant tammar, all four of the mammary glands are enlarged and potentially capable of secreting milk and the cell membranes carry specific receptors that can bind prolactin and initiate lactation. In the non-pregnant female the glands regress within one week and the concentration of prolactin receptors also declines.

Lactogenesis the marsupial way In all marsupials only those mammary glands lactate that supply the teats to which young become attached at birth. As in placentals, two factors are probably involved in the mainte- nance of milk synthesis and secretion, known as lactogenesis: withdrawal of milk from the gland and the stimulus of sucking itself. In the early stage of lactation in marsupials the volume of 80 Life of Marsupials

milk being withdrawn is minute, when compared to that of newborn placentals, and the direct stimulus of sucking is the critical factor in the control of lactogenesis. In placentals the stimulus of sucking by the young is conveyed as a neural signal to the brain, causing an increase in the secretion of prolactin from the pituitary gland, which in turn stimu- lates milk synthesis. Paradoxically, in several marsupials in which prolactin has been measured throughout lactation, although there is a peak in concentration at birth, there is no increase in its secretion after birth or for the next several weeks of lactation (Fig. 2.18a). Since prolactin is essential for continued lactation in the tammar, how is this achieved if the sucking stimulus does not increase prolactin secretion?

Pouchexit (a) Wean

>100

80

60

40

20

Plasma prolactin (ng/ml)

0

Birth (b) 10

5 Prolactin-receptor concentration Prolactin-receptor 0 70 140 210 280 350 Days from birth Birth Figure 2.18: Lactation in the tammar wallaby, Macropus eugenii, is controlled by the hormone prolactin and specific receptors for prolactin on cells of the mammary glands: (a) after an initial pulse of prolactin at birth the concentration of the hormone remains low throughout Phase 2, rises in Phase 3 and declines when the young is weaned. (b) The concentration of prolactin receptors in all four mammary glands (6) rises through Phase 1, under the influence of progesterone, and then declines after birth in all except the suckled mammary gland (O), in which the receptor concentration rises to a maximum at the end of Phase 2 and then plateaus through Phase 3. After Tyndale-Biscoe and Renfree (1987)

Sucking increases the target instead of the hormone In the tammar the concentration of prolactin receptors on the cells of the sucked gland increase rapidly after attachment of the young (Stewart 1984). One week after birth the weight of the sucked gland had doubled and the receptor concentration per cell had also doubled. By contrast Reproduction and development 81

the adjacent non-sucked glands had shrunk to one-fifth their former weight and the prolactin receptor concentration had halved. Thus, one week after birth there is a 40-fold difference in prolactin receptor concentration between the sucked gland and the non-sucked glands (Fig. 2.18b). Fostering additional newborn young onto the other teats showed that this is entirely due to the presence of the young on the teat, because all four mammary glands developed equally and produced milk. Thus, in the tammar, and by inference all marsupials, the sucking stimulus of the newborn young increases the local target for prolactin without increasing the secretion rate of prolactin from the pituitary. This explains why only the glands to which young attach lactate. It is yet another way in which the tiny newborn marsupial redirects its mother’s physiol- ogy for its own survival. Phase 2 lactation in the tammar lasts until about day 200, during which time the young first voluntarily relinquishes the teat, its eyes open and it is becoming physiologi- cally independent (Fig. 2.17).

Late stage lactation Phase 3 of marsupial lactation is equivalent to the whole postnatal phase of lactation of placen- tals, during which prolactin is elevated and the young are growing rapidly to independence (see Growing up and leaving the pouch).

The lactating mammary gland grows with the suckling In both the tammar (Stewart 1984) and the brushtail possum (Smith et al 1969) the weight of the mammary gland increases three to four-fold in late lactation and remains at this size until the young is weaned, when it rapidly shrinks. The sucking stimulus still maintains lactation but the concentration of prolactin receptors does not increase any more after 85 days, although because of the larger size of the gland, the total number of receptors is substantially higher. More importantly, from about day 150 the level of prolactin in the circulation increases and this is directly due to the sucking stimulus of the young. When the young is experimentally removed the prolactin level falls to a low concentration, and after the young is returned to the pouch and sucking resumes, the prolactin level again increases. This response is the same as that seen in placental mammals, where the frequency of sucking episodes is the main factor in maintaining elevated prolactin in circulation. Further evidence for the importance of the sucking stimulus rather than milk withdrawal in late lactation comes from the Virginia opossum. Lactation normally ceases at about 90 days when the young begin to feed actively outside the pouch. By fostering a second litter 60 days old to a weaning mother, Reynolds (1952) was able to maintain lactation for 154 days. The mammary glands first regressed to the size appropriate to the younger litter and then began to enlarge again slowly in response to their increasing demands. The persistent sucking of the younger litter, although withdrawing less milk, was yet more effective in maintaining lactation than the infrequent sucking of the older litter. In two other examples the activities of the young determined the length of lactation. A pouch young of the swamp wallaby, Wallabia bicolor, was fostered into the pouch of a red kangaroo (Merchant and Sharman 1966) where it grew at a faster rate than normal swamp wallabies. However, it left the pouch at 267 days, which is the normal pouch life for a swamp wallaby but 30 days longer than that of a red kangaroo. Similarly, a grey kangaroo young, fostered into a red kangaroo, was retained in the pouch for 374 days, or 135 days longer than a red kangaroo young would have stayed. What causes the young to leave the pouch is discussed later (see Growing up and leaving the pouch). The accelerated growth of the swamp wallaby is particularly interesting because its rate of growth was equal to that of young red kangaroos and greatly exceeded that of normal swamp 82 Life of Marsupials

wallabies, and it reached sexual maturity much before them. Similarly, tammar young growing in large mothers grew faster than those growing in the pouches of small mothers (Merchant 1989), and the female young reached puberty earlier. Both observations suggest that the condi- tion of the mother is a significant factor in the growth and development of the young.

How mesotocin helps Dairy farmers call it let down. It is the preliminary response of the cow that allows the milk to flow easily and be drawn from the teat. In the cow this is known to involve the release of the hormone oxytocin from the pituitary gland into the circulation where it stimulates contractions of the muscle fibres that surround the alveoli of the mammary gland. The normal trigger for this signal is the butting of the calf but other external events associated with milking can also elicit release of oxytocin. Like the cow, the lactating mammary glands of marsupials and monotremes are highly responsive to oxytocin and within minutes of an injection of oxytocin milk begins to flow from the teat of the lactating gland. However, in most marsupials the hormone respon- sible for this is not oxytocin but the related molecule, mesotocin. Lincoln and Renfree (1981) examined the response to oxytocin in agile wallabies, Macropus agilis, during Phase 2 and 3 of lactation, by measuring the pressure in the teat ducts of the lactating glands. Since agile wallabies, like red kangaroos, can suckle two young of different ages concurrently, Lincoln and Renfree (1981) were able to measure the response to the same dose of oxytocin at two stages of lactation simultaneously. They were interested in how a dose of oxytocin that could stimulate the small gland at the start of lactation did not flood the pouch with milk from the mammary gland at Phase 3 of lactation. They found that the small gland is more sensitive to oxytocin than the late stage gland and responded maximally at doses that would not elicit a response from the mature gland. Indeed, they concluded that the background level of oxytocin is enough to stimulate the early stage gland sufficiently for the small young one to obtain the small volume of milk it requires. Conversely, the mammary gland supporting the older young requires a higher level of oxytocin to release the mature milk and this is induced by the vigorous sucking stimulus of the older young.

Composition of marsupial milk Because the milk of marsupials must support the young from its tiny size at birth until it is an independent animal, the composition changes profoundly through lactation (Fig. 2.19). At first the milk is a dilute fluid containing more sugars than fats, while at the end of lactation it is rich in fats and proteins but contains little sugar. Not only do the major components change, the constituent sugars, fats, proteins, salts and minerals also change in relation to the needs of the developing young. The milks of several species have now been sampled through lactation, and the sequence of changes is the same in all of them, although the time course differs in relation to diet and life history of each species. Monotremes are similar to marsupials in this matter. These profound changes through lactation in marsupials, and monotremes, are quite differ- ent from those in placental lactation. Apart from the first few days, when colostrum is produced in some species, the constituents and volume of placental milks change very little through lacta- tion when compared to the large changes in all marsupials. Conversely, the differences between species of placental mammals in the concentration of milk and its constituents are much greater than between different marsupials, from the very dilute milks of horse and human (10% solids) to the highly concentrated fatty milks of seals (78% solids). We will now look at the composition of marsupial milk in more detail before considering the development of the young (see How the young marsupial becomes independent). Once again the tammar is the main species against which others will be compared. Reproduction and development 83

Figure 2.19: The changing composition of milk through lactation in the tammar wallaby, Macropus eugenii; (a) total of carbohydrates, lipids and proteins as g/100 mL milk; (b) major components of the whey proteins; (c) daily energy intake calculated as product of milk intake (Fig. 2.1) and energy content of milk components, with area under the curve representing the total energy exported by the mother tammar through lactation. After Nicholas (1988) and Cork and Dove (1989).

Carbohydrates in marsupial milk Lactose, composed of one molecule of glucose and one of galactose, is the main sugar of placen- tal milk but it is a very transient component of marsupial milk: during days 1–4 of lactation, lactose is the only sugar but thereafter its concentration rapidly declines. For the next 200 days tammar milk contains high concentrations of several oligosaccharides of increasing size, most composed of one molecule of lactose and from one to seven molecules of galactose. After 210 days these oligosaccharides disappear from the milk and the only carbohydrates are galactose and glucose in very low concentration. The question that exercised people in the 1970s, when this was first discovered, was why the simple sugars are packaged in the milk as large molecules that then have to be broken down in the young animal? 84 Life of Marsupials

Part of the answer may be that the early milk of marsupials contains higher concentrations of sodium and potassium ions than late stage milk or placental milk (Green et al 1980). The change- over occurs at about the time when the kidneys of the young develop the ability to concentrate urine (see Water economy and kidney function). Throughout lactation in the tammar the osmotic pressure of the milk remains the same as the plasma and tissues of the mother (ie it is isotonic). If it were higher than the maternal tissues, it would require energy to drive the secretion against the higher osmotic pressure of the milk. Since one molecule of a simple sugar exerts the same osmotic pressure as a large oligosaccharide molecule, secreting the sugars as large molecules keeps the milk isotonic and reduces the energy required to transport it across the mammary cell walls, while providing the young animal with high levels of carbohydrate and essential salt. Second, were the large sugar molecules hydrolysed in the gut, the increased osmotic pressure would draw water into the gut and cause diarrhoea. Instead, the oligosaccharides may be trans- ported into the intestinal cells by pinocytosis (meaning cell drinking), ingestion of droplets containing the sugar molecules and maternal antibodies in the milk. The enzyme that breaks down the oligosaccharides to galactose and glucose, acid beta-galactosidase, occurs inside the cells lining the intestinal wall of the pouch young but not in the lumen of the gut. Once inside the cell, the droplets discharge their contents, and the complex sugars are broken down into their constituents, glucose and galactose, which enter the circulation and provide the main energy needs of the young (Messer and Green 1979, Messer et al 1989).

Lipids in marsupial milk The early milk of the tammar is low in lipids, which comprise about 5 g/100 mL; this rises to 24 g/100 mL by the end of lactation (Fig. 2.19a). The change from carbohydrate to fat as the main source of energy in the milk coincides with the first emergence of the young from the pouch and its accelerated growth and increasing metabolic rate (Fig. 2.17). In dasyurids the same changes occur in the milk when the young are left in a nest. For the tammar the changes are even more dramatic when converted to daily energy intake (Fig. 2.19c). This reaches a peak of 668 kJ/day at day 260, with lipid contributing more than 70%: such a large increase could not be supplied by carbohydrate. The most abundant component in the lipid fraction is triglyceride (Griffiths et al 1972, Green 1984, Green and Merchant 1988) but traces of phospholipids, cholesterol and free fatty acids are also found. More than half the triglyceride in early milk of tammars, red kangaroos and quolls is saturated palmitic acid (C16:0), which declines after about day 70. Palmitic acid may be impor- tant in the synthesis of surfactants, which as we saw earlier, are required for oxygen transport across the lung. In the second half of lactation the predominant triglyceride is unsaturated oleic acid (C18:1), also characteristic of placental milk. Oleic acid is important in the synthesis of myelin sheaths around nerve fibres during the second half of pouch life (see Box 2.5).

Proteins in marsupial milk The protein component of milk of all marsupials studied increases through lactation but not as dramatically as do lipids (Fig. 2.19b). In the tammar the two major protein fractions are casein and whey, which both increase gradually (Green and Renfree 1982, Nicholas 1988). However, for whey this gradual increase hides large changes in the individual proteins, and more subtle changes in the component amino acids (Renfree et al 1981). Of the whey proteins, alpha- lactalbumin remains constant throughout lactation while serum albumin increases during Phase 2 of lactation and declines in Phase 3. Transferrin, the protein that carries iron, increases substantially in Phase 3 but the most remarkable feature of this phase of lactation is an unique protein, late lactation protein (LLP), which first appears at 180 days and rapidly becomes the major component of the whey proteins (Nicholas 1988). Its secretion coincides with the rise Reproduction and development 85

in concentration of prolactin, which as we saw is due to the sucking stimulus of the advanced pouch young. However, secretion of LLP cannot be induced in mammary gland tissue during early lactation by raising the concentration of prolactin experimentally and expression of the gene for LLP is only induced between 200 and 240 days (Trott et al 2002), so it is probable that the protein can be synthesised only after the gland has passed through the earlier phases of lactation. Collett and Joseph (1994) showed by in situ hybridisation that gene expression for alpha-lactalbumin and for LLP occurred in the same mammary epithelial cells at the changeover period in mid lactation, which supports the idea of developmental change in the secretory cells themselves. Similar whey proteins appear in the milk of the quokka, red kangaroo and grey kangaroos, coinciding with the time when the young emerges from the pouch and begins to eat herbage. A similar protein is not found in the milks of non-macropod species, such as the brushtail possum and the , peregrinus, nor in the carnivorous eastern quoll. One idea is that it has a dietary role in kangaroos and wallabies when they change from milk to the fermentation of herbage in the forestomach (see Growing up and leaving the pouch). This is supported by the observation that an LLP-like protein does not appear in the blood of the young tammar or quokka (Jordan and Morgan 1968). The gamma-globulin fraction of whey milk of marsupials contains all three classes of immunoglobulins, namely IgG, IgM and IgA. In the tammar IgG from the mother appears in the earliest milk and in the serum of pouch young in the first two days after birth (Deane et al 1990), which may confer passive immunity on the young animal. In the quokka gamma-globu- lins remain constant throughout lactation at levels below those in the mother’s serum and it is thought that they are derived from the maternal blood by selective transfer, as in placentals (see First phase of pouch life – acquiring immune competence). In the brushtail possum IgA is only high at the beginning of lactation, while IgG only rises in late lactation (Adamski and Demmer 2000).

Inorganic elements and salts in marsupial milk One major difference between marsupials and placental mammals is the means of transport- ing iron and copper to the young. In placentals, such as humans and cattle, copper and iron are actively transferred across the placenta, so that the newborn has levels in the blood (bound to the serum protein transferrin) and in the liver much higher than those in the mother’s blood. The milk contains very little of these elements, so that the levels in the suckling fall progressively until weaning. Although there are transferrin proteins in the yolk sac fluid of the tammar (Fig. 2.4), it is impossible for the marsupial at birth to contain adequate stores of copper or iron to maintain it until it is weaned. Indeed, in the young quokka the levels of both elements rise progressively during pouch life, as do the transferrin proteins in the plasma: these elements are obtained from the milk. During the first 170 days the concentration of iron in the milk is about five times (Kaldor and Ezekiel 1962) and of copper, three times, the concentration of maternal plasma and then falls to maternal plasma levels when the young quokka begins to emerge from the pouch and eat grass. The same pattern occurs in the tammar, red necked wallaby, Macropus rufogriseus, common and eastern quoll (Green 1984, Green and Merchant 1988). In all these species the fall to lower levels coincides with the time, in each species, when the young first leaves the pouch. Like other macropods, quokkas can suckle two young differing in age by about 170 days, so such an animal must be transporting a far higher concentration of iron and copper across the newly suckled gland than across the other. Whatever the transport mechanisms are they must be intrinsic to the gland and be a function of the stage of development of the gland, just as for LLP. 86 Life of Marsupials

Other minerals show opposite patterns of secretion in milk. For calcium, phosphorus and to a lesser extent magnesium, the concentrations in late stage milk rise to double the concentra- tions in early milk (Green and Merchant 1998), perhaps reflecting the increasing requirements for these elements in bone growth, which is accelerating. Sodium and potassium concentrations change during lactation in all marsupials so far examined (Green and Merchant 1988), with high levels of sodium in early milk declining to low levels after about 170 days, and potassium being the opposite. The changeover coincides with the maturation of the young animal’s kidneys (see Water economy and kidney function).

Conclusions about marsupial milk The clear pattern that emerges from the foregoing is that there are profound differences in almost all constituents between the early milk that supports the very immature pouch young and the later milk that supports the young when it is becoming independent of its mother and growing fast. It suggests that lactation in marsupials, and probably monotremes, is a much more complex process than lactation in placental mammals. This conclusion is supported by the extraordinary phenomenon of concurrent, asynchro- nous lactation in kangaroos and wallabies. As mentioned earlier, in several species of marsupial, which have successive litters, the second litter attach to teats unused by the previous litter, after the first litter have weaned. However, in many species of kangaroo and wallaby the first young continues to feed from the elongated teat on the much-enlarged gland, while the succeeding newborn attaches to one of the three remaining teats. Thus, the female simultaneously secretes from adjacent mammary glands two kinds of milk, differing widely in volume, energy content and all the component parts that represent the beginning and the end of lactation. The broad outline of the endocrine control of this is now understood but many of the more subtle changes that take place are certainly not. Some of the changes, such as the appearance of LLP or the secretion of iron and copper, seem to be due to an intrinsic sequence of cellular changes in the mammary gland itself, not directly controlled by pituitary hormones.

How the young marsupial becomes independent

Three stages of postnatal development At birth the young tammar weighs about 350 mg, less than 0.008% of its adult weight. It remains continuously attached to one teat until 100 days, during which time it reaches a body weight of about 100 g. Although the overall rate of growth appears to be slow, the instantaneous growth rate – the daily increment as a proportion of total size – is high. Between 100 and 200 days big changes occur: the eyes and ears open, the young tammar can stand on its feet, it is finely furred, it can concentrate urine and maintain a constant body temperature (Fig. 2.17). After 200 days its instantaneous growth rate accelerates again and it begins to leave the pouch and eat grass. It is weaned off milk by 300 days and, if female, then comes into first oestrus. All marsupials follow this sequence, albeit with different intervals and sizes. It is convenient to divide pouch life of a marsupial into these three phases. Phase 1 is, broadly, about developing an immune system and a functional nervous system, Phase 2 about control- ling physiological functions of the body, and Phase 3 about physical growth, sexual maturity and independence. For the mother Phase 1 is not energetically costly but Phases 2 and 3 are very demanding. Reproduction and development 87

Phase 1 of pouch life – acquiring immune competence Foreign organisms or proteins entering the body of an adult mammal elicit a defence response, which involves the proliferation of special cells by the lymphoid tissues of the body and the production of specific immunoglobulins, known as antibodies. These appear in the gamma- globulin fraction of the serum and react specifically with the foreign matter or antigen to neutral- ise it. In placental mammals the ability to mount an antibody reaction develops at about the time of birth, coincident with the differentiation of the lymphoid tissue of the thymus. Before this stage the animal or fetus is tolerant to foreign tissues and proteins and moreover in subsequent life will remain tolerant to the same proteins if encountered again. During this early period the fetus or newborn receives immunoglobulins from the mother, either across the placenta, as in the rabbit and human, or through the milk, as in cattle and horses, in which the first milk, or colostrum, is enriched with maternal immunoglobulins. The newborn marsupial is far more immature than the newborn placental and enters an environment, either in a pouch or on the belly of its mother that is certainly not sterile, and bacteria colonise the gut from an early age (Yadav et al 1972): how then does it survive? As already noted, immunoglobulins provided by the mother are present in yolk sac fluid of the tammar and they are also secreted in very early stage milk of the quokka (Yadav and Eadie 1973), the euro, Macropus robustus (Deane and Cooper 1984), the tammar (Deane et al 1990) and the brushtail possum (Adamski and Demmer 2000). This continues in all four species until the young first relinquishes the teat. Throughout this period the maternal proteins are absorbed unchanged across the gut epithelium, along with the oligosaccharides mentioned earlier, and presumably they provide protection to the immature young during this lengthy period: this was demonstrated in the quokka and the short-tailed opossum by immunising the mothers with specific bacterial antigens and subsequently finding antibodies specific to the bacterial antigens in the serum of the young (Old and Deane 2000). The immune system of the young animal develops much earlier than in placental mammals and this has been most thoroughly studied in the Virginia opossum and four other species. No lymphoid tissue is present at birth in any species but within the first week after birth lymphocytes appear in the thymus, which is the first lymphoid tissue to develop. The thymus of the Virginia opossum consists of a pair of structures lying near the base of the aortic arch and on the day of birth consists of undifferentiated embryonic cells (Block 1964). Within a day or so of birth, the first lymphocytes and the first lymph nodes appear and by day 17 the spleen also contains differentiated lymphoid tissue; plasma cells and secondary lymph nodes appear by day 60. The development of lymphoid tissue was grossly affected by removing the thymus at day 7 (Miller et al 1965): the number of small and medium lymphocytes was reduced and they failed to appear in the spleen, in which myeloid tissue persisted and increased. This suggests that the thymus is important in the origin and maintenance of lymphoid tissue and the suppression of myeloid tissue. All polyprotodont marsupials have a single thoracic thymus, like the opossum but the dipro- todont marsupials, such as the brushtail possum, tammar and quokka have a superficial thymus in the neck as well, and this becomes the dominant lymphoid tissue in these species. The thymus has differentiated into a distinct cortex and inner medulla by day 14 and its adult structure was fully developed by day 120 in the tammar (Basden et al 1997). The development of immune competence has been tested in the Virginia opossum and the quokka by the response of the young to skin grafts and to infection with bacteria. Before day 6 young opossums injected with a bacterial suspension or infected by a dirty wound do not produce antibody, do not show an inflammatory reaction and rapidly succumb to the infec- tion (Rowlands et al 1964); but after this age reaction to infection progressively increases and antibody can be detected in progressively higher amounts, while wounds rapidly heal. 88 Life of Marsupials

Similarly, the young opossum less than 10 days old will accept a foreign skin graft (Laplante et al 1969), retain it permanently and, if later challenged with a second graft from the same donor, will not reject it either. However, after day 12 a foreign skin graft was rapidly rejected. These experiments show that immunological tolerance is lost and immune competence is achieved in the opossum within the first two weeks of pouch life, an extraordinarily early stage of develop- ment that is far in advance of the young of placental mammals: however, it is an equivalent time after birth. The quokka also develops immune competence at an early stage of pouch life, so it is very likely that precocious development of immune competence is common to all marsupials and is an important adaptation for survival outside the mother’s body.

Phase 1 of pouch life – getting wired up For the first 120 days after birth (Phase 1) the brain is the fastest growing part of the body and the first part to differentiate into its final form. The separate parts of the brain grow at different rates and the change to a slower rate also varies (Renfree et al 1982). Thus, the olfactory lobes grow fastest and the rate slows at 120 days, while the brain stem grows slowest and does not change until 180 days. The gross changes in the brain reflect changes to the internal structure of the brain. At birth the brain stem and olfactory lobes are partly differentiated but the cerebral hemispheres and the cerebellum, so important in adult life, are small, undifferentiated and lack neural connection to or from the brain stem. Also, only some of the 12 cranial nerves have made connection to the brain, notably I, V and VII, and possibly VIII. As we saw earlier, these are the parts of the nervous system that control the movements of the young as it travels to the pouch and attaches to a teat. The major developments during the first 100 days are the differentiation of the eye, ear and whiskers and the cerebral cortex. Also during this first period the connection of the remaining cranial nerves to the brain, and the relay systems that link the several parts of the brain to each other and to the spinal cord, take place. In placental mammals all of this development takes place before birth, whereas in marsupials it takes place in the pouch. This accessibility of the young has allowed detailed studies on the developing functions of the nervous system (Box 2.5).

The visual system The eyes begin as paired outgrowths from the front end of the brain. When they make contact with the outer layers of the head they invaginate to become two-layered cups. The inner layer of each cup becomes the light sensitive retina and the outer layer becomes the pigment layer of the eye. The other parts of the eye, such as the lens, form from the overlying layers of the head. Differentiation of the retina of the tammar and the quokka has been described in detail and both follow a similar pattern (Dunlop et al 1988, Mark 1997, Marotte and Sheng 2000). The retinal cells divide and differentiate into three layers of interconnected sensory neurones. The outer- most layer comprises the photosensitive receptors, the second layer are interneurones, which modify and accentuate various aspects of the light signal before passing it on to the innermost layer, the ganglion cells. Each ganglion cell sends an axon to the brain along the optic nerve. Cell division begins just before birth and continues until day 100, after which very few neurones are formed. There is, however, considerable loss of cells by spontaneous death during this period, a feature common to the differentiation of other parts of the nervous system as well. In both species the first axons from the retinal ganglion cells begin their growth along the optic nerve (II) at birth. By day 12 they have reached the two main visual centres in the brain, the superior colliculus (SC) in the midbrain, and the lateral geniculate nucleus (LGN) of the forebrain. Those that go to the superior colliculus synapse with motor neurones that later will control eye and neck movements and thus the direction of gaze and attention. Those ganglion cell axons that reach Reproduction and development 89

Box 2.5: Neurones: what they are and what they do The building blocks of the nervous system are specialised cells called neurones and supporting cells called glia. Neurones come in many shapes and sizes but all comprise a central cell body containing the nucleus and one long tubular outgrowth called an axon or nerve fi bre, and many short, much-branched outgrowths, called dendrites. The cell body and dendrites receive chemical signals from other neurones or from sensory organelles, which initiate an electrical disturbance of the cell membrane called a nerve impulse. The nerve impulse travels out along the axon to its end, where the axon forms a special contact with the dendrites or cell body of another neurone, which in turn excites another neurone, and so on. Eventually the nerve impulse reaches a specialised motor neurone, whose axon terminates on a muscle fi bre and initiates a muscle contraction. The contact between neurones is called a synapse and an individual neurone may have up to 100 000 synapses with other neurones. By far the largest number of neurones form connections only with each other and are called interneurones. Interneurones are of two sorts, excitatory and inhibitory, depending on the specialisation of the synapses made by their axons. Excitatory synapses start nerve impulses and inhibitory ones prevent or block them. Interneuronal connections form specifi c circuits, which are characteristic of the species and organ, and are set up during development before they become functional. The working of this immensely complex system depends on the precision of the anatomical circuitry and the physiological balance between excitation and inhibition. Sensory neurones, and those interneurones to which they are connected, convey signals from sense organs, such as touch and temperature receptors in the skin, photoreceptors in the eye, stretch receptors in muscle or chemoreceptors in the nose and mouth, to the brain. They are also called afferent nerves because the impulse is towards the brain. Motor neurones convey nerve impulses away from the brain or spinal cord, and are called efferent nerves. The axons of motor neurones form synapses with muscle fi bres and their signal excites the muscle cell to contract. The simplest circuit in mammals is the stretch refl ex, in which stretch sensitive sensory nerves in muscle form synapses in the spinal cord with the motor neurone serving the same muscle. Stretching the muscle sends impulses up the sensory axons to the dendrites of the motor neurone, which transmits them back to the muscle, causing it to contract to its stable length. Usually, however, there is one or more interneurone between the sensory and motor neurones. Complex behaviour, such as walking, requires an intrinsic rhythm to be set up by a specialised set of interneurones in the spinal cord, called a pattern generator, the operation of which may be modifi ed by sensory neurones and which gives out coordinated impulses to the motor neurones. The arm movements of the newborn marsupial, described earlier, are an example of the integrative actions of the nervous system. The speed of transmission of a neural signal depends on the diameter of the axon and the number of synapses across which it passes. It is also much faster if the axon is covered in a fatty sheath called a myelin sheath formed by the wrappings of membranes of supporting glia cells. Myelin sheaths develop during development of the central nervous system. the LGN, however, synapse with a second relay of neurones that grow from the LGN into the roof of the forebrain, later to become the cerebral cortex, which they reach by day 15. In common with all mammals, most of the axons from the retina of one eye connect to the centres on the opposite side of the brain. This large crossover site is called the optic chiasm, a prominent feature on the lower surface of the brain. It is thought that crossover of the optic nerves compensates for the inversion of the image in the retina and so brings it into register with information coming from the body, especially the hands. Despite this very early establish- 90 Life of Marsupials

ment of the main nerve tracts in the brain, none is functional for many weeks. During this long period other axons from the retina follow the first ones and synapses form at the relay stations. In addition, efferent neurones in the cortex send axons down to the LGN by day 40 and to the superior colliculus by day 85. Part of the delay is because the efferent neurones that form layers 5 and 6 of the cortex involved in vision do not reach their own final positions until about day 36 (see The cerebral cortex). So what is the significance of this long, drawn out process? Mark’s (1997) idea is that in their early migration from the retina the first axons are responding to spatial cues in the matrix of the brain and follow pre-ordained pathways that take them to the SC and the LGN. Here they are positioned in patterns that accurately represent the pattern of their cell bodies in the retina of the opposite eye. He showed that the outgrowing axons would reach the normal position even after the eye from which they arose had been rotated 180°. Once the basic plan is established by the pioneer axons, others follow the same course. The linking up of the several elements takes place when the ends of the axons make multiple synaptic connections with other neurones, and signals can begin to pass from the eye to the cortex. The earliest age at which neural signals can be recorded from the cortex of the tammar, in response to stimuli of the optic nerve, is on day 46 and from the SC on day 39. In addition signals from descending neurones from the cortex to the SC can first be picked up at day 130. Thus, the neural connections between the eye and the brain have assumed their full capacity at the same time as the eyes open for the first time at 140 days.

The auditory system The main studies of this system in marsupials have been done on the gray short-tailed opossum, the northern quoll and the brushtail possum. While the sequence of events is similar in each species, the times for development differ. The following description is for the quoll (Aitkin 1998). While the parts of the inner ear, concerned with detecting gravity, are developed to some degree at the time of birth, the cochlea, which detects sound, is absent: it develops during pouch life at the same time as the auditory pathways within the brain are being established. As with vision, the auditory pathway involves several relays from the auditory nerve that enters the brain stem to the part of the cerebral cortex concerned with hearing. The first neurones in the brain stem arise within a few days of birth and their axons cross to the opposite side of the brain stem, like the crossover in the visual pathway. The next set of neurones arises during days 7–22 and their axons grow out to the roof of the midbrain, the inferior colliculus (IC). During the same time other neurones arise in the thalamus and their axons reach the cerebral cortex, where the six layers are being formed by migration, as in the visual cortex. This is complete by day 42 in the quoll. While the main neural network is established early in pouch life, the synaptic connections between the several sets of neurones take much longer to form. For the first 45 days the number of synapses increases slowly but after this age they increase rapidly to a maximum by day 70. The change in rate coincides with the opening of the ears and the exposure of the eardrum to external sound at day 63, and it has been suggested that the sounds received actually stimulate the forma- tion of synapses. By this age the young quoll responds to loud noise and by day 80 hearing acuity and sensitivity have attained the adult state. This sequence of events occurs earlier in the short- tailed opossum, at about day 25 when it is first left in a nest. In the tammar, the first response to sound was detected at day 114, considerably later than in the smaller species (Liu et al 1997).

Whiskers Whiskers are a good example of the sensory input from the body surface – the somatosensory system. They are highly specialised hairs around which are grouped several kinds of recep- Reproduction and development 91

tor cells: in the tammar, about 200 sensory fibres serve each whisker. In the adult animal they respond to bending of the hair shaft by anything that touches it, different receptors being tuned to different directions. Development of whisker innervation in the tammar takes the first three months of pouch life and is a three-stage process (Waite and Weller 1988, Waite et al 1998). Whisker buds appear at birth with outgrowth of the whisker itself occurring at day 35. The cell bodies of their sensory neurones lie just outside the brainstem in the trigeminal ganglion (cranial nerve V). Each

Figure 2.20: (a) The surface pattern of whiskers on the right muzzle of the 91-day tammar wallaby pouch young, Macropus eugenii and (b) the corresponding spatial representation of whiskers A and E in the somatosensory cortex of the left cerebral hemisphere After Waite et al (1998), photos by PME Waite. 92 Life of Marsupials

neurone gives rise to a long dendron that connects with the receptor cells on the whisker, and to an axon that enters the brain stem. The axons synapse with a second set of interneurones that send axons from the brain stem to the opposite thalamus region of the brain. Here, third stage neurones grow out from the thalamus and enter the developing cerebral cortex. This three-stage nerve framework is in place by day 15 after birth but the highly complex synaptic arrangement at each relay station does not become functionally connected until day 72. During this time the axons make many synapses with dendrites of the neurones of the next relay in a highly charac- teristic pattern, called a barrel. Each barrel comprises the cell bodies of the next relay arranged in a circle, with their dendrites pointing inwards and the ends of the axons of the previous relay in synaptic interaction. In the brain stem the barrels appear at day 40 and those in the thalamic relay appear at day 55. The final relay form synapses in layer 4 of the developing cerebral cortex, where they form the definitive barrel, at around day 80–90 (Fig. 2.20b). The actual whiskers have appeared by day 35 and are now connected, through the three relays, to the cerebral cortex. Each barrel in the cortex corre- sponds to an individual whisker on the opposite side of the snout of the young tammar and, remarkably, the pattern of the barrels in the cortex is the same as the pattern of the whiskers on the snout (Fig. 2.20a). In the brushtail possum when single whisker follicles were experimentally removed before day 50, the corresponding position in the cortex lacked a barrel, which suggests that the differentiation of each barrel is controlled in some way by the developing whisker, or the first signals emanating from it.

The cerebral cortex All the sensory input from the body is projected onto the cortex in specific sites, which can be mapped. If one looks at a section taken through the cerebral hemispheres of any mammal, the cell bodies of the neurones are arranged in six distinct layers, each with a separate function. The neurones in layer 4 receive axons ascending from lower centres in the brain, while neurones in layers 5 and 6 send axons to relay stations in the lower brain and, hence, to the motor neurones of the cranial nerves and the spinal cord. Likewise the centres that control motor function can also be mapped to the surface of the cortex. Layers 2, 3 and 5 are connected to other neurones in the same and in other layers of the cortex. The two halves of the cerebral cortex are connected by transverse bundles of axons arising from neurones in layers 3 and 5. In marsupials the most important of these interhemispheric tracts is the anterior commissure, with 20 million fibres in the adult tammar (Shang et al 1997). In placentals the anterior commissure is small and another structure, the corpus callosum, carries the main traffic from one cerebral hemisphere to the other, with up to 200 million fibres in the human brain. At birth, the cerebral cortex of placental mammals already has six layers of neurones, but in marsupials the cortex consists of only two embryonic layers: the six layers of the mature cortex will be formed during the first 70 days after birth. This has enabled researchers to study the formation of the cerebral cortex and the development of the neural connections between the six layers much more easily in marsupials than in placentals. The main technique used is to label the nucleus of dividing cells by incorporating a radioactive marker into one of the bases in the DNA molecule. Uridine or thymidine labelled with tritium (3H) is the usual marker, which is injected into the animal at a known age and, as the cells divide, the radiolabelled base is incorporated into the DNA of the newly formed neurone. If the cell does not divide again after being labelled in this way, the signal from the cell is of maximum strength but if it divides, the label is shared between the daughter cells and the signal becomes weaker with each division. By injecting a series of tammar pouch young of increasing age up to 70 days, it is possible to identify the age at Reproduction and development 93

which cells reach each of the six layers of the cerebral cortex and cease to divide (Reynolds and Saunders 1988). The neurones migrate to their final positions from the formative layer of cells lining the inner surface of the brain. The first cells to divide migrate to layer 6 during days 16–28, wheras the successively later born cells migrate through the lower layers to their positions further out. Each layer is established about 10 days after the last so that all the layers can be recognised by day 120 (Marotte and Sheng 2000). After day 120 no further neurones are formed from the basal layer and the young animal now has its final complement of neurones that will suffice for the rest of its life. From day 120 the cerebral cortex is ready to coordinate the activities of the young tammar. Connections between the cortex and the underlying centres of the brain take place during the same period that the six layers are forming, with axons progressively reaching the furthest, layer 1 (Mark and Marotte 1992). Axons from the eye, the whiskers and the ear make synaptic connec- tion mainly in layer 4, while the efferent neurones reside in layers 5 and 6. Overproduction of neurones in early development of the brain and subsequent cell death is common to the establishment of all sensory pathways in the brain. For instance, in the tammar the number of axons in the anterior commissure peaks at 60 million by day 140 and then declines to the adult number of 20 million (Ashwell et al 1996). It is supposed that this allows for the final sculpting of the tracts and connections between the different parts of the brain by experi- ence and repeated use. Myelinisation of fibres of the anterior commissure begins at day 160 and is completed by day 300 (Ashwell et al 1996). This is when oleic acid, which is an important component in the synthesis of myelin, becomes one of the main lipids in the milk.

The pyramidal tract The pyramidal tract is a large tract of motor neurones that convey signals direct from the cerebral cortex to the spinal cord. The cells are in layers 5 and 6 of the cerebral cortex and their axons pass down to the brain stem, where they cross over to the opposite side and synapse with the ventral column of the spinal cord. The pyramidal tract first appears at day 30–40 in the Virginia opossum. It is especially important in species that use their hands, since its role is to convey motor control from the cerebral cortex directly to the forearm and hand. It is well developed in the brushtail possum and tammar but not in dasyurids. In addition, there are other relays that connect the cerebellum to the spinal cord, which are involved in the control of movement and position, as well as the direct motor tracts that connect the cerebral cortex to the limbs and trunk via the spinal cord. Maturity of the nervous system is shown by the ability of the young tammar to hear at day 114, to see at day 140, to call and to stand at about day 200 and to hop at day 210.

Phase 2 of pouch life: becoming physiologically independent Phase 2 of development in the tammar lasts from about day 100 to day 200. It begins when the young first relinquishes the teat and lasts until it makes its first excursion from the pouch and begins to nibble grass. During this phase it develops control of its body temperature, the thyroid gland becomes functional as the fur thickens, and the kidney matures so that the young can concentrate its urine and conserve water. By day 200 it is physiologically independent of the mother (Fig. 2.17).

Thermoregulation, respiration and metabolism The ability to maintain body temperature at a constant level regardless of the surrounding air temperature has allowed birds and mammals to occupy a wide range of habitats and for the body to function at a high rate (see Chapter 1). But it comes at a high cost in terms of food intake, 94 Life of Marsupials

and the costs are inversely related to body size, the smallest species having the highest relative food requirements and oxygen consumption (Table 1.2). For the very small young of birds and mammals, maintaining a constant body temperature would require a very high input of energy, because of their small size. Instead, they rely on the mother’s body heat, either during incuba- tion in a nest, in the uterus or in the pouch. For the young marsupial the pouch is an environ- ment with a high stable humidity and temperature and, in species with a closed pouch, a carbon dioxide concentration of about 3% (Hulbert 1988).

Virginia opossum, pouch young 94 days

69 days 75 days 81 days

3

84 days

Virginia opossum Quokka

2 94 days 153 days

1 178 days

Oxygen consumption (ml/g/h) 64 days 100 days

Figure 2.21: Development of thermoregulation in the Virginia opossum, Didelphis virginiana, and the quokka, Setonix brachyurus. Top panel. In very young opossums body temperature falls rapidly to the low ambient temperature but at 81 days thermal stability has begun and is fully established at 94 days. Lower panel. Oxygen consumption of young opossums and quokkas at a range of ambient temperatures: before thermoregulation has developed oxygen consumption is directly related to the ambient temperature but as thermoregulation is established the pattern reverses, so that oxygen consumption is then highest at the lowest ambient temperature. After Reynolds (1952) and Shield (1966). Reproduction and development 95

If removed from the pouch during the first half of pouch life, the body temperature of the young marsupial rapidly changes to the surrounding temperature. If this is a low temperature, the animal becomes torpid and may cease to move, but will recover as quickly if returned to the warmth of the pouch. As it grows older the fall in body temperature after removal from the pouch is slower and, in the second phase of pouch life, the young may hold a steady tempera- ture for several hours, before falling to ambient (Fig. 2.21). Finally, at about 94 days, when the young Virginia opossum first emerges from the pouch, it can maintain a stable body tempera- ture against an ambient gradient as well as the adult animal can: now it shivers in response to low temperature and pants and licks its arms when exposed to a high temperature. Maintaining the temperature of the body above or below its surroundings requires the animal to generate or dissipate its own heat and to have sufficient insulation to prevent heat loss. The necessary adaptations for this develop during Phase 2 of pouch life in the marsupial. The daily energy intake from milk in the tammar young increases steeply from 200 days to a peak at 269 days, which dramatically reflects its increasing energy needs and the increasing burden for the mother of supporting it (see Lipids in milk). By measuring the oxygen consumed by an animal in a closed system, we can assess its net expenditure of energy at a given temperature. The pouch young of opossums, quokkas and tammars have been subjected to this procedure. The oxygen consumption of tammars and quokkas less than 100 days, and of opossums less than 60 days old, increases directly with an increase in the ambient temperature up to 35°C. However, as older animals begin to control their body temperature this is reversed: they now consume more oxygen when held at a low temperature than when held at a high temperature (ie oxygen consumption is inversely proportional to ambient temperature, Fig. 2.21). In opossums and quokkas held at low temperatures, the young that were just becoming furred consumed more oxygen than older animals with a full pelage of under-fur and guard hairs, presumably because the older ones had better insulation. In the quokka the ability to shiver develops at 120 days (Shield 1966), and in the tammar at 150 days, which is the age in each species when thermo- regulation is beginning. Conversely, when tammar young were held at an ambient temperature of 37.5°C, all showed a rise in body temperature but only those about 200 days old displayed panting and licking of their fur, whereas younger ones did not (Janssens and Rogers 1989). The major organs involved in the development of thermoregulation and water conservation in the young marsupial are the thyroid, liver and kidney, and the hormones that control the functions of each one. We will now consider the development of each of these systems in the tammar.

The thyroid gland Mammals from which the thyroid gland has been removed are unable to respond to cold temper- atures by increasing their metabolic rate. The thyroid synthesizes two hormones, thyroxine and tri-iodothyronine, which both contain iodine as a component of the molecule. Hence, the uptake of radioactively labelled iodine can be used to measure thyroid activity. When secreted into the circulation the general effect of these hormones is to stimulate glucose oxidation and protein synthesis. The way that the thyroid hormones stimulate endogenous heat production is by uncoupling oxidative phosphorylation and so channeling energy into heat rather than into synthesis of adenosine triphosphate (ATP). Because the development of thermoregulation occurs during pouch life in marsupials, it was inferred that thyroid function would also develop at this time. In tammar pouch young the thyroid tissue has differentiated into the characteristic follicles by day 75, but secretion did not accumulate in the follicles until day 180 (Setchell 1974). This coincided with a sharp rise in the uptake of iodine by thyroid tissue, reflecting the onset of active synthesis of thyroxin and tri-iodothyronine (Fig. 2.22). At the same time the level of iodine 96 Life of Marsupials

loosely bound to serum proteins fell, again indicating that the thyroid was sequestering all of the injected iodide. Both hormones rose to peak concentrations in the blood between day 160 and 180 and had dropped to adult values by day 300 (Janssens et al 1997). This transient peak may be associated with the higher oxygen consumption of the finely furred young before it is fully insulated. The importance of the thyroid gland for the development of thermoregulation in the tammar was critically shown by Setchell (1974), who surgically removed the thyroid glands of young tammars before day 140. These young subsequently failed to grow at the normal rate, to differentiate and grow a full coat of fur, and were unable to respond to low temperature by increased oxygen consumption, like normal young can do.

Tammar

Figure 2.22: Thyroid development in the tammar wallaby, Macropus eugenii. Uptake of the radioactive isotope of iodine, as Na131I, by the thyroid gland, liver and blood plasma of pouch young from 60 to 220 days. After Setchell (1974).

The increased metabolism of the older young requires a substrate to provide the energy for oxidation. At first this comes from the abundant sugars in the milk but after 180 days it comes from the increasing amount of lipids and proteins in the milk (Fig. 2.19).

Liver function Enzymes in the liver control the release of the energy substrates. Complex sugars, proteins and lipids in the milk are digested in the intestine to simple sugars, amino acids and short and long- chain fatty acids and pass directly to the liver. The liver is the main clearing house for the products Reproduction and development 97

of digestion, storing some and releasing others, so that the level of each in the blood remains fairly constant. Glucose is the main sugar arriving from the intestine and excess glucose is stored in the liver as glycogen. Under the influence of specific enzymes the process can be reversed when glucose is released into the blood stream. In well-fed pouch young the level of glycogen in the liver is generally high but after 24 h off the teat the glycogen is much reduced. The ability of the liver to maintain blood glucose at a constant level by storing or releasing glucose as required develops during Phase 1 of pouch life, while carbohydrate is plentiful in the milk. Towards the end of Phase 1 lipids become more abundant and these enter the blood stream as short-chain fatty acids. They are converted in the liver either to acetate (2-carbon) and enter the citric acid or Krebs cycle, generating energy in the form of ATP, or to propionate (3-carbon), which results in amino acids and glucose. If carbohydrate reserves are insufficient to maintain blood glucose levels, tissue and dietary protein may be converted to glucose by a process called gluconeogenesis and the amino groups appear in the blood as ammonia or urea. This can be detected in unfed pouch young by the rise in ammonia in the blood as the liver responds to the lowered blood glucose by converting protein. During the transition that starts at day 200 the liver enzymes responsible for using milk carbohydrates decline to a fraction of their former level while the enzymes responsible for gluco- neogenesis increase several fold (Wilkes and Janssens 1988, Janssens and Rogers 1989). This reflects the profound changes in the constituents of the milk and the increasing demands of the young animal as it becomes physiologically independent, and begins to eat grass.

Water economy and kidney function For very small pouch young the high humidity of the pouch may be important, as mentioned earlier, to help it to respire through the skin during the first few days of pouch life. However, any water loss across the skin is more than balanced by the continuous sucking of the dilute early milk. More important is the loss of water through the immature kidneys, which cannot concentrate urine. When denied milk, young tammars can lose up to 12.5% of their body mass from this cause, but will recover when given milk (Wilkes and Janssens 1988). Passing dilute urine could potentially cause a severe loss of water from the mother–young unit and, in the dry conditions of inland Australia this could compromise their survival. However, it does not happen because the lactating tammar, and probably the females of all kangaroos, ingests all the wastes produced by the young in the pouch. Indeed, a small young will not normally void urine or faeces unless the mother licks its cloaca. This response thus conserves precious water through the concentrating abilities of the mother’s kidneys. This remarkable behaviour was discovered 2 in tammars by labeling the water ingested by the mother with the deuterium isotope ( H20) and 3 labeling water injected into the young with the tritium isotope ( H20). By sampling the blood of mother and young at a later time the tritium label was found in the mother and the deuterium label in the milk and in the blood of the young one (Dove et al 1989). Not only does this behaviour conserve water, it also protects the young from bacterial infec- tions in its gut: any bacteria in the faeces of the young that enter the mother’s body induce an immune response in her and antibodies specific to the pathogen later pass back to the young in the immunoglobulins of the milk, effectively neutralising the infection. This phenomenon is known to occur in cattle and probably also plays a protective role in human infants living in unhygienic conditions. While water recycling is important in the early phases of pouch life, the young must be able to regulate its own urine concentration before the end of pouch life: this requires a functional kidney and hormonal controls. The kidney is composed of hundreds of blind tubules, called nephrons, all connected at their open ends to the ureter, which conveys the urine to the bladder for eventual voiding. The closed 98 Life of Marsupials

end of each nephron is expanded into a cup containing a network of blood capillaries, called a glomerulus, where water, urea and electrolytes pass from the blood into the tubules. The fluid then begins a slow passage along the tubule, where selective reabsorption of water and sodium ions takes place, and urea is progressively concentrated. The development of the kidney and its functions have been studied in three marsupials, the opossum (Krause et al 1979), the quokka (Bentley and Shield 1962) and the tammar (Wilkes and Janssens 1988). For the first week after birth the fetal kidneys are the functional excretory organs, as they were before birth (see Development of the yolk sac, amnion and allantois and Fig. 2.11). By day 20 of pouch life the fetal kidneys have shrunk and have been replaced by the definitive kidneys. For the first 100 days in the quokka and for 140 days in the tammar, new nephrons are being formed in the cortex of the kidney but not thereafter, so that the ratio of kidney mass: body mass declines to the adult value of about 0.6%. At the same time the glomeruli at the closed ends of the nephrons begin to enlarge, which means that filtration across the glomerular membrane from the blood increases. In the tammar this flow is 1 mL/kg body mass per minute at day

2000 a. Urine 1500

1000

500 Plasma Osmolality (mOsmoles/l) 0 b. 800 120 Urea 600 80 400 Urea (mM) 40

Ammonia (mM) Ammonia 200

0 0 0 100 200 300 Adult Age of young (days) Figure 2.23: The development of kidney function during pouch life in the tammar wallaby, Macropus eugenii: (a) urine concentrating ability of the kidney begins after day 200 when the concentration, or osmolality, of the urine can greatly exceed the osmolality of the blood plasma; (b) before day 200 the young tammar excretes ammonia in its dilute urine but after this age it excretes waste nitrogen as urea. After Wilkes and Janssens (1988) and Janssens and Messer (1988). Reproduction and development 99

140 increasing five-fold by day 250 (Wilkes and Janssens 1988). Second, the nephrons increase greatly in length, becoming much coiled in the cortex and with a long loop extending into the inner medulla of the kidney. Now the concentration of urine begins, as a result of the selective reabsorption of sodium in the cortex and of water in the medulla. The main factor that prevents the young tammar from concentrating its urine earlier is the immaturity of the kidney tubules. The hormone that controls fluid retention, vasopressin, is produced by the pituitary gland from a very early age but the kidney tubules cannot respond to it until after day 200 (Fig. 2.23a). As the kidneys develop the ability to concentrate urine, the excretion of nitrogenous waste changes from ammonia to urea. While the urine is very dilute in early pouch life ammonia can be excreted as it is, just as it can in fish and frogs that have no shortage of water with which to dilute it. However, ammonia is highly toxic when concentrated so, as the urine of the growing pouch young becomes progressively concentrated, ammonia in the urine declines and urea increases as the main nitrogenous waste (Fig. 2.23b). From the earliest stages the concentration of sodium ions in the blood plasma exceeds potas- sium ions, the normal ratio being 15:1. This ratio is necessary for the proper functioning of nerve conduction and muscle contraction and is achieved in the adult animal by selective reabsorption of sodium ions across the kidney tubules, so that the ratio of the two ions in the urine is about 1:1. Before 120 days in the quokka and 140 days in the tammar the kidney cannot reabsorb sodium, which is lost in the dilute urine. After this age in both species sodium is conserved and potassium excreted at a higher rate. As mentioned earlier, the early milk of both species has elevated concentrations of sodium, which falls to a low level when the kidney develops the capacity to reabsorb sodium. Presumably the higher concentration in early milk replaces the sodium lost in the urine, which is then recycled back to the mother when she takes in the urine of the pouch young. One consequence of the need to secrete sodium at high concentration in the early milk is the secretion of carbohydrate as long-chain sugars, as discussed earlier (see Carbohydrates in marsupial milk).

Phase 3 of pouch life: growing up and leaving the pouch At 200 days the young tammar is able to maintain a steady body temperature but its preferred temperature, when it uses the least amount of oxygen, is several degrees lower than the tempera- ture in the closed pouch. It needs to dissipate heat but it cannot do this inside the pouch and so it begins to put its head out to cool off. With time more of its body is exposed until it leaves the pouch entirely and begins to hop about and feed on grass. It leaves the pouch permanently at about 250 days, earlier if raised by a large female with more abundant milk, and it is fully weaned by about 300 days. The young tammar continues to take milk for several more weeks after first putting its head out of the pouch, with peak milk intake at 240 days, but herbage progressively forms a greater proportion of its diet. This is a critical period, as the diet changes from rich fatty milk to a diet of grass and herbs to meet the increasing energy demands and increasing independence of the young animal. The herbage must be fermented in the forestomach before being digested in the intestine. Bacteria and protozoa ferment the carbohydrate from the vegetation, especially the cellulose, to volatile fatty acids, and synthesise bacterial protein from urea recycled to the stomach (see Chapter 9). How do these microbes invade the forestomach? Croft (1981b) observed that red kangaroo young at foot would lick the lips of the mother for extended bouts of several minutes, investigating food items in mother’s mouth but more often taking saliva dripping from her mouth. Since this behaviour only occurs in young being weaned, it is possible that its main function is the transfer of microbes to the young. Young and wombats display an 100 Life of Marsupials

analogous behaviour during weaning when they take matter from the mother’s cloaca, which has come from the caecum and is quite different from normal faeces (see Chapters 7 and 8, respectively). In these species the hind gut, rather than the forestomach, is the site of bacterial fermentation. The transition from milk to a diet of grass is complex for all species of kangaroo. The decline in complex sugars in milk is further accelerated by the arrival of bacteria, which consume any remaining carbohydrate. Instead the bacteria produce volatile fatty acids by their anaerobic metabolism of cellulose and this becomes the main source of energy for the tammar, requiring different liver enzymes. The change in diet may actually induce the changes in energy metabo- lism, with gluconeogenesis becoming the main pathway for the production of glucose in the blood. As mentioned earlier, several enzymes that catalyse this pathway show much increased activity in the liver after day 200. It is at this stage of lactation that LLP becomes a major component of the milk of all species of kangaroo that have been investigated (see Proteins in marsupial milk). It is not clear what the role of LLP is in kangaroos but the coincidence of its appearance at the time of transition from milk to a grass diet that is fermented in the forestomach, suggests that it may be important in facilitating the transition. One idea is that it is a source of dietary protein that can be hydrolysed to glucose by gluconeogenesis during the transition to bacterial protein (Janssens and Messer 1988), but this may be too simple an explanation for the function of a complex protein. Another important factor at this transition period, especially for desert kangaroos, is the increased need for water by the young animal as it becomes physiologically independent. Because of its small mass its SMR could be expected to be twice as high as its mother’s: it is actually three times as high because of its rapid growth rate, and its body temperature and evaporative heat loss are also significantly higher, especially at the highest ambient temperatures in summer (Munn and Dawson 2001), so its need for water much exceeds that of adults in the same environment. During hot, dry conditions this puts a great demand on the newly weaned red kangaroo for water, energy and protein. When conditions are good these needs are adequately met from the pasture and milk is not required as a supplement: however, if the pasture quality is low in protein and water, milk can supplement this if the mother can still supply it (Munn and Dawson 2003a), but under such conditions she usually cannot. Hence, the high mortality of young at foot among desert kangaroos (see Chapter 9). Growth in the young tammar accelerates after it leaves the pouch. At 200 days it weighs about 0.5 kg, 100 days later it is 2 kg and at one year it is 3 kg (Janssens et al 1997). By this age the growth of females has slowed down and they enter their first oestrus before they are one year old, whereas the young males continue to grow for another year and do not become sexually mature until the beginning of their third year (Williamson et al 1990). This pattern of sexual dimor- phism in growth and sexual maturity is seen in all macropod species over 5 kg and in the large opossums. In the smaller wallabies and rat kangaroos, however, there is no difference between the sexes in final body size.

Relationship of reproductive processes to body size While this chapter has focussed largely on the development of the tammar wallaby, the general sequence of events, though not their duration, holds for other species that have been studied. Russell (1982) reviewed the available information on 56 species, representing all major families of marsupials, and showed that all parameters, such as litter size, neonatal size, development stages during early life and the length of lactation, correlate with adult body mass and the pattern of maternal care. Thus, species that leave the litter of young in a nest at an early stage of develop- Reproduction and development 101

ment, before the eyes open, such as dasyurids and small didelphids, have a shorter lactation than kangaroos that carry their young to an advanced stage of development in the pouch but, within each family, the correlation with maternal size is close. Likewise, within families, the weight of the whole litter at weaning correlates closely with maternal body mass but the weight of a single young at weaning does not. From this it follows that smaller species, which invariably have more than one young in a litter, make a larger invest- ment proportionately in reproduction than do larger species. Russell (1982) expressed maternal investment as the weight of the litter at weaning as a percentage of maternal body weight, and it varies from over 300% for the smallest dasyurids to less than 30% for the largest kangaroos, wombats and the koala: it is not, therefore, surprising that the smallest species leave their young in a nest from a relatively early stage of development.

Energetics of reproduction Another way to assess the cost of reproduction is to compare the food consumed, the time spent foraging, or the field metabolic rate of lactating and non-lactating females, all of which will be considered in Chapters 3, 4 and 6. For other species the total investment by the female has been measured as the product of the amount of milk produced through lactation and its changing composition. Cork and Dove (1989) provide an interesting comparison between the tammar

Figure 2.24: A comparison of the investment in pregnancy and lactation (expressed as metabolisable energy, ME) in 500 kg dairy cows, 50 kg ewes, 9 kg koalas, Phascolarctos cinereus, 5 kg tammar wallabies, Macropus eugenii and 1 kg common ringtail possums, Pseudocheirus peregrinus. To compare species of very different body mass, time from conception to weaning has been converted to ‘metabolic days’ (d/kg0.25). The three marsupials from different environments and life styles make similar energy investments in reproduction (MJ/kg under the curves) as do the two very different placental mammals, but the investment is differently distributed throughout reproduction. After Cork and Dove (1989), Krockenberger (1993) and Munks and Green (1997). 102 Life of Marsupials

and sheep and cattle (Fig. 2.24): while sheep and cattle make a large investment in reproduc- tion over a relatively short time, tammars make their investment over a much longer time but the total investment, adjusted for body mass, is almost the same in all three species. Subsequent studies on the common ringtail possum (Munks and Green 1997) and the koala (Krockenberger 1993) show the same pattern as the tammar, which supports the idea that this may be a common feature of marsupial reproductive investment: minute investment in gestation, low investment in early lactation and major investment in late lactation. Because the investment in pregnancy and the first half of lactation is small and extended in time, the female marsupial can terminate it in unfavourable conditions with a minimal loss of reproductive investment, and the early phase can be repeated again without much cost. Desert kangaroos have exploited this in their superb adaptation to uncertain conditions and their oppor- tunistic reproductive strategy (see Chapter 9). By contrast the placental female, once committed to pregnancy is unable to terminate without a serious loss of investment. In temperate climates with highly predictable times of food abundance this is not a significant risk, compared to the great advantage that accrues from the rapid exploitation of abundant resources: in the Australian environment of infertile soils and uncertain climate, however, the highly seasonal reproductive strategy of the sheep leads to severe loss in years when rains do not come and is thus not sustain- able in the long term. McNab (1986) considers that marsupials in South America and Australia are competitively equal to their placental counterparts where the resources are limited and mammals need to be parsimonious in energy expenditure: where the resources available permit high rates of energy expenditure marsupials are competitively at a disadvantage compared to ecologically equivalent placentals. These ideas will recur in subsequent chapters as we consider the reproductive strate- gies of different species in various environments. Chapter 3

Opossums of the Americas: cousins from a distant time Pre-European terracotta opossum from Colombia. Opossums of the Americas: cousins from a distant time 105

eventy-six species of marsupial live in South America and Central America, from sea level to the subalpine zone at 4200 m and from the tropics to the cool temperate climate of south- Sern Chile and Patagonia: most are forest dwellers, living on fruits, insects and other small animals. Although widespread, they comprise only 7% of all mammal species of the American tropical region, being far outnumbered by bats (46%) and rodents (27%). In size they range from 10 g to 2 kg and resemble one another in body form, diet and life history (Fig. 3.1, Plates 4 and 5). Nevertheless, the four extant families are not closely related, each having a separate ancestry back to the early Tertiary period: they are the living twigs of a very old tree. One twig consists of a single species from southern Chile, which has closer affinities with Australasian marsupials than with the other American species (see Chapter 1); the second twig comprises seven species of shrew opossums in the northern Andes and southern Chile; the third twig comprises five species of woolly opossums and the remaining 63 species comprise the large opossums and mouse opossums, most of which live in the tropical rainforests, and a few in the grasslands and two in the southern cone. It was not always thus: before the American continents became united 3 million years ago, South America had a more diverse marsupial fauna, which included large carnivores up to 200 kg.

A brief history of marsupials in South America

Origins When South America became separated from North America at the close of the Cretaceous period, 65 million years ago, the few small mammals – placental and marsupial – isolated there evolved independently for the next 35 million years. The marsupials filled the ecological niche of small insectivore, shared the niche of large carnivore with birds, and also edged into the niche of small rodent. Conversely, placentals in South America became large herbivores of various kinds but none became a large carnivore. Unlike in Australia, there is an abundant and continuous fossil history dating back to the very beginning of the Palaeocene epoch in South America, so that the lineages of the present day marsupials can be traced through the whole span of the Tertiary. In one extraordinarily rich fossil site at Tiupampa, Bolivia (see Fig. 1.12), formerly dated to late Cretaceous but now considered to be very early Tertiary, 11 species of marsupials and seven species of placental are known from teeth, skulls and even whole skeletons (Marshall and de Muizon 1988, Goin 2003). Among the marsupials, all of which were between the size of a rat and a small , are representatives of the major lineages that will become dominant in later periods: the opossum-like didelphids, very similar in size and dentition to the living species; a microbi- otheriid, the presumed ancestor of Dromiciops; and another species that could be the ancestor of the shrew opossums, or caenolestids. In addition, there is a species of borhyaenid, the family that was to become abundant later as the large marsupial carnivores. Thus, from the earliest fossil site in South America the lineages that would predominate for the next 65 million years were already established (see Fig. 1.13). In the Palaeocene and Eocene epochs, which encompass the first 25 million years of the Tertiary, 24 genera of marsupials representing six distinct families are known (Marshall et al 1990). The one group that had teeth adapted to eating plants, the polydolopodids, did not survive past the Eocene but the carnivorous species persisted. Polidolopodid and microbiotheriid species have recently been discovered in a middle Eocene formation on the Antarctic (formerly Palmer) Peninsula (Goin et al 1999): all are small and none of the larger borhyaenid carnivores have been found there. Total isolation of South America ended in the Oligocene epoch, about 30 million years ago, when the earliest rodents and primates entered the southern continent (Marshall et al 1979). 106 Life of Marsupials

These were the ancestors of today’s capybaras, Hydrochaeris, and guinea pigs, Cavia, and New World monkeys (Platyrhinni). How they came is still conjectural: presumably by way of an island chain from North America or, possibly, from Africa while that continent was much closer to South America. They may have displaced some of the smaller arboreal marsupials or even prevented marsupials from evolving into the small to medium niches that they came to dominate thereafter. This was the time in Australia when the great expansion of marsupials into arboreal and browsing niches took place. In the Miocene epoch (23–5 million years ago) the shrew opossums (caenolestids) and the large carnivorous borhyaenids, became abundant, both being represented by eight genera. The borhyaenids flourished right through to the Pleistocene epoch (3 million years ago), when they all became extinct; the caenolestids also declined to a tiny remnant of their former numbers and now only persist as seven species in the high Andes and southern Chile. In the Pliocene epoch, small placental carnivores, the ancestors of the coati, Nasua nasua, and kinkajou, Potos flavus, also reached South America and their arrival may have been a factor in the decline and extinction of the borhyaenids. At the same time as these two groups declined the didelphids, forerunners of today’s opossums, increased in number of species, and the present day genera began to appear (see Fig. 1.13). The smaller species, such as Marmosa and Monodelphis, made their first appearance in the Miocene, Lutreolina, Thylamys and Philander somewhat later in the Pliocene, but the other species, including the largest species of opossum, Didelphis, do not appear until the Pleistocene, just 2 million years ago. The apparent ‘relay’ of various carnivorous marsupials through time in South America, with borhyaenids replaced first by large carnivorous birds and later by placental carnivores, and the rise and diversification of smaller carnivorous opossums, is paralleled in Australia, where the early carnivores were large thylacinids that were replaced in the last five million years by the smaller dasyurids (Marshall 1982, Reig et al 1987) (see Chapter 4).

Marsupial top carnivores Borhyaenids were the large mammalian carnivores of the continent for more than 20 million years: they had short limbs and were probably not fast pursuit hunters like wolves but more like badgers or wolverines. Their teeth were very similar to those of the Australian thylacine, cynocephalus, so much so that they were at one time thought to be closely related: but by the Miocene there was no land connection with Australia. The thylacine is now known to be closely related to dasyurids (Krajewski et al 1997), so this is a case of independent convergence in adaptations for predation and eating flesh. But the most bizarre convergence to arise from borhyaenids in South America were the sabre-tooth carnivores that flourished in the Pliocene (5 million years ago), at the same time as placental sabre-tooth tigers, Barbourofelis and Smilodon, were living in North America. Thylacosmilus atrox was a large carnivore with a pair of huge upper canines that sprang from roots lying on either side of the nose and swept across the lower jaw, bearing against a large flange of bone and a tiny pair of lower canines; it had no incisor teeth but a normal complement of premolars and molars (Fig. 3.2). The enormous canines had very thin enamel and had open roots, so that they could grow continuously through life. While its teeth and skull resembled those of the sabre-tooth tigers of North America, Thylacosmilus did not have a tiger’s feet and limbs: instead it had short, strong forelimbs, which suggest that it was not a pursuit hunter but ambushed its prey and stabbed it to death. Its ability to stab was much helped by its enormous (67°) gape. Churcher (1985) estimates that it could have stabbed a body of 400 mm diameter, an animal the size of a deer. The prey available to it was a variety of large placental herbivores, some with long necks like giraffes and others with trunks and some that could run like horses; none was related to the modern animals that they superficially resembled but had evolved independently in South America, along with the carnivorous marsupials. Both Opossums of the Americas: cousins from a distant time 107

the variety of prey species and the marsupials that hunted them disappeared just before or soon after the continent became joined to North America about 3 million years ago. Opinions differ on whether Thylacosmilus was displaced by the placental sabre-tooth tigers coming in from the north, were starved out as their prey were extinguished by the new predators, or disappeared before the great interchange from other causes unknown.

Figure 3.2: Evolutionary convergence of sabre-tooth predators of South and North America from the Pliocene epoch: (a) the marsupial Thylacosmilus atrox; (b) the North American felid Barbourofelis fricki. Note the enormous gape of the marsupial species. After Churcher (1985). 108 Life of Marsupials

One other family of marsupials appeared in the Pliocene, which were different from all other marsupials in South America: Argyrolagus were small, hopping forms, with very short forelegs and elongated hind legs and the toes reduced to the 3rd and 4th and partially fused together. They also had a reduced number of teeth, with only five cheek teeth and one incisor in each jaw, and all open rooted like rodents’ teeth. They were remarkably similar to present day hopping mice but became extinct in the Pleistocene about 2 million years ago.

The great American faunal interchange The greatest change to the mammals of South America occurred when the isthmus of Panama arose from the sea in the Pliocene, 2.5–3 million years ago: this provided a continuous land link between the two American continents for the first time in more than 60 million years (Marshall et al 1979). There had been a few exchanges before this, with sloths, Bradypodidae, entering North America, and small procyonid carnivores, such as the coati and the kinkajou entering South America, but once the land connection was established a flood of species moved south and a lesser number of species moved north: this is called the great American faunal inter- change. Those placental mammals that went south evolved into new species of all kinds in the southern environments, especially in the rainforests but also in the higher altitudes and in the southern grasslands. It is not clear how much this invasion affected the indigenous marsupi- als of the forest but the large carnivorous marsupials became extinct at or before this time. One group of placental mammals that did not penetrate further than northern Colombia were the shrews and moles, Order Insectivora, which are represented in North America by 16 species in 10 genera: their absence further south may be because the small insectivore niche was already fully occupied by many small to medium marsupials, especially the shrew opossums. From South America a smaller number of indigenous mammals, both marsupial and placental, moved north into Central America and a few, like the the armadillo, Dasypus novemcinctus, and the Virginia opossum, Didelphis virginiana, extended far into the northern continent.

Relationships of living American marsupials The fossil history shows that the four families of present day marsupials are not closely related and this has been corroborated by recent evidence from DNA hybridisation and DNA sequence comparisons (see Fig. 1.10) (Kirsch and Palma 1995, Kirsch et al 1997, Jansa and Voss 2000, Patton and Costa 2003). On these criteria the Microbiotheriidae, represented today by Dromiciops, and the Didelphidae separated more than 60 million years ago, during the Palaeocene. Likewise, the differences between the caenolestids and the didelphids put their separation at about 50 million years ago. Both these figures agree well with the evidence from the Tiupampa fossils that these three families were already distinct then.

Natural regions of South America There are four major habitats of the continent (Fig. 3.3). Almost all of the northern two-thirds is the Brazilian zone, comprising the Orinoco and Amazon River basins, from the eastern slopes of Colombia and Peru, across all of Venezuela and Guiana to the eastern highlands of Brazil; it also extends into Central America. The dominant vegetation of this vast region is tropical rainforest in the central part and semiarid grasslands (llanos) on the periphery. It is bounded on the west by the Andean cordillera, which extends all the way to Patagonia in the far south; and on the east by the dry highlands of Brazil, from the Gran Chaco to the Cerrados and Caatinga in the north east. Along the Atlantic seaboard of Brazil there was a coastal forest, the Restinga, now largely cleared for agriculture. The extensive grassland pampas of Argentina extends to latitude 40°S; Opossums of the Americas: cousins from a distant time 109

beyond this is the dry rainshadow region of Patagonia to the east of the southern Andes. Most marsupial species dwell in the northern rainforests, while a few species live in the drier regions of the Caatinga and the southern pampas; only 10 species live in the cool temperate regions of the Andes and the far south, and their collective distributions leave much of the southern cone devoid of marsupials. Thus, despite their long occupation of South America, few of its marsupials have

Venezuela 10 10

Llanos French Guiana

Cali Eastern Colombia

0 0

NE Peru Semi-arid Amazon Basin Caatinga Andes

s -10 o -10 d a rr Pacific e lands Coastal C Plain TIUPAMPA Eastern High Atlantic coastal forest -20 -20 Rio de Janeiro

ITABORAI

-30 -30

Southern Buenos Cone Aires

-40 -40

RIOCHICAN

CASAMAYORAN

Patagonia -50 -50

Figure 3.3: Major vegetation regions of South America, separated by lines and with the high altitude regions shaded. Sites (O) where research mentioned in the text has been conducted; major Tertiary sites (V) where fossil marsupials have been found. After Patterson and Pascual (1972). 110 Life of Marsupials

truly adapted to life in a cool temperate environment (McNab 1986). This is curious because most of the sites of fossil marsupials from the distant past are located in the south or west, away from the present centre of distribution (Fig. 3.3). This may merely be because the tropical rainfor- est is not a suitable site for fossilisation or to recover fossils from. Alternatively, it may indicate that there have been considerable changes in the climate of the southern sites since the fossils were laid down and that the southern cone and western Antarctica was the centre from which mammals spread northwards into the rest of the continent and also into Australia. Indeed, the great Argentinian palaeontologist, Carol Ameghino, who discovered many of the fossil sites, postulated this more than 100 years ago and was derided by other scientists for it.

The cool temperate marsupials of South America The shrew opossums, Caenolestidae, and the sole member of the Microbiotheriidae are restricted to the cool temperate region but only two species of the Didelphidae occur in it.

High altitude species The climate above 3000 m in the Andes is cold and wet and the vegetation is sometimes described as ‘elfin forest’: it consists of low, dense shrubs covered in mosses and ferns beneath which are tunnels and galleries made by the small mammals that live there. This is the parãmos, the home of the obscure silky shrew opossum, Caenolestes fuliginosus, and four related species. They are obscure because they are very difficult to catch and almost impossible to study alive. The few people who have attempted to study them have been frustrated because none has lived for more than a few days in captivity, and no long series of animals have been collected from one site so as to understand their ecology, behaviour or reproduction. From the first two specimens collected during the 19th century, however, it was recognised that these animals were very differ- ent from all the other marsupials of South America because of the pair of large forward directed incisors in the lower jaw, which resemble those of the Diprotodontia in Australia. These and other features of the skull pointed to their close relationship to fossil marsupials from the early Tertiary Casamayoran formation in Patagonia (Fig. 3.3). The name given to the first species reflected this: Caeno is derived from a Greek word meaning new or modern and lestes (a thief) is a suffix often applied to small predaceous fossil marsupials, so the name means a modern member of an ancient group. Since the first two specimens were collected from the northern end of the Andean range in Colombia and Venezuela, other species of Caenolestes have been described from the parãmos of Colombia, Ecuador and northern Peru, and another species, Lestoros inca, from southern Peru (Fig. 3.4). These species live at altitudes between 1800 and 4200 m; the only other caenolestid is the Chilean shrew opossum, Rhyncholestes raphanurus, which lives at a lower altitude but in a similar habitat in southern Chile and is separated from the northern species by the desert regions of northern Chile and Peru.

Caenolestes fuliginosus The best known species is the silky shrew opossum (Fig. 3.1h, Plate 5), the first to be discovered (Osgood 1921). Based on stomach contents of recently caught specimens, its diet consists of a variety of insects but when captive animals were presented with live rats they killed them by stabbing with their procumbent lower incisors and then bit off pieces of the head and body with the cheek teeth (Kirsch and Waller 1979). At the end of its meal the silky shrew opossum rubbed its muzzle on the cage floor to remove blood and washed off the rest by licking the forepaws and rubbing them over the snout. This suggests that they probably eat a variety of Opossums of the Americas: cousins from a distant time 111

small vertebrates and invertebrates. Only four sets of data on reproduction of the silky shrew opossum have been collected, all from the parãmos of Colombia, by different people and in separate years (Tyndale-Biscoe and Renfree 1987). Adult males had enlarged testes and very large prostate glands in all months, and far more males were captured than females on each occasion. One pregnant female with three embryos was obtained in February 1910, another female at oestrus with four large follicles was collected on 30 April 1971 and four lactating females, without attached young, were collected in August–September 1969. Females do not have a pouch and

10 N 10 N

Caenolestes

0N 0N

10 S 10 S Lestoros inca

20 S 20 S Thylamys elegans

30 S 30 S

Dromiciops gliroides 40 S 40 S Rhyncholestes raphanurus

Lestodelphys halli

50 S 50 S

Figure 3.4: Distribution of marsupials that live in the cool temperate regions of the Andes and the southern cone of the continent. After Hershkovitz (1972), Hunsaker (1977), Patterson and Pascual (1972) and Streilein (1982b). 112 Life of Marsupials

only have four teats in the inguinal region, three or four of which were enlarged in the lactating females. Taken together, these scattered observations point to a seasonal breeding, beginning in February and ending in September. If this is correct, it implies one small litter a year and a very slow rate of development.

High latitude species

Rhyncholestes raphanurus Unlike the other caenolestids, the Chilean shrew opossum is found from sea level to about 1000 m from latitude 40°30'S to 43°30'S in southern Chile and on the island of Chiloe. It occurs in a range of habitats including sub-Antarctic rainforest, and is most frequently caught on the ground in dense cover or near den sites. Its diet includes , insects, and fungi. The snout is very narrow and the pair of long stout incisors in the lower jaw may be used, as forceps, to pull worms and other prey from crevices. At the back of the mouth there are two flaps of skin, as in other caenolestids, past which the food is directed onto the molars. In animals collected in late summer and autumn the base of the tail was enlarged with stored fat, while other animals collected after winter had thin tails. This fat storage is not seen in other caenolestids and has been interpreted as evidence that these shrew opossums may undergo winter torpor or hibernation (Patterson and Gallardo 1987). Temple-Smith (1986) had evidence that males live for only one breeding season and that females produce a single litter of up to seven young (there are seven teats but no pouch).

Thylamys elegans The elegant fat-tailed opossum, which weighs about 30 g, has an extensive distribution on the western side of the Andes from 14°S in Peru to 36°S in Chile, north of Rhyncholestes (Fig. 3.4). They have a very small home range of less than one hectare and they build nests of hair and leaves in dens in rocks or trees or in the burrows of guinea pigs. Like the other southern species, their tails fatten in winter, reaching maximum thickness in August when they enter torpor (Palma 1997). There is no evidence that they share nests during torpor, as do small hibernating possums in Australia (see Chapter 6). Females have 11–13 teats but no pouch and the breeding season begins in September, in early spring, and extends to the following March, during which time two litters are produced. In one 10-year study in south-western Chile the population fluctuated from peaks of 20/ha in three of the years to a minimum of less than 3/ha in the intervening years (Lima et al 2001). These marked changes correlated tightly with the prevailing rainfall, associ- ated with the events of the Southern Oscillation Index, or El Niño. Because the habitat where the species lives is semi-arid thorn scrub, rainfall is critical to the abundance of insect food on which the species depends: larval insects rich in energy become available in the winter and support the growth of adults prior to the onset of reproduction, while adult insects in spring and summer provide the protein requirements of lactation and the growth of the young. Conversely, in years of low rainfall population increase is small because reproduction is reduced and the maturation of young animals is slowed down. During the dry seasons adult opossums survive by undergoing torpor and drawing on reserves in the fat deposited in their tails, and reproduction ceases.

Lestodelphys halli The Patagonian opossum is closely related to Thylamys elegans, but larger, and is the only didel- phid marsupial that lives south of 40°S (Fig. 3.4). It is known from fewer than 10 specimens, collected in Patagonia between 33°S and 48°S, the most southerly locality of any marsupial in South America (Marshall 1977) and indeed the world, since it is further south than most of Tasmania. Little is known about it but it is thought to prey on small mammals and birds: it has Opossums of the Americas: cousins from a distant time 113

larger canines than other didelphids in a shorter and more massive jaw, its feet are stronger than those of Thylamys and it has exceptionally large auditory bullae, the part of the skull that encloses the inner ear, and presumably contributes to acute hearing. In these respects it more closely resembles the smaller dasyurids of Australia, such as Phascogale and Dasyurus (see Chapter 4). Like the other southern species its tail becomes fattened before winter.

Dromiciops gliroides The monito del monte, or little monkey of the mountains, Dromiciops gliroides, is the sole living representative of the Family Microbiotheriidae. Fossils species included in this family are known from the earliest formations of the Tertiary of South America, going back 60 million years. In addition, fossils that may be related to Dromiciops have been found at Murgon, Queensland, in eastern Australia that are 55 million years old. This marsupial holds an unique place in the history of marsupials because several important features link it more closely to Australasian marsupials than to South American species (see Chapter 1): it does not have paired sperm like all didelphids and caenolestids and its DNA has greater affinity to the diprotontid marsupials than to the other South American groups. Curiously, Oldfield Thomas anticipated this affin- ity in 1894, when he named it Dromiciops because of its close resemblance to the Australian , Dromicia, and –ops meaning ‘appearance of’ (Marshall 1978a). (Please note that Dromicia is now called , see Chapter 6). The two species do look superficially alike: both are less than 30 g body mass, both have short fat tails and both hibernate in the winter months in a nest. However, Dromiciops has more than one pair of incisors in the lower jaw and it does not have fused toes on its hind feet, like its Australian namesake. Dromiciops occurs only in south-central Chile from Concepcion (36°S) to the island of Chiloe and east to the Andean border with Argentina (Fig. 3.4). In this region it lives in bamboo thickets of the moist, cool temperate forests, which share many plant species with the same forest type of Tasmania and New Zealand. Like most other South American marsupials its diet consists of insects, especially beetle larvae and pupae, but in captivity it will eat a variety of other food, including fruit, oats, potato and minced meat. It constructs a nest 200 mm in diameter from the water repellent leaves of the bamboo, which protects it from rain. The nest chamber is reached through a small opening and the outside is covered with moss, which provides excellent camouflage (Mann 1958). Nests may be shared by a pair of animals and are also used by breeding females for the suckling young. Animals have been found torpid in a nest as have animals held in traps but the degree to which they hibernate in winter is still unclear. Females have a well-developed forward directed pouch containing four teats, again like Australasian marsupials. The scattered observations on reproduction by nine different people since 1893 indicate that breeding begins in the southern spring in November and ends in May as the first snow falls. The litter size in the pouch ranges from 1 to 4; after they leave the pouch the young stay together in the well-formed nest and older young may be carried on the mother’s back. It is not clear from these observations whether a female can produce more than one litter in a year, or what the life span of the adults is.

Conclusion These four marsupial species from the extreme southern cone of the continent show similar adaptations for low temperatures and food shortages in winter: all are small, store fat in the base of the tail, and can enter torpor and fairly prolonged hibernation. Since they are from three distinct families of American marsupials, each with a long separate ancestry, these common features of their biology are examples of adaptive convergence to an extreme environment. The only other American marsupial that has adapted to a cool temperate climate is the Virginia opossum in North America (see The Virginia opossum goes north). 114 Life of Marsupials

The neotropical marsupials

Family Didelphidae This is the largest family of American marsupials. The dentition is common to all species: adults have 50 teeth, which comprise five pairs of incisors in the upper jaw and four pairs in the lower, followed by a pair of prominent canines in each jaw then three pairs of premolars and four pairs of molars (Fig. 3.5). The dentition of sharp pointed simple teeth reflects the generalised diet of insects, small vertebrates and soft plant tissue. The sequence of eruption of the premolars and molars and their subsequent wear has been used in several species to determine the age of specimens caught in the field and so understand the rate of turnover of the population. For species of Didelphis the seven dental age classes have been related to actual age (Fig. 3.5). The full complement of teeth is achieved at about 9 months (class 4) and the subsequent three classes are

Figure 3.5: The upper premolar and molar teeth of a series of the common opossum, Didelphis marsupialis, to show the eruption of successive molars, the replacement of the deciduous premolar (dP3) with the permanent P3, and the wear on the cusps of the oldest animals that can be used to tell the relative age of the specimens’ and the differential growth of the upper canines in males and females after sexual maturity at dental class 4. After Tyndale-Biscoe and MacKenzie (1976). Opossums of the Americas: cousins from a distant time 115

more than 10 months old. Males grow to a larger size than females and have larger canines, these differences appearing at sexual maturity. The 63 species in the family Didelphidae are grouped on size and associated characters into three subfamilies (Table 3.1), which have had separate lineages for about 12 million years (Kirsch and Palma 1995, Kirsch et al 1997).

Table 3.1: Genera of American marsupials From Kirsch and Waller (1979), Harder and Fleck (1997), Kirsch and Palma (1995), Fonseca et al (2003).

Family Subfamily No. of No. of Mass (g) chromosomes species (N) (N) Microbiotheriidae DromiciopsA 14 1 16–30 Caenolestidae Caenolestes 14 5 20–30 Lestoros 14 1 20–30 Rhyncholestes 14 1 20–30 Didelphidae Marmosinae Marmosa 14 9 10–130 Micoureus 14 4 80–150 Monodelphis 18 16 20–150 Thylamyinae Thylamys 14 5 20–40 Lestodelphys ? 176 Marmosops 14 11 50–80 Gracilinanus ? 6 20–30 Didelphinae Metachirus 14 1 300–750 ChironectesA 22 1 650 DidelphisA 22 4 500–5000 LutreolinaA 22 1 500–800 PhilanderA 22 4 200–660 Caluromyidae Caluromyinae CaluromysA 14 3 200–400 Caluromysiops ? 1 [200]B Glironinae Glironia ? 1 [<200] Totals: 4 7 19 76 AFemales carry young in a pouch B[x] values in [ ] = mass estimated from body length.

The subfamily Didelphinae includes 11 species over 500 g body mass. The four species of Didelphis are the largest living American marsupials: the common or black-eared opossum, Didelphis marsupialis, occurs throughout the northern parts of South America below about 1000 m and extends through Central America as far as Mexico; it is replaced in eastern Brazil by the very closely related big-eared opossum, Didelphis aurita, which extends to 27°S (Fig. 3.6). The third species is the white-eared opossum, Didelphis albiventris (Fig. 3.1a, Plate 4), which partly overlaps the other two species but tends to occupy the higher altitudes in the west of the continent, drier regions in the east and cooler regions in the south: in Colombia it only occurs 116 Life of Marsupials

above 1000 m and in eastern Brazil it occupies the semi-arid Caatinga region in the northeast and extends to the deciduous forests of Argentina at 38°S. The fourth species, Didelphis virgin- iana, occurs in Central and North America and is thought to have evolved from the common opossum, D. marsupialis, within the last 2 million years. However, when all four species were compared by DNA hybridisation (Kirsch et al 1997) and DNA sequence (Patton and Costa 2003), the northern species appears to have branched off from the other three species much earlier than

1958

1982 40 N 40 N 1500 1870 Didelphis virginiana

Didelphis marsupialis 0N 0N

Didelphis aurita

Didelphis albiventris

40 S 40 S

Figure 3.6: Distribution of the four species of Didelphis across both continents, and showing the northward spread of the Virginia opossum, Didelphis virginiana, since 1500. After Hershkovitz (1972) and Gardner (1982). Opossums of the Americas: cousins from a distant time 117

this and before the southern species had separated from one another. At present this discrepancy between the evidence of the fossil record and molecular techniques cannot be reconciled. The most closely related to these are the four species of gray four-eyed opossum, genus Philander (Fig. 3.1b, Plate 4), which are smaller than the opossums but very similar in appear- ance to them and have the same distribution as the common opossum. Two other related species are morphologically different: the comadreja or marsupial weasel, Lutreolina crassicaudata, which has short legs, a broad tail base and very short, rounded ears, like a weasel or ferret; and the water opossum, Chironectes minimus, which is the only marsupial adapted to an aquatic life (Figs 3.1c, d, Plate 4). Unlike the rest of the South American marsupials, these 10 species have 22 chromosomes and the females have a well-formed pouch. More distantly related to these eight species is the brown four-eyed opossum, Metachirus nudicaudatus, which is the same size as the gray four-eyed opossums but is wholly terrestrial; unlike them, it has the primitive number of 14 chromosomes and the females do not develop a pouch (Table 3.1). Although it is less common than the gray four-eyed opossum, it is very widespread in primary rainforest throughout northern South America and Central America. Lemos et al (2001) compared the shape of the skull in these five genera and concluded that their evolution has been rapid and occurred during the past 5 million years, which certainly corroborates the evidence from the fossil record. Three of these genera are each represented throughout northern South America by a single species, and Philander by one very widespread species and three rare and highly localised species; only Didelphis is represented by three species in the Neotropics and a fourth species in North America. This is in sharp contrast to the members of the two other subfamilies, the Marmosinae and the Thylamyinae in which most genera contain many species (Table 3.1). There are 52 species of pouch-less mouse opossum in the size range of 10 g to150 g, all but two of which are forest dwellers. Formerly, all these species were included in two genera, Marmosa and Monodelphis (Fig. 3.1f, Plate 5), but they are now divided into two subfamilies and seven genera on the basis of DNA relationships (see Fig. 1.10) (Kirsch et al 1997, Jansa and Voss 2000). The subfamily Marmosinae includes the true mouse opossums Marmosa, the woolly mouse opossums Micoureus and the short-tailed opossums Monodelphis. The second subfamily Thylamyinae includes five species of Thylamys; the single species of Lestodelphys, already mentioned; 11 species of slender mouse opossum, Marmosops (Fig. 3.1g, Plate 5); and the six species of gracile mouse opossums, Gracilinanus. Apart from Monodelphis species, which have 18 chromosomes, all the other mouse opossums that have been studied have the ancestral number of 14 chromosomes.

Family Caluromyidae The fourth family of South America marsupials are closely related to the Didelphidae: while the earliest known fossils go back only to the Pleistocene (Marshall 1982), DNA hybridisation puts the separation at 49 million years (Kirsch et al 1997). The family includes three species of woolly opossum, Caluromys (Fig. 3.1e, Plate 5), which are distinguished from all other South American marsupials by their predominant use of fruit, nectar and gum as food and by their strong prefer- ence for living in the canopy of large forest trees: they seldom come to the ground and, for this reason, are rarely caught in traps either set on the ground or 2 m up in trees. Related to these three species are two rare species, the black-shouldered woolly opossum, Caluromysiops irrupta, which is found in southeast Peru and adjacent parts of Brazil, and the very rare bushy tailed opossum, Glironia venusta. These species do not have a permanent pouch but Caluromys females develop one from folds of skin on either side of the abdominal teats during lactation. 118 Life of Marsupials

The neotropical rainforest environment The tropical rainforests of Central and South America are highly diverse ecosystems with a greater number of plant and animal species than anywhere else in the world. The marsupials that live there are no exception, and at any one site the number of species can range from 5 to 10, with the record number being 17 from a single river valley in eastern Peru (Fleck and Harder 1995). The longest continuous study of one of these forest communities has been conducted at two sites in French Guiana, one in undisturbed primary rainforest and one in second growth forest altered by human occupation (Charles-Dominique et al 1981). Situated north of the Equator (5°N, 53°W) the temperature throughout the year is about 25°C, with marked dry and wet seasons: the dry season lasts from August to October, when the rainfall is about 100 mm per month, whereas for the rest of the year the rainfall is about 400 mm per month. This determines the abundance of fruit and insects on which the mammals depend: fruit of many species is abundant from September to May and scarce from June to August, when however, many trees are in flower and nectar is available. Insects are least abundant at the end of the wet season, April–June. The body mass of the marsupials, especially the terrestrial species is lowest in June–August, recovering by September. The incidence of breeding rises in October at the beginning of the wet season. Such a bare statement, however, does not well describe the complexities of a forest habitat in which most of the 130 species of trees, shrubs and lianas are uncommon and widely dispersed and only a few species are common. Each species in the primary forest has its own cycle of flowering and fruiting, so there are successive and overlapping periods of fruit ripening, which ensures an almost continuous supply of fruit for the mammal dispersers. Some species produce tiny seeds enclosed in the fruit, which are ingested mainly by rodents and widely distributed; these are generally the pioneer species, commonly found in the secondary forest. Other species produce fruit with large seeds, which are ingested with the surrounding pulp, but chemical deterrents prevent unripe fruit being eaten. In primary forest the fruiting sites are less numerous, more patchy, but most trees have greater biomass and shorter fruiting periods, so that crop size is much higher in primary forest than for secondary forest.

Diversity and relative abundance of marsupials in tropical forests Ten species of marsupial live in the different layers of the undisturbed forest of French Guiana: some live on the forest floor, some in the understorey shrubs and some live exclusively in the canopy of the emergent forest trees, which are up to 30 m tall. Furthermore, within each layer of the forest there are pairs of marsupial species of similar size that appear to occupy the same space, but one is rare and the other relatively common: and it is only four of the commoner species that are found in the disturbed, secondary forest environments (Table 3.2) (Charles- Dominique 1983, Julien-Laferriere 1991). Thus, in the undisturbed forest the variety of species is greater, although the common species are much less abundant than they are in the altered forest habitat. The three smallest species are the delicate slender mouse opossum, Marmosops parvidens (15–30 g), and the short-tailed opossum, Monodelphis brevicaudata (Fig. 3.1f), which are rare; and the murine mouse opossum, Marmosa murina (40–60 g), which is common. These species species live in the lower levels of the forest in dense vegetation, feeding on small insects. The next pair are the rare long furred woolly mouse opossum, Micoureus demerarae (= cinerea) (80–150 g), which extends from the understorey to the canopy and overlaps with the larger, common, bare-tailed woolly opossum Caluromys philander. The next group of large species are the gray four-eyed opossum, which is common in the understorey and on the ground, and the uncom- mon and terrestrial brown four-eyed opossum and the aquatic water opossum. The remaining pair is the common opossum, which occupies every level, and the rare, terrestrial, white-eared opossum. Of these 10 species the four common species, namely the common opossum, the gray four-eyed opossum, the bare-tailed woolly opossum and the murine mouse opossum are the Opossums of the Americas: cousins from a distant time 119

habitat generalists and the more adaptable species, and it is only these species that are common in secondary forest.

Table 3.2: Marsupials in forests of French Guiana

From Julien-Laferriere (1991). Habitat: 1, primary forest; 2, altered secondary forest. Strata: 1, terrestrial; 2, scansorial, understorey vegetation; 3, canopy.

Species Mass (g) Habitat Strata Abundance Food Marmosa parvidens 20 1 1, 2 + { M. murina 40 1, 2 1, 2 ++

insects, fruit Monodelphis brevicaudata 80 1, 2 1 +

Micoureus demerarae 150 1, 2 2, 3 + Caluromys philander 300 1, 2 2, 3 +++ fruit, nectar, gum Metachirus nudicaudatus 350 1 1 + { Philander opossum 450 1, 2 1 +++ fruits Chironectes minimus 450 1 1, water + invertebrates Didelphis albiventris 1100 1 1 + small vertebrates D. marsupialis 1100 1, 2 1, 2, 3 +++

A similar habitat in lowland eastern Peru is the rainforest bordering the Marañon River, a tributary of the Amazon, except that an annual cycle of flooding during the wet season compli- cates matters for the terrestrial mammals. As in Guiana, the main fruiting occurs during the wet season and, while the arboreal species can feed on the fruit, the species on the forest floor have to migrate to higher ground and live on other food. Eight species of marsupial were trapped on the 11 ha study site, the most abundant being, as in French Guiana, the common opossum, the gray four-eyed opossum, its close relative, Philander andersoni, and the white bellied slender mouse opossum, Marmosops noctivagus (Fleck and Harder 1995). At a second site down river from this one, only three species were trapped and these were the larger species: the difference between the two sites may have been due to the regular flooding of the down river site during the wet season, when the terrestrial species are obliged to leave the low lying areas and move to higher ground.

Restinga forest of the Atlantic seaboard of Brazil One of the last remaining large tracts of Atlantic forest, at Teresopolis, near Rio de Janeiro (23°S), carried a community of seven species of marsupial, of which the big-eared opossum is the most abundant, while the nearby secondary forest carried fewer species but the numbers of each species were greater, as in French Guiana (Davis 1945; Cerqueira et al 1993; Gentile and Cerqueira 1995). Similar communities of marsupial species of varying body mass and abundance living together in one piece of forest, as in French Guiana, Peru and the Brazilian restinga, obtains at other sites within the Amazonian region in Panama, Nicaragua and Venezuela (Harder and Fleck 1997). While the individual species vary from one site to another, all primary lowland forests tend to have a range of genera of different body size represented by one or more species: there will be one or two species of mouse opossums of each of the genera Marmosa, Microureus and Marmosops with one species of each of the larger opossums Didelphis, Philander, Metachirus, Caluromys, and sometimes one of the short-tailed opossums, Monodelphis. Two species occurred at all six sites and two others at four sites, while the rest of the 18 species listed were present at one to three sites each. 120 Life of Marsupials

The curious feature of all these forest communities is the large number of species that appear to share the same habitat and to live on the same food resources, namely a variety of insects, small vertebrates and any fruit that happens to be ripe and available. From detailed studies on the food of the common opossum, the white-eared opossum and the gray four-eyed opossum in the restinga forest of Brazil, the species of insects and plant fruits selected during both wet and dry seasons of the year were almost identical (Caceres and Monteiro-Filho 2001; Caceres 2002; Santori et al 1997). With such a broad overlap how have different species managed to evolve in this environment? The conventional wisdom is that one species will displace a less competitive species for the same ecological niche but this clearly does not happen in the South American forest communities. Perhaps the precise niche of the various species needs to be better defined, either in the type of food eaten or the space within the forest. For instance, the Matses people of the Peruvian rainforest recognise at least 40 habitat types within flood-prone and upland forest and know that opossums move from one to another through the year in relation to flooding and the fruiting of particular species of tree (Harder and Fleck 1997). Cunha and Vieira (2002) attempted to define habitat use more precisely: using the spool and line technique they traced the movements of the several species within the forest layers and the amount of time spent in each part of the forest. Three species lived exclusively in the canopy and could not be studied by this technique but four species that lived on the forest floor or in the understorey were accessible: as in French Guiana, the grey slender mouse opossum, Marmosops incanus, occupied the understorey, while the larger big-eared opossum and gray four-eyed opossum, Philander opossum, were more terrestrial, although both also extended into the under- storey and the big-eared opossum reached the canopy; the brown four-eyed opossum was exclu- sively terrestrial. The subtle differences between the four species seemed to be in their respective use of tree trunks – the smaller species used small branches while the common opossum used larger branches and more vertical trunks of trees. Thus, by each using different parts of the forest plants they may be able to share the resources in the three-dimensional world they occupy. To get a clearer understanding of these interactions six representative species, which have been well studied, will now be looked at in more detail.

The gray short-tailed opossum, Monodelphis domestica Of the 16 species of short-tailed opossums, two are widespread: the red-legged short-tailed opossum Monodelphis brevicaudata (Fig. 3.1f, Plate 5) occurs in the western half of the Brazilian zone, where it lives exclusively on the forest floor and, unlike other marsupial species, is active during the day, especially the late afternoon and evening. The larger gray short-tailed opossum lives in the semi-arid Caatinga of eastern Brazil, where it occurs in all habitats from granite outcrops, low and tall thorn scrub to the remnants of the Atlantic rainforest (Streilein 1982b). It has also followed people into cleared land and lives in and near human habitation; but it occurs most frequently in rocky areas, where it finds shelter, which presumably enables it to reduce its water loss from evaporation in the dry environment. In a laboratory study of its water balance Christian (1983) concluded that it could survive without access to free water on a high protein diet, which was up to 66% water. Since its normal diet is insects, frogs, lizards and small rodents, as well as fruits and seeds, it is likely that it meets its water needs from its prey and is not dependent on free water in its natural environment. Because of its wide choice of food it forages over a considerable area and is not restricted to a small home range: this was reflected in the low number of trapped animals that were recaptured in Streilein’s (1982b) study. He observed that the main period of activity was 1–3 h during the early evening with briefer bouts throughout the night. It is an active hunter of live prey, which it pins to the ground with its forepaws before biting on the neck or base of the skull. Its forefeet are also nimble enough to capture insects as small as 1 mm and moths in flight. Opossums of the Americas: cousins from a distant time 121

From limited observations it seems that gray short-tailed opossums are solitary and intoler- ant of other members of their species, which they repel with an open mouthed threat or, if that is insufficient, direct attack. They mark objects in their environment by rubbing the side of the head or chin against them and males will also rub the chest and scrotum against an object or the ground; whether this is a territorial or sexual signal is not known. However, in captivity the growth of young females is slower and the onset of sexual maturity retarded in the absence of adult males, and adult females will only come into oestrus and ovulate in the presence of a male, an effect induced by a pheromone released by the male (VandeBerg 1990).

Breeding, development, and longevity For a marsupial, the gray short-tailed opossum has an unusually fast development. As in other opossums, the gestation period is very short, 15 days, but after birth the young only remain attached to the teats for 16 days before being left in a nest by the female. However, they continue to depend on milk entirely for the next 20 days and thereafter are intermittently suckled for a further 20 days, before being weaned at about 60 days of age. Thus, the entire cycle from oestrus

Figure 3.7: Changes in body mass and resting metabolic rate of gray short-tailed opossums, Monodelphis domestica, during gestation and lactation, compared to non-reproductive females. After Harder et al (1996). 122 Life of Marsupials

to weaning is only 75 days: this is about half the duration of the breeding cycle in the larger opossums and less than half the duration of the bare-tailed woolly opossum. Like all other small opossums, the female does not have a pouch but carries the young for the first 16 days attached to the teats. As they grow the female adopts a particular posture, like female Marmosa also do, which holds them off the ground. Females have up to 13 teats but the litter size is usually less than the total, the mean being eight young. The combination of large litters and a short breeding cycle means that the reproductive potential of this species is very high. In captivity on a constant day length of 14 h and with unlimited food, females breed continuously and produce three or four litters in a year (VandeBerg 1983). Males and females are sexually mature at five months and in captivity live for less than four years. This rapid rate of development clearly taxes the resources of the female, a matter investigated by Harder et al (1996). The mean body mass of females increased from 65 g before oestrus to 77 g at the end of gestation and 87 g in the last week of lactation (Fig. 3.7). At the same time the mass of a litter of eight young increased from less than 1 g at birth to 140 g at weaning, or almost double the mass of the mother supporting them. The rich milk supplied during the middle period of lactation, when the carbohydrate reaches 13%, sustains this large growth (Crisp et al 1989). In achieving this reproductive investment the metabolic rate during gestation was 40% higher, and in late lactation 77% higher, than the rate of non-breeding females. In the natural environment food resources must determine whether the female can make this large investment or not. There are two studies of reproduction in wild populations, from the same region of eastern Brazil where the captive animals originated. The climate in this region (8–10°S) has an even temperature of about 25°C, with a dry season from August to January and a somewhat wetter season from February to July. However, the maximum rainfall is only 80 mm per month and even this is uncertain, so it is an unpredictable and semi-arid environment. Streilein (1982a) collected lactating females in all months of the year except August and concluded that breed- ing was continuous. However, in different years Bergallo and Cerqueira (1994) concluded that breeding began at the onset of the wet season in January and ceased in September, with independent young appearing in the population from April until October. Gray short-tailed opossums were most abundant in July, when the older age classes were well represented, but declined steeply during the dry period after October, which suggests that most adults die in the dry season and the population recovers the next year when the young animals breed. The large reproductive potential of the species combined with a severe attrition during the dry season probably accounts for the pattern described. What it does suggest is that the gray short-tailed opossums in this harsh environment only breed in the year following their birth. A similar pattern occurs in another species, Monodelphis dimidiata, much further south in Argentina (37°S) (Pine et al 1985): a single large litter is produced in December–January and by May all the adults have disappeared from the population. The weaned young grow slowly through the winter, their growth accelerates in the spring and they then produce the next genera- tion in the summer. This is similar to the unusual strategy of the Australian species of Antechinus (see Chapter 4), except that female as well as male opossums apparently die at the end of the single breeding episode.

From thorn scrub to biomedical research The gray short-tailed opossum was given the specific name, domestica, because it is common in and around human habitation (Streilein 1982b, Emmons 1990). Perhaps because of this it was easier to domesticate than other species of American marsupials. In 1978 nine animals, collected in Pernambuco, were successfully bred in captivity at the National Zoological Park, Washington, and from a nucleus of their offspring an inbred laboratory colony was established in Texas Opossums of the Americas: cousins from a distant time 123

(VandeBerg and Robinson 1997). Although the gene pool was subsequently augmented with a further 29 animals from Brazil and Bolivia, the nine founder animals have contributed almost half of the eventual gene pool of the colony of several thousand animals, which is now highly inbred. Many other colonies have been derived from it so that this small opossum is now a widely used laboratory species with more research papers published each year on it than on any other marsupial. While several other species of American and Australian marsupials have been bred in captivity, none has adapted to long-term captivity as well as the gray short-tailed opossum has done: perhaps being a commensal of human habitation in eastern Brazil may have predisposed it to captivity. In this respect it is like the laboratory rat and mouse, which also arose from species that lived in the human environment before being domesticated for research purposes.

The bare-tailed woolly opossum, Caluromys philander The bare-tailed woolly opossum (140–300 g) (Fig. 3.1e, Plate 5), lives in the canopy foliage more than 10 m above the ground and feeds largely on fruit, nectar and insects. Its main competitor in the forests of French Guiana is the kinkajou (Charles-Dominique et al 1981), which is about six times its mass (2.5 kg) and also feeds on the fruit of the larger trees of the primary forest. The kinkajou preferentially feeds in trees more than 25 m tall, whereas the bare-tailed woolly opossum will feed in the lower storey as well.

Social behaviour and home range The bare-tailed woolly opossum is exclusively nocturnal and solitary, its social behaviour being limited to mating and mother–young interactions. No mutual contact, such as mutual grooming, has been observed between adults, even by mating animals. Their repertoire of calling consists of a weak sexual call by the male and an aggressive distress call by both sexes, which varies in strength from a weak hiss to a loud staccato call audible up to 200 m distant; no other marsupial in these forests makes such a long-distance call (Charles-Dominique 1983). Adults will mark branches by dribbling urine, which may help to identify the home range but is not a territorial signal declaring to other bare-tailed woolly opossums to stay out, since home ranges of individu- als of both sexes overlap. The home range has been measured by following individuals carrying radio transmitters and by traps set 10–30 m above the ground and baited with banana. The size of the home range is constrained by the availability of den sites and abundance of fruiting trees: in the primary forest, where fruiting trees are widely dispersed the home range varied from 1.3 to 8.9 ha (mean 3.3 ha), with the animals using about 1.1 ha of their home range each night (Julien-Laferriere 1995). The core area, in which they spent about 70% of their time, contained significantly more feeding sites but not more den sites than the part of the home range outside it. In secondary forest, where fruiting trees are smaller and more numerous, the home range was much smaller (0.75 ha) (Atramentowicz 1982), presumably because the animals do not need to move so far to find food: during the dry season, when fruit is scarce in the secondary forest, the home range doubled. The use of space in the canopy by woolly opossums is thus highly opportunistic, which may explain why the species has colonised such a wide range of habitats from northern Venezuela to southern Brazil. It is also of interest that the home range of the bare-tailed woolly opossum and its density in the forest are very similar to the values for the similar sized sugar gliders of Australia, which also live on a diet of plant exudates and insects (see Table 6.1).

Living on a diet of fruit Bare-tailed woolly opossums become active at dusk and visit up to four trees in a night, seeking ripe fruit or flowers. Of the 39 available species of fruiting trees bare-tailed woolly opossums visited 23, some very rarely and others frequently: they prefer the less brightly coloured fruits, 124 Life of Marsupials

while birds and monkeys that feed by day prefer the coloured fruits (Julien-Laferriere 1999). When the fruit supply was abundant, from November to May, females in early lactation and males spent about 5 h per night feeding, whereas females in late lactation spent 9 h feeding. Conversely, in the dry season from June to October when fruit is scarce all animals spent 9 h feeding (Atramentowicz 1982), mainly on nectar from flowering trees and the gum oozing from the trunks of one tree species, Fagara rhoifolia. This diet of fruit pulp, nectar and gum, which is rich in sugars and lipids, provides bare-tailed woolly opossums with all their energy and water requirements, but it is very deficient in protein (0.39–0.71% N dry weight) (Julien-Laferriere 1999). Insects in the canopy provide up to 25% of the diet during the wet season and so provide part of the protein needs but they also are least abundant in the dry season, so that bare-tailed woolly opossums lose up to half their weight at this time of the year. The protein deficiencies of the diet affect the bare-tailed woolly opossum’s digestive physiology, its reproduction and the growth and development of its young.

Digestive physiology In order to discover what the minimum nitrogen (N) requirements of bare-tailed woolly opossums are, captive animals were fed a diet of mango pulp mixed with graded amounts of casein to achieve diets containing 0.45%, 0.9% and 2.5% N (Foley et al 2000). On these diets the opossums modified their food intake so that they maintained N balance: on the low N diet they ate more of the mixture, which increased the sugar intake, but passed it through the gut faster. The faster rate of passage caused more cells lining the gut to be sloughed, however, so the non- dietary faecal N increased, which aggravated the problem of achieving N balance. Despite these three effects of eating a low N diet, the animals were still able to maintain a positive N balance on a very low daily intake of protein, namely 176 mg dietary N/kg0.75. This is substantially lower than daily values of dietary N measured in leaf-eating marsupials of a similar size, such as ringtail possums, Pseudocheirus peregrinus (380 mg/kg0.75) and greater gliders, Petauroides volans (700 mg/kg0.75) (see Table 7.1) but is greater than the values for sugar gliders, Petaurus breviceps (87 mg/kg0.75), which also depend on a diet rich in carbohydrate and relatively low in protein (see Chapter 6). Conversely, bare-tailed woolly opossums given the high N diet ate less of it to stay in positive N balance and their body mass increased during the trial. Since there is so little protein in the fruit, which is their staple diet, insects must be their main source of protein and are probably the limiting resource for the animal’s survival and ability to breed, especially for females in the second half of lactation. While there may be sufficient food available for the female to live on, there may not be sufficient protein to support the additional demands of the growing young in the final stages of nursing (see Chapter 2).

Capturing the fruit for breeding success Reproduction in the bare-tailed woolly opossum is affected by the quality of the diet in three ways: the average litter size of three or four young is less than the six available teats in the pouch and smaller than the litter size of the other opossums in the forest; the development of the young is also slower than in its relatives that live on meat; and a second litter in the year is a rare event. The main breeding season begins in October and 80% of the females have young in the pouch by November, after the onset of the wet season; the main burden of late lactation comes 40 days after birth of the litter and coincides with the time when fruit and insects are most abundant in the canopy (Fig. 3.8). The increased burden for the female during late lactation is evident in the greater time she spends foraging each night in an enlarged home range. The rate of development in the bare-tailed woolly opossum is also slower than in other opossums. The length of gestation is 25 days, almost twice that of all other didelphids; growth rate of the young in the pouch is initially slow but accelerates after day 40, coinciding with the Opossums of the Americas: cousins from a distant time 125

Figure 3.8: Success of breeding in the bare-tailed woolly opossum, Caluromys philander, in relation to food abundance through the year in French Guiana. The first litters (G1) are reared to independence and go on to breed in the following year but the second litters fail as food declines in June. After Atramentowicz (1982). increase in the mother’s foraging activities and food intake; the young make their first appearance outside the pouch at day 80 and at day 100 they are left in a nest where they cling to each other; the mother returns to the nest to feed them in the middle of the night and again at dawn for 2–3 h. They are weaned at 120 days and almost all then disperse: only 1 of 60 young marked in the pouch was trapped on the study site after weaning. Females are sexually mature at nine months but do not breed successfully until their second year (Charles-Dominique 1983, Atramentowicz 1982, Atramentowicz 1995, Julien-Laferriere 1995). This whole rate of development is much slower than that of the ground-living common opossum in the same forest, which usually has two litters of six young, each produced in 90 days, and that of the gray four-eyed opossum, which has three litters of four to five young in a year, each taking 70 days. In both these species the young females are mature at six to seven months (Charles-Dominique 1983). An interesting point explored by Atramentowicz (1992) is what the extra cost is to the mother bare-tailed woolly opossum of carrying a large litter, using all six teats, instead of the usual litter of three. She found that there was no measurable cost to the mother bare-tailed woolly opossum, which neither foraged for longer nor lost weight through the extra burden, but the growth of the young was slower: the available resources were shared between the larger litter and they were weaned at a smaller size. In a good season they might be able to recover from this slow start but in a poor season or a second litter their survival might be lower than members of a small litter. After weaning the first litter a female bare-tailed woolly opossum will start a second litter but the fate of these young depends on the available fruit crop at the end of the wet season: unless the fruiting season is exceptionally heavy the young fail to complete pouch life or the subsequent phase in the nest (Fig. 3.8). Thus, the reproductive rate of the bare-tailed woolly opossum in the canopy is less than half that of the gray four-eyed opossum on the forest floor, but the life expectancy of those that survive to independence is considerably greater. After one year only one-third of the population had been replaced on the study site compared to the much higher turnover of gray four-eyed opossums and common opossums (Atramentowicz 1982, Julien- Laferriere 1995). 126 Life of Marsupials

In summary, it takes a female bare-tailed woolly opossum 110–125 days to produce a litter of four young, so that her second litter usually succumbs to food shortage and the young of the year do not breed until the next year, whereas the female gray four-eyed opossum and the common opossum produce up to 14 young in a year and the young of the first litter reproduce in the year of their birth (Table 3.3). However, the bare-tailed woolly opossum grows slower, has a longer period of dependency, a higher metabolic rate and lives longer than the gray four-eyed opossum, which breeds fast and dies soon.

Table 3.3: A comparison of the three commonest marsupials in primary and secondary forest in French Guiana, showing the much greater carrying capacity of the secondary forest for these three species From Julien-Laferriere and Atramentowicz (1990), Julien-Laferriere (1991). Numbers in italics indicate values that have been derived from measured data.

Caluromys Didelphis Philander philander marsupialis opossum Litters/year 1 2 3 Primary forest Density, No./ha 0.51 0.22 0.17 Biomass, g/ha 160 210 70 Litter size 3.5 5.9 4.8 Young/& per year 3.5 11.8 14.4 Secondary forest Density, No./ha 1.43 0.45 1.37 Biomass, g/ha 350 410 560 Litter size 4.4 4.3 4.3 Young/& per year 4.4 8.6 12.9

The gray four-eyed opossum, Philander opossum At 300–700 g the gray four-eyed opossum (Fig. 3.1b, Plate 4) is the same size as the bare-tailed woolly opossum that lives above it in the canopy: it can range into the understorey but lives mainly on the forest floor where it takes invertebrates and fallen fruit, as well as small rodents, grasshop- pers and eggs; it commonly forages near streams but is not otherwise aquatic (Atramentowicz 1988). It is agile, quick and more alert than other opossums and is generally solitary. For daytime shelter it builds a nest of dry leaves in hollow trees, fallen logs and the burrows of other animals (Castro-Arellano et al 2000). Gray four-eyed opossums get their odd name from the bright spots of white fur above their eyes, like the markings on spaniel dogs, but whether this deceives other animals into thinking their eyes are open when they are asleep is not known. Like the common opossum, it is not territorial: the home ranges of individual animals overlap and vary in size according to available resources. In Panama the home range was 0.34 ha while in French Guiana it was between 0.8 and 2.5 ha, which is relatively large, possibly because it is searching for animal prey. There was a high degree of mobility, with less than 20% of the animals staying in the study area for more than 200 days (Charles-Dominique 1983). In primary forest of French Guiana the density of the gray four-eyed opossum was 0.17/ha, which equates to a biomass of 70 g/ha (Julien-Laferriere 1991), whereas in the secondary forest the density was eightfold greater (1.37/ha) and the biomass was 560 g/ha (Table 3.3). At other parts of its range the density fell within these limits: in Mexico the density was 0.48/ha and in Panama it was 0.55/ha in primary forest and 0.65/ha in secondary forest. Opossums of the Americas: cousins from a distant time 127

The gray four-eyed opossum has three successive litters in the year during the months of abundant food, each taking 68–75 days, followed by 8–15 days in a nest; the litter size is four to five, which means that it has a potentially high fecundity, but mortality in the pouch determines the rate of increase: breeding starts during the dry season, and may be initiated by the seasonal conditions. In French Guiana the first litter is produced in August–September, coinciding with the main fruiting season, and survival of the young is very high. The second litter follows in November–December and the third at the end of the wet season in April. In the primary forest, where the density was low, all the second and third litters survived through to weaning but in the secondary forest, where the density was eight times higher, 60% of the second and third litters were lost and in another 10% some in the litter were lost. Here is a paradox: in the primary forest the resources are more scattered and the population is low so that competition for food resources is low and the smaller number of litters have a higher chance of survival, while in the secondary forest with abundant food and a population eight times as great, mortality in the pouch is much higher. Males are sexually mature at seven months, females at six months, so the young of the first litter can breed in the season of their birth. However, this first litter of the young female usually dies because the food resources are in decline by then. After one year the entire population had been replaced on the study site compared to only one-third of the population of bare- tailed woolly opossums in the same forest. Life expectancy in the wild is about 2.5 years (Julien- Laferriere and Atramentowicz 1990).

The water opossum, Chironectes minimus Closely related to the gray four-eyed opossum and of the same mass is the more specialised water opossum, which feeds on aquatic animals, such as small fish, frogs and small mammals. It is rare in the primary forest and not present in the secondary forest. It is the only marsupial that has special adaptations for life in water (Fig. 3.1c, d, Plate 4). It has webbed hind feet, which it uses exclusively for propulsion when swimming, the front feet being splayed out in front of it ready to seize any prey that it may encounter as it swims. It can dispense with the use of its forelegs in swimming because it is unusually buoyant: air trapped in its unwettable fur enables it to float with its entire head and body at the water surface, at an angle of about 3°. This enables it to breathe continuously while swimming without any special exertion to stay afloat (Fish 1993). Its maximum swimming speed of 2.6 km/h is almost double the maximum speed of the Virginia opossum (1.7 km/h), which uses all four legs to swim. Because its fur becomes wet and provides no buoyancy, the Virginia opossum must paddle with its forelegs to keep its nose above water and swims at a steeper angle of 11° to the water surface. The movement of its limbs while swimming is the same as if it were walking on land except that opposite fore and hind limbs sometimes make contact, which disrupts the stroke and slows its movement. The water opossum has other adaptations for an aquatic life: its body is streamlined and its tail is flattened laterally and is used as a rudder; its eyes are positioned high on the head, its whiskers are robust and its ears are small and closable; the pouch of the female is said to be especially water- proof, due to contraction of strong muscles in the pouch wall that close its mouth and, together with the waterproof fur, prevent water entering the pouch and affecting the young within (Enders 1937). However, the pouches of most marsupials, including the Virginia opossum can also be tightly closed, as Reynolds (1952) showed when he found the gas composition within the pouch to be different from the outside air. But the more extraordinary adaptation is that the adult male water opossum has a pouch anterior to the scrotum and when it swims the testes are drawn up against its belly and the mouth of the pouch is partly closed around the scrotum, supposedly to aid in streamlining the body in the water (Enders 1935). However, this oft quoted and remarkable feature of the male water opossum has only once been recorded from actual observation. 128 Life of Marsupials

The large opossums, Didelphis marsupialis, Didelphis aurita and Didelphis albiventris The three southern species of Didelphis have been studied at several localities from Panama, Colombia, Venezuela and Guiana in the north, through the dry Caatinga of eastern Brazil, the Atlantic forest of southeastern Brazil and as far south as Buenos Aires (Fig. 3.3). There is a common pattern in the diet, density and home range of the three species through this broad range of habitats, which reflects their opportunistic nature. In the forests of French Guiana the common opossum, which lives predominantly on the ground, is an opportunistic feeder, eating carrion, insects, fallen fruit or large earthworms that emerge after rain; in another study in southern Brazil the big-eared opossum and the white-eared opossum (Fig. 3.1a, Plate 4) lived together in the same 5 ha of forest where they ate the same range of insects, small vertebrates and fruits (Caceres and Monteiro-Filho 2001, Caceres 2002); in the dry Caatinga the white-eared opossum was opportunistic in what it ate, concentrating its efforts on whatever fruit was abundant. White-eared opossums will kill small vertebrates, which they find by smell and sound: when offered small venomous pitvipers, which live in the same habitat, the snake would be seized near the head, shaken violently and killed in 9s; it would then be eaten, beginning at the head (Oliveira and Santori 1999). The same approach is used to kill mice and small birds (Streilein 1982b). The other species may also feed on snakes since the serum of common opossums contains a fraction that neutralises snake venom and even pouch young may be protected, as the fraction has been detected in opossum milk (Jurgilas et al 1999).

Abundance and home range In the primary forest in French Guiana the abundance of the common opossum was about 0.22/ ha, biomass 210g/ha (Table 3.3), and in the secondary forest about twice as abundant, 0.45/ha (410 g/ha), reflecting the greater amount of food in the disturbed environment (Julien-Laferriere 1991). Nevertheless, the density at both sites is very much lower than was found in lowland forest (0.9/ha) and grasslands (1.5/ha) in Venezuela by O’Connell (1989), or of the white-eared opossum in the Caatinga (0.4–4.4/ha). In the latter the variation was because, during the dry season, the opossums were more concentrated in rocky outcrops, presumably to conserve scarce water (Streilein 1982b). The home ranges of opossums have been difficult to determine because the animals are not readily recaptured: this may be because they do not stay in the same place for long or because they do not live very long. In a one-year study of the big-eared opossum in Brazil (Caceres and Monteiro-Filho 2001), the females retained small home ranges of about 2 ha each throughout the year, which did not overlap with adjacent females.The males were present only during the wet season, when breeding occurred, and their larger ranges overlapped those of the females. Apart from the brief sexual encounters, opossums are solitary and repulse other opossums. By follow- ing common opossums fitted with radio transmitters, Sunquist et al (1987), discovered that the nightly movements in a Venezuelan llanos habitat were more than 1 km and the home ranges of females were 16 ha and males 123 ha, very much greater than estimates based on trapping. Like the other studies the much larger home ranges of males overlapped those of several females. There is a common pattern of populations fluctuating through the year with a peak in numbers when the young of the year first appear and then a decline as the adults disappear: most authors conclude that opossums do not live for more than one breeding season. The rapid turnover is associated with a relatively high reproductive rate, keyed to the annual cycle of plant productivity, as with the bare-tailed woolly opossum.

Breeding patterns in opossums Reproduction in the three species of opossum has been studied in eight populations north of the Equator and in eight populations south of the Equator. The observed variations are in the litter size, the number of litters per year and the time of onset of breeding (Table 3.4). Opossums of the Americas: cousins from a distant time 129

Table 3.4: Summary of breeding in Didelphis marsupialis (m), Didelphis albiventris (al) and Didelphis aurita (au) in relation to latitude and the annual dry season (shaded) at each site

Extent of breeding period represented by B, with those in bold indicating main peaks of birth.

Latitude Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Species Litter Reference 12°N B BBBB BBB B m Biggers (1967)

B BBB BBBBB m 6 Fleming (1973) 10°N BBBB m 7.5 { O’Connell 8°N BBBB m 5.9 (1989)

Julien-Laferriere,

Atramentowicz 5°N B B BB B BBBBm 5.9(1990) 4–5°N B B B B B B B B B m 6.5 { Tyndale-Biscoe, 4°N B BBB BBB Bm4.5MacKenzie

4°N BBBB B al 4.2 (1976)

8°S B B B B B al 6.2 Streilein (1982a) 8°S B B B B al 4.5 {

Cerqueira 9°S BBBBBBB al 4.5 (1984) 12°S BBB B B B al 5 Talamoni, Dias 21°S B B B B B B B al 6.2 (1999) Cerqueira et al 23°S B B B B B B B au – (1993) 8.5, 23°S B B B B B B B au 7.1 Davis (1945) Regidor, Gorostiague 35°S B BB B al 6.9 (1996)

In French Guiana (5°N) the common opossums have two litters per year; in the main one beginning in August more than 90% of females breed while at the second in February less than half of the females had young in the pouch. But for those that bred at either site there was little loss of young during pouch life. However, there was a significant difference in the mean litter size between sites and between the first and second litter, with the first litter at both sites being smaller than the second. Similarly at O’Connell’s (1989) two sites in Venezuela (10°N) the litter size was 7.5 in the forest site and 5.9 in the grasslands. The onset of breeding varies in different parts of the continent but begins at the end of the dry season and late lactation occurs during the wet season when fruit is abundant: in French Guiana the dry season is from August to November; in eastern Peru from June to October; in Panama, eastern Colombia and Venezuela the dry season is from December–January to March– April. Conversely, in Nicaragua and in western Colombia where the rainfall is uniform through- out the year, breeding is more extended and common opossums have three peaks of birth in January, May and August. 130 Life of Marsupials

South of the Equator the pattern is different: in eastern Brazil at 8°S the dry season is from August to December and the females of the white-eared opossum produce a single litter during January to March, whereas near Rio de Janiero (23°S) the dry season is from June to August and the main litter of the year is produced in July and a second one in October; and near Buenos Aires (35°S) two litters are produced, one in September and a second in November. The question is whether the trigger for the start of breeding in these three species of opossum is the beginning of the rainy season and the associated abundance of fruit that follows, or whether opossums are sensitive to the changing photoperiod, which becomes more pronounced further south. No experiments have been done to differentiate between these two possibilities. However, in the gray four-eyed opossum, which has a similar pattern of breeding, one experiment was conducted in which captive animals were provided with abundant food throughout the year and were subjected to the ambient photoperiod of 23°S (Cerqueira et al 1993). Despite having abundant food at all times none of the females bred during March to June, the normal non-breeding period in the wild for that latitude, which suggests that photoperiod signals are more important than nutritional state of the females. While this may be so at the Tropic of Capricorn, where there are marked changes in photoperiod through the year, it is most unlikely that seasonal breeding near the Equator is regulated by the small difference in photoperiod: it is more likely that the predictable sequence of dry and wet seasons provides the necessary signal for the various species of marsupial to initiate breeding at the close of the dry season.

Invasion of Central and North America since the Pleistocene

Common marsupial species that went north After the Panamanian isthmus arose eight marsupial species that are widely distributed in rainforest communities in South America spread northwards: the common mouse opossum, Marmosops impavidus, extended only into Panama, the brown four-eyed opossum and the slender mouse opossum, Marmosops invictus, reached Costa Rica; however, the Central American woolly opossum, Caluromys derbianus, the water opossum, Chironectes minimus, and the mouse opossum, Marmosa robinsoni occupied all available rainforest habitat in Central America to its northern limit in the Yucutan Peninsula and central Mexico. The common opossum and the gray four-eyed opossum also occupied the rainforest habitat but extended their ranges beyond it to the northeast, into the humid montane forests of eastern Mexico up to elevations of 1200 m (Fig. 3.6).

Four marsupial species arose in Central America Four marsupial species, not known from South America, have presumably evolved in Central America since the Pleistocene. Three of these are mouse opossums: Micoureus alstoni, which is closely related to the widespread long furred woolly mouse opossum, occurs in rainforest from Panama to the Yucatan; Marmosa mexicana extends further into the lowland forest of eastern Mexico, like the common opossum; and Marmosa canescens occurs in this habitat and in the lowlands of western Mexico.

The Virginia opossum goes north The fourth newly evolved species is the Virginia opossum, which is the only species to have extended its range far beyond the ancestral tropical forest habitat (Fig. 3.6). It was formerly thought to be a subspecies of the common opossum, but Gardner (1973, 1982) showed clearly that it is a distinct species. His attention was first drawn by the observation that six pairs of its Opossums of the Americas: cousins from a distant time 131

chromosomes are very different from those of the common opossum, the white-eared opossum and the gray four-eyed opossum. Whereas these three species have 22 single arm chromosomes (ie the centromere of each is at the end of the chromosome), 12 of the 22 chromosomes in the Virginia opossum each have a long and short arm with the centromere between them. These differences would make it difficult, if not impossible, for this species to produce fertile offspring with any of the other species and so there is now a genetic barrier between it and its nearest relative, the common opossum. Having made this discovery, Gardner then noticed other more subtle differences in the anatomy, behaviour and physiology of the Virginia opossum, when compared to the common opossum, and showed that both species occur together in a wide zone of Central America: this runs from the northern boundary of the common opossum in eastern Mexico to the southern boundary of the Virginia opossum in western Nicaragua (Fig. 3.6). The question is how did the new species arise and then become established in what was presumably an isolated population of the common opossum? In Mexico the Virginia opossum occupies a much wider range of cooler and drier habitats, from sea level to 3350 m than the common opossum does. Gardner’s (1973) hypothesis was that small populations of the common opossum became isolated in western Mexico at various times during the four glacial periods of the last half million years and became adapted to the prevailing cooler climates; at the same time chromosomal changes occurred in the isolated population, which established the genetic barrier to subsequent interbreeding with the common opossum in the next interglacial period. During these warmer periods the new and now more adaptable species spread out in two directions into temperate habitats that the common opossum could not occupy: one was along the northwest Sonoran coast of Mexico as far as the deserts of southern California, and the other was northeast around the Gulf of Mexico to Florida. While the Virginia opossum is the best-known American marsupial and has sometimes been referred to as ‘a living fossil’ because of its presumed ancient nature, it is actually a very recently evolved species; the oldest known fossils being from the third or Sangamonian intergla- cial period, between 120 000 and 75 000 years ago, so the present fossil evidence goes back less than half a million years. However, according to DNA hybridisation and sequence comparisons, the Virginia opossum separated from the three South American species of Didelphis 4.5 million years ago (Kirsch et al 1997; Jansa and Voss 2000). These two discordant pieces of evidence could be reconciled if the ancestors of the Virginian opossum reached Mexico by island hopping, well before the great American faunal interchange and before the South American opossums had become three species: it could then have adapted to the cooler climates in the uplands of Mexico, as Gardner suggests, and only after the land connection formed and the other marsupial species spread into the rainforests of Central America and Mexico, would it have mingled with the common opossum. The salient fact, however, is that the Virginia opossum now occupies a huge area of cool temperate habitat that is different from that occupied by its tropical cousins. So, what in its ecology and physiology enabled the Virginia opossum, with such an heritage, to become so firmly established in the northern continent against the competition from pre-exist- ing placental species; and what now limits its further spread north?

Distribution The Virginia opossum may have reached Florida 50 000 years ago but its further expansion northwards probably did not occur until the end of the fourth Ice Age after people arrived from Asia 12 000 to 20 000 years ago (see Chapter 10). The disappearance of many of the large herbiv- ores within 1000 years of human arrival must have caused profound ecological changes in the continent, which would have benefitted a generalised omnivore like the opossum. Support for this view comes from the occurrence of opossum bones in Amerindian kitchen middens between 4000 and 500 years ago. Before European arrival the opossum was a common item of food in 132 Life of Marsupials

the lower and middle Ohio valley as far as the southern shore of Lake Erie but was not on the Appalachian Plateau or north of latitude 40°N on the eastern seaboard (Fig. 3.6) (Guilday 1958). Since European settlement, the range of the opossum has continued to extend northwards: in the 19th century it had reached latitude 42°N and in 1903 it was reported that, ‘with the deforest- ing of the mountains [opossums] are invading large areas of the Allegheny regions previously unknown to them’. (Guilday 1958). The northward movement has continued through the 20th century, albeit slowly (about 10 km/year), reaching 43°N by 1958 and 45°N by 1982 (Gardner 1982). The northern limit appears to be climatic, since it coincides approximately with the –7°C January isotherm, and in Michigan this is approximately the pine–hemlock ecotone between northern coniferous forest and the southern deciduous forest. For different reasons the desert country of northwest Mexico and New Mexico were also natural barriers to the spread of the Virginia opossum through the Rocky Mountains into California. If compared with species adapted to arid environments, the opossum has only a moderate ability to concentrate urine in its kidneys and conserve water and so could not penetrate into those parts that have a low rainfall. However, when a few opossums were released near Los Angeles and San Francisco, between 1870 and 1915, the species spread rapidly through the west: by 1932 it was widespread from the southern border of California to a few kilometres north of San Francisco and extended into the foothills; by 1958 it had spread at an average rate of 50 km/ year north into British Columbia and eastwards everywhere to an altitude of 1500 m. But it did not extend north of the –7°C January isotherm in either Washington State or British Columbia; nor in the south has it joined up with the Mexican population across the desert regions of south- ern California. As in the eastern half of the continent, the two separate constraints of low winter temperature and aridity have determined its final distribution in the west.

Ecology In the southern parts of its distribution in Nicaragua, Texas and Florida, the Virginia opossum is very like its South American congeners: the average adult body mass ranges up to 2 kg and females produce two or three litters per year, each of about six young. However, in the northern parts of its range, adult opossums weigh up to 5 kg, very much larger than any other species of opossum (Gardner 1982). The tails of northern opossums are relatively shorter than those of southern animals, but the ears are not and, in winter, they become frost bitten. Thus, despite its recent arrival in North America, the Virginia opossum follows Bergman’s and Allen’s Rules that mammal species at high latitudes are heavier and have relatively smaller extremities than the same species near the Equator. Other adaptations for a cold climate are not well developed: the fur is sparse and although it is thicker in winter and opossums put on a layer of body fat in autumn, the loss of heat, or thermal conductance, is higher than in better adapted species (Hsu et al 1988). Also opossums cannot hibernate, like the small marsupials of the southern cone of South America, or placental mammals in the far north. The lack of these adaptations in the Virginia opossum supports the view that the species has only recently moved north from its ancestral tropical environment. Other features of its southern origins, however, that have favoured its spread north are its opportunistic omnivore diet, its high fecundity and its rapid population turnover. Its size and generalised dentition enable it to use a wide range of food from insects and earthworms to small vertebrates, as well as fruits and grasses. The remains of adult rabbits and larger mammals found in opossum stomachs are probably carrion since they are unlikely to kill animals of this size. However, like its southern cousins, it can kill and eat venomous snakes; and like them its food preferences change seasonally, insects predominating in the warmer months and mammals in the cooler. Virginia opossums are active at night and live alone by day in a variety of dens: they do not excavate burrows themselves but appropriate those of other species, such as woodchucks, Opossums of the Americas: cousins from a distant time 133

Marmota monax, or use old bird nests or rock shelters. Like their South American cousins, Virginia opossums are not territorial and the home ranges, which are much larger, overlap with those of other opossums. This may reflect the lesser availability of food or the greater require- ments of animals living in a cool temperate climate. Indeed, the main challenges for the Virginia opossum have been adapting physiologically to the extremes of the northern winters and repro- ducing successfully in such environments.

How Virginia opossums survive the winter Virginia opossums can maintain their body temperature at 34°C when exposed to an ambient temperature of 0°C, but they use four times as much oxygen to do it as they do at 30°C (Lustick and Lustick 1972). During the winter in Iowa opossums were active and emerged from dens to forage on days when the temperature was above –7°C but were seldom active if it was colder than this and never if the temperature was below –12°C (Wiseman and Hendrickson 1950). Nardone et al (1955) found that opossums held at –10°C could only maintain their body temperature at 34°C for 20 minutes: thereafter it fell in 1 h to 25°C and in some cases to 7°C. At these low body temperatures the heart slowed down from the normal rate of 240 beats/minute to 60 beats/minute and the electrocardiogram became abnormal. Opossums cannot hibernate and as these results show, they cannot remain active, or indeed survive, at ambient temperatures below –10°C; so sheltering in dens where the temperature is well above 0°C is essential to avoid winter extremes and maintain a normal body temperature economically. Brocke (1970) calculated that opossums held at 0°C required 845 kJ per day to maintain normal body temperature. He also estimated that the energy intake from a winter night’s foraging by an opossum would be about 1380 kJ. Thus, at the moderate temperature of 0°C an opossum would have to forage on two nights out of three in order to retain a positive energy balance. Since winter temperatures at the northern limit of its range are considerably lower than this its survival in this environment is only possible if the den temperature is near the opossum’s thermoneutral temperature of 25°C, and if it can forage every night. But at –10°C this is impossible, so the need to forage on warmer nights becomes impera- tive. In Iowa during the winter many opossums had frost damaged ears and tails, indicating that they were forced to forage on cold nights, and by spring they were in very poor condition. If an opossum starts the winter in good condition, it can also draw on its body reserves, and Brocke (1970) calculated that one-third of its winter energy needs were met from these reserves; but it still needs to forage on at least half the nights in the three months of winter. He showed that its distribution in Michigan closely followed the line connecting those parts of the country where at least half the nights of the winter months had a daily minimum temperature above 0°C. But merely surviving the winter is not enough: opossums must also breed, and at the northern limit of their present distribution the summers are short, so with their three month lactation they must start to breed soon after winter, when they are still in poor condition.

How opossums breed in the north

Breeding season and number of litters In southern Texas and Florida, females come into oestrus in early January and most births occur before the end of that month, but in California, Missouri and Pennsylvania the first births occur one month later, and in New York State not until the end of March (Hossler et al 1994). After weaning the first litter the female will return to oestrus, so that a second peak of births occurs three to four months after the first. In Texas some females can occasionally produce a third litter in July but further north, where the second litter is weaned in August, all females enter winter anoestrus then. Furthermore, even the second litter is rare in Ohio, Maryland and New York State. The size of the litter, however, is larger in the north than in the south so the annual 134 Life of Marsupials

fecundity is about the same throughout the opossum’s range. The main constraint on repro- duction in this species, as in other marsupials, is the cost of supporting the growing litter in the second half of lactation by providing rich milk to them. The milk of the Virginia opossum changes through lactation with protein and lipid increasing after 60 days and until weaning at about 90 days (Green et al 1996). The cost of this reproductive effort by the female opossum was estimated by Fleming et al (1981) to be twice the metabolic cost of a non-lactating female. Thus, the availability of food is what determines the time of onset of breeding, the size of the litter, and whether there is a second litter.

Fertilisation and male competition Female opossums have an oestrous cycle of about 28 days and they will accept the male for less than one day in the cycle. Furthermore, the eggs are shed within 12 h of the onset of oestrus and are rapidly covered by a mucin coat that is impenetrable to sperm (Rodger and Bedford 1982a). For these reasons there is only a brief period when fertilisation can occur, so competition between males for access at this time is intense. In one study where the movement of males and females were followed by radio tracking (Ryser 1992), the home ranges of adult males trebled during the breeding season to encompass the home ranges of about six females, and they moved up to 9 km as they searched throughout the night for females coming into oestrus. During the night of oestrus several males would gather near the female and engage in fighting, the largest male usually gaining exclusive access before midnight. Copulation takes place between midnight and dawn, matings later than this being usually infertile. Nevertheless, the male follows the female into her den after mating and remains with her through the next day, denying access to other males. In this competition there is clearly a selective advantage in being large and this is presumably the reason that male opossums grow faster and larger than females and are armed with much larger canine teeth (Fig. 3.5).

Size of litter Hossler et al (1994), who studied the species near the northern limit of their range, found that the size of the litter was highly correlated with the nutritional state of the female in early lacta- tion, as measured by hind leg fat index: the thinnest females had litters of seven while the fattest had litters up to 14 (Fig. 3.9). But how does a female opossum regulate the size of her litter? Although females ovulate far more eggs and grow far more fetuses than the number of teats in the pouch, it is rare for all teats to be occupied, so the litter size must be determined very soon after birth. There are two ways by which this could occur: some of the mammary glands fail to lactate in response to the sucking stimulus and the young attached to them die, or the female might remove some young. While some Australian marsupials deliberately reduce their litters during lactation, in the opossum there is little change in the size of the litter through lactation and it is more likely that the litter size is set at the time the young are born. When the litter size is less than the number of teats it is the posterior ones that are more commonly occupied than the anterior ones, and young on these teats are larger than those on the anterior teats (Cutts et al 1978), which indicates a gradient in the potential of the mammary glands to lactate. Although not so far examined, there could be a gradient in the concentration of prolactin receptors on the mammary glands at the time of birth (see Chapter 2), determined by the nutritional state of the female, with the posterior ones having a higher concentration than the anterior ones: this could affect the size of the litter she carries or the order of subsequent reduction in the litter, if resources are insufficient to support all the young. There is a similar gradient in the performance of mammary glands in lactating pigs, in which each piglet has its own teat and those on the most posterior teats are the smallest in the litter and the first to die if conditions deteriorate – the so-called ‘tail end Charlie’. Opossums of the Americas: cousins from a distant time 135

Figure 3.9: Correlation of maternal condition, as measured by the thickness of the subdermal fat of the thigh, and litter size in Virginia opossums, Didelphis virginiana, in New York State. After Hossler et al (1994).

Longevity After the initial period when lactation is established, the survival of young in the pouch is very high – between 92 and 98% – but less than one-quarter of juveniles survive the period of weaning (Hossler et al 1994). After this exceptionally vulnerable period, survival through the winter is higher and about 10% of opossums survive to one year, when they participate in the next breeding season (Fig. 3.10). Although the maximum longevity is about four years, few opossums on the mainland live beyond the summer of their second year, so each animal that survives its first winter probably makes a contribution to only one breeding season and the population turnover is rapid. This is very similar to the other species of Didelphis and the short- tailed opossums that live in the tropics, and raises the question whether opossums intrinsically age fast or whether their high mortality is due to external causes, such as predators, diseases and bad weather. Austad (1993) examined this question by comparing the rate of mortality of opossums on the mainland with a population isolated on an offshore island, where predation was much less. While the climate at the two sites was similar, the life expectancy of female opossums over one year of age on the island was double that of female opossums on the mainland (Fig. 3.10); they aged more slowly, and they were more successful at breeding in their third year. Conversely, the average litter size on the island was significantly smaller than on the mainland (5.7 versus 7.6) and the rate of growth of pouch young was slower. This may have been caused by greater compe- tition among females for food, since the density of opossums on the island was four times that of the mainland. What these results show, however, is that opossums do not have an intrinsic propensity to die at the end of their first breeding season, and when the environmental condi- tions are more favourable, they can live longer and breed in two or even three successive years. This is very different from what occurs in the males of some dasyurid marsupials in Australia: 136 Life of Marsupials

100 Weaning

80 Independence

60

40 Survival (%)

20 Island Mainland 0 0 12243648 Age in months

Figure 3.10: Life expectancy of Virginia opossums, Didelphis virginiana, in northern regions of its distribution. The first part of the curve shows the severe mortality after weaning; the second part, assumes that 20% have survived to one year and shows the difference in the subsequent survival of adult females in a population on the mainland, and on an island where there were few predators. After Hossler et al (1994) and Austad (1993).

after a single brief breeding season the immune system of these males is severely compromised and they disappear from the population in a short time (see Chapter 4). Most Virginia opossum males die after their first breeding season but not because their immune system is compromised (Woods and Hellgren (2003)). Virginia opossums, like their South American cousins, have a high fecundity and sustain very high mortality after becoming independent, so that fewer than one- tenth of those born participate in breeding in their second year and very few survive thereafter.

Conservation

Prehistoric mammal in South America Many of the large mammals of South America disappeared very suddenly less than 20 000 years ago, coinciding with the first appearance of people on the southern continent (Patterson and Pascual 1972). These were both the placental immigrants from the north that arrived during the great faunal interchange two to three million years ago, as well as many survivors of the old placental fauna. The large marsupial predators had disappeared two million years before human arrival on the continent and there is no evidence that the surviving small marsupial species were adversely affected by human occupation. To the contrary, the invasion of North America by the Virginia opossum was probably facilitated by human occupation of the continent.

Recent effects of human occupation Emmons (1990) identified three ways in which people affect the survival of indigenous mammal species: by hunting for food or other products, which in South America may mean skins or Opossums of the Americas: cousins from a distant time 137

the pet trade; by destroying resources the mammals need to survive, such as habitat and food supplies; and by introducing exotic species that kill or out compete the indigenous species. The South American marsupials are too small to be worth hunting for food in a concerted manner and insufficiently attractive for the pet trade; and, except on islands, recently introduced species do not affect the survival of marsupials in South America. Hence, the only serious threat to the continuing survival of marsupial species is the alteration of habitat for human habitation and forest clearing to raise stock. As already noted, the rare species of forest marsupials do not survive in altered forest, although the commoner species thrive and actually increase in abundance, so the main threat of land clearance is to the rare, endemic species. Fonseca et al (2003) have summarised the distribution of endemic species in each of the four major habitats in the conti- nent (Table 3.5). The forest regions support more than half of all the endemic species, with smaller numbers in the grasslands, the Andean regions and very few in the arid regions. Likewise, most of the species threatened by loss of habitat are also in the forest regions. With the present widespread clearing of the Amazonian rainforest and the Atlantic forest in Brazil, the prospect is therefore very high that many of the mouse opossums will disappear. However, the widespread species, such as the common opossums, the gray four-eyed opossum and the bare-tailed woolly opossum will probably survive and flourish in the altered environments that prevail.

Table 3.5: Total number of marsupial species, and the number of these that are endemic and threatened, in each of four main habitats of Central and South America From Fonseca et al 2003, table 3.

Habitat No. of species Endemic species Threatened species Forest 56 37 14 Open grassland 21 4 2 Andean ecosystems 13 11 5 Arid regions 9 3 4

Conclusions The evolution of marsupials in South America has been extraordinarily conservative: the only adaptive variants on the generalised omnivore were the large carnivores of the Miocene and Pliocene, the jumping argyrolagids also of the Pliocene, and the water opossum since the Pliocene. Compared to the history of marsupials in Australia the adaptive diversity has been very modest: South American marsupials did not give rise to specialised honey eaters, leaf eaters, gliders, browsers or grazers; and the size range has also been modest, from about 10 g body mass to 5 kg today, and to 200 kg in the Pliocene. By contrast, in Australia the size range is from 5 g to 90 kg today and to over 2000 kg in the recent past. Nevertheless, there are some interesting examples of convergence between South American and Australasian marsupials. Among the Didelphidae the larger opossums and the gray four-eyed opossums resemble the middle-sized species of Dasyurus; the Patagonian species Lestodelphys is similar in habits and appearance to Phascogale; and it has been suggested that Monodelphis dimidiata and Marmosa incana have a life history pattern that is similar to the several species of Antechinus (see Chapter 4 for discussion). The extraordinarily fast rate of reproduction in Monodelphis domestica is similar to the reproductive strategy of bandicoots (see Chapter 5). Dromiciops has similar strategies for life in a cold climate to the Australian mountain pygmy possum, Burramys parvus, and the life history of the bare-tailed woolly opossum is similar to the sap feeding sugar gliders, Petaurus (see Chapter 6). However, there was no radiation in South 138 Life of Marsupials

America of marsupial herbivores, comparable to the wide variety of arboreal leaf eaters, brows- ers and grazers that arose in Australia in the Oligocene and expanded in the Pliocene. The usual explanation for the discrepancy is that the marsupials in South America were competing with an array of archaic placental herbivores, and also from the Oligocene with rodents and primates, and that it was only in Australia, where there were no placental competitors, that marsupials could evolve to their full potential in adaptive diversity. McNab (1986) ascribes lack of competi- tive success of marsupials in those parts of the world where placental mammals predominate to their lower metabolic rate and to their slower rate of reproduction. The discovery at Murgon, Queensland, of Tingamara, challenges this argument: if it was a placental mammal, and both kinds of mammals occupied ancient Australia, why did marsupials alone prevail? Part of the answer may lie in the adaptations of Australasian marsupials to their present environment. Chapter 4

Predatory marsupials of Australasia: bright-eyed killers of the night Stripe-faced dunnart; drawing by AG Lyne. Predatory marsupials of Australasia: bright-eyed killers of the night 141

he carnivorous marsupials of Australasia range in size from the tiny , , and planigales,Planigale, at 2–10 g (Fig.4.1a,Plate 6),to the Tasmanian devil,Sarcophilus harrisii, Tof 9 kg. Most are members of the large family Dasyuridae, with 47 species in Australia and 17 in New Guinea and surrounding islands, only two of which are common to both regions. The three other species are: the dog-sized thylacine, Thylacinus cynocephalus, now extinct; the numbat or banded anteater, Myrmecobius fasciatus (Fig. 4.1e, Plate 7); and the marsupial mole, Notoryctes typhlops. The thylacine is quite closely related to the dasyurids, but the numbat is more distantly related, and the marsupial mole an enigma. To complete the picture one should include the largest mammalian predator ever to have lived in Australia, the extinct marsupial leopard, Thylacoleo carnifex. However, it was not related to the other carnivores but to possums. Small dasyurids only eat prey that they themselves have caught and subdued. They orient the struggling insect, using their paws to tread on it and expose the soft intersegmental cuticle. Having gained a tooth-hold and inflicted a killing bite, they may then decapitate the prey, or dismember it by tearing and shearing the cuticle before pulling out the body contents (Fisher and Dickman 1993). The diet of dasyurids up to 160 g body mass is predominantly insects and other small invertebrates, with the very small species selecting the smaller prey and the larger species selecting the larger prey available. It is not that the smallest species cannot kill large prey, such as beetles, as quickly as the bigger dasyurids, but that they choose not to because of the longer time it takes them to break up and eat the prey. It is more efficient for them to catch many small species than one large one. Conversely, the larger dasyurids, 300–500 g, such as the brush-tailed phascogale, Phascogale tapoatafa, avoid very small prey unless these are abundant. However, the numbat, which is this size, specialises on small prey, particularly ants and termites. Moving up the size scale the quolls that weigh up to 1 kg also eat insects and other inver- tebrates but more of their diet is vertebrate prey, ranging from skinks and snakes to birds and mammals. The two largest species, the spotted tailed quoll, Dasyurus maculatus, and the Tasmanian devil rely largely on catching and eating small possums, Trichosurus , and pademel- ons, Thylogale. The Tasmanian devil is also a scavenger on the carcasses of larger mammals, such as wallabies and kangaroos, Macropus, and common wombats, Vombatus ursinus. The three large predators occur in Tasmania where their diets overlap. Competition between them mainly resolves to body size, with the Tasmanian devils and male spotted tailed quolls competing for prey, while the eastern quolls, Dasyurus viverrinus, avoid direct competition by their use of smaller prey species (Jones and Barmuta 1998). Until 1900 this competition for prey would have included the thylacine weighing 15–35 kg, but little is known directly about it. Paddle (2000) concluded from his comprehensive review of the 19th century records that the Australian thyla- cine, Thylacinus cynocephalus, was a generalist predator of small to medium-sized species and that family groups hunted together. Jones and Stoddart (1998) endeavoured to reconstruct its hunting behaviour from a comparison of its anatomy and dentition with that of other similar carnivores, such as hyenas and hunting dogs, and civet cats. The long snout and wide gape of the thylacine suggests that it had a weaker bite than a hyena and more resembled a fox, which is a solitary hunter of small prey. It probably could not have brought down a large animal such as a kangaroo or large wallaby, Macropus. The thylacine, like the Tasmanian devil, was a pounce–pursuit predator of fairly open habitats, which killed medium-sized prey of 1–5 kg with a crushing, penetrating bite. The first thylacine to be described by Europeans in 1805 weighed 20 kg and contained 2.3 kg of wallaby flesh in its stomach (Smith 1982). If that meal represented a 5 kg Tasmanian pademelon, Thylogale billardierii, then that is the size of prey that a European , Vulpes vulpes, brings down. At fossil sites in the commonest species found in association with thyla- cine bones were rat kangaroos, wallabies, possums and bandicoots (Case 1985), which supports the idea that it was a predator of small mammals. By contrast the bones of Australia’s largest 142 Life of Marsupials

predator, the marsupial leopard, were associated with the bones of the extinct giant kangaroos, and Macropus, and , suggesting that it was capable of killing much larger mammals. Milewska (2003) has argued that it could not have brought down a full-sized dipro- todon or sthenurus, because it lacked large canines to do so and that it only preyed on pouch young of the large marsupials, but others disagree with this interpretation.

Anatomy Dasyurid teeth closely resemble the primitive dentition of the earliest marsupials (see Fig. 1.4) and American opossums (see Chapter 3). The cheek teeth reflect the type of prey: the largest species that catch and eat mammals have teeth that can crush bone as well as cut meat, whereas the teeth of the smaller species, which eat insects, have little capacity to crush but the rows of

Figure 4.2: Three aspects of the lower and upper molars of the eastern quoll, Dasyurus viverrinus. Viewed from below, the closed jaws show how the opposing cusps provide a sharp zig-zag edge for shearing flesh. The detail of this is shown in cross section above and plan view below for the upper and lower third molars. Predatory marsupials of Australasia: bright-eyed killers of the night 143

molars act like a pair of sharp scissors, with the cusps of the upper molars shearing against the cusps of the lower (Fig. 4.2). As well as the cheek teeth, all dasyurids have prominent canines and four pairs of upper and three pairs of lower incisors in the front of the mouth (see Fig. 1.5, fat-tailed dunnart). The numerous incisors readily distinguish the smaller species from placental mice and rats, which have two pairs of upper and two pairs of lower incisors, and from small possums, which have three pairs in the upper jaw and a single pair of incisors in the lower jaw (see Fig. 1.5). The shape of the feet varies in relation to the preferred habitat. Tree-living species have broader feet than ground living species but, in all but one species, the legs are relatively short. The , Antechinomys laniger, has long hind legs and was earlier thought to hop, like hopping mice, Notomys, but high speed cinematography showed that it actually bounds like other dasyurids. Dasyurids are generalised carnivores with a uniform body plan and all species have 14 chromo- somes (see Table 1.2). Although there are some minor rearrangements, dasyurid chromosomes are very similar to the 14 chromosomes of the South American didelphids (see Fig. 1.8, Plate 1). The main variation between dasyurids is body size, which affects the type of prey eaten, their metabolism and reproductive strategy. The first classification of the family Dasyuridae was based on anatomical features of the skull, teeth and feet. On these criteria the 30 species examined by Archer (1976) were divided into three subfamilies of increasing body size. This broad division has been generally supported by later analyses, although some species have moved from one subfamily to another, based on penis morphology (Woolley 1982), serology and immunology. There is now good agreement at the species level between DNA/DNA hybridisation (Kirsch et al 1990) and DNA sequences (Krajewski et al 2000c), which provide an estimate of the time of separation of each group in the past (Fig. 4.3). Good descriptions of all known species are to be found in Mammals of New Guinea (Flannery 1995) and Mammals of Australia (Strahan 1995). The Australasian carnivorous marsupials can be considered under four groups of increasing body size.

Brief review of the living species

Sminthopsinae The subfamily includes the tiny planigales (Fig. 4.1a, Plate 6) and ningauis, weighing 2–13 g, the smallest of any marsupial, and comparable to the smallest placental, the 2 g Etruscan shrew, Suncus etruscus, which they superficially resemble. This group also includes the slightly larger dunnarts, Sminthopsis, that range in body mass from 15 to 50 g. There are five species of Planigale, 3 of Ningaui, 19 of Sminthopsis (Fig. 4.1b, Plate 6) and one species of Antechinomys. These small carnivores occur mainly in the desert and semi-desert regions of Australia. In New Guinea the group is represented by one indigenous species of Planigale and two species of Sminthopsis that also occur in Australia. All three species are restricted to the southern coastal savannas of the island, opposite the northern tip of Cape York. Their absence from the New Guinea rainforest and the highlands suggests that these three species are outliers of the Australian radiation, which became separated after the sea level rose 12 000 years ago.

Phascogalinae There are three genera in the subfamily Phascogalinae, two restricted to Australia and one restricted to New Guinea, and there are no shared species. All have body mass of 20–200 g and are mainly insectivorous. In Australia there are two species of Phascogale and 10 species of Antechinus. The largest species is the brush-tailed phascogale, which weighs 145–200 g and is widespread in 144 Life of Marsupials

Dasyurus

Sarcophilus Phascolosorex* Neophascogale* Parantechinus Dasykaluta * Dasyuroides Pseudantechinus Dasycercus

Antechinus

Phascogalinae

Murexia*

Phascogale

Ningaui

Sminthopsis

Sminthopsinae

Antechinomys

Planigale

Thylacinus Myrmecobius Notoryctes 35 30 2025 15 510 0 Oligocene Miocene y million years ago Pliocene Quaternar Figure 4.3: The relationships of all Australasian carnivorous marsupials, based on mitochondrial and nuclear DNA sequences, with estimated times of divergence of each group in million years before present. The three sub-families of the Dasyuridae are clearly separated from each other and from the more distantly related thylacine, numbat and mole. Note that on these criteria the four indigenous genera of New Guinea (*) diverged between 8 and 12 million years ago. After Krajewski et al (2000c). Predatory marsupials of Australasia: bright-eyed killers of the night 145

coastal woodlands around Australia. The smaller species, the red-tailed phascogale, Phascogale calura, is now restricted to a small region of western Australia. It is thought that its reduced range is due to clearing of its preferred habitat for agriculture (see Conservation). The more numerous and widespread genus is Antechinus, species of which occur in tall forests all around Australia. None is larger than 100 g. They are predominantly ground-living animals but some species also climb readily and forage in the canopy for insects. The best-known species is the brown antechinus, Antechinus stuartii and its close relative, the yellow-footed antechi- nus, Antechinus flavipes (Fig. 4.1c, Plate 7), of southern and eastern Australia. In the last decade what was formerly called the brown antechinus has been divided into four species in a south to north species complex. The species in Victoria and southern is now called the , Antechinus agilis, the brown antechinus extends from there to northern New South Wales, where it is replaced by the , Antechinus subtropicus. In far north Queensland this species is replaced by the rusty antechinus, Antechinus adustus. Because all the earlier work was referred to the brown antechinus this name will be used hereafter. Two larger species, the , Antechinus minimus and the dusky antechinus, Antechinus swainsonii, occur in Tasmania and southeastern Australia, and another three species in northern Australia. All of these species, as well as the , have a highly unusual pattern of repro- duction (see The big bang breeders). The third genus is (including four species formerly called Antechinus), which only occur in New Guinea. They range in size from Murexia habbema at 20–45 g, through three species of about 50 g and two that are larger. All occupy montane forests of New Guinea. No species of Murexia has the unusual reproductive pattern of Antechinus and Phascogale. There are no smaller dasyurids in New Guinea forests, as there are in Australian forests, possibly because a range of rodents occupies the niches of very small insectivore (Flannery 1995).

Dasyurinae The subfamily Dasyurinae contains the largest members of the Dasyuridae. Of the six species of the genus Dasyurus, two occur in New Guinea and four in Australia and Tasmania. The two species in New Guinea range up to 800 g in body mass and are the largest indigenous predators on the island (Fig. 4.1d, Plate 7). In Australia there are three species that are up to 1 kg body mass, the eastern, western and northern quolls, which were formerly widespread but now, through contractions of their ranges, do not overlap. The much larger spotted tailed quoll, occurs around the eastern seaboard of Australia and in Tasmania. It weighs up to 7 kg and is thus almost as large as the Tasmanian devil. There are eight smaller members of the Dasyurinae, four species in New Guinea in the genera Phascolosorex, Neophascogale and Myoictis, and four other species in Australia and, again, there is no overlap of species between the two countries. It is evident from this rapid survey that the Dasyuridae underwent two separate radiations, one in New Guinea and one in Australia, with only very slight interchange via Cape York during the land connection 12 000 years ago. This pattern is repeated for the bandicoots (see Chapter 5), possums (see Chapter 7) and the macropods (see Chapter 9).

Thylacinidae The thylacine is the only recent member of the family Thylacinidae, although fossil species are represented in deposits of the late Oligocene to late Miocene epochs (26–5 million years ago) (Fig. 4.4). Like the Tasmanian devil, it was formerly distributed across Tasmania, the Australian mainland and New Guinea, and is depicted in cave paintings in northern Australia. After the arrival of the dingo, Canis familiaris, about 4000 years ago, both species became extinct on the mainland of Australia and New Guinea. The most recent remains of a thylacine from the 146 Life of Marsupials

mainland of Australia is a desiccated corpse found in a cave on the Nullarbor, which is estimated by 14C dating to be about 2600 years old. The Tasmanian devil survived longer on the mainland, the most recent remains being about 300 years old. The thylacine survived in Tasmania until the 1920s, the last known animal died in Hobart Zoo in 1936. Regrettably, very little about this remarkable animal, apart from its anatomy, was recorded by biologists when it was common but a surprising amount has been gleaned by Paddle (2000) from contemporary records of farmers, trappers and those who kept animals in captivity. In museums around the world there are a few preserved specimens, which with modern techniques can reveal a bit more about its relationship to other marsupials. It almost certainly had 14 chromosomes like all the Dasyuridae. Protein and DNA sequences, extracted from museum skins and teeth, also show that its closest relatives are dasyurids, from which it separated about 25 million years ago (Fig. 4.3) (Krajewski et al 1997). If these estimates are correct, it means that thylacinids were not closely related to the borhyaenid marsupials of South America because the continents had been separated for much longer, and the apparent resemblances result from convergent adaptations for life as a top predator.

Murgon Genus 34 incertae sedis Thylacinidae 4444

Barinyainae Dasyuromorphia

e Dasyuridae incertae sedis Sminthopsinae 2528 Dasyurida Dasyurinae 4 10 20

Phascogalinae 716

Myrmecobiidae 2 55 26 23 5.2 0.01 16.3 10.4 million years ago Figure 4.4: The fossil record of Australian carnivorous marsupials from a possible ancestor in the Eocene (55 million years ago), through the later flowering of the thylacinids and the appearance in the Pliocene (5 million years ago) of the three dasyurid groups. Number of genera, if more than one, shown in the horizontal bars. After Wroe and Muirhead (1999).

Myrmecobiidae The numbat, Mymecobius fasciatus, (Fig. 4.1e, Plate 7) is only distantly related to the other families on biochemical criteria and there is no fossil record to help resolve its relationships. It is highly specialised for feeding on ants and termites and is one of the few marsupials that is active during the day. It has 14 chromosomes and, like the smaller dasyurids, females lack a pouch.

Origins While the dasyurids have anatomical similarities with the ancestral marsupials of South America, and they all have the original chromosome number, the fossil evidence suggests that they are a Predatory marsupials of Australasia: bright-eyed killers of the night 147

relatively recent radiation. From the late Oligocene and throughout the Miocene (26–5 million years ago) a wide variety of thylacinid fossils have been found but no dasyurid fossils (Fig. 4.4). Then in the Pliocene epoch, about 5 million years ago, the thylacinids declined to a single species and the three subfamilies of dasyurids appeared and blossomed into a great variety of species (Wroe and Muirhead 1999). It is not clear whether the great variety of dasyurids arose from thylacine ancestors then, or from other ancestors not yet discovered.

Metabolism of dasyurids Unlike the diet of herbivores, which is very varied in content, the diet of carnivores is similar in respect of water content, nitrogen and energy across a range of prey. Apart from insect cuticle and vertebrate bones and feathers, which must be discarded, the diet is easily digested. The rate of passage through the simple gut is extraordinarily fast – from an hour or two in the Julia Creek dunnart, Sminthopsis douglasi (Hume et al 2000) to several hours in the eastern quoll, and the Tasmanian devil. Compare these times to the 100 hours for food to pass through the gut of a koala, Phascolarctos cinereus (see Chapter 7). Thus, for carnivores the main physiological differ- ences between species relate to body size, which determines what kind of prey can be caught and

41

Antechinus

Ambient temperature Sminthopsis

39

Dasycercus

Body temperature (°C)

Sarcophilus

37 024 6 Exposure time (hours)

Figure 4.5: Responses of four dasyurids of increasing body mass when exposed to an air temperature of 40°C and high humidity: 17 g fat-tailed dunnart, Sminthopsis crassicaudata; 60 g yellow-footed antechinus, Antechinus flavipes; 60 g , Dasycercus cristicauda; 6.7 kg Tasmanian devil, Sarcophilus harrisii. After Robinson and Morrison (1957). 148 Life of Marsupials

eaten. The dasyurids, with a uniform anatomy and diet but a range of body sizes provide a nice model for examining the importance of body size in the economies of mammals. As mentioned in Chapter 1, the size of the body has a profound effect on the cost of living, from keeping cool in hot weather and warm in cold weather, to requirements for water, for food and for the energy to drive the whole system, and to reproduce.

Thermoregulation Body mass affects thermoregulation, as noted by Robinson and Morrison (1957), who tested the responses of four dasyurids to the high ambient temperatures that desert-living species might

38 Sminthopsis crassicaudata (0.017 kg)

36 Sminthopsis macroura (0.022 kg)

34

32

Antechinus flavipes (0.06 kg)

38 Dasycercus cristicauda (0.06 kg)

36

Body temperature (°C) 34

38 Dasyurus geoffroii (1.3 kg)

36 Sarcophilus harrisii (6.7 kg)

34

0 0600 1200 1800 2400 Time in hours

Figure 4.6: Diurnal fluctuations in mean body temperature of the same species of dasyurid marsupial as in Figure 4.5 and the , Dasyurus geoffroii. Only the Tasmanian devil, Sarcophilus harrisii, is active by day and only the two species with the lowest midday body temperature are able to become torpid. After Godfrey (1968), Morrison (1965) and Arnold and Shield (1970). Predatory marsupials of Australasia: bright-eyed killers of the night 149

encounter. When exposed to ambient temperatures from 5°C to 30°C all maintained a steady body temperature of 37.4–38°C but, when they were exposed to 40°C and 50% humidity (Fig. 4.5), all showed a rapid rise of about 2°C. The largest species, the Tasmanian devil (6700 g) soon resumed its normal body temperature of 37.5°C and maintained it for 6 h without appar- ent discomfort. It did drink water, indicating that evaporative cooling by sweating is probably the main route for heat dissipation. The mulgara, Dasycercus cristicauda (72 g) stabilised its body temperature at 38.6°C and also drank water. However, the smallest species studied, the yellow-footed antechinus, (48 g) and the stripe-faced dunnart, Sminthopsis macroura (11 g), were unable to cope and had to be withdrawn from the experiment after about 3 h, when their temperatures went above 40°C. The importance of body mass was demonstrated by these observations, since the Tasmanian devil is an inhabitant of a mild climate and would never experience such high temperatures, whereas the very small stripe-faced dunnart lives in an environment where such temperatures occur frequently. How then does it cope in normal circumstances? The adaptations for survival of the smaller species include avoiding direct exposure to high temperatures, a marked diurnal cycle of body temperature (Fig. 4.6), and torpor. Although the Tasmanian devil may be active during the day, all the other species are strictly nocturnal. Several live in burrows or other shelters during the day, where the microclimate is less extreme. In addition, in the three species that live in hot dry environments, the western quoll, Dasyurus geoffroii , the mulgara and the stripe-faced dunnart, the lowest point in the diurnal cycle occurs at midday, when the body temperature may be 3–4°C lower than during the active period at night.

Standard metabolic rate – the cost of being small The standard metabolic rates of a range of dasyurids of increasing size have been measured (see Hume 1999, Table 1.1), which illustrate the importance of body mass in the economies of these marsupials. Hourly oxygen (O2) consumption ranged from 2.13 mL O2/g body mass for the smallest species (Planigale ingrami, 7 g) to 0.28 mL O2/g body mass for a Tasmanian devil at 5050 g. Converting this to energy expended per day (1mL O2 = 21 Joules) the range is from 635 kJ/kg for the planigale to 141 kJ/kg for the Tasmanian devil. Thus, on a per kilogram basis, the cost of living for the planigale is 4.5 times greater than for a Tasmanian devil, which weighed 1000 times more. However, the daily gross energy expenditure of the Tasmanian devil (712 kJ) is very much greater than the 3.8 kJ of the tiny planigale. These differences become even more marked when field metabolic rate is compared.

Field metabolic rate in dasyurids The field metabolic rate (FMR) is an estimate of the total energy required over a given period for a free living animal to carry out all its normal activities of hunting, feeding, regulating its body temperature and reproducing (see Chapter 1). Water turnover, respiration rate and food consumption have been measured in a range of dasyurids, using the techniques of 3 18 22 isotope dilution, such as doubly labelled water ( H2O, H2 O) and an isotope of sodium ( Na) (see Box 1.1). As with the standard metabolic rate (SMR) the smallest dasyurids have the highest FMR per unit weight, there being an almost tenfold daily difference between the brown antechinus (4218 kJ/kg) and the Tasmanian devil (407 kJ/kg) (Table 4.1). Also, for all species the FMR is 3–6.6 times higher than the SMR, indicating the overall cost of living. The FMR also varies markedly between high winter and low summer values in all species.The FMR of three species, represent- ing the range of body sizes, have been studied in greater detail, so that the costs of living for them can be related to their behaviour and ecology (Table 4.1). 150 Life of Marsupials

Table 4.1: Field metabolic parameters of three free-living dasyurids of increasing mass Data from Green 1997, tables 9.2 and 9.3.

Brown antechinus Eastern quoll Tasmanian devil Winter Summer Winter Summer Winter Summer Mass (g) 26 16 1102 1029 7100 7900 Field metabolic rate, 4218 2526 1055 733 407 328 kJ/kg per day Daily energy needs, 72 45 1169 793 2890 2591 kJ/day Food equivalent (g): at 3.5 kJ/g insect 21 13 334 227 at 5 kJ/g mammal 234 159 578 518 Water flux (mL/day) 16 9 262 148 743 724

Field energetics of eastern quolls and Tasmanian devils The eastern quoll and Tasmanian devil are still abundant in Tasmania, where both have been well studied in the wild. Both species live in a moderate climate with well-marked seasons and a flush of insects and young mammals in spring, which is the time when their young vacate the pouch and are left in a nest. Their protein diet has a high water content (70%) and metabolisable energy content between 3.5 kJ/g for insects to about 5 kJ/g for wallaby flesh. In summer the eastern quolls weighed about 1 kg and their daily energy needs was 793 kJ/day and their daily water turnover was 148 mL. This is equivalent to 159 g of mammal flesh or 227 g of insects: two rat-sized mammals or 340 insects would provide all the energy and about 145 mL of the daily water requirements. The value for water turnover includes the water required for cooling in summer and the balance would have been made up by drinking water. In winter the energy turnover was 1169 kJ/day or 334 g of insects, significantly higher than in summer, with correspondingly higher water turnover as well. The higher turnover in winter reflects the greater demand for energy for the maintenance of a stable body temperature in the cool weather. By contrast the Tasmanian devil, which weighed 7.9 kg in summer, consumed 2590 kJ/day or the equivalent of 518 g of prey. In winter it weighed less but consumed more food (578 g of prey). Thus, the Tasmanian devil ate twice as much food as the eastern quoll, although it was eight times as large. This again illustrates the advantage of body mass for the conservation of body heat and hence energy requirement in the form of food.

Brown antechinus Going down in size in the same type of habitat are the four antechinus species, which live in temperate forests in Victoria, New South Wales and Queensland. These species are considered together because, until recently they were thought to be one species and have been studied across their range under this name. The brown antechinus in the summer weighed 16 g and had an energy turnover of 45 kJ/day (Table 4.1), which on its insect diet of 3.5 kJ/g, was a daily intake of 13 g of prey. This provided 7 mL of its daily water requirements of 9 mL, so it required very little drinking water. Again, during the winter, when its weight was 26 g the energy requirement was 72 kJ/day, so it needed to consume 21 g of insects. While these figures are much lower than those for the larger species, just considered, as a proportion of the animal’s body mass they are much greater. Thus, to meet them the small antechinus must consume in prey almost its own Predatory marsupials of Australasia: bright-eyed killers of the night 151

body weight each night, whereas the eastern quoll needs to consume only one-quarter and the Tasmanian devil one-fifteenth of its own body weight. These figures were obtained from adult males and females that were either not breeding or, if female, in the early stage of lactation. During late lactation the cost of living for the female is much higher in all three species, because of the increased volume and high fat content of milk secreted (see Chapter 2).

Cost of reproduction The cost of reproduction in free-ranging eastern quolls in late lactation was measured in the same way by isotope dilution (Green et al 1997) and, with up to six young, the daily energy requirement in late lactation increased by 30%. When the energy exported in the milk itself was included the energy requirements in late lactation were double those of early lactation or non- breeding females. Similar ratios have been measured in lactating Tasmanian devils and brown antechinus. While a doubling of food intake for the Tasmanian devil is still a small proportion of its body weight, for the eastern quoll it represents about half its own body weight, and for the antechi- nus female it represents more than its own body weight each day. It is not surprising, therefore,

1000

Lactating 800

600 Non lactating 400 Males

200 Water influx (mL/kg/day) Males die 0

20

15 Non lactating Males 10 Lactating

Percentage body fat 5

Males die 0 JFMAMJJASOND Month Figure 4.7: Changes in fat metabolism in free-living male ( ) and female (O) brown antechinus, Antechinus stuartii, through the year. Note (a) the increase in water turnover and (b) reduced body fat of lactating females. After Green et al (1991). 152 Life of Marsupials

that in each of these species the young are left in a nest during late lactation, while the mother goes hunting for prey. Nevertheless, female eastern quolls with litters of three or five young lost weight by the end of lactation, indicating that they were near the limit of available resources. For brown antechinus females supporting a litter of up to eight young to independence is an even bigger task. While the early stage of lactation, when the young are very small, is not a severe drain on the water and energy needs of the female, by the time the young are left in the nest, the total litter weigh three to four times the mother’s own weight. At this time water and energy consumption were significantly higher than earlier because of the water and energy diverted into milk for the young; the fat reserves of the females also decline during this time (Fig. 4.7). Food consumption of brown antechinus males, as measured by water influx, remained steady through the first half of the year and then declined as they lived on their fat reserves until they died at the end of August (Fig. 4.7). While the total lifetime energy intake for males and females was about the same (40 megaJoules, MJ), the larger males consumed more than females on a daily basis, but they died by September. If they did not die then, they would consume a further 3.5 kg of prey. This is more than the 2.5 kg of prey required by the female to feed her litter to independence. If she is unable to catch enough food, she must either draw on her reserves and lose weight, as shown by the drop in body fat of lactating females (Fig. 4.7b), or successively reduce the litter size. The second option is sometimes adopted and the female selects the sex of the young to be discarded (Cockburn 1994). Conversely, a female may prolong lactation, which enables the young to be weaned later if resources are scarce, or at a larger size if resources are abundant (Cockburn 1992). Either way, the presence of males after September, competing for the available prey, would further jeopardise the females’ chances of rearing their litters to independence. Green (1997) suggested that this may have been a key factor in the evolution of male die off at the end of the mating period (see The big bang breeders).

Control of the breeding season The only period of the year when the abundance of prey is sufficient for the females of all three dasyurids is in spring and early summer, from August to December: this is a regular and predict- able event and the reproductive cycles of the predators closely follow it. Since the length of lactation and the growth of the young to weaning are also related to body size, the mean time to weaning in the Tasmanian devil is 150–240 days, for the eastern quoll 135–140 days and for the brown antechinus, 90–110 days. In consequence the mating season and timing of births occurs earliest in the Tasmanian devil, in March and April, later in the eastern quoll, in May and June, and latest of all in the brown antechinus in July and August. Thus, each species comes into breeding condition in response, presumably, to a different component of the changing day length through the year, but this has only been tested in antechinus and will be discussed later (see Photoperiod control of breeding). If a female Tasmanian devil or eastern quoll becomes pregnant at the start of the breeding season, she will carry through to the spring and not return to oestrus again in that year. However, if she fails to become pregnant at the start of the season she can return to oestrus about one month later and will then carry her young later in the year, when food may be harder to get. Both of these species live for several years so that a female in her lifetime may have several litters. In other words, she does not invest all her reproductive effort in the first litter she produces. Likewise, if a male is unsuccessful in fertilising a female in his first breeding season, there will be more opportunities in future years. By contrast with the larger species, the female antechinus does not have a second chance if she fails to become pregnant, because there is only one brief period in the year when she will mate. All species of antechinus have a very unusual reproductive strategy in which the males invest everything in their first and only breeding season and the females make nearly all their Predatory marsupials of Australasia: bright-eyed killers of the night 153

investment in the first litter they produce. This is so unusual that it has attracted a lot of atten- tion over the past 40 years.

The big bang breeders Woolley (1966) discovered that all the males in populations of yellow-footed and brown antechi- nus near Canberra died suddenly in August before the females, with whom they had mated, gave birth. Wood (1970) independently discovered the same phenomenon in another population of brown antechinus (now called Antechinus subtropicus) near Brisbane. Wood followed the entire population on a 0.8 ha plot of rainforest for three years by repeated recapture within the plot and on surrounding trap lines. There was an annual fluctuation in numbers of animals on the plot (Fig. 4.8) with the new cohort of young males and females appearing in the traps in January, soon after they had been weaned and were moving away from the parental nest. By this time the only adults in the population were females that had weaned their litter, and they would make up about one-fifth of the female population in the next breeding season in June. At this time the number of males caught on the site increased greatly and males previously caught on the site were caught away from it, indicating an increased activity and enlarged range of the males. Then suddenly all the males disappeared from the study site in late September and the surrounding trap lines and Wood concluded that they had all died.

Antechinus stuartii

Figure 4.8: Total number of brown antechinus, Antechinus stuartii, on 0.8 ha of rainforest during three years, subdivided to show the contribution of each annual cohort of males and females. Note the disappearance of the male cohort in October each year. After Wood (1970).

One month later, after all the males had disappeared, each female had a litter of about eight newborn young. After being suckled for three months these young began to disperse, and the annual cycle was repeated. Woolley (1966) had already observed in captive antechinus that the males passed into a general decline at the end of the mating period and most then died. If she nursed a male through this crisis, it would recover but its testes were permanently regressed, so that it could never again contribute to reproduction. The phenomenon of a single cycle of spermatogenesis, a highly synchronised and brief mating period followed by die off of the entire male population has subsequently been found to occur in all species of antechinus from north- ern Australia to Tasmania and in both species of Phascogale: it also occurs in the , Dasykaluta rosamondae, (Woolley 1991) and in the , Parantechinus apicalis, and in the northern quoll, Dasyurus hallucatus (Dickman and Braithwaite 1992), which are all members of the subfamily Dasyurinae. 154 Life of Marsupials

The two main questions raised by this discovery were the physiological mechanisms that cause synchronous death in all males at 11 months of age, and the significance of this strategy in evolutionary and ecological terms.

The physiology of male die off The mortality of the males results from the combined endocrine, behavioural and immunologi- cal factors that severally follow the frenetic period of fighting and sexual encounters during the very brief mating period. In the months leading up to the mating period the males grow faster than the females and reach almost twice their weight. The levels of testosterone reach a peak just prior to mating, when the accessory reproductive organs reach their maximum size and the epididymides are packed with mature sperm. Unlike in other mammals, there is only one spermatogenic wave (see Chapter 2) so that there are no immature sperm in the testis at the start of the mating period, only mature sperm, which are progressively used up (Woolley 1975, Kerr and Hedger 1983). This is a serious commitment to a single throw at reproduction. As well as the elevated testosterone in circulation, cortisol secreted by the adrenal glands is also elevated. Normally, most cortisol is bound to a specific protein in the blood, which inacti- vates it. However, testosterone inhibits this blood protein so, in the male antechinus at the start of the mating period, all his cortisol is in the unbound active form. Cortisol has two functions: it mobilises free glucose by initiating the breakdown of protein and it also depresses the immune system. The first fuels the male’s activity, the second leads to his early death. As the frenetic mating period begins the males cease to hunt or feed and instead devote all their efforts to fighting other males and finding females with which to mate. Copulation in antechinus is a prolonged affair, the couple remaining together for up to 12 h. Since the males are not feeding, the energy to drive all this activity is mobilised from their own proteins, helped by the high levels of free cortisol. As a result the males go into a negative nitrogen balance and lose weight. This is reflected in their FMR, which is actually lower than that of the females at this time, because the males are not taking in food but using up their own protein and fat (Fig. 4.7). By the end of the mating period, when all the females are pregnant, the males have lost up to half their weight and their hair is falling out from the fighting they have engaged in. Because of the high levels of free cortisol, they also have a lowered immunity to foreign pathogens, so that their burden of various blood and intestinal parasites increases, and they rapidly succumb to the cumulative effects, and die. When males were castrated prior to the mating period they held their body weight and survived after all the intact males had died, which showed that the initial driver of this entire process is testosterone.

Ecological significance of male die off What is the selective advantage of this extreme reproductive strategy? As we have seen, the cost of producing about eight young is very high for a small marsupial and can only be achieved successfully when there is an abundance of nutritious food to support it. Another point is that the marsupial mode of reproduction, with a long period of lactation, means that it is generally not possible for a female to complete two cycles of reproduction within the favourable time of the year. In the temperate regions of Australia, where nutritious food sufficient for a small marsupial occurs in an annual pulse of a few months, selection will favour the female that makes one highly synchronised reproductive effort that taps into this resource, rather than the female that breeds over a longer period. While the selective force for highly synchronised breeding applies primarily to females, selection will also favour males that are prepared for that critical time and can maximise their chances of fertilising females. In Wood’s (1970) population in Queensland it appeared that the Predatory marsupials of Australasia: bright-eyed killers of the night 155

males achieved this by increasing their home range and chance of meeting a female. However, in another population of the brown antechinus near Canberra, in which individual males and females were fitted with tiny radio transmitters, the strategy was found to be more complex. Prior to the breeding season both sexes foraged in overlapping home ranges and shared nest sites in old trees. As the breeding season began the males formed large congregations of 15–40 in a few tree hollows and ceased to forage. The only movements they made were rapid excursions from one aggregation site to another one (Cockburn and Lazenby-Cohen 1992). The females continued to forage in their former home range and made occasional visits to the male congre- gations, where they could, presumably, select their mates. This pattern of mate selection is called lekking (Lazenby-Cohen and Cockburn 1988), but it is not clear how the female selects her mates or how many during her visits. From other studies on captive antechinus it is known that females can mate with several males over the 10-day period of oestrus and the sperm are stored in the oviducts until ovulation occurs later. This means that the important competition between males takes place there, and several studies have been done to establish which male fathers the litter. It turns out that the last male to mate with the female fathers up to 70% of the litter but most litters have mixed paternity (Shimmin et al 2000, Kraaijveld-Smit et al 2002). It is therefore imperative that the males be fully prepared for extended and repeated copulation during a very brief time. Because the chance of a small animal surviving to a second breeding opportunity a year hence is very small, selection will favour males that go all out in their first year at the cost of their own survival. And this they do. Braithwaite (1979) has argued that this ‘big bang’ strategy, once adopted by a species, is impossible to reverse. Any male antechinus that reverted to the normal male spermatogenic cycle and continued to produce sperm beyond the brief time when females are available to be fertilised would be at a disadvantage in that season, compared to the males that put all their resources into the first season. And, if they survived to the second year, these males would again be at a disadvantage from the new vigorous young males, sired by their competitors the year before. So, on both counts such males would leave fewer genes than the big bang males, and would not persist. Another thing that may favour the big bang males is that by dying immediately after the mating period they are not competing for scarce food resources with their progeny, who will need the food to reach sexual maturity 11 months later (Green 1997). The most recent studies on the relationships of the dasyurids separates the phascogales from the antechinus species, which has led to the idea that the big bang breeding strategy arose independently in the two groups (Krajewski et al 2000b). Furthermore, the occurrence of sudden male die off in three species of the dasyurine subfamily in northern Australia adds further strength to the idea that the strategy has arisen several times from the common strategy in which males survive to a second season and females have more than one oestrus in a year. The annual die off of some didelphids (see Chapter 3), indicates that the same strategy may have arisen independently in American marsupials (Pine et al 1985, Lorini et al 1994). What has not been determined for any of these species is whether the males have a single spermatogenic cycle, as do Antechinus and Phascogale.

Ecological consequences of the big bang strategy While the big bang strategy is evidently successful in those habitats where the seasonal flush of insect food is predictable from year to year, it is a serious disadvantage when disaster befalls the population, as happens after a severe forest fire: this happened in southern New South Wales when a hot fire destroyed a forest in December 1972 (Newsome et al 1975). Prior to the fire two species of antechinus and two species of rat were common in the forest but all disappeared after 156 Life of Marsupials

the fire. Within a year house mice, which had never been trapped before, were abundant. In the next four years the rats recovered to their former abundance and the mice declined. However, the two species of antechinus recovered more slowly: the agile antechinus began to reappear three years after the fire but the dusky antechius did not reach its former abundance for eight years (Catling and Newsome 1981). The antechinus breeding strategy is also an effective barrier to interbreeding between related species in the same habitat. The brown antechinus and the yellow-footed antechinus occur in the same woodland habitat and differ in their ecology, and only slightly in appearance, the latter being slightly the larger. However, the yellow-footed antechinus breeds one month earlier than the brown near Canberra, so that the males have died before the brown antechinus females come into oestrus. Seasonal breeding in the brown antechinus and the larger dusky antechinus, is more complex: both species breed after the winter solstice but this is modified by altitude, latitude and by whether there is competition from the presence of the other species. Where the two species occur together, the dusky antechinus breeds five weeks earlier than when it occurs alone, and the brown antechinus breeds two weeks later than when it is alone (Dickman 1982). As with the yellow-footed antechinus, it is the larger species that breeds earlier when the two species overlap. These larger species are terrestrial, whereas the brown antechinus can forage in trees as well as on the ground. When there is competition the smaller species relies on flying insects, which appear later than the terrestrial larval stages.

Photoperiod control of breeding Whatever the ultimate cause of the timing of the breeding seasons of the several species, the strict timing of each species strongly implies that each uses some aspect of the change in day length as the trigger for the start of breeding. McAllan and Dickman (1986) analysed 162 populations of brown antechinus from latitude 37°S to 27°S and found no difference in relation to day length per se (Fig. 4.9a). However, when they reanalysed the data according to the daily rate of change of day length, the northern and southern populations separated clearly (Fig. 4.9b). The north- ern populations mated when the rate of change was 90–100 seconds per day while the southern populations (subsequently recognised as agile antechinus) mated when the rate of change was 120–130 seconds per day. Since the rate of change varies from zero at the winter solstice to a maximum at the equinox it accounts for the different timing and duration of the breeding season in the brown antechinus at different latitudes (Fig. 4.9c). Rate changes are faster at higher latitudes, where the breeding season is shorter, than at lower (northern) latitudes. Three related species in the north Queensland rainforests (19°S–16°S) do indeed have a longer breeding season of six weeks compared to two weeks in the south (Watt 1997). While this evidence for a photoperiod signal that controls breeding is persuasive, it has not been tested experimentally. It does not take into account that for successful mating to take place both males and females must have undergone prior development of their gonads for two months (see Chapter 2). Hence, the actual photoperiod signal that the animals are responding to may be the rate of change before the winter solstice, when the days are getting shorter, rather than after it. This gains support from one experiment conducted on the agile antechinus by Scott (1986) in Victoria (38°30'S), where the females ovulate and the males reach peak levels of testosterone in early August. From 1 May three groups of animals were subjected to either naturally changing day length, constant long days (14 h light:10 h dark) or constant short days (10 h light:14 h dark) until mid August. Long day inhibited ovulation in females and the testosterone rise in males, whereas females on the short days ovulated three weeks earlier than the controls, and the males had elevated levels of testosterone at the normal time. Predatory marsupials of Australasia: bright-eyed killers of the night 157

Figure 4.9: Timing the breeding season in the northern brown antechinus, Antechinus stuartii (dark shading) and the southern agile antechinus, Antechinis agilis (grey shading) in relation to: (a) length of day; (b) daily rate of change of day length; (c) the rate of change of day length as it varies with latitude (from zero at the solstices to maximum at the equinoxes). Thus, for the brown antechinus, which responds to 100 s/day, the mating period is earlier in the south than the north, but the duration is longer in the north. After McAllan and Dickman (1986).

The small dasyurids in New Guinea, four species of which were previously thought to be in the genus Antechinus, show no pattern of seasonal breeding and in at least seven species young appear to be born in all months of the year (Woolley 2003). New Guinea is so near the Equator that length of day would not be a reliable indicator of the time of year and prey species are presumably available throughout the year, so that there is no selective advantage for seasonal breeding. 158 Life of Marsupials

Dasyurids of the semi-arid rangelands

The fat-tailed dunnart The fat-tailed dunnart, Sminthopsis crassicaudata, occurs over a vast area of central and southern Australia, where its preferred habitat is open woodland, stony plains, saltbush steppe and grass- lands. It has been intensively studied in several parts of this range (Morton 1978a,b,c), as well as being bred for many years in captivity (Smith et al 1978, Woolley and Watson 1984). It serves, therefore, as the exemplar of the small to minute dasyurids that live in the semi-arid habitats of Australia, where dense cover is absent and where the climate can be extreme. As we noted earlier, the related stripe-faced dunnart from the same habitat cannot tolerate high ambient tempera- tures for more than two hours (Fig. 4.5), so the first question is how do these small species thrive in climates where such temperatures are a daily occurrence? Furthermore, we also noted that the smaller dasyurids, such as the brown antechinus, expend proportionately more energy during winter than larger species, so the second question is how do the smallest dasyurids cope with extreme temperatures in their preferred habitats and find sufficient food to support life and to reproduce? In his study site near Melbourne Morton (1978a) found that fat-tailed dunnarts were active throughout the night and sheltered under rocks during the day, thus avoiding the highest temper- atures. In summer the temperature beneath the rock was 15–40°C, whereas during winter the temperature under the rocks was 4–15°C. In the coldest months, from April to August, several animals would huddle together under rocks and so conserve body heat. More remarkably, he found mixed groups of fat-tailed dunnarts and house mice huddling together under the same rock shelters. In other circumstances fat-tailed dunnarts will kill and eat mice but in the field it seems that the energy savings through nest sharing are of more benefit to the fat-tailed dunnarts than the immediate value of eating the mice! Likewise for the mice the risk of being eaten was outweighed by the benefits of energy conservation by huddling with potential predators. After the winter solstice, while the ambient temperature still remained low, nest sharing declined rapidly, due to increasing intolerance among the fat-tailed dunnarts, as breeding began. The food of the fat-tailed dunnart consists almost entirely of small arthropods, with a prefer- ence for spiders, termites, ants, cockroaches and weevils (Morton et al 1983). When measured by water and energy turnover (Nagy et al 1988) a fat-tailed dunnart consumes its own weight (16 g) of insect food each night. While it has a varied diet to choose from, the prey species are scattered and hunting for enough of them means that the fat-tailed dunnart is very active for most of the night, covering considerable distances in search of food. A consequence of this great activity is that fat-tailed dunnarts do not settle in one place but have drifting home ranges. In capture–recapture studies this leads to overestimates of the standing population (Morton 1978b, Read 1984), because many animals are only caught once as they pass through the study site. It also accounts for the extraordinarily low rate of trapping. For instance, in 5100 trap nights and nearly 500 km of spotlighting, Morton et al (1983) caught a total of 25 fat-tailed dunnarts, 37 stripe faced dunnarts, 7 narrow-nosed planigales, Planigale tenuirostris, and one Giles’ planigale, Planigale gilesi. Two factors help fat-tailed dunnarts to maintain this high activity. They cut up the outer chitin of the prey into very fine pieces with their scissor-like molars, which increases the acces- sibility of the contained tissue by 70% (Moore and Sanson 1995), compared to the same prey ground up by a sugar glider, Petaurus breviceps. They also have a very rapid rate of digestion. More importantly, their FMR is 6.6 times their SMR (Nagy et al 1988). Why is this important? It means that the animal while at rest can conserve energy because of its low SMR, but when required it can call on large reserves. In other words it has a large metabolic scope, the highest of the four dasyurid species measured and higher than all other marsupials except Leadbeater’s possum, Predatory marsupials of Australasia: bright-eyed killers of the night 159

Gymnobelideus leadbeateri (see Chapter 6). In good times the fat-tailed dunnart develops a fat tail (hence the specific name) and the energy stored in this is sufficient to meet its requirements for a few days of food shortage. However, its large metabolic scope enables it to conserve energy by reducing activity, lowering its body temperature, basking in the sun, huddling in nests and briefly entering torpor. Of these several responses, torpor is the most significant.

Torpor Torpor occurs in a wide variety of small mammals, including all small to medium-sized dasyu- rids (Fig. 4.10) (Frey 1991, Geiser 1994) and is a physiological response to either extreme ambient temperature, which requires the expenditure of energy for cooling or warming the body, and/or a shortage of food. Torpor involves a considerable reduction in activity and body functions but is distinguished from the cold torpor of lower vertebrates by the mammal’s ability to arouse from torpor to full thermoregulation, even while the ambient temperature remains low. In its least developed form the species shows a daily fluctuation in body temperature and metabolism and, at the low point of the daily cycle, it need only stop shivering in order to cool to a body temperature characteristic of torpor: the stripe-faced dunnart and the mulgara do this (Fig. 4.6). Desert-adapted species avoid high daytime temperatures and the consequent water loss by entering a burrow or crevice. This reduces evaporative water loss immediately and by dropping the body temperature to that of the burrow a further saving is made as a result of the reduced respiratory rate. Together these can reduce water loss to less than 5% of that at normal body temperature outside. Not only does torpor conserve water during high summer temperatures, it

Figure 4.10: Oxygen consumption over 24 h in the fat-tailed dunnart, Sminthopsis crassicaudata, undergoing bouts of activity, torpor and arousal, to show the energy savings of torpor. After Holloway and Geiser (1996). 160 Life of Marsupials

can conserve energy during low winter temperatures when insect food is scarce. In the fat-tailed dunnart oxygen consumption fell during daylight hours, when the animals never forage, to one- eighth of the consumption during the active time at night (Fig. 4.10) (Holloway and Geiser 1996). However, arousal from torpor can be energetically costly and offsets to some extent the advantage of torpor. Fat-tailed dunnarts enter torpor only when low ambient temperature and food shortage occur together. If food is abundant, the dunnarts remain active even at ambient temperatures as low as 5°C. Food shortage is the critical factor that provokes torpor, since the animal can only carry sufficient reserves in tail and body fat to meet its energy requirements for one to two days at normal metabolism. This degree of torpor is not nearly as profound as in the eastern pygmy possum, Cercartetus nanus (see Table 6.2). Although the critical period for the fat-tailed dunnart is the early winter, when insect food is scarce and body weight is minimal, torpor is apparently not controlled by changes in photope- riod, as it is in some northern hemisphere rodents (Holloway and Geiser 1996). The conclusion from this, and studies on other small dasyurids, is that food availability is the main driver, since their demands are so intense and continuous. While they are active they must constantly hunt for food and if it is unavailable they can conserve energy and water for only one or two days. As Nagy et al (1988) wrote ‘free ranging dunnarts were working very hard during spring, even though this is when their insect food is expected to be most abundant’. Spring is also when they are breeding and face increasing metabolic demands.

Breeding in fat-tailed dunnarts Throughout its wide range the fat-tailed dunnart is a strictly seasonal breeder with two peaks of birth, in August and October. The first births occur in July, shortly after the winter solstice, and the last young are weaned in February, having been born before the summer solstice. In captivity dunnarts will breed when exposed to long day length and are inhibited by short day length. However, the critical signal to commence breeding in females or develop enlarged testes in males is the change from short to long day length (Godfrey 1969, Holloway and Geiser 1996). The response does not seem to be as subtle as the signal that provokes antechinus into breeding and this may be because the period of reproductive activity in these species is so prolonged. Teat number is 10 but not all are occupied, so the litter size varies from 1 to 10 and declines through lactation. In this way the reproductive rate can respond to food availability within the six-month breeding period. A similar pattern of breeding through the second half of the year occurs in three other species of dunnarts as well as in the smaller planigales but it is not known whether they are likewise controlled by photoperiod.

Conservation Except for the thylacine in Tasmania and the eastern quoll in eastern Australia, no species of carnivorous marsupial has become extinct since European occupation of Australia, although the range of several has been reduced drastically. While the thylacine was brought to extinction by persistent shooting throughout the 19th century and into the 20th century, it is not clear what caused the demise of the eastern quoll. In the early 20th century it was relatively common around and coastal New South Wales, and it is still common in Tasmania but has now probably gone from eastern Australia. The decline of the eastern quoll was not monitored but has presumably been hastened by the spread of the fox, which became established in New South Wales between 1900 and 1910 (Short 1998). Also, the development of logging tracks and other roads has enabled foxes to penetrate Predatory marsupials of Australasia: bright-eyed killers of the night 161

deep into forested regions in New South Wales, the preferred habitat of the eastern quoll (Catling and Burt 1995). The larger spotted tailed quoll is still present in these forests but being larger than a fox is better able to withstand it. The western quoll has disappeared from most of its former range in inland Australia (Burbidge et al 1988) but still occurs in the forested parts of Western Australia, where a different factor may be protecting it from fox predation. In Western Australia some native plants contain sodium fluoroacetate (1080) as a protective chemical, which even in very small amounts is lethal to foxes. Native mammals, however, have evolved a tolerance to the chemical, so that the herbiv- ores can eat the species of Gastrolobium and Oxylobium and thereby acquire quite high concen- trations of 1080 in their tissues. Quolls that prey on these species also have evolved a tolerance to the chemical. Since foxes have a very low tolerance for 1080, they cannot survive much beyond the margins of the native forests, where the native species persist. This difference in tolerance is now being successfully exploited in Western Australia to reduce foxes, by distributing meat baits that contain a dose of 1080 that is lethal for foxes but will not affect native predators that eat it (King et al 1978). The red-tailed phascogale has likewise suffered a huge reduction in its range in Western Australia, through loss of habitat in land cleared for wheat and sheep. It now persists only in the coastal woodlands, where Gastrolobium occurs and stock cannot survive, because of their lack of tolerance for 1080. Of the smaller species of dasyurids none has gone extinct in Australia, although the range of some may have contracted. To the contrary, as interest in dasyurids has increased in the last 40 years, several new species have been discovered or recognised, and other very rare species have been rediscovered. The dibbler was rediscovered in 1967 (Strahan 1995) in the coastal scrub of southern Western Australia, where it had been found by Europeans 83 years before. It has now been bred successfully in captivity and some of these animals are being returned to the wild. The Julia Creek dunnart, the largest dunnart (Fig. 4.1b), was identified as a distinct species from four specimens collected between 1911 and 1972 and lodged in the Queensland and Australian Museums. It was rediscovered in western Queensland by Woolley (1992), who revis- ited the original collection sites and searched for its remains in pellets; being very good collectors of small mammals. She found remains of the species in the pellets and subsequently caught live specimens on grazing properties near Julia Creek. The dunnarts shelter in cracks 2 m below the ground and they depend for their food on insects from the grasslands above them. The species has a very restricted distribution on the Mitchell grass downs on the cracking clay soils of northwestern Queensland, but the factors that limit its distribution are not yet clear. Predation by cats, as well as overgrazing by stock may be important, but natural events such as flooding and fire may also affect the populations of this rare species.

Remaking the thylacine? The extermination of the thylacine in Tasmania was deplorable and the wish to see the species alive understandable. So the idea that a thylacine might be resurrected from the DNA stored in museum specimens is intriguing. But is it possible? Jules Verne imagined travel under the north polar ice cap and a visit to the Moon, both of which became realities within a century of his writing. However, the obstacles to creating a thylacine are far greater than either one of those technical achievements. How so? First, the DNA of the specimens preserved for a century in alcohol must be extracted in unbroken strands and retain all the genetic information to make a new individual animal. This is extremely unlikely because breakdown products from alcohol, over time, become attached to the DNA. So far a few thousand base pairs have been sequenced, which seems a lot until one realises that about 65 million base pairs make up the genome of a mammal. Furthermore, this 162 Life of Marsupials

genome must be in its correct order and connections in order to be read and transcribed by the machinery of the encompassing egg cell. Second, if this prodigious task could be accomplished there are no viable thylacine eggs that could respond to the commands of the thylacine genome, so the egg of the nearest living relative would have to be used, namely that of the Tasmanian devil. Obtaining eggs from Tasmanian devils at the right stage of development may be feasible, although it has never been done, but the cytoplasm of a mammalian egg carries a considerable amount of developmental information, including all the mitochondrial DNA for the offspring. So, even if the cytoplasm responded to the commands of the thylacine DNA, the resultant embryo would not be a thylacine but a hybrid. A slightly easier option might be to fertilise a Tasmanian devil’s egg with thylacine DNA from a testis in which only one set of chromosomes is introduced – rather like the technique developed for infertile human couples called intracellular sperm injection (ICSI). Although this might have a greater likelihood of success, it still presupposes that the chromosomes of the two species have sufficient common regions to undergo the process of syngamy, where complementary chromo- somes unite, exchange DNA and separate. If that could happen, the result would be a hybrid between the two species, having half its chromosomes and all its mitochondrial DNA from the Tasmanian devil and half its chromosomes from the preserved thylacine. It would be a sort of mule, except that the relatedness of the Tasmanian devil and thylacine is much more distant than between a horse and a donkey – more like the relatedness of a horse to a rhinoceros. Third, supposing such an animal was to be produced, what then? The very word mule is synonymous with infertility. But much before this stage is reached the hybrid egg would have to undergo successful development in the uterus of a Tasmanian devil. Female devils habitually shed about 50 eggs of which a large number fail to develop normally during the short gestation period. It may soon be possible to support the intra-uterine stages of marsupial development in culture but that has not yet been achieved in any species. If it can be done, then the hybrid embryos might be taken through to full term. The newborn Tasmanian devil weighs 18 mg, which is about the same as all dasyurids, so it is likely that the thylacine newborn was little if any bigger than this. If the fetus is not rejected by the Tasmanian devil and was to be carried to term, its birth would not pose any serious problem. The more risky part would be its survival against the competition from other newborn Tasmanian devils, all competing for the four teats. Nurturing the embryo in the pouch of a Tasmanian devil or other dasyurid is probably the easiest part of the whole improbable exercise. In other cross-fostering experiments with macropods the young have grown to independence in the pouch of a foster mother of a different species. To breed out the Tasmanian devil component in the hybrid offspring and select for thylacine attributes would then require a great deal more time and funds. Just consider the time it has taken dog fanciers to reselect the qualities of the King Charles spaniel, using animals that were members of a single species. Finally, what is to be the fate of the reconstituted thylacine? Returned to the forests of Tasmania? Most of those where it was last seen have been logged or chipped. Kept in zoos? Is a new addition to a menagerie worth so much effort and cash? Surely the truly huge effort that this project entails is misplaced: it is squandering limited resources that would be far better used to ensure that other species of marsupial, still alive today, do not go the way of the thylacine. But dreams are more powerful than less spectacular and far more achievable goals.

Conclusions Throughout the long sweep of marsupial occupation of Australia marsupial predators covered the whole range of body size from the minute ningauis and planigales to the massive Thylacoleo. Predatory marsupials of Australasia: bright-eyed killers of the night 163

In the Miocene, thylacinids dominated the predator niche but in the last five million years the diversity of species appears to have increased, with the rise of the medium to small dasyurids and the appearance of Thylacoleo. Whether the rise of the dasyurids is really as recent as the fossil record suggests or whether they were present earlier but have not so far been discovered, remains for future research to disclose. Since the arrival of four new placental predators – humans, dogs, cats and foxes – the largest marsupial predators successively went extinct. The first was Thylacoleo after humans arrived; the thylacine after dogs arrived; and the smaller ones experienced severe reductions in range after cats and foxes spread across the continent. Only the smallest species seem to be holding their own. Direct competition with the new predators was probably the major factor in the earlier extinctions, whereas loss of prey through alteration of habitats by grazing stock may now be as important a factor in the continuing survival of the smaller species, like dunnarts. For these species, whose staple diet is invertebrates, a constant and regular supply of food is essential for survival. Integrity of the natural habitats, not genetic reconstruction of extinct species, is there- fore the key to long-term conservation of the remaining marsupial carnivores. This page intentionally left blank Chapter 5

Bandicoots: fast-living opportunists Pig-footed bandicoot, ecaudatus; lithograph by John Gould. Bandicoots: fast-living opportunists 167

he word bandicoot comes from India. In its South Indian (Telegu) form, pandi-kokku, it means pig-rat and refers to the large rat, Bandicoota bengaliensis. George Bass first Tused the term for some Australian marsupials in 1799, during the circumnavigation of Tasmania with Matthew Flinders. The term is now so commonly used in Australia for members of the marsupial family Peramelidae that no other common name is considered. However, it still confuses people who have lived in India, leading to complaints about bandicoot diggings in suburban gardens and demands to destroy these introduced pests from India. But the Australasian bandicoots are marsupials, very different indeed from the Indian bandi- coots, which are rodents. All are medium-sized animals of a fairly uniform shape with a pointed nose (Fig. 5.1, Plates 8 and 9), which they use, much as pigs do, in searching for ground-living organisms. Indeed, in New Guinea the Motu name for the largest marsupial bandicoot, Peroryctes broadbenti, translates as ‘pig’s younger brother’. Bandicoots are still widely distributed on New Guinea and Tasmania but on mainland Australia they have undergone severe contractions of range since European occupation, and several species are now extinct or very rare. Bandicoots are opportunists, which exploit their environment by their wide choice of food, fast growth, rapid reproduction and short life span: they live fast, die young and leave large families. Several unique characters associated with their life style differentiate them from other marsupials and trying to understand their relationship and past history is a key interest in marsupial biology today. While bandicoots are undoubtedly marsupials, fitting them into the family tree has always been a puzzle. They have some features in common with the carnivorous marsupials and others in common with the possums and kangaroos – some features, including their DNA, seem to link them more closely to South American marsupials.

Special characters of bandicoots The most distinctive character of bandicoots, which sets them apart from all other marsupials, is their complex allantoic placenta (see Fig. 2.10), the cells of which invade and fuse with the cells of the uterine wall, as in many placental mammals. Yet they cannot be classed with placental mammals, because they bring forth very small young and nurture them in a pouch, like other marsupials, and their reproductive tracts are unequivocally marsupial. Bandicoots in both Australia and New Guinea have the ancestral marsupial chromosome number of 14, apart from the bilby, Macrotis lagotis, which has 16 chromosomes. This is proba- bly the result of the splitting of one pair of chromosomes into two pairs. However, bandicoots differ from other marsupials in losing one sex chromosome from all somatic cells during early embryo development – one X chromosome in females and the Y chromosome in males. Like the carnivorous opossums and dasyurids, bandicoots have many incisors in the upper jaws and three or more pairs of incisors in the lower jaw in front of the larger canines (see Fig. 1.4). This sets them apart from the possums and kangaroos, which in the lower jaw have only a single pair of incisors and no canines. Their molar teeth reflect their omnivorous diet (Fig. 5.2), being squared up with rounded cusps for grinding, rather than the sharp pointed, shearing molars of dasyurids. However, the feet of bandicoots are more like those of kangaroos and possums, than of dasyurids and didelphids. On their hind feet digits 2 and 3 are small and partly fused to form a comb, used for grooming (see Fig. 1.4). Whether this shared character of fused toes, or syndac- tyly, is because they have a common ancestry with kangaroos and possums, or whether it evolved independently, remains uncertain. The bandicoots seem to occupy an intermediate position between the two other major groups of Australasian marsupials, although other characters, such as unconjugated sperm and their pattern of ankle articulation (Szalay 1982), place them clearly with the rest of the Australian marsupials rather than with the American marsupials. 168 Life of Marsupials

Figure 5.2: Molar teeth of bandicoots: (a) upper molar of , Isoodon obesulus; (b) upper molar of bilby, Macrotis lagotis; (c) upper and lower molar rows in partial occlusion of the extinct pig-footed bandicoot, Chaeropus ecaudatus, to show its apparent adaptations for eating grass. After Rich (1991) and Wright et al (1991).

Bandicoot antecedents The position of the bandicoots among marsupials has become even more anomalous since the analyses of mitochondrial and nuclear DNA sequences and DNA hybridisation studies. There is general agreement on these criteria that the bandicoots are a very ancient separate group, co-equal with the didelphid and caenolestid groups in South America and not closely related to either the dasyurid or diprotodont marsupials of Australasia (Springer et al 1998). On all these criteria they have closest affinity to the caenolestids (see Fig. 1.10) (Retief et al 1995, Krajewski et al 1997, Palma and Spotorno 1999). This relationship was first suggested by Osgood in 1921 on anatomical criteria but was not seriously considered by others workers: now it is being reconsidered. A serious difficulty with this is that the caenolestids, like the didelphids, have conjugated sperm (see Fig. 1.6), whereas no Australasian group, including bandicoots, does. However, if on other criteria, there is a closer relationship to these South American marsupials than to any Australian marsupials, it implies that the bandicoots have been a separate group since before the time that South America and Australasia were connected via Antarctica, 45 million years ago. Based on all the DNA criteria, Bandicoots: fast-living opportunists 169

a. New Family 9 7

New Family

Peroryctidae 10

Peramelemorphia Peramelidae 577

Thylacomyidae 32 2 55 23 26 5.2 16.3 10.4 0.01 million years ago

b. 8 Microperoryctes longicauda 11 Echymipera kalubu New Guinea 16 Echymipera clara Peroryctes raffrayana 25 4 Isoodon macrourus 7 Isoodon obesulus Isoodon auratus Australia 12 Perameles nasuta Macrotis lagotis

Figure 5.3: The relationships of Australian and New Guinean bandicoots: (a) the distribution of the fossil ancestors of bandicoots during the Tertiary (numbers in bars indicate number of genera more than one); (b) probable relationships of living species, based on DNA hybridisation data, with estimated time of separation of branches in millions of years from present. After Muirhead (1999) and Kirsch et al (1997). the calculated time of separation of bandicoots from other Australian marsupials is estimated to be about 60 million years ago, or well before the separation of the southern continents. In support of their long separate lineage, bandicoot-like molar teeth have been found at Murgon, southeast Queensland, dated to 55 million years ago (Fig. 5.3a). Fossil bandicoots have also been described from the late Oligocene to the late Miocene epochs, 26–10 million years ago, but these fossils cannot be linked directly to any present-day bandicoots. The three present day families are known only as far back as the Pliocene epoch, about 5 million years ago (Muirhead 1999). This last great expansion of the bandicoots is thought to have been a response to the major climatic changes in Australia and the emergence of New Guinea during the Pliocene and Pleistocene epochs. 170 Life of Marsupials

Living bandicoots Living bandicoots form three easily distinguished groups but there is disagreement as to whether they represent three families or subfamilies. Here we will treat them as two Australian families and a New Guinean family of the Order Peramelomorphia (Fig. 5.3). The best-known family is the Peramelidae, which comprises seven species in two closely related genera, Perameles and Isoodon. In body mass they range from 200 g to just less than 1 kg, although in captivity they may grow much heavier. There were four species of long-nosed bandicoots, which were distributed in eastern and southern Australia and through the western desert country (Fig. 5.4). The long-nosed bandicoot, Perameles nasuta is still relatively common on the eastern seaboard, and the eastern barred bandicoot, Perameles gunnii, is still common in Tasmania but is reduced to a single relict population in Victoria. The , Perameles eremiana, is extinct, as is the , Perameles bougainville (Fig. 5.1b, Plate 8)

Figure 5.4: Distribution of living bandicoots in New Guinea and Australia, and former distributions of rare and extinct species shown in outline. Note the presence of only one New Guinea species in Cape York and one Australian species in southern New Guinea. After Strahan (1995) and Flannery (1995). Bandicoots: fast-living opportunists 171

except for two populations on Bernier and Dorre islands, off the West Australian coast. The three short-nosed bandicoots are the , Isoodon auratus, the southern brown bandi- coot, Isoodon obesulus (Fig. 5.1a, Plate 8), and the northern brown bandicoot, Isoodon macrourus, which are severally distributed around Australia and in Tasmania (Fig. 5.4). Of all seven species only the northern brown bandicoot extends into southern New Guinea. The second wholly Australian family is the Thylacomyidae, which comprises the two species of rabbit-eared bandicoots, the formerly widespread but now endangered bilby, Macrotis lagotis (Fig. 5.1c, Plate 9), and the now extinct lesser bilby, Macrotis leucura, known only from central Australia. Bilbies are the only bandicoots that excavate burrows, using their powerful broad- nailed forefeet. Their burrows are easily identified by the characteristic tracks and tail mark, as well as the unusual construction of the burrow, which descends in a fairly steep and winding spiral for about two metres and has no exit hole. They occupy the burrow during most of the daytime and forage at night. Their large ears are folded forward during sleep and open out during activity. Originally widespread from the Great Dividing Range in New South Wales to Western Australia (Fig. 5.4), the bilby is now confined to those parts of its former range that are north of the permanent distribution of rabbits (Morton 1990). The bilby’s diet consists of insects, bulbs and fungi. The pig-footed bandicoot, Chaeropus ecaudatus, previously included in the Peramelidae is, on biochemical criteria, probably more closely related to the bilbies (Groves and Flannery 1990, Westerman et al 1999). Now extinct, it formerly occurred over a wide area in central and south- ern Australia. It was distinguished by its long thin legs and reduced digits, hence the common name. On its forefeet the digits were reduced to two (digits 2 and 3) and on the hind feet to one main one (digit 4) and the tiny comb of digits 2 and 3. Sadly, very little is known of its habits because it disappeared so quickly from any country occupied by Europeans. Aborigines in central Australia told Andrew Burbidge and his colleagues (Burbidge et al 1988) that the pig-footed bandicoot was found in sandplains and dunes with spinifex and tussock grass, sometimes with a mulga overstorey. It lived in a grass-lined nest in a scrape or short burrow and it ran or scampered but did not hop. Gerard Krefft (Krefft 1862), who collected the first speci- men near the , thought its diet was solely grass, although later writers thought it fed also on other plants and insects. A recent analysis of the stomach contents of museum specimens and the wear facets on its molar teeth supports Krefft (1862) (Fig. 5.2c) (Wright et al 1991). Like other bandicoots, the females had eight teats in a backward opening pouch but the litter size was two and its breeding season was estimated to be May–June. It had disappeared from South Australia by the 1930s but in central Australia it remained common until the 1950s. The third family is the Peroryctidae of New Guinea. Because the hair of some species is coarse they are sometimes called spiny bandicoots. There are 10 endemic species in four genera, all of which occupy forested habitat, ranging up to 3900 m (Zeigler 1977, Flannery 1995). Peroryctes comprises one widespread species, Peroryctes raffrayana (Fig. 5.4) and one species, Peroryctes broadbentii, restricted to the south-eastern tip of the island). It is the largest of all bandi- coots, in which males may weigh 5 kg, greatly exceeding the weight of any Australian species. Microperoryctes longicauda has a widespread distribution, whereas Microperoryctes papuensis only occurs in the south-eastern end of the island. Microperoryctes murina is the smallest New Guinean species, which at 120 g weighs less than any Australian species, and it is restricted to the western end of the island. The third genus, Echymipera, contains four species (Fig. 5.1d, Plate 9), of which three are widely distributed across the island, and one, Echymipera rufescens, extends into the northern rainforest of Cape York (Fig. 5.4). These four species have adapted to human habitations but whether they are as opportunistic in their habits as the Australian bandicoots is still unclear. The fourth genus is represented by a single rare species, Rhynchomeles prattorum, found only on Seram Island where it is the only species of bandicoot. 172 Life of Marsupials

As with the dasyurids, there is very little overlap of species between Australia and New Guinea: the overlap is restricted to the area of the last land connection between the two land masses. In both groups the inference is that the ancestral dasyurids and bandicoots extended through the whole continent and, after the emergence and separation of New Guinea from Australia, separate adaptive radiations occurred on each. When land connection was subsequently restored during low sea levels in the late Pleistocene a small amount of exchange occurred in both directions. A similar conclusion is arrived at for the arboreal marsupials (see Chapter 7) and the kangaroos (see Chapter 9).

Lifestyle of bandicoots Most of the remainder of this Chapter will consider the five species whose physiology and ecology is best known. They are the long-nosed bandicoot, the eastern barred bandicoot, the southern brown bandicoot, the northern brown bandicoot and the bilby.

Diet and activity Australian bandicoots are omnivorous animals that will eat insects, other invertebrates, under- ground fungal fruiting bodies, as well as grasses, seeds and roots of plants. They make character- istic scratches as they forage for soil invertebrates and fungi, using their long noses. By feeding on soil fungi and passing their spores in their faeces, they are thought to aid in the spread of mycorrhizal fungi that live in roots of eucalypt forest trees and help them to acquire nutrients from the soil: rat kangaroos are far more effective at this (see Chapter 9). Bandicoots are generalists in their diet and in the habitats they occupy. These range from tussock grasslands and heaths to open forest. They also respond to altered habitats such as the regenerating stages after fire, rubbish tips and open paddocks, which provide an abundance of insect larvae. As opportunists they respond to favourable conditions by a rapid rate of reproduc- tion and equally rapid colonisation of adjacent habitat. Their solitary habit, lack of any discerni- ble social structure and early dispersal of juveniles help this to occur (Lee and Cockburn 1985).

Physiology Hume (1999) has described in detail the digestive physiology and metabolism of bandicoots and the bilby. By comparison with dasyurids, bandicoots have a longer gut, including a short caecum, and the passage of food through it is substantially slower (10–24 hours, compared to 5–7 hours in the eastern quoll, Dasyurus viverrinus, and Tasmanian devil, Sarcophilus harrisii). These differences reflect the omnivorous diet of bandicoots, particularly the longer time required to digest plant material than insects. The bilby has a larger caecum than the bandicoots, further enlarged by an expanded colon, in which the ingested plant material is retained and possibly subjected to bacte- rial fermentation, as in some possums (see Chapter 7). The pig-footed bandicoot also had a large caecum and a partly divided stomach, in both of which chambers bacteria have been found that may have been associated with fermentation of the predominantly grass diet of this bandicoot. Bandicoots and bilbies are too large to be exclusively insectivorous, because of the time required to collect and process large quantities of small insects, and too small to be exclusively herbivorous. By having a digestive tract that is more differentiated than the simple tube of a carnivore and the ability to retain food in the hind gut, they can switch between plant, animal and fungal material according to the relative abundance of each type of food. The omnivore niche may also explain the limited size range of bandicoots, compared to the carnivorous dasyu- rids: the larger carnivores can tackle large prey such as birds and mammals, which bandicoots cannot do because of their different kind of teeth. Bandicoots: fast-living opportunists 173

Bandicoots fill the opportunistic generalist niche among marsupials and one component of this is their rapid reproductive response to changing circumstances. It might also be inferred from this that bandicoots have a higher metabolic rate than other marsupials and nearer to the rate of placentals. However, this is not so in the four species of bandicoot in which the standard metabolic rate (SMR) has been measured (see Hume 1999). The SMRs of the long-nosed and northern brown bandicoot are right on the average for marsupials, 70% of the placental stand- ard, while the golden bandicoot and the bilby are 47% and 58%, respectively, which may be an adaptation to the desert environment in which they live. On this slender evidence bandicoots are not different from other marsupials. Field metabolic rates (FMRs) have also been determined in three species (Green 1997) and they range from 1.4 to 4.3 times the SMR. Compared to dasyurids the bandicoots do not display a large metabolic scope but the amount of information is not sufficient to say whether there are big differences between seasons or in the cost of reproduction.

Reproduction Reproduction is much more rapid than in any other group of marsupials, even including the gray short-tailed opossum, Monodelphis domestica (see Chapter 3). The length of pregnancy in the long-nosed bandicoot and the southern brown bandicoot is 12.5 days, the second shortest of any mammal, and lactation lasts only 45 days, so that the production of a litter to independence is accomplished in two months. The female returns to oestrus during the last 10 days of lacta- tion so that she gives birth to a new litter as the older litter is weaned. In this way a succession of litters can be produced every 60 days, while conditions are favourable. While all Australian species have eight teats in the backward directed pouch, mean litter sizes are less than four and the new litter preferentially use the teats that were not used by young of the previous litter. However, in the northern brown bandicoot, Merchant (1990) found that some young do become attached to regressing teats and they survive. Initially they grew slower than their litter mates, possibly because the composition of the late stage milk still being secreted was unsuitable for the newborn. While the different size of the teats may play some part in the small litter size in bandicoots, the main reason is more likely that the lactating female cannot raise a litter of more than four young because of the very large energy requirements of late lactation (see Growth rates and longevity). Unlike in opossums and dasyurids, in which the number of eggs shed at ovulation far exceeds the number of teats in the pouch, in bandicoots the number is about half the number of teats in the pouch, 4.9 in the northern brown bandicoot and 3.3 in the long-nosed bandicoot (Lyne and Hollis 1977). Two factors contribute to the extraordinarily fast rate of reproduction in bandicoots, when compared to other marsupials of comparable size: the complex placenta and the composition of bandicoot milk. By comparing two other Tasmanian marsupials of approximately the same size as the eastern barred bandicoot, namely the eastern quoll and the Tasmanian bettong, Bettongia gaimardi, we can see how different the bandicoot investment in reproduction is (Table 5.1). Despite their very short gestation, bandicoots deliver several young that are each the equal in mass of the single young of the Tasmanian bettong and 20 times the mass of the newborn eastern quoll. This suggests that the intimate placenta of the bandicoot and bilby is a more efficient organ of exchange than the yolk sac placenta of the other species. This must give the young a quick start, since the duration of lactation is also the shortest of all marsupials. The embryonic vesicle of the long-nosed and northern brown bandicoot becomes attached to the uterine wall on day nine of pregnancy by fusion of fetal and maternal cells to form a true allantoic placenta, which supports it for the remaining 3.5 days. This is very similar to the first three days of placental formation in the rabbit, which begins on day eight. However, in the rabbit, 174 Life of Marsupials

the placenta develops into a much more complex structure that supports the fetus for another 20 days. The result is that the similar sized wild rabbit produces several far larger neonates after a gestation only twice the length of the bandicoot’s gestation, and they then require a far shorter period of postnatal support (Table 5.1). An interesting question raised by this comparison is what prevented bandicoots from exploit- ing the allantoic placenta to the same extent as did placental mammals? Was it some particular aspect of the marsupial heritage that truncated this pathway? For instance, the convoluted repro- ductive tract (see Chapter 1), may have prevented larger fetuses from being born. It has also been suggested (Amoroso and Perry 1955) that marsupials did not evolve a means to protect the fetus from rejection by the mother’s immune system and so their young are perforce delivered when still minute, before such rejection could take place. Experiments to test this idea in the tammar wallaby did not support the idea of immunological intolerance of the fetus (see Chapter 2), but no one has tested it in bandicoots. Whatever the reason for the truncated pregnancy of bandicoots, their young are born at a relatively advanced stage for marsupials: the brain and eye are as advanced in development as those of kangaroos, which have a gestation more than twice as long, and their growth rate after birth is faster than all other marsupials.

Table 5.1: Comparison of maternal investment in fi ve species of marsupial and a placental, all of about the same adult size Data from Tyndale-Biscoe and Renfree (1987).

Species Mass (kg) Gestation Neonate Litter size Lactation (d) (mg) (d) Eastern quoll 0.9 19 13 6 140 Tasmanian bettong 1.2 21 320 1 160 Barred bandicoot 0.8 12.5 250 3 60 Bilby 1.1 14 350 2 90 Wild rabbit 1.2 28 60 000 6 21

Growth rates and longevity While the growth rate of all marsupials is slower than for equivalent sized placental mammals, bandicoots have the fastest growth of any marsupial (Lee and Cockburn 1985) and the shortest period of lactation. This rapid rate of growth is aided by milk that is richer than that of any other marsupial, reaching 55% solids at the end of lactation (Merchant and Libke 1988): this is richer than that of all placental mammals except seals (78%) (Oftedal 1984). As in other marsupials the composition of the milk changes through lactation, with the early milk being low in fat and high in carbohydrates (see Chapter 2). The change to milk rich in lipid and protein occurs at about day 25, 15–20 days before the young are weaned (Fig. 5.5a). Not only is the milk very concentrated and high in lipid during this phase, but also the volume ingested daily by each young rises to 20 mL: this is much more than the volume taken in by the eastern quoll or tammar wallaby young at day 50. For a litter of three or more young this is a very substantial outlay by the mother, so how is the mother able to deliver this? During the early stage of lactation the mother bandicoot increases in body mass by 10–20%, depending on the time of year, by laying down body fat, including two large strands in the body cavity, which reaches a maximum at day 35 (Fig. 5.5b). In the second phase of lactation this fat depot is mobilised into the copious rich milk that supports the young. As a consequence the mother bandicoot loses up to 200 g, or about one-third of her body weight. Bandicoots: fast-living opportunists 175

60

50 Lipids

40

30 Proteins

20 Solids fraction (%)

10 Carbohydrates

0 200 Weaned 200 Mother

150 150 Pouch Young

100 100

50 50 Pouch young body weight (g)

0 0 Mean change in body weight of mother (g) 0 10203040506070

Days since birth

Figure 5.5: Lactation in the northern brown bandicoot, Isoodon macrourus: (a) changing composition of the milk through lactation; (b) changes in body weight of mother and young to show the large transfer of maternal lipid reserves in the last third of lactation. After Merchant and Libke (1988), Merchant (1990) and Merchant et al (1996).

Simultaneously with this last phase of lactation the female bandicoot returns to oestrus and conceives again, so that the next litter is born as the previous litter is weaned. Clearly, such a strategy requires continuous and adequate supplies of food, which can be rapidly converted into independent young. Bandicoot breeding reflects this and the hormonal control of the process is also unusual. The corpora lutea formed at ovulation secrete into the circulation high levels of progesterone from day 9 of pregnancy and for the first three weeks of lactation (Fig. 5.6). Even after this time they apparently suppress ovulation while young remain in the pouch. From day 35 prolactin is elevated and remains so until about day 50, when the young are left in a nest and prolactin returns to basal levels. This is a similar pattern to the tammar wallaby (see Chapter 2) and prolactin probably stimulates the secretion of the large volumes of rich milk during this 176 Life of Marsupials

16 80

14 Progesterone Prolactin 12 60 ) 10

8 40 Weaned 6 rolactin (ng/ml

4 20 P

Progesterone (ng/ml) 2

0 0 -10 010203040506070

Oestrus Days since birth

Figure 5.6: Hormone profiles in the northern brown bandicoot, Isoodon macrourus, through pregnancy and lactation. Note that progesterone in circulation is elevated from day 4 of pregnancy (8 days before birth) until day 19 of lactation and prolactin is elevated from day 35 until day 50 of lactation, during the period of maximum milk production. After Gemmell (1981) and Hinds (1988).

time. Decline of prolactin coincides with the onset of oestrus and ovulation in the female and the start of the next pregnancy. If young are lost prematurely, the return to oestrus occurs earlier. In the southern regions of Australia and Tasmania, which have a regular winter rainfall, the breeding seasons of bandicoots are restricted to the months of July–March when insect and invertebrate food is most abundant. However, in Queensland, where food supply is less variable, the northern brown bandicoot breeds throughout the year, with a mean litter size of 3.4. In New South Wales breeding began when the females gained weight and ceased when their weight declined. Further south in Victoria breeding by the southern brown bandicoot was highly synchronised in three successive years, beginning after the winter solstice (Fig. 5.7). This suggested that they were responding to a photoperiod signal, either as a direct response in the bandicoots or indirectly through greater abundance of prey species responding to photoperiod. In Tasmania eastern barred bandicoots produced three or four litters in one season but the earliest and latest litters were smaller than the middle ones, which correlated with the changing body weight of the females (Heinsohn 1966). However, survival of the first litter was higher than the subsequent litters because the young were weaned at the time of most abundant food: they were able to establish themselves and breed by the end of the season of their own birth, further contributing to the high rate of increase of the population. The high potential rate of reproduction of bandicoots is matched by a high rate of mortality of young. In the northern brown bandicoot 50% of the young had died by 100 days and 80% by one year, after which the residual population survived until the third year (Gordon 1974). Similarly, in the southern brown bandicoot in Victoria the mortality in the first year was 50% and no animal lived beyond 3.5 years (Fig. 5.8) (Lobert and Lee 1990). And in the long-nosed bandicoot near Sydney less than 10% of the young survived to become sub adults and only 3% survived to become adults (Scott et al 1999). Bandicoots: fast-living opportunists 177

100

80

60

40

20

Percentage of pouch young Percentage females carrying 0 MJ J ASONDJ FMAMJ J ASONDJ FMAMJ J ASONDJ FMAMJ 1974 1975 1976 1977

Figure 5.7: Annual breeding in the southern brown bandicoot, Isoodon obesulus, in Victoreia, to show the highly synchronised start and close of breeding each year. From Stoddart and Braithwaite 1979.

100

80

60 Males 40 Females

20

Percentage minimum survival 0 non-br br non-br br non-br br non-br 1st Year 2nd Year 3rd Year 4th Year

Figure 5.8: Lifetime survival curves for 34 male and 28 female southern brown bandicoots, Isoodon obesulus, in Victoria, measured during the breeding and non-breeding seasons of each year. After Lobert and Lee (1990). 178 Life of Marsupials

Thus, bandicoot reproduction is highly responsive to environmental and nutritional conditions. Under adverse conditions the breeding season is delayed or the litter size is reduced, while under good nutritional conditions reproductive performance is influenced directly by the body weight of the female. This is not surprising when one considers that the successful rearing of a litter depends on delivering large volumes of rich milk in the final weeks of lactation. What little is known of reproduction in the bandicoots of New Guinea suggests that they also have small litters of one to three young and that they breed throughout the year (Aplin and Woolley 1993). All these features of bandicoot life history are clear adaptations for rapid exploitation of habitats that are temporarily favourable and give them a distinct advantage over other mammals. Cockburn (1990) asked why, with such advantages, bandicoots have been such a conservative group of marsupials through their long evolutionary history. He concluded that the selection for very rapid growth and development of the young, within the constraints of the marsupial organisation, closed off the options for subsequent diversification, as seen in the other major family groups: bandicoots have specialised in being generalists and opportunists. Seen in this context the apparently advanced placenta of the bandicoots may be viewed as an adaptation that provides the most rapid development of the fetus within the constraints of the marsupial plan. While this evolutionary pathway was a successful one for 60 million years, it has been singularly unsuccessful when responding to the massive environmental changes of the last 200 years in Australia.

Conservation As mentioned at the beginning of this Chapter, bandicoots in Australia have suffered severe decline since European occupation of the country, with all species showing reduction in range, three species going extinct and two other species surviving only on offshore islands. The same cannot be said of New Guinea, where the variety of species remains as it was. In Tasmania the two original species of bandicoot are still flourishing, despite large changes to the habitat. So, what caused the massive decline of bandicoots in mainland Australia? As we have seen, these are the opportunists among marsupials, which it might be supposed, would have been the most likely to survive the habitat changes caused by land clearing and grazing stock. But they did not. Because the decline began so early it is largely unrecorded and the causes of it are now difficult to determine precisely. First, there are the records of early explor- ers and naturalists, then the memories of the Aboriginal people and finally the ideas of ecologists today, who have seen the last stages in the decline. John Gould travelled in Australia between 1838 and 1845, collecting material for his Mammals of Australia (1863). Species such as the bilby, the pig-footed bandicoot and the western barred bandicoot were then distributed across the southern half of the continent from the Swan valley settlement in Western Australia to New South Wales and Victoria. Twenty years after Gould’s observations Krefft (1862) wrote that these species were disappearing from Victoria in the face of settlement and the arrival of grazing sheep and cattle. Jones (1924) recorded the decline of these species from South Australia by the turn of the century. However, further north, according to the descendants of the Aboriginal people, who lived there until the 1960s, these species survived until the 1950s and 1960s in the desert and semi-desert country of Western Australia and Central Australia. A century after Gould, Troughton (1957) concluded that bandicoots ‘appear to be rapidly disappearing over the whole continent’. This is fortunately not quite true, since three species appear still to be secure. Bandicoots: fast-living opportunists 179

Table 5.2: Status of Australian bandicoots Data from Strahan (1995), Seebeck et al (1990), Burbidge et al (1988).

Species Original distribution Period of Current status decline Perameles bougainville Southern Australia 1860–1900 2 Island populations Perameles eremiana Central Australia 1940–1970 Extinct Perameles gunnii Victoria 1930–2000 1 Relict population Tasmania – Abundant Perameles nasuta Eastern Australia – Secure Isoodon auratus North western Australia 1940–1970 3 Island, 1 relict population Isoodon macrourus Northern Australia – Reduced but secure Isoodon obesulus Southern Australia – Reduced, patchy distribution Tasmania – Secure Macrotis lagotis Southern and central 1930–2000 Northern part of range only Australia Macrotis leucura Central Australia 1920–1960 Extinct Chaeropus ecaudatus Southern and central 1860–1950 Extinct Australia Echymipera rufescens North Queensland – Secure

The present status of the original 11 species of Australian bandicoots is summarised in Table 5.2. In southern Australia, from Victoria to Western Australia, the early disappearance of bandi- coots closely followed the establishment of pastoralism and the attendant clearing of the woodland vegetation. Grazing by stock and later by rabbits would have reduced the available plants and associ- ated insects that the bandicoots depended on, and also removed the cover that gave them shelter from the sun and predators. However, extensive clearing for agriculture occurred in Tasmania but did not affect the two species of bandicoots there, which suggests that the fox, which was not released in Tasmania, may have had a large part in the extinctions on the mainland. Another factor in the vulnerability of bandicoots is their body size. Among Australian indig- enous marsupials and placentals the critical weight range in body mass is from 35 g to 5.5 kg. All Australian bandicoots fall within it and the only species that have survived well are those that live in high rainfall areas in southern Australia and Tasmania, and in the tropical north. What is the basis of the critical weight range? These are the species that are too large to be able to escape the extremes of the climate in crevices, like small dasyurids can, and so require more energy for thermoregulation. However, they are not large enough to survive food shortages or move long distances to favourable habitat, as kangaroos can. Because bandicoots have the fastest rate of growth of any marsupial they require abundant food during the extended period of repro- duction. It may be that the competition with introduced herbivores for the available resources affected bandicoot reproduction more directly than it affected other marsupials because of this requirement. Thus, the trait that enables them rapidly to exploit favourable conditions was their undoing when the habitat was severely altered. In Central Australia several species persisted in regions that were never brought into produc- tion for pastoralism and then declined rapidly after about 1950. Indeed, the decline of the central 180 Life of Marsupials

Australian species was more extensive and profound than anywhere else and took place in less than 25 years, which is extraordinarily fast. It also took place before foxes reached there, so fox predation cannot have been the cause of the decline. It has led to the interesting conclusion that the decline may have resulted from the abandonment of traditional life by Aboriginal people at about this time (see Burbidge et al 1988 and Chapter 10). The relict populations of the golden bandicoot on Barrow Island and the western barred bandicoot on Dorre and Bernier islands have survived well (Fig. 5.4). This has been ascribed to the absence of cats and foxes but also to the higher rainfall and absence of fire (Friend 1990). Indeed, a small population of the golden bandicoot on Hermite Island near Barrow Island was wiped out by 1914, after cats reached the island.

Consequences of living on small islands The western barred bandicoot had an extensive distribution from the Western Australian coast to the tablelands of New South Wales but now occurs on only two islands off the coast of Western Australia. The mainland form of the species was similar in size to the eastern barred bandicoot, but the two forms on Bernier and Dorre islands are very much smaller and are the smallest of the Australian bandicoots. Short and his colleagues (1998) think that this is an adaptation to survival under the special conditions of the island habitat. The islands have been isolated from the mainland for 8000 years and from each other for about 3000 years (see Chapter 9). The total number of bandicoots on the two islands fluctuates between 4000 and 5400, according to prevailing rainfall (see Table 9.7). Unlike other species of bandicoot the females on Bernier and Dorre islands are significantly larger than the males, although, like in other species, males outnumber females in the population. Furthermore, although the males are small, their testes are disproportionately large. These unusual features of the island populations have probably arisen from the competing pressures of dwarfism and the reproductive needs of the species. When small mammals are isolated on islands there is often a tendency towards dwarfism: the advantage is presumably that the total biomass of the species is divided among more individuals on the limited area available, thus favouring survival of the species, which might otherwise fall below a viable population size. Against this trend, however, the females must be large enough to carry young to independence and support lactation. On the islands the litter size is less than two, although the females have eight teats like all other bandicoots, and there is some suggestion that they carry the young for a shorter time than other related species. These constraints do not operate on the males, hence the smaller size of males than females. However, the excess of males over females, the high degree of overlap in the home ranges of males in the two populations and the relatively large size of the testes, suggest that the males have a promiscuous mating strategy and very strong competition for access to females. This also favours genetic viability in a small population. The golden bandicoots on Barrow Island are also smaller than the mainland animals of the same species, which may indicate the same trend to dwarfism. However, the island is five times the size of either Bernier or Dorre and the population is estimated to be about 70 000. The two small populations of the southern brown bandicoot on the Franklin Islands off South Australia are each about 500 animals and the individuals are also smaller than mainland and Tasmanian animals of this species (Copley et al 1990). These populations are of interest in showing how small a population can be and still survive in the long term. What can be learnt from the mainland extinctions and island survival of bandicoots, for the long-term conservation of relict populations of other species of bandicoot? Clearly, the major causes of bandicoot decline across the continent have been the loss of food and shelter from land clearing, pastoralism and altered fire regimes, and the increased risk of predation from Bandicoots: fast-living opportunists 181

introduced foxes. Islands have provided the best long-term survival for several species, while other species appear to be secure in those parts of the continent where rainfall is sufficient to ensure food and shelter. Nevertheless, species that appear to be abundant can decline rapidly, as occurred with several species of bandicoot in the first half of the 20th century. One way to monitor the continuing status of such species has been described by Opie et al (1990) for the long-nosed bandicoot and the southern brown bandicoot in eastern Victoria. Based on all present records of the two species and what is known of their ecology, all potential areas that met their requirements were identified, using the computer program BIOCLIM. This disclosed large suitable areas that are unoccupied by either species. The conclusion they drew was that, while the climate is optimal in these areas, the habitat has been so altered by farming that the species have gone. Of more concern was the apparent absence of the species from suita- ble habitat. This could be because the species are trap shy or because other unknown factors are causing their decline. From past experience there may not be much time to redress the decline to extinction of yet more species of bandicoot in Australia. This page intentionally left blank Chapter 6

Pygmy possums and sugar gliders: pollen eaters and sap suckers Feathertail glider on one cent coin (Permission of the Royal Australian Mint). Pygmy possums and sugar gliders: pollen eaters and sap suckers 185

arsupials that feed on plants fall into three groups, based on body size and what part of the plant they eat. First, the smallest species, up to 0.7 kg, use plant exudates and nectar Mfor their energy needs, and fungi, pollen or small invertebrates for protein; the second, from 0.5 kg to 15 kg, live on the leaves of forest trees; the third, are the ground-dwelling browsers and grazers, which range from 1 kg rat kangaroos, 50 kg wombats and 80 kg kangaroos. In this chapter we will consider the first group, which includes three families of tiny possums, ranging in size from the 7 g honey possum or noolbenger, Tarsipes rostratus, to the 40 g mountain pygmy possum or burramys, Burramys parvus, and the gliding possums and their relatives, the Petauridae, up to 700 g (see Fig. 6.1, Plates 10 and 11). None eats leaves: they depend on sap and other products of several species of Eucalyptus, Acacia, Protea and Banksia, and the insects that live on them (Fig. 6.2). Plant sap is rich in soluble sugars but contains little protein. A variety of insects, such as cicadas, scales and lerps, aphids and leafhoppers, exploit this resource by pierc- ing the plant ducts that contain the sap. In order to extract the limited protein in the sap they imbibe large quantities and excrete the surplus sugars as honeydew. After the insect has left, sap may flow from the pierced ducts and when dry this forms a white encrustation of sugars, called manna. Small marsupials feed on honeydew, manna and on the fungus that grows on the sugar. The larger species can also incise the tree and imbibe the sap directly. However, when wounded, Eucalyptus trees produce an inedible substance called kino, so that the possums and gliders have to make new cuts to get fresh sap. Acacia, however, when incised, produces gums, which contain complex carbohydrates and some protein, and these are the staple diet of several species. In addition to these products, the flowers of these trees and other flowering shrubs produce nectar, which is a rich source of simple sugars; and pollen, which is a source of protein for those species that can rupture the pollen coat and get the contents inside. Some species obtain their

8 Foliage

7

6

5 Sap, gums, nectar, pollen, 4 Insects, nectar, pollen insects

3

Number of species 2

1

10 100 1000 10000 Body mass (g)

Figure 6.2: Plant food eaten by marsupials of different body mass. After Smith (1982b). 186 Life of Marsupials

protein nitrogen by eating beetles, moths and other insects, which they catch under the bark or in the canopy of the trees, or larvae, which they extract from the wood itself. This wide variety of carbohydrates, all derived ultimately from plant sap, and wide variety of proteins from fungi, pollen and insects supports 17 species of small diprotodontid marsupials in Australia and New Guinea. Some species specialise on nectar and some on sap or gum but most species can use whichever resource is abundant at any particular time of the year. These 17 species belong to four distinct families, with three having fossil lineages going back to the late Oligocene epoch, 26 million years ago (Fig. 6.3) (Brammall and Archer 1999). The fourth family, Tarsipedidae, contains a single species, the honey possum, of which the oldest fossils are only 2 million years old, but its very unusual anatomy, chromosome number of 24 (see Table 1.4) and distant relationship to the other three families makes it certain that it has a much longer separate ancestry than the fossil record implies.

Chapter

Ailurops 10 Phalanger 7 4 Trichosurus 2 3 Wyuldia 2 Buramys 2 4 Cercartetus Acrobates

Distoechurus 6 2336 Petaurus Gymnobelideus 24 Tarsipes 7 22 Pseudocheirus 3 4 5 7 2 Petauroides Hemibelideus 2 Vombatus 8 22 Lasiorhinus 8 3 322 Phascolarctos 7 26 23 16.3 10.4 5.2 2 0.01 0 million years ago

Figure 6.3: On the left the relationships of living genera of possums, gliders, wombats and koalas (see Chapters 6–8), based on DNA hybridisation criteria in Figure 1.10. These have been matched, on the right, to their respective fossil records, according to Archer et al (1999). Note the numbers above the horizontal bars refer to species (if more than one) known from that time period. Pygmy possums and sugar gliders: pollen eaters and sap suckers 187

The long separate lineages of the four families are also supported by comparisons of their proteins (Baverstock et al 1990) and DNA sequences and DNA/DNA hybridisation (Edwards and Westerman 1992, 1995). Both kinds of evidence suggest that all four families arose in the Oligocene after Australia became separated from Antarctica and the vegetation changed from rainforest to eucalypt woodland and heath associations. This change in vegetation provided new kinds of food plants and very different environments for marsupial herbivores. The similarities seen today in their life forms, physiology and behaviour reflect independent adaptations to the particular environment and food resources that they exploit. Three of these families include some of the smallest of all marsupials (7–10 g), which show special adaptations for small size, notably torpor and very prolonged pregnancies. The best- studied representatives are the honey possum (Tarsipedidae), the feathertail glider, Acrobates pygmaeus (Acrobatidae), and the eastern pygmy possum, Cercartetus nanus, and mountain pygmy possum, Burramys (Burramyidae): each displays a different pattern of thermoregula- tion and reproduction, again reflecting their separate origins. The fourth family, the Petauridae, includes the larger species, which weigh between 70 g and 700 g, the best-studied species being the sugar glider, Petaurus breviceps.

The tiny nectar and pollen specialists The Tarsipedidae The honey possum or noolbenger is one of the smallest of all possums and almost as small as the smallest dasyurids, Ningaui and Planigale (Fig. 6.1a, Plate 10). Adult males weigh 7–9 g and adult females 10–15 g, the females being always dominant to the males. Very little was known about them until 1977, because they are normally active at night and were difficult to catch. In that year biologists learnt from farmers in southwest Western Australia that they were finding tiny marsupials at the bottom of holes dug for fence posts. Since then artificial postholes lined with plastic pipe have enabled much to be learnt about this very active and engaging animal. One thing this discovery showed was that honey possums travel across the ground at night in search of the flowers that provide all their food. Formerly widespread through the heath lands of south-western Australia (Fig. 6.4), the honey possum is now restricted to the southwest tip of Western Australia, where it subsists exclusively on nectar and pollen of several species of flowering shrubs, especially Banksia (Proteaceae). It has a rudimentary set of teeth that are inadequate for eating insects but it has a long slender snout and highly specialised tongue and palate with which it collects nectar and pollen (Fig. 6.5). The tongue has a brush-like surface at the tip and is stiffened by a keratinous keel. The pollen is scraped from the upper surface of the tongue by transverse combs on the roof of the mouth and conveyed to the stomach. The simple stomach has an accessory pouch, which may be analogous to the crop of birds or cheek pouches of hamsters, serving to store nectar collected during the night (Richardson et al 1986). In some mammals and birds that feed on pollen a high sugar content in the stomach induces the pollen to extrude its pollen tube, as it would do on the stigma of a flower, and so gives access to its contents. The stomach pouch of the honey possum may similarly be a site for sugar storage, since pollen grains with emerging pollen tubes have been found in the stomach (Turner 1984b), although others are still intact when they leave the stomach. It is still unclear how honey possums get the contents out of the pollen grains that leave the stomach intact, although 12 h after inges- tion all pollen grains appearing in the faeces are empty. The Banksia species they favour have pollen grains with a large pore with a thin covering, through which the pollen tube grows, and presumably this is breached in the small intestine. 188 Life of Marsupials

Figure 6.4: Distribution in Australia and New Guinea of pygmy possums and petaurid gliders. Data from Flannery (1995) and Strahan (1995). Pygmy possums and sugar gliders: pollen eaters and sap suckers 189

(a) (b)

Figure 6.5: Scanning electron micrographs of the honey possum to show: (a) the brush-like surface of the tongue tip; and (b) the combs on the palate that scrape off the pollen collected on the tongue. Original electron micrographs from RD Wooller.

Nectar is 25% sugar and provides the honey possum with its water and some of its energy needs and the pollen contents, which are 20% protein and 37% carbohydrate, provide all its protein needs and additional energy (Wooller et al 1984, Bradshaw and Bradshaw 1999, 2001). And its energy needs are considerable. The standard metabolic rate (SMR) of the honey possum is the highest of any marsupial by a large margin, being twice the average marsupial rate and 158% of the average placental rate (Withers et al 1990). The field metabolic rate (FMR) of males and non-lactating females in winter and summer, measured by doubly labelled water (see Box 1.1), was found to be 2.7 times this (Nagy et al 1995, Bradshaw and Bradshaw 1999), or an energy requirement of 29–33 kJ per day. Other estimates for lactating females are double this figure (Turner 1984b). A Banksia inflo- rescence can produce nectar equal to 2 kJ per day and, depending on the site, this translates per square metre of heath to 0.05–5 kJ. So, to meet its daily energy requirements, a honey possum must feed from at least 14–17 inflorescences in an area of 7–700 m2 (0.7 ha). It is, therefore, not surprising that honey possums fall into postholes in their busy search for food to drive their metabolism and reproduction. When Bradshaw and Bradshaw (2002) attached minute radio transmitters (0.9 g) to honey possums they found significant differences in the short-term movements and habitat use between males and females. The females had small home ranges of 0.14 ha, centred on the prime food sources, whereas the males spent the day up to 200 m away from prime food sources and had an overall home range of 0.8 ha. The females might actively exclude the smaller males from the prime habitat, where flowering shrubs are common and shelter is more secure from predators. During the daytime honey possums rest in 190 Life of Marsupials

the tangled heads of grass trees, Xanthorrhoea, and they feed mainly at night: in summer their main activity is around dusk and dawn, with no activity between 0900 h and 1700 h; in winter the peak periods are closer together and some animals remain active throughout the day (Arrese and Runham 2002). The unique features of the southwest tip of Australia that enables honey possums to survive are the maritime climate and mild winters: the numerous species of woody shrubs flower at different times throughout the year, thus providing year round food for the honey possums at night and nectar-feeding birds during the day (Fig. 6.6). Indeed, nectar-feeding birds cannot visit some of the Banksia species because the inflorescences are hidden within the thickets and it is thought that these species have co-evolved with the honey possum as their main pollinator (Wooller et al 1983).

Beaufortia anisandra

Banksia nutans

Banksia quercifolia

Dryandra formosa

Banksia gardneri

Banksia coccinea

Banksia attenuata

Banksia grandis

Adenanthos cuneatus

100

80

60

40

20 with pouch young pouch with Percentage females

0 125

100

75

50

25 Number per 100 trapnights

0 ND JF MAM J J ASOND JF MAM J J ASO ND J 1977 1978 1979 1980 Figure 6.6: Honey possums and their food. The abundance of pollen of nine species of Proteaceae collected at different times of the year by foraging honey possums, and how this may influence the three peaks of breeding and the total size of the population. After Wooller et al (1981, 1984). Pygmy possums and sugar gliders: pollen eaters and sap suckers 191

Honey possums have two important adaptations for living in this special environment, torpor and embryonic diapause.

Torpor conserves energy when nectar is scarce Torpor in the honey possum is different from that seen in the dunnart (see Chapter 4), although it is probably provoked also by a shortage of food. The honey possum does not show a daily fluctuation in its body temperature but when it enters torpor its temperature falls to 5°C, much lower than in the dunnart, and its oxygen consumption and energy needs are correspondingly low (Collins et al 1987). When honey possums were held in metabolism chambers with plenty of nectar and the ambient temperatures and day length were kept similar to their natural environ- ment, their metabolism remained at the normal level throughout the day and night. However, if nectar was withheld for one night the animals became torpid within a few hours and remained so until the next day when they spontaneously recovered normal metabolism. When FMRs were measured in free-ranging honey possums they varied widely (Nagy et al 1995, Bradshaw and Bradshaw 1999), which suggests that honey possums may enter torpor and reduce their FMR whenever they are unable to find sufficient nectar in a night to maintain their energy needs. This is more likely to occur in smaller honey possums and is commonly observed in those that have spent a night in a pit trap without nectar. However, if they are unable to find nectar for several consecutive days, their chances of survival diminish. Natural mortality is particularly heavy during the summer, when the number of flowering shrubs is lowest.

Embryonic diapause spaces out the reproductive effort The honey possum has a reproductive pattern that is highly adapted to its unique habitat. Females give birth to one to four minute young that weigh only 4 mg, and each attaches to one of the four teats in the pouch. Survival of singletons is very low, presumably because one young cannot exert sufficient sucking stimulus to maintain lactation, and four young seldom survive to independ- ence: two is the commonest number of young in the pouch (Wooller et al 1981). Immediately after giving birth the female mates with several attendant males and conceives again. These new embryos develop to the blastocyst stage but unlike in kangaroos, they continue to grow slowly in the uterus for the next three months. This is a form of embryonic diapause seen also in badgers, Meles meles, and roe deer, Capreolus capreolus, among placental mammals. The mating system of the honey possum is also unusual. Males are smaller than females and there is strong competition between them for oestrous females, which is reflected in two remark- able attributes. They have the largest sperm of any mammal, and they have relatively enormous testes: together with the epididymides they are 5.4% of the body mass. Were men to have testes of this proportion, each would weigh 2 kg. After birth the young grow and develop in the pouch for two months by when they are furred and weigh 2.5 g each. They leave the pouch at this stage and are carried on the mother’s back for another month, by which time they weigh 4–5 g. If she has raised three young, they now weigh as much as she does and, as with all marsupials, this last month of lactation is the most demanding period in terms of energy and protein requirements for the young. Turner (1984b) calculated that a female honey possum at this period would need to visit 2400 florets in at least eight Banksia inflorescences each night to obtain sufficient protein for her young. It is perhaps not surprising that few females raise more than two young of the litter to independence. However, young of the first litter of the year attain sexual maturity at 7–9 months and these females contribute to the main growth of the population in the spring (August–September). During the last month of lactation the next litter of embryos in the uterus enter the final brief fetal stage of development, although this has rarely been seen and never measured, and the minute young are born soon after the previous litter have weaned. It is still not clear how 192 Life of Marsupials

the period of embryonic diapause during the early stage of lactation is controlled. It is certainly different from that in kangaroos (see Chapter 2), since it is not shortened by experimental removal of the pouch young. Initiation of the final stage of pregnancy may be controlled by the availability of nectar and pollen. This varies through the year, with a few species flowering between February and April and most flowering between May and November, followed by a dearth of nectar and pollen in the summer. The honey possum population reflects this pattern, as judged by the number of possums caught per trap night through the year (Fig. 6.6). The lowest numbers occur between December and March and the peak occurs in August. In the main study on the species by Wooller et al (1981, 1984), half the females caught in February–March had pouch young and there were two subsequent peaks in May–July and September–October. The young born at these times make their greatest demands on their mothers one month later, respectively in April, August and November, which coincide with the successive flowering peaks of Banksia nutans, Banksia coccinea and Banksia attenuata, respectively (Wooller et al 1983). The interval between each peak is about the duration of embryonic diapause and it may be that the final stage of pregnancy is initiated by the nutritional condition of the female at the time. What was not clear until recently was whether the females carry dormant embryos through the long summer dearth of flowering shrubs until February: from the observations of Bradshaw et al. (2000) on three females held in open cages, it seems that they do. Males were removed from these females while they were suckling litters in October and their young were weaned in November. None had a litter during the summer months but they gave birth on 18, 19 and 25 February of the next year. Two other females kept with males also gave birth on 10 and 18 February, which seems to indicate that the first births of the year are synchronised by a photope- riod signal and the later less synchronised peaks reflect prevailing food resources. Since honey possums probably do not live for much longer than one year the three sets of offspring they can potentially produce in that year is, for most, their total lifetime production. So close is the relationship between the year round flowering of the heath plants and its life cycle that it is unlikely that honey possums ever lived in any other habitat. Certainly no fossils have been discovered anywhere other than in southwest Western Australia.

The Acrobatidae and Burramyidae Both species of the family Acrobatidae and the five species in the family Burramyidae have the primitive chromosome number of 14 (see Table 1.4) and, because of other similarities in their biology, they will be treated together here. Those species that have been studied in any detail display true hibernation, as distinct from the transient torpor of the honey possum and the small dasyurids, and all except one species have a reproductive pattern similar to the honey possum. The Acrobatidae includes the tiny (10–14 g) feathertail glider (Fig. 6.1b, Plate 10), which is the eastern equivalent of the honey possum, living in forests and woodlands of eastern Australia, and the larger (40–50 g) feathertail possum, Distoechurus pennatus, which lives in the forests of New Guinea (Fig. 6.4). Both are distinguished by the unusual laterally flattened arrangement of hairs along the tail, which make it look like a feather, but only the Australian species can glide. Much more is known about this species than about the New Guinea species. As with the honey possum, the opportunity to study this tiny marsupial also came about by the chance discovery that feath- ertail gliders choose to make their communal nests in telephone junction boxes (Fanning 1980, Fleming and Frey 1984). While this is inconvenient for telephone maintenance staff, the gliders will readily use other boxes of similar design that are put out especially for them. The feathertail glider is the smallest gliding possum and one of the three smallest gliding mammals worldwide (as distinct from bats, which are true fliers). Its gliding membrane, or Pygmy possums and sugar gliders: pollen eaters and sap suckers 193

femoralis

platysma Feathertail glider semi- tendinosus

forearm

ankle humerodorsalis

Sugar glider

fifth finger tibiocarpalis first toe

Greater glider

elbow

Figure 6.7: Musculature under the gliding membrane or patagium of the feathertail glider, Acrobates pygmaeus, the sugar glider, Petaurus breviceps, and the greater glider, Petauroides volans. Note the different points of attachment of the patagium to the limbs in each species and the different arrangement of the underlying muscles. After Johnson-Murray (1987). 194 Life of Marsupials

patagium, extends from the elbow of the forearm to the knee of the hind leg (Fig. 6.7a). Within the patagium the arrangement of the muscle components are different from other marsupial gliders and the limbs are not elongated to support it, which suggests that the gliding mode has probably evolved independently in this species (Johnson-Murray 1987). The ends of all the feathertail glider’s toes are expanded into pads that aid it in climbing the smooth bark of eucalypt trees (Fig. 6.8b). Because of its very small size it can exploit the forces of surface tension to counter gravity: the toe pads are deeply ridged and moistened by sweat glands

Figure 6.8: Adaptations of the hands of small possums: (a) Left hand and foot of the western pygmy possum, Cercartetus concinnus, to show the special pads on the digits and soles (after Jones 1924); (b) scanning electron micrograph (courtesy of H. Rosenberg) of the thumb pad of the feathertail glider, Acrobates pygmaeus, to show the fine ridges and grooves, and a section of four ridges of the same to show the sweat glands at the base of each groove (@) that provide the fluid adhesive. Pygmy possums and sugar gliders: pollen eaters and sap suckers 195

that open at the base of each groove, which provides such good adhesion to smooth surfaces that they can walk up even a vertical pane of glass (Rosenberg and Rose 1999). Similar structures are found on the toes of the other small possums (Fig. 6.8a), although they are not as well developed as in the feathertail glider. The family Burramyidae comprises five species of small, non-gliding possums (Fig. 6.4). There are four species in the genus Cercartetus, which weigh between 7 g and 24 g and live in the wet sclerophyll forests around the coast of Australia and in Tasmania: the best-known species is the eastern pygmy possum. One tropical species, the long-tailed pygmy possum, Cercartetus caudatus, lives in the central montane forests of New Guinea and in the rainforests of Cape York. The fifth species, the largest in the family, is the rare burramys (40–80 g), which lives in the Snowy Mountains of eastern Australia (Fig. 6.1c, Plate 11). It has a fossil record going back to the Oligocene but, until 1966, was thought to be extinct. When the skull and teeth of this species were first discovered in 1894, in a cave deposit in New South Wales, it was thought to represent the ancestral stock from which kangaroos had evolved, because of its large blade-like premolars that resemble the same teeth of rat kangaroos. However, when Ride (1956) re-examined the fossil material, he concluded that it was most closely related to the pygmy possums and sugar gliders. Then, in August 1966, a small, torpid, possum was found in a wood heap in a ski hut at Mt Hotham in the Victorian Alps. This caused great excitement when it was realised that it was the controversial fossil, Burramys parvus, alive! It is very rare for a palaeontologist’s conclusions to be put to the ultimate test of resurrection, but David Ride’s conclusions were fully vindicated. Not only is the burramys remarkable for the manner of its discovery but also because its only known habitat is in boulder banks above the tree line, which are covered with snow for seven months of the year. This poses special problems for a small mammal that cannot rely on food stores or body reserves to take it through such a long period. The seven species in these two families differ in their ancestry and life patterns but all live on plant exudates or, in the case of burramys, on seeds and berries; and pollen or insects provide their protein nitrogen. Except for the two New Guinea species, they occupy regions of Australia and Tasmania that experience cool winters and periods of inclement weather.

Diet and daily activity Although nectar and pollen, mainly from eucalypts, is a major part of its diet, the feathertail glider is not such an extreme specialist as is the honey possum. Like the honey possum, the upper surface of its tongue is covered in a weft of fine papillae, which presumably helps it to lick up nectar from Banksia and Eucalyptus flowers that are its main food source (Turner 1984a). It also feeds on manna and honeydew and, while it cannot incise the bark of eucalypts for sap, it will feed from sap sites made by larger gliders. Unlike the honey possum, it has a full dentition and forages beneath the bark of eucalypts for insects, especially moths. Fungus also contributes to its protein needs. Its simple stomach and gut reflects its diet of highly digestible carbohydrates, pollen and insect protein. Thus, it has a wider dietary repertoire than the honey possum and can switch between food sources through the changing seasons of the year. In its preferred habitat of mixed woodland and deep forest, food resources do not appear to be limiting, except for the additional needs of lactation. Nevertheless, being about the same size as the honey possum, it faces the same constraints of size and must feed each night to remain active. Feathertail gliders emerge from their daytime den within one hour after dark and remain active most of the night. In one study that tracked four feathertail gliders with attached radio transmitters (Kirk et al 2000), their home ranges varied from 0.5 ha for a lactating female to 2 ha for one of the males, and the minimum distances travelled in one night were between 125 m and 220 m (Table 6.1). 196 Life of Marsupials olivores Fleming and Frey (1984), Kirk et al (2000) and McKay (1989) (1990), Turner Ward Smith (1984a) References Wooller et al (1983), Russell and Renfree Wooller (1989), Bradshaw and (2002) Ree and Bennett (2003) (1995) and Kavanagh (1990) Comport et al (1996) (1984), Kerle (1984, 2001) (1990) 20 Henry (1984), Kehl and Borsboom Feeding time (% active) 22 Thomson and Owen (1964) 155.6 Martin and Handasyde (1999), Mitchell Dwiyahreni et al (1999) & , ( Distance per night (m) <50 125 200–500 551025 60 Suckling (1984), Jackson (2000b) Quin (1995), van der Ree (2002), 125–220 500 50–75 Atramentowicz (1982), Julien-Laferriere 590–3420 60590–2350 90 Jackson (2000b) Craig (1985), Goldingay (1992), 200 210–500 15 Dunnet (1964), Crawley (1973), Green Home range of group (ha) 0.14–0.8 0.3–1.0 1.0–3.3 Biomass (g/ha) Density (No./ha) 0.12 1.2kg Body mass & (g) 13 900 9 ( (g) 26120 29120 160213 14–20 110 1.6–2.9 380–540 192 192–479 0.2–6.1 0.5 1.3–1.9 17–885 0.9–1.5 0.5–4.0 171–328 3.0–3.5 15 15 19 285400 0.4–2.1 250 0.5–1.4 125–560 370620 340 530 0.2–0.3 0.1–0.2 63–100 21–122 18–20 30–85 0.7–1.1kg 0.7–1.1kg 0.3–4.02.5 kg 2.8–3.3kg10 kg 2.5 kg 1.3–2.6 1–10 7 kg 2.5–25 kg 2–4 1.0–7.4 20–40 kg 2.0–4.0 10 kg Tarsipes rostratus Species Body mass Cercartetus nanus Gymnobelideus leadbeateri Petaurus breviceps Petaurus norfolcensis Acrobates pygmaeus Caluromys philander Petaurus gracilis Petaurus australis Pseudocheirus peregrinus Petauroides volans Trichosurus vulpecula Phascolarctos cinereus Ailurops ursinus Table 6.1: Relationship between body mass and home range, distance travelled, per cent of time feeding for marsupial exudivores f Table Pygmy possums and sugar gliders: pollen eaters and sap suckers 197

The eastern pygmy possum has a similar distribution to the feathertail glider and lives on the same variety of food, switching from one species of plant to another as nectar, pollen, manna, seeds, fruit or insects become abundant. Where the two species occur together, they seem to make the same choices (Huang et al 1987). It also has a tongue with a brush-like surface for collecting nectar and pollen, and its activity patterns and home range are similar to those of the feathertail glider (Table 6.1). Unlike feathertail gliders, eastern pygmy possums fatten in autumn and lose weight through the winter. Burramys is different, for it must live on the seasonally limited food available in its alpine habitat. This includes seeds, which it cracks with its large premolar teeth, and the berries of small shrubs, particularly the alpine podocarp and several species of heath that are most abundant at the end of the summer. However, its main food resource is a single species of insect, the bogong moth, Agrotis infusa, which breeds on the western plains of New South Wales and southern Queensland. The adults are carried in vast numbers each year on high altitude jet streams to the alps. There they spend the summer, from October to February, torpid in crannies in the boulder banks. The moths are very fat when they arrive and they provide an abundant source of protein and energy-rich food for burramys. Between January and March, when the bogong moths depart, the adult burramys double their body weight from about 40 g to over 80 g (Geiser et al 1990). From May until October the land is covered with deep snow and burramys hibernate in nests in the boulder banks, while their body mass steadily declines as they live on their stored fat. In captive animals the pre-winter fat depot was sufficient for the whole winter without additional food. This cycle of summer fat deposition and winter use resembles the pattern seen in small arctic mammals that undergo long hibernation and in them it is known to be under hormonal control, regulated by the changing length of day. However, burramys that have been bred in captivity with regularly available food, do not put on weight in autumn or lose it during the winter, which suggests that it is not so tightly controlled by day length as in arctic species. The use of seeds and berries by burramys is very unusual for a marsupial, seeds being gener- ally taken by native rodents in Australia. Smith and Broome (1992) suggested that burramys is the last representative of a group of seed-eating marsupials that were displaced when rodents entered Australia in the Pliocene epoch; and that it can only compete with rodents for this resource in the high alpine areas, because of its ability to hibernate, which the rodents cannot do.

Behaviour and energy metabolism Feathertail gliders spend the daylight hours in a nest, which they construct from leaves and other dry material, including old birds’ nests (Fanning 1980). They do not appear to have a well-developed social structure, sometimes sharing a nest with other gliders or frequently living alone. In the cooler months up to eight feathertail gliders may huddle together in a nest. The composition of these groups is predominantly female, with only one adult male and weaned young of a recent litter and subadults of the previous year (Fleming and Frey 1984, Fleming 1985). Huddling together conserves heat and such gliders are tolerant of each other, especially in the early part of the breeding season at the end of winter. The eastern pygmy possum is also generally solitary, finding a site in a decayed log to rest in during the day, and in the colder periods of the year huddling together in larger groups, especially during bouts of hibernation. Burramys use their limited habitat in a different way. Females preferentially live on bogong moths in the boulder banks, which provide essential fat and protein for lactation, while the males live below them in the heaths, feeding on seeds and berries. Movement between the boulder banks and heath land is therefore essential for the survival of the population (Smith and Broome 1992). 198 Life of Marsupials

Hibernation as an energy conservation strategy While the honey possum enters brief spells of torpor if it fails to obtain nectar, it does not face long periods without food during prolonged cool weather. By contrast, these three small gliders and possums live in habitats where high-energy food is not available in all seasons and where winter temperatures may be very low. Their strategy is to undergo very prolonged periods of torpor lasting several days or weeks. This condition is called hibernation. It occurs most frequently in the winter months and may be preceded by the animal laying down substantial stores of fat, so that it can survive without additional food for several months. How does hiber- nation differ from torpor? The important features are that body temperature can fall to a very low level for a longer duration in each episode, and the species has the ability to arouse spontane- ously to an active state. An eastern pygmy possum when torpid is curled up in a tight ball with the long tail coiled on one side. The large ears are limp and folded, and the animal feels cold to the touch. The body temperature is only a few degrees above the ambient temperature, the oxygen consump- tion is reduced to less than 2% of the active state and respiration is correspondingly reduced (Geiser 1993). The heart rate becomes very slow during torpor, especially at low temperatures (Bartholomew and Hudson 1962). For instance, an active eastern pygmy possum at 30°C has a heart rate of 300 beats per minute. This rises to 650 beats per minute for an active animal at an ambient temperature of 5°C, whereas the torpid animal at 5°C has a heart rate of only 60–80 beats per minute, or about one-tenth the active rate. Despite this very slow heart rate, the electro- cardiogram shows a normal pattern without irregularity or missed beats. The ability of the heart of a torpid animal to function normally, albeit very slowly, is one of the most important adaptations for hibernation and distinguishes natural hibernators from non-hibernators exposed to very low temperatures, as we saw in the Virginia opossum, Didelphis virginiana, in the northern winter (see Chapter 3). In eastern pygmy possums the duration of hibernating bouts is inversely proportional to the ambient temperature, so that at 5°C it can remain torpid for up to five weeks, which is among the longest known for any hibernating mammal. Likewise burramys can be torpid for up to three weeks, during the seven months hiber- nation period (Geiser 1993) (Table 6.2). The reason the individual bouts do not last throughout the winter in burramys, or other hibernators, is because the animal must arouse occasionally to discharge the metabolic products that accumulate during torpor, and for this the kidneys must function at normal body temperature.

Table 6.2: Aspects of hibernation and torpor in small possums and gliders, and in one dasyurid After Geiser (1993, table 1, p. 73), Morton (1978a), Holloway and Geiser (1996).

Species Minimum Tb Duration (days) O2 consumption (%SMR) Sminthopsis crassicaudata 18.0 0.2 12.5 Tarsipes rostratus 5.4 0.4 5.2 Acrobates pygmaeus 2.0 5.0 1.0 Cercartetus nanus 1.3 23.0 1.6–3.8 Burramys parvus 2.4 14.0 3.0 Petaurus breviceps 15.5 0.4 –

Feathertail gliders also save energy by hibernating. When active their body temperature varies from 33°C to 38.5°C but in winter, or when food is scarce, body temperature will fall below 30°C and the animals become torpid (Fig. 6.9). When a whole group becomes torpid in a Pygmy possums and sugar gliders: pollen eaters and sap suckers 199

nest their body temperatures are only a degree or two above the ambient temperature down to 2°C and can remain at this very low temperature for up to five days (Table 6.2). During this time the metabolic rate falls to 1% of their SMR, and the respiration rate also declines with up to 12 minutes between breaths (Jones and Geiser 1992). This physiological response provides a very great saving in energy when food is limited or when foraging would be energetically too costly because of low ambient temperatures.

40 (a) active

30

20

torpid Body temperature (C) 10

0 (b)

6 )

active 4

2 espiration (ml oxygen/g/h R

torpid

0 010203040

Ambient temperature (C) Figure 6.9: How torpor conserves energy for the feathertail glider, Acrobates pygmaeus. The relationship between ambient temperature and: (a) body temperature; (b) respiration in active feathertail gliders (6) and in those undergoing torpor (Q). After Fleming (1985). 200 Life of Marsupials

The real danger for an animal in torpor is that it may freeze to death if the ambient tempera- ture falls below 0°C. However, an important adaptation of hibernators is the ability to arouse spontaneously if the body temperature falls below a critical low point. For the eastern pygmy- possum the critical temperature is 1.3°C and for feathertail gliders and burramys it is about 2.0°C. It is still not clear how arousal is achieved when the ambient temperature falls below the critical point. At first disturbance the hibernating eastern pygmy possum utters a faint hiss, which indicates that some sensory perception is functional. During arousal there is an activa- tion of heat-producing mechanisms with a concomitant rise in oxygen consumption and body temperature (Fig. 6.10). The rate at which the body temperature can rise varies from 0.33°C to 1°C per minute, which is comparable to the rates during arousal of placental hibernators. However, marsupial hibernators differ in one respect. In placental species the anterior end of the body warms up faster than the hind end, due to the release of heat by a special tissue called brown fat that lies across the shoulders and around the heart. Such discrete depots of brown fat are not found in marsupials (Hayward and Lisson 1992), and in eastern pygmy possums both ends of the body warm up at the same rate. There is some evidence that tissue like brown fat may occur diffusely throughout the body of marsupials (Loudon et al 1985), which could provide the necessary heat during arousal, but the matter is not yet resolved.

Pygmy possum

4.0 TAir = 18°C TAir = 9°C

2.0

3.0

Respiration (ml oxygen/g/h) 1.0

0 0 40 80 120 160 0 40 80 120 160

Time after entering torpor (min) Time after start of arousal (min)

Figure 6.10: Oxygen consumption during hibernation and arousal in the eastern pygmy possum, Cercartetus nanus, when the ambient temperature was held constant at 9°C (O) and at 18°C ({). Body temperature at the start of arousal was 12°C and 20°C, respectively, and reached 33°C in both cases at peak oxygen consumption. After Bartholomew and Hudson (1962).

The capacity for energy conservation of these three hibernators is, thus, considerably greater than that of the honey possum, both because of the lower temperatures to which the body can fall and because of the longer duration of hibernating episodes. Nevertheless, the differences between all four species suggest that torpor, as an adaptation for energy conservation, has evolved independently in these three families of tiny marsupials, as it probably did in the four species in the southern cone of South America (see Chapter 3). Pygmy possums and sugar gliders: pollen eaters and sap suckers 201

Fitting reproduction into the annual seasonal cycle How is reproduction accommodated with winter hibernation and fat deposition? Each species does it in a different way. There is no difference in the size of adult male and female feathertail gliders and there is no evidence for long-term bonding of either monogamous pairs or male-dominated breeding groups. The mating system is, thus, polygamous, as it probably also is in the honey possum, although the testes of feathertail males at their maximum weigh only 1.5% of the adult body weight, much less than those of the honey possum. In the southern part of its range, in Victoria, the species is a seasonal breeder (Ward and Renfree 1988a,b). In males the testes enlarge in April and reach their maximum size in June, at which time the prostate gland is also maximally enlarged. Both organs remain large until January, when males cease to produce sperm. Females come into oestrus in June–July and the first births occur in August, immediately followed by another oestrus. There are four teats in the pouch and up to four young are born but by the time of weaning at 100 days the usual litter size is two. The young remain in the pouch for 65 days, are suckled for another 35 days in a nest until their eyes open and remain in the parental social group after weaning. The embryos conceived after the birth of the first litter grow slowly to become small vesicles of about 2000 cells during lactation of the first litter. Completion of embryo development has not been observed but birth of the second litter occurs over an extended period between October and January. If this occurs early, the female may mate again and a third litter is possible, but the chance of this declines as the season draws to a close, either because ovulation does not occur or because the males are by then infertile. By contrast with the honey possum, reproduction in the feathertail glider is more conservative, with a well-defined breeding season that begins in mid winter and ends at mid summer: fewer offspring are produced in a year, but life expectancy at two to three years is longer. The pattern of reproduction in the eastern pygmy possum is similar to the honey possum and feathertail glider, except for some differences in emphasis. Males and non-breeding females are the same size, while lactating females are significantly heavier. Once they have attained sexual maturity the testes of males remain large and, like in honey possums, they are capable of fertilis- ing females in all seasons. Likewise, the females breed in all months, producing litters at intervals of about three months. Unlike the feathertail glider female, however, the number of eggs shed at each ovulation is greater than the six teats in the pouch, although the initial litter is seldom more than four and this progressively reduces through lactation to two. The critical factor in the successful rearing of a litter is the availability of pollen of Eucalyptus and Banksia species during the later stages of lactation. These pollens contain a suitable balance of amino acids for pygmy possums (van Tets and Hulbert 1999) and it was shown by Turner (1984b) that they are capable of extracting the contents of the pollen grains during passage through the gut. As in the honey possum and the feathertail glider, eastern pygmy possum females mate soon after giving birth and development of these embryos proceeds slowly until the end of the previ- ous lactation. This pattern of embryonic diapause probably also occurs in the other three species of Cercartetus, including the New Guinea species, and in the feathertail possum of New Guinea. Curiously, in none of these small possums and gliders that display embryonic diapause has the final stage of fetal development been observed, so at present it is not known what stimulates the embryonic vesicle to proceed to the final stage of pregnancy and be born. The reproductive pattern of burramys is different and follows the pattern seen in most marsupials that have a very short gestation: there is no evidence for post partum oestrus, embry- onic diapause or a second litter after the first is weaned. This may be because of the special conditions of its alpine environment. Young are born in November after a gestation of 13–16 days and generally all four teats in the pouch are occupied. The young leave the pouch at 20–30 202 Life of Marsupials

days, before the eyes open, and then remain in a nest until weaned at 55–65 days. They reach adult weight at 90–120 days by March, before winter conditions set in. This extraordinarily rapid rate of growth is comparable to the gray short-tailed opossum, Monodelphis domesticus (see Chapter 3) and bandicoots, (see Chapter 5) and much faster than in any other diprotodont marsupial. This suggests that it is a special adaptation to the extreme conditions of the alpine habitat, where reproduction and fat deposition must be accomplished during the brief summer when bogong moths and seeds are abundant (Kerle 1984). Both adults and young must be well prepared for hibernation through the long winter snow cover. Indeed, without the annual pulse of bogong moths, burramys could not survive in its only present day habitat. Burramys was formerly more widespread through south-eastern Australia, during the last glacial period 18 000 years ago, when alpine conditions were presumably more extensive. This raises the question of the future prospects for a species that numbers only a few thousand individuals in scattered populations at the highest elevations in Australia. The predicted climate changes in the next 100 years could see the alpine environment disappear in Australia and the migratory habits of the bogong moths change, either of which would spell the extinction of this rare marsupial, which has already experienced one resurrection.

Small possums and gliders: conclusion The genera of Tarsipes, Acrobates, Cercartetus and Burramys are only distantly related to one another and yet they all undergo deep torpor and three of them display embryonic diapause. The ancestors of each genus almost certainly acquired both attributes independently, as adaptations by very small mammals living in seasonal climates with low winter temperatures. The patterns of torpor displayed range from brief but quite deep torpor in honey possums, to true hibernation of varying duration from five days in feathertail gliders to five weeks in eastern pygmy possums. In Burramys, particularly, the breeding season is fitted into the post hibernation part of the year. The pattern of embryonic diapause is very similar in the three genera in which it occurs and is distinct from the phenomenon in the kangaroos and rat kangaroos. Its duration seems to be controlled more by the immediate availability of food than by an external signal, such as changing day length or the sucking stimulus of the pouch young. This is shown by the longer interval between litters when resources are scarce and the shorter interval in summer when these are abundant. Burramys is the only species in the whole group that does not show embry- onic diapause, presumably because the very brief opportunities for breeding have not selected for it. Similar patterns of hibernation and embryonic diapause also occur in a variety of small placental mammals, representing several orders, which live in extreme environments. We may conclude that both phenomena are expressions of basic mammalian attributes that can be selected for in conditions that confront species of small size, be they placentals or marsupials.

Gum, sap and insect specialists The Family Petauridae includes five species of small gliding possums and four non-gliding species, ranging in weight from 70 g to 700 g. The non-gliding Leadbeaters’s possum, Gymnobelideus leadbeateri, is considered to most resemble the ancestral form, from which the gliding species evolved in Australia, and the non-gliding striped possum, Dactylopsila trivirgata, and trioks, Dactylopsila megalura and Dactylopsila palpator, evolved in New Guinea (Edwards and Westerman 1992, Flannery 1995). Like the bandicoots, the striped possum later spread south to York Peninsula and two species of glider later spread to New Guinea. Pygmy possums and sugar gliders: pollen eaters and sap suckers 203

Leadbeater’s fairy possum The story of Leadbeater’s possum, Gymnobelideus leadbeateri, has some parallels to burramys. During the last glacial period it occupied a much wider distribution than it does today, both in New South Wales and Victoria, but not Tasmania. The first two specimens were collected in Victoria in 1867 and named after John Leadbeater, who was the taxidermist at the Museum of Victoria. Nothing further from beaten lead can be imagined than this vibrant, feisty little animal, but the alternative ‘common’ name, fairy possum, belies its highly aggressive nature. Until there is a more appropriate name it seems it will continue to be called Leadbeater’s possum. Only three other specimens were collected by 1910 and none thereafter, so that by 1950 it was deemed to be extinct. It was rediscovered in 1960 in an entirely different habitat from the earlier records and has since been found to live in a very circumscribed area of mature ash, Eucalyptus regnans, forests of Central Victoria (Fig. 6.4). Like burramys its very limited distribution, habitat requirements and small surviving number of about 4000 animals mean that its future under a changed climatic regime is highly uncertain. Its preferred habitat is a particular successional stage of the forest that develops after a severe forest fire has converted many of the largest trees into dead standing stumps with many hollows, and a profuse understorey of wattle and other shrubs has grown up beneath them. In January 1939 immense fires engulfed the great ash forests of Central Victoria, causing the death of many of the mature trees. This opened the canopy and a thick regeneration of ash saplings and wattle, Acacia dealbata, grew up during the subsequent 40 years and now provides optimum habitat for Leadbeater’s possum. Leadbeater’s possums select nest sites in large, old trees or large dead stumps about 10–30 m above ground in cavities with entrance holes of 3–5 cm diameter. This size excludes larger compet- itors, such as mountain brushtail possums, Trichosurus caninus, and owls, Ninox sp., while smaller competitors like Antechinus can be displaced. In the deep hollow inside the trunk a substantial nest of 20–30 cm diameter is fashioned from shredded eucalypt bark, which provides excellent insula- tion. This nest is occupied by a social group of 2–8 animals and forms the centre of their relatively small home range of 1–2 ha (Table 6.1). Nest sites and their occupants are spaced out in the forest, so that there is little overlap between adjacent groups, although some fighting occurs along territo- rial boundaries or when a group loses some members (Smith 1984a). Leadbeater’s possums are about the same size as sugar gliders, a near relation, but their social organisation is very different. Unlike all other petaurids, adult females at 160 g are significantly heavier than adult males at 120 g, and an adult female is the dominant member of the group. Each group comprises a mated pair, which remain together for lengthy periods, as well as one or two other adult males, and subadult males and juveniles of the family. The dominant female defends the nest site with the aid of her male partner and she drives away other females and her own subadult daughters. This leads to a population that is male biased, and the opposite of the social structure of other petaurids. This unusual pattern ensures that the resources of the group territory are available for raising young. What are the resources of the home range?

Diet of Leadbeater’s possum For their energy needs Leadbeater’s possums take manna and probably honeydew from lerps, as well as incising the bark of Acacia for gum. They do not appear to incise the bark of eucalypts for sap as other gliders do, and there is little evidence for them taking nectar and pollen (Smith 1984b). During late winter and spring there is an absolute shortage of energy resources in the form of exudates, which is probably the limiting resource for them. Their diet also includes insects throughout the year, especially large tree crickets under bark. To a less extent, moths, small beetles and other insects are also taken but the crickets are the staple source of protein. 204 Life of Marsupials

Because insects are available throughout the year, protein resources are not limiting like gum is and so do not influence the time of breeding. Leadbeater’s possums are only active at night, emerging from their nest site at sunset and returning permanently at dawn. However, through the night they make up to 50 brief visits to the nest of a few minutes to one hour (Smith et al 1982). Outside the nest they are extremely active, foraging for food. The nest site is usually near the canopy of wattles and they begin the night’s activities by leaping into the canopy and then running along branches and leaping one metre or more to the next. Like the pygmy possums, broad pads on the ends of the digits of their feet enable them to make a firm grip on the smooth bark of eucalypt trunks. This extreme degree of activity is reflected in their FMR. Although their SMR is within the normal marsupial range for their body mass, the FMR is five times higher, which is exception- ally high for a possum (Smith et al 1982). The reason is that 73% of the daily energy budget is expended in activity and travel, while only 10% is for thermoregulation (Fig. 6.13). The low value for thermoregulation is due to the well-insulated nest and because the group members huddle together in it. However, they do not become torpid, as does the sugar glider. Because of the high FMR, water turnover is also high, despite the high humidity of the forests. Leadbeater’s possums obtain water from their food, from dew licked off the branches and as a product of oxidation of carbohydrates in the gum and manna that they eat.

Reproduction The reproductive pattern is similar to that of the brushtail possum (see Chapter 2), with a gesta- tion of less than 20 days. Embryonic diapause does not occur. Like the feathertail glider, however, the ovulation rate of up to 12 eggs much exceeds the four teats available in the pouch. The litter size, despite the high ovulation rate, is only one or two young, which remain in the pouch for 80–93 days and are then left in the family nest for up to a further 40 days, so they make their first forays from the nest at three to four months. Two litters are produced in one year, the first is born in May–June and weaned in late September, coincident with the new flow of acacia gum: the second litter is born in October–November and weaned in late January. Juveniles disperse from the natal nest at about one year. Because of the unusual social structure, dispersal is earlier and mortality higher in young females than in young males. Life expectancy after the first year for females is less than two years and somewhat longer for males – again a reversal of the normal pattern. This indicates the considerable burden that reproduction places on the female to defend a territory with sufficient resources to provision her young to independence. Leadbeater’s possums must also compete with sugar gliders, which are the same size and feed on the same resources. The distribution of sugar gliders completely surrounds the confined forest environment of the Leadbeater’s possum, so how does the latter species survive? It seems that the two species use different parts of the habitat, sugar gliders choosing nest trees with long narrow cracks in the main trunk, while Leadbeater’s possums prefer stumps of large old trees with many holes. Also, in the dense vegetation and abundant resources of the regenerating ash forests the Leadbeater’s possum’s greater agility may give it a competitive advantage where gliding is no advantage for extensive travel. However, in more open woodland, Leadbeater’s possum cannot compete for dispersed resources with the sugar glider’s greater mobility (Lindenmayer 1996).

Victorian mountain ash forests and the survival of Leadbeater’s possum The future for the Leadbeater’s possum epitomises the controversy between conservation and commerce in Australia’s old growth native forests. Thought for decades to be extinct, this extremely rare possum then reappeared in the ancient ash forests of the Central Highlands of Victoria. Since the beginning of European settlement these forests have been logged for high- grade timber and continue to be logged intensively, now mainly for pulpwood. After Andrew Pygmy possums and sugar gliders: pollen eaters and sap suckers 205

Smith’s ecological studies on the species (Smith 1984a) it seemed that its essential requirements in the successional stage of regenerating forest after wildfire could be readily accommodated into the harvesting regime for timber. However, Lindenmayer (1996) has shown that it is much more complex than this. Current logging practices require that all decayed and dead trees, called stags, be removed as well as the commercial stock and this denies the possums one of their two essential require- ments, hollows of a particular dimension in decaying old trees. The thriving populations in the mountain ash forests that were burnt in the great fires of January 1939 are there because an unusually large number of stags were formed after the death of large trees in those fires. It is now evident that a mountain ash tree provides shelter for Leadbeater’s possums only when it is more than 150 years old and beginning to decay (Smith and Lindenmayer 1988, Lindenmayer 1996). Thereafter the likelihood of the dead tree remaining upright declines over the next 50 years. This means that the forests burnt in 1939 that are now optimal habitat will progressively become less suitable as the stags decay and fall, as they are already beginning to do. In past centuries, presumably, Leadbeater’s possums would gradually move to another part of the forest that was at a different successional stage of the process and thrive there. Now, however, silvicultural practices that result in even-aged stands do not provide the range of hollows, so there will be steadily diminishing opportunities for the species to move elsewhere, unless suffi- cient middle-aged trees are left to decay. Since Leadbeater’s possums live in the most productive forests for timber and this cycle takes two or three centuries, there is an acute conflict of interest between ensuring the long-term survival of this unique possum and the immediate demands of the timber industry for the resource. The normal compromise of setting aside an area of forest as a perpetual reserve for the possum will not satisfy its requirements for a particular successional stage of the forest, unless the reserve is very large indeed. With the knowledge of the species requirements Lindenmayer and Possingham (1995) have attempted to model the minimum area of forest required to retain Leadbeater’s possums and other dependent wildlife into the long term. Their estimate is 200 000 ha. They further suggest that if the forests are to continue to be harvested for timber, far more rigorous restrictions on the length of the rotation cycle and on the pattern of logging must be implemented. Lindenmayer (1996) uses a quotation from a 1989 enquiry into the Victorian Timber Industry: ‘failure to conserve Leadbeater’s possum would be unlikely to engender public confidence in the govern- ment’s ability to protect other rare and endangered species’. Indeed.

Trioks and striped possums The other three non-gliding species of petaurids occur in New Guinea and adjacent islands, with one represented in far north Queensland (Fig. 6.4). These are the insectivorous striped possum, Dactylopsila trivirgata, and the trioks, D. megalura and D. palpator (Fig. 6.1d, Plate 11) (Flannery 1995). This genus is represented in the Oligocene (Fig. 6.3) and on molecular criteria the living species are most nearly related to Leadbeater’s possum. They are specialised to search and feed on adult insects by tearing away the bark of trees and by gouging out the wood to search for wood-boring insect larvae. They have two remarkable adaptations for this. Their skull is markedly different from that of other petaurids and indeed all other phalang- eroid marsupials. The incisor teeth protrude forwards from strong jaws, which are supported in a skull with a shortened snout, reinforced to withstand strong forces applied by the teeth when gouging wood. On their hands digit 4 is much longer than the other fingers (Fig. 6.1d): it is used to tap the wood and detect the presence of larvae, and when the larvae are exposed by the teeth, it is used to extract them from the wood. What is really remarkable is that a very similar set of adaptations is seen in a small lemur that lives in Madagascar, the aye-aye, Daubentonia madagascarensis. The aye-aye has a similarly shaped skull, reinforced for gouging wood, like that 206 Life of Marsupials

of the trioks and it also has a pair of elongated fingers for tapping the wood and extracting insect larvae (Cartmill 1974). However, its elongated finger is digit 3 rather than digit 4. In other parts of the world woodpeckers occupy this particular niche and they also have a skull reinforced for withstanding strong forces applied by the bill. But there are no woodpeckers in Madagascar or in New Guinea and Australia. This is a beautiful example of convergent evolution of primates and marsupials, responding independently to a vacant niche in Madagascar and New Guinea. In Queensland the striped possum lives in tropical rainforest and adjacent woodlands and its diet, based on a preliminary study by Handasyde and Martin (1996), includes wood-boring insects excavated from rotten tree trunks and wild figs, which it forages for in a home range of 5–21 ha.

The sugar glider By far the commonest of all the petaurids is the smallest species, the sugar glider, Petaurus brevi- ceps (Fig. 6.1e, Plate 11). It has an extensive distribution through southern, eastern and northern Australia (Fig. 6.4), where it occupies open woodlands, eucalypt forest and rainforest, and even small patches of forest in farmland and along roadsides. It also occurs throughout most of New Guinea, except for the very high montane areas, and on some of the adjacent islands. It does not occur in Western Australia and formerly did not occur in Tasmania. However, it was introduced in about 1835 and is now widely distributed across the island. Thus, its present range extends from just north of the Equator to 43°S in Tasmania. Like many other small mammals it follows ‘Bergman’s Rule’ – being largest in the higher latitudes of its range and smallest in the lower. In New Guinea males weigh about 90 g and females about 80 g, whereas in Victoria males weigh about 140 g and females about 115 g. Which factors determine this remains a puzzle, although it is usually thought to be a response to the cooler temperatures at higher latitudes and the economy that larger size provides for thermoreg- ulation. When Quin et al (1996a) attempted to relate this to factors other than latitude for the sugar glider, ambient temperature and rainfall correlated only slightly with body size and not nearly as strongly as latitude itself. Since sugar gliders are only active during the night and are very dependent on feeding for much of that time, it may be that the greater duration of feeding time in higher latitudes enables the animals to grow to a larger size. A related factor is the effect of a larger species of glider in the same habitat, competing for the same resources (see Larger petaurids and Table 6.3). What has made this small gliding possum so successful? The major factors seem to be its strong communal family structure, its opportunistic diet, and its energy economies in transport and thermoregulation.

Social behaviour of sugar gliders Sugar gliders are gregarious, living together in communal group nests of seven to 10 animals in preference to living in smaller, more dispersed groups. Each group has one, or very rarely two, dominant males, which are 20% heavier than subordinate males and females in the group. They carry out social activities such as patrolling and defending the territory and mating with females. All the adult females in the group breed and the young produced stay with the group until one year old, when they disperse. Some of the female young may remain in the group but all the young males disperse. Many males die at this time, so that the prevailing sex ratio in communal groups is always biased to females (Sadler and Ward 1999). This pattern is the exact opposite of the social structure of Leadbeater’s possum, its near relative. During daylight hours the members of a group nest together in a hollow tree and at night they forage separately in a territory of 0.5–4 ha, which they jointly defend against other sugar gliders. Their most vigorous defence is reserved for special trees that provide sap or gum. This communal Pygmy possums and sugar gliders: pollen eaters and sap suckers 207

structure is maintained by complex communications between members of the group. In the nest they mutually groom each other and when foraging they utter various calls that provide general recognition and alarm at the presence of a predator. According to Flannery (1995) the calls include buzzing, droning, screaming, hissing and yapping sounds. But the main communication is by specific odours (Schultze-Westrum 1969, Stoddart and Bradley 1991). The dominant male in the group develops a patch on the front of his head that secretes a pungent odour and he rubs this onto the fur of all other members of the group and their surroundings. All the adults in the group also have a sternal gland on the chest, which secretes a different odour, and there are also several glands around the cloaca that provide other odours. Some of these odours enable recognition of the species, some are specific to the communal group and some are individual signatures. If a strange adult enters the territory of the group, it is quickly detected by its different odour and the group expresses aggression towards it, especially the dominant males. Each adult female has her own odour, which her own offspring can recog- nise and differentiate from that of other females in the group. The degree of development of the frontal gland is greatest in the most dominant male and is controlled by testosterone (Mallick et al 1994). The dominant male is usually the heaviest animal in the group and has the highest level of testosterone. Furthermore, when challenged with an injection of the brain hormone, gonadotrophin releasing hormone, GnRH, which indirectly stimulates secretion of testosterone by the testes, the level rose highest in the dominant male. This pattern was dramatically reversed if a dominant male was put into a foreign communal group (Bradley and Stoddart 1997): in this situation his weight and testosterone level declined and the response to GnRH was less. At the same time a subordinate male in his former group assumed the dominant role and his weight and testosterone level increased, as did his response to GnRH. Clearly, the change in response of these males was determined by their changed social conditions. The factor that controls this is the odour of the dominant male, shown by recording the responses in heart rate and respiratory rate of subordinate males presented with the odour of strange males. When presented with the odour of castrated males or females there was no detectable response, but when the odour of a dominant male was presented there was a rapid rise in heart rate and an increase in the level of hormones associated with stress, such as cortisol and adrenalin (Stoddart and Bradley 1991). In species in which males are larger than females it is generally supposed that the dominant male sires all the young of the females in his group. In communal groups of sugar gliders when two males share the dominant position, do they both contribute to the progeny or does one of the pair tolerate the other because he is closely related? This has been investigated by comparing the relatedness of the dominant males to each other and to the females in the group. In one study of captive sugar gliders the co-dominants were more closely related to each other than either was to the females (Klettenheimer et al 1997), which suggests that the males cooperated with each other in defending the territory and in rearing the young in the group because they were related. Cooperation rather than competition avoids costly fighting within the group and allows for the benefits in energy savings of the whole group sleeping together in one nest. However, in three wild populations of sugar gliders, where subordinate males could readily disperse, only one out of 27 groups contained co-dominant males, a father and son that had joined another group together (Sadler and Ward 1999).

Diet of sugar gliders In Victoria sugar gliders get most of their energy from gum exudates of Acacia and the sap of Eucalyptus, and obtaining it occupies 40–50% of their feeding time (Smith 1982a). Since the annual production of gum is about 780 g/tree, seven trees could provide one sugar glider with its annual energy needs. Manna, honeydew, nectar and pollen made up only a minor part 208 Life of Marsupials

of the diet and they got very little energy from insects. However, insects of the forest canopy and from under bark were a major component of the diet during spring and summer. This is when the young of the year are growing fastest and their protein requirements are greatest. The breeding season appears to be timed to coincide with the spring flush of insects. Further north in coastal New South Wales nectar and pollen from Banksia and Eucalyptus flowers were the prevalent food source through winter, spring and summer, with Acacia gum and insects making only a minor component of the diet in autumn (Quin 1995). Pollen here was the main source of protein for reproduction, as it is in north Queensland, where they foraged in the understo- rey acacias (Jackson 2000b). In the different environment of New Guinea sugar gliders feed on fruits, lowland pitpit, Saccharum edule, and beetle larvae but nothing is known about seasonal differences in the diet and the relative amount of time spent gathering it. Thus, sugar gliders can exploit whatever food resource is most abundant in the area at the time and they are not restricted by any one. They are small enough to feed on nectar and pollen but strong enough to cut into a tree to draw sap and open beetles for their contents. Sugar gliders feed only during the hours of darkness, emerging at dusk from their daytime communal nest in a hollow tree, and returning to the same or a different hollow before dawn. The only sugar gliders that have been seen to feed in daylight have been undernourished animals excluded from the communal group that occupies the territory at night. Their survival is low because they are easy prey to predators during the daytime. When held in captivity under condi- tions of controlled light cycles, sugar gliders became active soon after lights went out and ceased activity before the lights came on. On short nights, mimicking summer nights, they were active throughout the dark phase but on long nights they became inactive two hours before the lights came on (Goldingay 1984). In captivity they were not constrained by food shortage, which may have been the reason that they did not remain active throughout the night – they had satisfied their metabolic needs. Their body temperature and oxygen consumption reflected their active and resting phases, with daytime body temperatures when they were sleeping being about two degrees lower than during the active periods, and oxygen consumption being about half the night time level (Dawson and May 1984). It is difficult to gain an idea of the amount of time that sugar gliders spend foraging at night but Carthew (1994) devised a way to do this, by attaching a reel of fine thread to their backs and following where the thread was left. She found that they moved around their woodland habitat by gliding, alighting directly on Banksia trees to visit the inflorescences and gather nectar and pollen. On successive nights they would follow the same route, revisiting the same trees but feeding from different inflorescences, so that most of the available flowers were visited. Because of this their role in pollination may be important to the plants. She found that some sugar gliders covered 265 m (the length of the thread) in 1 h of foraging, while others took all night to cover this distance. This means that they can, in theory, travel 1 km in a night’s foraging. The actual distance covered, however, depends more on the density of flowering shrubs, with distance between shrubs visited ranging from 1–60 m. The teeth of sugar gliders reflect their diet and feeding activities. Although they eat insects and other invertebrates, as do small dasyurids (see Chapter 4), their dentition is very different. Their incisors are large chisel-like teeth, which are used to tear open bark and to cut the living tissues of trees to reach the sap, and their molar teeth are squared off and rounded to form grinding surfaces. Unlike in dasyurids, neither the front nor back teeth are sharp pointed, nor do the molars have the shearing facets that enable small dasyurids to cut up insect bodies into very small fragments. Instead, sugar gliders discard the outer parts of insects and only consume the soft inner tissues. As a result they are unable to extract as much protein from insect bodies as can small dasyurids (Moore and Sanson 1995). But since they do not rely on insects for their energy needs but only for their protein needs during reproduction, this is not a limitation for them. Pygmy possums and sugar gliders: pollen eaters and sap suckers 209

One consequence of a diet of exudates and pollen and the soft tissues of insects is that little abrasion is caused by the passage of food through the gut, so the loss of protein nitrogen from the cells lining the gut is reduced. As in the eastern pygmy possum (van Tets and Hulbert 1999), this is reflected in a very low nitrogen requirement in sugar gliders (Smith and Green 1987), and probably all sap feeders. Indeed, the daily maintenance nitrogen (N) requirements of these two species (46 and 87 mg N/kg0.75 respectively) are lower than for any other species of marsupial (Hume 1999). This is another advantage of their diet compared to that of the larger ringtail possums and gliders that rely on protein in eucalypt leaf (see Table 7.1). However, the dilute solution of sugars in the exudates and nectar and the free water on the trees from which they feed, means that their water turnover rate is about 40% higher than for carnivorous species of the same body size. This water must be discharged through the kidneys, which have a relatively large glomerular zone for filtration of the blood and a small medullary zone where water is reabsorbed. This is the opposite of desert-living carnivorous species in which most water is reabsorbed in a large medullary zone.

Energy conservation Sugar gliders conserve energy in two ways: gliding reduces the energy requirements for travel as mentioned above, and communal nesting conserves energy otherwise required to keep warm. Fleming (1980) measured oxygen consumption in sugar gliders held at different temperatures from 4°C to 40°C, either singly or in groups of four (Fig. 6.11). Single gliders increased their consumption of oxygen from a minimum at 28°C to a maximum at 4°C, as expected. However, the sugar gliders that could huddle together consumed half as much oxygen at each temperature

3.0

2.5

2.0

1.5

1.0 Respiration (ml oxygen/g/h) 0.5

0.0 010203040 Ambient temperature (C)

Figure 6.11: How sugar gliders, Petaurus breviceps, save energy by huddling at low temperatures: respiration of single gliders ({) and huddled gliders (V) at a range of ambient temperatures. After Fleming (1980). 210 Life of Marsupials

and reached the minimum at an ambient temperature of 16°C. Huddling thus represents a very substantial saving in maintenance energy requirements. In addition, sugar gliders can lower their body temperature 20°C below normal and enter shallow torpor, like small dasyurids do, but not as profoundly as the pygmy possums or the honey possum (Table 6.2). In this condition they use even less oxygen in thermoregulation: this is an important strategy during inclement weather, when rain prevents them from foraging. Körtner and Geiser (2000) showed how dependent free-ranging sugar gliders are on favour- able weather conditions during the night to meet their daily energy requirements. They placed tiny radio transmitters in the bodies of two or three gliders in each of three different communal groups and recorded their body temperatures continuously for several days (Fig. 6.12). On clear nights the gliders were active throughout the night, as indicated by high body temperatures (37°C). However, after cold rainy nights, when the gliders could not feed, their body tempera- tures fell to 15°C during the following day in the nest and they were torpid. They recovered normal body temperature in the evening prior to emerging for feeding that night. Becoming torpid was evidently a common response to conserve energy during inclement weather and lack of food: it indicates how crucial it is for such a small mammal to feed frequently. The SMR of sugar gliders is within the normal range for marsupials and their FMR is 3.8 times higher (Nagy and Suckling 1985). This is considerably less than the 6–10-fold difference of small dasyurids that must hunt insects for their energy and protein needs (see Chapter 4) or of

40

35

30

25

20

15 temperature (C) 10

5 Ambient

0 30 20 10

rain (mm) 0 5 August 6 August 7 August 8 August 9 August

Figure 6.12: Body temperatures in four free-ranging sugar gliders, Petaurus breviceps, measured continuously over four days. Rainfall on 5–6 and 7–8 August induced torpor with lowered body temperatures during the following daytime, whereas after the dry nights of 6–7 and 8–9 August none of the gliders entered torpor. After Körtner and Geisler (2000). Pygmy possums and sugar gliders: pollen eaters and sap suckers 211

Leadbeater’s possum. Only part of the explanation is that plant exudates provide a more acces- sible source of energy than do insects. The greatest cause of energy saving by the sugar glider is the lower transport costs of gliding. Sugar gliders and the related species have a gliding membrane that extends from digit 5 of the hand to digit 1 of the foot, which gives them the maximum area for lift (Fig. 6.7b). They cannot gain altitude while gliding but they can travel up to 20 m in a single glide, losing about 10 m in height in that distance (Jackson 2000a). The only energy cost is the vertical climb to regain height for the next glide, which for a small animal like the sugar glider is very much less than the cost of travelling across the ground. The advantage of gliding is well illustrated by comparing the energy costs of the sugar glider with those of the non-gliding Leadbeater’s possum, which is the same size and consumes the same diet (Fig. 6.13). The energy costs in the nest are about the same for both species but the total energy costs outside the nest are one-third more in Leadbeater’s possums. While the costs of thermoregulation are higher in the sugar glider, because of heat loss from the larger surface area when gliding, activity costs for the Leadbeater’s possum are twice those of the sugar glider. Leadbeater’s possums can meet their food requirements within a small area, whereas the sugar glider can travel considerable distances to visit particular trees and exploit scattered food sources.

250

200

150 activity

(kJ/d) 100

thermogregulation 50

Daily energy consumption standard mettabot lism resting in nest 0 Leadbeater s Sugar glider possum

Figure 6.13: A comparison of the daily energy budgets of sugar gliders, Petaurus breviceps, and Leadbeater’s possums, Gymnobelideus leadbeateri, in Victoria to show the energy savings of gliding. Data from Nagy and Suckling (1985, table 4).

Reproduction and life history of the sugar glider In the southern parts of its range the sugar glider is a seasonal breeder, most young being born between August and November after a gestation of 15–17 days (Suckling 1984). Like burramys and the brushtail possum, Trichosurus vulpecula, females will return to oestrus 12 days after the young are removed or lost from the pouch. There are four teats but the litter size is either one or two. The young remain in the pouch for 70 days, and then ride on the mother’s back for 20 days before being left in the nest until weaned at 120 days, the same as the Leadbeater’s 212 Life of Marsupials

possum. Pouch emergence coincides with the spring flush of insects in the diet, which provides essential protein for late lactation and the demands of the young. If conditions are favourable, the female will have a second litter in the season. Males can produce sperm all year and females are potentially capable of breeding whenever conditions are favourable. Further north, females may produce three litters and in New Guinea they breed throughout the year. This pattern of reproduction is more like the dunnarts, Sminthopsis, and bandicoots, Isoodon, than the strictly seasonal species, such as the brown antechinus, Antechinus stuartii. There is heavy mortality among the young gliders, especially young males, when they disperse from the family group at about seven months. However, for those sugar gliders that survive and become established in a group, they may live for four to six more years. For such a female her lifetime production is potentially as high as 30 young, considerably exceeding the productive capacity of the female Leadbeater’s possum.

Larger petaurids There are four other species of Petaurus, all larger than the sugar glider but none nearly as widespread: indeed, two species have very restricted distributions (Fig. 6.4). The very rare species, Petaurus abidi, which is more than twice the size of the sugar glider, lives on the northern slopes of New Guinea between 800 m and 1200 m in undisturbed rainforest, but very little is known about it. The mahogany glider, Petaurus gracilis, rediscovered in 1989, is restricted to a very small area of 108 km by about 7 km in coastal woodland below 200 m, in north Queensland. Its diet is similar to the sugar glider’s and both species coexist in the same area, but it is restricted to the nectar and pollen of a very few tree species that occur in this area and it relies on complex seasonal cycles of food availability (Jackson 2000b). It occupies the higher strata of the forest, while sugar gliders in the same forest occupy the understorey and second growth. By divid- ing the resources the two species can coexist in the same forest. However, the sugar gliders in this area are much smaller than elsewhere (Table 6.3) and their density is one-tenth of that in Victoria, where the mahogany glider does not occur. Combining these two parameters, the total biomass of sugar gliders in Victoria was between 120 and 885 g/ha, while in north Queensland the biomass of sugar gliders was 17–37 g/ha, and that of mahogany gliders in the same area was 63–100 g/ha (Jackson 2000b). Unfortunately, the available habitat for the mahogany glider is inexorably diminishing because the land is also suitable for sugar cane cultivation, so the prospects for the long-term survival of the mahogany glider are low.

Table 6.3: Effects of competition on Petaurus breviceps when it shares habitat with a larger species of Petaurus Data from Quin (1995), Jackson (2000c).

Species Mass Mass Density Biomass ( (g) & (g) (n/ha) (g/ha) P. breviceps (Glengarry, Vic.) 125 110 1.1–1.2 121–275 P. breviceps (Willung, Vic.) 140 125 2.9–6.1 305–885 P. breviceps with P. norfolcensis 120 104 0.24–0.54 25–64 P. breviceps with P. gracilis 80 70 0.24–0.46 17–37

The squirrel glider, Petaurus norfolcensis, occurs in the woodlands of eastern Australia, from the northern plains of Victoria (Menkhorst et al 1988, van der Ree 2002) to Queensland (Fig. 6.4). In most of this country its preferred habitat is now restricted to the edges of country roads Pygmy possums and sugar gliders: pollen eaters and sap suckers 213

and the few trees left in open paddocks. Nevertheless, it has adapted to living in extremely long, narrow home ranges of up to 800 m by 20 m wide, with a total area of 1.4–2.8 ha (van der Ree and Bennett 2003). While this is about the same size as the home ranges in uncleared woodland, the distance required to travel each night is much longer, although defending the 20 m end of an elongated home range is easier. The squirrel glider does not occur in the area occupied by the similar-sized mahogany glider but its habitat requirements and diet are very similar to those of the sugar glider with which it does overlap. In Quin’s (1995) study, where the two species lived in the same area of northern New South Wales, both species had the same diet, lived in communal groups of two to seven or two to nine animals and bred twice in the year. The home ranges of groups of both species overlapped to a considerable degree and there appeared to be no defence of territory either within species or between the two species. At a finer level there may be partitioning of the habitat, as there is between the mahogany glider and the sugar glider. However, as in that case, the density of sugar gliders was about one-quarter of Victorian populations where squirrel gliders do not occur, and the total biomass was also much less (25–64 g/ha) and much less than the biomass of squirrel gliders (171–328 g/ha) (Tables 6.2 and 6.3). It seems that where they overlap and there is competition between the species, the sugar glider is effectively displaced by the larger species, which occupies the best hollows (Traill and Lill 1997).

The yellow-bellied glider The largest of all the petaurids is the fluffy or yellow-bellied glider, Petaurus australis (Fig. 6.1f, Plate 11), which weighs between 500 g and 700 g. It lives exclusively in the wet sclerophyll forests of the eastern seaboard of Australia, from western Victoria at 37°S to 17°S in north Queensland (Fig. 6.4) and has been the subject of much recent research because of its promi- nent status in the native forests and its vulnerability to logging practices. Its vulnerability is partly because it is at the upper size limit for subsistence on plant exudates. Why is there an upper size limit to living on exudates of plants and how does the yellow-bellied glider manage to live at the limit? One way to look at the problem is to compare this species with the sugar glider, which is one- quarter its size but lives on the same food resources, and then in Chapter 7 to compare it with the greater glider, Petauroides volans, which never takes sap or other exudates but lives entirely on leaves of the same species of trees in the same forest.

Diet of the yellow-bellied glider Yellow-bellied gliders are the true specialists on sap feeding: with their powerful lower incisors they cut the bark of particular trees within their home territory and feed for many hours of the night on the sap that flows out. Compared to the sugar glider, they spend much less time feeding on other exudates and never feed on acacia gum. They do hunt for insects under bark and this probably constitutes their main source of protein. Like the sugar glider, they are strictly noctur- nal and spend the daylight hours in a nest hollow in a large tree, which may be a different species than the sap trees they feed on. Goldingay (1989) observed how yellow-bellied gliders spent their time, by following individ- ual animals over many nights. Like sugar gliders, they were active for almost the whole night in summer, emerging from their daytime den 45 minutes after sunset and retiring only 10 minutes before sunrise, but in winter they were active for only three-quarters of the night, emerging one hour after sunset and retiring up to two hours before sunrise. However, during the active period, they spent most of the time (81%) feeding, or climbing and gliding to food sources (8.7%), which left very little time for grooming, resting or social activities. Of the feeding time in New South Wales, 50% was spent on sap, 33% on manna, 13% on honeydew and 5% on 214 Life of Marsupials

insect foraging. In Victoria, at the extreme western edge of its range, 83% of the feeding time was spent on sap, except for two months of the year when honeydew and nectar were predominant (Carthew et al 1999). Likewise, at the other end of its range, in north Queensland, sap feeding accounted for 80% of feeding time, with Banksia nectar a small but important component of the diet in winter and autumn, and all other sources of food being very minor components (Quin et al 1996b, Bradford and Harrington 1999). Nevertheless, glider faeces found below sap trees contained the pollen of 16 species of flowering trees, including six rainforest species that were sufficiently common to be a source of protein for the gliders. While sap is clearly the most important source of energy for yellow-bellied gliders every- where, they seem to be very selective in the trees they tap. In any locality they select only one or two species of Eucalyptus from all those present and then only certain trees, which will be visited frequently for a time and then abandoned. The main factors that influence their choice to feed on sap are wet conditions in the forest, a lack of other sources of food, such as nectar or honeydew, and a high sap flow in the particular tree at the time (Goldingay 1991). Sap flow evidently varies between individual trees, and in the same tree at different times, and the gliders select a tree by making an initial incision to test the flow before cutting a full-size incision and feeding. Other members of the group may then join to feed at the same tree. Thus, the main factor that determines whether an area of forest can support yellow-bellied gliders is the number and distribution of suitable sap-producing trees (Eyre and Goldingay 2003). This influences the size of the home range and the social structure of the group, as well as the long-term survival of the species in the forest.

Home range and how to defend it Like sugar gliders, yellow-bellied gliders live in social groups but the group has exclusive use of a much larger home range. It contains several favoured sap trees and one or more den trees, which may be widely scattered through the home range (Fig. 6.14). In the Victorian highlands the home ranges of pairs of gliders were between 30 ha and 55 ha (Craig 1985) and in southern New South Wales up to 85 ha (Goldingay and Kavanagh 1993). By contrast, at another site in New South Wales, the home ranges occupied by groups of four or more gliders were 34 ha and in north Queensland even larger groups of gliders occupied home ranges of less than 30 ha. The difference in group size and home range reflect the relative abundance of food trees. Where these are dispersed a pair of gliders requires a large exclusive home range (Henry and Craig 1984, Kavanagh 1987), but where the preferred sap species is abundant and produces copious sap through most of the year, as bloodwood, Eucalyptus resinifera, does in north Queensland, yellow- bellied gliders live in larger family groups occupying much smaller home ranges. Nevertheless, even in this prime habitat, the density of this largest sap-feeding glider is much lower than of any other sap feeder (Table 6.1). Despite the larger body weight of the gliders, the total biomass in three studies of the species ranged from only 21 g/ha to 122 g/ha. A large home range must be regularly occupied and defended. Yellow-bellied gliders travel considerable distances each night to visit sap trees and trees in flower and to patrol the bounda- ries of their territory. They do this by gliding. A single glide can range from 10 m to 100 m (mean 39 m), which enables them to travel more than 2 km in one night (Goldingay 1989). This is considerably more than the sugar glider can accomplish and allows rapid cross-country movement to patrol the group’s large territory. Yellow-bellied gliders also make loud calls that can be heard up to 0.5 km away, which adver- tise their presence to other groups (Kavanagh and Rohan-Jones 1982, Goldingay 1994). When not gliding they make either a full call of two shrieks and a long gurgle, or a short call of one shriek and many repeated gurgles. They also emit a low frequency moan as they glide from a tree and alight at another, and a long gurgle while gliding. These calls are made only during the night Pygmy possums and sugar gliders: pollen eaters and sap suckers 215

Group 1

Group 2

Group 3

0 300m

Figure 6.14: The large separate home ranges of three groups of yellow-bellied gliders, Petaurus australis, along two streamlines (B) in eucalypt forest of eastern Australia. Each group has its own (Q) den trees, (O) sap trees, and (V) honeydew trees. After Goldingay (1992). and are more common during the first 3–4 h of activity as the gliders make their way from their dens and forage for food requiring short periods in trees. Later in the night there is a period of lesser calling, while they are feeding on sap, manna or honeydew. Calling increases again at the end of the night as the gliders make their way back to their den trees. Both males and females call and they do so more frequently at the periphery of their home range (18 calls/h) than when in the core (4 calls/h). This suggests that an important function of calling is to define territory boundaries and ensure exclusive use to the group. This was tested experimentally by playing a full call on a cassette recorder, when gliders responded from 200 m away and came to within 15 m of the cassette. Among marsupials this function of calling appears 216 Life of Marsupials

to be unique to the yellow-bellied glider, although loud calls are used for this purpose by howler monkeys and gibbons in other parts of the world. It is another example of convergent evolution, in this case of a specific behavioural repertoire, in marsupials and placentals.

Social behaviour The yellow-bellied glider is a gregarious species in which groups comprise one male and one or several females and their weaned young. In Victoria a single pair in serial monogamy occupies each home area with their single offspring of the previous year (Craig 1985). Further north in New South Wales the social groups varied from monogamous pairs and offspring, to polygy- nous groups with a dominant male and several breeding females and their offspring, as in north Queensland populations. Monogamy is thought to evolve when limited food resources prevent social grouping, and where breeding success depends on territorial defence by the male partner. However, polygyny can occur where abundant food resources allow the male glider to defend an area that can support more than one female (Goldingay 1992). In the populations in north Queensland the dominant male in the group, as in sugar glider groups, has a prominent scent gland on the top of his head and another scent gland under the tail, some distance from the cloaca (Russell 1984). Scent is transferred most frequently when the group is feeding together at a sap site. Any member of the group can approach another member and rub the top of its head against the underside of the tail of the other, thereby mutually exchang- ing scent. Only the dominant male produces the scent and it is transferred to all members of the group by this mutual head-to-tail rubbing. He directs head rubbing to the females of higher rank more than to subadults and so the intensity of the odour of each member’s tail is directly proportional to its rank in the group. As with sugar gliders, the scent provides the means for differentiating members of the group from foreign gliders, who may attempt to feed from the sap site. However, a foreign female can join the group and feed at the sap site if she performs head rubbing on the dominant male and he accepts her advance. But an adult male that attempts to intrude on a sap site is immediately challenged by the resident male and driven off, followed by loud calls. If the intruder does not retreat, a fight ensues and one male eventually claims the site for his group. For this reason the dominant male is very active through the night defending sap sites against challengers. This may explain why a home range only ever has a few active sap sites, since he could not defend more of them. And it also explains why he always feeds first at the sap site to get the energy needed for the constant patrolling of the territory.

Reproduction of yellow-bellied gliders The yellow-bellied glider’s pattern of reproduction reflects the limitations of life as a large sap feeder: females have two teats but usually only one young is born each year and it is carried in the pouch for 100 days and placed in a nest for a further two months. This is a much longer period to independence than in sugar gliders. Although breeding is seasonal, food availability at critical times of lactation may determine it. Breeding in New South Wales began when exudates were abundant in July–September (Goldingay and Kavanagh 1990), and late lactation and weaning coincided with the highly predictable bark shedding of Eucalyptus viminalis in December– January, when abundant insect protein was available. At another site where nectar and pollen were available through most of the year, most gliders gave birth in February–April and late lacta- tion coincided with the main flowering of Eucalyptus maculata in August (Goldingay 1992). The question that remains is whether reproduction is initiated three months before a predict- able flowering period or whether breeding is more opportunistic and, if the flowering fails to eventuate, the young die. Because of the variability in the time of breeding at different sites it seems unlikely that a regular signal like changing photoperiod can be involved. It is interesting to Pygmy possums and sugar gliders: pollen eaters and sap suckers 217

compare this very low rate of reproduction in the largest sap feeder with the two litters of up to four young by the common ringtail possum, Pseudocheirus peregrinus, living in the same habitat on leaf and bacterial protein (see Chapter 7). If it survives pouch life, the single young remains within the family group for two years and then disperses to join or form other groups. Young females can more readily join existing groups, whereas young males may only be able to do so if a group has lost its dominant male. In Craig’s (1985) study in Victoria there was a considerable turnover of adults in the small groups, which suggests that mortality is high. For older animals the need to spend 90% of their active time feeding, gliding and maintaining sap sites is so taxing that when they are unable to continue they probably decline quickly – in one study a dead adult was found at the base of the main sap tree. This is another consequence of the limits of this kind of life for the largest sap-feeding glider.

Conservation Because of its critical dependence on forest trees for its food and because of the large area needed exclusively by each group, the yellow-bellied glider is an indicator species for forest conservation. Goldingay and Possingham (1995) have attempted to estimate the minimum area needed for the long-term survival of the species. Based on the results of the three main studies of this glider in Victoria, New South Wales and Queensland they estimated the minimum area to be 18 000 ha. However, most of the forest reserves designated for wildlife conservation are less than 10 000 ha and on this estimate are, therefore, inadequate. In north Queensland Bradford and Harrington (1999) made a similar analysis of the needs of the isolated population of yellow-bellied gliders north of the Herbert River gorge above 700 m. The gliders are restricted to 1400 ha of suitable habitat, where they are wholly dependent on sap of Eucalyptus resinifera, and for nest sites on Eucalyptus grandis. But rainforest is encroaching on the habitat and the population is already declining within this small area of suitable habitat. This is evident from old sap site scars, which remain obvious for many years, on trees where gliders no longer live. On the basis of Goldingay and Possingham’s analysis, albeit in another region, this area is wholly insufficient to sustain this northern subspecies of the yellow-bellied glider in the long term.

100 Exudivore / Insectivore Folivores 80 Yellow-bellied glider Squirrel glider 60 Mahogany Sugar glider 40 glider Common ringtail possum Greater 20 Bear glider Brushtail cuscus possum 0

Percent of active time spent feeding 100 1000 10000

Body mass (g)

Figure 6.15: A comparison of the proportion of active time required to forage for either plant exudates and insects, or eucalypt leaves, and how this is related to body mass. After Jackson and Johnson (2002). 218 Life of Marsupials

Earlier the question was asked of why there should be an upper limit on the size of a marsu- pial that lives on sap and other exudates. Comparing the very successful sugar glider with the yellow-bellied glider, the larger species has to spend almost twice as much of the active time during the night foraging for the dispersed food and so it needs the exclusive use of a far larger area to ensure sufficient sap sites to provide it (Table 6.1). As a consequence its density and total biomass are very low and, at one young per female per year, its ability to reproduce is minimal. Very small but highly dispersed populations are much more vulnerable to alterations in the availability of food and major changes to the forest ecosystem, such as logging, than are the smaller species and it seems likely that they rapidly succumb if these are unavailable. By contrast those marsupial species that are adapted to feeding on the leaves of the same forest eucalypts have an abundant source of food and do not need to move long distances at night and so can occupy very small home ranges (Fig. 6.15). Although the highest biomass of the yellow-bellied glider is 122 g/ha, that of the greater glider in the same forest is 3300 g/ha (Table 6.1). Furthermore, greater gliders spend a much shorter part of the night feeding and much time, apparently inactive, as they process their leaf diet (see Chapter 7).

Figure 1.7: (a) (Opposite, top right) One example of how chromosome painting can resolve homologies of chromosomes between related marsupials. The 11 chromosomes of the male swamp wallaby, Wallabia bicolor, coloured blue, have been incubated with fluorescent-labelled tammar wallaby, Macropus eugenii, chromosomes 2 (green), 7 (red) and X (white). The labelled DNA has hybridised with the complementary regions of the swamp wallaby X and Y2 chromosomes: the long arm of the X chromosome and all of Y2 are the homologue of tammar wallaby chromosomes 2 and 7 and only the short arm of the X chromosome is the homologue of the tammar X. (JAM Graves.) (b) (Opposite, top left) Schematic summary of the full homologies between the 11 chromosomes of the male swamp wallaby, Wallabia bicolour (4 pairs of autosomes + Y1, Y2 and X) and the 16 chromosomes (7 pairs of autosomes + X and Y) of the tammar wallaby, Macropus eugenii. After Toder et al (1997).

Figure 1.8: (Opposite, bottom) A composite diagram built up from many separate studies, using the same colour code as in Fig.1.7, to show how chromosome painting helps to resolve the differences between species with different numbers of chromosomes in the same family (Macropodinae), and species with the same number of chromosomes in different families (Huon tree kangaroo, Dendrolagus matschiei, southern hairy-nosed wombat, Lasiorhinus laitfrons, and stripe-faced dunnart, Sminthopsis macroura). For the Macropodinae, the chromosomes of the four species on the right can be derived from an ancestral 22 chromosomes as seen in the New Guinea wallabies, , the , Thylogale, and the rock wallabies, Petrogale. Adapted by Robert Rapkins from data in Glas et al (1999) and De Leo et al (1999). PLATE 1

Macropus rufus Macropus eugenii 1 2 3 4 5 6 7 8 9 X 1 2 3 4 5 6 7 X Y

Ancestral macropod

Dorcopsis spp. Thylogale thetis Petrogale lateralis Dendrolagus matschiei 1 2 3 4 5 6 7 8 910X Y 1 2 3 4 5 6 X

Ancestral Wallabia bicolor Australian 1 2 3 4 Y Y X marsupial 1 2 1 2 3 4 5 6 X Y

Lasiorhinus latifrons

126X 345 Vombatidae

Sminthopsis macroura 12X 3456 Y Dasyuridae PLATE 2

Figure 1.9: The principle of DNA/DNA hybridisation between closely related and distantly related species. Hybrids are formed from small amounts of radioactively labelled DNA and 1000 times as much unlabeled DNA from the same or a different species (top panel). The DNA is melted to separate it into single strands, which are incubated to allow the single strands to reassociate. Much of the unlabeled DNA reassociates with (a) complementary unlabeled strands and, because these products are not radioactive they do not affect later measurements. Some of the single strands (b) fail to reassociate and (c) only about 1% of the labelled strands hybridise with other labelled strands. (d) Some of the DNA forms hybrid duplex molecules, consisting of one labelled and one unlabeled strand. In hybrids of DNA from two species, the proportion of the nucleotide bases along a strand that are paired with complementary partners on the adjacent strand depends on the genetic similarity of the two species: because a greater number of bonds link the strands in well-matched hybrids (bottom panel left), they melt at a higher temperature than poorly matched ones (bottom panel right). From Sibley and Ahlquist (1986, with permission of Scientific American and Patricia J Wynne, artist). PLATE 3

(a)

(b)

(c) (d) Figure 2.9: Stages in tammar wallaby, Macropus eugenii, and brushtail possum, Trichosurus vulpecula, development: (a) the unattached vesicle of the tammar on day 16, showing spreading mesoderm (m), primitive streak (ps) at tail end, paired somites (s) on each side of the mid line, heart (h) and early head fold (hf) (Ivan Fox); (b) a later stage possum vesicle with the vascular yolk sac enshrouding the embryo and the non-vascular yolk sac beneath it; (c) a full-term tammar fetus enclosed by the translucent amnion (am) and with a large, non-vascular allantois (al) and vascular yolk sac (ys) extending from the umbilical cord (uc) (Ivan Fox); (d) a newborn tammar soon after attaching to a teat in the mother’s pouch, with conspicuous blood vessels under the moist skin, large forelimbs and conspicuous claws on its fingers (LA Hinds). PLATE 4

(a) (b)

(c)

(d)

Figure 3.1: Some American marsupials (see Chapter 3): (a) the white-eared opossum, Didelphis albiventris; (b) the gray four-eyed opossum, Philander opossum; (c) the water opossum, Chironectes minimus swimming with its forelegs spread out to catch prey, and (d) feeding on an and displaying its webbed hind feet; (e) the bare-tailed woolly opossum, Caluromys philander (photo D. Julien-Laferrière – CNRS/MNHN); (f) the pouchless mouse opossum, Marmosops impavidus, with attached young; (g) the brown short-tailed opossum, Monodelphis brevicaudta; (h) the silky shrew opossum, Caenolestes fuliginosus. PLATE 5

(e) (f)

(g)

(h) PLATE 6

(a)

(b)

Figure 4.1: A quiddity of carnivores: (a) the narrow-nosed planigale, Planigale tenuirostris; (b) the Julia Creek dunnart, Sminthopsis douglasi, with 60-day-old litter (PA Woolley and D Walsh); (c) the yellow-footed antechinus, Antechinus flavipes; (d) the New Guinea spotted quoll, Dasyurus albopunctatus (PA Woolley and D Walsh); (e) the banded anteater or numbat, Myrmecobius fasciatus (GB Sharman). PLATE 7

(c)

(d)

(e) PLATE 8

(a)

(b)

Figure 5.1: Bandicoots and bilbies from Australia and New Guinea: (a) the southern brown bandicoot, Isoodon obesulus (Ederic Slater); (b) the western barred bandicoot, Perameles bougainville; (c) the bilby, Macrotis lagotis (Esther Beaton); (d) the common spiny bandicoot, Echymipera kalabu, of New Guinea (PA Woolley and D Walsh). PLATE 9

(c)

(d) PLATE 10

(b)

(a)

(c)

Figure 6.1: Pygmy possums and gliders. (a) Male honey possum, Tarsipes rostratus, on Banksia inflorescence. Note the disproportionately large scrotum (PA Woolley and D Walsh). (b) the feathertail glider, Acrobates pygmaeus (Ederic Slater). (c) the mountain pygmy possum, Burramys parvus (Ederic Slater). (d) the long-fingered trioks, Dactylopsila palpator, from New Guinea (PA Woolley). (e) a sugar glider, Petaurus breviceps, about to launch off in flight (Esther Beaton). (f) the yellow-bellied glider, Petaurus australis, the largest sap feeding marsupial (Ederic Slater). PLATE 11

(d)

(e) (f) PLATE 12

(b)

(a)

Figure 7.5: The best studied marsupial forest folivores: (a) the koala, Phascolarctos cinereus (Esther Beaton); (b) the common ringtail possum, Pseudocheirus peregrinus, the smallest marsupial folivore (Ederic Slater); (c) the pale and dark colour phases of the greater glider, Petauroides volans (Esther Beaton); (d) the brushtail possum, Trichosurus vulpecula eating leaves of red ironbark, Eucalyptus sideroxylon (Karen Marsh); (e) the spotted cuscus, Spilocusucus maculatus, of New Guinea (PA Woolley). (c) PLATE 13

(d)

(e) PLATE 14

(a)

(b)

(c)

Figure 9.2: Representative species of the Macropodidae: (a) the burrowing bettong, Bettongia lesueur, from Bernier Island; (b) a male tree kangaroo, Dendrolagus goodfellowi, from New Guinea (Esther Beaton); (c) a yellow-footed rock wallaby, Petrogale xanthopus, from Queensland (Esther Beaton); (d) female tammar wallaby, Macropus eugenii, from Kangaroo Island (Lyn Hinds); (e) adult female red-necked wallaby, Macropus rufogriseus, from south-eastern Australia grooming itself with its ‘comb,’ digits 2 and 3 of the hind foot; (f) adult female and one-year-old young eastern grey kangaroo, Macropus giganteus, from south-eastern Australia. PLATE 15

(d)

(e)

(f) PLATE 16

(a)

(b) Figure 10.3: Rock paintings of extinct species in , northern Australia: (a) a large quadruped with single young, which may represent the tapir-like marsupial, Palorchestes; (b) a representation of the long-beaked echidna, Zaglossus, that occurred in this region 20 000 years ago but now only occurs in New Guinea. Photos: George Chaloupka collection. Chapter 7

Life in the trees: koala, greater glider and possums Brushtail possum, Trichosurus vulpecula; pen and ink drawing by AG Lyne. Life in the trees: koala, greater glider and possums 221

eaf-eating marsupials inhabit the forests and woodlands of Australia, New Guinea and surrounding islands. They consume the leaves of Eucalyptus and other tree species, and Loccupy those niches that in other parts of the world are filled by flying squirrels, lemurs, leaf-eating monkeys and sloths. They range in body mass from 700 g to 15 kg, the lower limit set by the nature of their leaf diet, and the upper limit set by the energetic costs of climbing to, and being supported by, the smaller branches so as to reach the leaves. Body size also affects the sort of food eaten and the life style of folivorous marsupials as much as for the omnivores, and the poor quality of their diet exacerbates this. Eucalyptus leaves are distinctly uninviting as a source of food because of their low nutritional value and the many potentially toxic compounds they contain as their protection against browsing. However, there are about 600 species of Eucalyptus in Australia and southern New Guinea, where they are a major component of the flora of both countries, so any species that can eat Eucalyptus leaves has a constant and abundant food supply. First we will consider this uniquely Australian food source and then look at how four repre- sentative species are adapted to live on it.

Eucalyptus foliage as a source of food The family Myrtaceae, to which the genus Eucalyptus belongs, first appeared in the Eocene epoch, when Australia was warm and moist and the marsupials were small insectivorous animals. However, the first species of Eucalyptus did not appear until the dry, cool Oligocene epoch, about 35 million years ago. They evolved on the highly weathered, leached soils of the Australian conti- nent and today the leaves of Eucalyptus are characterised by high levels of indigestible fibre and lignin and low levels of essential nutrients, such as phosphorus, potassium and nitrogen. As a food source for mammals the levels of protein and non-cellulose carbohydrates are correspond- ingly low (Fig. 7.1). By contrast grasses contain higher levels of structural carbohydrates, mainly cellulose and hemicelluloses, which can be used by mammals only after bacteria in the gut have hydrolysed them. Other differences between grasses and Eucalyptus foliage are the secondary chemicals, present in the latter but absent from grasses. These chemical defences vary between species and between localities, the metabolism varying according to plant nutrition. Where soils are rich in nitrates and phosphates the plants tend to use nitrogen-based compounds, such as cyano- gen, for defence, whereas where soils are infertile, plants tend to use carbon-based terpenes and phenolics, such as tannins, because these can be synthesised during photosynthesis. Indeed, it is now considered that the predicted rise in atmospheric carbon dioxide (CO2) may so increase photosynthesis that the higher level of phenolics will make Eucalyptus species growing on poor nutrient soils unsuitable for folivorous marsupials (Lawler et al 1997). Before 1980 it was thought that arboreal marsupials were uniformly spread through all Eucalyptus forests. Then surveys conducted in southern New South Wales (Braithwaite 1983, 1984, Braithwaite et al 1983) showed that two-thirds of the arboreal marsupials occurred in only nine of the 22 forest associations: the remaining 13 forest associations, which comprised 90% of the forest, were almost devoid of arboreal marsupials. The arboreal marsupials lived in only those Eucalyptus species with a high concentration of foliage nutrients – peppermint gums, Eucalyptus radiata, manna gum, Eucalyptus viminalis, and mountain ash, Eucalyptus dalrym- pleana. One critical component of the foliage nutrients was the potassium concentration (Fig. 7.2), although phosphorus and nitrogen were also important. This, in turn, was related to the soil fertility and the substrate formations from which it was derived, and granites and shales. These findings suggested that leaf-eating marsupials do not occur in the 222 Life of Marsupials

Figure 7.1: How to live on eucalypt leaves: (a) chemical constituents of Eucalyptus foliage to show the very high proportion of inedible material and secondary chemicals (unshaded); (b) the energy budget of the greater glider, Petauroides volans, feeding exclusively on eucalypt leaf: 52 % of the gross available energy in the leaf is used to detoxify the secondary chemicals, leaving less than half for all the glider’s needs. Note how little of this is used for gliding travel. Data from Cork and Sanson (1990) and Foley et al (1990).

other forests because the foliage does not provide them with sufficient nourishment to live or, more importantly, to support young through the final stages of growth and development before weaning. Some forests that appear to have small populations of arboreal folivores may only be supporting non-breeding animals, termed sink populations. For arboreal mammals to benefit from feeding on the preferred Eucalyptus species, however, they must detoxify the other secondary compounds. For instance, tannins can bind to proteins released during digestion, making the protein unavailable to the animal and some phenolics Life in the trees: koala, greater glider and possums 223

80

70

60

50

40

30

20

10 Animal density (numbers per 100 ha)

0 25 30 35 40 45 50 55 60 65

Foliage potassium (g % dry wt x 100)

Figure 7.2: Relationship between the mean concentration of potassium in Eucalyptus foliage in 22 forest communities in south-eastern Australia, and the density of arboreal marsupials, mainly greater gliders, Petauroides volans. Only nine plant communities can support more than 30 animals/ha. After Braithwaite (1984). have to be combined with glucuronic acid to be excreted. This is energetically costly with more than half the available energy in the leaves being used in these processes (Fig. 7.1). However, the biggest obstacle to the use of Eucalyptus leaves for food is the time spent waiting for sufficient of the secondary metabolites to be eliminated before starting to eat again: the animals cannot merely increase the intake of Eucalyptus leaf to compensate for its low nutritional value because of the additional time required to denature the toxic chemicals. Some secondary compounds in the leaves are so nauseating that the animals avoid those trees altogether. Furthermore, in a eucalypt species that is usually palatable, some individual trees will be eaten and others alongside them will be ignored, or sometimes one or a few branches on a tree will be left untouched while the rest of the tree is severely defoliated. This has been approached by observing what the animals will or will not eat, and then isolating the anti-feedant chemi- cals responsible for making leaves unpalatable (Pass et al 1998). Leaves that are not eaten were subjected to chemical analysis and, at each step of the separation they were mixed into test food and offered to the animals to see whether they would eat it or not. By this means the important compound was identified as a formylated phloroglucinol, which has a phenolic ring bonded to a terpene, and was named sideroxylonal after red ironbark, Eucalyptus sideroxylon, in which it is abundant. Common ringtail possums, Pseudocheirus peregrinus, would eat leaves contain- ing low concentrations of sideroxylonal but when the concentrations were high they ate much less of the leaves, or refused to eat any at all. When the leaves of various species of Eucalyptus 224 Life of Marsupials

of differing palatability were analysed these also were highly correlated with the concentration of sideroxylonal in the leaves. In woodland of mixed Eucalyptus the concentration of sideroxy- lonal in the leaves of individual trees ranged from low and palatable to high and unpalatable. Since common ringtail possums will not eat leaves containing more than 14 mg/g dry weight sideroxylonal but they must eat a minimum amount of leaf each day, the carrying capacity of woodland can be estimated from the number of palatable trees in the area (Lawler et al 2000). Sideroxylonal is especially abundant in the Symphyomyrtus group of eucalypts but is absent from the Monocalyptus group, so common ringtail possums predominantly select eucalypts from the latter group (Fig. 7.3). The closely related greater glider, Petauroides volans, has the same preference for Monocalyptus species. Conversely, koalas, Phascolarctos cinereus, and brush- tail possums, Trichosurus vulpecula, can tolerate sideroxylonal concentrations of up to 50 mg/g dry weight and they feed almost exclusively on Symphyomyrtus species (Moore et al 2004).

Koala Greater glider

Common ringtail possum

Symphyomyrtus Monocalyptus Corymbia

Figure 7.3: A comparison of the selection of foliage from the three major groups of Eucalyptus species by four leaf-eating marsupials: each marsupial selects a different suite of species in its diet. From Moore et al (2004).

Size constraints on arboreal folivores All herbivores face similar challenges in obtaining their food from leaves: they must break the cell walls in order to obtain the cell cytoplasm, which contains plant proteins and some carbohy- drates; they must dispose of the indigestible fibre, which is mainly cellulose or lignin; and they must detoxify the plant’s defensive chemicals, the phenolics, terpenes and cyanogens. Life in the trees: koala, greater glider and possums 225

One way this is done is to have bacteria ferment the material in a special part of the gut, either in the forestomach as ruminants and kangaroos do, or in the caecum and colon, as horses, rabbits, possums and wombats do (Fig. 7.4). Bacteria hydrolyse the cellulose of cell walls to simple sugars, which are then fermented to short-chain fatty acids (SCFA), such as acetic, propi- onic and n-butyric acid. These are absorbed across the gut wall and provide an additional energy source to the animal. Breakdown of the cell walls in this way also gives further access to the cell contents. Bacteria in the gut confer other benefits as well: they can synthesise new proteins from urea produced by the host, thereby conserving nitrogen and water, and may also denature plant

input: volume x raw leaf rate of passage

Mill (molar teeth)

Digester pH 4 (stomach)

carbohydrates Absorber amino acids re-ingestion of (small intestine) lipids caecal pellets

short chain fatty acids

Sorter/Fermenter pH 8 (caecum)

Dehydrater urea (colon) water

lignin bacterial faeces fibre protein

Figure 7.4: Generalised diagram of the gut of a leaf-eating marsupial with a hindgut fermentation chamber, which can breakdown cellulose to short-chain fatty acids and recycle urea into bacterial protein. The protein can only be used by those species that return it to the stomach by the process of reingestion. After Hume (1999). 226 Life of Marsupials

secondary compounds. However, bacterial fermentation is slow and for the herbivore to obtain sufficient resources by this means the food must be stored in the gut for a relatively long time and must contain little lignin, so that the bacteria can reach the cellulose. Because a large animal has a lower metabolic rate than a small one it can afford to ferment the herbage in a large chamber for a considerable time. Conversely, the gut capacity of a small folivore (about 1 kg body mass) is proportionately too small to allow it to gain sufficient nourishment for its daily metabolic needs by fermenting plant material. Instead, the small folivore breaks the cell walls and reduces the particle size by fine chewing and then selectively excretes the larger particles. In this way it obtains access to a larger proportion of the cell contents and does not rely on bacterial fermentation to breakdown the cell walls further. To do this it must eat more food: but the quantity that can be processed is limited by the level of secondary compounds, mentioned above, and by the rate of passage through the gut, because the harsh Eucalyptus leaf abrades the gut wall with consequent loss of tissue nitrogen. For these reasons it is theoretically impossible for a mammal of less than about 1 kg body mass to meet its metabolic needs of energy and protein nitrogen solely from a leaf diet. The smallest exclusively leaf-eating mammals, such as lemurs, flying squirrels and possums, weigh about 1 kg: those that are smaller than this, feed on insects, pollen, nectar or plant exudates (see Fig. 6.2). Paradoxically, arboreal herbivores that are larger than about 2 kg cannot make effective use of insects, pollen and nectar because the energy expended in searching for them exceeds the benefit to be derived from them (see Fig. 6.15). The starkest contrast is between the common ringtail possum (Fig. 7.5b, Plate 12), which at 700–900 g is the smallest leaf-eating marsupial, and the yellow-bellied glider, Petaurus australis (see Fig. 6.1f, Plate 11), which at 500–700g is the largest sap feeder (see Chapter 6). The marsupial leaf eaters to be discussed in this chapter are all hindgut fermenters: they have small stomachs but all have a large caecum, and the koala also has an expanded proximal colon in which bacteria are present (Fig. 7.4). A modest amount of cellulose digestion occurs in the hindgut of these species, as well as some protein synthesis and denaturing of secondary compounds. Hence, the species that live on Eucalyptus leaves show adaptations of their digestive system, their metabolism, their behaviour and their reproduction that enable them to exploit this abundant but special food supply.

The folivore–eucalyptus arms races The prevalence and relatively high levels of secondary chemicals may have evolved in Eucalyptus as a protective response to the browsing pressure of insects and arboreal marsupials. The earliest relatives of the koala, ringtail possum and brushtail possum made their appearance at the end of the Oligocene, 26–23 million years ago (see Fig. 6.3), about 10 million years after the first species of Eucalyptus appeared. From their dentition these animals were already adapted to eating leaves: they had high crowned molar teeth for cutting and crushing leaves and a well-developed gap, or diastema, between the molar teeth and the incisor teeth at the front of the jaw, which is charac- teristic of plant-eating mammals. From this evidence we may surmise that marsupials, similar in life form to present day leaf- eaters, were already exploiting the abundant food to be found on species of Eucalyptus. What is of particular interest is that these fossils are recognisably members of separate families, so that they must have already followed separate evolutionary pathways to become leaf eaters. From the differences in the way that the arboreal marsupials use Eucalyptus leaf as food today adaptations to exploit Eucalyptus as a food resource probably evolved independently several times. At the same time the various species of Eucalyptus evolved an array of defence mechanisms in response to herbivory, which place an additional metabolic burden on the folivores. Life in the trees: koala, greater glider and possums 227

The interaction between a food species and the animal that feeds on it is termed an ‘arms race’ and many such arms races have been described in natural systems. For instance, there are those between large cats and their prey, between infectious organisms and the immune system of host species, and between intestinal parasites and their hosts. In any such arms race we only see the end game in living organisms today and, from this, surmise the ways in which it may have come about. To understand the arms races that have taken place between Eucalyptus and arboreal marsupials we can look at the composition of the leaves and other products of the plants of the several preferred food species: and we can examine the ways different marsupial species use different plants to support them and enable them to breed (Fig. 7.3). Thus, the leaf- eating marsupials today are the culmination of several independent arms races, rather than all being derived from a single ancestral group of arboreal leaf-eaters.

Relationships and past history of arboreal folivores The three arboreal folivore groups that have living representatives are the koala, the common ringtail possums/greater glider group and the brushtail possums/cuscus group. They have been placed into three Families, based on bone and tooth structure, blood proteins, the shape of sperm, and chromosomes (see Chapter 1). All are adapted to climbing in trees with opposable big toes on their feet, and on the hand the thumb and first finger oppose the remaining three digits. The fossil record shows that koalas have been a separate lineage since the Oligocene (26 million years ago), and so have the common ringtail possums, while the direct lineage of the brushtail possum reaches back to the middle Miocene epoch (16 million years ago). Other fossil groups that were probably leaf eaters have left no living descendants. Since the Pliocene epoch (5 million years ago), the common ringtail and the brushtail possum families have radiated into the many present-day species (see Fig. 6.3), with each family undergoing separate radiations in New Guinea and Australia. The conclusions from the fossil record are supported by recent results of DNA/DNA hybridi- sation and DNA sequence data (see Chapter 1), which indicate that the three groups that eat Eucalyptus foliage (Fig. 7.3) have been separated for a long time and each has closer relatives that are not leaf eaters. Thus, the closest relatives of the koala are the wombats, which are grass eaters and have a long separate fossil history; the closest relatives of the brushtail possums are the small Burramyidae, which are seed, sap and insect eaters; and the closest relatives of the ringtail possum/greater glider family are the smaller sap-feeding glider possums of the family Petauridae (see Chapter 6). This is further evidence that adaptations to eat Eucalyptus leaves evolved independently at least three times.

A quick survey of marsupial leaf eaters

Phascolarctidae The sole living representative of this ancient family is the koala, which occurs in open woodland and eucalypt forests of eastern Australia from Queensland to Victoria (Fig. 7.5a, Plate 12). It does not occur naturally in Tasmania or Western Australia, nor does it occur in the rainforests of Cape York or in New Guinea and adjacent islands, where the nearest equivalent species are the , Ailurops ursinus, and tree kangaroos (Dendrolagus see Chapter 9).

Pseudocheiridae The second family has one widely distributed species and several other species with restricted distributions (Strahan 1995). In Australia, the common ringtail possum (Fig. 7.5b, Plate 12) 228 Life of Marsupials

occurs in woodland habitats from north Queensland to Tasmania and across into South Australia, with an isolated subspecies, the western ringtail possum, Pseudocheirus peregrinus occidentalis, in southwest Australia (Fig. 7.6). Four other species, restricted to rain forests in Queensland, are the Herbert River ringtail, Pseudochirulus herbertensis, the Daintree River ringtail, Pseudochirulus cinereus, the lemuroid ringtail, Hemibelideus lemuroides, and the green ringtail, Pseudochirops archeri. The rock ringtail possum, Petropseudes dahli, is restricted to rocky breakaways in north- ern Australia.

Figure 7.6: Distribution of some of the forest folivores in Australia, New Guinea and islands as far west as Wallace’s Line. Note the minor extension from New Guinea into northern Australia of the southern common cuscus, Phlanger intercastellanus, and the common spotted cuscus, Spilocuscus maculatus, but no overlap of species of the Pseudocheiridae. From Flannery and Schouten (1994), Flannery (1995) and Strahan (1995). Life in the trees: koala, greater glider and possums 229

In addition to these non-gliding species, there is the greater glider (Fig. 7.5c, Plate 12), most closely related to the lemuroid ringtail but is distributed in wet sclerophyll forests from north Queensland to southeast Victoria (Fig. 7.6). Unlike the common ringtail possum, the greater glider does not occur in similar forests in Tasmania or Western Australia. Hume (1999) divides the Australian Pseudocheiridae into three groups on the basis of their gut anatomy and diet: the green ringtail and the rock ringtail have a short intestine, a small or simple caecum and a long colon, and their diet consists of some leaf as well as other plant material; species of Pseudocheirus and Pseudochirulus share a similar gut anatomy and patterns of nitrogen concentration in differ- ent parts of the gut (see Common ringtail possum, Nitrogen metabolism); the greater glider and the lemuroid possum have a larger caecum than the previous group, a greater degree of folivory and similar long limbs. Although only the greater glider has functional gliding membranes between the elbow and the ankle (see Fig. 6.7c), the lemuroid possum has rudimentary gliding membranes and makes leaps of 2–3 m in the canopy of its food trees. In New Guinea and adjacent islands another adaptive radiation occurred, where there are nine species of ringtail possum: four species of Pseudochirops and five species of Pseudochirulus, (Flannery 1995), none of which overlaps with the Australian species. The latter species are much smaller than the Australian species and it is unlikely that they can subsist solely on leaves but insufficient is known of their feeding habits.

Phalangeridae The third family includes cuscuses and brushtail possums whose range extends from Tasmania in the south to the limit of Wallace’s line in the north (Fig. 7.6). There are 22 species in the family, the commonest being the brushtail possum, Trichosurus vulpecula (Fig. 7.5d, Plates 12 and 13), in Australia and the northern common cuscus, Phalanger orientalis, in New Guinea. The latter occurs on numerous islands from and Mollucas in the west to the Solomon Islands in the east, almost certainly taken by people more than 10 000 years ago; it was the first Australasian marsupial to be described by European travellers (see Chapter 1). Now the range of this family includes the three islands of New Zealand, where the brushtail possum was first introduced in 1858. Several species of brushtail possum have been named from separate regions of Australia but these are now all considered to be regional varieties of the one species (Kerle et al 1991). They range in body mass from 1.5 kg in northern Australia to 4.5 kg in Tasmania. The closely related mountain brushtail possum or bobuck, Trichosurus caninus, is restricted to the tall Eucalyptus forests of eastern Australia, but not Tasmania or Western Australia; and the scaly-tailed possum, Wyulda squamicaudata, lives in rocky outcrops in north-western Australia (Fig. 7.6). No species of Trichosurus occurs in New Guinea but nineteen species of cuscus occupy equiv- alent niches there and on surrounding islands (Flannery and Schouten 1994, Flannery 1995). Three species are very rare with highly restricted distributions, while two of the commonest species, the southern common cuscus, Phalanger intercastellanus, and the common spotted cuscus, Spilocuscus maculatus (Fig. 7.5e, Plate 13), also occur in the most northerly region of rainforest of York Peninsula (Strahan 1995). The largest species of all is the 10 kg bear cuscus, Ailurops ursinus of Sulawesi (Dwiyahreni et al 1999). Flannery (1995) considers that cuscuses evolved from possum-like ancestors in Australia and, after the separation of New Guinea from Australia in the Pliocene epoch (5 million years ago), they underwent an adaptive radiation in the tropical forests of New Guinea. The two Australian species probably extended south onto the extreme northern tip of York Peninsula during the last time there was a land connection 12 000 years ago. Interestingly, the northern race of the brushtail possum does not extend as far north on York Peninsula, so that there is almost no range overlap between brushtail possums and cuscuses (Kerle 2001). 230 Life of Marsupials

How five species live in trees For five Australian folivore species sufficient is now known about their ecology and physiology for comparisons to be made and so to begin to understand the limitations of a leaf-eating life in the trees and how each species has adapted to it. The koala, the greater glider and the common ringtail possum live exclusively or predominantly on eucalyptus leaf and all three are selective in the species of Eucalyptus and in the type of leaf that they will eat. While eucalyptus leaf is also a major component of the diet of the brushtail possum and the bobuck, they supplement it with other plant material. Cuscuses do not eat eucalyptus leaf but the more nutritious leaves and fruits of tropical rainforest species.

Koala The nearest relatives of the koala are the ground-living wombats and it is probable that koalas evolved from ground-living ancestors (see Fig. 6.3). They will come down to the ground to travel between trees but they feed exclusively on the leaves of several species of Eucalyptus. Although

Figure 7.7: Occlusal view of the 2nd upper and lower molar teeth of the koala, Phascolarctos cinereus, the common ringtail possum, Pseudocheirus peregrinus, and the ground cuscus, Phalanger gymnotis, to show the squaring of the upper molars by the addition of the metaconule and the distinct modifications of the cusps that aid in cutting plant tissue. After Long et al (2002). Life in the trees: koala, greater glider and possums 231

they are able to consume large quantities of eucalyptus leaf, they are selective in the species they will eat and vary their choice at different times of the year. When offered leaf and shoot of various species of Eucalyptus at San Diego Zoo they chose species that had significantly higher concen- trations of crude protein, and significantly lower concentrations of fibre and lignin (Ullrey et al 1981). Similarly, in a Victorian forest the koalas varied their preference between three species seasonally, selecting new foliage from each when the crude protein level was maximal (Hindell et al 1985, Martin 1985). In the past 20 years much has been learnt about the way that koalas manage to live on this diet, and how they acquire their essential requirements of energy (in the form of carbohydrates and fats), digestible nitrogen (in the form of plant protein) and water. The koala is the largest of the marsupial folivores, which gives it an advantage in living on eucalyptus foliage, compared to the smaller species. In southern Australia koalas weigh 9–14 kg, while at the northern limit of their range in Queensland they weigh about 5 kg. Being large, their metabolic requirements are relatively low and the volume of the gut relatively large, which means that the throughput of chewed leaves is more than sufficient for its nutritional and water needs.

Processing its food Koalas have a battery of high crowned, ridged molar teeth (Fig. 7.7a), which cut the leaf into small pieces but do not grind it further. Animals in middle life, when the teeth are partially worn, break the leaf tissue most effectively into small particles, so releasing more of the cell contents for digestion in the stomach and absorption in the relatively short small intestine. Koalas gain 90% of their energy requirements from digestion of cell contents in the stomach and absorption in the small intestine: they do not, therefore, rely on bacterial fermentation of cell wall cellulose with the release of SCFA from the hindgut. This is surprising because the hindgut of the koala is proportionately the largest fermentative chamber of any herbivore. While the large caecum and proximal colon, which comprise the hindgut, contain large quantities of anaerobic bacteria capable of anaerobic fermentation, and movement through the gut is very slow (100 h for solutes and 32 h for particulate matter (Cork and Warner 1983, Krockenberger 1993), the actual amount of SCFA that result from breakdown of cell walls provides less than 10% of the koala’s energy requirements (Cork and Hume 1983). What then is the function of the very large caecum and proximal colon? One idea is that fine particles and solutes are selectively retained in them, while indigestible fibre is cleared quickly, which allows faster throughput of fibre and access to more of the digestible cell contents. Another function may be the retention of water in the gut: koalas have a relatively high level of body water at 77%, much of it contained in the caecum (Degabriele et al 1978).

Water balance The water turnover rate of koalas per day ranges from 71 to 91 mL/kg0.71, which is less than for the ringtail possum and brushtail possum (Table 7.1). This suggests that either koalas live in a habitat where water is always available, or they are very good at conserving the water they consume: this comes from water in the food, metabolic water produced during oxidation of food and free water. Female koalas are able to obtain all their water needs from plant material in the trees but large male koalas supplement this by drinking water on the ground. In trees the only free water is dew, rain on leaves or amounts collected in tree hollows, so koalas depend on the nutritional composition and palatability of the leaves eaten. Indeed, when offered foliage of varying water and nitrogen content koalas selected leaves with the highest water and nitrogen content (Pahl and Hume 1990). Water is lost from urine, faeces and by evaporative cooling for thermoregulation. By storing water in the caecum and by passing relatively dry faeces high in undigested fibre, koalas can conserve water and, hence, maintain a low water turnover rate. 232 Life of Marsupials

Table 7.1: Comparisons of diet, digestibility and maintenance nitrogen requirement (MNR) in four arboreal folivores, feeding on Eucalyptus foliage, and the southern hairy-nosed wombat, Lasiorhinus latifrons, on straw Data from Chilcott and Hume (1984), Cork (1986), Cork and Sanson (1990), Foley and Hume (1987c), Hume (1999), Krockenberger (2003).

Parameter Unit Koala Brushtail Greater Ringtail Wombat possum glider possum Body mass kg 9.2 2.5 1.1 0.7 25 Diet E. punctata E. melliodora E. radiata E. andrewsii straw Nitrogen percentage 1.1–1.5 1.6 1.9 1.1 content dry weight Dry matter g/kg0.75 per 41 [54]A 36 44 41 [88] A 33–40 intake day Dry matter percentage 55 33 57 58 41 digestibility Mean fluid, hours 100 51 51 70 49 retention particulate, h 32 49 23 37 69 MNR, truly mg N/kg0.75 271 420 560 290 [620]B 116 digestible per day Water mL/kg0.71 71–91 100–117 84–92 98–159 61–86 turnover per day

AIntake during late lactation. BRequirement without reingestion of caecal contents.

Conservation of water becomes important for koalas only at ambient temperatures above 30°C, when evaporative cooling is needed for thermoregulation. Below this temperature, metabolic water adequately compensates loss by evaporation, because the fur of the koala is highly reflective and so acts a good insulator. In Queensland, with ambient temperatures in summer of 40°C, koalas select leaves with a moisture content higher than leaves selected in the winter (Ellis et al 1995). This may be an important factor in defining the northern boundary of koala distribution in north Queensland, where the ambient temperature is high and high humidity makes evaporative cooling inefficient.

Nitrogen metabolism Among arboreal folivores the daily nitrogen requirements of the koala at 271 mg N/kg0.75 is half that of the greater glider and brushtail possum (Table 7.1) and is easily met by its daily dietary nitrogen intake (520 mg N/kg0.75, Cork 1986). Two factors may be involved in this extraordinar- ily low nitrogen requirement: first, its very low metabolic rate means that its requirements are low and second it may gain some protein nitrogen from the bacteria in its large hindgut, thereby reducing excretion of nitrogen as urea. Although there is no direct evidence that adult koalas can use bacterial protein, if they can do so this would not only conserve nitrogen but also save water otherwise needed to excrete urea in the urine.

Energy metabolism The SMR of the koala at 161 kJ/kg0.75 per day is very low, being only 55% of the standard for placental mammals (see Fig. 1.3) and, hence, is lower than most other marsupials (Degabriele Life in the trees: koala, greater glider and possums 233

and Dawson 1979), and the FMR of males and non-lactating females is only between 2 and 2.8 times SMR (Ellis et al 1995) (Table 7.2). For this reason, koalas can obtain all their energy needs from non-cell wall constituents of eucalyptus leaf (Cork et al 1983).

Table 7.2: Daily energy metabolism (kJ/kg0.75) in four arboreal folivores Data from Nagy and Martin (1985), Green (1997).

Koala Brushtail Greater Common possum glider ringtail possum Body mass (kg) 9.2 2.5 1.1 0.7 Gross energy intake 740 886 Digestible energy 340 436 [182]A Metabolisable energy 260 345 [91]A SMR 161 188 210 266 FMR, males 330 (2.0) 527 (2.5) 646 (2.4) FMR, NLB 444 (2.8) 515 (2.5) 575 (2.2) FMR, L3C 763 (2.9)

AWithout reingestion of caecal contents. BNL, non-lactating or early lactation females. CL3, females in lactation phase 3; FMR, fi eld metabolic rate; SMR, standard metabolic rate. Figures in parentheses represent the metabolic scope between SMR and FMR.

Figure 7.8: Dorsal views of the brains of the koala, Phascolarctos cinereus, and the , Vombatus ursinus, at the same length, to compare the unusually small cerebral hemispheres and cerebellum of the koala with the more normal anatomy of the wombat. Photos courtesy of John Nelson. 234 Life of Marsupials

However, the key issue for the koala, as for the greater glider and common ringtail possum, is the metabolic cost of late lactation, including the export of milk with a high nitrogen and energy content, which reaches 400 kJ/day in late lactation (Krockenberger et al 1998). This is the period of greatest investment in reproduction of any species of marsupial, and folivores are no excep- tion: they must consume enough high nutrient eucalypt leaf on which to maintain late lactation and produce independent young. Another unusual feature of koalas, which may be associated with their very low metabolic rate, is their disproportionately small brain (Haight and Nelson 1987). Whereas the brain volume of most mammals is about equal to the space in the skull where it is housed, the koala’s brain occupies only 61% of the available volume. The cerebral hemispheres have a smooth surface, which is unusual in an animal of its size and are so small that they do not meet in the mid line (Fig. 7.8). Flannery (1994) has suggested that its unusually small brain size is an adaptation for conserving energy, since brain has the highest energy requirement of any tissue in the body. The trade off is a reduction in the ability of the koala to carry out complex behaviour of an unfamil- iar kind. For instance, the feeding repertoire of the koala is to draw a small branch to its mouth, sniff it carefully and, if acceptable, begin to eat the leaves from the branch. If, however, plucked leaves of the same species are presented to the koala on a flat surface, it is unable to respond to the unusual situation and will not eat the leaves. In their natural environment this rigid behav- iour may not be a serious disadvantage, compared to the advantages of energy conservation.

Behaviour and ecology The social life of the koala reflects its low metabolic rate (Mitchell 1990). For about 20 h of the day a koala rests or sleeps in the fork of a tree. It feeds for two to three hours, mainly at night, and moves about and reacts to other koalas during the remaining hour. Within the preferred habitat an adult koala lives for long periods in a few trees within a small area, most of the time alone. While home ranges overlap, there is little interaction between resident koalas, except at mating. Females have smaller home ranges than males in the same habitat and are tolerant of other females. However, young koalas are more mobile, dispersing from the natal area at one to two years. Among males there is a dominance hierarchy, with dominant males excluding younger males from their area. They mark their territory by rubbing trees with a secretion from glands on their chest and signal their presence by loud bellowing. Holding a home range that overlaps the home ranges of several females gives the dominant male greater access to females but also provides an optimum habitat in times of stress. In Queensland Gordon et al (1988) showed that dominant koalas survived drought and heat wave conditions better than subordinates and young animals. Their study was conducted near the northern limit of the koala’s range, in mulga–poplar box woodland and river red gums, Eucalyptus camaldulensis, favoured by the koalas. In January 1980 there was a severe drought and high temperatures (over 40°C) and the eucalypt trees lost their leaves. Because foliage is their main source of water this deprived the koalas of food and water and 63% of the popula- tion died, mortality being most severe in young koalas, which were excluded from optimal sites by older dominant koalas. Trees near water holes retained their foliage and their resident koalas survived in good condition, but subsequent recovery of the koala population was slow. While such dramatic population crashes may be rare, they may be an important factor in determining the long-term distribution of the species.

Reproduction of koalas The breeding strategy of koalas is consistent with its slow pattern of life: females begin to breed at three years of age; the single young is born in summer; it is weaned at one year and continues to associate with its mother for another year (Table 7.3). Life in the trees: koala, greater glider and possums 235

Table 7.3: Reproductive data on arboreal folivores Data from: Martin and Handasyde (1999), How (1976), Dunnet (1964), Thomson and Owen (1964), Tyndale-Biscoe and Renfree (1987), Tyndale-Biscoe and Smith (1969a), Smith (1969).

Parameter Koala Brushtail Bobuck Ringtail Greater possum possum glider Young/female/year 0.5 1.3 0.5 1.7 0.7 Pouch life, days 240–270 145 175–200 120 120 Length of lactation, days 360–380 230 275 210 240 Sex. maturity, &&, yr 3 1-2 2–3 1 2 1st yr mortality, (( 0.83 0.57 0.68 0.40 1st yr mortality, && 0.50 0.64 0.68 0.20 Adult sex ratio,% (( 40 50 48 39

Krockenberger (1996, 2003) and Krockenberger et al (1998) have investigated the cost of reproduction for koalas by determining the composition of milk at different stages of lactation and by measuring the FMR of the mother and the young through development. As in other marsupials the milk changed in composition through lactation, with a proportionate increase in lipid and protein and maximal energy content of the milk in the last phase of lactation. However, unlike in terrestrial marsupials, such as tammars, dasyurids and the bandicoot, Isoodon macrou- rus, the lipid content plateaus when the young koala has left the pouch and is beginning to supplement its diet with foliage (Fig. 7.9a), and this phase is longer than in terrestrial marsupials. The pattern in the other leaf-eating marsupials is similar and Krockenberger (1996) thinks it is an adaptation to the low nutrient quality of the diet. The female koala obtains the additional resources for synthesising and secreting the concen- trated milk, and carrying the young, by increasing her food intake by 27% and her FMR by 40%. But she also reduces her home range, concentrating her feeding on trees with the highest nutritional content and lowest phenolic content, and possibly saves energy further by holding the young close to her body, thereby reducing heat loss from both bodies. The young koala first puts its head out of the pouch at about 180 days and soon after it begins to eat eucalyptus leaves. It leaves the pouch permanently at about 260 days when it moves to its mother’s back, and is weaned off milk at about 380 days. During the time when it is in and out of the pouch it feeds on a special faecal pap, which it takes from its mother’s cloaca. This material is quite unlike the normal faeces and resembles the contents of the caecum, including a high concentration of bacteria (Osawa et al 1993). It has long been thought that its function is to infect the gut of the young koala with bacteria and thereby facilitate digestion of Eucalyptus foliage. However, since bacterial digestion of cellulose only accounts for 10% of the adult koala’s energy requirements, this seems an unlikely explanation for the behaviour. A more important function may be to provide a supplementary source of protein, from bacterial protein in the caecum, at the critical period of changeover from milk to a leaf diet. Whereas the adult koala cannot digest bacterial protein produced in its hindgut, as mentioned earlier, the young one could do so in its stomach. Common ringtail possums almost double their access to leaf protein by taking material from the caecum back to the stomach (see Common ringtail possum, Nitrogen metabolism). The growth rate of the young koala increases soon after it begins to take pap, which it does for about one month, before progressively coming to rely on a leaf diet. Pap may have the same function in koalas as late lactation protein does at the time of weaning in kangaroos and wallabies (see Chapter 2). 236 Life of Marsupials

40 a

30 Total solids

20 Lipid 10 Protein

Carbohydrate 0 150 200 250 300 350 400 40 b

30

20

10

Percent milk consumption 0 0 30 60 90 120 150 180 40 c

30

20

10

0 30 60 90 120 150 180 210

Age of young (days) Figure 7.9: Relative constituents of the milks of three forest folivores through lactation. (a) the koala, Phascolarctos cinereus after Krockenberger (1996); (b) the brushtail possum, Trichosurus vulpecula, after Cowan (1989a); (c) the common ringtail possum, Pseudocheirus peregrinus, after Munks et al (1991). Arrows indicate the age when young leave the pouch in each species. Compare with more detailed treatment of the tammar wallaby, Macropus eugenii in Figure 2.19.

Some consequences of koala reproduction Because of the unusually long dependent period of the young, female koalas tend to breed in alternate years (Martin and Lee 1984), although in favourable circumstances, such as abundant high quality food trees, they can breed every year. Female koalas have a life expectancy of about 15 years, males somewhat shorter, so that their potential lifetime fecundity is high, although not always realised. However, when it is realised it can be disastrous: several cases have been recorded in Victoria where the wholesale defoliation of preferred food trees has led to widespread death of the koalas and decline of the population. Martin and Handasyde (1999) discuss the causes of these events in their excellent book The Koala. They point out that the potential fecundity and longevity of the females can lead to a Life in the trees: koala, greater glider and possums 237

very high rate of increase if not matched by juvenile mortality. In Victoria, when koalas were introduced to islands that previously had no koala population but did have their preferred food trees, the populations outgrew the resources in 20–30 years. However, if the founder population was infected with the genital fungus Chlamydia, which made females infertile, the population remained small and defoliation and death of the trees did not occur. In the light of this, the question has been asked why koala populations in other parts of Australia have not destroyed their habitat by over browsing? In Queensland, as mentioned above, extreme conditions in summer may reduce the survival of young koalas to such an extent that the rate of increase is curbed. However, another factor may be the lifting of human hunting pressure. At the time of European settlement, Aboriginal people were hunting koalas for food, and early observers noted that koalas were rare. As the Aboriginal people were displaced by European settlers and ceased to hunt, koalas increased and, until the beginning of the 20th century, they were hunted for their pelts. For instance, in 1920–21 over 205 000 pelts were sold in the Sydney market. This trade was stopped in 1925, since which time koala populations have increased and defoliation of some forests occupied by them has occurred. If this is a correct interpretation of events, it suggests that the proper management of koala populations, which are over brows- ing their food trees, is to cull them. But regulatory authorities will not countenance this, being mindful of hostile publicity, especially from outside Australia.

Koalas on Kangaroo Island Six koalas from Victoria were released on the western end of Kangaroo Island in 1923 and a further 12 two years later, at the time when there was a perception that the species might soon be extinct. By 1960 the population had eaten and destroyed most of the manna gums, Eucalyptus viminalis, on which the koalas depend, and some koalas were translocated to the only other stand of manna gum on the island at Cygnet River. By 1996 the total koala population was estimated to be more than 5000 and the existing manna gums and other food trees were so over browsed that they would die in the next few years. It was predicted that there would be widespread starvation and death of koalas, as occurred at several sites in Victoria (Martin 1997). Thus, it had become urgent that the population of koalas be reduced to less than 1000 animals and held at this level thereafter. But the State authorities explicitly refused to allow this to be done. Landowners on Kangaroo Island are allowed to destroy wallabies and kangaroos, which are native to the island, but are forbidden to exercise the same control over the exotic koalas. This is because the koala is a national icon beyond the bounds of ordinary species. Being unable to recommend culling, the Koala Rescue Strategy in 1997 began to translo- cate some koalas off the island, and surgically sterilise some of the resident population. Fertility control is now being implemented as the main plank in the Strategy, at a cost of $140 per animal. While it may seem to be a humane means to solve a difficult problem, it will not reduce brows- ing of manna gum and other species for many years to come. Even if most of the koalas are successfully sterilised, this will not begin to affect the size of the population until other mortality factors come into play. Since sterilisation may have the effect of increasing the life expectancy of females, this could be many years away. Barlow et al (1997) modelled the different effects of sterilisation, culling and a combination of both, on brushtail possums in New Zealand and concluded that sterilisation by itself would not lead to a long-term reduction in the population for at least 15 years. However, if combined with an initial 80% cull of the population, sterilisa- tion would maintain a small population for much longer than by culling alone. Possums and koalas have similar life expectancy and reproductive potential, so these models have consider- able relevance to koalas on Kangaroo Island. They suggest that fertility control will not reduce the current population in the short term, but in the longer term could be useful in preventing a 238 Life of Marsupials

small, sustainable population from increasing. In other words, it will only be effective in combi- nation with substantial population reduction, either by culling or translocation. Translocation would be effective if up to 4000 koalas could be removed from Kangaroo Island in the immediate future, but no one has contemplated that level of interference. What is envis- aged are a few koalas being translocated to southeast South Australia, where koalas already occur. This level of translocation will have no benefit on Kangaroo Island. Thus, the current policy will see the continuing loss of vegetation and the progressive decline of the health and condition of the koala population. This sequence is not new. To paraphrase Charles Elton (1942): There is dismay, followed by outcry, and demands to authority. Authority remembers its experts or appoints some: they ought to know. The experts advise a cure. Authority rejects the cure. Disaster follows.

Conclusion The koala displays special adaptations for living on the leaves of Eucalyptus. Its size allows it to consume large quantities of leaf and its special dentition allows it access to the cell contents, which provide it with sufficient energy and nitrogen to sustain its exceptionally low metabolic rate. Its very large caecum and colon provide a mechanism for separating the fine particles from the coarse and retaining the former for further use, and as a means of conserving water. Its exceptionally low metabolic rate, behaviour and reproductive strategy are other adaptations for a life sustained by Eucalyptus foliage. However, its reproductive potential and relatively long life can be disastrous for the forests, and the koala population, if not curbed by mortality of the young or by external predation. This now poses an acute problem for the management of this iconic species.

The greater glider The greater glider (Fig. 7.5c, Plate 12) is the commonest arboreal species in the tall Eucalyptus forests of eastern Australia. The southern form weighs 1–1.7 kg and is variable in colour from dark brown to creamy white, while the northern form weighs 0.7–0.9 kg, is less variable in colour and has shorter ears and tail (Comport et al 1996).

Diet Greater gliders live in hollow spouts of tall Eucalyptus trees and feed exclusively on the foliage of these trees. Night-time observations of free-living greater gliders has confirmed that they do not feed on any other components of the trees, such as gum, sap or manna, and insect remains are not found in their stomachs (Kavanagh and Lambert 1990). In captivity, gliders have maintained body weight for three years on a diet of fresh Eucalyptus radiata foliage alone (Hume et al 1984, Foley and Hume 1987b). The diet is, thus, very similar to that of the much larger koala, although the smaller body mass is at the theoretical lower limit for a species to live and reproduce on Eucalyptus foliage. How do greater gliders obtain sufficient protein from the leaves and how do they obtain sufficient energy in the form of soluble carbohydrates from the plant cells? Like the koala, the greater glider is very selective in the species of Eucalyptus it will eat at any particular time of the year. As mentioned earlier, greater gliders are not uniformly distributed through the tall forests but are found only in 10% of the forest that contains nine particular Eucalyptus species. In these forests greater gliders feed only on the foliage of the Monocalyptus group with the highest nutritional content, particularly Eucalyptus viminalis, and Eucalyptus fastigata (Kavanagh and Lambert 1990). This is most marked in summer when these species put out new shoots, which the gliders select. This is the time of the year when the young are being weaned and require an abundant supply of nutritious food for growth. At other times selection Life in the trees: koala, greater glider and possums 239

for these two species is less marked and the gliders forage in other species, such as Eucalyptus radiata, Eucalyptus obliqua or Eucalyptus cypellocarpa, that are growing new shoots. As well as selecting the most nutritious leaves, the molar teeth of the greater glider have sharp transverse crescentic ridges, like those of common ringtail possums (Fig. 7.7b) with which it can cut the leaves into very small particles, finer than the koala does. It also has a large, sacculated, caecum in which the coarse indigestible fibres are separated and excreted relatively rapidly from the hindgut (in 23 h), thereby reducing the gut filling effects of its high fibre diet. At the same time the fine particles and fluid contents of the food are retained in the caecum for up to 51 h (Table 7.1). Here the cell walls are further broken down by bacterial fermentation, releasing more of the cell contents, and converting the cellulose to SCFA. However, as in the koala, SCFA provide less than 10% of the energy requirements of the animal (Hume et al 1984, Foley and Hume1987a) because the high content of phenolics in Eucalyptus foliage inhibits fermentation and so reduces the available carbohydrate from this source (Foley et al 1989). Another metabolic cost to the greater glider is detoxifying the pungent oils, or terpenes, that are such a conspicuous component of the leaves that it feeds on. These readily pass into the blood stream and can potentially destroy cells of the body. However, less than 1% gets into the general circulation because most are detoxified in the liver and excreted in the urine. The phenolics are detoxified by conjugation with glucuronic acid and the terpenes are converted into polyoxygen- ated compounds. As a result of these necessary processes less than half the gross energy ingested in the foliage is available as useful energy for the glider (Fig. 7.1b) (Foley 1987).

Nitrogen metabolism Greater gliders fed on Eucalyptus radiata require more protein nitrogen than other folivores, including the brushtail possum (Table 7.1) (Foley and Hume 1987c): their daily maintenance nitrogen requirement (MNR) is almost twice that of the common ringtail possum and koala fed on similar foliage. While greater gliders recycle much of the urea into bacterial protein, they also use nitrogen in the form of ammonium to neutralise the strong organic acids resulting from detoxification of the essential oils to labile glucuronides.

Energy metabolism The greater glider has little margin to satisfy either its nitrogen or energy requirements. The question then is how does it use the residual energy to sustain its daily life? Unlike the koala, its SMR is within the normal marsupial range for its body mass (Table 7.2), and its daily maintenance energy requirement of 350 kJ/kg0.75, is about 1.7 times the SMR. Foley et al (1990) examined its balance sheet by measuring the energy and water metabolism of 3 18 free-living greater gliders, using the technique of doubly labelled water ( H2O, H2 O, see Box 1.1). Although water is not readily available in the canopy to drink, greater gliders have a water turnover rate similar to animals with a ready supply of water. This suggests that they obtain water, like koalas do, from tree hollows, the surface of leaves and from water in the cell contents, and as metabolic water from the oxidation of sugars and lipids. The FMR of male gliders, measured by doubly labelled water turnover, was 2.5 times SMR, similar to the koala and the common ringtail possum (Table 7.2). However, as already mentioned, this figure represents less than half the gross energy in the leaves consumed, 52% being lost in faeces and urine after detoxifying secondary chemicals in the leaves. How they use the balance of metabolisable energy for their daily activities is shown in Fig. 7.1b. Standard metabolism accounted for 39% and a further 7% was heat generation for thermoregulation. This assumes that the tree hollows that the gliders occupy during the day are well insulated against high air temperatures and the fluffy pelage provides good insulation 240 Life of Marsupials

against low temperatures during the active periods at night. A further 6% would be used in digestive processes, such as gut motility and fermentation, while 23% is required to hold onto branches, to chew and to groom during the night-time feeding. Greater gliders expend very little energy (2%) moving about, partly because they can travel between the canopy of adjacent trees by gliding, an adaptation seen in arboreal folivores in other parts of the world. The main cost during travel is regaining height after a glide from one tree to another: males visit about 22 trees each night and females about 14. This leaves a residual balance of about 20%, some of which is due to variation among individuals, some to errors in comparing laboratory and free-ranging gliders, but some is the energy required to maintain social interactions between animals. Greater gliders are at about the minimum size for an animal subsisting exclusively on Eucalyptus leaves and it is clear from the analysis of their energetics that they are only able to live on this diet by leading a slow life. The cost of reproduction was not included in the study by Foley et al (1990) but by analogy with the common ringtail possum (see The cost of lactation in ringtail possums) the cost of late lactation would probably increase the FMR by 30–40%. Because only half the food intake is useable, this translates into an increase in leaf intake of 60–80%, or selection of leaves with higher protein and energy content and/or lower phenolics and terpenes. For a species living so close to the limits of sustainability the nutritional quality of the food, both its energy content and nitrogen content, are critical to survival. The additional cost of late lactation must, therefore, be the main factor that determines the distribution of the species in Eucalyptus forests and why the distribution in the forests is so patchy. What are the behavioural and ecological adaptations of greater gliders in this regard?

Ecology and social behaviour In their preferred habitat of tall Eucalyptus forests of south-eastern Australia, greater gliders are evenly spread through the forest at a density of about 1/ha in the south and up to 4/ha in north Queensland (Tyndale-Biscoe and Smith 1969a, Comport et al 1996). They generally live alone except during the brief highly synchronised breeding season in April–June when the single young is born. Young lost prematurely are not replaced and there is no second peak of breeding because the males are no longer producing sperm (see Chapter 2). The young leave the pouch in October and disperse at the end of the year. The prevailing sex ratio of the adult population is 39% male (Table 7.3), which is imposed on the population at the time of dispersal, when a large proportion of the young males disappear, presumed dead. More interestingly, the number of females with pouch young is about the same as the number of adult males, so that there is a pool of non-breeding females. This is because gliders form monog- amous pairs (Henry 1984, Kehl and Borsboom 1984). In both studies the home ranges of adult females did not overlap in the forest but those of males were larger and overlapped the home range of one or two females, depending on the quality of the forest. Where conditions were good the male covered the home range of two females, otherwise only one, which suggests that monogamy is facultative rather than obligatory. Henry (1984) also observed that weaned young were most vulnerable to aggressive encounters from resident males but young females continued to share dens with their mothers, a feature also found in brushtail possum and bobuck populations. This combination of monogamy and selective loss of male pouch young may be an impor- tant adaptation that moderates the rate of increase of the population under stable conditions and prevents the over exploitation of the forest that can occur in koala populations. Conversely, if forest fire or predation should decimate the population, more male juveniles would survive, a larger proportion of females would breed in subsequent years, and the rate of increase rise. Spotted tailed quolls, Dasyurus maculatus, and powerful owls, Ninox strenua, are important predators of greater gliders: at a site in southern New South Wales Kavanagh (1988) observed Life in the trees: koala, greater glider and possums 241

a pair of powerful owls reduce a glider population to one-tenth of its former size in four years (Fig. 7.10). He calculated that the 100 ha study area provided 13% of the annual requirements of the pair and that after they had depleted the area, they moved on to other parts of the forest and would not revisit for several years.

140

120

100 arrival of pair of powerful owls 80

60

40

20 Number of greater gliders / 100 ha gliders Number of greater 0 1981 1982 1983 1984 1985 1986 1987

Figure 7.10: Decline in the density of greater gliders, Petauroides volans, after 1983, when a pair of powerful owls, Ninox strenua, hunted in the forest. From Kavanagh (1988).

Response of greater gliders to habitat loss Populations of greater gliders are able to recover from predation and natural calamities such as forest fires, but how do they respond to the more profound changes brought about by logging and clear felling for pine plantations? In the Buccleuch State Forest, near Tumut, NSW, we were able to determine how they responded to the loss of their habitat by following the subsequent fate of animals marked when their home trees were pushed over (Tyndale-Biscoe and Smith 1969b). During the five years of the study 1105 greater gliders were handled but only 40 were harmed at tree fall, because most glide free as the tree goes over. However, less than one-quarter of those released were ever seen again, most during the next eight days. By then they had lost weight and, if female, had lost their pouch young. In the subsequent year only 6% of the gliders that had been marked in the previous year were recaptured, predominantly in the Eucalyptus forest adjacent to the area felled the previous year. While almost all gliders survived tree fall very few survived the next week, unless their home range extended into the forest that was not felled. It was not clear whether they could not survive because they were killed by predators – power- ful owls, quolls, foxes or wedge-tailed eagles, Aquila audax – or had moved into adjacent forest and been unable to find empty refuges there. When an adjacent block of forest was depleted of gliders before felling began, so that there would be unoccupied sites for displaced gliders to move into, the number of them later recovered from this block was no greater. So it is not a lack of unoccupied den sites that prevents the gliders surviving but rather that they are unable to move to and occupy strange habitat. Because greater gliders are living so close to the limit of their resources the total disruption of their habitat, especially loss of den trees and their prime food 242 Life of Marsupials

source, would very quickly put them into negative protein and energy balance. With little or no fat reserves they would be physiologically unable to survive for more than one week and would not have the energy to travel long distances to other food trees. Kavanagh and Wheeler (2001) tested this by fixing radio transmitters to nine greater gliders and tracking them for 11 months in two areas of forest, one of which was logged half way through the study. Each glider used up to three den sites and the average home range was 1.8 ha for males and 1 ha for females. In the logged area the home ranges of the greater gliders were reduced to the unaffected parts of their former home ranges and the remaining den sites: none moved into other unlogged forest. While the greater glider has physiological limits to leaving destroyed habitat, common ringtail possums and bobucks also show strong site attachment and will not move to unfamiliar habitat when their own habitat is destroyed.

Survival in small patches When the clear felling occurred near Tumut in the 1960s some small patches of native Eucalyptus forest were left. Thirty years later these are now isolated patches of original forest, varying in size from 1 ha to 25 ha, inside a large monoculture of mature pine trees. Since greater gliders cannot live in pine forests and probably cannot travel more than a few kilometres through such forest, the interesting question is whether they have survived in such small areas of suitable habitat, as isolated populations for 30 years. Since the density in the original forest was about 1 adult per hectare, even the largest patches could not expect to carry more than a few gliders, which on theoretical grounds would be unlikely to be viable over a period that represents about six generations. Lindenmayer et al (1999a) have found that small numbers of greater gliders still occupy the larger patches and the densities of common ringtail possums are actually greater in the mixed forest than in the nearby native forests. Using microsatellite DNA sequences, Lindenmayer and Taylor (2003) have found that greater gliders in some of the oldest and most isolated patches have lower genetic diversity and are more closely related to each other than to gliders in native forests more than 10 km distant. More interestingly, they are also distinct from the original population that lived in the forest before it was clearfelled in 1966, which suggests that the gliders in the patches are direct descendants of the original population and have suffered genetic erosion through inbreeding for five or six generations. This conclusion is based on an analysis of micros- atellite DNA extracted from 50 study skins of the greater gliders, which had been collected at the time of clear felling before, and were lodged in museums around the country. These specimens were collected long before there was any thought that such a genetic study could be undertaken, and illustrate rather nicely the great importance of museum collections that are properly looked after.

The concept of dependency on forest habitat The koala and greater glider depend wholly on Eucalyptus leaf for their diet, and the bobuck also depends on the forest habitat. These species are limited to habitat where their preferred tree species occur, whereas other species with less restricted diet, such as the brushtail possum and the common ringtail possum, can be found in a wider range of habitats. Nevertheless, all five species depend on the quality of the foliage and this, in turn, depends on the quality of the soils on which the trees grow. Not surprisingly, the best quality timber species occur on the best soils and these are chosen for logging. Similarly, most other areas, where highly fertile soils are accessible, were cleared for farming in the 19th century and many lost their wildlife populations then. For example, in October 1883, William Caldwell came to Australia to collect embryos of the koala and other marsupials. At the invitation of the owner, he visited Manar Station in the southern tablelands of New South Wales, where koalas ‘were Life in the trees: koala, greater glider and possums 243

exceedingly numerous’. He reported (1884) that he obtained ‘nearly 100 embryos in all stages from the unsegmented ovum onwards’. To get that series he must have shot many more than 100 koalas. Likewise, Mick Gowen (Gowen 1978) as a boy in the early 1900s, walked four miles to school in Braidwood through thickly wooded country, which carried a large population of koalas. Today there are no large eucalyptus trees at Manar or around Braidwood, and there are no koalas. The forests that are now left on fertile soils are the prime sites for timber production or, if on private land, prime sites for clearing for additional farmland. A major concern for the long-term conservation of native wildlife is that most of the remaining prime forest habitat is in private hands. In their study of land use in the Bateman’s Bay forests, Braithwaite et al (1993) found that 19% of forests were in National Parks, 48% in State Forests and 32% were on freehold land. However, when the productivity of the forest was considered, 80% of the remaining highly productive forest was in private hands and less than 1% was in National Parks. Their conclu- sions were that remaining forests on the most productive soils continue to be at the greatest risk from intensive exploitation and clearing, with a concomitant loss of fauna: the greatest threat to the survival of forest-dependent species now comes from decisions made by private landown- ers, rather than the State forest authorities. The only implement available to regulate this was the control of export licences by the Commonwealth Government but in 1995 that power was relinquished to the Australian States.

Conclusion The greater glider is exquisitely adapted to the special environment of the tall Eucalyptus forests of eastern Australia, where it is the most abundant species. But it can live nowhere else. This is reflected in its distribution, which excludes the same tall forests in Tasmania and in Western Australia, presumably because the intervening habitat of the former land bridge in Bass Strait and the Nullarbor plains never carried forest in which the species could live. Likewise the north- ern limit of its distribution is set where Eucalyptus forest is replaced by rainforest. Its nearest relative, the lemuroid ringtail, which has a more varied diet and is not dependent on the foliage of Eucalyptus trees, occupies the rainforest. The down side of the greater glider’s total depend- ence on tall eucalypt forests is that it is very vulnerable to forestry activities.

The common ringtail possum The common ringtail possum, Pseudocheirus peregrinus, is the smallest of the arboreal folivores that have been closely studied and it is below the theoretical lower size limit for a species that relies solely on foliage for its diet. Like the koala and the greater glider it also shows a strong preference for eucalyptus foliage but will take foliage, flowers and fruits from other species of trees and shrubs in the lower layers of the forest (Thomson and Owen 1964). Nevertheless, the highest densities of the common ringtail occurs where the preferred Eucalyptus species are found. Common ringtails select young leaves over old ones and in one study (How et al 1984) the young emerged from the pouch at the time of peak flush of plant growth and the flowering and fruiting of tea-tree, Leptospermum. Young eucalypt foliage contains higher concentrations of nitrogen, representing protein, and lower cell-wall concentrations than older leaves, but also contains higher concentrations of tannins (phenolic fraction), which make the protein less avail- able (Cork and Pahl 1984) (Fig. 7.1a). Adaptations for a diet of Eucalyptus are similar to those of the two other species already considered. The molar teeth of common ringtails are similar to those of the koala and greater glider in having the four major cusps shaped into crescentic blades (Fig. 7.7b), and with them it cuts the leaves into very small pieces. When the teeth were experimentally ground (Gipps and Sanson 1984) the leaf intake increased, presumably because the common ringtails were unable to 244 Life of Marsupials

obtain as much of the cell contents from the larger pieces of leaf. Like the koala and the greater glider, the common ringtail has a large caecum, where fine and coarse particles are separated. The very fine particles in the solute fraction are retained in the caecum for up to 70 h, where massive populations of micro-organisms become attached to the tissue components and partial digestion of cell walls and tanned cytoplasts occurs (O’Brien et al 1986). The common ringtail retains solutes and very fine particles for longer than does the greater glider but shorter than the koala (Table 7.1). However, the most important difference between the common ringtail and the previous two species is its practice of taking caecal contents directly back to the stomach, where they begin a second passage through the gut (Fig. 7.4): by this behaviour the common ringtail gains far greater access to the protein and energy products of bacterial fermentation that occur in its caecum. This behaviour, variously termed coprophagy, caecotrophy or reingestion is also practised by rabbits, Oryctolagus, Sylvilagus, hares, Lepus and pikas, Ochotona, and a leaf-eating lemur of Madagascar, Lepilemur leucopus (Charles-Dominique and Hladik 1971): it was discovered in the common ringtail possum by Mervyn Griffiths in the 1960s (Tyndale-Biscoe 1973). Chilcott (1984) found that the hard faeces, produced during the night while foraging, are not eaten, but soft faeces produced during the day while resting, were taken directly from the cloaca. By fitting a wide collar on them, Chilcott and Hume (1985) prevented common ringtails from ingesting soft faeces during the day, and so were able to collect the pellets, analyse their composi- tion and determine the contribution they make to the common ringtail’s diet. Unlike the normal faeces passed at night the caecal pellets have three times the water content and more than twice the nitrogen content (Fig. 7.11). Reingestion is not practised by the greater glider or by the closely related lemuroid ringtail possum, so what benefit is it to the common ringtail possum?

hard pellets caecal pellets 60 7

50 6

40 5

30 4 Dark

20 3 Percent dry matter content in pellets Light 10 2 Nitrogen content in pellets (% of dry matter)

0 1 1800 2400 0600 1200 1800

Time of day (hour) Figure 7.11: Daily pattern of faecal pellets produced by common ringtail possums, Pseudocheirus peregrinus: during the night, while they forage, they void dry pellets with low nitrogen content; during the day, while resting in a nest, they produce moist caecal pellets with high nitrogen content, which they eat. After Chilcott and Hume (1985). Life in the trees: koala, greater glider and possums 245

Nitrogen metabolism The daily MNR of common ringtails is 290 mg N/kg0.75, which is similar to the koala but half the requirements of the greater glider (Table 7.1). However, the caecal pellets, with their high nitro- gen content, contribute more than twice the estimated MNR of the common ringtail. Without this additional access to metabolic nitrogen from its diet the common ringtail would need to obtain 620 mg N/kg0.75 per day, very similar to the requirements of the greater glider, which does not practice reingestion. Common ringtails recycle 96% of the urea that is produced in the liver from their own metabolism to the caecum, where it is synthesised into bacterial protein. But this is only useful to the animal through reingestion, because the bacterial protein must be digested in the stomach and the amino acids absorbed in the small intestine. Recycling urea also conserves water that would otherwise be excreted as urine; instead, it is returned to the stomach in the caecal pellets. Water intake was made up largely of free water in the leaf, supplemented with drinking water, the mean daily intake being 98 mL/kg0.71, rising to 159 mL/kg0.71 during late lactation (Table 7.1). Reingestion is, therefore, a critically important factor in enabling the common ringtail to conserve water and thrive on a diet consisting solely of eucalyptus leaves with a nitro- gen content of only 1.1%. It is especially important during late lactation when 1.4 g of protein is exported each day in the milk to support growth of the young (Munks et al 1991, Munks and Green 1997). For a folivore dependent on eucalypt leaf of very low nitrogen content, maintain- ing late lactation is the most critical factor in determining the species’ long-term survival.

Energy metabolism Reingestion of caecal contents is also highly important in the energy balance of common ringtails. Chilcott and Hume (1984) found that common ringtails preferred Eucalyptus andrewsii to other species and lived exclusively on this in captivity. From their daily dry matter consumption of 41 g/kg0.75 they obtained 436 kJ/kg0.75 per day, or 59%, of the gross energy in the leaves. However, caecal pellets contributed 58% of this (Table 7.2). If this contribution is subtracted, the daily balance (182 kJ/kg0.75) is less than the SMR of the common ringtail possum, so clearly the species could not exist on a diet of Eucalyptus leaf without reingestion. This becomes more acute when we consider the FMR of common ringtails, which are 2.2–2.9 times the SMR, including the costs of reproduction. How do they achieve this on Eucalyptus foliage?

The costs of lactation in ringtail possums Munks and Green (1995, 1997) measured the field energetics of non-lactating common ringtails and male ringtails, and of lactating ringtails in early, mid and late lactation, using the techniques of isotope dilution (see Box 1.1). They were able to differentiate between the energy requirements of the non-reproductive common ringtail and the added burden of reproduction, especially the last phase of lactation. By using deuterium-labelled water in the mother and tritium-labelled water in the young, they could also differentiate between the energetic costs of the mother and the amount of resources exported to the pouch young. The FMR of male common ringtails varied through the year, with the highest rate (2.4 times SMR) during the autumn when the females come into oestrus and males are defending their home territories. Conversely, females at this time and those in the first phase of lactation, when the pouch young are minute, had the lowest FMR, only 2.2 times SMR. This rose to 2.5 and 2.9 times SMR for lactating females in the second and third phases of lactation, respectively. Thus, between early and late lactation the overall energy metabolism of the mother common ringtail increased by about 33%. In addition, the energy exported in the milk also increased 10-fold, from 7 kJ/day in early lactation to 75 kJ/day in late lactation, thereby increasing the total energy used to 50% above 246 Life of Marsupials

non-lactating and male common ringtail possums. The extra energy for this came from an increase in the daily intake of foliage, from 135 g to 180 g, and an increase in the time spent foraging for it: also about 100 g of stored fat, laid down during the earlier phases of lactation, which passed to the young in the milk. Another way to look at the cost of reproduction for the common ringtail is to calculate the total energy expended throughout the year and compare that for a non-breeding female (212 MJ/kg per year) with that for a female that produces two young to independence (234 MJ/kg per year). The difference (22 MJ/kg per year) represents the total investment in reproduction. This is very similar to the value for the tammar wallaby, Macropus eugenii, and the koala, Phascolarctos cinereus (see Fig. 2.24). However, in the very much smaller common ringtail, the investment is spread over a relatively longer time and the growth rate of the young is also relatively slow, when compared to terrestrial species of the same body mass. Thus, the time to weaning for the young common ringtail is about 210 days, whereas for the small rat kangaroos it is 130–160 days and for the bandicoot only 60 days. Why is the growth rate of the common ringtail so slow? The slow growth rate is not because the young are less efficient at converting milk into body tissue but to the quality of the milk provided to them: like in other marsupials, the milk of common ringtails changes through lactation (Munks et al 1991) but it is more dilute. The carbo- hydrate component declines but protein and lipid do not rise as markedly as in terrestrial marsu- pials in the last phase of lactation (Fig. 7.9c). This phase is prolonged in the common ringtail, as it is in the koala and brushtail possum, and the young are left in a nest, which conserves energy for the mother. While this may appear to be less efficient than the terrestrial species, the low yield of energy and protein from common ringtail milk may be dictated by, and be an adapta- tion for, living on eucalyptus foliage. Another advantage of the long lactation is the opportunity it provides for the young to learn social skills in the communal nest and develop the ability to climb and forage in the trees.

Social behaviour of common ringtails Unlike the koala and greater glider, common ringtails are gregarious and their social behaviour is centred upon the communal nests, which they build in the branches of small trees, or in the hollows of larger trees. In Victoria, Thomson and Owen (1964) found that the communal group usually consisted of an adult male with one or two adult females, their dependent progeny and the immature offspring of the previous year. Such a group might build several nests at different heights and localities and would repulse any strange common ringtail that attempted to enter an occupied nest. The group has a strong attachment to its site, so much so that when a group of common ringtails was experimentally removed from their territory, it was not recolonised for the next two years (Thomson and Owen 1964). The distribution and abundance of nests was highly correlated with particular types of habitat. They were most abundant in low scrub, or areas regenerating after partial clearance, but were much less common in heavily timbered country where the understorey was sparse. Since nests are essential for the survival of young after they relinquish the mother’s back, as well as being a daytime refuge for adults, abundance of common ringtails correlates closely with requirements for nest sites. Because of the limitations of their diet, successful reproduction by common ringtails is critically dependent on habitat with suitable food trees that will sustain the costs of late lactation.

Reproduction As with the greater glider, the constraints of the diet affect the reproductive strategy of the common ringtail possum, but in a different way. The breeding season of common ringtails in southeast Australia is more extended than that of greater gliders, and about 90% of females successfully produce young each season. Older females breed during April–June, and one-year Life in the trees: koala, greater glider and possums 247

old females in June and July. If young are lost prematurely, a female can return to oestrus and replace her litter and, in favourable conditions, females can produce a second litter in October. There are four teats in the pouch and four young can be reared, although the average litter is two. Thus, the potential rate of increase of common ringtails is considerably higher than for the greater glider (Table 7.3). However, the realised reproductive rate depends critically on the avail- ability of nutritious foliage to support late lactation. Life expectancy is considerably shorter than for the other two species, with one half dying in the first year and few male ringtails surviving their fourth year and females their fifth year, so that population turnover is rapid (Thomson and Owen 1964, How et al 1984). Compared to the greater glider the common ringtail has a more opportunistic reproductive strategy, which is reflected in its much wider distribution and adaptation to a variety of habitats and food resources.

Conclusions In contrast to the previous two species, the common ringtail is at the extreme lower limit for a strictly folivorous marsupial, but it ‘breaks the rules’ by recycling its food, thereby doubling its access to protein nitrogen, water and energy resources. Like other smaller species it has a higher rate of increase and turnover, which possibly makes it more adaptable to varying environments.

Other ringtail possums As mentioned earlier, the greater glider and all the ringtails of Australia and New Guinea can be put into three groups on the basis of their gut anatomy (Table 7.4). The greater glider and lemuroid ringtail are the only species that certainly do not practise reingestion of caecal contents. The green ringtail and the rock ringtail possum have a small or simple caecum and a long colon and their diet includes a greater variety of plants than the common ringtail. However, there is some evidence that the green ringtail practises reingestion. All the other species of ringtail from tropical Australia and New Guinea have a large sacculated caecum and all are predominantly folivorous. Hume (1999) thinks that they probably all practise reingestion, although there is only direct evidence for the largest New Guinea species, Pseudochirops cupreus, and the smallest, Pseudochirulus mayeri. In New Guinea there are five species that are considerably smaller than the common ringtail, weighing less than 500 g, which raises the question of how such small species can maintain themselves on a diet largely of tropical leaves. The two smallest species, Pseudochirulus canescens, and Pseudochirulus mayeri, include mosses, lichens and ferns in their diet. Not enough is known of them to exclude fungi and insects from their diet. If included, these could supplement the nitrogen requirements of such small animals. The common ringtail possum is the most fecund of all the ringtails. The Western ringtail has litters of one or two and may breed twice in the year (Jones et al 1994), while the Herbert River and Daintree River ringtails each produce a single litter of two young per year. All the other species of ringtails in Australia and New Guinea appear to produce only single young and have only two teats in the pouch.

Rainforest species Four species of ringtail possum are restricted to the rainforests of North Queensland, as well as the spotted and southern common cuscus and a subspecies of the brushtail possum, the coppery brushtail. The particular habitat requirements of these species have interesting parallels with the arboreal folivores of the Eucalyptus forests. The only species that occurs in both types of forest is the common ringtail possum. The north Queensland rainforests occur in two regions separated by a wide barrier of differ- ent habitat in the Laura basin (Winter 1984, Kanowski et al 2001). The two species of cuscus, which also occur in New Guinea, are restricted to the northern region, and the four species of 248 Life of Marsupials

Table 7.4: Diet and occurrence of reingestion in species of the Pseudocheiridae

Species Locality Mass (kg) Litter size Diet Caecum; Reingestion Pseudocheirus species: peregrinus E. Aust 0.7–1.1 2–4 Eucalyptus + Large, yes occidentalis W. Aust 0.9–1.1 1–2 Eucalyptus + Large, yes Pseudochirulus species: herbertensis NQ 0.8–1.45 2 Tropical foliage Large cinereus NQ 0.7–1.2 2 Tropical foliage canescens NG 0.22 1–3 Tropical foliage Large caroli NG 0.44 forbesi NG 0.4–0.6 1 Tropical foliage Large mayeri NG 0.15 1 Tropical foliage + Large, yes bryophyta schlegeli NG 0.3 Pseudochirops species: archeri NQ 0.7–1.35 1 Tropical foliage Small, yes alberti NG 1.2 1 Tropical foliage Large corinnae NG 1 1 Tropical foliage Large coronatus NG 1.5 cupreus NG 1.2-2 1 Tropical foliage Large, yes Petroseudes dahli NAust 1.3–2 1 Blossom, foliage Simple Petauroides volans EAust 0.7–1.7 1 Eucalyptus leaf Large, no Hemibelideus NQ 0.8–1.4 1 Tropical foliage Large, no lemuroides

E. Aust, eastern Australia; N. Aust, northern Australia; W. Aust, western Australia; NG, New Guinea; NQ, north Queensland. From Hume (1999), Flannery (1995).

rainforest ringtail possums and the coppery brushtail possum, none of which occurs in New Guinea, are restricted to the southern region. This disjunct distribution indicates a long-standing isolation of the two regions of rainforest since at least the last union of New Guinea with Cape York Peninsula 12 000 years ago. However, other elements of the New Guinea marsupial fauna have crossed the barrier into the southern region of rainforest, notably the rufous spiny bandicoot, Echymipera rufescens (see Chapter 5) and the striped possum, Dactylopsila trivirgata (see Chapter 6). Within the southern region the four species of ringtail and the coppery brushtail are restricted to the upland rainforest above 400 m and are most abundant above 800 m. The coppery brush- tail and the lemuroid ringtail are limited to forest above 600 m, while the Herbert River and Daintree River ringtails have lower limits of 350 m and 450 m, respectively. And within these limits the species are almost twice as abundant in forest growing on rich basalt soils than in forest on poorer metamorphic soils. The green ringtail has a lower limit of 300 m and the two species of cuscus range down to sea level. These altitudinal limits determine the distribution of each species because they cannot traverse deep valleys that intersect the forested region if these are lower than their altitude limit. Thus, the Daintree River ringtail in the forest north of Cairns is isolated by intervening lowlands from the closely related Herbert River ringtail in rainforest south of Cairns; and the lemuroid ringtail has isolated populations on the uplands of separate Life in the trees: koala, greater glider and possums 249

ranges. Intolerance of the higher temperatures in the lowland forests sets the lower altitudinal limit of each of these five species (Kanowski et al 2001). If this is so, the anticipated rise in global air temperatures may drive the species out of the lower parts of their existing ranges, leading to further fragmentation of the populations. A more immediate threat to the survival of isolated populations of these species is the fragmentation of the rainforest brought about by land clearing. Much of the rainforest around Atherton between 600 m and 900 m was cleared for farmland between 1909 and 1940. By 1983, 76 000 ha of rainforest had been removed, leaving some large tracts of intact forest that had been selectively logged, many small patches of 2–600 ha, and corridors of secondary growth connect- ing the fragments. Laurance (1990) assessed the impact of fragmentation on the four species that occur in that region by spotlight census counts over 13 months in seven large tracts, 10 small patches and three corridors. Almost all the sightings were in forest above 700 m. The lemuroid ringtail was found exclusively in the large tracts, the coppery brushtail and the Herbert River ringtail were found in all sites and the green ringtail was most abundant in the corridors. The overriding determinant of survival of arboreal marsupials in fragments of rainforest is their ability to disperse across unfavourable country, just as it is for other species in eucalypt forests. Indeed, the lemuroid ringtail is as dependent on large tracts of forest as its southern relative the greater glider, whereas the green ringtail and coppery brushtail are more adaptable, like their southern relatives the common ringtail and the brushtail possum. Laurance (1990) concludes that because the smallest forest tract that contained the full complement of species was 3600 ha, reserves should be larger than this. Fortunately, for the long-term survival of the arboreal species, the whole of the North Queensland Wet Tropics has since 1990 been declared a World Heritage conservation area and no further land clearing is taking place.

Cuscuses The cuscuses are less specialised herbivores than the species so far discussed, and most feed on a variety of fruits and the leaves of some rainforest trees. However, in captivity, the common spotted cuscus (Fig. 7.5e, Plate 13) readily ate fresh chicken as well as a variety of fruits and leaves, and other species while in captivity have been observed to kill and eat rats, birds and small reptiles (Flannery 1995). The canines of these species are proportionately larger than those of the strictly folivorous species and the first pair of premolars are sharp-pointed teeth, which may indicate that live animals are a common part of the diet of cuscuses. The digestive anatomy of the ground cuscus, Phalanger gymnotis, is less specialised for a leaf diet than that of the common ringtail or greater glider (Hume 1999). The stomach is larger, the small intestine longer and the caecum smaller and the colon is not enlarged as a fermentative chamber. Indeed, the production of SCFA contributed less than 5% of the digestible energy, compared to 10% for the koala and almost 50% for the common ringtail possum. A study on the spotted cuscus by Dawson and Degabriele (1973) found that its metabolism, like that of the koala, is less than the predicted value for a marsupial of its body mass and that its fur provides good insulation, thereby conserving energy. In this it resembles the koala and the placental sloths of the South American rainforests, which live on a similar diet. Knowledge of the reproduction of cuscuses is also scant. The litter size is one to three but there is no substantive evidence for seasonal breeding or how long the association between the mother and the young is.

The mountain brushtail possum Compared to the three species so far considered, the mountain brushtail possum or bobuck and the brushtail possum are generalist folivores. Both species can live on eucalyptus foliage but also 250 Life of Marsupials

feed on a wide range of other plants, many of which are found in the lower storey or on the forest floor (Seebeck et al 1984). However, in their ecology, the two species are very different from each other. In his study of both species in the same area of northern New South Wales, How (1972) observed that 90% of the diet of the bobuck was lower storey plants, whereas that was only 20% of the diet of the brushtail possum. The bobuck, like the greater glider, is conservative in its reproduction and is restricted to the tall wet Eucalyptus forests of eastern Australia (Fig. 7.6). Where they overlap, the bobuck is larger than the brushtail possum and better adapted to tall Eucalyptus forest and displaces it from the lower levels of the forest in this habitat. However, its very adaptations of low fecundity and high site attachment make it less resilient and adaptable to other habitats, or to alterations of its own habitat. Its absence from Tasmania, despite the presence of suitable forest, suggests that it failed to penetrate a barrier of unsuitable habitat at the time the island was joined to the mainland. The more versatile brushtail possum did do so and, in the absence of competition from the bobuck, occupied the dense wet forests, which are the prime habitat of the bobuck on the mainland of Australia. It is not clear what factors restrict the bobuck but its social organisation and breeding strat- egy are clearly adapted to survival in a stable ecosystem. Like the greater glider the home ranges of females do not overlap but those of males overlap with either one or two females with whom they share dens. Martin (2000) found that the paternity of the pouch young corresponded in almost all cases with that of the resident male, thus strongly supporting the inference that the species is monogamous. In polygynous species, where the males compete for access to several females, adult males are considerably larger than females, but there is no difference in body size between male and female bobucks, both being in the range of 2.8–3.5 kg. Females reach sexual maturity at three years and there is a single period of births in early autumn (How 1976). Pouch life is six months and the young does not become independent until nine months, much longer than the brushtail possum (Table 7.3). Even after this time, if the young one survives, it continues to associate with its mother and share her den for another year. During this lengthy association the mother either does not breed or, if she does, the next young seldom survives beyond early pouch life. Thus, the annual fecundity of the bobuck varies between 0.3 and 0.8 young/female. However, for those young that do survive to independence, life expectancy is higher than for young brushtail possums. At another site in Victoria (Viggers and Lindenmayer 2000) the longevity of bobucks was found to be over 10 years and some of these animals remained in the same small area throughout their adult lives. Like koalas and greater gliders they also show very strong site attachment even after their habitat is severely altered by logging. In brief, the bobuck has a conservative breeding strategy, appropriate to a stable environ- ment. This contrasts with the breeding strategy of the brushtail possum in Australia. Because so much more is known about its physiology and ecology, the last section of this chapter is about the brushtail possum.

Paragon or pest – the secrets of success The brushtail possum, Trichosurus vulpecula (Fig. 7.5d, Plate 13), like the common ringtail possum, is more fecund, more adaptable than the bobuck, and occurs in a wide range of habitats throughout continental Australia and all the larger islands (Fig. 7.6). Its fecundity and adaptability has been most emphatically demonstrated by its colonisation of the whole of New Zealand and several offshore islands since its introduction in 1858. In New Zealand, some long established possum populations display the bobuck’s conservative strategy, like they do in Tasmania, while in newly occupied areas they display the more fecund strategy seen in Life in the trees: koala, greater glider and possums 251

continental Australia. What is it about this species that has enabled it to thrive in the altered conditions of Australia since European occupation, and that has made it the prime mammalian pest species in New Zealand? The brushtail possum is an animal of the forest and woodland and is common in the winter rainfall areas of Australia, as well as on Tasmania and Kangaroo Island. However, it also occurs in the drier regions along water courses, where it lives in the river red gums, Eucalyptus calmaldu- lensis, and feeds on the surrounding grasslands. It also occurs in the far north of the continent in open forest, and the coppery brushtail possum occurs in tropical rainforest. Across its wide range it varies markedly in body mass from about 1 kg in northern Australia to 4.5 kg in Tasmania, and from pale grey pelage and long ears in the north to dark brown and short ears in the south. The dark colour is common in high rainfall areas of Tasmania and the same pattern is found in New Zealand. Brushtail possums in Western Australia, like other marsupials there, are tolerant of high concentrations of sodium fluoroacetate, whereas possums from elsewhere in Australia are highly susceptible to the compound. The differences are very great: the lethal dose for possums in Western Australia is 125 mg/kg body mass compared to 1 mg/kg in eastern Australia. This is a local adaptation to the occurrence of sodium fluoroacetate as a protective compound in some native plants, particularly the genus Gastrolobium, of Western Australia (King et al 1978).

Diet of brushtail possums In marked contrast to the koala, greater glider and common ringtail possum, brushtail possums thrive on whatever vegetation is available. In Eucalyptus forests they eat eucalypt leaves, but they have adapted to several introduced species such as the shoots and male cones of pine trees in plantations, pastures of clover and grass, and in suburban gardens will eat ornamental and orchard species. In New Zealand they have occupied all the native forests, where they eat many of the broad-leaved tree species, especially the several species of rata and pohutakawa of the genus Metrosideros, a distant relative of Eucalyptus in the same Family Myrtaceae. In New Zealand they have also been observed to eat fungi, snails and the eggs of native birds. Both in Australia and in New Zealand possums eat three or four different species in a night’s foraging, rather than restrict their diet to a single species, as do koalas and greater gliders, and to a lesser extent ringtail possums. This may reflect differences in the gut of possums compared to the more specialised eucalyptus feeders. Freeland and Winter (1975) argued that secondary compounds in the foliage determine the quantity of Eucalyptus that a possum can eat and that this regulates possum populations and affords the forest trees some protection from browsing. However, in other forests, 90% of the dietary intake was Eucalyptus foliage and in laboratory experiments brushtail possums were able to maintain their body weight on a diet of Eucalyptus melliodora, which was low in nitrogen and silica but high in lignified fibre and phenolics (Foley and Hume 1987a). The possums preferred mature leaves to young shoots, so the diet had the same nutritional content throughout the year. Compared to the specialist leaf eaters already discussed, how are these generalist feeders adapted to eating eucalyptus foliage? The molar teeth of possums are more rounded than those of greater gliders and ringtails (Fig. 7.7c), so they cannot cut the leaf tissue as finely, and as a consequence the digestibility is lower (Table 7.1). However, they can crush plant tissue, which is more effective when eating fruit or herbs on the forest floor, and on this diet the digestibility is high. Unlike in the greater glider and the common ringtail possum, internal ridges do not divide the possum’s caecum, so it cannot separate coarse and fine particles to the same extent as they do, and is more like the koala in this respect. However, the mean retention times for both fluid and particulate material on a Eucalyptus diet were about 50 h; the retention of fluid fraction was the same as the greater glider and faster than the ringtail possum and koala. 252 Life of Marsupials

Nitrogen metabolism The MNR of brushtail possums on a diet of Eucalyptus melliodora were greater than those of the ringtail and nearer to the greater glider (Table 7.1). Because they cannot selectively retain the fluid and fine particles of the diet, much potential nitrogen is lost in the faeces (Wellard and Hume 1981, Foley and Hume 1987b). Since possums do not reingest the caecal contents they cannot make use of any bacterial protein synthesised from recycled urea because the amino acids cannot be absorbed across the colon. However, when possums were fed on a mixed diet, not Eucalyptus, with more accessible protein nitrogen, their MNR were halved (Harris et al 1985). In other words, on a mixed diet possums can obtain sufficient protein for the maintenance of life on less food and can still raise their daily intake to meet the extra protein requirements of reproduction, especially late lactation. This may be the reason that both the brushtail possum and the bobuck select a variety of foods rather than living exclusively on Eucalyptus. In New Zealand forests Williams (1982) noted that possums would travel considerable distances to eat the fruits of certain species, such as hinau, Elaeocarpus dentatus, which has high protein content. And female possums in this forest were heaviest in autumn, when hinau is in fruit, especially in years when the fruit was abundant.

Energy metabolism Conversion of digestible energy (340 kJ/kg0.75 per day) to metabolisable energy (260 kJ/kg0.75 per day) was also lower in the brushtail possum than in the common ringtail possum, and for the same reason: the high content of essential oils and phenolics require energy to be detoxified (Table 7.2). Nevertheless, brushtail possums maintained their body weight on this diet, which suggests that energy intake, at 1.5 times SMR, was sufficient. Despite the lack of good separa- tion of fine material in the gut of brushtail possums, some bacterial digestion of cellulose does occur in the caecum, with production of SCFA. This contributes about 15% of the possum’s energy needs (Foley and Hume 1987a), although, as in the koala, the main source of energy is the contents of plant cells, not their walls. As with nitrogen metabolism, possums can compensate for their less efficient digestion of Eucalyptus leaf by having a more varied diet of high-energy food, such as fruits and ground living plants (Harris et al 1985). The FMR of the brushtail possum has not been determined, so comparisons cannot be made with the three other species of arboreal folivores on this important aspect of its life. However, it is likely that its FMR will be about 2.5 times its SMR (Table 7.2) but this can only be resolved by more comprehensive studies of possum energetics.

Breeding biology of the brushtail possum The reproduction of the brushtail possum has been studied in a wide range of habitats and latitudes in Australia and in New Zealand. In northern Australia births occur at all times of the year and there is no obvious peak: in all other parts of its range females come into oestrus during February–April and the main peak of births occurs in April–June. There is some evidence that the onset of breeding is influenced by changes in the day length, operating through the pineal gland. In a population in the Orongorongo range near Wellington, which has been studied since 1947 and continuously since 1966, there has been very little variation from year to year, with the mean date of birth being between 3 May and 15 May (Bell 1981, Efford 1998). This supports the idea that changing day length at the autumnal equinox may be the initiating signal for females to come into breeding condition (see Chapter 2). In all populations that have been studied, 80–90% of females over one year old produce a single young at this time and it is suckled in the pouch for about 145 days (Table 7.3). If the young survives, it then rides on its mother’s back for about six weeks and becomes fully independent at six months. Some females may then return to oestrus and produce second young in October. The Life in the trees: koala, greater glider and possums 253

proportion that have a second young varies from 50% near Canberra, 20% near Brisbane and Christchurch, to none in Tasmania, parts of New South Wales and the Orongorongo population near Wellington (Kerle 1984, Fletcher and Selwood 2000). The determining factor does not seem to be latitude or locality but rather the availability of abundant nutritious food in spring. In New Zealand a second birth is more common in populations that are invading new habitat and does not occur in long-established populations, such as the Orongorongo population. This is consonant with available resources being the main driver. Thus, the fecundity of the species varies from less than one to a maximum of two young per female per year (Table 7.3). Since females can become sexually mature at one year, compared to three years in the bobuck, the reproductive potential of brushtail possum populations is very much higher. But, while the potential is high, survival of the young is contingent on adequate resources during late pouch life and the first months of independence. The composition of milk and the changes through lactation in the brushtail possum are similar to those of the koala and the ringtail possum (Fig. 7.9b), with the lipid and protein concentration not rising as much as in terrestrial herbivores, and late lactation being prolonged. However, since neither the FMR through lactation nor the total volume of milk secreted each day during lactation has been determined, it is not possible to estimate the total cost of reproduction in this species. If it is similar to the koala and the ringtail possum, the major burden of reproduc- tion is the final prolonged phase of lactation during winter and spring, supported from fat stores developed in autumn. Like the ringtail possum, female brushtail possums increase their lipid stores in the autumn, and their body mass reaches a maximum in winter. In the Orongorongo population, the body mass of females in autumn critically affected the subsequent survival of the pouch young (Bell 1981, Efford 1998). Pouch young carried by females

young reared to 2.4 independence

2.3

2.2 young lost from pouch

Mean body mass (kg) 2.1 no young born

2.0 Summer Autumn Winter Spring

Figure 7.12: The relationship between body mass of adult female brushtail possums, Trichosurus vulpecula, in autumn and winter, and their subsequent breeding success in spring in the Orongorongo forest, New Zealand, during 1966–75. After Bell (1981). 254 Life of Marsupials

over 2.5 kg and more than four years old had the highest chance of surviving to the end of pouch life, whereas the survival of young carried by smaller, younger females was less than 50% and many failed to breed altogether (Fig. 7.12). The main mortality of pouch young occurred at three months, which is the time when the milk composition changes from mainly carbohydrate to lipid (Fig. 7.9b) and is energetically much more costly to produce. It also coincides with the middle of winter when the condition of all brushtail possums is low and the main mortality of the adults occurs. During the 30 years at Orongorongo the greatest number of young were weaned in years when there had been a large crop of hinau fruit in the previous autumn and pouch young mortality greatest when this crop was sparse (Brockie et al 1981, Ramsey et al 2002). The inference is clear that the deposition of fat stores at the start of the breeding season is the key to carrying a young through the winter successfully and providing sufficient milk in late lactation for it to become independent. The overarching determinant for successful reproduction in the possum, as for the other folivores, is therefore the availability of sufficient resources at the onset of breeding in autumn. The attributes of high fecundity and early dispersal of the young provide the means for this species to exploit new habitats rapidly, or to re-establish in an old habitat after the population has been reduced by natural factors or human activity. An unexpected finding is that the fecun- dity of the species is generally higher in Australia than in New Zealand, despite much higher densities in New Zealand than in Australian populations. This paradox will now be examined.

Social behaviour and ecology The most detailed and long-term studies on the ecology of the brushtail possum have been done in New Zealand, because of the importance of the species in the forests and farmland of that country. However, shorter-term studies in Tasmania and across mainland Australia largely corroborate them. As with diet and reproduction, the ecology of the brushtail possum varies across its wide distribution, again reflecting its adaptability.

Density varies with habitat and time Possums are difficult to count, especially in New Zealand forest, because they are living in a three- dimensional world of tangled, dense vegetation. Indirect measurements have been attempted, such as counting the number of droppings left on a cleared area of the forest floor. Although this may give a relative estimate, there is considerable variation in the number of droppings produced each day that affect the estimate. Mark and recapture studies, especially those done over many years, give much better estimates; and estimates based on the number of bodies picked up along a poison trail on successive days can give an estimate of the standing population at the time of the operation. In 31 New Zealand populations the density varied greatly between major habitats (Efford 2000). The highest densities (5–25 per ha) occurred in broadleaf–podocarp forests that have a rich understorey of nutritious species or are adjacent to farmland pastures. The highest recorded densities were 25/ha in a previously logged forest adjoining pasture, and 19.8/ha in a modified forest with regenerating understorey species, highly favoured by brushtail possums. Much lower densities of 1–3/ha were found in southern beech, Nothofagus, forests and pine plantations that lack understorey species. By contrast, most densities in Australia are less than 2/ha, ranging from less than 0.3/ha to a maximum of 6/ha in a Tasmanian rainforest. Are the densities of brushtail possums stable over time? The only good evidence for long- term stability is the Orongorongo study near Wellington, where the density fluctuated around 10/ha from 1966 to 1999, and trap catches 20 years earlier had been similar (Fig. 7.13). It has often been said that New Zealand populations are much higher during the initial colonisation of an area but the evidence for this is not precise. Life in the trees: koala, greater glider and possums 255

18

16

14

12

10

8

6

Possums per hectare 4

2

0 1940 1950 1970 1980 1990 2000 Year Figure 7.13: Variation in density of brushtail possums, Trichosurus vulpecula, between 1967 and 1998 in the Orongorongo population, based on live trapping data over five nights, three times per year. Also shown is the estimated density by kill trapping in the same forest in 1946–47. After Efford (2000).

Several factors may be involved in densities being so much lower in the native habitats of the brushtail possum in Australia than in New Zealand forests. While a lower carrying capacity of the sparser vegetation in eucalypt forests is certainly important, it is not reflected in a lower fecundity. So are the lower densities due to a higher mortality rate, especially among the young animals? Young possums have to contend with high summer temperatures in Australia and several predators that do not occur in New Zealand. If secure den sites are essential to survival, competition for them from other possums and from other marsupials, nesting birds and large reptiles, such as goannas, Varanus, may be an important factor in the lower densities of possums in Australia. Wren Green (1984) put forward the interesting idea that competition for suitable den sites is the main limiting factor on possum populations. Since den sites are more abundant in New Zealand forests than in Australian Eucalyptus forests and woodlands, higher densities of possums can be sustained in New Zealand than in Australia, despite higher fecundity in Australian than in New Zealand populations. Green and Coleman (1987) tested this in a detailed analysis of dens and den use in a popula- tion of 47 possums carrying radio transmitters in a New Zealand broadleaf–podocarp forest. Half of the 282 dens were on or under the ground, the others being in tree hollows or dense vegetation. Over one year each possum used 10–15 dens, the same number as used by possums in the Orongorongo population (Cowan 1989b). This is a far higher den use than the 1–3 dens per possum found in three studies in Australia (How 1972, Winter 1976, Hocking 1981). In Australian studies the high mortality of the dispersing young males has been interpreted as being due in part to competition from other possums and other marsupials for limited den sites, but there is no evidence of this in New Zealand. 256 Life of Marsupials

Home ranges, stability, overlap, movements Density is not the same as home range of an animal. Home range is the normal area that an animal occupies in its daily activities and it may or may not share it with other members of the same species. The home ranges of males and females overlap in all the studies done on brushtail possums, so that the home ranges are larger than the measures of density might suggest. But home ranges, like density, vary with the quality of the habitat, being largest in poorer habitat and smaller in better habitat. Where possums have been followed by radio tracking for a considerable time (Brockie 1992) the individual home range is very stable. Furthermore, when an adjacent area has been depleted of its population by a poisoning or trapping campaign, the resident brushtail possums do not move into it but remain in their familiar home range. The possums that eventually colonise vacant land are the young animals, especially males, that have left their natal area and are dispers- ing over long distances . Although possums will share dens, they are generally solitary and sleep in separate dens during the daytime. Nocturnal encounters between adults of either sex are mild, the usual response being for both animals to avoid direct contact, but to peer silently at each other with erect ears. Possums avoid each other by marking branches and other objects with scent from sebaceous glands on the chest and chin and two sets of paracloacal glands: one produces a holocrine secre- tion that is distributed by urine dribbling and the other an apocrine secretion that is produced in response to fear. Brushtail possums do not show any pattern of mutual grooming or other contact activi- ties such as gregarious species like common ringtail possums do, the only longer contact being between a female and the young riding on her back until it is weaned. Adult possums only come together for the brief period of mating: there is no close association between mated pairs, such as occurs in the bobuck.

Mating patterns If there is a large overlap of home ranges, what are the mating patterns of brushtail possums? Two recent studies in New Zealand have examined this indirectly by assessing the paternity of all pouch young, using minisatellite DNA profiles of the adults and pouch young in the entire population (Taylor et al 2000, Ji et al 2001). Although carried out in different places and years, both studies showed that while some males sired several young in a season, more than half sired none. But the second young of females that lost young were seldom sired by the same male, and in subsequent years other males sired the progeny, so that mating appeared to be promiscuous and random. This is quite different from bobucks, where most matings are between the pair that shares a home range. The fact that a male brushtail possum could sire up to seven offspring in the short breeding season, suggests that any consort behaviour in these populations must have been very short. This is different from Winter’s (1976) observations in a Queensland population, where it took the males nearly one month of consorting with a female before mating took place. However, both the New Zealand populations had been severely reduced in previous years and this could have affected male behaviour: it is another example of the varied and flexible responses of this most adaptable species.

Juvenile dispersal and mortality Young brushtail possums seem to behave differently in Australian and New Zealand popula- tions, although this perception may change with more thorough studies in Australia. In the first study by Dunnet (1964) near Canberra, between 70 and 90% of the young survived through pouch life but after weaning and during the subsequent dispersal phase there was considerable Life in the trees: koala, greater glider and possums 257

(a) 200

160 ) m ( 120

80

istance moved 40 D

2505

20 Onset of sexual maturity

15

10

5 Number of young surviving 0 5 7 9 11 13 15 17 19 21 > 24

Age in months Emergence from pouch

(b) 200

) 150 m (

100

50 istance moved D

0 0123456> 6

Age in years

Figure 7.14: The fate of young brushtail possums, Trichosurus vulpecula, in Australia and New Zealand: (a) in Australia the number of male (O) and female ({) young surviving on the study site each month after emergence from the pouch is equal until sexual maturity at 11 months; after this age males move further from their natal territory than females, and their mortality is greater; (b) in the Orongorongo population, New Zealand, the same pattern of dispersal occurs but one year later than in Australia. Data for Australia from Dunnet (1964), and Johnson et al (2001) and for New Zealand from Efford (1991)). mortality, which fell most heavily on the male young (Fig. 7.14a). In Australian populations the young males begin to disperse at one year, when they are becoming sexually mature, whereas males in the Orongorongo population do not disperse until after the age of two years (Fig. 7.14b). In both countries the females move much less than the males and most of them stay within the mother’s home range and continue to share her den sites. 258 Life of Marsupials

The males that disperse suffer a higher mortality and this occurs earlier in Australia than it does in New Zealand. Part of the reason for this may be the several predators that can take dispersing young possums in Australia, such as spotted-tailed quolls, foxes, cats, goannas, carpet snakes and powerful owls, while in New Zealand the only predators of possums are cats and humans. But as already mentioned, dispersing possums in New Zealand have little difficulty in finding an unoccupied den. One reason for the higher survival of female young than of males is also related to dens. In the Orongorongo population female young continued to associate with their mothers after weaning and some eventually inherited the prime den sites. This was repeated over three generations (Brockie 1992). Similarly, in Queensland, daughters settled in their mother’s home range but sons moved away. Johnson et al (2001) proposed that there is competition between mothers and daughters for dens and the daughters are excluded from a den that is occupied by the mother. In forests where there was a shortage of good dens the females produced more sons, who did not compete directly for dens, but where den trees were plentiful the number of female young produced was greater. How a female can adjust the sex of her offspring in accordance with exter- nal factors, such as abundance of dens, is not known, nor was it postulated.

How long do possums live? Life expectancy affects the lifetime fecundity of females. A few records from near Canberra and in the Orongorongo study indicate that female possums can live for up to 13 years in the wild and still breed: males live for up to 10 years. However, the average life expectancy from the end of year 1 is much less than this. The best estimates again come from the Orongorongo study (Efford 2000) where the life expectancy to six years is 90% but declines steeply after seven years, with most of the older animals dying during the cold wet winter. This is indirectly supported by the increase of possum remains in the diet of feral cats at this time. In How’s (1981) study in northern New South Wales, the life expectancy of brushtail possums was much lower, with a severe mortality between independence and two years and few of the resident animals living to six years. Part of the reason in this case was because the brushtail possums were excluded by bobucks from the Eucalyptus forest and good dens, and were living in peripheral habitat.

Conclusion This brief review of the brushtail possum emphasises the ability of the species to live in many habitats: and its ability to mature at one year, promiscuous mating pattern and ability to breed twice in a year, and continue to do so for its full life span, provides it with a high lifetime fecun- dity. This potential is very responsive to prevailing nutritional conditions: where these are optimal the population can increase very rapidly but the potential can be moderated at several points to achieve stability in established populations. Brushtail possums share the habitat and its available food but defend den sites, which are essential to individual survival from climatic extremes and, in Australia, from predators. Where den sites are scarce this is the major factor limiting the population and young animals that cannot find a den when they disperse suffer heavy mortality, especially in Australia. The reproductive performance of females is very responsive to available food resources, especially in autumn and spring. If these are insufficient to build up fat reserves, the young will die during mid pouch life because the mother cannot sustain her investment. After weaning, mortality operates disproportionately on the young that cannot establish a home base and a den refuge. Males disperse further than females and they are the vanguard in colonising unoccupied habitat. Because den sites are so important to survival of the new cohort of possums, a popula- tion can overshoot the carrying capacity of the forest if these are not a limiting resource. This can lead to over browsing and mortality of favoured species of forest trees and understorey shrubs. Life in the trees: koala, greater glider and possums 259

Brushtail possums and New Zealand forests: an arms race in action New Zealand separated from Gondwana 80 million years ago, before marsupial or placental mammals reached that continent. Its forests, which are similar to the forests that originally covered Australia, evolved without leaf-eating mammals, although leaf-eating insects take a heavy toll of the plant growth, and ground-dwelling moas, Dinornis, Euryapterix, Pachyornis, probably browsed the understorey species. Polynesian settlers probably induced the grasslands of the eastern South Island by extensive burning of the broadleaf–podocarp forests and they brought the rat, Rattus exulans, and domestic dog. However, the advent of European settlement caused far more profound changes to the forests through extensive clearing for agriculture and the introduction of many species of browsing and grazing mammals. Among these species was the common brushtail possum, a highly regarded furbearer. All the introductions were probably the common brushtail possum, although introduction of the bobuck was recommended as late as 1936. Acclimatisation Societies were responsible for the earli- est successful introductions between 1858 and 1900 from Tasmania and from near Melbourne and Sydney (Fig. 7.6). They also made many secondary introductions from New Zealand bred stock, so that the species was firmly established on both islands by 1900 (see Montague 2000). The few places where initial introductions failed could be attributed to insufficient numbers released and to lack of suitable dry shelter at the site of introduction. The mixture of genotypes from Tasmania and continental Australia, and subsequent mixing in New Zealand by the secondary introductions, enhanced the genetic diversity, so that traits adapted to the New Zealand conditions could be rapidly selected for and further aid the early establishment and spread of the species. One example of the rapid selection for favourable genotypes is the frequency of dark-furred possums in high rainfall regions of New Zealand (Brockie 1992). Also, once a new population was established in a district, the potential rate of increase of the possum was probably fully realised very quickly. Thus, in the Catlins forest of the South Island, 42 possums were liberated in 1894–95 and 18 years later trappers took 60 000 skins from that district, and no doubt left many more possums behind (Wodzicki 1950). Similar rates of increase must have occurred across the country because a large export trade in possum skins flourished for many years, reaching 500 000 in 1945 and three million in 1980. Since then the number of skins exported has steadily declined because of a weak market, not lack of possums. The rabbit and the brushtail possum were both introduced to New Zealand in about 1858. The rabbit had become a devastating plague 20 years later, the brushtail possum not until 90 years later. The initial build up of the possum was much slower than the rabbit, presumably because of its lower rate of increase. But its ability to use a great variety of plants as food and to occupy a greater diversity of habitats enabled it to occupy New Zealand far more extensively than the rabbit. In the long term its marsupial mode of reproduction was no disadvantage and it is now a far more intractable problem than is the rabbit. By 1900 the first adverse effects of possums were reported by orchardists, but not until 1920 were effects on the native forests noticed. The first study concluded that the damage to the native forests was negligible and was far outweighed by the benefits of the fur trade. In the late1940s when the effects of possum browsing in native forests became more apparent, official opinion changed: the possum was then declared to be a pest of equal importance to the rabbit and red deer. This was the impetus for the beginning of ecological research on the possum at the Orongorongo site and subsequently in other centres.

Impact on forests – spread of tree and shrub mortality as possums arrived Widespread mortality of forest trees was first reported from the North Island in 1947 and soon after the same pattern was reported from the broadleaf–conifer forests of the western slopes of the South Island. In some places, notably the Kokatahi valley, not only were the dominant tree 260 Life of Marsupials

species defoliated and dying but also many of the understorey species were equally affected. The dominant species most affected were southern rata, Metrosideros umbellifera, and kamahi, Weinmannia racemosa, and the conifers, Podocarpus totara, and Libocedrus bidwillii. Fifty years later old dead stumps represent the former rata and kamahi, but the Kokatahi valley is again clothed in forest. A different suite of species has replaced the original vegetation but the total biomass is almost the same (Bellingham et al 1999). Since the brushtail possum population has never been measured in this valley it is not possible to say whether the carrying capacity of the forest has been changed by the transformation. The same pattern of forest mortality has occurred successively in other parts of Westland since the 1950s, as possums have spread up the valleys, the most recent being along the Adams River. The dramatic changes that have occurred are generally ascribed to large invading popula- tions of brushtail possums, causing the mortality by defoliating preferred species of trees and shrubs. While the possums are undoubtedly a major factor, there is growing evidence that the possums may precipitate changes that were inherent in some, but not all, the indigenous forests. Some indigenous forests, such as the Catlins, that have had large populations of possums for more than 100 years still show no mortality and retain the original plant species. In the Westland forests, which are the most severely affected, mortality is confined to the zone above 600 m. The upper limit of the forest, where the plant species are presumably most vulnerable to change, is about 1000 m. Hence, the arrival of browsing mammals and changes in the climate in the past century may both have contributed to the widespread mortality in the vulnerable zone above 600 m. Another factor in Westland that may underlie the current forest mortality is the evidence of widespread collapse of the forest after an immense earthquake in 1717 (Wells et al 1999). The rata trees, which recolonised the slopes in the aftermath of that earthquake, are now reaching the end of their life span and their inevitable death is contributing to the mortality seen today. Thus, forest collapse, when it occurs, may have more to do with pre-existing susceptibility to dieback, than to the size of the brushtail possum population. In susceptible forest it may only be

30 25

25 20

20 15 nt y a h ce / r 2

15 10 m Pe

10 5 orest mortalit F Possums per hectare

5 0

0 200 400 600 800 1000

Altitude (m) Figure 7.15: The relationship between density of brushtail possums, Trichosurus vulpecula, at different altitudes in a New Zealand broadleaf-podocarp forest and tree mortality. Possum density was measured by kill trapping, and tree mortality at each site by (V) percentage of dead trees and by (O) basal area of dead timber (m2/ha). After Coleman et al (1980). Life in the trees: koala, greater glider and possums 261

necessary for the possum population to pass a certain low threshold density for the balance to be tipped to dieback. Conversely, in another forest, where the vegetation is not susceptible, larger possum populations may be supported without serious effect on the vegetation, as for example in the Catlins. What is the evidence for this? The best attempt to answer this question was made by Coleman et al (1980), who measured possum abundance and plant damage independently on the same study sites. They measured possum abundance by trap capture and counts of faecal pellets on cleared areas, independ- ently measured degree of leaf browse, and determined vegetation damage in terms of basal area of dead trees. As previous studies had shown, the severest degree of browse and of tree death occurred in forest at 600–900 m (Fig. 7.15). Possum density at this altitude was between 1.9 and 7.5/ha, whereas in the forest below 300 m, where the degree of browse and number of dead stems was low, the possum density was 25/ha. At the lower altitudes the vegetation comprised more understorey species and bordered on farmland, both of which could be factors in supporting the higher density of possums. Nevertheless, rata and kamahi were present throughout the study site and only showed severe mortality at the higher altitudes. Thus, forest mortality and possum density are not necessarily closely related. But what effect do possums have on the forest ecosys- tem in the longer term and how does that affect the possum population itself?

Long-term responses in the Orongorongo forest The long-term study at the Orongorongo site has provided the most detailed understanding of the responses of a New Zealand forest to the presence of brushtail possums. They were introduced to this forest in 1893 and for the next 50 years the population provided a livelihood for trappers. Mason (1958) examined the stomach contents of 135 possums trapped during 1946–47 and at the same time observed the relative abundance of plant species near the trap lines. She concluded that there was a marked preference in the diet for a few species of plants, which were more frequently present in the stomachs than their frequency in the forest. The four most favoured species in descending order were fuchsia, Fuchsia excorticata, northern rata, Metrosideros robusta, titoki, Alectryon excelsa, and kamahi (Fig. 7.16). Twenty-five years later, in 1969–73, Fitzgerald (1976) carried out a similar study at the same site but using plant remains in possum faecal pellets, instead of stomach contents, to determine food preferences. The possums still selected a few species in a night’s foraging but the preferred species had changed markedly: fuchsia and titoki were absent, rata and kamahi comprised half the intake and several other species, not present in possum stomachs 25 years before, now comprised up to 10% of the diet. Except for a few heavily browsed plants, the two highly preferred species in 1946–47 were no longer to be found in the study area: selective browsing had removed the species from the forest and the possums had turned to other less favoured food species. In a third study during 1975–89, the dietary selection by the possums continued to alter (Campbell 1990, Allen et al 1997). Northern rata continued to be a main component of possum diet but two other preferred tree species, kamahi and tawa, Beilschmiedia tawa, continued to decline, the latter species to virtual extinction. Since this is marginal habitat for tawa the added stress of brushtail possum browsing may have been sufficient to eliminate it from the forest. While these species declined, some other understorey species, such as supplejack, Rhipogonum scandens, increased in the possum diet. For other species, such as hinau, which possums do not browse, the fruit is an important source of high quality food in autumn but, because it is not browsed, hinau remains common in the forest. Possums have never browsed several other species in the forest and these contain up to 12% by weight of terpenes, compared to less than 2% in the favoured species (Brockie 1992), which may contribute to their unpalatability. One of the major species in the diet, northern rata, is the dominant tree species in the forest and during 1970–75 Meads (1976) studied possum browsing on 24 large trees. In that time 262 Life of Marsupials

40

1946-47 1969-73 1976-85 30

20 Occurrence (%)

10

0

eus

florus

epanda

Alectryon excelsus Beilschmedia tawa Fuchsia excorticata Leptospermum spp. Melicytus rami

Metrosideros fulgens Metrosideros robusta Brachyglottis r Weinmannia racemosa Pseudopanax arbor Rhipogonum scandens Geniostoma ligustrifolia Figure 7.16: The changing diet of the brushtail possum, Trichosurus vulpecula, over 40 years in the Orongorongo forest, New Zealand. In the first study stomach contents were analysed for relative abundance of each plant species, and in the latter two studies fresh faecal pellets were analysed. Because of wide variations in relative abundances in different months of the year the figure only represents the broad changes that have occurred, with species on the left having declined and those on the right having increased in the diet of possums. Data from Mason (1958) (1946–47), Fitzgerald (1976) (1969–73) and Allen et al (1997) (1976–85).

one-third of the trees were heavily browsed to the point of defoliation and three trees died, while other nearby trees were untouched or only lightly browsed. Several possums would be seen browsing together in the heavily browsed trees, indicating that they were especially attrac- tive to the possums, possibly because browsing stimulates new growth, which is more palatable. Conversely, it may be that the unbrowsed trees contained secondary compounds that made them unpalatable, as is known to occur in species of Eucalyptus. Meads (1976) predicted that all the rata trees at Orongorongo would progressively die from over browsing within a few years, but during the next 20 years this did not happen, with 21 trees still alive in 1990 (Cowan et al 1997). That was a year when the possum density was high (Fig. 7.13). In the next three years seven of the surviving trees were severely browsed but only one died by being blown over. While possums have browsed these trees for 30 years, most of them have survived. Despite these many changes in the composition of the forest over 50 years, the possum population has remained at about the same density throughout (Fig. 7.13). This suggests that the plant species that have replaced the earlier preferred species can supply the essential nutrients for possum reproduction. What is clear from the Orongorongo study, and from studies in other Life in the trees: koala, greater glider and possums 263

parts of New Zealand, is that brushtail possums actively select their food plants rather than rely solely on the most available species. In many different studies, tree species of rata make up a large proportion of possum diet and yet the nutritional content of rata leaf is low and secondary compounds quite high. The values for dietary nitrogen of northern and southern rata are less than 1% dry weight, which is at the low end of the range found in Eucalyptus species in south-eastern Australia (Braithwaite et al 1983, Lambert et al 1983). However, the values for potassium and phosphorus were in the upper part of the normal range for Eucalyptus species that support arboreal marsupials in Australia (Braithwaite 1984, Cork 1992). The levels for total phenolics are also in the range of values measured in Eucalyptus in south- eastern Australian forests but Eucalyptus foliage with the low nitrogen values and high total phenolics found in rata would not support populations of arboreal folivores in south-eastern Australian forests (Cork 1992). Why then is a species of such low nutritional value such a preva- lent species in possum diet? One important difference between eucalypt forests and New Zealand forests is the far greater species diversity of the latter, so that the possums are not dependent solely on the low nutrient foliage of rata. All dietary studies on possums in New Zealand agree that two or three species comprise the bulk of the diet, with about 12 other species of plants making up small proportions of the total. Rata and kamahi may provide the base energy intake for possums, but the nitrogen requirements of late lactation must come from the minor species of the diet with a high nitrogen content, or from insects (Cowan and Moeed 1987), fungi or birds eggs. The recent discovery that possums take the eggs of rare native bird species has caused alarm in New Zealand. It would be interesting to know whether the predation on birds coincides with the period of late lacta- tion. This recalls the observations that New Guinea cuscuses will kill and eat small vertebrates (Flannery 1995). Since conditions favourable for sustained breeding determine the long-term size of a possum population, species diversity in the understorey must be a critical factor. Forest that lacks the diversity of food plants needed for successful breeding, such as southern beech forest, will only support a small population or become a sink for surplus young possums bred in more favour- able forests.

Responses of the plant species Another long-term response of palatable species in heavily browsed forest associations may be rapid selection for unpalatable genotypes with higher levels of secondary compounds, like sider- oxylonal in Eucalyptus mentioned earlier. The potential for polymorphism in palatability of food trees can be seen in Meads’ (1976) study of rata trees at Orongorongo. It can also be seen in the patchy mortality of southern rata at sites like Adams River in Westland, where possums have only recently arrived. One reason that some forests have not undergone defoliation and mortal- ity could be that the understorey species, essential for successful reproduction, contain levels of secondary compounds that make them unpalatable to possums. For instance, in the Catlins forest and on Banks Peninsula fuchsia has not disappeared from the understorey, as it has in the Orongorongo forest, perhaps because it is less palatable there.

An arms race in process? For New Zealand it is unlikely that any indigenous forest areas are unoccupied by possums and what is happening is the first skirmish in an arms race between an adaptable folivore and a new food resource, in which both herbivore and plants are undergoing rapid selection. In this arms race some plant species are being eliminated or reduced to a rare status, especially those at the ecological limit of their range. Species that are naturally unpalatable survive and replace them, 264 Life of Marsupials

and other edible species undergo selection for individuals that are genetically less palatable to the folivore. Eventually, what will remain will be either, forests that were always less susceptible to browsing and will survive, or forests in which genotypes for unpalatability have undergone strong selection. In both cases the folivore must adapt its breeding strategy to the changing resources available from less palatable or less nutritious species. From the evidence in Australia and New Zealand, the brushtail possum is very responsive to nutritional factors and, if it cannot support lactation, its population will decrease. When the brushtail possum was introduced to New Zealand it was, in a sense, returning to the vegetation of its distant ancestors; this is the latest interaction of an arboreal marsupial and a forest ecosystem, comparable to the arms races that presumably occurred in Australia during the Miocene, as Eucalyptus species replaced rainforest. But in New Zealand the arms race is being played out much faster and it offers unique opportunities to study evolution in action.

Possum control in New Zealand – is it worthwhile? If this conclusion is correct, what can be achieved by reducing possum populations in the native forests of New Zealand? In 2000 NZ$45 million was spent on possum control, four times as much as on all other pest control activities. About half of the expenditure on possum control is aerial distribution of 1080 poison for conservation of the forests and native wildlife. While the poison bait is designed to minimise the poisoning of non-target species, especially native birds, there is evidence that species other than possums are killed (Spurr 2000). But more pertinent is how effective is widespread poisoning at reducing or eliminating possum populations and thereby restoring the forests and their original inhabitants? According to Veltman (2000), the evidence that possum control helps native forest animals is weak, because the studies have been done as part of control operations, not as designed experiments. One study that was designed to test the effect on forest vegetation of reducing the possum population (Payton et al 1997) showed that reducing the possum population by 87% and maintaining it at that low density prevented further decline in the forest but not recovery to its former state. Part of the uncertainty about the effectiveness of control measures is that if small popula- tions of possums can affect susceptible plant species or forest communities, poison campaigns to reduce the population may not have much or any effect. Conversely, other forests that can support large populations of possums without serious effects will also not respond if the possum population is reduced. Without a better knowledge of how the arms race between possums and forest is being played out in different parts of New Zealand, attempts to alter it in favour of the forest and its original inhabitants will continue to be inconclusive. The only certain control is total removal of possums from an isolated area, as was done on Kapiti Island. Recovery of the vegetation and some of the wildlife has followed. But this is only possible on an island. It is surely more realistic to accept that the brushtail possum is now a permanent component of the New Zealand biota, to which the native vegetation and wildlife will accommodate as the arms race is played out.

General conclusion Living on the leaves of forest trees has severe limitations for a small marsupial. It cannot respond to low nutrient leaf by increasing its intake because its gut cannot contain it, and it cannot increase throughput because of the damage to the cells that line the gut and consequent protein loss. If the plant species produces secondary chemicals, there is an additional energy cost in detoxifying them. But the critical factor for species survival is the final stage of lactation: it is energetically costly and the demands for protein by the growing young require that nitrogen Life in the trees: koala, greater glider and possums 265

intake by the female is increased. In Australia these constraints determine the distribution of arboreal folivores in eucalypt forests. Species of eucalypts that have nitrogen values less than 1% dry weight and high levels of phenolics either do not support arboreal folivores or support only non-breeding animals. Each of the species that has been studied thoroughly has responded to the challenge of eating Eucalyptus leaf differently: the koala exploited the metabolic advantages of large size and low metabolism; the greater glider conserved energy expenditure on travel by gliding; the ringtail possum has gained much added benefit from the leaf by recycling; and the brushtail possums and the cuscuses have exploited the advantages of being generalist feeders, able to use whatever is available, including animal protein. Of the species we have considered the koala, the greater glider and the bobuck are resource conservers, whereas the common ringtail possum and the brushtail possum are more versatile and adaptable: they occur in a wide range of habitats and are represented by several subspecies or closely related species across the Australian mainland and Tasmania. Conversely, the conservative species are wholly restricted to the moist habitats of tall Eucalyptus forests of Eastern Australia: none occurs in the same type of forest in Tasmania or Western Australia, presumably because they were unable to cross the heathlands of the former Bass Strait land bridge or the treeless Nullabor of Western Australia. Being so dependent on the integrity of the forest habitat, the conservative species have been much more vulnerable to change brought about by forestry practices and land-clearing for agriculture. This also applies to the arboreal folivores of the rainforest. The only way to ensure the continuing survival of such species is the retention of large areas as nature reserves on high nutrient soils. By contrast, the opportunistic species have adapted to the changes brought about by European settlement, are generally secure and, in some cases, have thrived under the altered conditions. This is especially so of the northern common cuscus, which has become established in many Pacific islands after human introduction, and of the brushtail possum, which is transforming the indigenous forests of New Zealand. This page intentionally left blank Chapter 8

Wombats: vegetarians of the underworld Common wombat, female and young; pen and ink sketch by Tricia A Wright. Wombats: vegetarians of the underworld 269

ombats are the only survivors of a diverse range of large herbivores that evolved in the mid Miocene epoch to exploit the grasslands of Australia. They are bulk feeders that Wferment plant tissue in a large colon at the back end of the gut. The equally diverse kanga- roos also evolved in the Miocene to feed on grasses, but they ferment it in a capacious forestom- ach. While only three species of marsupial hindgut grazers survive today, there are 72 species of kangaroo. Similarly, on other continents in the Miocene, placental mammals evolved to process grasses in either the foregut or the hindgut. As among marsupials, the placental foregut fermenters – antelopes, cattle and sheep – have greatly succeeded, and the hindgut fermenters – the horses and rhinoceroses – have declined. Why is this so? The history of the wombats, and by inference that of the extinct diprotodonts and zygomaturans, may help to understand why fermenting grass and herbage in the colon is less effective for a large mammal than fermentation in the stomach.

Wombat antecedents Fossil wombats are recognised by two features that distinguish them from the other marsupial grazers of the Miocene, namely a single pair of incisors in the upper jaw and open-rooted teeth (Murray 1998). The first of these features is seen in the earliest fossil species, Rhizophascolonus,

Phascolonus gigas, 250 kg (5my – 40k)

50 Phascolomys medius, 80 kg (2my – 40k) 40 30

20

10

0 CM

Southern hairy-nosed wombat, 40 kg (2my – present)

Warendja wakefieldi, 20 kg (10my – 40k)

Figure 8.1: Outline reconstructions of the three main genera of extinct wombats, compared to the southern hairy-nosed wombat, Lasiorhinus latifrons. From Murray (1998). 270 Life of Marsupials

from the late Oligocene epoch, 23 million years ago, and the open-rooted teeth occur in all later species from the late Miocene. The smallest and least specialised species is Warendja from the late Miocene to the Pleistocene epoch, which in its dentition and adaptations for burrowing most nearly resembles a very large rabbit (Fig. 8.1). Then in the Pliocene epoch, 5 million years ago, giant wombats arose, as did giant diprotodonts and kangaroos. Ramsayia and Phascolomys medius were widespread from south and central Australia to northern Queensland and western New South Wales, while the giant of them all, Phascolonus gigas, estimated to have weighed up to 250 kg, had a wider distribution, which included Western Australia and Tasmania. The living genera Vombatus and Lasiorhinus first appear in the Pleistocene epoch, about 2 million years ago (see Fig. 6.3), when they had a much wider range than they do now (Fig. 8.2). Vombatus ursinus, the common wombat, is the largest of the living species with a body mass up to 40 kg. It is an inhabitant of the moist forests of south-eastern Australia. At the northern limit of its present distribution it occurs only above an altitude of 600 m, but further south it extends from sea level to the tree line of the alpine regions. It formerly occurred in Queensland and throughout south-eastern Australia, as far as the Murray River in South Australia, in Tasmania and several islands in Bass Strait. Until about 40 000 years ago the closely related Hackett’s wombat, Vombatus hacketti, occurred in Western Australia (Fig. 8.2), which implies that the common ancestor extended right around the southern part of the continent. As the climate warmed during the past 12 000 years both species were presumably driven southwards, until the western species was deprived of suitable habitat and went extinct, and the living species at its northern limit retreated to higher altitudes.

Northern hairy-nosed wombat

Southern hairy-nosed wombat

Hackett's wombat (extinct)

Common wombat

Figure 8.2: Present and former distributions of living wombats. After Wells (1989) and Strahan (1995). Wombats: vegetarians of the underworld 271

Likewise, both the southern hairy-nosed wombat, Lasiorhinus latifrons, (32 kg) and the north- ern hairy-nosed wombat, Lasiorhinus krefftii, (35 kg) were formerly more widespread across inland Australia and are now restricted, respectively, to small isolated populations in South Australia and Queensland. These shrinking ranges have been further reduced since European settlement by the competition for native grasses from introduced stock and rabbits. Indeed, the northern hairy-nosed wombat is now reduced to less than 100 animals in a single reserve of 3000 ha and, perforce, much less is known about it than about the other two species.

Saved by burrowing? One important reason why the three living species of wombat have survived when all the other hindgut fermenters have gone may be their habit of making and occupying underground burrows. Indeed, Johnson (1998) thinks that burrowing was a relatively recent adaptation in the history of wombats, being a response to the increasing aridity of the continent in the last 2 million years, and that the ancestors of today’s wombats did not burrow. No other large mammal that lives on grass makes burrows. There are many small grass-eating burrowers (10 g–10 kg body mass) but all larger mammals that burrow are either carnivores, like badgers, Meles, or feed on concentrated food: armadillos, Dasypus, pangolins, Manis, and aardvarks, Aardvark, live on colonial insects; and porcupines, Hystrix, live on roots, tubers and fallen fruit. This is because the work required to excavate a big burrow increases by the cube power of spoil removed, while the strength of muscle to do it increases by the square power. Thus, large burrowers have powerful short limbs for digging and, to provide the necessary energy, they need a more concentrated fuel than grass. Also, a large grazing herbivore needs to range widely to find sufficient food to maintain itself and it cannot do this if tethered to the vicinity of a burrow, with short limbs adapted for burrowing. So, how have wombats managed to overcome these two constraints that other large herbiv- ores have not? We can approach this question by seeing how wombats live on grass; then what the costs and benefits of burrowing are; and then consider whether burrowing is a recent adapta- tion among wombats.

Diet of wombats When given a choice, all three species of wombat prefer to eat tough native perennial grasses and sedges, but in adverse conditions they will eat introduced pasture species, as well as forbs and the leaves of woody shrubs, including eucalypts, such as mallee (Hume and Barboza 1998). The common wombat living in eucalypt woodland favours species of Poa, Danthonia and Themeda, and the rush Lomandra, which it finds within a home range of about 20 ha. The southern hairy-nosed wombat feeds mainly on Stipa nitida, which grows around the warren complex and, because it grazes nearby, the grass becomes close cropped, which induces it to put out new shoots favoured by the wombats. As a result the home range of this species is about 4 ha, which is extraordinarily small in comparison with the 380 ha home range of the western grey kangaroo, Macropus fuliginosus, in the same environment. These grazing halos around the burrow complexes expand as the plant growth declines and the wombats range further to feed. The northern hairy-nosed wombat, likewise, feeds on tough native perennial grasses that occur in its natural environment, preferring species of Heteropogon, Enneapogon and Aristida. Its home range is about 25 ha with a core area of about 4 ha. The fact that wombats can obtain sufficient food within such small home ranges suggests that their metabolic requirements are much less than those of other herbivores, which in turn explains how they have avoided the second of the two constraints to burrowing for a large herbivore. 272 Life of Marsupials

A dental mill for treating bulky plant food The structure of the wombat skull and the form of its teeth eloquently express its adaptation for processing tough plant material. At the front the single pair of incisors in the upper jaw are set transversely to form a sharp cutting tool that bears against the single pair of incisors in the lower jaw. As the upper lip is split the wombat can crop the vegetation very close to the ground. There are no canines so there is a substantial gap, or diastema, between the incisors and the cheek teeth, which comprise one premolar and four molars in each jaw. All the teeth are long and curved with open roots, which mean that they continue to grow throughout life, like the teeth of rabbits and rodents. The upper molars curve outwards and the lower molars curve inwards (Fig. 8.3), the two sets meeting very close to the midline of the skull (Murray 1998).

Section through wombat snout to show position of molar teeth

The chewing sequence

grass stem

6 5 4 3 2

7 1

Figure 8.3: Arrangement of the molar teeth in the skull of the southern hairy-nosed wombat and how they are used to cut herbage very finely by a side to side movement. After Murray (1998).

In a young wombat the molars are four-lobed, somewhat like the molars of the koala, Phascolarctos cinereus (see Fig. 7.7a), but the enamel is soon worn off the crown so that the inner dentine is exposed and gets worn away faster than the hard enamel on the sides of the teeth. As a result, the enamel on the inner curvature of each tooth forms a strong, sharp blade and that on the outer curvature a thinner crest. When the wombat chews food it moves the lower jaw sideways as well as upwards, so that the blades of the lower teeth meet the blades of the upper teeth and cut the grass stems. The cut pieces are then held within the opposing basins and, under increasing pressure, are sheared off Wombats: vegetarians of the underworld 273

as the ridges pass each other. The wombat makes these short powerful chewing strokes with its large, specialised jaw musculature, chewing on alternate sides, using only the muscles of that side. It has very large masseter muscles, which originate on the zygomatic arch under the eye socket and insert on the back end of the lower jaw. But because the zygomatic arch is set well away from the midline and the molar row is set near the midline, the masseter muscles can exert a very strong lateral force to the jaw, as well as the usual compression force. This disposition of the chewing muscles and skull gives the wombat head its broad, flat appearance. All these features of the chewing complex of wombats are very similar to those of large rodents, such as marmots, Marmota, and different from the disposition of teeth and jaw muscles of other grass eaters, such as kangaroos (see Chapter 9) and sheep. Since the extinct wombats from the mid Miocene had similar teeth and skull anatomy, it is reasonable to conclude that they also processed a similar diet of tough grasses and similar plants.

Adaptations for feeding on coarse bulky herbage The dental mill, especially of the southern hairy-nosed wombat, grinds the food to a mean parti- cle size half that of the western grey kangaroo, grazing on the same species of plant (Wells 1989). This releases more of the plant cell contents for digestion. Proteins and soluble carbohydrates are digested in the simple stomach and absorbed in the relatively short small intestine (Hume 1999). This provides all the wombat’s protein needs, but it obtains much more of its energy needs from bacterial fermentation of the residual plant matter, mainly cell walls, in the hindgut, like the koala and the greater glider, Petauroides volans, do (see Chapter 7). However, unlike in those species, the caecum is tiny and the colon, which is the largest part of the wombat’s gut, is the fermentation chamber. It has two clearly defined parts, the anterior being the main fermentation chamber and the hind part being the site of water reabsorption. In the desert-adapted southern hairy-nosed wombat the fermentation chamber is small and the water absorbing part is large, while in the forest-dwelling common wombat the proportions are reversed, presumably because it has less need to conserve water. Fluid passes through the gut of southern hairy nosed wombat in 49 h and particulate matter in about 70 h, which is slower than in eastern grey kangaroos (15 h and 35 h, respectively, Hume 1999) and the opposite of the koala (see Table 7.1). The combination of a large capacity colon and a relatively slow rate of passage through it maximises the opportunity for bacteria to hydrolyse the plant cellulose to short chain fatty acids, SCFA, and to return them to the wombat’s circulation. As in the koala and common ringtail possum, Pseudocheirus peregrinus, urea produced by protein breakdown is returned to the gut, where bacteria in the colon use it to synthesise new protein. Although the wombat cannot make use of this bacterial protein itself, like the common ringtail possum does by reingesting caecal contents, it does benefit indirectly in two ways: the recycled urea allows the bacteria to proliferate and this increases the rate of fermentation in the colon, making more SCFA available to the wombat; and since less urea is excreted in urine, nitro- gen is conserved, which is reflected in the extraordinarily low digestible maintenance nitrogen requirement of both species of wombat (common wombat, 71 mg N/kg0.75 per day, southern hairy-nosed wombat, 116 mg N/kg0.75 per day). These values are at the low end for herbivorous marsupials (see Table 7.1) and lower than for any large placental herbivore. Both aspects go a long way to explaining how wombats can thrive on grasses with very low energy and protein content. Recycling urea to the colon also conserves water that would otherwise be required to excrete urine. Although wombats do not excrete very concentrated urine, their water turnover rate is among the lowest recorded for herbivorous mammals (Wells and Green 1998). This is very important for a desert living species like the southern hairy-nosed wombat and is achieved by 274 Life of Marsupials

efficient resorption of water in the posterior half of the colon, so that the water content of the dung can fall to under 40%, especially in the summer. This is as dry as camel dung and lower than that of donkey, Equus asinus, dung (64%) under comparable conditions (Schmidt-Nielsen 1964). Because of their very low water turnover, southern hairy-nosed wombats can derive all the water they need from the water content of the plants they eat and the metabolic water from the oxidation of the carbohydrates in the diet. They do not need to drink.

Wombat energetics The energetics of the southern hairy-nosed wombat further demonstrates the special adapta- tions of this species to its harsh environment. Its standard metabolic rate (SMR), measured in captivity was 130 kJ/kg0.75 per day. This is very low, being only 64% of the marsupial average (see Table 1.2), or only 42% of the placental average (Wells 1978, Wells and Green 1998). Its thyroid hormone levels, which are a reflection of SMR, are the lowest recorded for any mammal. By contrast, the SMR of the common wombat, living in a more benign environment, is near the marsupial average, 210 kJ/kg0.75 per day (see Table 1.1, Gowland 1973). Field metabolic rates (FMR) have not been obtained for any species of wombat but for the southern hairy-nosed wombat, Wells and Green (1998) calculated the margin between the minimal energy require- ments and the wombat’s actual daily intake of perennial grasses, based on water turnover data and the water content of the plants consumed. They showed that the amount of forage consumed would provide 2.8 times the maintenance energy needs of the wombats, which is very similar to the ratio between SMR and FMR for the koala, greater glider, and for the common ringtail possum during late lactation (see Table 7.2). Thus, its preferred diet is sufficient to support this most critical period of the life cycle, provided there is enough food available. Southern hairy-nosed wombats are extremely frugal in their nitrogen, water and energy needs and this enables them to maintain body mass on very low quality food, even during drought. They outperform the donkey, a placental hindgut fermenter adapted to a similar habitat, mainly because of their lower metabolic rate and hence lower food intake and slower rate of passage (Hume 1999). However, the donkey lives in the open and can tolerate high air temperatures, whereas southern hairy-nosed wombats become severely distressed at air temperatures above 30°C. Wombats can survive in their present environment only because of their burrows, which enable them to avoid ambient temperatures outside their thermoneutral zone. Furthermore, during drought they can reduce their activity and conserve energy and water by staying in the burrow, and in cooler weather they can emerge to bask in the sun, which provides heat that would otherwise be provided by food. Although less is known of the thermoregulation of the other two wombat species, neither can tolerate high ambient temperatures, which they also avoid by retreating deep into their burrows (Brown 1984). For the common wombat the burrow also provides protection against extreme temperatures during forest fires: when McIlroy (1973) followed the fate of common wombats as their forest was cleared-felled and then burnt, none succumbed to the intense fires above their burrows.

Burrow architecture and environment While many observations on wombats have confirmed the central importance of burrows, the interior of wombat burrows was for long only known indirectly, by probing from the entrance or by destructive excavation. In Furred Animals of Australia, Ellis Troughton wrote in1957: ‘Explored by “open cut,” burrows have measured up to 100 feet (30 m), and a child might crawl through to Wombats: vegetarians of the underworld 275

the nesting chamber in some of them’. Three years later a schoolboy in Victoria did exactly that and provided a remarkable description of the interior of burrows and of the common wombats that lived in them. Peter Nicholson, aged 14, was able to crawl along the burrows and turn around in the wombat’s rest chambers, and he described what he found in his school magazine (Nicholson 1963). He discovered that the burrows were up to 18 m long, had short branches and one or more nest chamber, some lined with bracken and eucalypt leaves. He also observed that the wombats moved deeper into the burrow during the day and gradually moved to the entrance

(a) Entrance

Plan of warren 14

4 10 3

1m 18

Port Section of warren 1m

Shade Entrance 3 10 14 18

4

(b)

40 Summer Ambient 3 Wombats out temperature of burrow 30 2 Portholes 3-18 temperature 1 20

0 Winter 3 30 Temperature (C) 20 2

10 1 Number of wombats out warren

0 0 0 4 8 12162024

Time of day (hour)

Figure 8.4: How the southern hairy-nosed wombat, Lasiorhinus latifrons, uses its underground shelter: (a) the plan and section of a large burrow complex to show how portholes are used to measure the burrow environment; (b) throughout the diurnal cycle the temperature within the burrow remains constant, at 26°C in mid summer and 14°C in mid winter compared to the temperature outside the burrow; and their daily pattern of activity enables the wombats to avoid the extremes. From Taylor (1998) and Shimmin et al (2002). 276 Life of Marsupials

as the light intensity in the burrow declined before sunset. In later studies on the same species, McIlroy (1973, 1977), who could not emulate Nicholson by crawling in himself, determined burrow dimensions and pattern with a plumbers ‘snake’ and by triangulating the position in the burrow of wombats to which he had attached radio transmitters. He confirmed the maximum length was about 20 m and most large burrows were without branches and contained two or more resting chambers. Usually there was one resting site about 2 m from the entrance, where the wombat would lie up until the light intensity and/or air temperature outside was suitable for it to emerge. Most burrows were dug in sloping hillsides and near watercourses, with the result that they can be much deeper underground than the burrows of the hairy-nosed wombats. The burrows of the latter two species (southern and northern) are built on flat land and usually begin with an entrance crater from which one or more burrows lead away, descending to a maximum depth of 2 m in the southern species and up to 4 m in the northern species. Shimmin et al (2002) have developed a different way to study the burrow environment, and the physiology and behaviour of southern and northern hairy-nosed wombats in their burrows, without disturbing them. Narrow boreholes (50 mm diameter) are sunk vertically from the surface to penetrate the roof of the burrow and are lined with close fitting plastic pipe. Then a device with a light source and a video camera is lowered down the pipe to view the direction and structure of the burrow and its contents. This shows where the next borehole should be sunk, and so on, until the entire burrow complex has been mapped and a vertical profile obtained (Fig. 8.4a). Most burrows had a single entrance and ranged in length from 4 to 28 m, but large warrens were far more complex with the largest having 28 entrances and a total length of tunnels of 89 m on several levels. The depth of the tunnels depended on the soil type, being about 1 m in firm soil and up to 2 m in sandy soil. In five warren complexes of the northern hairy-nosed wombat (Shimmin 2001) the archi- tecture was simpler, with four having one entrance and a tunnel length of 10–15 m, and the fifth having three entrances and a tunnel length of 54 m. All the tunnels were dug deeper than those of the southern hairy-nosed wombat, up to 4 m underground. After completing the survey of a warren complex the port holes are capped on the surface and at any future time devices can be lowered down to view the burrow and its inhabitants, or to measure air temperature, airflow, humidity and gas composition.

Burrow environment During the first two years of the study by Shimmin et al (2002) on the southern hairy-nosed wombat, the air temperature above ground ranged from 2°C to 38°C, while air temperatures in burrows remained within the range of about 14°C in mid winter to 26°C in mid summer (Fig. 8.4b), which is within the thermoneutral zone of the wombat. Temperature stability is even more apparent when measured through the day: beyond 3 m from the entrance the temperature remained constant, despite wide variation outside. In the simple burrows with a single entrance airflow is negligible, but even in the complex warrens it was slight. As a result the humidity a few metres in from the entrance was higher than outside and rose progressively along the tunnel to almost full saturation: this also contributes to water conservation for the wombat. Since there were no differences in air temperature along the burrows, it is probably the higher humidity that influences the wombats to tunnel further than 3 m. Wells (1978) had observed previously that when a wombat returns from foraging outside it retreats into the depths of the burrow. When it is preparing to emerge the following night it moves towards the burrow entrance but if it encounters high temperatures or low humidity it retreats again. In winter when the ambient temperatures are much lower it remains about 3 m from the entrance in the twilight zone until sunset. Taylor (1998), in a study extending over Wombats: vegetarians of the underworld 277

20 years, confirmed this pattern by fitting all the entrances of several warren complexes with swinging flaps, so that he could record the traffic in and out of the burrows throughout the 24 h and during different seasons of the year (Fig. 8.4b). This showed that wombats emerge from the burrow in the evening when the outside temperature is the same as the burrow temperature and return in the early morning while the ambient temperature is still below the burrow temperature. Thus, wombats take full advantage of the microclimate inside the burrow complex to escape the climatic extremes outside and, by remaining within their thermoneutral zone at all times, conserve water and energy. The burrow provides wombats with the essential means to survive all extremes of the climate where they live.

Breathing in a burrow Another important matter for a life underground in a confined space is how concentrated the carbon dioxide (CO2) from expired air becomes in unventilated burrows during the day when the wombats are resting, and how the wombats cope with high levels. In unoccupied closed burrows the CO2 and oxygen composition of the air is not different from outside, so there must be sufficient diffusion through the soil for the gases to equilibrate. However, in occupied burrows,

CO2 levels rose to as high as 2.6%, well above the normal 0.04% in air, and oxygen levels fell to 16% from the normal 21%. Such changes are understandable when a large animal is confined in a small space, especially when engaged in digging at the end of a long burrow. Most mammals, including humans, exposed to such high levels of CO2 would rapidly increase their respiratory rate and die. Even though burrowing mammals can tolerate low oxygen and high CO2 better than non-burrowers, wombats digging at the end of a tunnel have to abandon their efforts after about five hours and seek fresh air (Frappell et al 2002). This is, nevertheless, a long time and their very low metabolic rate must be one factor that lets them tolerate high concentrations of CO2.

Consequences of using burrows Two consequences of this absolute reliance on the burrow environment are the costs of construc- tion and maintenance of burrows, and the limits they impose on the distance that a wombat can forage at night.

Burrows represent a large capital investment of time and energy The cost of burrow construction is considerable. Steele and Temple-Smith (1998) calculated that the volume of large warrens was between 8 and 13 m3 and they observed one subadult wombat dig a tunnel 4 m long in 30 minutes. This is about 1 m3, or more than one tonne of soil, a consid- erable amount to move, especially if it has to be pushed out some distance along the pre-existing burrow. This was probably exceptionally fast because the wombat was endeavouring to get away from the destructive excavation of the warren behind it. In undisturbed warrens, the longest new work that Shimmin et al (2002) recorded was 10 m in four weeks. Most burrows only undergo minor enlargements and alterations. Steele and Temple-Smith (1998) estimated the volume of a burrow to be greater than the volume of the mound of spoil outside it and concluded that the soil in the burrow must be compressed. Wombats compress the floor by frequent passage through it and they can raise the height of the tunnel by pressing their flattened head and hips against the roof. It is said that wombats also use this technique as a defence against predatory dogs, where the dog’s head is crushed against the roof of the burrow. G. Shimmin (pers. comm. 2003) has estimated the cost of burrow construction at 14 times SMR, which is far more than the cost of normal daily activities (2.8 times SMR). Burrow construction, therefore, represents a very large investment of time and energy and might be 278 Life of Marsupials

expected to be rarely undertaken; and sharing existing burrows a common occurrence. It is not known how old the large warren complexes are but it is likely that they are used by many genera- tions of wombats. While young wombats make short burrows within existing complexes, estab- lishing a wholly new burrow would be a large undertaking, while the benefits of being allowed to remain in the natal territory and inherit the burrows would be very great. In the northern hairy- nosed wombat there is some evidence that the young wombats remain in their natal warren and adult females establish new burrows (Johnson and Crossman 1991) but in the southern species two-year old wombats are driven from the home warren and few of this age group survive (Wells 1978).

How wombats use their investment Southern hairy-nosed wombats live in large warrens with many interconnected entrances, which are their prime refuges, shared by up to 10 wombats. Around this central warren is a circle of small, simple burrows 100–115 m from the central one (Wells 1978, Steele and Temple-Smith 1998). The resident wombats graze as far as the perimeter of small burrows, an area of about 4 ha,

N

0 100 200 300 metres

Figure 8.5: Pattern of burrows (Q) and their use by five clusters of northern hairy-nosed wombats, Lasiorhinus krefftii, as disclosed by the individual home ranges of the males (—) and females (--) in the Epping National Park. After Johnson and Crossman (1991). Wombats: vegetarians of the underworld 279

thereby causing the halo of a close-cropped sward of new shoots. The small perimeter burrows provide a temporary refuge to extend the grazing area further, or are used by young wombats that have been driven from the central warren. The social structure within the warren is not known precisely but there is some separation of sectors of the common grazing area between members of the group, and males display territorial behaviour towards other warren groups, presumably to defend the essential food resource in the grazing halo and also perhaps the warren refuges. The territory is also marked by scratches and droppings, which give off a recognisable odour. The northern hairy-nosed wombat is similarly adapted to living in a hot dry climate on perennial grasses and avoids extreme temperatures by use of its large, deep burrows, which it only leaves to graze for a few hours each night (Johnson 1991). In the summer the wombats only graze outside for 2 h each night, being able to obtain sufficient nourishment in that time because the grasses are more productive in summer than winter, when they graze for 6 h. The individual home range is up to 25 ha but, like the southern hairy-nosed wombat, its core area is about 6 ha. These are extraordinarily small areas to support such large animals. Using radio- telemetry Johnson and Crossman (1991) followed the movements of 10 adult wombats for one year, during which time they used 28 burrows. Their use was not random but small groups of three to five wombats exclusively shared a cluster of burrows within a discrete area (Fig. 8.5). Within the cluster wombats occupied a burrow alone for 70% of the time and shared with one or two wombats at other times. Females shared more frequently with females than with males, which were usually solitary. Outside the burrows the feeding ranges of wombats of the same

1820 0220 1945 2215 2155

0400 0100 1950 1830 0030 0140 2040 2030

0045 2130 2045 1915 2200

19 November 20 November

1930 0305

2000 1940 2050 1935 0150 0105 2350 0205 0308 2115 2010 2240 2020 2300 2100 2340 0201

2400 2320

0150 24 November 25 November 200 metres

0045 Location at specified time Creek Route Parkland (semi-open feeding area) Burrow

Figure 8.6: Movements of an individual female common wombat, Vombatus ursinus, wearing a radio transmitter, during four successive night’s grazing, to show how she used the five available burrows to sleep in each day and travelled to the same feeding site each night. After McIlroy (1976). 280 Life of Marsupials

sex did not overlap but the ranges of females overlapped those of one or more males. Piles of dung and strong smelling urine marked the boundaries of feeding ranges and burrow entrances. While this indicates stable patterns of use, 30 months later more than half the wombats had moved to a different cluster of burrows. The pattern of burrow use of the common wombat is similar. The home range is from 5 to 23 ha, which is achieved by using different burrows on successive nights (Fig. 8.6). By radio- telemetry, McIlroy (1973, 1976) found that each individual used or shared from 3 to11 burrows throughout its home range with five to seven other wombats, and together they progressively used more than 40 burrows. At some times fewer than 12 burrows were known or suspected of being occupied and population size was estimated directly from the number of wombats caught, or indirectly from the number of burrows used. In Tasmania the home range of males was about 10 ha and of females about 20 ha, with overlap between sexes (Taylor 1993). This pattern of use, and the limited home ranges disclosed, is very similar to the other two species, despite the very different environment in which the common wombat lives. It suggests that living underground imposes real constraints on how far such an animal can move away from its large capital invest- ment and its refuge against extremes of climate. It also imposes constraints on reproduction.

Wombat strategies for breeding All three species have a very conservative breeding strategy. There are two teats in the backward directed pouch but only one young is produced. Pouch life is long and, after it leaves the pouch, the young one remains in close association with its mother for another year. This means that a female wombat can, at best, produce one young in alternate years. But even this level of reproduction is very dependent on high quality forage near the burrow complex to support late lactation.

Southern hairy-nosed wombat Breeding in the southern hairy-nosed wombat is a drawn out affair, consonant with the extreme nature of the environment the animals live in, the limitation of the food and the uncertainties of the climate. In years when there is sufficient winter rain to induce plant growth, females ovulate between August and October and males have elevated levels of testosterone and enlarged prostate glands at the same period. However, in years of low rainfall (<200 mm) females do not ovulate and males have low levels of testosterone and small prostate glands. In one study males had sperm in the ejaculates during late August–March but were aspermous at other times (Taggart et al 1998). Copulation occurs in the warren and after a gestation of 22 days the young is born and attaches to one of the two teats in the pouch. Most births occur in October, with a spread of two months either side, and the young remains in the pouch for six months, by which time it is lightly furred, its eyes are open and it weighs about 0.45 kg. It leaves the pouch permanently about three months later, in July or August, when it makes its first appearance outside the burrow and begins to feed on grass. It is not fully weaned until one year old and does not reach adult size until three years. The critical times in this life cycle are during the winter and spring in the year after its birth, when it depends on rich milk and is changing to a grass diet. However, in the few studies of this species, young between one year and three years have not been found and it is thought that many do not survive (Gaughwin et al 1998). Severe drought causes complete failure of recruitment of young due to either a failure of ovulation in females or lack of sperm in males. In one 3-year study during 1976–78 there was a drought in the middle year (Gaughwin et al 1998). In the first and last years 90% of females Wombats: vegetarians of the underworld 281

produced young, but during June 1977 to February 1978 only 1 out of 28 females examined had a young, all the others were anoestrus. Despite the failure to breed that year the body mass and mesenteric fat of the females was little different from the two breeding years, which suggests that anoestrus was not induced by poor nutritional condition but by some factor associated with plant growth. Indeed, there was a good correlation between the proportions carrying pouch young in December–July and an index of plant growth during the previous six months. Such an inhibition of reproduction may conserve nutrient resources otherwise wasted on young that will not survive the drought. And drought is a regular feature of the land the southern hairy- nosed wombat occupies: in the last 100 years there have been only 20 years when the rainfall would have been sufficient for young wombats to survive pouch life (Wells 1989). These climatic constraints on wombat populations are further exacerbated when sheep or rabbits compete with the wombats for the scarce vegetation. On a longer perspective the cycle of wet and dry years may be an essential component of this dryland ecosystem: during favourable years the vegetation grows and herbivore numbers increase; this transfers nutrients into the biomass and depletes the soil; the drought then returns the nutrients to the soil and provides the basis for the next cycle (Wells 1982).

Northern hairy-nosed wombat The northern hairy-nosed wombat is now one of the most endangered species of marsupial in Australia. Known from fossil remains in Victoria, south-western New South Wales and central Queensland, the species had declined to only three isolated populations by the time of European occupation and now lives at only one site in Queensland. Like the southern hairy-nosed wombat, females produce a single young in the spring or summer, which then spends almost a year in the pouch. Also, like the southern species, it is very dependent on rainfall of more than 200 mm in late summer, to support breeding, and its severe decline has been accelerated by competition with stock for essential food plants. The species died out from two of its former sites in Queensland during droughts in 1921 and 1939, and declined to less than 30% of its former range in the remaining site at Epping Forest National Park. Since 1981, when cattle were excluded from the Park, plant biomass has risen to 1721 kg/ha, compared to 138 kg/ha outside the Park, and recruitment of young is slowly increasing (Crossman et al 1994). Destocking is undoubtedly the most effective way to help the species to survive in the longer term. At present it occupies one-tenth of the 3000 ha of the Park and the immediate goal is to encourage it to reoccupy the rest.

Common wombat Common wombats can breed at any time of the year but more births occur between December and March than at others. This fits with McIlroy’s (1973) observations that the testes of males were larger during September to December than at other times. Copulation has twice been observed in the wild (Triggs 1996) and once at a zoo in Germany (Böer 1998). On all three occasions it occurred outside the burrow and began with much vigorous chasing, followed by the male biting the female and bringing her to the ground, whereupon she lay prone and the male lay on his side behind the female when copulation took place. If the female conceives, birth follows about 22 days later but if not, she can return to oestrus 32–34 days later, and have another opportunity to conceive. Females carry small pouch young from May to August and fully furred pouch young in September–October. The young wombat emerges from the burrow for the first time at eight months when it begins to eat grass. A captive-reared wombat at this age was seen to eat small, moist faecal pellets produced by the mother, which were different from the normal dry cubes (Böer 1998), and Triggs (1996) observed an orphan wombat of the same age to consume fresh faecal pellets. These observations 282 Life of Marsupials

recall the caecal pap that the young koala takes from its mother at the stage when it begins to eat eucalyptus leaves. In both species it may serve two functions: to infect the gut with fermentative bacteria and to provide semi-digested plant food during the transition from a milk diet. At this stage the young weigh about 4 kg, compared to the adult mass of 22 kg, and they continue to associate with their mother for up to another year. McIlroy (1973) concluded that females do not return to oestrus and breed again until the close association with the young ceases, which means that a female can only produce a single young in alternate years. This pattern of protracted growth and development is similar to the southern hairy-nosed wombat, even though the common wombat lives in a more benign environment, where drought is not such a prevalent event. Nevertheless, the period of late lactation and early independence of the young wombat is critical to successful reproduction and, as with the other two species, is severely affected by competition from other herbivores. This was demonstrated at the extreme western limit of the species’ distribution in South Australia.

Competition with stock At the time of European settlement the common wombat occurred throughout the southeast of South Australia, as far as the bank of the Murray River (the southern hairy-nosed wombat still occurs on the north-western bank). For 50 years after European arrival common wombats coexisted with cattle and sheep, even as these two species attained their peak populations.

Rabbits present Rabbits excluded 20 Introduced forbs 10

0 50 Introduced grasses 40 30 20 10 0 20 Native forbs 10 Percent cover 0 60 Native grasses 50 40 30 20 10 0 75 76 77 78 79 80 81 82 75 76 77 78 79 80 81 82 Year Year Figure 8.7: Changes in abundance of introduced and native forbs and grasses when rabbits were removed from an area of the Coorong, South Australia. From Cooke (1998). Wombats: vegetarians of the underworld 283

However, 20 years after rabbits arrived their additional grazing pressure suppressed the native pastures and the range of the common wombat contracted south and east. Failure to rear young to weaning was more probably the cause of the decline after rabbits came, rather than simple competition with stock (Cooke 1998). Today the common wombat survives in the extreme southeast of the State, in the Coorong National Park, from where sheep were removed in the 1960s: rabbits are still abundant and wombats scarce. Cooke (1998) excluded rabbits experimentally from an area for seven years and compared it with adjacent areas that still had rabbits. In the exclusion area the introduced grasses and forbs declined while perennial native grasses and forbs increased from 73 kg/ha on rabbit-infested land to 1085 kg/ha on areas cleared of rabbits. Hence, the rabbits accounted for more than 90% of the grazing pressure (Fig. 8.7). The reason that wombats and kangaroos respond to the increase in native perennial species is subtle and relates to the difference between placental and marsupial investment in reproduction (see Fig. 2.24). The introduced species of annual grasses and forbs reach their peak of produc- tion in the spring, which coincides with lambing and the greatest investment in reproduction by sheep, and they dry off in the summer months. However, the native grasses grow more slowly in the spring but remain productive well into the autumn and winter, thereby providing the resources for the main investment of the marsupials in late lactation. By favouring the intro- duced annuals rabbits altered the pasture, making it less able to support the more protracted reproduction of marsupials

Effects of forestry activities on common wombats In their prime habitat of eucalypt forest common wombats must contend with the effects of forestry activities and the transformation of eucalypt forest to pine plantation (Fig. 8.8). In a mixed Eucalyptus forest the response of the wombat population to the clear felling of the forest, the burning of the fallen debris and the later growth of pine trees was followed over 25 years and projected to 38 years by observing other wombat populations in older plantations (McIlroy 1973, McIlroy and Rishworth 1998). The initial road making and clearing of the large trees destroyed about half of the minor burrows and one-third of the intermediate burrows that the wombats used infrequently; but less than 5% of the important major burrows were affected and the wombats quickly repaired them. After the forest was felled the fallen debris was burned in a huge conflagration, which destroyed all the plants on the forest floor as well as the debris. However, none of the resident wombats died and some were seen moving around even before the embers were cool. Recovery of plant cover on the ground was rapid and all the resident wombats gained weight and females carried pouch young. Throughout the subsequent years as the pine trees grew to a large size the major burrows remained intact, although less often used as the wombat numbers declined. The population size was maintained at the pre-felling level until year 5 after establishment of the plantation, when the canopy of the pine trees closed over and the growth of grasses declined. The wombat popula- tion reached its lowest level at year 15, when there was no ground vegetation: subsequently it recovered to half its pre-felling level. Although not specifically investigated, the inference is that once the canopy closed over there was insufficient grass in the plantation to support successful breeding at the previous level. In a later study in the same area (Lindenmayer et al 1999b), few wombats lived in the mature pine plantation, except in isolated patches of eucalyptus forest within the pines: here the density was the same as in the Eucalyptus forest surrounding the pine plantation. 284 Life of Marsupials

3

2

1 Number of wombats per hectare Number of wombats

0 1235791114151619213638

Burnt Pine plantation age (years)

Clear felled

Eucalypt forest

Figure 8.8: Responses of common wombats, Vombatus ursinus, to transformation of eucalyptus forest to pine plantation over 40 years. After McIlroy and Rishworth (1998).

The origin of burrowing in the Vombatidae The question of when burrowing first began among the ancestors of today’s wombats has intrigued people since the discovery in the 19th century of the giant fossil wombat, Phascolonus gigas. The pertinent questions are whether such a large species could burrow and if it did, what benefit it would derive from a burrow; and if not, was it the ancestor of the living species, or were they derived from an earlier, smaller species that has so far not been discovered. The main argument in favour of it being a burrower is that, like all fossil wombats, the giant wombat had relatively short, stout limbs, and the humerus of the forelimb was highly distinctive with a low broad head and large tuberosities to which strong digging muscles would have been attached. The arguments against it being a burrower are based on the very much larger cost of burrowing for an animal four times the size of the living species, and what the benefits would be for such a large animal. The chest girth of Phascolonus gigas has been estimated to be twice that of the living species, so its burrow would have needed to be proportionately larger and longer to achieve the same internal environment as existing burrows (Woolnough and Steele 2001). This means that the burrows would have been very large (14.4–26.4 m3) and long (54–98 m), necessitating the removal of up to 30 tonnes of soil. The limitations of soil strength would have meant either digging deeper or into denser soil to avoid risk of burrow collapse, both of which would increase the cost of construction considerably. Since the species occurred across southern Australia and Tasmania in places with very different soil types, soil type was clearly not a constraint on its distribution: for this reason it probably did not burrow. An alternative idea is that it may have behaved like capybaras, Hydrochoeris, of South America, living near water and excavating shallow pits to lie in; but at 250 kg it was much larger than living Wombats: vegetarians of the underworld 285

capybaras (90 kg). It had extraordinary wide upper incisors and nostrils directed dorsally (Fig. 8.1), so it may have been a semi-aquatic grazer, like the hippopotamus, which also has short, stocky legs. Its remains have been found at Lake Callabonna, South Australia, which in the past was part of a large inland water system. Those that accept the semi-aquatic conclusion propose that burrowing either evolved much earlier in the ancestors of today’s wombats and remains of their direct ancestors have so far not been discovered, or it arose more recently as an adaptation to the increasing aridity of continental Australia in the last 2 million years. The resolution of this question must await more evidence from hitherto undiscovered fossil wombats. This page intentionally left blank Chapter 9

Consummate kangaroos Aboriginal depiction of female kangaroo and pouch young; logo of the Australian Mammal Society. Consummate kangaroos 289

angaroos have been astonishingly successful in Australia and New Guinea. The super- family Macropodoidea includes 19 species in New Guinea and 49 species in Australia, Konly three of which are common to both countries. In Australia another six species have gone extinct since European settlement, so 150 years ago the grand total was 72 species. This is one-third of all marsupial species in Australasia. Members of the kangaroo superfamily range in size from less than 0.5 kg to 80 kg and are found in tropical rainforests, woodlands, open grass- lands, spinifex deserts, mountains, islands and rocky cliffs. Some can travel long distances; others climb trees; some live in caves and two live in burrows underground. Equally remarkable is that much of the diversity in this very successful group of marsupials has only emerged in the past 5 million years – a relatively brief time in the long history of marsupials. Three attributes distin- guish these marsupials from all others and have presumably contributed to their success: foregut fermentation, hopping and embryonic diapause as an integral component of reproduction. Having an enlarged forestomach in which plant material, especially grass, can be fermented by bacteria gives far greater access to the products of bacterial metabolism than hindgut fermenta- tion does, without the need to recycle the contents back to the stomach, as some ringtail possums do. Among placental mammals those groups that evolved a foregut fermentation chamber have also been extraordinarily successful. But no other group of large mammal has evolved bipedal hopping, although it provides economies in transport, especially for desert-adapted species that need to cover great distances with minimum expenditure of energy. The kangaroo pattern of reproduction must have been adopted early in the evolution of the family, since it is common to all living species except one: its selective advantages were presumably greater flexibility in the control of breeding, and enabling the young at birth to be more advanced than those of other marsupials (see Chapter 2). These three attributes of kangaroos cannot be directly discerned from the fossil remains of their ancestors, although fossil teeth and skulls can tell us something about the food they ate and limb structure can tell us whether they could have hopped. As these three features are held in common by the living descendants we can infer that they either arose in the common ancestor and have been retained ever since or, less likely, evolved independently more than once. Fortunately, the history of the Macropodoidea has become much better known in the past 40 years, as a result of the incredible wealth of fossils discovered at sites in central and eastern Australia, and at the Riversleigh World Heritage area in northern Queensland. The oldest known macropod fossils are from the late Oligocene epoch, 26 million years ago, while others are known from the Miocene epoch to the present. The relationships among the living species of kangaroos, wallabies and rat kangaroos have also become clearer, through comparisons of their anatomies, their chromosomes and their protein and DNA sequences. This chapter begins with an introduction to the living species, their relationships to each other and past histories. It then considers how the three common attributes have contributed to their adaptive success, and then looks at how some representative species have exploited their inheritance to live in different parts of the land.

Relationships, distribution and origins The relationships within the superfamily Macropodoidea are best explained in terms of their evolution from forest-dwelling, omnivorous ancestors, through browsers, browser–grazers, to specialised grazers. This evolution is largely reflected in the pattern of the teeth and to a lesser extent in the changing proportions of the limbs from small, running ancestors through quadru- pedal bounders to fully bipedal hoppers. Living macropods are divided between two very unequal families: one small rat kangaroo is the only living member of the Hypsiprymnodontidae; all the 290 Life of Marsupials

Balbaridae 13 7

Propleopinae 3

Hypsiprymnodontinae 2 2 Macropodoidea Hypsiprymnodontidae

Potoroinae 8 9

Bulungamayinae 5442

Macropodinae 11 28 61

Macropodidae Sthenurinae 5 12

Oligocene Miocene Pliocene Pleistocene Holocene 26 23.316.4 10.4 5.220 0.01 million years ago Figure 9.1: Summary of the fossil history of kangaroos and their ancestors during the last 26 million years, and relationships between them based on fossil material. Numbers in blocks indicate number of species if more than one. After Cooke and Kear (1999).

others are placed in a second large family, the Macropodidae. The latter family comprises three subfamilies with living species; Potoroinae, Macropodinae and Sthenurinae. There are three extinct groups, the Balbaridae, Propleopinae and Bulungamayinae (Fig. 9.1).

Hypsiprymnodontidae The tiny musky rat kangaroo, Hypsiprymnodon moschatus, of the Queensland rainforest floor is the only living member of the family Hypsiprymnodontidae and is considered to be closest to the ancestral stock from which all other kangaroos evolved. Although it superficially resembles the other rat kangaroos in having larger hind legs than forelegs, it is only distantly related on genetic criteria, such as immunoglobulins (Baverstock et al 1990) and mitochondrial and nuclear DNA sequences (Burk et al 1998, Burk and Springer 2000). Furthermore, it does not hop but bounds using all four legs and it retains digit 1 on the hind foot, as possums do, but no macropod does. It is omnivorous in its diet, and its teeth and stomach resemble those of possums more than other macropods; and it does not exhibit embryonic diapause. Its ancestors are known back to the early Miocene, 23 million years ago, although it may have diverged much earlier. On the basis of similarity of teeth the musky rat kangaroo is related to an extinct subfamily, the Propleopinae. One of the last species of this group, which died out with the other megafauna about 30 000 years ago, was oscillans, known from parts of a skull, one humerus and several separate teeth. Because the teeth look similar to those of the tiny musky rat kangaroo but very much larger, it is depicted as a very large rat kangaroo with big hind legs, but this is a leap of faith because the skeleton is not known. The details of the skull and teeth suggest that it was a predator, and the one humerus that it was an animal that ran and probably occupied a similar niche to the present day fox, Vulpes vulpes (Ride et al 1997). was a smaller species in Consummate kangaroos 291

which a complete skull is known from Queensland and it also is thought to have been a carnivore (Wroe et al 1998). It is because Hypsiprymnodon has a simple stomach, does not hop and does not show the kangaroo pattern of reproduction, that it together with Propleopus and Ekaltadeta, are now placed in a separate family, the Hypsyprymnodontidae.

Rat kangaroos Of the three subfamilies with living species, the Potoroinae includes nine species of rat kanga- roos ranging in size from 1 to 3 kg and restricted to Australia and Tasmania: none occurs in New Guinea (Table 9.1). Rat kangaroos hop, but, in contrast to kangaroos and wallabies, they can use their tails to pick up objects such as grass. The front pair of upper incisors are large and overhang the lower incisors, while incisors 2 and 3 and canines in the upper jaws are small; the premolars are very large with prominent flutings on the sides and the four molars behind them decrease in size from front to back, as shown for the burrowing bettong, Bettongia leseuer (Fig. 9.3a). The cheek teeth erupt early and remain in the same position in the jaw throughout life. The shearing premolars and the molars with rounded crowns reflect the species’ varied diet of ground-living herbaceous plants, subterranean fungi and occasional animal flesh. Rat kangaroos have a sacculated forestomach (Fig. 9.7) in which bacterial fermentation occurs; and all species so far examined display post-partum oestrus and embryonic diapause.

Table 9.1: Living and recently extinct species of the Hypsiprymnodontidae, Potoroinae and Lagostrophus CEN, central Australia; NG, New Guinea; NE, north-eastern Australia; NW, north-western Australia; SE, south-eastern Australia; SW, south and south-western Australia; TAS, Tasmania; Strahan (1995); *, present; @, formerly present, now absent.

Hypsiprymnodontidae No. of NG NE NW SE TAS SW CEN Body mass chromosomes (kg) Hypsiprymnodon 22 * 0.5 moschatus Potoroinae Aepyprymnus rufescens 32 * * 3.5 Bettongia gaimardi 22 @* 1.7 Bettongia lesueur 22 @ @ * 1.0 Bettongia penicillata 22 @ * @ 1.3 Bettongia tropica 22 * @ 1.2 Caloprymnus campestris – @ 0.9 Potorous longipes 24 * 2.0 Potorous platyops –@@ Potorous tridactylus 13/12 *** 1.2 Sthenurinae Lagostrophus fasciatus 24 @ @ * 1.3–2.5

Before European settlement the nine species were widely distributed across continental Australia, and two species are still common in Tasmania. They occupied grasslands predomi- nantly, where they made nests in tussocks or, in the case of the burrowing bettong (Fig. 9.2a, Plate 14), excavated extensive warrens, the mounds of which can still be seen across inland Australia (Noble 1997). 292 Life of Marsupials

Figure 9.3: Lateral view of the skulls of (a) the burrowing bettong, Bettongia lesueur, (b) the swamp wallaby, Wallabia bicolor, and (c) the red kangaroo, Macropus rufus, at the same basi-cranial length. Note presence of canines and the large fluted premolars in the bettong and the large diastema in the wallaby and kangaroo. In rat kangaroos the molars diminish in size posteriorly and have a fixed position in the jaw, whereas in the wallabies and kangaroos the molars are of equal size, high crowned, and move forward in the jaw, being shed progressively from the front. In the kangaroos the premolars are also shed.

Since European settlement, two species have gone extinct and three other species only remain on islands. The other species have much reduced ranges and all are endangered, except the two species in Tasmania, the Tasmanian bettong, Bettongia gaimardi, and the long-nosed potoroo, Potorous tridactylus. In Western Australia the brush-tailed bettong, Bettongia penicillata, has recov- ered in numbers since the control of foxes by aerial spread of the poison sodium fluoracetate Consummate kangaroos 293

(1080) in several reserves, but this is still a tiny fraction of its former range across the whole of southern Australia. In New South Wales the numbers of four species were severely reduced during 1890 to 1910 when over three million were killed for bounty (Short 1998) and three of these species disappeared soon after the arrival of foxes in about 1910. Only the largest species, the rufous bettong, Aepyprymnus rufescens, is secure in part of its former range, in northern New South Wales and southern Queensland. The burrowing bettong, which once occurred across most of the continent, still thrives on Barrow, Bernier and Dorre Islands off the coast of Western Australia. These small macropods have followed the same pattern of demise as the dasyurids (see Chapter 4) and the bandicoots (see Chapter 5).

The anomalous banded hare wallaby The banded hare wallaby, Lagostrophus fasciatus, has some features in common with potoro- ines and some with macropodines. These anomalies led Flannery (1983) to propose that it may be a living member of the otherwise extinct subfamily Sthenurinae. Among his reasons were the relative positions of the upper and lower incisors. As in sthenurines the two lower incisors bear directly on the arc of upper incisors, unlike in macropodines, where the upper incisors overhang the lower incisors; also the molar row is straight and flat with all teeth occluding at the same time, and the small premolar (P3) moves forward during life so that the diastema becomes progressively shorter (Flannery 1983). The banded hare wallaby also has an unusual chromo- some number (24) and the anatomy of the female reproductive tract is different from all the Macropodinae and is more like that of rat kangaroos. And now the sequence of some of its DNA has been found to be distinct from both potoroines and macropodines (Westerman et al 2002) (Fig. 9.6). Although its relationship to other living macropods seems remote, some people challenge the interpretation that it is related to the sthenurines. This question may soon be resolved, however, if ancient DNA can be recovered from the beautifully preserved sthenurine skeletons recently discovered (2002) in caves on the Nullarbor Plain. The banded hare wallaby formerly occurred across southern Australia but, like the burrowing bettong, is now extinct on the mainland and survives only on Bernier and Dorre Islands.

Kangaroos and wallabies By far the most abundant subfamily is the Macropodinae, which includes 61 species (Tables 9.2– 9.4). They include six species of New Guinean forest wallabies, Dorcopsis and ; 10 species of tree kangaroo, Dendrolagus (Fig. 9.2b, Plate 14), eight of which live in New Guinea and two in the rainforests of northern Queensland; and six species of forest-dwelling pademelons, Thylogale, four of which live in New Guinea and three in Australia (Flannery 1995). The only other macropods that live in New Guinea are the recently-discovered spectacled hare wallaby, Lagorchestes conspicillatus, and the agile wallaby, Macropus agilis, which both occur in northern Australia (Fig. 9.12). The exclusively Australian macropodines are the small quokka, Setonix brachyurus, of Western Australia; 16 species of rock wallaby, Petrogale (Fig. 9.2c, Plate 14); three species of hare wallaby Lagorchestes; three species of nailtail wallabies, Onychogalea, the swamp wallaby, Wallabia bicolor; and eight other large wallabies and six large kangaroos, Macropus (Fig. 9.2d,e,f, Plate 15). Of all these species only three occur on Tasmania – one pademelon, Thylogale billar- dierii, Bennett’s or red-necked wallaby, Macropus rufogriseus, and the eastern grey kangaroo, Macropus giganteus. Four Australian species are now extinct and four others have very restricted distributions. As with the rat kangaroos and the banded hare wallaby, it was the smaller species that were most vulnerable to European occupation, whereas the larger species remain secure and some have increased their ranges and abundance. 294 Life of Marsupials

Table 9.2: New Guinea wallabies, hare wallabies and nailtail wallabies CEN, central Australia; NG, New Guinea; NE, north-eastern Australia; NW, north-western Australia; SE, south-eastern Australia; SW, south and south-western Australia; TAS, Tasmania; mass of males, females; *, present; @, formerly present, now absent. Flannery (1995), Strahan (1995).

No. of NG NE NW SE TAS SW CEN Body mass (kg) chromosomes (, & Dorcopsis atrata * 7.5, – Dorcopsis hageni * – , 5.5 Dorcopsis luctuosa 22 * 9.3,3.6 Dorcopsis muelleri * –, 5.0 Dorcopsulus * –, 3.0 macleayi Dorcopsulus 18 * –,1.8 vanheurni Lagorchestes @ asomatus Lagorchestes 15/16 *** * 1.6–4.5 conspicillatus Lagorchestes 20 @ * 1.2, 1.3 hirsutus Lagorchestes @ leporides Onychogalea 18 * @ 6.5, 4.5 fraenata Onychogalea @ 3.5 lunata Onychogalea 20 ** 7.5, 5.8 unguifera Setonix brachyurus 22 * 3.6, 2.9 Despite the large number of species, and the variety of habitats in which they live, all 61 species of the Macropodinae are so closely related to each other that there is an ongoing debate as to whether they should be considered a separate family or should be treated as a subfamily equivalent to the Potoroinae. Until recently the relationships between members of the subfamily, based mainly on dentition, were difficult to discriminate because of a plethora of shared charac- ters, so that taxonomists referred to the macropodine ‘family bush’ rather than ‘family tree’ (Flannery 1989). However, DNA/DNA hybridisation (Kirsch et al 1997) and mitochondrial and nuclear DNA sequences (Burk and Springer 2000) are helping to sort this out, and the techniques of G-banding and chromosome ‘painting’ have largely resolved the anomalies in chromosome number between the species (see Fig. 1.8, Plate 2) (Rofe 1978, Glas et al 1999). The difficulty in sorting out the relationships reflects the rapid and relatively recent evolution of the subfamily – which is most clearly seen in the rock wallabies, where rapid speciation is still in progress – and to adaptations for feeding on grass, which seem to have evolved independently more than once.

Classifying wallabies and kangaroos by their teeth The macropodine dentition differs from the potoroine in several respects. The upper incisors are wider, especially the third upper incisor, I3, and form an arc within which the large pair of lower Consummate kangaroos 295

Table 9.3: Pademelons, rock wallabies and tree kangaroos CEN, central Australia; NG, New Guinea; NE, north-eastern Australia; NW, north-western Australia; SE, south-eastern Australia; SW, south and south-western Australia; TAS, Tasmania; mass of males, females; *, present; @, formerly present, now absent. Flannery (1995), Strahan (1995).

No. of NG NE NW SE TAS SW CEN Mass (kg) chromosomes (, & Thylogale billardierii 22 @* 7.0, 3.9 Thylogale browni * 6.6, 4.3 Thylogale brunii 22 * Thylogale calabyi * Thylogale stigmatica 22 ** * 5.1, 4.1 Thylogale thetis 22 * 7.0, 3.8 Petrogale assimilis 20 * 4.7, 4.3 Petrogale brachyotis 18 * 4.4, 3.7 Petrogale burbidgei 16 * 1.2 Petrogale coenensis 22 * 5.0, 4.0 Petrogale concinna 16 * 1.3 Petrogale godmani 20 * 5.2, 4.3 Petrogale herberti 22 * 6.0, 4.3 Petrogale inornata 22 * 5.0, 4.2 Petrogale lateralis 20,22 * * 4.5, 3.5 Petrogale mareeba 18 * 4.5, 3.8 Petrogale penicillata 22 * @ 7.9, 6.3 Petrogale persephone 22 * 7.2, 5.2 Petrogale purpuriecollis 22 * Petrogale rothschildi 22 * 6.1, 4.1 Petrogale sharmani 20 * 4.4, 4.1 Petrogale xanthopus 22 * 6–12 Dendrolagus bennettianus 14 * 12.5, 9.3 Dendrolagus dorianus 12 * 10.4, 8.8 Dendrolagus goodfellowi 14 * 7.9, 8.1 Dendrolagus inustus * 15.5, 11.4 Dendrolagus lumholtzi 14 * 7.5, 6.0 Dendrolagus matschiei 14 * Dendrolagus scottae * 11.5, 9.3 Dendrolagus spadix * 9.1, – Dendrolagus ursinus 12 * Dendrolagus mbaiso * –, 8.8 296 Life of Marsupials

incisors fit. In order to make contact with the upper incisors they must either be pulled to one or the other side or be splayed out (Fig. 9.4). Because the two halves of the lower jaw, the dentaries, do not become fused together, their point of union, the symphysis, is flexible and the long pair of lower incisors can spread apart and shear against the three pairs of incisors in the upper jaw. By contrast sthenurines had more rigid or even fused dentaries and could not make this movement, their lower incisors instead meeting the upper ones directly. Most species of macropodines lack canines and all have a longer gap between the front and back teeth and smaller premolars than the potoroines. The first two cheek teeth in each jaw of a young kangaroo are a large cutting or sectorial premolar, called P2, and a large molar-like tooth,

Figure 9.4: Jaw muscles of the pademelon, Thylogale stigmatica. Superficial layers and portions of bone removed to show the deeper muscles that move the jaws during chewing. Also shown (in the red-necked wallaby, Macropus rufogriseus) is the flexible joint in the lower jaw which allows the lower incisors to be spread to meet the upper incisors. After Abbie (1939) and Ride (1959). Consummate kangaroos 297

called deciduous premolar 3 (dP3). These two teeth provide the very young macropod with a shearing tooth and a grinding tooth in each jaw. Later, when M1 and M2 erupt behind them, the first two teeth are shed and are replaced by a single cutting tooth called P3 at the front of each row. As the animal ages the remaining molars M3 and M4 erupt sequentially behind M2. Premolar 3 is retained throughout life in most wallabies, which are predominantly browsers, but is shed during middle life in the larger grazing kangaroos such as the red kangaroo, Macropus rufus (Fig. 9.3c). The main difference between potoroine teeth and macropodine teeth, however, is the size and structure of the molars. In all macropodines the four cusps are joined in pairs to form high, transverse ridges or lophs, unlike the rounded teeth of potoroines. The extinct species also had lophodont molars, which are thought to have evolved for feeding on low growing, siliceous grasses. In the macropodines additional ridges, called links, also develop at right angles to the lophs, which further aid in the breakdown of the plant tissue. The evolution of lophodont molars heralded a different way of chewing. Among the possums and more particularly the wombats, Lasiorhinus, chewing involves a side-to-side motion of the molars, with the shear facets passing each other transversely to cut the plant tissue (see Fig. 8.3). In the potoroines lateral chewing is also possible but among the macropodines lateral movement of the lower jaw is limited by the position of the lower incisors inside the upper incisor arc: instead there is an additional forward- and-back motion of the molar teeth during chewing. This is clearly seen on the crests of the lophs, where all the scratches in the tooth enamel are antero-posterior, not transverse. However, as the lophs of the lower molars engage the lophs of the upper molars, they slide into the trough between, where they strike the links and then move sideways (Sanson 1980); the scratches on the enamel of the links being transverse. This movement of the molars is accomplished by an alternate rotation of the lower jaw first on one condyle, as the opposite condyle moves forward, and then on the other at the next bite. This rotatory movement shifts the lower molar row on one side a few millimetres forward so that the crests of the lophs of the upper and lower molars shear against each other like a well-set pair of scissors. The major jaw muscles control this complex chewing cycle (Fig. 9.4). These are the pterygoid muscles on the inside of each dentary, which move the jaw to one side or the other and forward; and the outer masseter and temporalis muscle complex, which pulls it up and backwards. The deepest layer of the large masseter muscle penetrates forward into the masseteric canal, a characteristic feature of the macropodine dentary not seen in other herbivorous marsupials. Ride (1959) studied the movements of the macropod jaw during chewing using X-ray cinema- tography of an adult Bennett’s wallaby. The forage is grasped, not cut, and as it comes away the lower jaw is drawn back and the lower incisors close together, by contraction of the masseter and temporal muscles. The grass stems are arranged in the diastema by the tongue and lips and the ends are then fed into the molar row, where they are ground between the lophs and links, first on one side and then on the other. Sanson (1989) recognises three grades of dentition in macropodine species on the basis of the degree of their adaptation for processing grass: these are the size of P3, the height of the molar lophs and links, and whether the molar rows are straight or arched.

The browser grade The least specialised browsers are the New Guinea forest wallabies, Dorcopsulus and Dorcopsis, and the six species of pademelons, Thylogale, which browse on shrubs and soft plants of the forest floor, and the tree kangaroos, which feed on tree foliage. Only the New Guinea forest wallabies and the tree kangaroos retain a small canine in the upper jaw but all of them have moderately large premolars, which are used for cutting stems of shrubby plants, and are retained through life. The four molars in each jaw erupt early in life and form a flat occlusal surface with 298 Life of Marsupials

Figure 9.5: Variation in the dentition of rat kangaroos, wallabies and kangaroos, to show how different kinds of food are processed, and the degree of occlusion of the molar teeth. The arrows indicate the relative importance during the bite of the respective teeth. After Sanson (1989).

the opposing set of molars in the lower jaw, but the links between the lophs are small (Fig. 9.5b). These teeth can exert some shearing force as well as crushing on tougher plant tissue, such as grass. The quokka and the swamp wallaby, which are also browsers, have this type of dentition: so did the earliest known macropodine fossil, Dorcopsoides and the fossil wallabies of the genus . The rock wallabies, Petrogale, probably evolved from pademelon stock but became adapted to living in rocky breakaways and cliffs and so were limited to the plants that grow near these Consummate kangaroos 299

sites. Their diet includes more grasses than the diet of pademelons and their molars are higher crowned, reflecting this greater dependence on grass. However, they still retain P3 at the front of the molar row and this leads to excessive wear on the front molars and much less wear on the back molars. One small species, the nabarlek, Petrogale concinna, sheds P3 so the molars move forward in the jaw and, because the rows are arched, the molars do not all occlude at the same time. The enamel on these teeth is thin, the teeth wear down quickly and are replaced by others behind them. This unusual pattern of tooth replacement in this one species of rock wallaby is probably an independently acquired adaptation for feeding on grass, even though it resembles the same condition in the large kangaroos.

The intermediate browser–grazer grade The intermediate group between browsers and grazers, which includes the hare wallabies and the eight species of wallaby in the Notomacropus subgenus (Table 9.4), are predominantly grazers but will also browse. Their premolars are small and are occasionally shed, thus allowing some forward movement of the molar rows. The molars have higher lophs and well-developed links between them. The opposing molar rows are arched so that occlusion and wear occurs succes- sively through life, from M1 to M4 (Fig. 9.5c).

Table 9.4: Large wallabies and kangaroos CEN, central Australia; NG, New Guinea; NE, north-eastern Australia; NW, north-western Australia; SE, south-eastern Australia; SW, south and south-western Australia; TAS, Tasmania; mass of males, females; *, present; @, formerly present, now absent. Flannery (1995), Strahan (1995).

No. of NG NE NW SE TAS SW CEN Body mass chromosomes (kg) (, & Wallabia bicolor 11/10 * * 17, 13 a. Notomacropus group Macropus agilis 16 **** 19, 11 Macropus dorsalis 16 * * 16, 6.5 Macropus eugenii 16 * * 7.5, 5.5 Macropus greyi @ Macropus irma 16 * 8, – Macropus parma 16 * 5, 4 Macropus parryi 16 * * 16, 11 Macropus rufogriseus 16 ** 20, 14 b. Osphranter group Macropus antilopinus 16 ** 37, 17.5 Macropus bernardus 18 * 21, 13 Macropus robustus 16 *** * 46, 25 Macropus rufus 20 * 66, 27 c. Macropus group Macropus fuliginosus 16 * * 54, 28 Macropus giganteus 16 * ** * 66, 32 300 Life of Marsupials

The grazer grade The species that feed predominantly on grasses are the true grazers. They are the nailtail wallabies and the six species of large kangaroo (Table 9.4). The large kangaroos fall into two subgroups (Dawson and Flannery 1985), four in the main Osphranter subgenus are the red kangaroo, the euro, Macropus robustus erubescens, antilopine wallaroo, Macropus antilopinus and black wallaroo Macropus bernardus, and the extinct Macropus ferragus; and the grey kangaroo cluster, eastern grey, western grey, Macropus fuliginosus, and the extinct Macropus titan. In all these species P3 is very small and is usually shed before all the molars have erupted. The molars have very pronounced lophs and well-developed links between them. Each molar row is strongly curved, the upper and lower rows curving in opposite directions so that only two pairs of molars occlude on each side at any one time. They erupt sequentially from the front and each whole row moves forward in the jaw through life, with the most worn down molars being shed from the front (Fig. 9.5d). In very old animals the molar row may consist of only the last two worn down molars (M3 and M4). Occasionally old eastern grey kangaroos may have a fifth set of molars (Kirkpatrick 1965). The rate at which the molars move forward is related to the quality of the grasses eaten; if coarse, siliceous grasses are being eaten, the wear on the molar crowns is greater and the forward progression faster than when softer grasses are eaten (Sanson 1989). Before this was appreciated it was thought that the rate of forward progression was constant for a species and could be used as an accurate method for estimating the age of the animal. While it is still useful for relative ageing within a population, comparisons between populations must be made with caution. The classification of macropodines, based on dentition, can now be compared with the other criteria: chromosomes, and protein and DNA sequence comparisons, and DNA/DNA hybridisation.

Classifying wallabies and kangaroos by their chromosomes As mentioned in Chapter 1, chromosome numbers among species in the superfamily Macropodoidea range from 10 to 32, that is to say, from 5 to 16 pairs of chromosomes of which one pair are the sex chromosomes; this diversity is thought to reflect the fairly recent and rapid evolution of the group. Despite the wide range of chromosome numbers, however, the techniques of G-banding and chromosome painting now make it possible to understand how the changes may have taken place between related species. The musky rat kangaroo and four species of Bettongia have 22 chromosomes (10 pairs of autosomes plus two sex chromosomes), as do the pademelons, most of the rock wallabies, the quokka and some of the New Guinea forest wallabies (Tables 9.1–9.4, see Table 1.4). On this evidence it seems likely that 22 is the ancestral number for the whole superfamily and that other numbers have been derived from this by fission or by fusion between pairs of chromosomes and rearrangement of bits within chromosomes (Rofe 1978). This has been clearly established for the subfamily Macropodinae by chromosome painting (see Fig. 1.8, Plate 1). The chromosomes of three species with 22 chromosomes were shown to be the same. These were the pademelon, Thylogale thetis, a New Guinea forest wallaby, Dorcopsis sp. and the black-footed rock wallaby, Petrogale lateralis. In Fig. 1.8 one of each of the 10 pairs of autosomes are individually coloured along with both sex chromosomes. By following the colour code the 20 chromosomes of the red kangaroo can be derived from this ancestral pattern by fusion of the chromosome pairs 1 + 10 to give nine pairs of autosomes plus two sex chromosomes, and the 16 chromosomes of the tammar are derived by two further fusions of the chromosome pairs 5 + 8 and 6 + 9 (Glas et al 1999). The tammar pattern is common to all the seven living wallaby species of the Notomacropus group with 16 chromosomes. However, in the euro and antilopine wallaroo (Osphranter group), which also have 16 chromosomes, the second set of fusions is between 5 + 6 and 8 + 9, while in the two species of grey kangaroo the three fusions are between 1 + 8, 6 + 10 and 5 + 9. Consummate kangaroos 301

Thus, while most species of the genus Macropus have 16 chromosomes, they have arrived at this number by three different patterns of fusion of the ancestral 22 chromosomes. This supports the subdivision of the genus on dental criteria. Likewise, the tree kangaroos with chromosome numbers of 14 can also be derived from the ancestral 22 chromosomes by a different series of four fusions (1 + 7, 4 + 6, 5 + 10, 8 + 9), as can the smallest chromosome number of all in the swamp wallaby, Wallabia bicolor, by six fusions (see Figs 1.7b, 1.8, Plate 1).

Classifying wallabies and kangaroos on DNA criteria Comparisons of DNA/DNA hybridisation and DNA sequences between different species of kangaroo support the relationships derived from chromosomes (Kirsch et al 1995, 1997, Burk and Springer 2000). The musky rat kangaroo is clearly different from all other species, including the other rat kangaroos (Burk et al 1998, 1999). Within the Potoroinae the four species of Bettongia are closely related to Aepyprymnus, despite the latter species having 32 chromosomes rather than 22; the

Hypsiprymnodon Potorous Caloprymnus Table 9.1 Aepyprymnus Bettongia Lagostrophus Dorcopsis Dorcopsulus Setonix Table 9.2 Onychogalea Lagorchestes Dendrolagus Thylogale Table 9.3 Petrogale Wallabia Macropus Table 9.4 Notomacropus Osphranter

Figure 9.6: Relationships among the 18 genera of the Macropodoidea, as disclosed by mitochondrial and nuclear DNA sequences, using the brushtail possum, Trichosurus vulpecula, and koala, Phascolarctos cinereus, as the out group (not shown). This analysis includes the extinct desert rat kangaroo, Caloprymnus campestris, now shown to be related to the , and the banded hare wallaby, Lagostrophus fasciatus; it supports the separation of the latter species from both the Macropodinae and the Potoroinae. For further details see Tables 9.1–9.4. After Westerman et al (2002). 302 Life of Marsupials

three species of Potorous, including the extinct species, are closely related to each other and are distinctly different from the and from the extinct desert rat kangaroo, Caloprymnus campestris (Westerman et al 2004) (Fig. 9.6). As already mentioned, the banded hare wallaby is distinct from all other species of rat kangaroo, kangaroo and wallaby, which is consistent with the idea that it is the single remaining representative of the otherwise extinct Sthenurinae. Within the Macropodinae there are three fairly distinct groups, which agree with the groupings already recog- nised from teeth and chromosomes. The New Guinea forest wallabies, Dorcopsis and Dorcopsulus, the quokka, the hare wallabies and nailtail wallabies are distantly related to each other; pademel- ons are closely related to rock wallabies and tree kangaroos; the swamp wallaby is on its own; the seven living wallabies of the Notomacropus group are closely related and distinct from the six species of kangaroo. Among the latter, the two species of grey kangaroo are distinct from the red kangaroo, euro, wallaroo, and antilopine wallaroo which form the Osphranter group. Thus, the relationships of kangaroos, based on such different criteria as teeth, chromosomes and DNA, are in general agreement, which suggests that the current classification reflects the reality of their evolution. Where there is disagreement, as in the relative position of the nailtail wallabies or nabarlek, it is likely that adaptations for grazing evolved independently in more than one line of descent. As these estimates of similarity become more accurate and agree with each other more closely, it also becomes possible to estimate the time of separation between related species more precisely. These suggest that Hypsiprymnodon must have been separated from all species of the Macropodidae for more than 30 million years, although its lineage is only known since the Miocene, 23 million years ago. This supports the view that the separation occurred before the acquisition of a hopping gait, foregut fermentation and embryonic diapause in the main line leading to the Macropodidae.

Kangaroo origins So far 20 macropod genera, representing four distinct families, have been described from the earliest Oligocene fossil beds (Fig. 9.1). This indicates that the origin of the group occurred earlier than this, but how much earlier must await the discovery of earlier fossil sites. The two most numerous groups of the early period, the Balbaridae and Bulungamayinae, died out in the late Miocene between 5 and 10 million years ago. The relationships of these two groups to modern kangaroos and rat kangaroos is still unclear: some people argue that the balbarines gave rise to the rat kangaroos and the bulungamayines to the main branch leading to kangaroos and sthenurines (Cooke and Kear 1999), while others hold the contrary view (Flannery 1989). The crucial question in these arguments is, when did the hopping gait first appear in the ancestral kangaroos? This is difficult to answer because few intact skeletons have been discovered, most fossils being bits of skull or separated teeth. Transition from quadrupedal gaits to hopping probably occurred during late Oligocene to early Miocene. An Oligocene balbarine at Riversleigh, Balburoo, had forelimbs and hind limbs of similar length and it also had digit 1 on the hind foot as in the quadrupedal musky rat kangaroo today. Because the attachment sites on the femur for the main driving muscles used in hopping (quadratus and adductor magnus, Fig. 9.11) were small, it is probable that this species did not hop. By contrast, the middle to late Miocene bulungamyines, Gunguroo and Wanburoo, were probably bipedal because they had lost digit 1 of the hind foot and digits 2 and 3 were reduced and partly under the large digit 4, as in the typical kangaroo foot (see Fig. 1.5). Also the bones of the ankle had an articulation that would have allowed less sideways movement, another adaptation for bipedal hopping. Since bipedalism is common to all except the musky rat kangaroo and the balbarines, either all had a common bipedal ancestry from the bulungamayines, or bipedalism evolved independently more than once. Kirsch et al (1997) think that bipedalism may have arisen Consummate kangaroos 303

independently in early potoroines and later bulungamayines, whereas Cooke and Kear (1999) think that the balbarines were related to Propleopus and Hypsiprymnodon, distinct from other macropods. If they are correct, the bulungamayines were the stem group for the whole of the Macropodidae (hopping bipedalism) and declined or were replaced in the late Miocene by the radiation of the Potoroinae, Sthenurinae and the Macropodinae, coinciding with the spread of grasslands in response to increasing aridity. The earliest known sthenurines were wallaby-sized creatures but later species in the Pliocene and Pleistocene grew large (up to 3m tall) and were distinguished from the true kangaroos by a further reduction of the hind foot to a single large 4th toe, rather as the limb of the horse has been reduced to a single digit. The tail was short and stocky and the forelimbs were probably not used for locomotion at all. Their arms were more flexible than those of kangaroos and could be raised above the head to grasp branches in a manner similar to the extinct ground sloth of South America. This resemblance was further accentuated by the very short muzzle and massive skull and fused halves of the lower jaw. Having flourished for more than 5 million years, all the sthenurines went extinct between 10 000 and 30 000 years ago, unless the banded hare wallaby is a surviving member of the group. As mentioned earlier, the major radiation of both Potoroines and Macropodines occurred in the last 5 million years. In the latter subfamily there were about 11 species 4.5 million years ago, and then there was an explosive radiation with all modern genera being represented by 2 million years ago. The Macropodidae have been extremely dynamic throughout their relatively brief history, rapidly adapting to new ecological opportunities and producing large radiations with wide diversity of forms (Flannery 1989). As already suggested, their main attributes for this success must have been their digestive capabilities, their unique mode of locomotion and their manner of reproduction.

Three steps to success

Foregut fermentation gives greater access to plant resources Fermenting herbage in the front part of the gut has several important advantages for a mammal. The main advantage is that the bacteria that live in the foregut can break down the cellulose of the plant cell walls and release the contents more effectively than can be accomplished by the sharp- est teeth. Second, the short-chain fatty acids (SCFA) produced by the anaerobic fermentation of the cellulose are a rich source of energy for the mammal, which would otherwise be unavailable because mammals cannot synthesise the enzyme to hydrolyse cellulose. Third, the bacteria in the forestomach can use urea as a source of nitrogen (N) to synthesise their own protein, which subsequently becomes available to the mammal when it digests the bacteria in the small intes- tine. This has two advantages for the mammal; it conserves nitrogen that would otherwise be excreted as urea in urine, and it conserves water that would otherwise be needed to flush out the urea. Thus, the evolution of this symbiotic relationship with bacteria in foregut fermentation has given ruminants in other parts of the world and macropods in Australasia great economies in energy, nitrogen and water conservation. How do the macropods achieve this?

The macropod stomach The macropod stomach is very large and when full can weigh 15% of the total body mass of the animal: this is also a feature of other mammals that digest herbage by bacterial fermentation in the foregut, such as ruminants and camels. However, the stomach of macropods is very different 304 Life of Marsupials

in its anatomy from that of sheep and cattle. It is an elongated curved bag, deeply indented on the outer curvature and smooth on the inner shorter side. Four parts can be distinguished exter- nally, each with distinct functions: the two anterior parts are unique to the macropods, while the two posterior parts correspond to the stomach anatomy of other marsupials. The oesophagus opens into a funnel-shaped region, which extends along the inner curvature as a partly closed spiral groove and is lined with the same type of squamous epithelium as lines the oesophagus (Fig. 9.7). Its function is to separate fine from coarse plant material and direct the fine material to the posterior region of the stomach, which is lined with an epithelium that secretes hydro- chloric acid, producing a low (acid) pH of 1.8–3.0, and proteolytic enzymes that break down protein from plants and bacteria.

Figure 9.7: The stomach anatomy of the eastern grey kangaroo, Macropus giganteus, showing the large fermentation chamber and structure of the wall of the forestomach, compared to the different anatomy of the omnivorous long-nosed potoroo, Potorous tridactylus, and the browsing red-necked pademelon, Thylogale thetis. After Griffiths and Barton (1966) and Freudenberger et al (1989). Consummate kangaroos 305

The coarse plant material is directed to the forestomach, where bacterial fermentation takes place. It is the largest part of the macropod stomach and is divided by a deep permanent fold into the anterior sacciform forestomach and the posterior tubiform forestomach. Both are lined with a glandular epithelium that secretes mucus but neither acid nor proteolytic enzymes. The pH ranges from 4.6 before feeding to 8.0 after, a range that is maintained by the buffering action of the mucus and by saliva from the parotid salivary glands released during chewing. The copious parotid saliva contains high concentrations of bicarbonate, phosphate and sodium ions, which act to buffer the fluid at an alkaline pH and support bacterial growth in the forestomach. Both parts of the forestomach contain dense populations of bacteria (1010 per mL) and ciliated protozoa (106 per mL). The bacteria digest the cellulose of the plant cell walls anaerobically to SCFA, thereby releas- ing the contents of the plant cells for further digestion by them and by the host macropod. The component parts of the forestomach vary between species, reflecting their different habits and food preferences (Fig. 9.7). As already mentioned, the musky rat kangaroo has a simple stomach with a low pH, no development of the forestomach regions and no bacterial fermentation. In the rat kangaroos, tree kangaroos and pademelons the sacciform forestomach is much larger than the tubiform forestomach, which may reflect their predominant diet of digestible browse. In the pademelon, as food enters the stomach it is preferentially directed to the sacciform forestomach and only reaches the smaller tubiform forestomach several hours later. By contrast, in the grazing kangaroos the sacciform forestomach is much smaller than the multilocular tubiform forestomach. The coarse herbage, mainly grasses, that constitutes their preferred food is directed to the tubiform forestomach, where it is retained for many hours, being thoroughly mixed in each compartment before being moved to the next by contractions of the partitions between each pocket and by muscles of the stomach wall. When a bolus of food eventually passes from the tubiform forestomach to the hindstomach it encounters a very acidic (low pH) environment and proteolytic enzymes, which digest the protein of the micro-organ- isms and the plant cell contents. These then pass to the small intestine where they are assimilated into the animal’s blood stream. Unlike the possums, wombats and the koala, Phascolarctos cinereus, which ferment plant material in the hindgut, the caecum and proximal colon of macropods are relatively small and, although they contain some fermentative bacteria, they contribute little to the animal’s economy. However, the distal colon is important as the site of absorption of water from the faeces, as it is in the other herbivores, and its length varies in relation to the habitat of each species: in forest- dwelling species, such as pademelons, the distal colon is short and the faecal pellets have a high water content, while in desert-adapted species, such as the euro, it is long and the faecal pellets have a very low water content (54%), similar to dung of the southern hairy-nosed wombat, Lasiorhinus latifrons, and the camel, Camelus dromedarius. During pouch life, before the young macropod begins to eat herbage, all regions of the stomach have a variable but low pH and are proteolytic (Griffiths and Barton 1966). When the young macropod leaves the pouch permanently the forestomach becomes a functional fermen- tation chamber with a neutral pH. It is also colonised by bacteria and protozoa, presumably from its mother. At this time the young often makes contact with its mother’s muzzle (Russell 1973, Croft 1981b) and in some instances the mother has been seen to pass chewed matter to the young. This behaviour resembles the passing of caecal contents to young koalas and wombats at the time that they are beginning to eat herbage, and it may be how bacteria and protozoa are transferred to the young animal.

Carbohydrate metabolism – getting energy from cellulose The bacteria in the forestomach are living in an environment of low oxygen, so they cannot hydrolyse the cellulose of the plant cell walls to simple sugars but only to the intermediate 306 Life of Marsupials

products of anaerobic fermentation. These are the short-chain fatty acids, already encoun- tered in the hindgut fermenting koala, wombat and greater glider, Petauroides volans, namely acetic, propionic and n-butyric acid, plus ammonia, methane, carbon dioxide and hydrogen. Production and absorption of short-chain fatty acids has been measured in the tammar wallaby (Fig. 9.8). Production begins in the sacciform forestomach and continues at a diminishing rate as the bolus moves along the tubiform forestomach. At the same time they are being absorbed across the wall of the forestomach, so that less than one-fifth of the production is still present in the hind stomach. The butyric acid is used by the stomach epithelium as the source of energy for the transport of the other fatty acids into the circulation, where they are metabolised in the liver to glucose; this is detected in the blood soon after the macropod has fed.

Figure 9.8: Daily net production of short-chain fatty acids (in mmoles) in the stomach of the tammar wallaby, Macropus eugenii, to show that they are synthesised in the forestomach, released progressively into the circulation along the length of the forestomach, and the balance passed to the small intestine. After Dellow et al (1983).

Rate of passage through the tubular stomach of the macropod is a crucial factor in the fermen- tation of cellulose by bacteria: the slower it travels the more breakdown can occur and hence less food, or less nutritious food, is required to maintain a positive energy balance. Although the rate of passage is faster than through the ruminant’s chambered stomach, the use of low-grade grasses by kangaroos is, paradoxically, superior because less time is spent fermenting the coarser material, which continues on to the caecum and proximal colon. One consequence of this in the larger macropods is that their intake is not affected by the quality of the forage to the same extent as it is in the sheep. In one experiment that compared euros with sheep, the euros ate 30% less than the sheep on a low fibre diet, because of their lower metabolic rate, but they were able to consume a higher fibre diet than the sheep could (Hume 1999). In this the euros are similar to the hindgut fermenting southern hairy-nosed wombat. However, this advantage of the tubiform forestomach for the large macropod does not apply to the smaller species with a larger sacciform forestomach, because they cannot process poor quality forage fast enough to benefit from it. There may be a pointer here to why so many herbivorous marsupials were large during the arid periods in the Pleistocene – large size favoured bulk feeding. Consummate kangaroos 307

How to conserve nitrogen and water by recycling urea An important advantage of forestomach fermentation is the greater opportunity it gives to conserve nitrogen. Unlike hindgut fermentation the protein synthesised by the bacteria can be readily digested after it leaves the forestomach. This not only provides an additional source of protein to the mammal but also enables it to re-use nitrogenous waste products, such as ammonia and urea. Amino acids cannot be stored in the body, so that amounts in excess of those required for maintenance, growth or reproduction are broken down in the liver to ammonia, which is then converted to non-toxic urea and normally excreted in the urine. However, the urea in circulation can also pass back into the forestomach of ruminants and kangaroos, where the bacteria use it to proliferate and so enhance the hydrolysis of cellulose, especially if the forage being eaten is low in protein. In addition the parotid saliva of macropods, like that of ruminants, contains urea, which is added to the contents of the forestomach. Indeed, the protein content of the sacciform forestomach can actually contain more protein than was present in the food just eaten, because the bacteria have synthesised new protein from the urea delivered in the saliva (Hume 1999). The converse of this also holds: when kangaroos and wallabies eat forage with a high concen- tration of protein nitrogen, such as legumes, the bacteria use less urea and more of it is excreted

Tammar 600

500

400

High nitrogen diet, 300 1.2% N dry weight 28 days

200

Low nitrogen diet, 100 0.4% N dry weight 28 days

Fluid urea concentration (mMolar) 0

Cortex Outer Middle Inner Inner Papilla Stripe Stripe Stripe Medulla

Outer Medulla Plasma Kidney Urine

Figure 9.9: The difference in urea retention in the kidneys of tammar wallabies, Macropus eugenii, fed on diets with high and with low concentrations of nitrogen, to show that the kidney is able to retain urea when the tammar is fed a low nitrogen diet. After Lintern and Barker (1969). 308 Life of Marsupials

in the urine. When they eat low-protein forage, much less urea is excreted and this is achieved by selective reabsorption of the urea in the kidney (Lintern and Barker 1969) (Fig. 9.9). Not only is this an important adaptation for conserving nitrogen, it also conserves water that would other- wise be required for excretion of the urea. This is an important adaptation, since water shortage and poor quality forage often go together in the habitats occupied by kangaroos and wallabies. As a result of this ability to conserve nitrogen, some macropods have much lower mainte- nance nitrogen requirements than some of the arboreal marsupials previously considered, such as the greater glider (see Chapter 7). For instance, the daily value for the tammar wallaby (230 mg N/kg0.75) is similar to the ringtail possum which practices reingestion and half that of the greater glider, which does not (see Table 7.1). However, the daily nitrogen requirements for the forest-dwelling pademelon is similar to the greater glider (530 mg N/kg0.75), whereas those for the desert-adapted euro are much lower (160 mg N/kg0.75) (Hume 1999). What this comparison shows is that the macropod pattern of digestion is capable of wide variation, depending on the environment and food source available to the species: in temperate climates the abundant nutri- tious forage passes through the gut rapidly and most urea is excreted in the urine; but where the forage is high in fibre and low in protein, it passes through the gut slowly, most of the urea is recycled and scarce water is conserved. This will become more apparent when the adaptations of individual species are considered.

The economies of hopping The extraordinary mode of locomotion was remarked on by members of Captain Cook’s party on their first encounter with kangaroos, and they noted how fast the animals could move, compared to dogs. A later observation of the endurance of a male eastern grey kangaroo being coursed by foxhounds in Tasmania was reported by Gould (1863):

I recollect one day in particular, when a very fi ne Boomer jumped up in the middle of the hounds, in the open; he at fi rst took a few high jumps with his head up, looking about him to see on which side the coast was clearest, and then, without a moment’s hesitation, he stooped forward and shot away from the hounds, apparently without an effort, and gave us the longest run I ever saw after a kangaroo. …The distance he ran, taking in the bends in the line, cannot have been less than eighteen miles [29 km], and he certainly swam more than two. I can give no idea of the length of time it took him to run this distance, but it took us more than two hours; and it was evident, from the way in which the hounds were running, that he was a long way before us; and it was also plain that he was still fresh, as, quite at the end of the run, he went over the top of a very high hill, which a tired kangaroo never will attempt to do.

If the estimate of two hours to cover the distance is correct, the average speed of the kangaroo was 14.5 km/h, which is well within the measured speeds of kangaroos of 40 km/h. This early, vivid account epitomises the unique nature of kangaroo locomotion: speed, endurance and long distance travel without exhaustion. How is it done? The first serious attempt to answer this was made when Dawson and Taylor (1973) trained two red kangaroos to use a treadmill and tolerate a mask, so that they could measure the oxygen consumption at different speeds. They found that while the kangaroos were walking, the cost of locomotion increased steeply as the speed increased until they began to hop at about 10 km/h. Thereafter the oxygen consumption did not increase up to a speed of 35 km/h (Fig. 9.10). This is very different from a quadrupedal animal of the same size, like a foxhound, in which oxygen consumption increases linearly with increasing speed. Thus, a dog chasing a kangaroo at 35 km/h Consummate kangaroos 309

consumes almost twice the amount of oxygen consumed by the kangaroo and therefore cannot sustain the chase for long. The trade off for the kangaroo is that at very slow speeds its ungainly walk, in which it also uses its tail as a fifth leg while bringing its huge hind legs forward, is much less efficient than the dog’s.

Figure 9.10: A comparison of the energetics of walking and hopping in the 18 kg red kangaroo, Macropus rufus, the 5 kg tammar wallaby, Macropus eugenii, and the 1 kg brush-tailed bettong, Bettongia penicillata. In the two larger species oxygen consumption (—) does not increase at speeds between 11 km/h and 30 km/h, compared to similar sized quadrupeds (- -); even the smaller bettong is more economical than a 1 kg quadruped. After Dawson and Taylor (1973), Baudinette et al (1992) and Webster and Dawson (2003).

The economy of hopping also holds for the pademelon and the tammar wallaby, both more than 5 kg body mass. However, the quokka, the brush-tailed bettong, Bettongia penicillata, and the long-nosed potoroo, which weigh from 1 to 2.5 kg, behave more like quadrupeds (Baudinette 1989), with the oxygen consumption rising with increasing speed. Nevertheless, the rate of increase is slower than for a quadruped of the same size, so there is still a definite energy saving when small macropods hop (Webster and Dawson 2003) (Fig. 9.10). Indeed, the smallest of these, the long-nosed potoroo (1 kg), only hops at high speeds, preferring to bound with all four feet at lower speeds (Baudinette et al 1993). The slow walking gait of the smaller species is less ungainly than that of a kangaroo and they do not use their tail to support the body while bring- ing the hind feet forward. Another feature of macropod locomotion is that they do not increase their speed by increasing the hopping frequency but by increasing the stride length. Since the length of the leg determines maximum stride length, smaller species have proportionately shorter stride lengths and faster hopping frequencies. Thus, for the brush-tailed bettong the stride frequency at all speeds was 3.5 strides/s and the stride length was from 0.3 m to a maximum 1.8 m (Webster and Dawson 2003), whereas for the red kangaroo the hopping frequency was 2.5 strides/s and the stride length from 310 Life of Marsupials

0.8 to 4 m (Dawson and Taylor 1973). The tammar is intermediate with a hopping frequency of 3.5 strides/s and a stride length 0.8–2.4 m (Baudinette et al 1987). The important muscles of the lower leg used by kangaroos in jumping are the calf muscles, gastrocnemius and plantaris. The first of these is the larger and the main agent responsible for extending the foot and so lifting the body off the ground. It begins on the underside of the femur (or thigh bone) and swells into two large muscle masses, which then narrow at the ankle into a powerful tendon that is attached to the heel. The second smaller plantaris muscle is also attached to the thigh bone but its tendon passes round the heel and attaches to the sole of the foot, especially to the large 4th digit, the great toe of the kangaroo with which it drives off from the ground (Fig. 9.11). At the end of a stride, while the kangaroo is in the air, the two hind feet are fully extended forward; then as the feet strike the ground they are quickly flexed, so that the two muscles and their tendons are stretched to their full extent. The muscles themselves change only slightly in length, because the muscle fibres are arranged in a pennate manner, but in a large kangaroo the 200 mm long tendons are stretched by 11 mm. The elastic energy stored in the stretched tendons is rapidly released in the next jump as the tendons shorten.

retractor muscles Sartorius Rectus driving Vastus muscles

Gracilis Adductor Biceps Quadratus Plantaris Semitendinosus Gastrocnemius

Tibialisanterior elastic storage tendons

Figure 9.11: Diagram of the hind limb of a kangaroo to show the disposition of the main muscles involved in hopping: the muscles that deliver the main driving force and those involved in retraction are grouped around the hip, while the muscles of the lower limb store the elastic energy when the kangaroo first lands at the end of a hop. After Alexander and Vernon (1975).

It is the special geometry of the leg and the elastic properties of the tendons that enable the kangaroo to move with such economy, once it has reached a speed of more than 10 km/h. The energy stored in a spring or a tendon is proportional to the square of the force acting on it: as the kangaroo’s speed increases, and the forces at landing increase, the amount of energy that is stored in the tendons at each stride also increases. Alexander and Vernon (1975) showed the importance of this by calculating the total energy needed for a kangaroo to hop, which at 10 km/h is 24 Watts/kg and at 22 km/h is 36 Watts/kg. However, the actual cost of hopping measured by Dawson and Taylor (1973) was 20 Watts/kg at both speeds, so the energy stored as elastic energy at each jump is respectively 4 and 16 Watts/kg, which is a very considerable economy, especially at the faster speed. When this phenomenon was first discovered some people thought that the kangaroo calf muscles and tendons might have special properties not found in other large mammals that do not hop. However, Proske (1980) found that there was no difference in either: the change in length of the lower leg muscles during hopping is negligible, because of their pennate struc- ture, compared to the substantial changes in tendon length, but the muscles provide an essential component by maintaining the tendon in tension, rather like a shock absorber. Another idea to Consummate kangaroos 311

explain the paradox was that during hopping the kangaroo may incur an oxygen debt by respir- ing anaerobically and accumulating lactate in its limb muscles, which it would then repay later by panting, as athletes do. However, the work done by the limb muscles and the lactate levels in them do not increase during hopping at increasing speeds (Baudinette et al 1992), so elastic energy in the tendon remains the most likely explanation for the economy of hopping. Baudinette and his colleagues (1987) examined two other aspects of hopping in the tammar wallaby: are heart rate and respiratory rate phase locked with hopping frequency? When a tammar or kangaroo lands all the large muscles of the hind legs are in an active state simultane- ously, unlike in quadrupeds in which the muscles in one leg are relaxed while those in the other leg are contracting. This simultaneous contraction of all the muscles might be expected to affect the return of blood to the heart and so cause the heart rate to be phase locked to the hopping frequency. In the event, as speed increases, the heart rate increases 1.8 times faster than the hopping frequency. However, the breathing cycle is phase locked to the hopping frequency in a 1:1 coupling, so that the tammar breathes in as it leaves the ground and breathes out as it lands. This is probably due to the mass of the gut being thrown forward at each landing against the diaphragm like a piston pump to expel the air. It may also explain why the tammar’s diaphragm is lighter and less muscular than in other mammals of the same size. Hopping is an extraordinarily economical way to travel for a large mammal but there is less mechanical and physiological advantage for species of less than 5 kg body mass. So what was the evolutionary driver to a bipedal gait in the small ancestors of today’s great hoppers? Small bipedal mammals can make brief vertical jumps, which are energetically expensive but provide very rapid acceleration and may be an effective means of avoiding predators. Baudinette (1991) thinks that hopping may have been selected in the ancestral macropods because it provided a better means of escape from predators on the forest floor, but once entrained it was irreversible; for instance, tree kangaroos remain predominantly bipedal despite the disadvantage of this gait for climbing in trees. However, when later descendants increased in size their bipedal inherit- ance became highly advantageous or preadaptive for travel; hence, perhaps, the rapid adaptive radiation of the larger species of macropod when they passed the 3–5 kg threshold. It may be no coincidence that most of the macropodine species are at or above this body mass (Tables 9.2–9.4). It may also be a factor in the greater decline of the smaller species, which were unable to outpace quadrupedal predators, such as foxes. While there is a lower limit above which hopping is energetically efficient, there may also have been an upper limit above which size became a disadvantage. Both species of grey kangaroo were formerly much larger than now and Flannery (1994) suggested that they became smaller in response to human hunting. Conversely the size of the red kangaroo has not changed. Perhaps today’s large kangaroos are the optimal size for hopping locomotion and the larger species were either slower or less energetically efficient. If this were so, there would have been selective pressure for smaller, fleeter body size in grey kangaroos. Likewise, the large sthenurines may have been slow and thus easier prey for people.

Embryonic diapause in macropod reproduction The pattern of reproduction described in detail for the tammar wallaby in Chapter 2 is typical of all members of the Macropodidae. Only the musky rat kangaroo is different, in that it bears twin young and does not exhibit embryonic diapause. In all non-macropod marsupials, except possibly the pygmy possums (see Chapter 6), the length of pregnancy is half to two-thirds the length of the oestrous cycle in the unmated female. Progesterone secreted by the corpus luteum reaches a peak concentration in the middle of the oestrous cycle and the young is born as the concentration declines. Subsequent activity in the ovaries is suppressed during lactation, so there is no post-partum mating or embryo in diapause. By contrast, in all species of the Macropodidae 312 Life of Marsupials

the length of gestation is between 80 and 110% of the length of the oestrous cycle, so the next oestrus and mating occurs shortly before or after the female gives birth. The longer gestation is associated with a corpus luteum that continues to secrete progesterone at a high level until the young is born, when it abruptly falls (see Fig. 2.3). A second feature of the macropodid reproductive cycle is a transient pulse of progesterone that occurs during the first week after mating. This early pulse of progesterone is important in synchronising uterine and embryo development for pregnancy (see Chapter 2). In females that give birth and begin to lactate, however, the corpus luteum formed on the ovary after ovulation remains small and inactive while the young suckles in the pouch, the early pulse of progesterone does not occur and the resultant embryo remains in diapause until the pouch is vacated by the older offspring. If the older offspring is lost or dies through failure of lactation, the progesterone pulse occurs, the dormant embryo awakens, completes its development and is born, and the process is repeated. This unusual pattern of reproduction allows the female macropod various breeding strate- gies from continuous breeding, through opportunistic breeding to highly seasonal breeding. Continuous breeding is the commonest and most widespread pattern, with successive young being produced at the end of the previous lactation. All species of the Potoroinae display this pattern and, because the interval between young is short, up to three young a year can be produced. The same pattern is seen in the desert kangaroos during favourable times but when the conditions deteriorate they become opportunistic or eventually anoestrus. In southern latitudes, where nutritious forage occurs predictably after winter rains, macropods show seasonal breed- ing, either by prolonging diapause until after the summer solstice, called seasonal quiescence, or by undergoing seasonal anoestrus in which no follicles grow in the ovaries. How the basic pattern is adapted to particular conditions and environments will become evident when particu- lar species are considered in the next section.

Sensory attributes of kangaroos The main sensory attributes of kangaroos and wallabies are vision and hearing but both vary between species in accordance with their respective life styles. The large kangaroos and wallabies have visual capacities similar to animals such as rabbits and and are different from those of cats and humans. Their eyes are positioned high in the skull, which gives them a wide field of vision but at the same time the forward field of vision of each eye overlaps that of the other by about 25°. For instance, the tammar wallaby has a peripheral field of vision of 324°, which includes 50° of binocular vision. By comparison the rabbit has 360° peripheral vision but only 24° binocular vision, while the cat has 186° periph- eral and 98° binocular vision. The wide field of peripheral vision enables the tammar to see movement in almost every direction around it, as can the rabbit, but its greater degree of binocu- lar vision enables it to see precisely what it is doing with its hands (Wimborne et al 1999). The capacity of the eye to detect the visual world varies across the retina because the density of light sensitive cells, or photoreceptors, varies from a very high density on the central horizon- tal plane of the retina, called the visual streak, to very few at the periphery; and the highest density of all is in that part of the visual streak that receives the forward binocular field of vision (Hemmi and Grünert 1999). Similarly, the ganglion cells that receive the nerve signals from the photoreceptor cells are densest on the visual streak and less dense in the periphery. Thus, the acuity of the tammar’s vision is greatest when light falls on the visual streak, where each ganglion cell receives signals from only a few photoreceptor cells, but its ability to detect movement in a wide field of vision is higher in the periphery where each ganglion cell receives input from many widely scattered photoreceptors. The reason why the most sensitive part of the retina is a long Consummate kangaroos 313

horizontal streak in the tammar and the large kangaroos is thought to be because their normal field of vision is from the horizon to their immediate surroundings: arboreal species, including tree kangaroos, lack a visual streak and the densest concentration of ganglion cells and photore- ceptors are in a central region of the retina (Dunlop et al 1988). The visual acuity of the tammar has been tested by its ability to discriminate black/white gratings of different width and in different light intensity (Hemmi and Mark 1998): it is higher than most other small mammals, including the rabbit, probably due to its hand–eye coordina- tion, but it is slightly less than the cat, and much less than human visual acuity. More remarkable has been the demonstration that the tammar has colour vision (Hemmi 1999), particularly in the blue to green part of the spectrum (420–500 nm), where it can discriminate between two monochromatic colours only 20 nm apart. It cannot discriminate higher wavelength light in the yellow to red region because it only has two types of photoreceptor cones, blue sensitive and green sensitive, which are both concentrated in the visual streak. Nevertheless, the demonstra- tion of colour vision in a macropod marsupial is interesting and raises the question whether rock wallabies also have colour vision and, if so, do they have yellow sensitive cones: some rock wallabies are brightly coloured yellow and others russet, so it might be expected that they can recognise such colours. Kangaroos and wallabies have large ears, each of which can be moved independently through about 180° (see Fig. 9.2f, Plate 15), so both ears can scan 360°, just as their eyes can scan most of the visual field. Coles and Guppy (1986) studied the acoustic sense in tammars by placing sensi- tive sound receivers near the eardrum of each ear. When the ear was pointed in the direction of the sound source the acoustic pressure at the eardrum was increased 25-fold in the optimum frequencies of 3–8 kHz and to lesser degree at lower and higher frequencies than these. The degree of amplification fell off sharply when the ear was not pointed at the sound source, so moving the pinna can give the tammar a precise knowledge of where the sound is coming from. Furthermore, because each ear can move independently of the other the tammar can detect sounds of varying intensity and frequency coming from different parts of its environment. By contrast to the large kangaroos and wallabies, tree kangaroos have short rounded ears, which are not mobile, another indication of the return to arboreal life where there is not the same need for sensing the movements of ground-dwelling predators.

Case studies in macropod adaptation

Ara, the red kangaroo The species that nicely demonstrates the three steps to success, discussed above, is the red kanga- roo: its home is the whole of inland Australia, where high temperatures, uncertain rain and lengthy drought are the ever-present threats to survival. To the Aranda people of Central Australia, who also lived in this land, Ara was one of their totemic spirits. The richness of their ceremonial and mythological life and their intimate knowledge of, and attachment to, the land was a response to the unpredictable climate, with its alternating bounty and famine. In the Aranda tradition Ara made two great Dreamtime journeys. The daytime journey encompassed 14 totemic sites from Ajaii on the western end of the MacDonnell Ranges, to the most sacred site of Ara at Krantji, and thence to Ara-perka at the eastern end. The night-time, underground, journey went from there across the desert, where Ara could not live, to Ara-ngurunja in the far north. Krantji was where the most important ceremonies were carried out: the rocks, trees and sacred hollows were struck so that, when the next rains came, every grain so dislodged arose as a kangaroo. 314 Life of Marsupials

Alan Newsome studied the ecology of the red kangaroo in the MacDonnell Ranges between 1959 and 1962 and learnt the Aranda stories from elders, who still knew the traditions, and from an earlier generation recorded by Theodor Strehlow (1947, 1971). Strehlow estimated that as much as seven-eighths of the tribal lands in Central Australia would be evacuated during drought, as the people retreated to permanent waters. Newsome’s study began in the longest drought experienced in Central Australia since European arrival and ended as the drought broke in 1962. He witnessed the massive decline of cattle from their pre-drought peak of 300 000 to less than one-third of that number by the end of the study. In that process of decline the impact of stock grazing on the native mammals was made plain: the smaller species disappeared; and the red kangaroos disclosed their remarkable reproductive response to the extreme conditions. Out of his experience of its ecology, Newsome (1980) realised that the Dreamtime songs and creation stories were grounded in a deep understanding of the red kangaroo and of its adaptations to the uncertainties of the climate. What impressed him was that 10 of the 14 totemic sites in the overland journey of Ara were places that he had independently identified as drought refuges for kangaroos. These were stream lines and grassy plains near the main ranges, to which the kanga- roos retreated during the drought years of his study. The underground journey of Ara was more difficult to understand because it traversed the desert, which his informants said was never occupied by Ara. Newsome interpreted the story to mean that no one knew how Ara could travel to the distant totemic site except by a magical underground route. And the reason that Ara could not travel across that country was because it was devoid of watering places and trees to provide shelter in hot weather. The first cattle brought into Central Australia after 1870 were likewise restricted to the floodplains and were unable to graze on the Mitchell grass plains until 1938, when bores were sunk to provide water for stock. This provided water for kangaroos as well, but more importantly, grazing by cattle stimulated the plants to put out green shoots, which is the favoured food of red kangaroos. This enabled the red kangaroos to extend their range greatly and had led to large recruitment of kangaroos in the 1950s. But it also took them away from shade on the treeless plains. Male kangaroos that experi- enced a few weeks of very hot weather on the plains had impaired spermatogenesis, and the interstitial cells that secrete testosterone were reduced in size. Also the percentage of females that had ovulated but not become pregnant was much higher in the hot weather. In other mammals high temperatures can affect spermatogenesis in this way, as can the loss of the pituitary gland or its secretions (see Chapter 2). Whatever the cause, the reproductive condition of the kanga- roos was impaired in the country that had formerly been outside their range. Whereas the food supply for kangaroos had been improved greatly, the landscape remained open and exposed, with subfertility the consequence. Traditionally, the totemic sites commemorated the creation of Ara and from the beginning no hunting was permitted for some distance around these sites, which are prime kangaroo habitat and the source of recovery after drought. In the modern idiom this is a network of conservation reserves sited in prime habitat: traditional knowledge and ecology are congruent. A century after the first cattle and sheep were brought to the inland plains an aerial survey of the three most abundant species of kangaroos was undertaken in the pastoral zones of the whole continent. The 1981 census estimated the total numbers to be 15 million, of which 8.4 million were red kangaroos, 4.9 million were eastern grey kangaroos and 1.7 were western grey kangaroos. In the same year the pastoral zone carried 25 million beef cattle and 20.5 million sheep, as well as unknown numbers of feral camels, horses and donkeys. There were also another 113 million sheep and 2 million cattle in the southern and eastern parts of the continent (Table 9.5). Later estimates of the total number of kangaroos in the rangelands have fluctuated up to a total of 35 million after years of above average rainfall. The story of kangaroos is, thus, intimately Consummate kangaroos 315

connected with the establishment of the pastoral industry in Australia, and the competition that developed between the indigenous mammals and the new species from very different environ- ments in the northern hemisphere. Indeed, all the early field studies on kangaroos between 1955 and 1965 were instigated by concerns of pastoralists that kangaroos were competing directly for resources that could otherwise be used by stock. This was the reason for Ealey’s study of euros in the of northern Western Australia (1967a,b), for Newsome’s study of red kanga- roos in Central Australia (1965, 1975), for Frith’s study of red kangaroos in New South Wales and Victoria (1964) and Kirkpatrick’s study of eastern grey kangaroos in southern Queensland (1965) (Fig. 9.12).

Table 9.5: Numbers of kangaroos and stock (millions) in the rangelands and the rest of Australia Figures are based on 1981 aerial surveys of kangaroos, and stock returns for 1981; from Caughley et al (1987a).

Red Eastern Western Total Sheep Cattle kangaroos greys greys kangaroos Rangelands 6.6 3.9 1.3 11.8 20.5 25 Other lands 1.8 1.0 0.4 3.2 112.8 2.0 Totals 8.4 4.9 1.7 14.0 133.3 27.0 Body Mass (t) 0.05 0.05 0.04 0.03 0.4 Biomass (t) 420 000 250 000 70 000 740 000 4 000 000 10 800 000

The 1981 census showed a large disparity between the total biomass of kangaroos and of stock in the pastoral zone (Table 9.5). One explanation might be that stock had displaced formerly much more abundant kangaroos, but the reports of the first explorers contradict this: kangaroos were then uncommon in inland Australia and, as Newsome indicated, their numbers increased greatly after the provision of artificial watering points. Indeed, the four common species of kangaroo have been less affected by the changes wrought by the pastoral industry in this region of Australia than have the smaller native species. So, if kangaroos were scarce before European occupation, how was the land able to support 18 times the biomass of mammalian herbiv- ores when domestic stock were introduced? Were cattle and sheep exploiting a vacant niche in the ecosystem that had formerly been filled by the extinct megafauna, as Flannery (1994) has suggested? Or was the spectacular growth of stock numbers a relatively transient phenomenon, as fossil water in the artesian basin and the plant biomass was consumed? This question will be considered after looking at the adaptations of kangaroos in this environment.

The large kangaroos of Australia Of the six large kangaroos, the red kangaroo, Macropus rufus, is the most abundant and the only one exclusively restricted to the arid zone, where it is found mainly in the better-watered plains country and low open woodlands. Its heartland is in western New South Wales and, except for the true deserts, it occurs through most of the central regions and as far as the north-western coast (Fig. 9.12). It occurs where the mean annual rainfall is less than 800 mm and unpredictable, and the mean annual temperature is more than 15°C: it does not occur in the south-west corner or the southern and eastern areas, or Tasmania, all of which have predictable winter rainfall; and it does not occur in the tropics north of latitude 14°S. About 12 000 years ago it occurred as 316 Life of Marsupials

Figure 9.12: Distribution of the genus Macropus: three widespread wallabies, including the only one to occur in New Guinea, and five of the large kangaroos restricted to Australia. Data for the red kangaroo, M. rufus, the eastern grey kangaroo, M. giganteus, and the western grey kangaroo, M. fuliginosus, across Australia are based on aerial surveys and show where the highest densities of each species occur. Also shown are the sites of the main field studies referred to in the text. After Caughley et al (1987b), Flannery (1995) and Strahan (1995).

far as the southern coast of the continent, when presumably the pattern of rainfall or the mean temperature was different from today (Flannery and Gott 1984, Caughley et al 1987b). The distribution of the euro overlaps that of the red kangaroo, where it occupies the hills and ranges, rather than the plains. It is a stockier animal with shorter hind legs more appropriate for hill climbing. A subspecies, the eastern wallaroo extends the species’ distribution beyond the Consummate kangaroos 317

rangelands, as far as the coast of eastern Australia. The euro also extends into the tropics, which it shares with two closely related species, the antilopine wallaroo and the black wallaroo. The populations of euros and wallaroos have not been estimated but are presumed to be similar to red kangaroos (Edwards 1989). By contrast, the two species of grey kangaroo predominantly occupy the eastern and south- ern regions of the continent that have seasonal, predictable rainfall. The eastern grey kangaroo (Fig. 9.2f, Plate 15), lives in eastern Australia and Tasmania, its heartland being in southern Queensland and northern New South Wales (Fig. 9.12), where summer rainfall predominates; it does not occur in South Australia or Western Australia, despite a climate favourable to it in the west, presumably because the intervening Nullarbor Plain has long been a barrier to its western spread. The western grey kangaroo is the least abundant of the four species and occurs only where winter rainfall predominates. Its heartland is in southern New South Wales, with progressively lower densities extending through South Australia to Western Australia; it occurs on Kangaroo Island but not in eastern Victoria or Tasmania. One piece of evidence suggests that the species originated in Western Australia and subsequently spread eastwards: as mentioned earlier, sodium fluoroacetate occurs in 33 species of native plants in the genera Gastrolobium and Oxylobium in western Australia but it does not occur in any plant species in eastern Australia. Like other marsupial species in the west (see Chapters 4 and 7), western grey kangaroos have a high tolerance of the poison (lethal dose 20 mg/kg body weight, Oliver et al 1979): but so do western grey kangaroos from Kangaroo Island and New South Wales. This suggests that the species originated in Western Australia, where the tolerance evolved, and subsequently spread eastwards, retaining its tolerance despite the absence of the poisonous plants. On the basis of its present distribution the species should be called the southern grey kangaroo, but on the grounds of its probable origin it is properly named. Neither the red kangaroo nor either species of grey kangaroo lives north of latitude 14°S, each for different reasons: it is too hot for the eastern grey, too wet for the red, and both too hot in summer and too dry in winter for the western grey (Caughley et al 1984, 1987b); but in the southern half of the continent there is considerable overlap of the four main species, and in parts of western New South Wales all four species occur together. While the patterns of annual rainfall largely determine the distributions of each species, this measure is less important to kangaroos than the occurrence of effective rainfall. This is a fall of rain that, after evaporative loss, is still sufficient to induce germination of seeds and the growth of ephemeral, nutritious plants, on which the kangaroos depend for food and successful reproduction. The inland kangaroos must also contend with very high summer temperatures, limited shade and the low nutritional worth and high salt content of much of the vegetation. To live in such an environment kangaroos must have effective means to control the heat load, and conserve water and energy resources. These in turn affect the breeding responses of each species and hence their abundance. The red kangaroo and the euro are especially well adapted to these conditions, each in different ways, and they will be considered first.

How red kangaroos and euros cope with heat There are two sources of heat that an animal living in a hot climate has to cope with: the heat generated by its own metabolism and the heat load from its surroundings, especially high air temperatures of 40°C and solar radiation, which may reach 126°C. Like most marsupials, kanga- roos have a relatively low standard metabolic rate (SMR), which is 70% of the placental mean (see Table 1.2), and a low body temperature, both of which are advantageous in a hot dry climate because there is less metabolic heat to be dissipated to the environment. While this is part of the kangaroo’s heritage, the desert-adapted camel also has a SMR 73% of the placental mean. The camel has evolved the ability to let its body temperature rise through the day to 46°C and 318 Life of Marsupials

discharge the heat load at night. Red kangaroos can also do this, although to a lesser extent than the camel: soon after dawn the body temperature drops 1–3°C below the mean and rises during the day to 1°C above – ie from 35°C to 37°C (McCarron and Dawson 1989). Coping with the external heat load is a much greater challenge. An unprotected person in central Australia will not survive for much more than one day of summer temperatures, so how do kangaroos live there? Euros avoid the extremes of the external environment by living in or near caves, while red kangaroos rely on greater mobility to find nutritious, succulent food and the shade of bushes. Euros are active in the day during the cooler periods of the year but in hot weather they seek the shelter of rock outcrops and caves, where available, or excavate shelter sites under dense mulga bushes. The microclimate of these sites, especially caves, is not only cooler and more humid than the outside but also removes the animal from the severe heat load of solar and other radiation. In two studies the air and radiant temperatures in caves were both less than 30°C and humidity 40%, while a site under a mulga bush had a solar radiation of 40°C, which though high is still a considerable protection (Fig. 9.13) (Dawson and Denny 1969a, Ealey 1967a). Thus, euros avoid the extreme solar radiation of midday, but in late afternoon, while the air temperature is still high, the solar radiation has declined and euros emerge from under bushes and seek the open shade of cliff faces, where they can discharge radiant heat.

Figure 9.13: Daily variation in the air temperature, solar radiation and humidity of the resting sites of: (V), red kangaroos, Macropus rufus, compared to (O), the cave sites of euros, M. robustus. After Dawson and Denny (1969a). Consummate kangaroos 319

By contrast, red kangaroos are never seen around rocky outcrops or caves and avoid the extreme outside temperatures of midday by taking shelter under small saltbushes or mulga bushes, where the air temperature is 32–38°C and humidity about 20%, but the solar radiation temperature is about 60°C. While this is much higher than in the euro shelter, in the kangaroo site there is a greater flow of air, which allows for evaporative cooling. Part of the heat load is also avoided by reducing activity to a minimum and by the posture adopted while sheltering: red kangaroos may stand with the large tail drawn under the body, so that it also is shaded from the heat but air can flow around the legs and body held off the ground. The pale colour of the red kangaroo reflects a higher proportion of solar radiation than does the darker coat of the euro. Also the fur of the red kangaroo is finer and more than three times as dense as that of the shaggier euro – 62 fibres/mm2 compared to 20 fibres/mm2 – and this greatly increases the insula- tion under windy conditions, in both hot and cool conditions (Dawson and Brown 1970). Both types of behaviour reduce the heat load and the need for evaporative cooling and hence conserve body water. Nevertheless, cooling is necessary and kangaroos have three means of evaporative cooling, which they employ when the air temperature is high or they have been exercising: they can pant, sweat or spread saliva on the skin. By far the most important of these is panting: when the air temperature was experimentally increased from 25°C to 45°C the respira- tory rate of both species increased 15-fold. Initially the airflow is through the nose, with evapo- rative cooling on the nasal epithelium, but as the air temperature increases the kangaroos begin open mouth panting, and evaporative cooling occurs across the surface of the throat. The partic- ular advantage of panting, over sweating, is that the cooling affects the deep body temperature of the animal, especially the brain, even while the skin temperature may remain high from direct contact with the high air temperature. The importance of panting for desert-living kangaroos is reflected in the nasal turbinals, which are much larger in red kangaroos than in grey kangaroos that live in less extreme temperatures. Red kangaroos and euros only sweat during exercise and stop as soon as they rest, even though the body temperature may still be high and they continue to pant. This remarkable ability seems to be another adaptation for conserving water, since sweating only cools the skin, which will be rapidly negated by inflow of heat from the high air temperature; thus panting uses precious water as a coolant more efficiently. The third way that kangaroos cool themselves is by spreading saliva on the forearms (Dawson 1973). This saliva is secreted by the maxillary salivary glands and has a different composition from the parotid saliva secreted during feeding; it also contains less protein and phosphates than parotid saliva. As it evaporates it cools the blood returning to the heart through large veins directly under the skin of the forearm. While the dryness of the air aids evaporative cooling by this means and it does affect the deep body temperature more directly than sweating does, it is an extravagant use of water, which must be made good from succulent vegetation or by drink- ing. It is used as a supplement to panting by kangaroos experiencing high air temperatures but, if their water is restricted, they cease to produce the saliva. Antilopine wallaroos in the tropics also use this means of evaporative cooling but spread saliva on the inner thighs, rather than the forearms. It is clear from the above that water is essential for temperature regulation of desert kanga- roos. This they obtain from the plants they eat or by drinking at soaks or stock dams; so the frequency of drinking is a nice measure of the wellbeing of euros and red kangaroos.

How euros live in the arid environment Euros are more sedentary than red kangaroos and occupy smaller home ranges and, in one study where 63 euros were collared, none moved more than 7 km, even during drought, compared with 20–30 km for red kangaroos at the same time (Croft 1991b). Their social structure is also 320 Life of Marsupials

different from the red kangaroos (Croft 1981a): apart from the mother–young relationship, which is the most enduring social relationship, adults avoid direct contact when sharing shelter, and live alone. Nevertheless, because the population is sedentary, animals sharing a common area meet frequently, but aggressive behaviour is more evident than among red kangaroos. Euros compete for caves and shady places, with high frequencies of agonistic interactions at these sites. Usually the largest euros get access to caves and lying up shelters, only sharing if neither can displace the other. Because females are smaller than males, they usually lose such encounters but they overcome this by occupying areas in caves or overhangs too small for large males. The mating system is probably polygynous with the alpha male mating with all or most of the females and retaining dominance for about one year. After being displaced as the alpha male, fertility may decline, for 38% of the largest males in one study were aspermous (Sadleir 1965). In the Pilbara of Western Australia, Ealey (1967b, Ealey et al 1965) found that the use of caves was closely related to the availability of water: when water was scarce or the moisture content of plants was low, euros sought relief from the heat in cool caves and so conserved water that would otherwise be required for thermoregulation. However, when water was abundant at artificial stock dams, euros did not use these heat refuges during the day, despite high ambient tempera- ture, most being found on the shaded side of granite outcrops. Likewise, the frequency of drink- ing at stock dams varied with the availability of caves and rainfall: some animals, predominantly lactating females, came regularly to stock tanks but about one-quarter of the population never did. Drinking at the tanks declined after falls of rain of 30 mm and was less intense in areas where caves were present. Taken together, the observations of Ealey and his colleagues suggest that euros can survive without free water if they have access to cave shelters; but more euros can occupy an area if water is available.

Euro nutrition – getting the most from the plants The euro strategy for survival in the arid regions is similar to that of the southern hairy-nosed wombat: by exploiting the lower temperatures and higher humidity of caves euros conserve water and obviate the need to drink, but like the southern hairy-nosed wombat they are restricted to the country near the refuges and to the plants that grow nearby. Also, like the wombat, euros can conserve nitrogen to a greater degree than red kangaroos or sheep (Freudenberger and Hume 1992) and this enables them to obtain their essential protein requirements from plants that have a very low nitrogen content. Ealey and Main (1967) compared the responses of two populations of euros in the Pilbara, one living on nutritionally poor vegetation of spinifex, pungens, which is only sparingly eaten by red kangaroos or sheep, and had a mean nitrogen content of less than 0.6% dry weight, and another population feeding on mixed vegetation with an average nitrogen content of 1.1%. The euros feeding on spinifex had urea concentrations one-fifth of those on the better site, presumably because they were reabsorbing the urea in the kidneys, as tammars do (Fig. 9.9). Like all macropods, euros can respond to a low nitrogen diet by recycling urea to the forestomach, so that lower urea concentrations in urine reflect a poorer diet and greater urea recycling. The critical factor for euros is not so much the total amount of vegetation but its protein value, which in turn is related to rain. Euros can live and even breed on a low protein diet but when drought is prolonged the protein content of the plants declines, the animals go into negative nitrogen balance and fail to maintain lactation: nutrition rather than water is the major factor in regulating density of euros. An interesting corollary of this in the Pilbara in 1954 was that where the prevailing vegetation had a low protein content, the population remained low through regular mortality, whereas where there was other more nutritious vegetation the euro population increased much above the long-term carrying capacity and, during the next severe dry period, many euros died (Ealey et al 1965). Consummate kangaroos 321

In Central Australia there are few euros but red kangaroos are common (Newsome 1975). Permanent rock holes are rare and far apart and because the soils are sandy, seepage sites are also rare. Euros are faced with little good shelter from heat and little chance to recoup any lost water through drinking. Although its preferred food, Triodia sp., is widespread, the euro is uncom- mon. Therefore, the reason that the euro is rare in Central Australia, compared to the Pilbara, is probably the scarcity of heat refuges and drought fodder.

How red kangaroos live in the arid environment Red kangaroos cannot escape the heat to the same extent as euros and thus water lost through evaporative cooling during panting and licking the forearms must be replenished by selecting better quality forage than that used by euros, and by better conservation of body water. The kidneys of red kangaroos have a greater concentrating ability than those of the euro and, where the two species were studied together, their urine was consistently more concentrated than that of euros, especially in summer (Dawson and Denny 1969a,b). Red kangaroos have a strong preference for green feed, especially newly sprouted grasses and forbs, which they are able to select even when the vegetation seems dry and brown. In Central Australia green grass comprises 75–95% of their diet and consists mostly (54%) of kangaroo grass, Eragrostis setifolia, which remains green during dry times (Newsome 1975). Similarly in western New South Wales this grass and bottlebrush, Enneapogon avanaceus, comprised 21–69% of food eaten and, since their proportion in the stomach was greater than in the pasture, the kangaroos were deliberately selecting these two species (Bailey et al 1971, Caughley et al 1987a). When condi- tions become dry red kangaroos congregate on the open plains and along watercourses because only there can they find enough green grass growing on seepage lines to meet their water and protein needs. Although they prefer to eat grasses and forbs, when these become scarce they switch to chenopods, Bassia diacantha, and black bluebush, Maireana pyramidata, and in some areas will even browse shrubs. However, they never eat some perennial chenopods, such as round-leaf chenopod, Kochia, which euros eat, even when these plants are abundant (Ellis et al 1977). Because they are not constrained to living close to caves, red kangaroos occupy much larger home ranges than euros, living on different types of country in different seasons to obtain their preferred food. In western New South Wales Croft (1991a) found that red kangaroos had weekly home ranges of 259–560 ha, with large animals, mainly males, occupying the largest areas. Moreover, when forage is poor and rainfall patchy they will travel 25–30 km to more favourable sources of nutritious food (Frith 1964, Bailey 1971, Caughley et al 1987a). In Central Australia Newsome (1965) observed that two-thirds of the red kangaroos stayed within 1 km of persistent forage during dry times but after effective rain fell they moved up to 28 km to feed on the new growth of ephemeral forbs and grasses. These directed movements of kangaroos when rain falls may involve many kangaroos and be rapid: Croft (1991a) followed 188 radio-collared red kangaroos in a larger population that was also being counted from aircraft. Rainfall had been low on his study site through 1981 and early 1982 and in February there were 5211 kangaroos on the study site. When in April good rain fell about 30 km away none of the radio-collared kangaroos remained on the study site, and the total population on 20 April had dropped to 1336. Then in late June good rain fell on the study site and by 30 June the population was back to 7508 and was still 5707 in August. Eight radio- collared female kangaroos tracked to the distant site came back but three marked males were shot there. In this example about 4000 red kangaroos moved 25–30 km to local patchy rainfall when conditions in their home range were poor, with females showing high fidelity to their home range by returning after recovery of the pasture. These two examples show what happens when rain falls while the kangaroos can still move. However, when rain is long deferred, starving kangaroos are unable to muster the resources 322 Life of Marsupials

to escape the deteriorating conditions and eventually die on their home site. Nevertheless, red kangaroos have an extraordinary capacity to endure starvation and dehydration and recover: they can lose more than 20% of their body weight, which is much more than most large mammals, such as cattle, and comparable to camels, which can lose more than 30%. While most marked red kangaroos show this nomadic pattern of movement within a 30 km radius of the point of release, a few kangaroos travel huge distances in a short time. The fastest record is of a young female that moved 338 km in 15 weeks in Western Australia (Oliver 1986), while the longest duration was a male kangaroo, marked when two years old in western New South Wales and recovered 25 years later 300 km south of the original study site (Bailey and Best 1992). This pattern of long-distance movements has been observed in all studies of red kangaroos but it is still unclear why some animals travel so far: no particular age group or sex is involved.

Social behaviour of red kangaroos Red kangaroos associate in small groups of two to four animals and show less agonistic behav- iour than euros. Females with young at foot are the commonest group and, at higher densities, female kangaroos coalesce in larger groups, usually with one male (Frith 1964, Croft 1981b, Johnson 1983). The males reach sizes almost twice that of females, and develop proportionately much larger shoulders and arms than females (Jarman 1983), which they display while establish- ing a dominance hierarchy. The main agonistic behaviour is between young adult males, which engage in spectacular ritualised ‘boxing’. Each member of a pair holds its adversary’s shoulders and then, supported on its tail one rakes the other with both hind feet. These ritualised bouts establish dominance relationships between aspiring males, leading eventually to alpha status and first access to oestrous females. Alpha males display much less agonistic behaviour and more sexual behaviour, until displaced by a challenger, whereupon they live alone and avoid close contact with other animals. In one study (Sadleir 1965) a high proportion of the largest males had no sperm in their ejaculates and it may be that these were old displaced males, although that has not been established in any subsequent study. By contrast, subordinate males do mate with females after the dominant male has mated and their sperm presumably compete to fertilise the egg: but this presumed competi- tion has been tested critically only in the tammar wallaby (see Chapter 2). Thus, these two very successful, widely distributed and closely related species of kangaroo have exploited different aspects of their common macropod heritage in adapting to the arid environment: the euro keeps near its caves and thereby conserves water, but must perforce be sedentary; it has adapted to the nearby and less nutritious herbage by greater use of urea recycling and a very low nitrogen and basal metabolism. Conversely, the red kangaroo can travel long distances to exploit the nutritious and succulent plants that bloom after rainfall but has a higher water turnover required for thermoregulation on the open plains. Newsome (1975) thinks that the two species, which can interbreed to produce sterile hybrids, have only recently diverged from a common ancestor and that the key factor in their divergence was diet: one species, able to survive on a hard diet of spinifex, became sedentary while the other, with a dietary requirement of green herbage, became nomadic. But no species can long survive unless it leaves offspring. Euros and red kangaroos reconcile the substantial costs of reproduction and an uncertain climate by exploiting their macropod heritage of embryonic diapause in an unusual and remarkable way.

Reproduction in an uncertain climate Both euros and red kangaroos are opportunistic breeders, so that their reproduction must be considered in relation to good conditions, to drought and to the conditions after breaking rains Consummate kangaroos 323

(Fig. 9.14). Under favourable conditions adult females of both species breed continuously and have high fecundity throughout the year. When examined in the field they have simultaneously one offspring running at heel, which suckles an elongated teat from outside the pouch, and a very much smaller offspring attached to another teat inside the pouch; while upon dissection, a tiny blastocyst can be flushed from one of the two uteri (see Chapter 2). By the time the pouch young of the red kangaroo is 120 days old the young at heel is fully weaned but still associates with its mother; when the pouch young is about 200 days old the uterine blastocyst resumes develop- ment, which it completes in about 33 days, and is born a day or so after the older young has been permanently excluded from the pouch. Post-partum oestrus and ovulation follow and another egg is fertilised then. So, although it takes a particular offspring nearly 600 days to develop from conception to independence, the female kangaroo can produce an independent offspring every 240 days, while good conditions prevail.

Prevailing good conditions Red kangaroo Independent

At heel

In pouch

In uterus

Oestrus Embryonic diapause Gestation Drought breaks

Temporary improvement At heel Deteriorating conditions Severe drought

In pouch

In uterus

0 100 200 300 400 500 14 600 days Days

Figure 9.14: Diagram to illustrate the opportunistic breeding strategy of the red kangaroo, Macropus rufus, and how females and their young respond to the prevailing conditions as influenced by rainfall, indicated by shading.

As conditions deteriorate, and for red kangaroos this means primarily the quality and quantity of available feed, reproductive activity is progressively affected. The first effect is on the young at foot: its diet is changing from milk to herbage, it has to maintain its body temperature and it is growing fast. Its SMR is twice as high as its mother’s and its body temperature and evaporative heat loss are also significantly higher, especially at the highest ambient temperatures that it is likely to encounter in the summer (see Chapter 2) (Munn and Dawson 2001), so its need for water much exceeds that of adults in the same environment. The high metabolic rate required for growth, and the high costs of regulating its body temperature during the summer, puts a great demand on the newly weaned red kangaroo for water, energy and protein. When conditions are good these needs are adequately met from the pasture and milk is not required as a supplement: however, if the pasture quality is low in protein and water, milk can supplement 324 Life of Marsupials

this if the mother can still supply it (Munn and Dawson 2003a), but under such conditions she usually cannot. This leads rapidly to the death of the young kangaroo: but the female kangaroo still carries a small young in the pouch and has another elongated and dry teat as evidence of the lost young at heel. If the drought persists, secretion of late stage milk declines, growth rate of the small pouch young slows, it becomes emaciated and dies; this is usually at the age of about two months when maternal investment in the young is about to increase (see Chapter 2). The decline of the pouch young triggers development of the uterine blastocyst and birth occurs as the older one dies. This next pouch young also develops for two months before undergoing the same fate as the previous one. This process may be repeated several times unless rain falls: then plants grow, milk produc- tion increases and the young continues to develop to the end of pouch life. This unusual adapta- tion of the macropod reproductive pattern allows breeding by red kangaroos to continue well into a drought and for females to take immediate advantage of an improvement in conditions when they occur. This accounts for the common observation that soon after a drought breaks most female red kangaroos and euros have small young in their pouches, which gave rise to the popular but erroneous idea that kangaroos have some extraordinary foreknowledge of when a drought will break and they anticipate the coming rain. What they actually do is much more remarkable. A similar response to inadequate nutrition was observed in eastern wallaroos on two sites in New South Wales, one with improved pasture and fertiliser and the other with Poa and no fertiliser (Taylor 1982). There were no differences in age of maturity or incidence of breeding at the two sites but on the untreated native pasture only 12% of the females had young at foot compared to 30% on the improved pasture, which suggests that the vegetation on the former site was insufficiently nutritious to maintain late lactation. Should drought persist for more than six months, female red kangaroos and euros eventu- ally cease to breed altogether; at the end of the final pregnancy they do not ovulate and instead enter a state of anoestrus until they either die or the drought breaks. The condition of the ovaries and uterus in such animals resembles the condition seen in tammar wallabies after the pitui- tary gland has been removed, and the inference is that secretion of pituitary gonadotrophic hormones ceases under these extreme conditions. When the rain comes and the plants grow again the response in the kangaroos is immedi- ate: in two studies, made two weeks after good rainfall had broken a year-long drought, 65% of the female kangaroos examined were in oestrus (Frith and Sharman 1964, Sharman and Clark 1967). Since it takes 10 days for ovarian follicles to mature, this indicates that the kangaroos were responding directly to the rain, rather than to the new feed that would follow. The strong scent given off from dry ground after rainfall – called petrichor (Bear and Thomas 1966) – may provide the immediate stimulus to the pituitary gland via the olfactory nerve, rather like pheromones do, although this has not been investigated. Thus, the main regulator of red kangaroo populations is mortality of late stage pouch young and young at foot, which depends in turn on sufficient feed for the females to maintain lactation and the young at foot to sustain their very high metabolism. Frith and Sharman (1964) found at three sites in New South Wales that 83% of young leaving the pouch died at the driest site (Gilruth plains), 53% at the less severe climate of Mt Murchison, and only 15% at Toganmain, where forage was abundant. Because of this, recruitment into red kangaroo populations is very uneven, with large numbers in good years and none in other years: in Newsome’s study between 1959 and 1962 the major recruitment to the population had occurred 12 years before in the exceptionally good years of 1947–49 (Newsome 1975). Consummate kangaroos 325

Grey kangaroos of woodland and forest: adaptations and speciation By contrast with the red kangaroo and euro, both species of grey kangaroo live in those parts of Australia where seasonal changes are more regular, rainfall predictable and the vegetation more abundant. Nevertheless, there is a broad zone of overlap where two or more species occur together, and this is also the region where sheep are the main commercial species. In these circumstances, do the different herbivores compete for the same plant food or do they use differ- ent species? Although there have been many studies on food preferences of kangaroos and stock, this question is still difficult to answer. The reason is that the selection of different plants by each species varies in relation to rainfall and to the interactions between the species grazing the land. When rainfall has been sufficient for plant growth, each species prefers different classes of plants and there is no apparent competition: when dry conditions prevail all species are forced to share the same diminishing resources. The findings of 23 studies were summarised for the rangelands, where the four main herbiv- ores are sheep, red kangaroos and both species of grey kangaroo. This showed that in good condi- tions red kangaroos and eastern grey kangaroos prefer to eat grasses and annual forbs, while western grey kangaroos and sheep have a broader diet that includes, as well as grass and forbs, chenopod shrubs, saltbush and mulga. During dry periods, however, the preferences change: red kangaroos and eastern grey kangaroos still seek grasses and, even in severe drought, will not eat saltbush and shrubs; indeed red kangaroos starve to death before irreversibly modifying the vegetation. By contrast sheep and western grey kangaroos shift to eating chenopod shrubs and mulga. The first studies conducted in southern Queensland, where eastern grey kangaroos occur together with red kangaroos and sheep (Caughley 1964, Griffiths and Barker 1966, Griffiths et al 1974), is instructive of these interactions on the ground. In Caughley’s (1964) study faecal pellets from eastern grey kangaroos were more abundant in the denser vegetation and almost absent from the open treeless grasslands, which agreed with sight traverses through the same habitats: eastern grey kangaroos were more common in the woodlands than the grasslands. Conversely, red kangaroos were seldom encountered in the denser vegetation but frequented the open woodlands and grassland; so, on this station the two species of kangaroos were dividing the habitat rather than competing for the same areas, and this was reflected in their selection of food plants. Plant preferences were determined either by collecting samples from the mouth and stomach of kangaroos and sheep shot while feeding at different times of year or from fresh dung (Griffiths and Barker 1966, Griffiths et al 1974). Notwithstanding their preference for dense cover the eastern grey kangaroos never ate the foliage of the main woody species mulga, Acacia aneura, or berrigan, Eremophila longifolia, but predominantly (64–79%) ate the grasses, such as spinifex, Triodia mitchelli, growing in the woodlands. The predominant species of the open country, Mitchell grass, Astrebla pectinata, was absent from their diet but as already noted, comprised a major part of the grassy diet of the red kangaroos. Like the grey kangaroos, the red kangaroos eschewed the mulga and berrigan, but unlike the greys, more than half their diet consisted of small dicotyledonous plants, as noted before in comparing them with euros. By contrast, sheep browsed mulga and berrigan in all seasons, and like the red kangaroos selected succulent dicoty- ledons. However, the particular species selected differed: sheep chose species of Malvaceae not eaten by red kangaroos; and the red kangaroos selected Portulaca olevacea, which was never found in sheep stomachs. Thus, in favourable conditions the three species were each selecting different species in the pasture; but as the pasture deteriorated due to lack of rain and an influx of additional sheep, all 326 Life of Marsupials

three species were obliged to eat more of the perennial grasses, such as Triodia, and showed less diversity in their diet. This is similar to the remarkable associations of mixed herds of wild herbiv- ores on the east African grasslands, where each species has particular food preferences, which do not overlap with those of other species. But unlike the African grasslands, where the annual growth of grasses is highly predictable, the Australian rangelands have an uncertain climate. In the Australian rangelands pasture growth is closely predicted by rainfall over the previous three months: rainfall is the prime driver but is unpredictable, therefore the pasture biomass is highly variable and can vary up to 100-fold in two years (Fig. 9.15). At Kinchega (Caughley et al 1987a) the kangaroo populations increased when pasture biomass exceeded 200 kg/ha and decreased when it was less: population size was tightly controlled by pasture biomass that was controlled by rainfall. Two negative feedback loops control this process: the pasture biomass loop and the pasture–herbivore loop. At high pasture biomass the pasture grows less vigorously or stops growing, so grazing stimulates growth; and the more pasture available the more each

Western grey kangaroo 30

20

10

0 30 Red kangaroo

20 Births per 100 adult females adult 100 per Births

10

0 1200

1000

800

600

400

200 Pasture biomass (kg/ha) 0 1980 1981 1982 1983 1984

Figure 9.15: A comparison of breeding by red kangaroos, Macropus rufus, and western grey kangaroos, Macropus fuliginosus, at Kinchega, western New South Wales, during and after a severe drought in 1981–84, to show the different responses of the two species. After Caughley et al (1987a). Consummate kangaroos 327

kangaroo eats and the more rapid is the reproductive response (faster growth rate of pouch young and high success rate); so the kangaroo population increases and pasture biomass declines, which stimulates plant growth. Thus, while the short-term changes in plant biomass and kangaroo density may appear chaotic they are tightly deterministic and the chaos is due to the unpredictable rainfall. In good seasons, when the pasture biomass exceeds the long-term average, both sheep and kangaroos can be supported without affecting each other, but when the plant biomass falls below 200 kg/ha they are in competition.

Competition between eastern and western grey kangaroos Although eastern and western grey kangaroos have probably diverged from a common ancestor quite recently, the ecological differences between the two species are subtler than those between red kangaroos and euros. Indeed, until 1972 they were not recognised as separate species and in the wide zone of overlap in New South Wales (Fig. 9.12) it is impossible to separate the two species from the air and difficult even on the ground. Nevertheless, they are clearly distinct species: when crossed, male hybrids are sterile and female hybrids have gestation lengths inter- mediate between the two species. So, the question is how two such similar species can coexist. Do they share the habitat and so avoid direct competition, like red kangaroos and sheep do, or are we witnessing today the invasion by western grey kangaroos of habitat formerly occupied exclusively by eastern grey kangaroos? Caughley and his colleagues began to address these questions by selecting two sites in New South Wales where both species occurred (Caughley et al 1988): the southern location was in the heartland of the western species’ distribution and at the western limit of the eastern species’ distribution; while the northern location was in the heartland of the eastern species’ distribu- tion and at the northern limit of the western species’ distribution. Females of both species were sampled at both sites in mid winter when they would be carrying pouch young. There was little difference between the species at the two sites in the plants eaten or the time of breeding, except that both species bred a few weeks earlier in the south. As was expected, the density of each species was highest in its core area and low in the core area of the other species. Each species was also fatter in its core area than in its marginal area but both species were heavier in the south than in the north, reflecting Bergman’s Rule that mammals are larger at higher latitudes. The only factor that was significantly different was the recruitment rate, measured as the percentage of females carrying pouch young. For both species 90% of the females in their respective core area had pouch young, whereas the percentage with pouch young in the marginal part of their ranges was less: in the case of the western grey kangaroo, only 62%. There was no evidence of varying mortality factors, such as predators or disease, so the tentative conclusion was that failure to replace animals lost through death or emigration is the main factor limiting the distribution of each species. For the western grey kangaroo, which relies on plant species that respond to winter rainfall, the decline of these species in the north, where summer rainfall patterns prevail, may be the cause of the lower recruitment. Conversely the eastern grey kangaroo has been moving westward for the past 70 years due partly to the increase in watering points for sheep and cattle, as in Central Australia. For both species the western boundaries of their respective ranges are proba- bly maintained by competition with red kangaroos and euros, because the latter species have a better tolerance of high temperatures and uncertain rainfall. The factors that affect the range overlap between the two species of grey kangaroo are harder to understand. Perhaps one must look to the past history of the three species. As mentioned earlier, the red kangaroo occurred as far south as Melbourne 12 000 years ago, when the climate was hotter and drier than today. Presumably grey kangaroos were also displaced from the south- 328 Life of Marsupials

ern region and therefore separated from each other. Since they are clearly closely related, this would suggest that even earlier, when the climate was more benign, a single species of grey kanga- roo would have extended right across the southern regions of Australia and Tasmania. During the arid period they were separated by the red kangaroo, which subsequently retreated northwards, allowing both wings of the former grey kangaroo species to reoccupy Victoria. But now the differ- ences in the two populations were sufficiently distinct for them to have become separate species.

Grey kangaroos are seasonal breeders, not opportunistic Unlike the desert kangaroos, the grey kangaroos are seasonal breeders. Most females give birth in early summer. Growth of the pouch young is slower than in red kangaroos, so that the young leaves the pouch after about 280 days in the following spring, after predictable rainfall and plant growth (see Fig. 2.1), and is still running at heel when its mother next breeds. Post-partum ovulation does not occur in either species and grey kangaroos generally do not carry dormant embryos in the uterus, so that the whole pattern of breeding in these two species is different from the desert-adapted kangaroos. However, under very favourable conditions eastern grey kanga- roos in Queensland may come into oestrus and conceive during the second half of lactation: when that happens the embryo remains dormant until the end of pouch life of its older sibling. Thus, the potential for diapause is there but it does not occur as a normal feature of repro- duction, as in the desert species. Conversely, the western grey kangaroo, which lives entirely within the regular winter rainfall region of Australia, has lost the capacity altogether: on the very rare occasions when a female western grey comes into oestrus during lactation and conceives, the embryo completes its development without pause, enters the pouch with its older sibling and dies soon thereafter (Poole 1975). The relative advantage of seasonal and opportunistic breeding was nicely demonstrated at Kinchega National Park, in western New South Wales, where Caughley and his colleagues compared the responses of red kangaroos and western grey kangaroos to drought and recovery (Caughley et al 1987a). Like Newsome’s study, this one began after several years of substantial rainfall, which had resulted in a large recruitment to the populations of both species. Then, throughout 1982 the rainfall was very low and insufficient for the germination of seeds of ephemeral grasses and forbs. The pasture biomass declined from more than 800 kg/ha in late 1980 to 9 kg/ha in September 1982 and remained low for the rest of the year (Fig. 9.15). Since kangaroos need about 200 kg/ha of perennial grasses just for maintenance, the situation was dire. The kangaroos changed from grasses and forbs to saltbush and bluebush perennials, but these were insufficient to support them and 14 000 died by the end of that year. This was about 35% of the red kangaroo popula- tion and 75% of the western grey kangaroo population. Rain fell in March 1983 and continued through the rest of the year, so that by year’s end the plant biomass had recovered to 1100 kg/ha. Each species responded to these conditions differently. Through the drought red kangaroo females continued to give birth but the young did not survive for more than 50 days in the pouch and were then replaced by another. After the breaking rains in March 1983 the current batch of small pouch young survived, so that by the end of 1983, juveniles made up 25% of the popula- tion. This first pulse of recruitment was followed at eight-month intervals by two more pulses, so that, by the end of 1984, most red kangaroo females had produced three successive young to independence (Fig. 9.15). By contrast, the western grey females did not replace lost young during the drought and then did not respond for seven months after the breaking rains in March. Then, in October 1983, which is the normal time of the year for this seasonally breeding species, more than half of them gave birth: they were apparently unable to respond to the earlier, favourable conditions until they had also received an appropriate photoperiod signal. By the end of 1984 the western grey females had each produced two young, compared to the three produced by the red Consummate kangaroos 329

kangaroo females. However, because their young grow slower, juveniles only began to make up 25% of the population by the end of 1984, a year later than the red kangaroos. Being opportunistic breeders, the red kangaroos could exploit embryonic diapause and respond immediately to the favourable conditions after breaking rains, whereas the western grey kangaroos, being locked into a pattern of seasonal breeding, could not respond opportunistically and so took much longer to recover from the drought. At Kinchega the western grey kangaroos are living near the northern limit of their range and in the drought they suffered much higher mortality than the red kangaroos, and recovered much more slowly after rain fell. This may be a major reason why the heartland of the western grey kangaroo is in the southern winter rainfall region of Australia and it does not occur in central Australia. During those same years, Norbury et al (1988) observed the same pattern in another population of western grey kangaroos further south, in Victoria. While their seasonality is a disadvantage at the edge of their range, it is highly advantageous in the southern parts, such as on Kangaroo Island, where the winter rainfall is predictable and the spring flush of growth very beneficial to the emerging young (see Fig. 2.1).

Social behaviour of grey kangaroos The social behaviour of grey kangaroos reflects their seasonal breeding and preference for woodland habitat. There are three common associations related to essential life functions: male– male agonistic behaviour to establish hierarchical rank; males courting oestrous females; and the mother–young association. During autumn and winter western grey males at Kinchega live in large groups away from females. They engage in agonistic behaviour, especially boxing to determine male dominance and use of arms in fighting and threat displays. These displays involve an exaggerated and stiff-legged walk that emphasises the male’s size and chest: this is further emphasised by picking up grass and rubbing it against sweat glands on the chest to spread the odour (Johnson 1983). These displays and sparring bouts in winter establish a dominance hierarchy among the males and, when females come into oestrus in the spring and summer, the largest males get first access to the females. Lactating females live apart from other adults except for an associated immature male or female offspring from the previous year. When her young is about to leave the pouch a female grey kangaroo takes herself away from the mob and during this time the young gets in and out of the pouch and does little circular runs: this period establishes the close bond between both of them, including recognition of each other’s special call. At this time the mother and young engage in mutual nuzzling, which may aid in the transfer of bacteria and protozoa to the young animal’s forestomach. The association continues for at least another year but the mother’s relationship to her offspring differs between male and female young: in good seasons male young grow faster than female young and dissociate from their mother earlier; conversely, when condi- tions are adverse a female grey kangaroo may discard her male pouch young but not a female young. In this way the sex ratio of pouch young differs in relation to rainfall: more male young are produced in good years and more female young in dry years. Female young associate with their mother longer than male young and may continue to do so even when they themselves have begun to breed (Jarman and Coulson 1989).

Competition with grazing stock We now return to the question posed at the beginning of this section: were cattle and sheep exploiting a vacant niche, or was the spectacular growth of the stock numbers a transient phenomenon, as fossil water and the plant biomass was overexploited? Newsome (1975) reviewed the evidence from the Pilbara and Central Australia and later Caughley et al (1987a), with more direct evidence, reviewed the situation in western New South Wales. 330 Life of Marsupials

Prior to European settlement euros and red kangaroos were evidently in low numbers in the Pilbara, constrained by lack of water and predation by Aboriginal people. Similarly in Central Australia the first explorers reported red kangaroos to be rare, and euros even more so. Sheep were introduced to the Pilbara in 1866, initially around natural waters and then, as bores were sunk every 10 km or so, the sheep were moved into drier country to feed on the native pastures of the plains. By 1920 there were 700 000 sheep and a few thousand cattle in the region and they remained at this level until the drought of 1935–36, after which sheep numbers fell to 300 000: by 1945, after the arrival of blow flies, Lucilia cuprina, many of the properties were abandoned. While some of the sheep decline was due to mortality during dry years, the main cause of the decline was the failure of the ewes to breed after 1930, because of the loss of the most favoured plant species with the highest protein content: after 1945 numbers were only maintained by importing sheep to offset the failure of local breeding. The practice of the time was to overstock in the anticipation of good rains and a large profit from wool: if rain did not come, the stock were left to die and, in the process, reduced the vegetation further. In today’s terms this was asset stripping, in which fossil ground water was used to convert plant protein to keratin in the highly profitable form of wool. Australia ‘rode on the sheep’s back’, but at the cost of losing the accumu- lated capital of ground water and the most nutritious plant species. An unexpected consequence of sinking bores was to remove the main constraint on the euro and red kangaroo populations of the region, which then increased greatly and were blamed for the decline of sheep. After the drought of 1934–35 all three species starved but only euros and red kangaroos recovered: sheep did not because by then the nutritious grasses and forbs had gone and spinifex dominated the pasture. The next drought in 1944–45 caused extensive mortal- ity of both euros and red kangaroos, after which only the euro recovered because it alone could thrive on spinifex. By 1959 some properties were carrying 10 times as many euros as sheep and the industry ceased soon thereafter (Ealey 1967a). In 1960–62 poisoning at stock water points, and their subsequent closure, drastically reduced the euro population, which has never returned to its previous high numbers. In 1978 part of the ravaged land was transferred to the Western and in 1991 it was returned to the Aboriginal owners. In the 40 years since pastoralism ceased and the euro population was reduced, there has been some regeneration of perennial grasses and a partial return to the pre- pastoral state, but red kangaroos remain extremely rare (Berry 1991). In Central Australia cattle were introduced in the 1870s along the river flats adjacent to permanent water, as in the Pilbara (Newsome 1975). The numbers remained at less than 60 000 until after 1938, when bores were sunk in the extensive grassy plains. Helped by a long run of good seasons during the next 20 years, the number of cattle rose to a peak of 353 000 in 1958. As mentioned earlier, there was substantial recruitment into the red kangaroo population at the same time, brought about by the added water points and the alteration of the grasses, especially the new green shoots induced by cattle cropping the dry stems. During a succession of dry seasons that began in 1958, the cattle numbers fell to 100 000 by 1965 and the red kangaroos also declined, due to their failure to breed. Whereas both cattle and kangaroos feed largely on grass, the cattle eat more species and also eat shrubs and tree foliage that kangaroos do not eat, so while good conditions prevail there is a degree of separation of the two herbivores. Competition between them occurs when rain fails and the plant biomass declines. In their ability to feed on mulga bushes and saltbush, which kangaroos do not eat, cattle may, indeed, be tapping into a resource that was formerly eaten by the large extinct marsupials. Unlike the situation in the Pilbara, cattle continue to be grazed in Central Australia decades after wells were sunk. The highest densities of all four species of large kangaroo are on the sheep rangelands of eastern Australia, where they are probably now more abundant than when the sheep arrived in 1860, because of the establishment of artificial watering points throughout the pastoral country. Consummate kangaroos 331

The conflict with pastoralism arises from the perception that there is direct competition between sheep and kangaroos for the available plants. As already discussed, when plant biomass is above 200 kg/ha, competition between sheep and kangaroos is minimal, but below this level competi- tion increases: since this condition depends on rainfall, it is impossible to predict when such competition will occur; but because pastoralism is a marginal activity, kangaroos need only affect sheep occasionally for them to be perceived as a pest. How marginal is the pastoral indus- try in western New South Wales? In a report prepared in 1983 for the New South Wales government one-tenth of the proper- ties were assessed for 1977–80 (quoted by Caughley et al 1987a). At the time there were 7.2 million sheep, 0.1 million cattle and 8.2 million kangaroos present in the region. The report showed that 46% of the sampled properties generated no disposable income and that finan- cial difficulties were most common on properties with low authorised stocking totals; but even the largest properties were unable to generate surplus income for capital investment and debt servicing. So even if the number of properties was reduced by amalgamation into fewer larger properties, the rangelands would still not contribute significantly to the wealth of New South Wales, because of the marginal nature of the enterprise. Because the large number of kanga- roos has an adverse effect on this enterprise, should the kangaroos be drastically reduced to increase stocking rates, or should sheep be phased out and the kangaroos be used commercially? This question has been raised with increasing frequency in the last decade (see Grigg 1988) but remains intractable for several reasons. The three options for kangaroo management are full protection; damage mitigation; or harvesting as a renewable resource. Protection would lead to higher numbers of kangaroos if enforced but it would be well nigh impossible to enforce because of prevailing attitudes in the country. A possible compensation would be to pay landholders an annual fee for the estimated number of kangaroos on their property. Damage mitigation should be assessed for prevention of future damage, not as a response to past damage, which is merely punitive. However, the unpre- dictability of the system in the rangelands makes it impossible to predict the future impact of kangaroo competition, since that only occurs when rainfall is low. Another problem in damage mitigation is the cost of carrying it out: salaried government shooters would be prohibitive, farmers do not have the time to do it themselves, so commercial shooters are the only feasi- ble option. The third option is resource harvesting, either by farming kangaroos or by harvest- ing wild kangaroos. Farming would require a change to the law, since at present a landholder does not own the wildlife, and there is the problem of kangaroo movement off one property to another, which would need to be resolved. So far, farming kangaroos has not been found to be feasible, because costs exceed income from hides and meat, although the rising prices for kanga- roo meat at restaurants may help to change the equation. That leaves commercial harvesting. Kangaroos have been used commercially since the early days of European occupation but only became important after refrigeration became possible for meat as well as hides. The four large species are harvested in only four States, which have kangaroo management programs and annual quotas set by the federal government. While these are designed primarily for damage mitigation, the kangaroo industry is unable to expand or contract rapidly in response to seasonal changes (which drive the kangaroo population) and so setting quotas cannot be a sensitive agent for damage mitigation. Harvesting, however, would require an independent regulator to prevent over-exploitation of the resource so that kangaroos would be able to recover after drought. Market forces do not encourage sustainable harvesting because the discount rate favours current yield over future yield, as has been abundantly demonstrated with whaling and fisheries around the world. A kangaroo industry dependent on a sustained annual yield would founder in dry years and face a glut in wet years. This is why the issue is still unresolved 15 years after the debate began. 332 Life of Marsupials

The smaller macropods

The Notomacropus wallabies This group of seven large wallabies have the same chromosome number (16) and are closely related on DNA/DNA hybridisation criteria (Kirsch et al 1995). They are distributed around the periphery of the Australian continent but do not occur in the central semi-arid and desert regions, as the large kangaroos do (Fig. 9.12). The agile wallaby occurs in southern New Guinea and the tropics of northern Australia; the black-striped wallaby, Macropus dorsalis, and the whiptail wallaby, Macropus parryi, occur further south in eastern Australia; the parma wallaby, Macropus parma, is a rare occupant of the coastal forests of New South Wales and its distribu- tion is overlapped by the red-necked wallaby, Macropus rufogriseus, which extends into Tasmania as the subspecies known as Bennett’s wallaby. The toolache wallaby, Macropus greyi, formerly occurred in the eastern part of South Australia but was hunted to extinction by the 1920s. The former range of the tammar wallaby included the western part of South Australia and south- ern Western Australia but its present distribution is much reduced through land clearance and fox predation. Finally, the western brush wallaby, Macropus irma, has a similar distribution to the western grey kangaroo in the southwestern corner of the continent and its numbers have declined substantially since the spread of the fox, which probably preys on the young animals. Several of these species were introduced to New Zealand during the 19th century but only parma wallabies, tammars and Bennett’s wallabies persist there today. In the early part of the 20th century Bennett’s wallabies escaped from captivity in Britain and in Germany: they persist in the Yorkshire Dales but in Germany they were all eaten during World War II. The best studied of the seven extant species are the tammar wallaby and Bennett’s wallaby, which are maintained in captive breeding colonies in Australia and in Britain. The tammar will serve as the representa- tive of this group of the Macropodinae.

The tammar wallaby – a highly seasonal breeder with a different strategy The tammar wallaby (Fig. 9.2d, Plate 15) is the best studied species of all the kangaroo family and may be the first marsupial in which the entire genome is sequenced. At the time of European occupation the species occurred in southwestern and southern Australia, separated by the Nullarbor Plain, where it was absent. It also occurred on 10 offshore islands, where it was encoun- tered by some of the early Dutch explorers (see Chapter 1). Starting in the west, these islands are East and West Wallabi Islands in the Houtman’s Abrolhos archipelago and Garden Island off the west coast, Middle and North Twin Peaks in the Recherche Archipelago, St Peters Island in the Nuyts Archipelago, and Flinders, Thistle and Kangaroo Island in Spencer’s Gulf. Apart from Kangaroo Island, which is large and carries western grey kangaroos as well as tammars, all these islands are less than 10 km2 and each carries no macropod other than tammars. These island populations have been isolated from the mainland populations for 7000 to 15 000 years, since the end of the last Ice Age, and each population shows genetic differences from the other. This is particularly valuable for the proposed genome project. Like so many other medium-sized marsupials the mainland and some island populations of tammars have been severely reduced or gone extinct since European occupation. Shortridge (1910) wrote of the mainland population in Western Australia, ‘very plentiful in many parts of the south-west, but rapidly disappearing in the cultivated districts, especially towards the northern end of its range’. The decline continued as the region was cleared for wheat and sheep, especially after 1960, so that now tammars are restricted to a few isolated pockets where foxes are control- led. In the 19th century the tammar wallaby was so common on Eyre Peninsula and around Adelaide that battues were organised to destroy large numbers, so as to protect crops and pastures. Since then it has drastically declined and on the mainland around Adelaide it was extinct by the Consummate kangaroos 333

1920s and on Eyre Peninsula probably by the 1970s; likewise farming activities led to extinction of the populations on Flinders Island and St Peter’s Island. Only on Kangaroo Island is the species abundant and secure, and all the research colonies are derived from this population. The origin of its name is interesting: the specific name eugenii was given to a specimen collected in 1802 or 1803 from Ile Eugene, in the Nuyts Archipelago. Nicolas Baudin, Commander of the Geographe, named the island, presumably after Eugene Hamelin the Commander of his sister ship Naturaliste, in May 1802. However, Matthew Flinders had already named it St Peter’s Island three months earlier and Baudin’s name lapsed. Subsequently the island lost its tammars and the Paris Museum its type specimen, so it is no longer possible to confirm the identity of the tammars on Kangaroo Island, but by common usage they carry the same name. Despite becoming extinct on the mainland of South Australia, the genotype fortuitously survives in New Zealand. When Sir George Grey, formerly the Governor of South Australia, became Governor of New Zealand in 1861, he purchased Kawau Island in Auckland harbour, where he built a grand mansion and stocked the island with many exotic plants and animals, including tammars, black-striped wallabies and parma wallabies. He fell out of favour with Queen Victoria and was dismissed in January 1868 but as a private citizen retained Kawau Island until 1888. It is not clear whether he obtained the wallabies during the period of his Governorship or after his dismissal in 1868, nor is there any record of how, or from where, he obtained the wallabies. The black-striped wallabies died out but the tammar and parma wallabies increased ‘in an almost incredible manner’ on Kawau Island, and in 1912 some tammars were also released near Rotorua and are now established in about 3000 km2 of surrounding forest and farmland, where they are regarded as a pest (Sadleir and Warburton 2001). Since there are no records of where Grey’s animals came from in Australia, it was not clear whether the New Zealand tammars were obtained from Kangaroo Island, or from the now extinct populations on Flinders Island or the mainland. How different genetically are these several populations? While it is relatively straightforward to assess the genetics of the larger populations of living tammars, the relationship of these to the former island and mainland populations is fraught with difficulty because so few accurately recorded specimens were collected when the species was abundant. This is why the identity of the New Zealand tammars is now so important. The tammar on Flinders Island was recognised as having a greyer coat and thinner head than the Kangaroo Island tammar (Jones 1924) and was given a separate name; and the animals on East and West Wallabi Islands are consider- ably smaller than those on Kangaroo Island. Nevertheless, other earlier workers, such as Gould (1863), suspected that all the island populations were part of one species. Poole and colleagues (1991) attempted to resolve the relationships by comparing all the tammar skulls in museums around the world with skulls from the extant populations on the Abrolhos Islands, Garden Island, Middle Island and Kangaroo Island, and from tammars in New Zealand. Using nine measures they found that there were three distinct groups that corresponded with the species’ geographical distribution: the Western Australian mainland tammars and those from East and West Wallabi Islands, Garden Island and Middle Island formed one group; tammars collected on Flinders Island and the South Australian mainland in the 19th century, together with New Zealand tammars from Kawau Island and Rotorua, formed a second group; and tammars from Kangaroo Island comprised the third group. Between the Garden Island and Kangaroo Island populations, at the two ends of the species’ continuum, there are also differences in serum proteins, iron-binding transferrin protein and red cell antigens. But this does not establish whether the several populations are all members of one species; to test this the island races must interbreed. When brought into captivity Kangaroo Island tammars would not mate with any first generation tammars from Garden Island, East and West Wallabi Islands or Middle Island, but descendants that had been habituated in 334 Life of Marsupials

captivity would mate. Matings between Garden Island and Kangaroo Island tammars that had been bred in captivity were fertile and the offspring were also fertile, which indicates that they still belong to one species (McKenzie and Cooper 1997). The success rate of hybrid breeding was about half that of breeding within each geographic race, so there is some physiological barrier to interbreeding, as well as the behavioural constraints, indicating the first step towards the island races becoming separate species. The conclusion that New Zealand tammars are more like tammars collected on Eyre Peninsula in the 19th century than Kangaroo Island tammars (Poole et al 1991) is supported by molecular genetics, using differences in microsatellite DNA (Taylor and Cooper 1999); the Kawau Island tammars lack several genes that are very common in the Kangaroo Island population and, since it is highly unlikely that they could have lost them during their 100 year sojourn on Kawau Island, it is more probable that the tammars came from another place; and the Rotorua tammars are genetically a subset of the Kawau Island population. This conclusion is important for the conservation of tammar genetic diversity and, since the species is regarded as a pest in New Zealand, some Kawau Island tammars have been brought to Australia, for eventual release in their ancestral home near Adelaide. The discovery that New Zealand tammars came from the mainland of South Australia may also help to resolve the questions of how long western and eastern races have been separated and where the species originated.

Did tammars originate in western or eastern Australia? This question can be approached by looking at the regional differences in the species’ response to the metabolic poison sodium fluoroacetate (1080), which occurs naturally in a number of native plants in Western Australia but not in South Australia, where tammars formerly lived. Western grey kangaroos have a very high tolerance of this chemical (lethal dose 20 mg/kg), both in Western Australia and in eastern Australia, which suggests that the species evolved in Western Australia and carried the acquired tolerance eastwards. Tammars living on the mainland of Western Australia can also tolerate the poison but to a lesser extent than western grey kangaroos: the lethal dose for mainland tammars is >5 mg/kg, while for tammars on East and West Wallabi Islands and Garden Island the lethal dose is <2 mg/kg (Oliver et al 1979). Since plant species containing sodium fluoroacetate do not occur on the islands, the tammars must have retained their tolerance since their isolation 11 500 and 7000 years ago, respectively. By contrast, tammars from the eastern part of the range are highly susceptible to 1080 poison; the lethal dose for tammars from Kangaroo Island is <0.2 mg/kg (Oliver et al 1979). New Zealand tammars near Rotorua probably have a similar sensitivity, since 1080 is used to control them, but the minimum lethal dose has not been determined (Warburton 1990). Thus, unlike western grey kangaroos, tammars probably originated in South Australia and subsequently spread to Western Australia, where the tolerance to sodium fluoroacetate evolved; if the movement had been the other way the South Australian races would have retained their tolerance, as the East and West Wallabi Islands and Garden Island tammars have done. Since tammars are known from the Devil’s Lair cave deposit in Western Australia, which is 30 000 years old, the populations of South Australia and those of Western Australia have probably been separated for up to 50 000 years (Oliver et al 1979). The other character that all the races of tammars hold in common is their very unusual pattern of seasonal breeding, which is especially advantageous to a species living in a climate that is highly predictable. Although most studies of tammar breeding have been done on the Kangaroo Island population, New Zealand tammars and mainland Western Australian tammars display the same breeding pattern, so it must have evolved in response to pre-existing conditions on the mainland and not as an adaptation for survival on islands. Nevertheless, it is now an integral part of the tammar’s adaptation to an island environment. Consummate kangaroos 335

Adaptations to island life The largest population of tammars lives on Kangaroo Island, where regular winter rainfall and hot dry summers dominate the climate. Most of the annual rain falls between May and September each year and there is free water in streams at this time. This stimulates good predictable plant growth, which supports the major investment in reproduction by female tammars during late lactation (see Fig. 2.1). By contrast, the conditions in summer are often severe, when the ambient temperature can reach 40°C, although the mean summer temperature is substantially less. Then the streams are dry and the only available water is that in the vegetation, which has also dried off. On several of the other islands on which tammars live there is no fresh water at all and they subsist entirely on water in plants and on seawater. Thus, the main constraints on tammars in all the environments they occupy are the quality of the plant food, the availability of water and high temperatures in summer. Tammars are, however, well adapted to survive these conditions. On Kangaroo Island tammars shelter under the divaricating bushes that cover much of the island, and are most active after dusk and before dawn, foraging for acacia seed in the shrubs and feeding on grassy clearings or farmland pastures. By tracking tammars carrying small radio transmitters, Inns (1980) found that each tammar had a well-defined home range of 42 ha in summer and 16 ha in winter, which overlapped with the home ranges of other tammars, rather like pademelons do. Each tammar occupied a daytime area of scrub, where it could shelter from the sun, and moved to another area of grassland at night. During the summer months the grass on the main feeding area dried off and the tammars were moving over greater distances than in winter, probably searching for better quality food.

Temperature regulation and water turnover Like the desert-adapted kangaroos, tammars maintain their body temperature against high ambient temperatures by panting and by licking their forearms. Above 30°C their respiratory rate increases steeply and evaporative water loss also increases, while above 40°C they are no longer able to control their body temperature and die unless they can avoid the heat (Dawson et al 1969). Water is thus essential for cooling and when that is available, as in captivity, the water turnover rate is highest in summer and lowest in winter (Kaethner and Good 1975). Paradoxically, in free-ranging tammars on Kangaroo Island, the water turnover is the reverse: the highest rate is in midwinter, when the animals are eating forage with a high water content, and is lowest in mid summer when free water is scarce and the forage is dry. They avoid dehydration by passing less water in urine and by reabsorbing water from the distal colon and passing faeces with a water content of about 40%, like southern hairy-nosed wombats do. Nevertheless, the hot summer months are when they lose weight and suffer their greatest mortality of the year: young of the year are the most vulnerable and suffer heavy mortality in hot summers, while the oldest animals succumb during the following cold winter months. Tammars on Garden Island and the Abrolhos Islands face an even greater environmental challenge in summer because there is no fresh water. During hot weather tammars have been observed to go to the beach and drink seawater: in captivity tammars maintained their weight on seawater and low protein diet for up to 30 days (Kinnear et al 1968). They can do this because they have a renal concentrating capacity far greater than other small macropods, such as the quokka and pademelon (Hume and Dunning 1979). Very few mammals have such a great ability to concentrate urine and thrive on seawater. Those that can are smaller than tammars and most are desert-living species, such as the kowari, Dasycercus byrnei, whose renal capacity is associated with water conservation. 336 Life of Marsupials

Nutrition and nitrogen balance The second challenge to tammars is obtaining sufficient protein and energy from the forage, especially in the summer. As mentioned earlier (Fig. 9.9), tammars are able to conserve nitrogen by recycling urea back to the fermentation chambers of the forestomach where bacteria incor- porate it into their own protein and it becomes available to the tammar in the small intestine (Lintern and Barker 1969). By reducing the amount of urea that must be excreted through the kidneys the tammar effects a considerable saving in water and so can pass less urine without becoming dehydrated. In the winter months on Kangaroo Island, when nutritious, moist forage is available, the urea content of the urine is higher than in summer. But in summer the daily maintenance requirement for nitrogen is 230 mg N/kg0.75, which is half the value for the forest- dwelling pademelon and parma wallaby (477 mg N/kg0.75), although not as low as the euro (170 mg N/kg0.75) (Hume 1977, 1999). The combination of a very low maintenance nitrogen requirement and good kidney concen- trating ability enable the tammar to live in island environments with severe water shortages. Without this ability, tammars would be unable to withstand the simultaneous shortages of protein and fresh water that occur regularly every summer in their island habitats (Chilcott et al 1985).

Population dynamics Inns (1980) analysed the results of two 3-year study periods of marked tammars at one site on Kangaroo Island, 1966–68 and 1975–78, and found an annual fluctuation in the population, as the new cohort of young animals entered the population in October and then age-related mortality reduced it through the following year (Fig. 9.16). More than 90% of the adult females carried a pouch young born in January–February and, in most years, there was almost no loss

150 Females

1975 100 1976

1974

1973 1977 Tammar 50 >1972

0 150 Males Population size

1975 100 1976 1974

1973 1977 50 >1972

0 1975 1976 1977 1978

Figure 9.16: The annual pattern of the tammar wallaby, Macropus eugenii, population on Kangaroo Island. After Inns (1980). Consummate kangaroos 337

of young until they left the pouch in early October. Between then and the end of summer the following March, 35% of the young females and 44% of the young males died. Those that survive their first summer continue to grow through their second year and reach full adult size during their third year. Thereafter the annual mortality of both sexes was low, so that the life expect- ancy was about 10 years for males and 13 years for females. Some females that had been marked during 1966–68 were recovered during the second study period when 12 and 13 years old and were still carrying pouch young. In summary, if a tammar survives its first summer, its life expectancy is high; but its chance of surviving the first year is contingent on prevailing good conditions in its first spring and moderate conditions through summer. In each of the two study periods, one summer (1967–68 and 1977–78 respectively) was much drier than the others and a much higher mortality occurred in both these years; in late 1977 half the young were lost from the pouch before October and the young animals just out of the pouch also suffered high mortality over summer, while the old animals also had a high mortality and those that survived were often in poor condition. By contrast, in the more favourable years the young females, just one year old, also gave birth at the same time as their mothers in January or February, but none did so in the dry year of 1978.

How tammars have achieved a highly synchronised breeding season The basic pattern of reproduction in the tammar is the same as in the red kangaroo, but is adapted to a climate where winter rainfall is predictable and nutritious forage follows in the spring. The female tammar makes her major investment in lactation when the young is between 200 and 300 days old (see Fig. 2.1). For instance, protein nitrogen exported in the milk rose from 50 mg/d in early milk to 2500 mg/d in late stage milk. This period falls between late July and the end of November on Kangaroo Island. By giving birth 7 months before this, during the height of summer, the major burden of late lactation and

Figure 9.17: The annual cycle of reproduction of the tammar wallaby, Macropus eugenii, on Kangaroo Island and how it is related to the changing day length through the year. Note that most births occur six weeks after the summer solstice as day length decreases. 338 Life of Marsupials

the emergence of young from the pouch occur when food is abundant and survival is high. This benefit must have imposed heavy selection pressure on the ancestors of tammars to modify an opportunistic breeding strategy, like that of the red kangaroo, to a highly seasonal one. And, indeed, the time of birth is astonishingly constant from year to year and in different places, occurring between 14 January and 14 February (mean 28 January), on Garden Island, Kangaroo Island, Kawau Island, Rotorua and in captive colonies in Canberra and Melbourne. Now that we know that the New Zealand tammars originated on the mainland of South Australia, it is clear that the precise seasonality of the species is not a recent response to the special condition of life on islands but preceded their isolation. Indeed, it must have preceded the spread of the species to Western Australia 50 000 years ago if, as the 1080 data suggests, the species originated in South Australia. So, how is the basic reproductive pattern different? Like the red kangaroo, the tammar conceives a few hours after giving birth and suckles her young in the pouch for the next seven months. During the first five months of her lactation, when the days are shortening, the dormant embryo in her uterus will only develop and be born if the young in the pouch is lost or removed, just like in a female red kangaroo (see Chapter 2). However, after the winter solstice on 21 June the embryo will not develop if the pouch young is removed, nor does it develop when the older young leaves the pouch in October, as it would in a red kangaroo: instead it continues to remain dormant in the uterus until after the summer solstice on 21 December (Fig. 9.17).

40 Louisiana 30° N 30

20

10 Tammar 0

30 Longest Canberra 35° S day 20

10 Percentage of births Percentage 0

50 Longest Kangaroo Is. 36° S day 40

30

20

10

0 DJFMAMJJASOND Longest day

Figure 9.18: Reversal of the onset of breeding by tammar wallabies, Macropus eugenii, when taken from Australia to Louisiana in North America. After Berger (1970). Consummate kangaroos 339

Young females attain puberty when they leave their mother’s pouch in October and, in good years, experience their first oestrus and conceive then but, like their mothers, their embryos become dormant and they also do not give birth until after the summer solstice. Since the length of gestation in the tammar is about 26 days this means that the blastocysts carried by adult and young females resume development about two weeks after the summer solstice, an extraordinar- ily rapid, synchronised response to an external event. How is this rapid response achieved? It could be either the result of an intrinsic annual cycle within each animal – a circannual rhythm – or each tammar responds to a regular event in the environment that acts as a signal to initiate reproduction at a particular time of the year. There is some evidence for a circannual rhythm in female tammars and there is also strong evidence that they are sensitive to the changing length of the day and night after the summer solstice. Pat Berger, who discovered this phenomenon in tammars on Kangaroo Island (Berger 1966), dramatically demonstrated this when she took 25 female tammars to Louisiana, USA, in late August and they gave birth two months later in November. In subsequent years her colony continued to breed six months out of phase with Kangaroo Island, six weeks after the northern summer solstice (Fig. 9.18). Other seasonally breeding mammals that have been taken across the Equator, such as Merino sheep to Australia, also reverse the time of birth by six months in accordance with the

Tammar 16

15

14 Group A (control) 13

12

15

14

Hours of light per day Group B (solstitial) 13

12

11 Sep Oct Nov Dec Jan Feb Mar equinox summer equinox solstice

Figure 9.19: Response of female tammar wallabies, Macropus eugenii, to an experimental photoperiod imposed at the vernal equinox, in which the day length was first increased to 15 h for 6 weeks and then shortened to 12 h, compared to a control group held on a changing day length that mimicked the normal annual pattern: all the tammars in the experimental group gave birth (?) about 32 days after the change to 12 h, whereas the control group, which experienced the normal photoperiod changes, gave birth three months later, after the artificial summer solstice. After Hinds and den Ottolander (1983). 340 Life of Marsupials

prevailing pattern of photoperiod, but the actual hormonal control varies profoundly between species. How female tammars translate the environmental factors into hormonal responses that control the ovary and the embryo have been the subject of much study, so that the control of breeding is better known in the tammar than in any other marsupial, and better known than for all but a few species of placental mammal.

Components of the photoperiod signal The first experiments designed to understand the photoperiod signal that tammars respond to was to place them in a light-controlled room at the spring equinox (12 h light: 12 h dark) and expose them to mid-summer day length of 15 h light to 9 h dark. This did not induce them to give birth but, after they were returned to outside pens and experienced a shorter day length, all gave birth 30–32 days later (Fig. 9.19) (Sadleir and Tyndale-Biscoe 1977, Hinds and den Ottolander 1983). This showed that it is not long day in itself that is important but the change in day length from long to short that is stimulatory. However, when other tammars were held on summer day length from the equinox until after the summer solstice (ie for four months), they gave birth at the normal time in January–February, which suggests that a circannual rhythm may also be involved in the precise timing of birth on Kangaroo Island. From many subsequent experiments some general conclusions can be made about the response of female tammars to photoperiod. Conditions of winter day length (9 h light: 15 h dark) and continuously increasing day length are inhibitory, and the tammars remain quies- cent indefinitely, whereas decreasing day length and equinoctial day length is permissive, and the tammars behave like red kangaroos, breeding continuously. When Pat Berger took tammars across the Equator in August they responded to the change from increasing to decreasing day

Light change 15L:9D to 12L:12D

First signal Melatonin to blastocyst message read Birth Progesterone Prolactin peak occurs peak absent

Days

Melatonin Prolactin Corpus luteum profile remains irreversibly changes absent committed for 72 h

Figure 9.20: Summary of the cascade of hormone responses in female tammar wallabies, Macropus eugenii, that follow an experimental change from an inhibitory photoperiod of 15 h light: 9 h dark to a stimulatory photoperiod of 12 h light: 12 h dark. Consummate kangaroos 341

length. With this understanding of the tammar’s response it is now possible to manipulate them to breed or not to breed at any time of the year. This has enabled the train of events from reading the message in the day length to the birth of the young to be worked out quite precisely (for a summary, see Tyndale-Biscoe and Renfree 1987).

The hormonal cascade that follows the photoperiod signal Three endocrine organs are involved in the response of the tammar to a stimulatory photope- riod: the pineal gland on the roof of the mid-brain, which secretes the hormone melatonin; the pituitary gland beneath the mid-brain, which secretes the hormone prolactin; and the corpus luteum on the ovary, which secretes progesterone.

The progesterone profile The change in progesterone levels, where the early pulse of progesterone from the corpus luteum was shown to be a good indication of the resumption of development by the blastocyst, leading to birth 22 days later, was described in Chapter 2. When reactivation is induced by a change in photoperiod, timing of these events is subtly different: the progesterone pulse occurs on day 8, 9 or 10 rather than on day 5, 6 or 7, but birth still occurs 22 days after the pulse on day 30–32 (Fig. 9.20). Thus, embryo reactivation is delayed for an additional three days but the subsequent events occur in the same order. What delays it?

The melatonin profile The pineal gland secretes the hormone melatonin only during hours of darkness, the information about light reaching it by a circuitous route from the eyes to the brain and by sympathetic nerves that reach the cells of the pineal. The first hypothesis to account for the three-day lag was that it would take the pineal three days to change its secretion of melatonin to the new light regime. However, the profile of plasma melatonin alters on the first succeeding night after a photoperiod change and thereafter continues to correspond to the lengthened period of darkness. The essential components of the melatonin message for tammars are: that its duration is more than nine hours; that the duration of elevated melatonin increases by more than one hour (eg a change from 9 to 10 h is stimulatory); and that the change is experienced by the tammar for more than three successive nights. If these conditions are met, the corpus luteum reactivates, the progesterone pulse occurs on day 8–10 and the young is born on day 30–32. What is more remarkable is that it is not necessary for melatonin to be elevated for the whole of the 10 h; a brief pulse at the beginning and the end of the period will suffice. Thus, melatonin is not acting like a conventional hormone but more like a neural signal that is turned on and off to define a specific period of time. But this is not the whole story; the pituitary gland is also intimately involved.

The prolactin profile When the pituitary gland is surgically removed during the second half of the year, the corpus luteum reactivates, the progesterone pulse occurs seven days later and the blastocyst reactivates. This means that a hormone secreted by the pituitary gland continuously inhibits the corpus luteum in the second half of the year: this is prolactin, which is also involved in lactation (see Chapter 2). When tammars experience an inhibitory photoperiod (15 h light: 9 h dark) they secrete a single pulse of prolactin from the pituitary gland each morning at about the time that the lights come on, the equivalent of dawn. When the photoperiod is changed or is mimicked by a change in the duration of the melatonin profile, the prolactin pulse continues for the next three or four mornings and then ceases. Thus, the time it takes to lose the morning pulse of prolactin is the same as the time it takes to respond to the change in the melatonin profile. Is the prolactin pulse the effective agent of inhibition and does its disappearance allow the corpus luteum to 342 Life of Marsupials

reactivate? To test this, tammars were placed on an inhibitory photoperiod (15 h light: 9 h dark) for several days and then changed to a stimulatory photoperiod (12 h light: 12 h dark), which should provoke reactivation. Half were given a morning injection of prolactin to mimic the dawn pulse and half an injection of saline as a control; the controls reactivated normally and the prolactin treated tammars were delayed until the end of the injections. Thus, the morning pulse of prolactin that occurs while tammars experience a long day or an increasing day length is the essential element in maintaining seasonal quiescence, and it explains why the corpus luteum grows and the blastocyst reactivates in tammars that have the pituitary removed. Putting the three profiles together, the change in photoperiod from 15 h light: 9 h dark to 12 h light: 12 h dark causes an immediate change in the melatonin profile and this information is presumably stored and compared to the experience of the next two succeeding nights (Fig. 9.20). If the stimulus continues for three successive nights, the tammar makes an irreversible response, which is seen when the prolactin pulse is abolished on the 4th morning. Then, if there is no dawn pulse of prolactin for the next 72 h the corpus luteum reactivates irreversibly and, within one day of reactivating, it provides the signal that releases the blastocyst from diapause. One day later the pulse of progesterone occurs and stimulates the endometrium for the maintenance of the developing embryo for the remaining 22 days of pregnancy. This sequence of irreversible events shows that the hormonal control of seasonal quies- cence in female tammars is a subtle extension of the control that operates during lactation in other kangaroos and wallabies. In the first half of the year the prolactin pulse is induced by the sucking stimulus of the pouch young on the teat, via a neural pathway to the pituitary gland, whereas after the winter solstice the daily pulse of prolactin is maintained by the duration of the melatonin signal, itself determined by the length of the night. Thus, the transformation of the tammar, from being an opportunistic breeder like the red kangaroo, to a highly seasonal breeder was a change in either the secretion of prolactin by the pituitary, or the sensitivity of the cells of the corpus luteum to a daily pulse of prolactin. This genetic change must have occurred in the species in its site of origin in South Australia and was carried as part of its genetic heritage to all the other parts of its present range, including New Zealand.

Male tammars track the females If 90% of the females give birth within less than one month and all of them undergo oestrus a few hours later, the males must be synchronised for the challenging task of fertilising them. Unless the young is lost from the pouch in the first half of the year, and this is very uncommon, most females will not return to oestrus until a year hence. In some seasonal breeders the males are equally sensitive to photoperiod changes as the females but this is not so for male tammars. When males were experimentally isolated from females for several months before the summer solstice their level of testosterone and pituitary hormones in circulation was very low, whereas in another group of males that were penned with females during the same time the pituitary hormones and testosterone steadily rose to peak levels one month after the summer solstice (Catling and Sutherland 1980). This coincided with the rising levels of progesterone in the females that had responded to the photoperiod signals and would give birth at the end of January (Hinds and den Ottolander 1983) (Fig. 9.21). Testosterone is essential to stimulate the accessory reproductive organs, especially the prostate and Cowper’s glands that produce the seminal fluid, and on Kangaroo Island these organs were significantly enlarged in male tammars at this time (see Fig. 2.5) (Inns 1982). Thus, the male tammar is secondarily seasonal, through association with females in late pregnancy, presumably stimulated by a pheromone that is secreted under the influence of progesterone. This is further evidence that the seasonality of the female tammar has evolved relatively recently by a subtle modification of the opportunistic breeding pattern of other kangaroos and wallabies, and the males respond to the changing condition of the females. Consummate kangaroos 343

Tammar

15

300

10

200

Females 5 Testosterone (ng/ml) Males with females 100 Progesterone (pg/ml)

Males without females 0 0 Sept Oct Nov Dec Jan Feb Summer Main period of births solstice and post-partum mating

Figure 9.21: Male tammar wallabies, Macropus eugenii, do not respond to the change in day length but track the response in females: after the summer solstice only males associating with females, whose levels of progesterone increase until they give birth, show elevated testosterone, whereas males isolated from females at the same time do not. After Catling and Sutherland (1980) and Hinds and den Ottolander (1983).

Bennett’s wallaby, the only other seasonally quiescent breeder The only other wallaby that displays seasonal quiescence like the tammar is Bennett’s wallaby in Tasmania. It is the island subspecies of the red-necked wallaby, Macropus rufogriseus rufogriseus (Fig. 9.2e, Plate 15), that occurs throughout the eastern Australian seaboard. On the mainland of Australia the species is not a strictly seasonal breeder and behaves more like a red kangaroo, a euro or a rock wallaby (Merchant and Calaby 1981). However, the subspecies on Tasmania has a pattern of reproduction that is very similar to that of tammars. Most births of Bennett’s wallaby in Tasmania occur early in the year and after the winter solstice the females undergo seasonal quiescence. Although the females are sensitive to changing photoperiod in the same way as tammars, and melatonin is involved in transducing the signal, the role of prolactin is uncertain: the transient morning pulse of prolactin that is such an impor- tant component of the tammar’s response has not been detected in Bennett’s wallabies. It is not surprising that there are differences between tammars and Bennett’s wallabies, because seasonal quiescence has almost certainly evolved independently in the two species: it probably evolved in tammars 50 000 years ago, whereas Bennett’s wallabies have only been isolated from the mainland of Australia and their ancestral stock for about 12 000 years. Because red-necked wallabies and Bennett’s wallabies readily interbreed, it would be possible, with modern techniques, to identify the genetic basis of seasonal quiescence in this species.

Pademelons, tree kangaroos and rock wallabies Pademelons are small macropods that live in dense forested habitats adjacent to open glades. During the daytime they live in the deeper recesses of the forest browsing on shrubs and forbs. 344 Life of Marsupials

At dusk they move directly to another feeding area some distance away in an open glade where they feed on grasses. Both parts of the home range are shared with other pademelons, feeding together (Johnson 1980, Vernes et al 1995). With the clearing of the forests after European occupation of Australia, the grazing areas for pademelons increased and some species readily adapted to feeding on introduced pastures. From the early days of settlement they were regarded as pests and large numbers were destroyed (Lunney and Leary 1988). Because they live in a benign environment they do not display the features of water and nitrogen conservation seen in island wallabies and the desert-adapted kangaroos, and their breeding is non seasonal. Males are larger than females and their breeding strategy is to compete for access to females, like the larger species do. The six species of Thylogale are distributed from New Guinea in the north, along the eastern seaboard of the Australian continent to Tasmania at 43°S (Table 9.3). Pademelons have the basic macropod complement of 22 chromosomes, from which the chromosome patterns of other kangaroos and wallabies can be derived by the technique of chromosome painting (see Classifying wallabies and kangaroos by their chromosomes and Fig. 1.8 Plate 1). Also, when compared by DNA/DNA hybridisation techniques (Campeau-Peloquin et al 2001) the pademelons appear to be the ancestral group from which arose both the tree kanga- roos of New Guinea, Dendrolagus, and the rock wallabies, Petrogale, of continental Australia, which share similar adaptations for climbing in the structure of their feet and tails. Tree kanga- roos remained within the ancient forests but became secondarily specialised for a life in the trees (Fig. 9.2b, Plate 14), with shortening of the hind legs, a wider foot and short nail on digit 4, but they were unable to re-evolve a prehensile, which their possum ancestors, presumably, had. At the same time rock wallabies moved out of the forests and occupied rocky outcrops for shelter and fed in the surrounding woodlands and grasslands (Fig. 9.2c, Plate 14). The evolution of the two groups must have occurred independently about 3.5 million years ago when New Guinea was separated from Australia, since rock wallabies are absent from New Guinea, despite suitable rocky habitat in the highlands. However, two species of tree kangaroo do occur in northern Queensland rainforests, having presumably migrated south during the last land connection with New Guinea, in the same way as did the southern common cuscus, Phalanger intercastellanus, and the common spotted cuscus, Spilocuscus maculatus – and the spiny bandicoot, Echymipera rufescens.

Distribution and relationships of rock wallabies As their name implies, rock wallabies are restricted to rocky places, preferring steep slopes, deep gorges and boulder piles, and they feed on the plants that grow nearby. Their feet have unusu- ally large pads and short nails, which enable them to land on smooth rock without slipping, and their tails are slender, long and curved with a tufted end, which provide balance while leaping from one rock to another; they can also climb sloping tree trunks. Unlike most other macropods, some rock wallabies have beautifully coloured coats, but whether this is associated with their strong territorial behaviour remains unclear. Because of these several adaptations they are a highly distinct group of the Macropodinae, which are also closely related to each other. While the heartland of the tree kangaroos is New Guinea, the heartland of rock wallabies is continental Australia: none occurs on New Guinea, the Bass Strait islands and Tasmania, or on Kangaroo Island, but many small islands close to the continent carry populations of species that live on the nearby mainland (Fig. 9.22). These islands became separated during the past 9000 to 15 000 years, as the sea level rose, so the rock wallabies have remained isolated since then, which raises interesting questions about the long-term genetic viability of very small populations. For instance, on Barrow Island, 55 km off the Western Australian coast, there is a small population of about 150 black-footed rock wallabies, Petrogale lateralis, that have been isolated for 8500 years Consummate kangaroos 345

from the nearby mainland population. The genetic diversity of the island animals is only one- tenth that of the wallabies on the adjacent mainland, indicating that they are highly inbred, and yet they have clearly survived for a very long time (Eldridge et al 1999).

brachyotis

concinna coenensis

godmani burbidgei mareeba

sharmani assimilis

Barrow Is. persephone inornata rothschildi purpureicollis herberti

lateralis xanthopus

lateralis

penicillata

(to Hawaii, New Zealand)

Figure 9.22: The distribution of all 16 species of rock wallabies, genus Petrogale, in continental Australia and some offshore islands. The shading indicates the four main groups: pale shading the xanthopus group and the route they are thought to have taken in the initial spread from the east coast to the west coast; the second lateralis group spread through Western Australia and back across the continent, giving branches to the brachyotis group in the north and the most recently evolved penicillata group (darkest shading) along the eastern seaboard from Cape York to the southeast. After Eldridge and Close (1992, 1997) and Eldridge et al (1999).

Less extreme isolation also applies to many continental populations because of the fragmented nature of their preferred habitat, which is often separated by large areas of unsuitable country. Before European occupation rock wallabies could move between isolated sites but after the clearing of woodlands and the spread of the fox, this has become increasingly difficult, so that small populations from 100 animals to as few as five are severely isolated in their rocky refuges. The social behaviour of rock wallabies is also very different from pademelons, wallabies and kangaroos, being more territorial and monogamous, probably driven by strong competition for favourable sites on the rocky outcrops where they live. For 150 years the taxonomic relationships of rock wallabies, collected from various places around Australia, was very confused. In some regions the species looked distinct but shared almost identical chromosomes, while along the North Queensland coast eight species look the same and can only be distinguished by their very different chromosomes. With the develop- ment of the new techniques of G-banding and chromosome painting (Eldridge and Close 1997, 346 Life of Marsupials

O’Neill et al 1999) and molecular criteria, such as nuclear and mitochondrial DNA sequence data (Loupis and Eldridge 2001) and DNA/DNA hybridisation (Kirsch et al 1995, 1997, Campeau- Peloquin et al 2001), the relationships of all the rock wallabies are becoming clearer. At present, 16 species and 21 distinct chromosomal races are recognised and these are distributed among four main species complexes (Fig. 9.23). Furthermore, the time when each species complex became a separate entity can now be estimated and so we can begin to understand how rock wallabies spread across Australia. It is an extraordinary story, which begins with the four species complexes.

P. assimilis P. godmani P. inornata 0.8 0.7 P. herberti P. penicillata penicillata 0.5 1.2 P. sharmani P. mareeba P. coenensis 1.7 P. purpureicollis P. lateralis 3.4 lateralis P. rothschildi P. persephone 3.7 P. xanthopus xanthopus P. brachyotis 7.5 2.2 P. burbidgei P. concinna brachyotis Dendrolagus Figure 9.23: Relationships between rock wallabies, genus Petrogale, based on DNA/DNA hybridisation and chromosomes. Numbers indicate estimated time of branch in million years before present. After Eldridge and Close (1997) and Campeau-Peloquin et al (2001).

The Petrogale xanthopus species complex Across the continent the yellow-footed rock wallabies – the xanthopus group – comprises the Western Australian species, Petrogale rothschildi, the South Australian and central Queensland species, Petrogale xanthopus, and the relict population of the Proserpine rock wallaby on the Queensland coast and Whitsunday Island, Petrogale persephone. These three species have chromosomes that are similar to those of pademelons (see Fig. 1.8, Plate 1). Also, of all the rock wallabies, these are most similar by DNA/DNA hybridisation to pademelons and tree kangaroos (Campeau-Peloquin et al 2001) and are therefore thought to be nearest to the ancestral stock from which all other rock wallabies evolved about 3.7 million years ago. Yellow footed rock wallabies are the largest rock wallabies (Table 9.3) and are distinguished by the bright orange colouration of the feet and transversely striped tail (Fig. 9.2c, Plate 14). In the early 20th century they were severely hunted for their attractive pelts. Consummate kangaroos 347

The Petrogale brachyotis species complex North-western Australia, from Broome to Arnhem land, is occupied by the widespread short- eared rock wallaby, Petrogale brachyotis, and two much smaller species, the nabarlek and the tiny , Petrogale burbidgei, whose very restricted ranges are within that of the larger species. The nabarlek is unique among wallabies in growing supernumerary molar teeth through life, a feature that formerly was used to separate it from all other rock wallabies. However, chromo- some and DNA characters clearly relate it to the short-eared rock wallaby (Fig. 9.23). All three species have isolated populations on small islands near the coast of northern Australia (Fig. 9.22). This group diverged from the basal stock 3.7 million years ago and from each other about 2.2 million years ago, when they exclusively occupied northern Australia.

The Petrogale lateralis species complex The black-footed rock wallabies formerly extended across the southern region of Western Australia and South Australia, the MacDonnell Ranges in Central Australia, and isolated populations in the Flinders Ranges and central Queensland, the latter population now being recognised as a separate species, the brilliantly coloured purple-necked rock wallaby, Petrogale purpureicollis. The continental range has drastically contracted since the land clearing of the wheat belt of Western Australia, so that there are now only a few tiny colonies on isolated rocky torrs, surrounded by crops and sheep pastures. Small populations also occur on seven small islands off the coast of western and southern Australia, which show chromosome differences from the mainland stock (Eldridge and Close 1997, Eldridge et al 1999). Based on DNA/DNA hybridisation criteria the lateralis complex separated from the ancestral stock about 3.4 million years ago.

The Petrogale penicillata species complex The Petrogale penicillata species complex consists of eight species, extending from southern Victoria and along the Great Dividing Range of eastern Australian to far north Queensland. They are so alike in their form, coat colour and shape of skull that it is impossible to distinguish them on these conventional criteria. However, they all have chromosomal differences, such as fused chromosomes, additional bits of DNA or inversions within a chromosome, which separate them one from the other (Eldridge and Close 1992, 1997). Eastward flowing rivers appear to be the barriers that separate each species in this series from the next most northerly species, but all are very closely related to each other and at least two pairs form subfertile hybrids at their respective boundaries (Fig. 9.22). However, male hybrids between all of them are sterile so each species is maintaining a separate genetic identity and, therefore, should be regarded as a genuine species. Another feature that indicates their very recent separation is that at the boundary between adjacent species some populations carry chromosome anomalies of the next-door species. These are called chromosome introgressions. They are thought to arise when a rock wallaby of one species crosses the border and interbreeds with a wallaby of the neighbouring species: if the mating is fertile, it bequeaths part of its chromosomal signature to its descendants in the other species. While these introgressions only occur near the boundaries between adjacent species, they have been found between all adjacent species along the Queensland coast from Petrogale coenensis in the far north to Petrogale penicillata in the south, which suggests that speciation in this complex is very recent and still active. This conclusion is supported by molecular data (nuclear DNA markers), which could not differentiate the five most northern species from each other because they have separated so recently (Loupis and Eldridge 2001). On DNA/DNA hybridisation data this species complex is estimated to have diverged from the eastern branch of the lateralis species complex 1.2 million years ago and from each other only between 0.8 and 0.5 million years ago. 348 Life of Marsupials

It is a curious anomaly that the eastern seaboard of Australia is occupied by only two species of pademelon but by eight species of rock wallaby. One idea that has recently been proposed (O’Neill et al 2001) is that the rapid speciation within the penicillata species complex has been caused by a viral infection of the genome, which induces fragmentation and recombination of the chromosomes, leading to infertility between adjacent populations and the formation of new species.

How rock wallabies spread across Australia With the combined understanding of their chromosome patterns and the relationships and times of divergence that DNA/DNA hybridisation and DNA sequence data discloses, the timeta- ble of rock wallaby evolution is becoming clearer (Kirsch et al 1997, Eldridge and Close 1997, Campeau-Peloquin et al 2001). The divergence from pademelons occurred about 4 million years ago, probably in the forests of eastern Australia, at the same time as tree kangaroos arose in New Guinea. The ancestral stock of rock wallabies arose in the forests of the eastern seaboard and spread westwards across the continent, as far as Western Australia, where they occupied the rocky outcrops on the ancient pre- shield: their direct descendants are the xanthopus species complex; and the relict population of Petrogale persephone, near Proserpine, Queensland, may represent the original site of origin in eastern Australia (Fig. 9.22). Two offshoots of this ancient stock arose: in the north about 2.2 million years ago the brachyotis species complex diverged, later to become the three species that occupy that country today. In the southwest the lateralis species complex arose 3.4 million years ago and spread back eastwards across the southern and central regions of the continent, as far as western Queensland. The penicillata species complex of the eastern seaboard arose from the eastern branch of the lateralis complex 1.7 million years ago, and is still in the process of active speciation along the eastern seaboard (Fig. 9.22). As may be imagined, unravelling the relationships of this highly complex group of species across Australia has been fraught with difficulty and, while the main relationships are now understood, it is likely that there are still other species, yet to be discovered, in isolated rock fastnesses of the continent. The bewilderingly large number of species of rock wallaby and the great diversity of chromosome patterns that they display is in marked contrast to the uniformity of the four common species of large kangaroos with their continent wide distributions. Is the difference a consequence of the peculiar ecology and social behaviour of rock wallabies? This is best appreciated by looking more closely at the constraints of their unusual life style and how one species is adapted to it.

The allied rock wallabies of Black Rock: a case study The most comprehensive study of rock wallabies has been conducted in northern Queensland on a small isolated population of the allied rock wallaby, Petrogale assimilis, one of the penicillata species complex. Black Rock, 250 km northwest of Townsville, is an isolated sandstone outcrop of 32 ha, which on its eastern face rises to 20 m with fallen rock piles that provide daytime shelter for most of the rock wallabies. It is surrounded by mixed woodland and more open country, which cattle graze (Horsup and Marsh 1992). The next nearest rock wallabies are a very small population on a rock outcrop 8 km distant and a larger population 41 km away. This area of Queensland has a highly seasonal climate of hot, wet summers and warm, dry winters. Three-quarters of the annual rainfall (730 mm) falls in the four months December– March. The wallabies’ preferred food in all seasons is annual forbs and, unlike other rock wallabies, they actively avoid grasses. During the eight-month annual dry period they eat less Consummate kangaroos 349

nutritious plants, browsing on shrubs and fallen leaves, fruits from trees and, in extreme circum- stances eating cattle dung. During the wet season they feed only at night in the woodlands, and shelter in caves and rock shelters during the day, but in the dry season, as forbs become scarce, they forage beyond the rocks during the day, risking exposure to their main predator, the wedge- tailed eagle, Aquila audax. The critical time of the year for the Black Rock wallabies is therefore the end of the dry season in late winter; if there is a food shortage at this time, pouch young die and mortality of adults is high. The rock wallabies are opportunistic feeders that rely on thinly scattered food resources for much of the year, so that group feeding, as in pademelons, is not an appropriate strategy (see Jarman and Coulson 1989) and there is strong competition for both feeding areas and daytime refuge sites, which are shared only by mated pairs. By following the nocturnal movements of radio-collared rock wallabies each month for nearly two years, Alan Horsup (1994) determined the home ranges and core areas of 26 rock wallabies at Black Rock. The mean home range was 12 ha with smaller core areas of about 4 ha, but the shape was generally elongated at right angles to the line of the rock escarpment, with the narrow neck straddling the rock; this was more pronounced in dry seasons, as the female walla- bies ranged further from their rock refuges (Fig. 9.24).

Figure 9.24: Home range and core areas of a female allied rock wallaby, Petrogale assimilis, on Black Rock, Queensland, during two wet and two dry seasons. (O) the main den site in the escarpment rocks, ( . ) the position of the female at different times made by radio fixes. From Horsup (1994).

Fidelity to the home range was very strong, probably because knowledge of food sources, shelter sites and escape routes are crucial for survival, especially during the dry season. Unlike pademelons and the larger wallaby species, male and female rock wallabies are about the same size and at Black Rock they form strong bonds between a mated pair, which share a common feeding range, bask together on their rock refuge and regularly groom each other. The pairing may last for several years or until one partner dies. The male partners marked their feeding areas at night with urine and both sexes were observed to sniff bare rock where food was not present, presumably testing for intruders. Movements from the rock shelter to the pair’s feeding areas was very regular, with departures within 30 minutes of sunset and return within one hour of sunrise. However, they would not leave the rock pile during rainy nights. Maintenance of a strong pair bond is costly in terms of sharing limited food resources but advantageous in sharing shelter sites, which is the limiting resource, especially in the lengthy dry period when conserving water is important. At this time the home range of the female increases but not that of her male partner. Both males and females have hierarchical status but females are especially aggressive towards other females attempting to share shelter or food resources; this generally precludes a male from 350 Life of Marsupials

pairing with more than one female unless he occupies the home ranges of both, which occasion- ally happens, or the second female is the daughter of the female partner. The reason that males are not larger than females may be because they do not compete for access to many females, as in pademelons, large wallabies and kangaroos. Young unpaired males have larger home ranges and they attempt to consort with several females and are rebuffed by other males. When they eventually displace a resident male and establish a consort relationship with a resident female their home range diminishes in size. The question often asked in such apparently monogamous relationships is how many of the young are fathered by the male partner. Peter Spencer studied this at Black Rock over eight years by analysing the microsatellite DNA of 131 adults and 124 of their offspring (Spencer et al 1998). Of these young, the consort male had fathered two-thirds and another male the remaining one- third. More interestingly, in 11 females, where the consort fathered all the young, the probability of raising the young to weaning was 0.45, significantly higher than in the five where none was fathered by the consort male (0.25). But the highest reproductive success (0.58) was among the five females whose young had mixed parentage: females that reared their young to independ- ence mated with their consort, whereas females that lost a small pouch young had an extra-pair mating the next time. This may be a strategy for increasing the chance of reproductive success without disturbing the long-term partnership. It also points to a possible mechanism that may operate when the population has been reduced by sustained drought and serial loss of young: mating with a male from another colony may be more likely then and would facilitate outbreed- ing in the small population at Black Rock. While these studies were being done, Delaney (1997a,b) followed the reproductive activity of the population and its changing size, which varied according to the amount of rain each year. After several good seasons and high recruitment in the early 1980s, the population was at its peak in 1986, when there were about 60 adult animals but no young. One female caught in 1986 had been marked 10 years before and was at least 12 years old, and most of the population in that year were more than five years old; there had been no recruitment in the previous two years (Delaney and Marsh 1995). Four years later, after several more dry years, the total population was down to 26 adults. The next nearest colony of allied rock wallabies is 8 km away and there was no evidence of marked animals from Black Rock moving there; beyond that colony the next one is 41 km away, so the decline in the population must have been due to mortality of the adults, not emigration. At the same time there was almost no recruitment because 88% of the pouch young and young at heel died. After the start of high rainfall in April 1991 the population recovered quickly, so that two years later the adult population was back to 42, half of which were breeding for the first time, and there were many immature wallabies (Spencer 1996). Thus, the breeding pattern in this small population on Black Rock is of very infrequent pulses of heavy recruitment, followed by years when few if any young survive to breed and the adult population slowly decreases. How is this achieved? The reproductive strategy of the allied rock wallaby is similar to that of the red kangaroo and euro: most females produce a succession of young through the year, regardless of the season. Despite the low rainfall during 1986–89, 90% of the females carried pouch young throughout Delaney’s study but less than 10% also had a young at foot, indicating heavy mortality in the pouch; moreover, those that were weaned disappeared before reaching sexual maturity at two years. The critical factor determining whether or not a young will continue to thrive in the pouch is rainfall and the plant biomass that follows, as translated into milk. Milk was sampled at intervals through lactation at Black Rock and total milk production estimated by the sodium isotope turnover technique (see Box 1.1) (Merchant et al 1996). The amount of milk produced was less than in the tammar (see Fig. 2.19) and the milk composition changed earlier, reflecting the faster growth rate and earlier weaning time of the rock wallabies. The characteristic fall in complex sugars and increase in proteins and lipids occurred at about 150 days when the young Consummate kangaroos 351

one began to relinquish the teat. From Delaney’s observations most pouch young die before this stage, which is when the energetic cost to the mother increases steeply. This pattern of mortality is similar to the red kangaroo, although the conditions are different: survival of the pouch young depends on a sufficient food supply being available for the mother to sustain late lactation and at Black Rock those conditions occur rarely. When these conditions do occur, as in the early 1980s and in 1991, there is substantial recruitment to the population, which sustains it for several years. During the years of low recruitment the population survives because of the longevity of the adults, which may live for more than 12 years and breed for at least 10.

Is the Black Rock population very inbred? The evidence from the integrated studies at Black Rock are of a self-sustaining, small population of rock wallabies with episodic recruitment every few years from the surviving breeding stock. Such a small population might be expected to be very inbred and have little genetic diversity. However, the genetic diversity of an isolated population can be considerably increased by the introduction of even a few animals from another colony of the species contributing new alleles to the gene pool. Genetic diversity is estimated by measuring the number of alleles at each of a number of gene loci along the chromosomes. In an outbred population there will be a large range of alleles at each locus and the population is said to have a high degree of genetic diversity or heterozygosity. Conversely, in a small population, which has lost many alleles through cross- ing between closely related individuals, the heterozygosity is low. Contrary to expectation, the genetic diversity at Black rock during 1987 to 1993 was high (3.8%), and it had 94% similarity with the small population 8 km distant, but only 43% genetic similarity with the next nearest colony 41 km away (Spencer et al 1997). While no rock wallaby was seen to range more than 1 km from Black Rock during the study, there clearly is genetic mixing with the colony 8 km away but not with the more distant one. Spencer estimated that the genetic similarity of the two colonies could be maintained if only one animal moved between them every five years. The larger home ranges of immature or young males may be the means whereby this occurs even though movement of marked animals was not detected. The social behaviour on Black Rock probably also helps to maintain genetic diversity through the mating system, in which most males contribute to the gene pool, and by the number of animals that are available to replace established animals.

Comparison between the Black rock population and other rock wallaby populations The pattern seen in the population at Black Rock seems to be common among populations of the other species complexes of rock wallabies. All that have been studied have a non-seasonal pattern of breeding, with successful reproduction keyed to rainfall; that is, rock wallabies are oppor- tunistic breeders, like the desert kangaroos. However, not all species show the strict monogamy observed at Black Rock, where mated pairs share a common home range that they defend from other wallabies; brush-tailed rock wallabies, Petrogale penicillata, do (Jarman and Bayne 1997), but yellow-footed rock wallabies, Petrogale xanthopus are polygynous, with a dominant male holding a territory with a group of several females (Sharp 1997). The size of the home range also differs within and between species. While brush-tailed rock wallabies in northern New South Wales have home ranges similar in size to the allied rock wallaby on Black Rock, the home range of this species on Motutapu Island, New Zealand, where a few animals were released in 1873, was one-third the size. By contrast, the home ranges of the larger yellow-footed rock wallaby in the harsh environment of the Flinders Range, South Australia, was 10 times larger. These marked differences probably reflect the availability of plant food near the rock refuge: where this is abundant the home range can be small, but where this is scarce the home range must be large. 352 Life of Marsupials

The use of rock refuges also differs between species; the smaller species can penetrate deep into the fissures in the rocks and thereby escape predators such as foxes, but the larger species are apparently reluctant to enter caves that do not have another exit by which they can escape a fox or dog. The larger yellow-footed rock wallabies tend instead to seek the shade of rock cliffs, as do euros. Sharp (1997) studied yellow-footed rock wallabies in western Queensland on an isolated torr, called the Hill of Knowledge, which held a total population of about 150 adults. They were distributed around the massif in 12 semi-isolated colonies and appeared to be highly sedentary: during three years of following radio-collared wallabies only one male was observed to move from one colony on the hill to another. As at Black Rock, there was heavy juvenile mortality, so that the separate colonies remained stable or declined. Nevertheless, the genetic analysis showed that there is regular interbreeding between the 12 colonies on the torr and with a small colony 10 km away; however, there was much less interbreeding with the next population 70 km distant (Pope et al 1996). Despite the differences in the mating behaviour of the two species, the genetic outcome is much the same.

Conservation strategies for rock wallabies The main threat to survival of rock wallabies across Australia are grazing stock, introduced preda- tors and genetic inbreeding due to isolation of very small populations. Grazing stock compete for the same forage in the countryside around the rock refuges but goats, Capra hircus, present a worse threat because they compete for the plants living within the rocks themselves. Predation especially by foxes, strikes rock wallabies in two ways: foxes can hunt them within the rocks and so deny them their essential refuge, but they also prevent rock wallabies from traversing the country between refuges, especially where land clearing has reduced the amount of cover. The combined effects of these threats have meant that surviving populations of rock wallabies are small and isolated, so that there is a trend to lower genetic diversity, as in island popula- tions. It is assumed on theoretical grounds that this must lead to inbreeding depression and eventual extinction. The fact that the Barrow Island population and other small island popula- tions have almost no genetic diversity and yet have survived for several thousand years suggests that the theory may need to be revised. Furthermore, thriving populations of several hundred brush-tailed rock wallabies, Petrogale penicillata, have arisen on Oahu Island in Hawaii and on Motutapu Island in New Zealand, each from a founder stock of less than 10 animals, which suggests that, genetically, rock wallabies may be unusually resilient. If this is so, the main cause of their decline across inland Australia must have more to do with competition for food and preda- tion itself. A good example of this comes from Western Australia, where sodium fluoroacetate (1080) can be used as a selective poison because foxes are highly susceptible to it and the native mammals have a high tolerance of it (see Chapter 4). The black-footed rock wallaby was formerly widespread in Western Australia but is now reduced to a few isolated localities with populations of less than 100 animals at each site (Kinnear et al 1998). The sites where they still occur are granitic outcrops with relict woodland, surrounded by cleared farmland. The cleared land is inimical to the survival of rock wallabies and this has reduced their opportunities to move from one site to another. In addition the wallabies are preyed upon by foxes, which have become abundant in Western Australia since the arrival of rabbits. Kinnear and colleagues (1998) demonstrated the profound effect of fox predation on black- footed rock wallabies by following the fate of five small populations from 1979 to 1990. From 1982 poison was regularly laid around two sites, Mt Caroline and Nangeen Hill, and not at three other sites. The number of adult rock wallabies at Mt Caroline and Nangeen Hill quadrupled by 1990 (42–166), whereas at the three sites without fox control the populations continued to decline and one population went extinct (Kinnear et al 1998). Not only did the populations Consummate kangaroos 353

increase after foxes were eliminated but also rock wallabies reoccupied areas of the rock torrs at Mt Caroline that they had previously vacated because foxes could penetrate the crevices in them. Since the two populations at Mt Caroline reserve have increased, seven rock wallabies now occupy another rock outcrop 8 km to the north. The origin of this small colony has been identified as the eastern population on Mt Caroline: molecular genetic markers showed it to be a subset of the founder population with a lower genetic variation (Eldridge et al 2001). The country between the two sites is highly altered farmland and, even without foxes, would have been a hazardous journey for a rock wallaby to make. This is the longest recorded distance traversed by any species of rock wallaby, so it is likely that it was a single event; the minimum number required to make the trip would have been a female carry- ing a pouch young and a uterine blastocyst, one of which was male. In addition to this natural episode of recolonising, five rock wallabies were transferred in 1990 from Nangeen Hill to the site where they had become extinct and, after fox control was imposed there, the population grew to 45 animals in eight years.

Island wallabies With the rise of sea levels that began 14 000 years ago, thousands of islands were formed around Australia and New Guinea, from tiny rock islets above high tide to Kangaroo Island and Tasmania. Some of these islands carry kangaroos, wallabies or rat kangaroos and in some cases these are now the only surviving populations of the species. No island of less than 1 km2 carries macro- pods; islands up to about 8 km2 carry no more than one species, while larger islands may carry between one and four macropods as well as other species of marsupial; only Tasmania carries five species of macropod. Presumably, when an island became separated from the adjacent mainland it would have contained all the species that occurred in the same range of habitats on the mainland, but as time passed some species died out because the isolated population was too small to survive long term or because of competition for resources with one of the other species. This is illustrated by Aru Island, south of New Guinea, (7700 km2), which became isolated about 10 000 years ago. When it was part of the larger land mass it carried one species of Dorcopsis, the agile wallaby and two species of pademelon, Thylogle brunii and Thylogale stigmatica. The two largest species died out at or before its final separation and Thylogale stigmatica about 7000 years ago, so today it carries only one species of macropod. The islands of Bass Strait, which were formerly part of Tasmania, provide another indication of the past because some of them have fossil deposits from a time before European arrival in Australia (Hope 1974). Tasmania (67 900 km2) supports five species of macropod, but is the only one that supports eastern grey kangaroos, although fossil remains of the eastern grey kangaroo have been found on Flinders Island (1330 km2). Of the other islands only the six largest islands (110–1330 km2) – Flinders (not South Australian island of same name), King, Cape Barren, Clark, Robbins and Hunter – carried Bennett’s wallabies, Tasmanian pademelons and potoroos. Interestingly, the long-nosed potoroo did not occur on any of the smaller islands and the Tasmanian bettong occurred on none of the islands. It is not surprising that the eastern grey kangaroo did not survive on the islands because species of large size require a large area to support a minimum number of animals in perpetuity. The commonest species on the Tasmanian islands is the pademelon, Thylogale billardierii, which formerly occurred on 15 islands, the smallest being only 1.4 km2. The smallest island that carried two species – the pademelon and Bennett’s wallaby – was Erith-Dover Island (7.8 km2). Most of these islands lost their macropods in the early 1800s, during the occupation by sealers. Eight islands off the coast of South Australia carried macropods in 1800: only two do today (Robinson et al 1996). The largest, Kangaroo Island (4400 km2), is the only one to support 354 Life of Marsupials

western grey kangaroos and the only one that still supports tammar wallabies, but it formerly had seven other species, comprising the red kangaroo, the red-necked wallaby, the toolache wallaby, the eastern hare wallaby, Lagorchestes leporides, the burrowing bettong, the brush-tailed bettong and the long-nosed potoroo (Kitchener et al 1980). These are only known from fossils and it is not clear when they died out. The three next largest islands (34–39 km2) carried tammars, and brush-tailed bettongs also occurred on one of these (St Peters Island) and on the much smaller St Francis Island (8 km2) but all have now lost their macropods. The most thorough analyses of the distribution of macropod species on different islands has been done on the 24 islands around the coast of Western Australia that carried one or more species of macropod at the arrival of Europeans (Main 1961, Main and Yadav 1971, Kitchener et al 1980, Burbidge 1990, Strahan 1995). Tammars, quokkas and black-footed rock wallabies inhabit islands in the southwest, while three species of hare wallaby, the burrowing bettong and five species of rock wallaby inhabit various islands in the northwest. Among the southern islands tammars occur on five islands, as already discussed, ranging in habitable area from 2.8 to 11.7 km2; black-footed rock wallabies occur on four granite islands, ranging from 1.04 to 9.3 km2; and quokkas occur on two other islands of 7.8 and 15.5 km2, respectively. Since tammars, quokkas and black-footed rock wallabies occurred on the mainland adjacent to the southern islands, what determined which species survived on each island? One factor may have been that the distribution of macropod species on the mainland changed with time, the oldest islands being occupied by the earlier inhabitants and the younger islands by the more recent. For instance, rock wallabies occur on islands more than 10 500 years old, whereas with one exception tammars and quokkas occur on islands younger than this. The exception is the Houtman’s Abrolhos archipelago off the coast from Geraldton, WA, which are 11 500 years old but have no habitat suitable for rock wallabies. Rock wallabies occupy islands with good rock shelter and, because they tend to occupy drier environments than do quokkas and tammars, they were probably present when the climate of Western Australia was drier and so were able to occupy the older islands at the time of their separation; quokkas and tammars arrived later but before 7000 years ago when the younger islands separated. The shelter of rocks allows rock wallabies to occupy islands with less plant cover than tammars and quokkas need: the islands they occupy have fewer plant species than the islands occupied by tammars and quokkas (Main and Yadav 1971). As with the Tasmanian islands, none of the Western Australian islands carried brush- tailed bettongs, even though the species was formerly common on the adjacent mainland. But what led to quokkas surviving on two islands and tammars on five others? Rottnest Island and Garden Island are both derived from the same arc of consolidated limestone dunes and yet Rottnest Island (15.5 km2) carries a population of about 5000 quokkas and Garden Island (12 km2), a smaller population of tammars. Garden Island has no fresh water and, as already mentioned, tammars can survive without fresh water, whereas quokkas cannot; Rottnest Island has fresh water soaks, which are used by the quokkas. Did the competition on each island lead to one species surviving and one going extinct? There is no answer to this question, but analysis of five larger islands in the north, each with more than one species of macropod, sheds some light on it.

Barrow Island Barrow Island is a large island of 236 km2 with a varied vegetation cover and habitats. It carries a dwarf euro (Macropus robustus isobellanus), the spectacled hare wallaby (Lagorchestes conspicilla- tus) the black-footed rock wallaby (Petrogale lateralis) and the burrowing bettong, as well as the northern brushtail possum (Trichosurus vulpecula arnhemensis), the golden bandicoot (Isoodon auratus) and a small dasyurid, the fat-tailed (Pseudantechinus macdonnellensis). These are all species with distributions in northwest Australia, rather than the southwest. One Consummate kangaroos 355

absent species is the red kangaroo; at its current density on the adjacent mainland, Barrow Island would only accommodate about 200 adults, which may have been too small a population to survive long term. There are 1800 euros on Barrow Island (Table 9.6) but they are about half the size of euros on the adjacent mainland, males about 20 kg and females less than 10 kg. Reduction in body mass is a common adaptation of large mammals isolated on small islands, presum- ably because the same biomass can be distributed between a larger population. The other three macropods on Barrow Island are the same size as their congeners on the mainland.

Table 9.6: Barrow Island marsupials Area of the island is 236 km2.

Species Density Total Mass (kg) Reference No./km2) population Pseudantechinus Very few Main and Yadav (1971) macdonnellensis Isoodon auratus 296.6 70 000 0.25–0.67 Strahan (1995) Trichosurus vulpecula Main and Yadav (1971) Bettongia lesueur 14.4 3 400 1–1.5 Short and Turner (1993) Lagorchestes 41.8 9 700 2–4 Short and Turner (1991) conspicillatus Petrogale lateralis 0.6 150 4.5 &/3.5 ( Eldridge et al (1999) Macropus robustus 7.9 1 800 20 &/8.3 ( Short and Turner (1991) Total macropods 63.8 15 050

The rock wallaby population on Barrow Island is very small (about 150 animals), is geneti- cally homogenous and is restricted to the rocky cliff faces on the western coast. The most numer- ous macropod is the spectacled hare wallaby, which lives in spinifex clumps; the next is the burrowing bettong, which lives in warrens excavated under the cap rock, both species being widely distributed across the island. Both species formerly occurred on small islands nearby. The spectacled hare wallaby occurred on two of the Montebello Islands, 24 km north of Barrow Island, but had died out by 1950, prior to the British atomic tests. The burrowing bettong may have shared the larger Hermite Island (10 km2) with the spectacled hare wallaby, but the evidence for this is a single lower jaw found there in 1958 (Serventy and Marshall 1964). If the jaw came from an animal that lived there, this is one of the smallest islands on which two species of macro- pod coexisted. An alternative explanation is that the jawbone was taken to Hermit Island from Barrow Island by a sea eagle. Boodie Island, near Barrow Island, carried a small population of burrowing bettongs but they declined after the introduction of black rats, Rattus rattus, and were then inadvertently destroyed when they ate the poisoned bait laid for the rats. While populations on very small islands are vulnerable to random events, species on a large island, like Barrow Island, can continue to thrive alongside the industrial activities of a major oil well.

Time capsules of Shark Bay Bernier and Dorre Islands represent a microcosm of what Australia was like before European occupation, and Dirk Hartog Island shows what happened afterwards. These three Shark Bay islands are derived from a single arc of consolidated limestone, which became separated from the continent about 8000 years ago and from each other about 3000 years ago. Aboriginal people did not occupy them and the only occupation of Bernier and Dorre Islands since was when hospitals 356 Life of Marsupials

were established on both islands for Aboriginal people between 1904 and 1911. Dirk Hartog Island became a sheep station in 1899 and has remained so. Since 1699 explorers and scientists have visited the three islands sporadically and recorded the condition of the fauna, so that there is a remarkably long history of a part of Australia that has remained unaffected by European occupation. Bernier Island (44 km2) and Dorre Island ((53 km2) both carry populations of three macropod species, as well as the western barred bandi- coot, Perameles bougainville (see Chapter 5). All four species formerly inhabited the adjacent mainland but began to disappear after sheep were introduced to this part of their range in 1880 (Shortridge 1910). They survived for longer in the southwest part of the continent into the first two decades of the 20th century, but none exists on the continent today. The banded hare wallaby used to occupy eucalypt woodlands, especially low, densely branched bushes beneath which they made runways, where they could shelter from predators and the midday sun. By contrast, the rufous hare wallaby or mala, Lagorchestes hirsutus, was an inhabitant of spinifex grasslands of the dry inland plains, and also of the sandplain heath in what became the wheat belt of Western

Dorre Is. Bernier Is.

Mala Burrowing bettong

Mala Burrowing bettong Banded hare wallaby

Banded hare wallaby

Figure 9.25: Distribution of the banded hare wallaby, Lagostrophus fasciatus, the mala, Lagorchestes hirsutus, and the burrowing bettong, Bettongia lesueur, on Bernier and Dorre Islands, Western Australia, to show how each species uses different parts of the island habitats. Transverse lines are the census transects for each set of histograms. From Short and Turner (1992, 1993). Consummate kangaroos 357

Australia. By 1990 this wide distribution on the continent had shrunk to two tiny populations in central Australia, one then being lost to foxes and the other to wild fire; now captive bred animals are being released into protected areas of its former habitat. The third species, the burrowing bettong, was also an inhabitant of the inland plains with a distribution across the continent, but now only exists on these two islands and on Barrow Island further north. Dirk Hartog Island (620 km2) formerly had the same suite of marsupials as Bernier and Dorre Islands but lost them soon after it became a sheep station in 1899. Despite its large size, Dirk Hartog Island apparently never carried either euros or red kangaroos: red kangaroos do not occur on the adjacent mainland but euros are abundant, so it is not clear why they were not on the island and have not even been found as fossils. The early explorers recorded banded hare wallabies and burrowing bettongs but no one reported the presence of the mala. Since mala survive on Bernier and Dorre islands, which were northern extensions of Dirk Hartog Island, the species may have died out after the separation of the outer islands and before explorers visited the island. How have the three macropod species on Bernier and Dorre islands survived in isolation for 8000 years without one or the other going extinct? Each species lives in a different habitat, as their ancestors formerly did on the mainland. Burrowing bettongs live in groups in burrow complexes they excavate in the cliffs along the edge of the islands, and they feed on the roots of bushes and carrion scavenged along the shore. The banded hare wallaby is also gregarious, feeding on spinifex and leaves of the low branching shrubs under which they shelter and which cover much of the centre of the two islands. The mala is a solitary animal that occupies the sand plains on the upper parts of the islands and either shelters under the spinifex tussocks or in short burrows about one metre long, which it excavates under or adjacent to a thicket. From the visits of explorers and scientists over nearly 300 years at different times of the year and at different points on the islands, it seemed that the populations of each species fluctu- ated greatly and the long-term survival of all three was uncertain. Short and his colleagues have estimated the size and distribution of the populations of all four marsupial species by counting the animals on 25 line transects across both islands in 1988–89, after a prolonged dry period (Short and Turner 1992, 1993), and in 1991–92 after two years of above-average rainfall (Short et al 1997). This showed that the populations of all four species do change markedly in response to previous rainfall (Table 9.7), and that the density of the macropod species varies across the islands in relation to habitat, so that where one species is abundant the others are less so (Fig. 9.25). Thus, some of the reported variation in abundance of the three species may have more to do with successive scientists visiting different parts of the islands than a real change in the animal populations. For instance, in 1988–89 banded hare wallabies were at their highest densities and the other two species at low densities at the north end of Bernier Island and the south end of Dorre Island, whereas the mala was most abundant at the south end of Bernier Island, where there were no banded hare wallabies. At the same time the burrowing bettongs were at relatively low densities on both islands, possibly because they had been worse affected by the preceding drought. At the first survey the total populations of the three species reflected this: banded hare wallabies were the most numerous on both islands (56%), malas the second most abundant and burrowing bettongs comprised only 12% of the total (Table 9.7). Dry conditions had prevailed on the islands for two and three years, respectively, before the first survey. Good rain fell in April 1989, just before the first survey of Bernier Island, and continued until the second surveys took place. All three species increased on both islands but to different degrees; the two hare walla- bies increased less than the burrowing bettong, which more than doubled from 1650 to 4290. However, this may indicate that burrowing bettongs had suffered a greater decline during the drought than the two larger species and recovered quicker after rain. 358 Life of Marsupials

Table 9.7: Populations of three macropods and a bandicoot on Bernier and Dorre islands in 1988–89, after three dry years and in 1991–92, after two wet years Data from Short and Turner (1992, 1993) and Short et al (1997, 1998).

Species 1988–89 1991–92 Per cent increase Perameles bougainville 4 400 5 380 22 Bettongia lesueur 1 650 4 290 160 Lagostrophus fasciatus 7 700 10 100 31 Lagorchestes hirsutus 4 300 6 270 46 Total macropods 13 650 20 660 51 Macropod density (No./km2) 14.1 21.3

The differences also reflect each species’ variation on the macropod pattern of reproduction. The mala is an opportunistic breeder, like the red kangaroo (Johnson and Burbidge in Strahan 1995), whereas the banded hare wallaby is a more seasonal breeder with a main cohort of young born in May and sometimes a second young in November (Tyndale-Biscoe 1965), reflecting its former origin from the winter rainfall region of the continent. The burrowing bettong, like other rat kangaroos, is a continuous breeder with females potentially able to produce a succession of three young in a year and young females reaching sexual maturity in their first year. This could account for the rapid recovery of the burrowing bettong as compared to the mala during favour- able seasons. Nevertheless, breeding was not continuous on Bernier and Dorre islands, with few females carrying pouch young during the summer but more than 90% of them doing so after the autumn rains. In 1988, when the islands were dry, none of the females were lactating but this changed after the beginning of the rain in August 1989. It seems that the continuous breeding strategy has been modified on the islands to the extent that females are regulated by rainfall or its lack. While the sex ratio of pouch young is equal, adult males outnumber adult females by 40%, so there must be a sex-biased mortality of females. The life span on the islands was estimated to be more than three years but how much more is not known. Other bettongs can live for five years. The burrowing bettong is the only macropod marsupial that constructs large underground warrens with many entrances, like wombats do. On the islands and in captivity burrowing bettongs preferred to use warrens with many entrances, rather than simple burrows like those made by the mala. This also means that several animals regularly share the large warrens during their daytime rest. These groups are fairly stable and comprise two or more females and their offspring but only one adult male: males do not share warrens with other males and, indeed, they use their time in the warren to check the sexual condition of the associated females. To this extent the mating system of the burrowing bettong is polygynous even though the males are not larger than the females; males defend the warren but do not engage in overt fights with other males for access to females. Unlike rock wallabies, though, the social cohesion of the burrowing bettong group does not extend to the night, when all animals disperse to feed alone. Solitary feeding is common among small herbivores that rely on searching for small items of scattered food, by comparison with the large macropods, which feed in groups on more widespread but less nutritious food. The burrowing bettong is unusual in showing social grouping during the day but this is probably because of the advantage of sharing the construction of warrens. By retreating underground during the day burrowing bettongs must gain economies in water conservation, as wombats do, and may also be more secure from daytime predators, such as sea eagles and wedge-tail eagles (Short and Turner 1999, Sander et al 1997). On the continent burrowing bettongs also made very large warrens with up to 90 entrances, which were occupied Consummate kangaroos 359

by as many as 20 animals in a warren. These large constructions are very durable and can still be seen in the mulga woodlands of New South Wales and Queensland (Noble 1999), although their makers have long since gone.

Future prospects for burrowing bettongs So long as the three islands where the burrowing bettong survives remain free of grazing animals and introduced predators, the burrowing bettong is secure. But the history of its total demise across the continent and on Dirk Hartog Island shows that it is very vulnerable to change. In Western Australia vigorous efforts are now being made to reintroduce the species to some of its former range, most notably on Heirisson Prong, a peninsula of the same formation as Dirk Hartog, Bernier and Dorre islands. Because of its configuration the peninsula has been secured from foxes but so far cats still remain. Burrowing bettongs from Dorre Island were introduced in 1992; initially in a fenced area and more recently the fences have been opened. The popula- tion was well established in 2000, with over 1000 young born (Short 2000), but predation by cats since then has severely depleted the population.

Life of the smallest macropods The smallest macropods are the rat kangaroos in the genera Aepyprymnus, Potorous and Bettongia (Table 9.1). Potorous tridactylus weighs about 850 g and Potorous longipes is somewhat larger. Apart from the burrowing bettong, already discussed, there are three other species of Bettongia, namely the brush-tailed bettong of southern Australia, the Tasmanian bettong (also formerly south-eastern Australia), and Bettongia tropica the northern bettong, restricted to North Queensland. All three species are similar in size (1.2 kg) and habits, are closely related and in their former distributions did not overlap. The brush-tailed bettong formerly occupied much the same large area of Australia as the burrowing bettong but also included the southern regions of the continent. Despite this the only offshore islands that carried this species were St Francis Island and St Peter’s Island, off South Australia. The populations of both islands were exterminated by the 1920s (Jones 1924) and no specimens from them have survived but it has been identified from recovered bones (Robinson and Smyth 1976, Robinson et al 1996). As mentioned earlier, no Tasmanian island carried the Tasmanian bettong either; so, did some aspect of their biology prevent bettongs from surviving on islands? Neither species constructs burrows and perhaps this was the reason that they could not survive on islands, where water was scarce or absent. A more critical factor may have been diet: potoroos and bettongs are too small to live on grass and forbs alone and depend for a maintenance diet on the fruiting bodies of subterranean fungi that are associated with the roots of trees, especially species of Eucalyptus and Casuarina. One consequence of island separation is the decline of tree species and, presumably, their associated fungi. Without an adequate supply of fungi these smallest macropods would have been at a serious disadvantage in competition with the larger, grass-eating species, such as tammars, rock wallabies and pademelons. The burrowing bettong is the exception, being less dependent on fungi than the other three species, because of the economies of living in burrows and because it is able to supplement its diet with animal protein scavenged on the beach.

Eucalypts, fungi and rat kangaroos: a productive partnership The significance of the mutual interdependence of potoroos and bettongs, hypogeal fungi and eucalypt forest trees has only become appreciated in the past 15 years and in this Chapter they 360 Life of Marsupials

stand as the final example of macropodid adaptation to the special conditions of the Australian environment. Indeed, this may be one of the oldest adaptations of macropodids, since species of Bettongia are known from the late Oligocene when the eucalypts were making their first appear- ance in the fossil record. The best known underground or hypogeal fungi are the truffles of Europe that are associ- ated with oak trees and are found by pigs and specially trained dogs. In Australia there are many species of hypogeal fungi, which form close, symbiotic relationships with species of Eucalyptus trees. The mycelium weft of the fungus infects the root system of the tree, drawing carbohydrates from it and providing access to soil water and nutrients by greatly increasing the absorptive area of the tree’s root system. Without the fungus the tree species would be very limited indeed; but the fungus depends on mammals to distribute its spores and infect other trees. The fruit- ing body or sporocarp of the fungus is formed 5–20 cm underground and comprises an outer inedible rind and an inner layer of spores and supporting tissues. In the main Australian family, the Mesopheliaceae, the layer of spores is sandwiched between the rind and a core of especially edible material, which is highly attractive to bettongs and potoroos. The sporocarp emits a pungent odour above the ground, which leads the rat kangaroo to burrow down and retrieve it. When it opens the sporocarp some spores are dispersed in the air and on its fur, and some are eaten along with the tasty core. In some fungal species passage through the mammalian gut is also important for germination of spores, which then more readily infect the roots of eucalypts: these spores are carried much further away from the host fungus and tree. Because the fungal fruiting bodies grow underground they are an important source of food for the mammal after fire, when other plant food has been destroyed. Indeed, several species of

Brush-tailed bettong 1.02

1.00

0.98 Condition index 0.96

0.94 67891011121314

Sporocarp production (no./m2)

Figure 9.26: Body condition of Bettongia gaimardii in relation to abundance of underground sporocarps of subterranean fungi. From Johnson (1994). Consummate kangaroos 361

Mesophelia are stimulated to form sporocarps by cool fires, either directly by the rise in ground temperature or more indirectly, by responding to alterations in the host tree caused by the stress of being burnt and the consequent changes to its water physiology. They also emit a different odour after fire and both these responses help to disseminate fungal spores at this time. These several adaptations suggest that hypogeal fungi, eucalypts and rat kangaroos have co-evolved in Australia over a long time. Apart from rat kangaroos other marsupials, such as bandicoots, wombats and possums, as well as rodents, eat hypogeal fungi but for none of them is it the major item of the diet. Potoroos and bettongs are the specialists in using this source of food, which forms more than half their daily intake. Since squirrels in North America and bandicoots in Australia are unable to thrive on a diet solely of fungus, how do rat kangaroos do so? The crucial difference is that they, like other macropodids, have a sacculated forestomach in which microbial fermentation takes place, with breakdown of the fungal cell walls, release of short-chain fatty acids and then digestion of the microbial protein in the small intestine. Johnson (1994) assessed the nutritional worth of hypogeal fungi for the Tasmanian bettong living free in Tasmania and found that they made up about 70% of the diet throughout the year and that the bettongs would prefer to eat fungus even when other plant food was abundant. Where there was an abundance of fungal sporocarps the condition of the adult bettongs was high, the growth rate of the young was rapid and the population was dense (Fig. 9.26). Female bettongs consumed more fungus than males and this was associated with accelerated growth of pouch young, since the growth rate of the young is related to the body mass of the mother (see Chapter 2). The same heavy reliance on hypogeal fungi has been observed in the brush-tailed bettong in Western Australia and in the long-nosed potoroo in Victoria, although both species supple- mented fungi with plants and some animal protein (Bennett and Baxter 1989). This led to the idea that fungal sporocarps cannot provide a balanced diet and must be supplemented with other foods. However, Claridge and Cork (1994) showed that it is a high energy source due to the abundance of fatty acids in the core and Wallis et al (1997) found that fungal sporocarps have a good balance of essential amino acids when compared to vertebrate muscle. They also noted that after destructive forest fires, when no vegetation remained above ground, the potoroos survived entirely on hypogeal fungi. Also their water turnover did not change through the seasons because they were not reliant on succulent vegetation above ground.

Rat kangaroos in North Queensland: how three species live in different habitats The vegetation of North Queensland is determined by the east–west gradient from high to low rainfall: rainforest extends from the coast up to 600 m and is replaced in the west by a narrow band of wet sclerophyll eucalypt forest, which is being encroached on by the rainforest. In its turn the wet sclerophyll forest gives place further west to open woodland with a grassy understo- rey. Each of these habitats is occupied by a different species of rat kangaroo, each with a different diet, which nicely illustrates the role of hypogeal fungi in the economy of rat kangaroos and how they respond to conditions after fire. The musky rat kangaroo, at 0.5 kg is the smallest species and, as mentioned earlier, is only very distantly related to the other two species. It lives in the coastal and upland rainforest and depends on seasonal fruit fall and ground-living insects; it does not eat hypogeal fungi, which are largely absent from the rainforest. Moreover, it has a simple stomach and could not ferment fungus even if it were available. This tiny marsupial has a clearly defined annual pattern of breeding that is closely linked to seasonal production of fruit on the forest floor in the second half of the year, when the females are in late lactation and the young are dispersing – the period of greatest nutri- tional demand for females (Dennis and Marsh 1997). The males also undergo seasonal changes 362 Life of Marsupials

with great enlargement of the testes in October–April, associated with aggressive behaviour only during this time. Each female has one litter of from one to three young in February–March and at no other time of the year; the young vacate the pouch in October but remain in the maternal nest until January of the following year when they disperse. Life expectancy is about four years, the main predators being pythons, goshawks, owls and quolls. The rufous bettong, which at 3.5 kg is the largest species of rat kangaroo (Table 9.1), lives in the grassy woodlands with low rainfall, where it makes a nest in grass tussocks and is generally solitary. Its former range across much of eastern Australia has, since European occupation and the spread of foxes, greatly diminished and it is now found only in northern New South Wales and Queensland, where foxes are scarce. Being larger than bettongs and potoroos it can meet 80% of its dietary needs from grass, as spectacled hare wallabies do: hypogeal fungi form only a small proportion of its diet, mainly as a source of protein nitrogen (McIlwee and Johnson 1998). Indeed, fungal sporocarps are so sparsely scattered in the woodlands that foraging for them would be energetically too costly for such a large animal. Female rufous bettongs produce a succession of up to three young a year with no particular breeding season. Males are the same size as females and maintain transitory contact with as many females as possible, visiting at their nest sites, of which they have intimate knowledge. Males that successfully mate are ones that have associated with the female previously (Strahan 1995, Frederick and Johnson 1996). The northern bettong at 1.2 kg is more than double the mass of the musky rat kangaroo and one-third the mass of the rufous bettong. It lives exclusively in the wet sclerophyll eucalypt forest between the rainforest and the grassy woodlands and feeds predominantly on fungal sporocarps, which are a high value food source and are abundant under the eucalypt trees throughout the year. It does not penetrate the rainforest, where the trees only associate with fungi that do not form sporocarps (Johnson and McIlwee 1997), nor can it compete with the larger rufous bettong in the woodlands, where hypogeal fungi are less abundant, because it is too small to survive on an exclusive diet of grass and forbs. It thus occupies a special intermediate niche, less than 10 km wide, where fungal sporocarps are plentiful throughout the year, allowing year round breeding and a stable population. Northern bettongs are solitary and feed upon a patchily distributed, high-quality food resource, so during a night’s feeding they move about 150 m per hour within a relatively large home range of 60 ha. This is considerably larger than the home range of potoroos (10 ha) and musky rat kangaroos (1–4 ha), which live in denser forested habitats, and of larger macropods, which feed on evenly distributed grasses (Vernes and Pope 2001). An important question is how northern bettongs survive low intensity fires, which occur frequently in their habitat. Karl Vernes investigated this by comparing the behaviour, body condition and breeding success of northern bettongs in burnt and unburnt portions of their habitat (Vernes et al 2001). In the immediate aftermath of the fire, when their nests and cover had been destroyed, they sheltered in boulder banks and all survived. There was no change in home range or movements of radio-tracked bettongs and almost no mortality as a result of direct or indirect effects of the fire. However, there was an immediate change in the species of fungus eaten after fire: bettongs eat at least 35 species but immediately after fire they ate more of the fire-adapted family Mesopheliaceae, which respond to fire by producing more sporocarps. Despite these changes in diet the body condition of bettongs did not change, indicating that they are not compromised by fire. Breeding was continuous in Vernes’ population, most females carrying pouch young and births occurring in all months. Furthermore, there was no change in breeding success or the growth of the young bettongs in burnt and unburnt areas, so fire did not affect either the likeli- hood of a female having a pouch young or of the young dying before pouch exit. As with other species of bettong, sexual maturity is attained in the first year and life expectancy is about five Consummate kangaroos 363

years. However, dispersal to other populations is low, suggesting high subadult mortality (Vernes and Pope 2002). Because the northern bettong is a fire-adapted species, like the potoroo in Victoria and the brush-tailed bettong in Western Australia, it has an advantage in this habitat over the larger rufous bettong that depends on grass for the major part of its energy needs. Thus, these three species of rat kangaroos in North Queensland each occupy one habitat type to the exclusion of the other two species. Of the three, the northern bettong is the most vulnerable because the wet sclerophyll forest, to which it is adapted, is being inexorably encroached on by rainforest (see Chapter 6 in regard to the yellow-bellied glider, Petauris australis). But there is another more serious factor: the narrow belt of wet sclerophyll habitat is now regarded as prime land for residential development, and this is likely to have deleterious effects on bettongs; modification and loss of habitat from subdivision, increased roads and vehicles leading to greater numbers of road kills and predation by domestic cats and dogs. As with the yellow-bellied glider in this habitat, the northern bettong has disappeared from areas that it formerly occupied and its present distribution is fragmented: if it is to survive, its special habitat must be preserved, and that includes regular low intensity fires (Laurance 1997). Why have these small macropods declined so much since European occupation? Clearing eucalypt woodland removes hypogeal fungi, their main dietary source, and grazing has removed cover, making the smaller species much more vulnerable to foxes. The only survivors are the three species that live north of the fox, and the Tasmanian bettong and long-nosed potoroo living in Tasmania. Conversely, the disappearance of all these species that were agents in the distribution of fungal spores may have adversely affected the eucalypt species that still grow in their former habitats: ‘dieback’ is a widespread phenomenon in eucalypt woodlands in Australia and the cause has been variously ascribed to infection with the fungus Phytophthera, changing water tables, and salinity. Loss of symbiotic fungi formerly carried by rat kangaroos may also be an important factor.

Conclusions This Chapter began with the notion that kangaroos are pre-eminent among marsupials because of three features of their common inheritance: foregut fermentation, bipedal locomotion and embryonic diapause. Foregut fermentation has allowed large desert-living kangaroos to balance the competing requirements of thermal load, scarce water and impoverished feed; small walla- bies to survive on islands devoid of fresh water; and rat kangaroos to live on underground fungi even after fire. The considerable economies of hopping have enabled red kangaroos to cover great distances in search of green feed and rock wallabies to scale rock cliffs. And their special repro- ductive physiology has enabled some species to respond to predictable episodic rainfall, some to regular abundant rainfall and some to highly irregular and wholly unpredictable rainfall. Despite the beautiful adaptiveness of all kinds of kangaroos, wallabies and rat kangaroos to the Australasian environment, most species have fared disastrously since 1800 when European occupation of Australia began. Without exception the species less than 5 kg in weight have either gone extinct or survive in a small part of their former range: often this is an offshore island. But many offshore islands have also lost their macropod species, either due to the introduction of sheep or alien predators. Only the largest kangaroos and wallabies are secure. This is despite competition with the enormous increase in herbivore biomass in the form of sheep, cattle, and to a lesser extent, feral camels, donkeys and horses. In some parts of inland Australia the land could not sustain the increased herbivore biomass for long and both the grazing stock and the kangaroo populations declined: in other parts the pastoral industry, based on sheep, is in severe 364 Life of Marsupials

decline and only survives with government subsidies, including the costs of controlling kanga- roo numbers. This suggests that the rangelands cannot sustain such a large herbivore biomass in the longer term. An alternative to use the native herbivores instead of the introduced species has been canvassed for 15 years but has not yet been seriously attempted. Chapter 10

Marsupials and people: past and present Bradshaw rock painting of people hunting kangaroos in Northern Australia. From Walsh and Morwood 1999. Marsupials and people: past and present 367

arsupials have had a large part in the culture of the peoples of Australia and New Guinea as sources of food and as part of the Dreamtime creation stories (see Flood M1997). In both places species now extinct can be recognised in ancient rock art, and the former distributions of species that have vanished since European settlement have been traced by interpreting the memories and creation stories of Aboriginal people today (Tunbridge 1991). Ecologists have also learnt to appreciate the significance of the sacred sites of species, such as the red kangaroo and the mala, as the critical refuges for the species. The present view is that people first reached Australia and New Guinea between 60 000 and 40 000 years ago, although indirect evidence suggests that it could have been considerably earlier. People entered North America much later: between 13 000 and 12 000 years ago, when there was a land bridge across the Bering Straits, and they spread to the southern-most extent of South America within 2000 years. Thus, neither Australasian nor American marsupials encountered people until very recently, in geological terms. Human occupation of Australia and New Guinea and later of North and South America was followed by the disappearance of many of the largest species of indigenous mammals. In Australia and New Guinea these were marsupials; in the Americas these were all placental mammals. Not only mammals were affected: in Fiji and Vanuatu giant crocodiles, rails, frogs and goannas disappeared at about the same time as the first occurrence of human settlement on these islands 26 000 years ago and in Australia giant birds, lizards, snakes and crocodiles disap- peared at the same time as the large marsupials. One view for the widespread extinction of many genera of large reptiles, birds and mammals at the close of the Pleistocene epoch is that it resulted from climatic changes at the end of the last Ice Age, 12 000 years ago, not from the arrival of hunter–gatherer people. This view has been challenged by the findings that extinctions around the world occurred at different times in relation to climate but all occurred after the arrival of human populations into new lands (Martin and Klein 1984). The rapid extinction of the horse and mammoths from North America is now well documented: in Michigan, USA, the decline to extinction of mammoths occurred in little more than 1000 years. In South America a similar disappearance of horses occurred, as well as the giant sloth and other large mammals of that continent. Before human arrival there was one marsupial in North America and the marsupial fauna of Central and South America consisted of the small forest-dwelling species that occur there today. Human impact on them was probably negligible and some species became commensal with human habitation. However, the wholesale loss of the large herbivores from North America 12 000 years ago must have had profound effects on the habitat and on the populations of the surviving species. The spread northwards of the Virginian opossum from Mexico through the eastern seaboard of North America may have been one such consequence (see Chapter 3).

Holocene changes and human impacts in New Guinea and Australia Pollen and charcoal samples taken from lake beds in Queensland and New South Wales show that there were marked changes in vegetation and indications of a new fire regime that began in eastern Australia between 40 000 and 100 000 years ago, the Pleistocene epoch (Flannery 1994). If the changes were the result of human occupation, and some species of the extinct megafauna survived in southern Australia until 11 000 years ago, then human populations coexisted with the megafauna of Australia for tens of thousands of years. The big question that has been debated for the past 30 years is what was the role of humans in the demise of the Australasian megafauna. Did people, either by direct predation or indirect habitat alteration, cause it? Or was it a natural 368 Life of Marsupials

event caused by increasing aridity after the last Ice Age 12 000 years ago? Or was it a combina- tion of both? Since the question is still unresolved we can only put the contending evidence for the two options. This includes the evidence for climate and sea level changes over the past 150 000 years, the time of arrival of human populations, the distribution of the megafauna and estimates of minimum ages of fossils; and any evidence for direct interaction between people and megafauna.

Palaeoclimates and human arrival The past climate of the world going back 300 000 years is now well established from past sea temperatures, determined by the ratios of the two isotopes of oxygen in deep sea cores and in the Antarctic ice sheet, which provide information about former sea levels. These levels are corroborated by independent evidence from the ages of former coral terraces on the Huon Peninsula, New Guinea (Chappell 1993) (Fig. 10.1). The Third Ice Age closed about 140 000 years ago, when the sea level was 110 m below the present sea level, and it was followed by a world wide warming with rising sea levels, which reached a maximum equal to the present sea level at 120 000 years ago (Fig. 10.1). The subsequent interglacial period lasted until 70 000 years ago, and was followed by the Fourth or last Ice Age, when the New Guinea highlands, the Snowy Mountains and Tasmania were glaciated. The world emerged from the last Ice Age about 12 000 years ago, with consequent rise in the sea level to its present high point.

Limits of accuracy Electron spin resonance of dating methods Uranium-Thorium and optical luminescence

14C thousands of years before present 140 120 100 80 60 40 20 0 Sea level 0

Huon Terraces

-50

Deep-sea cores

-100 m Earliest dog Tasmania isolated Megafauna extinction Last thylacine in Australia & New Guinea Mungo 3, New South Wales Devil's Lair, Western Australia Lake Menindi, New South Wales Cuddie Springs, New South Wales

Figure 10.1: Sea level changes during the past 140 000 years, based on two data sets, oxygen isotope measurements from deep-sea cores and ages of exposed coral terraces on the Huon Peninsula, New Guinea. After Chappell (1993). Above the sea levels are shown the limits of accuracy of four current methods of dating archaeological material; and below, the salient sites and events, based on these methods, in the prehistory of Australia and New Guinea. After Mulvaney and Kamminga (1999). Marsupials and people: past and present 369

The times of lowered sea levels reduced the width of sea separating New Guinea and Australia from South-East Asia, which would have made it easier for people to cross in small craft. The Third Ice age, 140 000 years ago, was one time when the distance would have been minimal, but there is no evidence of human occupation as far back as this. The next period when the sea level was 50 m lower than at present was between 60 000 and 70 000 years ago and it is likely that this was when people reached New Guinea and Australia, but not Tasmania. Between 17 000 and 20 000 years ago, during the Fourth Ice age, the sea level was 110 m lower than at present, so that a huge area between northern Australia and New Guinea was lowland, and the Australian coastal plains extended to the continental shelf and encompassed Tasmania and Kangaroo Island (Fig. 10.2). The shortest sea distance to South-East Asia was between Timor and the north-west coast of Australia, represented today by the Ashmore Reef, which was less than 100 km. This would have allowed new waves of people to cross the narrow gap, and this was when stone ground axe technology appears in Australia and people appear to have occupied Tasmania for the first time. This was also when there was a land connection between Asia and North America.

NEW GUINEA

WALLACEA Aru

SAHUL

Arnhem Land

GREATER AUSTRALIA

Cuddie Springs

Mungo Devils Lair

0 500 1,000 2,000 km

Figure 10.2: Map of Australasia with the outline of the continental shelf that would have been exposed 20 000 years ago, and (V) locations of important archaeological sites in Australia, and two probable routes of entry by people from South-East Asia. After Chappell (1993). 370 Life of Marsupials

An important effect of the increasing temperatures and rising sea levels that followed was that much of the most fertile, vegetated land between northern Australia and New Guinea was inundated and Tasmania was once again isolated by an open Bass Strait. This inundation would have greatly reduced the habitable land, with profound disruptions to the people’s lives, especially as it coincided with an increasing aridity of the remaining land. Opinions vary widely on the rate at which people occupied Australia. Mulvaney and Kamminga (1999) considered that the whole continent, apart from the centre, was occupied in the first few thousand years, whereas Horton (1984) thought that only the coastal strand was occupied, with people spreading southwards very slowly, relying on fishing and hunter- gathering for their livelihood. The oldest dated skeleton is one discovered by Jim Bowler on the shores of Lake Mungo in inland New South Wales. His original estimate, based on 14C dating was 25 000 years but this has been revised to 62 000 years, using the new technique of electron spin resonance (Thorne et al 1999). If this revised age for the Mungo 3 skeleton is correct, much of the continent was occupied very early. Other sites from 50 000 years ago are known from Arnhem land, southwest Western Australia and Victoria. No evidence for human occupation of Tasmania at this time has been found, which suggests that the sea level was not low enough for a land connection 60 000 years ago. It is also not clear whether there was only one occupation by people or whether later waves of people displaced the first ones. However, there is growing support for Birdsell’s (1993) trihybrid theory of three successive waves of people coming to Australia from Asia, the last being about 5000 years ago, when they brought the dog.

The megafauna The marsupial fauna in Australia and New Guinea formerly included a suite of very large herbivorous species and a few large carnivores. These include very large kangaroos, wallabies and wombats, as well as the rhinoceros-sized Diprotodon and the large short-nosed kangaroos or sthenurines, the carnivorous Thylacoloeo and a large thylacine. It also included the mihirung, a giant bird , not related to the emu, a giant python and a giant goanna. Fossil remains of 15 species of large marsupial are also known from the late Pleistocene epoch of New Guinea (Flannery 1995). These come from widely scattered sites across the island and include the thylacine, nine species of wallaby and five species of diprotodonts. The diprotodont species and seven of the largest macropods disappeared from New Guinea about 40 000 years ago, but two macropods and the thylacine survived until 2000 years ago. The earliest remains of the New Guinea singing dog are of about this age and Flannery (1995) supposed that the dog increased the hunting capability of the people and may have been a factor in these extinctions, and also of the extinction of two species of pademelon in the highlands about 3000 years ago. On Aru Island, which during the Fourth Ice Age was part of the lowlands between Australia and New Guinea, the agile wallaby was common until 10 000 years ago. Its disappearance then may have been due to vegetation change rather than human hunting. In Australia about 50 species of large extinct marsupials have been discovered, which we widely distributed across the continent and on Tasmania. Horton (1984) noted that the criterion for what is megafauna is somewhat loose: if the cut off point is 44 kg, which Roberts et al (2001) used, it includes the six species of living kangaroos and three species of wombat, and it excludes several of the smaller extinct species of Sthenurus and Propleopus. Also, some of the larger extinct species are thought to have transformed into the smaller living species (ie Macropus titan became the eastern grey kangaroo, M. giganteus, see Chapter 9). From 167 sites across eastern and south-western Australia, Horton (1984) plotted the distri- bution of the extinct species: there were three main assemblages. Among the most southerly were four species of Sthenurus, restricted to the southern coastal regions from Tasmania to Marsupials and people: past and present 371

south-western Australia. Four other species covered the same regions but extended further north in eastern Australia: these were Zygomaturus, Palorchestes, and another species of Sthenurus. Five other species of Sthenurus, two other species of Procoptodon, the giant wombat Phascolonus and a species of Propleopus occurred across the inland region of eastern Australia. The largest of all the extinct species, Diprotodon optatum, was the only species that occurred across the whole of eastern Australia. Horton (1984) thought that as the continent became more arid the vegetation, on which these grazing and browsing mammals subsisted, shifted south and they had to follow. Continuing severe aridity from 12 000 to 9000 years ago further restricted the movements of large herbivores, which would have been constrained by the need for continuous sources of water, and this finally led to their demise. In support of his argument Horton (1984) noted that fossil remains of the red kangaroo, the epitome of a desert-adapted species, have been found in southern Victoria, much further south than the present distribution of the species. He argued that the red kangaroo survived the great aridity that destroyed the other megafauna because of its special adaptations to desert condi- tions. This argument critically depends on knowing the youngest age of the extinct species. Some of the early measurements, determined by 14C dating, of Diprotodon bone from Lake Menindee, NSW, gave an age of 11 000 years ago for the most recent specimens, and at Cuddie Springs, near Brewarrina, NSW, estimates of bones found in matrix were dated to about 20 000 years ago. Roberts et al (2001) have challenged the idea that the megafauna survived into the last arid period 12 000 years ago. They determined the ages of 23 genera of extinct megafauna by two new methods, thorium–uranium decay in bone samples and optical luminescence of sand grains found in association with the bones. For ages greater than 20 000 years these techniques allow much more accurate measurements than by 14C dating (Fig. 10.1). Furthermore, they selected only bone from articulated skeletons, because of the greater likelihood that these had remained in the place where the animal died, and they sampled from fossil sites across the continent as well as in New Guinea. From this wide sample the age of the specimens varied from 80 000 to 37 000 years ago: none was found to be incontrovertibly younger than this. This new evidence strongly suggests most species of megafauna went extinct by 44 000 years ago. This is 10 000 years after the most likely time of human arrival in Australia and more than 20 000 years before the onset of the great arid period; so it does not support the climate change hypothesis but does allow that humans could have been involved in the demise of the megafauna. Others, notably Field and Fullagar (2001) and Wroe and Field (2001), are vigorously challenging the conclusion that no megafauna survived past 37 000 years ago, and the further conclusion drawn from this, so the verdict is still out.

Human interactions with megafauna Two distinct cultural phases are recognised in prehistoric Australia by Mulvaney and Kamminga (1999). The first is distinguished by unifacially worked pebbles and by the use of flake tools and the parent core. These were heavy, hand-grasped tools similar to those of the early to mid- Palaeolithic culture of other lands. The tools occur in the lowest archaeological strata in south- ern Australia and on Flinders Island, Kangaroo Island and Tasmania. The people who occupied the island sites became cut off from the mainland 11 000 years ago when the sea level rose. They survived on Tasmania but died out on the other two, presumably because the islands were too small to sustain a human population indefinitely. On the mainland the flake and core culture was superseded by a much more advanced stone culture, characterised by finely worked flakes, known as pirri and bondi points, which were hafted to wooden handles, and by mills for grind- ing seed. Mulvaney and Kamminga (1999) think both technologies were developed in Australia between 7000 and 5000 years ago. However, their resemblance to similar tools elsewhere has been thought by others to be evidence of a new wave of people bringing new techniques. 372 Life of Marsupials

Mammoth Cave in southwest Western Australia is more than 31 000 years ago and contains the remains of five extinct marsupial species, some of which show evidence of charring (Merrilees 1984), but it is likely that this was not due to human activity but was caused by natural fires. There is less doubt, however, about the earliest deposits around the shores of the Menindee Lakes, NSW. Two samples of charcoal from there were dated to 27 000–22 000 years ago, and were associated with flake and core artefacts, as well as the bones of 10 extinct and 13 extant species of mammals. Many of the bones show evidence of having been handled by people, being broken or charred, but the predominant articles of food were fish and crustacea. Other evidence for an association of people with extinct species are images of large animals on rock faces (Fig. 10.3, Plate 16), where the species no longer occur. The most dramatic is that of a large quadruped with a short trunk, which is thought to be the extinct Palorchestes. Rock carvings in the foothills of the Flinders Ranges (Basedow 1914) depict footprints similar in size and shape to actual footprints of Diprotodon found on the bed of Lake Callabonna 250 km away (Fig. 10.4). These images are among others depicting the footprints of Genyornis and smaller ones of extant emu, kangaroo and wombat. There are no dingo footprints, which supports the theory of the site’s great antiquity.

(a)

(b) (c)

Figure 10.4: Did people know Diprotodon? (a) Footprint of Diprotodon at Lake Callabonna, compared to (b) a reconstruction based on the skeleton, and (c) a rock drawing from Flinders Ranges, South Australia. From Basedow (1914). Photograph by David Ride.

Field and Fullagar (2001) have described disarticulated bones of marsupial megafauna with stone tools used for butchering intermingled with them at Cuddie Springs, NSW, which is dated at 36 400 years ago. There is no collagen in the bones that could provide accurate 14C dating but there is a suggestion of blood proteins on the stone tools, which is still to be analysed. This looks Marsupials and people: past and present 373

like very good evidence for people and megafauna being contemporaneous but none of the bone shows butchering marks and there were no articulated parts, so the bones might have washed in from somewhere else. Overlying the early strata at Lake Menindee is a much later bed containing point and blade artefacts of the later microlith culture. The ages of these sites are between 7000 and 4500 years ago and the important thing is that they contain none of the extinct species from the earlier strata, except the thylacine and the Tasmanian devil, Sarcophilus harrisii. Thus, between 10 000 and 6000 years ago there was a major cultural change in the human populations from heavy coarse stone implements suitable for flensing, to fine well-made microliths suitable for arming spears to use on smaller, fleeter game. Most of the large species had disappeared and the domes- tic dog or dingo had made its entry: the earliest date for this, based on good evidence in South Australia, is about 4000 years (Mulvaney and Kamminga 1999). Since the remains of dingo have not been found in other sites earlier than this it is most probably near to the time of its arrival. Calaby (1971) has argued, on the analogy of the fox, which crossed Australia in less than 70 years, that the dingo would have become widespread within 100 years of its introduction, and may have spread ahead of the people that brought it. Clearly it was not associated with the extinction of the megafauna but, as in New Guinea, it may have been a factor on continental Australia in the later extinction of the thylacine and the Tasmanian devil. Thus, there is no evidence of direct predation on the extinct megafauna in Australia by people, such as butchering sites or bones with stone projectiles embedded in them. This sort of evidence in North America, found in association with the remains of mammoths, mastodons and sloths, led to the theory of human overkill as the prime cause of the extinction of the North American megafauna. Nevertheless, the absence of evidence of butchering is not in itself suffi- cient to exclude human hunting as a prime cause of extinction. As Diamond (1999) has pointed out, the huge slaughter of bison in the 19th century in North America has left no trace, so that now only 150 years later, an archaeologist would be hard pressed to prove that it was caused by people.

How could people hunt Diprotodon to extinction? Apart from the question of whether there is direct evidence for humans causing the extinction of the megafauna, there is the uncertainty of a thinly scattered population of hunter–gatherers being able to extinguish many entire species. Choquenot and Bowman (1998) tried to estimate this for three classes of megafauna based on body mass of 250 kg, 500 kg and 1000 kg. They developed models for the rate of increase of each size class from the known values for contemporary large herbivores, and they estimated the hunting skills and meat requirements of the putative hunters, based on the hunting efficiency and use of buffalo (500–1000 kg) by present day Aboriginal people in northern Australian savanna woodland, where one buffalo provides meat for about 40 people for one week. They chose this habitat because in the last interglacial period the area extended across to New Guinea and was occupied by Diprotodon, and because it has been little affected by stock grazing. They estimated the rate at which the megafauna would have been killed in terms of the probable density of hunter–gatherers and the area that they would need to have covered each day in order to find sufficient large prey to support the group: the higher the density of the human population, the smaller the area that each person must cover in one day to find prey. Thus, at a density of about 0.1 person/km2 the minimum search area would be 70–100 ha/day for the smallest prey class and 120–200 ha/day for the largest prey class: at 0.4 person/km2 the search areas would be much smaller (Fig. 10.5). These estimates led them to predict the time to extinction in the range of 75–500 years. The shorter estimates require higher densities of people than at present exist and very high hunting efficiency, even when the prey became rare. 374 Life of Marsupials

Figure 10.5: Extinction isoclines for megafauna of body mass (···) 1000 kg, (---)500 kg and (—) 250 kg. Isoclines show the probability of extinction for each size group, above the line, in relation to human population density and rate of effective daily search. At a likely human density of 0.1/km2, each person would have to search 100 ha per day continuously to bring the largest species to extinction. After Choquenot and Bowman (1998).

The conclusions of Choquenot and Bowman (1998) are that hunter–gatherer communities could not have exterminated the largest species but might have had an effect on species in the size range of 250 kg. However, their models do not take into account the possibility that the hunters could hunt juveniles rather than adults, nor how the constraints of marsupial reproduction may have affected the rates of increase of the prey species. For instance, the model assumes that the rate of increase of Diprotodon would have been similar to buffalo, which can reproduce within a single year, but this is unlikely. If the rate of replacement was much slower and the young were taken for food, this would have had a more serious impact on the population than hunting adults. With the current knowledge of reproduction in living marsupials and the energetics of reproduction in the larger herbivore species, it is possible to make fairly informed estimates of the cost and rate of reproduction in the largest extinct species.

Did Diprotodon reproduction help its extinction? Diprotodon optatum was the largest species, with an adult mass of over 2 tonnes. In form it resembled a rhinoceros with a short trunk like that of a tapir and it browsed on shrubs and trees. All large marsupial herbivores produce one young at a very immature stage of develop- ment and then nurture it for an extended time, first in the pouch and then outside. The young is most vulnerable during the last phase of lactation and the period immediately afterwards. The reproductive performances of all marsupials are strongly correlated with maternal body mass, when plotted on a double log scale from 10 g to 100 kg (Russell 1982). By extrapolating by an additional order of magnitude to 1000 kg, estimates of the reproductive performance of Diprotodon can be made (Table 10.1). Marsupials and people: past and present 375

Table 10.1: Reproductive features of Diprotodon optatum, calculated by extrapolation from data for living marsupials in the appropriate fi gure from Russell (1982).

Reproductive parameter Diprotodon value Russell (1982) figure no. Body mass of adult (kg) 1000kg Main food source Grasses and herbage Litter size 1 Gestation length (days) About 60 (based on the grey kangaroo) Neonatal weight (g) 4 2a Eyes open (maturation of central 300 5 nervous system) (days) Pouch exit (days) 700–800 6 Mass at pouch exit 60 kg 8 Age at weaning 1000 days 7 Mass at weaning (10% adult mass) 100 kg Rate of growth, PE to weaning (g/day) 400 4a

The single young would probably have weighed between 2 and 5 g after a gestation of six to eight weeks. It is unlikely that such a large animal could have adopted the sitting posture that most of the smaller marsupials do at birth, but would have stood upright, as the grey kangaroo does. The pouch, as in wombats and the koala, was probably backwardly directed, so that the young would have travelled downwards or horizontally along the mother’s abdomen to enter it. As in all other marsupials, it would have attached to one teat and remained permanently attached to it until after its eyes opened and it was capable of thermoregulation. Extrapolating from Russell’s (1982) data the eyes of the pouch young would have opened at 300 days, which in macropods represents maturation of the central nervous system and occurs after one-third of lactation (see Chapter 2). Thermoregulation is achieved after two-thirds of lactation and so the young Diprotodon would have attained this stage at about 600 days, left the pouch at about 700 days and been weaned at 900 days. In the large kangaroos the mass of the young at weaning is about 20% of the mother’s mass. Extrapolating from Russell’s (1982, Fig. 8) the equivalent value for Diprotodon would have been 10% and so the young at weaning would have weighed 100–200 kg: this is considerably larger than an adult male kangaroo. However, unlike the grey kangaroo, in which the young is weaned at one year, the Diprotodon young would have taken nearly three years to reach independence. In the large kangaroos females reach sexual maturity soon after being weaned, while males do not reach maturity until double that age. Assuming a similar relationship in Diprotodon, females would have become sexually mature in year 4 and males not until year 8. If females began repro- ducing at four years and lived for 30, they could produce a maximum of eight young in a lifetime, or more importantly, four female young. However, this assumes that each reproductive effort was successful. As we know from reproduction in arid zone kangaroos, rainfall and availability of food during the second half of lactation is crucial to survival of the young (see Chapter 9). In all marsupials this is the most costly part of reproduction for the female, and in the case of the large kangaroos the critical period lasts for about five months: for Diprotodon that critical phase of lactation would have lasted for one year to 18 months. Since the species survived success- fully for several million years before people reached Australasia, the inference is that it lived in habitats where nutritious herbage was available in sufficient amount throughout the year rather than being seasonal. Its known distribution was the whole of eastern Australia and northern 376 Life of Marsupials

Tasmania and would have included the extensive lowlands that extended from Arnhem Land and Cape York to present day New Guinea. What would have been the impact of people on such a slow breeding species? Flannery (1994) has suggested that the main threat came from their lack of fear of the new predator, which meant they were readily killed off. The large young would have been easier prey and still a substantial source of food and, since the rate of replacement was so slow, this would have had an equally serious impact on the population. The other effect of human occupation was the introduction of fire, and Flannery (1994) has argued that this radically altered the vegetation of much of Australia, changing rainforest to fire-resistant species of Eucalyptus. Firing of the grasslands or herbage, on which the female Diprotodon depended to maintain lactation, could have had an impact as severe as the hunting of the young, so it is not necessary for people to have extermi- nated the megafauna directly by hunting them to extinction: alteration of their habitat could have had just as profound an effect on their survival. On current evidence (Roberts et al 2001) the overlap of the megafauna and people was of the order of 15 000 years. Even if people caused the extinction of the Australasian megafauna 45 000 years ago, it was a drawn-out affair, suggestive of a slow failure of the large species to reproduce successfully. It was not a blitzkreig as has been described for the North American megafaunal extinction (Martin and Klein 1984). This is consonant with the lack of any evidence in Australia of butchering sites, such as those described in North America.

European impact on Australian marsupials When we move from considering pre-historic extinctions of marsupials to the post-European period the rate of extinction greatly increases. We must ask why, if marsupials were able to thrive in Australia for more than 50 million years, did so many species succumb in less than 200 years of European occupation? Indeed, the rate of loss was much faster than this implies, because in many places the decline to extinction occurred within 20 or 30 years of an area being settled. The primary reason was competition for the land with agriculture. Agriculture simplifies an ecosystem in order to divert radiant energy into productive plants and animals and away from unproductive species. Inedible plants become weeds and compet- ing herbivores become pests. Pests and agriculture have been linked since antiquity and in this sense there were no weeds or pests in Australia before European occupation: Europeans brought the concept to Australia. The indigenous plants and animals, long isolated from the rest of the world, were vulnerable to the transformation of the landscape and to the exotic species that were introduced. As Europeans moved with their livestock across the continent to create a thriving agriculture in Australia, the indigenous fauna declined rapidly and waves of extinction followed. The earli- est losses were in eastern Australia (Dickman et al 1993), followed by southern Australia, while many arid zone species remained relatively abundant through the 1920s, or later in more remote regions. Then they, in turn, declined rapidly with the advent of pastoralism and the spread of the rabbit and the fox. For most of the continent the major changes occurred in the 20th century and are continuing in the southwest and northwest of the continent as settlement of these regions intensifies (Recher and Lim 1990). Land clearance and pastoralism, deliberate destruction of perceived pests, changes in Aboriginal living and the spread of introduced predators, especially the fox, all contributed in different parts of the continent to the decline or extinction of the native animals. Marsupials and people: past and present 377

Effects of land clearance and over stocking After 1830, as settlement spread across eastern Australia, land clearance was carried out at a rate faster than on any other continent, and by 1890 sheep numbers rose to a size never again attained. In western New South Wales the peak of 12 million sheep was reached in 1890 and, after a prolonged dry period from 1892 and a slump in the wool price, had crashed to 4 million sheep by 1895 and stabilised at about 5 million thereafter (Fig. 10.6).

Figure 10.6: The changing number of sheep in western New South Wales between 1860 and 1930, and, for the central slopes only, the number of bounties paid on rat kangaroos and foxes between 1888 and 1920. Rabbits reached this area in 1877 and sheep numbers peaked in 1890 and then stabilised at half that number. After Caughley et al (1987a) and Short (1998).

The sheep population behaved like any other herbivore occupying an empty but favourable niche, for instance, red deer in New Zealand (Caughley 1983). Rapid growth of the invader population overshoots the resources, leading to a population crash, with severe consequences for the native vegetation, followed by a cycle with a lower stable limit. In Australia the growth of the sheep population was aided by the wholesale conversion of the savanna woodland to pasture. But as the sheep population continued to rise, the palatable grasses were eaten down, the fuel bed was reduced, and unpalatable woody shrubs and young trees increased (Noble 1997). In addition, the introduction of fence wire in the 1870s (Rothery 1970) enabled stocking rates to be imposed that far exceeded the sustainable capacity of the soils and vegetation and contrib- uted to the devastating crash of the sheep population in the 1890s. This sequence of events had profound, long-term effects on the land: and this was predicted by people at the time. The crisis in the pastoral industry led to the Royal Commission of 1901 into the Western Lands Division of New South Wales. It is salutary to read the submissions by many of the landowners to the Commission, who admitted that they had been responsible for the calamity by their unsustainable practices. For instance, R. Griffiths, Manager of Nymagee Station, which carried 25 000 sheep in 1874 but only 3000 in 1900, wrote (Lunney 1994): 378 Life of Marsupials

If people went into that country now, with the accumulated experience of the last 16 years, probably the same mistakes would not be made in over-estimating the carrying capacity of the country. I spent £5000 in ringing and scrubbing [ringbarking trees and clearing shrubs] the leasehold area. That showed an immediate improvement in the growth of grass; but the rabbits came along, and assisted the sheep to eat the country out, and it has never recovered since; and, in my opinion, it never will. That is practically the condition of the whole Western Division.

Tiver and Andrew (1997) concluded from their study in eastern South Australia that the overgrazing events that occurred 100 years ago are still affecting plant regeneration today: of all the vertebrate herbivores, sheep have by far the most adverse effects on the regeneration of trees and shrubs in eastern South Australia. Their results are at odds with current conservation literature, which identifies rabbits and, to a less extent, goats as the most important factors in preventing the regeneration of woody perennial species. Although rabbits can cause damage for short periods when herbage is scarce, this does not affect the long-term regeneration of most plant species; neither do unrestrained feral goats cause the damage that they do in experimental enclosures. Throughout much of the rangelands, rabbit populations have again been brought to

Figure 10.7: Habitat fragmentation. The shaded areas are native vegetation in the Kellerberin district, Western Australia, and they show that most of it was cleared for wheat between 1920 and 1984. From Saunders (1994). Marsupials and people: past and present 379

a low level by the spread of the rabbit calici virus, and the native vegetation is showing spectac- ular recovery. However, this will not be sustained without changes to livestock management, especially reduced stocking rates. There is a rare opportunity to achieve a long-lasting recovery, but it will require major changes in the perception of what are sustainable stocking rates by local councils, pasture protection boards, lending agencies and land managers. Regrettably, it is more likely that landholders will increase stocking rates to exploit the new growth. Alteration of the plant cover, through land clearing and overstocking, was the main cause of extinction of small marsupials of the grasslands and dry woodlands because the stock ate the cover and food from over the small mammals, leading to their demise (Newsome 1971). The smaller species were not directly attacked but, through loss of food resources and shelter from predators and extreme temperatures, their decline was rapid: species such as the bilby and the pig-footed bandicoot were initially abundant in the southern regions of Australia, but it was already being remarked in the mid 19th century that they were disappearing from Victoria, as the land was occupied for sheep (Krefft 1862). A similar pattern occurred in the western division of New South Wales, where five species that were present before European settlement had disappeared by 1857 and 11 additional species by 1900 (Dickman et al 1993). This continued in the 20th century in all areas of settlement as the native woodland vegetation was inexorably cleared for agriculture: in Western Australia more than half the remaining woodland in the southwest was cleared for wheat and sheep after 1960 (Fig. 10.7). Another change in pastoralism was the extension of stock water- ing points throughout the semi-arid country (James et al 2000), which brought grazing stock on to land hitherto unavailable to stock and so extended their dire effects on wildlife.

Deliberate destruction of native mammals During the first half of the 19th century most native mammals were regarded as pests that interfered with farm animals or ate crops. For example, in the Bega district of southern New South Wales after settlement in 1830, initial clearing of the valley bottoms for agriculture led to an increase in the small marsupials, such as bandicoots, pademelons and wallabies, which were regarded as pests and were actively destroyed (Lunney and Leary 1988). As forest clear- ing extended up the hillsides and the natural cover was removed the marsupials declined and the introduced hare came to be regarded as the major pest species. By 1900 the hare also was in decline and the rabbit replaced it as the major pest species, followed by the fox (Fig. 10.6). In other places around the southern settled areas battues were held to destroy large numbers of wallabies, rat kangaroos and bandicoots. In New South Wales nearly three million rat kanga- roos were presented for bounty between 1880 and 1920 (Short 1998). These comprised four species recorded by Krefft as being present along the Murray River in 1860 namely, Bettongia lesueur, Bettongia penicillata, Bettongia gaimardi and Aepyprymnus rufescens. From 1892 to 1895 250 000 bounty payments were made each year (Fig. 10.6), thereafter the numbers declined until less than 2000 payments were made in 1915. This destruction was taking place at the same time as the massive effort to destroy rabbits that had reached New South Wales in the 1870s. Many more rat kangaroos were affected by trapping and poisoning directed primarily at rabbits, and by the netting fences introduced in 1895. Despite this huge slaughter the rat kangaroos survived until foxes appeared in 1900. Foxes entered New South Wales from the south, supported by the vast numbers of rabbits, and would have taken rat kangaroos almost incidentally. However, the decline in bounty payments for rat kangaroos coincided in each part of the state with the arrival of foxes: by 1900 in the south and by 1915 in the north, which suggests that foxes caused the final extinction of the three species of Bettongia. The fourth species survives in northern New South Wales and Queensland in habitat not frequented by foxes. In Tasmania the thylacine was regarded as a sheep killer by pastoralists as early as 1830 and private companies offered bounties. In 1888 the Tasmanian government, under pressure from 380 Life of Marsupials

farmers suffering from the crisis in the wool industry, declared it to be a pest and offered a bounty (Guiler 1985, Paddle 2001). From that year until 1910 one pound (equivalent to a labourer’s weekly wage) was paid for each scalp. About 100 adults were killed each year until 1905 when the numbers quickly fell and none was presented after 1909 (Fig. 10.8). However, many more were killed for bounties given by farmers and private companies, so that the official total of 2184 is a gross underestimate of the number actually killed (Paddle 2001). The species lingered on until the 1930s but the earlier slaughter was surely the main cause of the species’ extinction.

160

140 Thylacine 120

100

80

60

40 Number of thylacines killed Number of thylacines

20

0 1890 1895 1900 1905 1910 Year

Figure 10.8: Total of adult (grey) and juvenile (black) thylacines presented each year in Tasmania during the Government bounty scheme between 1888 and 1912. After Guiler (1985).

The only other species brought to extinction by direct predation was the Toolache wallaby, Macropus greyi, of South Australia, which was a favourite species for coursing by local hunting clubs. Wallabies continued to be hunted in Tasmania until the 1970s but those species have survived. Large numbers of kangaroos are shot each year from all States except Victoria and Tasmania, in part for ‘damage mitigation’ and part for the domestic and export market, but the populations of these species are unaffected and the harvest is sustainable (see Chapter 9).

Rabbits and foxes: the destructive duo In Australia the wild rabbit has been regarded as the major mammalian pest for the past 140 years. The rapid spread of the rabbit across the continent between 1860 and 1900 and the conse- quent devastating effect on the pastoral industry is part of Australian folklore. This was much faster than in Europe, where it was introduced by the Romans, and was the slowest of 28 intro- duced mammal species (Thompson and King 1994). What peculiar factor or factors in Australia so spectacularly favoured the rabbit? Rabbits came with the first European settlers, and wild-type rabbits were released more than 30 times in widely scattered locations before 1859 (Myers 1986). At least one of these early intro- Marsupials and people: past and present 381

ductions on the Cooks River near Sydney, NSW, has survived as a genetically distinct population (Phillips et al 2002). Why then did it take nearly 60 years for rabbits to break out and why did they then spread across the land so fast? The main invasion began from wild rabbits released at Barwon Park, near Geelong, Vic., in 1859, and the conventional view is that their success was because they were wild rabbits rather than the domestic rabbits that were brought previously. But this is clearly not so if wild rabbits became established near Sydney years earlier. The rabbit is a biological opportunist that responds to favourable conditions, especially the short-cropped pasture brought about by grazing stock: in this sense it is a human commensal, like rats and mice. Perhaps wild rabbits failed to spread across Australia before 1860 because favourable conditions did not develop until the transformation of the woodlands to grazing pasture. Extensive land clearance, widespread destruction of native predators and the severe overstocking that occurred after 1870 accelerated these changes and so favoured the spread of the rabbit across Australia between 1870 and 1920. Because the rabbit burgeoned after sheep had already reached very high numbers, it was probably part of the response to, rather than the cause of, the decline in productivity that devastated the pastoral industry at the beginning of the 20th century: while it has always been convenient to blame the rabbit for what happened, the real despoilers of the Australian vegetation and wildlife were unsustainable numbers of domestic sheep, as in the Pilbara of Western Australia (see Chapter 9).

Changes in Central Australia Some native species thought to be extinct since the early 20th century actually persisted in the central deserts until the 1940s or 1950s. Rabbits entered from the southeast in the 1890s and became widespread shortly afterwards in the southern areas of Central Australia. However, they never became common in the northern parts, except in particular habitats and none of the native species were restricted by them. According to the Aboriginal people interviewed by Burbidge and colleagues (1988), foxes became established in the area after the native species had gone, but cats had been present in the central deserts from a much earlier period and came from the west coast. The first European explorers to the Centre encountered cats and there is some suggestion that they were present before European settlement, possibly introduced to Australia by trading vessels from Asia. Grazing stock were never introduced to the desert country because of the lack of water, so they cannot have been the cause of the later decline, and several other factors have been considered (Burbidge et al 1988, Morton 1990). The decline began in the southern parts of the range and later in the north, consistent with the time when Aboriginal people moved from their traditional lands to European settlements. This led to the idea that the fire regimes practiced by Aboriginal people had been critical to the survival of the smaller marsupial and placental species. Aboriginal use of fire resulted in a tight mosaic of small areas differing in the time since fire, which provided a diversity of shelter and feeding areas for the mammals. With the departure of the people natural fires replaced the human fire regime, which changed the mosaic of vegeta- tion succession to a more uniform pattern that was more susceptible to widespread hot fires. While the very small species could escape into cracks and survive, and the largest species could move elsewhere, the intermediate size range could not survive extensive fires, or the vegetational changes that followed them. This critical weight group has been the most profoundly affected, with all now extinct or reduced to tiny enclaves in their former habitat. Nevertheless, it does seem extraordinary that a suite of mammals that had lived in that environment for millions of years before human occupation of Australia were now unable to survive without human fire regimes. Morton (1990) attempted to reconcile these and other factors in a model that could explain the extinction of the critical weight range mammal species in the central desert regions. First, the 382 Life of Marsupials

soils are unusually infertile with low phosphorus and nitrogen, and the rainfall is uncertain and extremely variable. The latter has the effect of causing flooding and washouts, which redistrib- ute nutrients to particular areas, making them more fertile. These two factors result in a highly patterned environment, comprising patches of fertile soil, with dependable and digestible plant production, dotted through a vast infertile landscape supporting relatively unpredictable and unusable plant growth. Second, the mammal species that evolved in this environment had relatively low rates of increase and low basal metabolic rates. In favourable seasons the populations would increase and spread out onto the less fertile parts, while in dry seasons small populations would persist in the more productive patches. Body size of the species is an important factor in long-term survival. The very small species (<500 g) cannot survive on herbage and instead rely on concentrated foods such as roots and invertebrates; they are also small enough to escape the extremes of temperatures in deep cracks or crevices or under rocks and so can persist in very small produc- tive patches and recolonise in subsequent favourable times. The largest species of kangaroo are exclusively herbivorous but they can travel large distances from one productive patch to another or to permanent water. The intermediate group of 500 g to 5 kg body mass are the least likely to persist: they are too large to escape down crevices and thus survive on a small productive patch, but not mobile enough to move to another patch when conditions deteriorate. By contrast, birds of the same body mass are mobile and so can persist. The indigenous species had been able to persist in the long term, despite severe droughts, until the arrival of other competitors. In the better country the impact of grazing stock was to deplete the herbage in the productive patches, especially during drought and/or overstocking, so that source populations of the medium-sized species were unable to survive to the next favour- able season. The extension of watering points through much of the inland further aggravated this. In the desert areas, where stock did not penetrate, rabbits spread further north in good seasons and had the same effect on the source populations of indigenous species. With the next drought the rabbits died out and some native species survived longer in the north than in the south. Foxes followed the rabbit but preyed on native species when encountered. The residual question raised by Morton’s hypothesis (1990) is why rabbits were able to survive in the inland environment better than the species that had evolved there? The following factors provide a likely answer: their high potential rate of increase enabled rabbits to respond much more rapidly to favourable conditions and exploit the herbage to the maximum, and so leave source populations to survive the next drought cycle. Also, because the rabbit is essentially a commensal species with the sheep, large source populations in the pastoral regions could more readily recolonise depleted areas when favourable conditions returned. While the native species may have been able to persist through one or two cycles of drought and plenty in competition with rabbits, their source populations were inexorably declining or becoming more scattered as productive patches lost their populations in competition with rabbits. Thus, while changes in the vegetation and pattern of fires were probably important, the arrival of rabbits and foxes provided the final act of extinction. The species that fall into the critical weight range now survive only in particular areas where food resources are still adequate and predators are rare or absent. These are the high rainfall areas of northern Australia and the southeast and southwest of the continent and on offshore islands of sufficient size to carry a viable population. These include Tasmania and Kangaroo Island. Marsupials and people: past and present 383

Rehabilitation of endangered species As the indigenous marsupials have declined to relict populations, there have in recent years been two responses: protecting the remaining habitat from further change and reintroducing species to part of their former range. Several attempts have been made to protect areas of natural habitat with vermin-proof fences so that the rare native species can flourish inside. The difficulty for long-term protection is maintaining the integrity of the fences, and where such attempts have failed it has been because foxes, cats or domestic dogs have got in. Private parks have been more successful because the income flow from visitors can support continuous vigilance on the perimeter fences. However, to attract visitors the density of native species must be maintained at levels above those in a natural habitat. While small parks provide the opportunity for people to see rare species, this is not the same as ensuring the survival of the genetic diversity of the species in self-perpetuating popula- tions. For this, very large areas of prime habitat for the species concerned need to be set aside as dedicated conservation reserves. The difficulty of this approach is that the areas required for effective protection are so large that it is politically unrealistic to expect that to happen. In Western Australia a different approach has been adopted, which exploits the high tolerance of native marsupials for sodium fluoroacetate (1080) because of the prevalence of plants with high natural levels of the chemical (see Chapter 9), and the great susceptibility of foxes to this poison. By regular aerial distribution of meat baits containing sufficient 1080 to poison a fox but not affect any native species, large tracts of forest and woodland habitat have been cleared of foxes and the native species have recovered to their former abundance. This approach is only effective in the special circumstances of Western Australia and it depends on long-term, substantial resources to distribute the poison sufficiently frequently to hold foxes at low densities. The other approach has been to introduce rare species to isolated islands that are within the former range of the species. Short et al (1992) analysed the success of 25 attempts to establish such populations, some on islands and some in large protected enclosures. The introductions to islands were far more successful (60%) than those to mainland enclosures (11%): on islands that had no mammalian predators the success was 82%. Hence, successful reintroduction requires the absence of mammalian predators. Since 1992 four other attempts at reintroduction of rare species of marsupial have been attempted to mainland sites (Serena 1994): all failed within a few weeks of release and almost all mortality was due to exotic predators.

Species that have benefited from change A few species have increased their range and abundance as a result of human alteration of the habitat. The interaction of the four species of kangaroo with sheep and cattle was discussed in Chapter 9, where it was noted that the food preferences of the several species only overlap under unfavourable conditions, such as drought or overstocking. However, in the semi-arid regions of Australia pastoral activities actually favour kangaroos by making water available at stock dams and by clearing woodland. For instance, the extensive grassland community surrounding the Murrumbidgee floodplain is maintained by grazing, and since red kangaroos prefer grassland to woodland, land clearance has improved the habitat for them and their range has increased. Likewise, in Central Australia red kangaroos were formerly restricted to the slopes and plains near permanent water but, when bores were put down for cattle, this enabled the kangaroo populations to increase and spread out onto the drier country. And cuscuses in New Guinea and surrounding islands, brushtail possums in New Zealand and the Virginia opossums in California have benefited from changes brought about by people. 384 Life of Marsupials

Conclusions People probably reached Australia before the megafauna became extinct. Whether the extinction was caused by human factors, such as hunting and habitat alteration by burning, or was the result of the extensive dry period after the last Ice Age, remains an open question. However, human occupation did not seriously affect the survival of the smaller marsupial fauna in Australia, New Guinea or South America. The overwhelming factor in the extinctions or range contractions of marsupials in Australia has been European occupation. Widespread clearing of the woodland vegetation and, initially, extreme overstocking with sheep held behind fences, profoundly altered the native vegetation on which the marsupials directly or indirectly depend. Rabbits and foxes followed, leading to further disappearances of small marsupials, aided by widespread destruction of native species that were perceived to compete with stock for the grass. The marsupial features of low metabolic rate and slow reproduction were pre-adapted to the special conditions of Australia – leached soils of low fertility and highly unpredictable climate with long cycles of drought and plenty – and marsupials thrived in the Australian environment for 50 million years. The rapid extinction of so many of the medium-sized species in the past 150 years, and the continuing decline of the surviving species, is a potent sign that the replacement system may not be sustainable in the long term. Marsupial species are like the canary in the coal mine: if we knew better how marsupials survived for so long and why they died out so rapidly, we might know better how to live in this country for the long term. In 150 years sheep and rabbits have caused profound and possibly irreversible changes to the environment that may be inimical to their long-term survival. And ours. References 385

References

Abbie AA 1939. A masticatory adaptation peculiar to some diprotodont marsupials. Proceedings of the Zoological Society of London 109, 261–269. Adamski FM, Demmer J 2000. Immunological protection of the vulnerable marsupial pouch young: two periods of immune transfer during lactation in Trichosurus vulpecula (brushtail possum). Developmental and Comparative Immunology 24, 491–502. Aitkin L 1998. Hearing – the brain and auditory communication in marsupials. p. 117. Springer, Berlin. Alcorn GT, Robinson ES 1983. Germ cell development in female pouch young of the tammar wallaby (Macropus eugenii). Journal of Reproduction and Fertility 67, 319–325. Alexander RM, Vernon A 1975. The mechanics of hopping by kangaroos (Macropodidae). Journal of Zoology, London 177, 265–303. Allen RB, Fitzgerald AE, Efford MG 1997. Long-term changes and seasonal patterns in possum (Trichosurus vulpecula) leaf diet, Orongorongo valley, Wellington, New Zealand. New Zealand Journal of Ecology 21, 181–186. Alvarez W (1998). T. rex and the Crater of Doom. Vintage Books, New York. Amoroso EC, Perry JS 1975. The existence during gestation of an immunological buffer zone at the interface between maternal and foetal tissues. Philosophical Transactions of the Royal Society B 271, 343–361. Aplin KP, Woolley PA 1993. Notes on the distribution and reproduction of the Papuan bandicoot Microperoryctes papuensis (Peroryctidae, Peramelomorphia). Science in New Guinea 19, 109–112. Archer M 1976. The dasyurid dentition and its relationships to that of didelphids, thylacinids, borhyae- nids (Marsupicarnivora) and peramelids (Peramelina: Marsupialia). Australian Journal of Zoology Supplementary Series 39, 1–34. Archer M, Flannery TF, Molnar RE 1985. First Mezosoic mammal from Australia – an early Cretaceous monotreme. Nature, London 318, 363–366. Archer M, Godthelp H, and Hand SJ 1993. Early Eocene marsupial from Australia. Kaupia 3, 193–200 Archer M, Arena R, Bassarova M, Black K, Brammall J, Cooke B, Creaser P, Crosby K, Gillespie A, Godthelp H, Gott M, Hand SJ, Kear B, Krikmann A, Mackness B, Muirhead J, Musser A, Myers T, Pledge N, Wang Y, Wroe S 1999. The evolutionary history and diversity of Australian mammals. Australian Mammalogy 21, 1–45. Arnold J, Shield J 1970. Oxygen consumption and body temperature of the Chuditch (Dasyurus geoffroii). Journal of Zoology, London 160, 391–404. Arrese C, Runham PB 2002. Redefining the activity pattern of the honey possum (Tarsipes rostratus). Australian Mammalogy 23, 169–172. Ashwell KWS, Marotte LR Lixin L, Waite PME 1996. Anterior commissure of the wallaby (Macropus eugenii): adult morphology and development. Journal of Comparative Neurology 366, 478–494. Atramentowicz M 1982. Influence du milieu sur l’activite locomotrice et la reproduction de Caluromys philander (L.). Revue d’Ecologie (Terre et Vie) 36, 373–395. Atramentowicz M 1988. La frugivore opportuniste de trios marsupiaux didelphidés de Guyane. Revue de Ecologie 43, 47–57 Atramentowicz M 1992. Optimal litter size: does it cost more to raise a large litter in Caluromys philander? Canadian Journal of Zoology 70, 1511–1515. Atramentowicz M 1995. Growth of pouch young in the bare-tailed woolly opossum, Caluromys philander. Journal of Mammalogy 76, 1213–1219. Austad SN 1993. Retarded senescence in an insular population of Virginia opossums (Didelphis virgin- iana). Journal of Zoology, London 229, 695–708. Baggott LM, Moore HDM 1990. Early embryonic development of the grey short-tailed opossum, Monodelphis domestica, in vivo and in vitro. Journal of Zoology, London 222, 623–629. Bailey PT 1971. The red kangaroo, Megaleia rufa (Desmarest), in northern New South Wales. I. Movements. CSIRO Wildlife Research 16, 11–28. 386 Life of Marsupials

Bailey PT, Best L 1992. A red kangaroo, Macropus rufus, recovered 25 years after marking in north-western New South Wales. Australian Mammalogy 15, 141. Bailey PT, Martensz PN, Barker R 1971. The red kangaroo, Megaleia rufa (Desmarest), in northern New South Wales. II. Food. CSIRO Wildlife Research 16, 29–39. Barlow ND, Kean JM, Briggs CJ 1997. Modelling the relative efficacy of culling and sterilisation for controlling populations. Wildlife Research 24, 129–141. Bartholomew GA, Hudson JW 1962. Hibernations, estivation, temperature regulation, evaporative water loss, and heart rate of the pigmy possum, Cercaertus nanus. Physiological Zoology 35, 94–107. Basden K, Cooper DW, Deane EM 1997. The development of the lymphoid tissues of the tammar wallaby, Macropus eugenii. Reproduction, Fertility and Development 9, 243–254. Basedow H 1914. Aboriginal rock carvings of great antiquity in South Australia. Journal of the Royal Anthropological Institute 44, 195–211. Bathgate RAD, Parry LJ, Fletcher TP, Shaw G, Renfree MB, Gemmell RT, Sernia C 1995. Comparative aspects of oxytocin-like peptides in marsupials. Advances in Experimental Medicine and Biology 395, 639–655. Baudinette RV 1989. The biomechanics and energetics of locomotion in Macropodoidea. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 245–253. Surrey Beatty, Sydney. Baudinette RV 1991. The energetics and cardiorespiratory correlates of mammalian terrestrial locomo- tion. Journal of Experimental Biology 160, 209–231. Baudinette RV, Gannon BJ, Runciman WB, Wells S 1987. Do cardiorespiratory frequencies show entrain- ment with hopping in the tammar wallaby? Journal of Experimental Biology 129, 251–263. Baudinette RV, Gannon BJ, Ryall RG, Frappell PB 1988a. Changes in metabolic rates and blood respi- ratory characteristics during pouch development of a marsupial, Macropus eugenii. Respiration Physiology 72, 219–228. Baudinette RV, Halpern EA, Hinds DS 1993. Energetic cost of locomotion as a function of ambient temperature and during growth in the marsupial Potorous tridactylus. Journal of Experimental Biology 174, 81–95. Baudinette RV, Runciman SIC, Frappell PF, Gannon BJ 1988b. Development of the marsupial cardiore- spiratory system. In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. pp. 132–147. Springer, Basel. Baudinette RV, Snyder GK, Frappell PB 1992. Energetic cost of locomotion in the tammar wallaby. American Journal of Physiology (Regulatory Integrative Comparative Physiology) 262, R771–R778. Baverstock PR, Krieg M, Birrell J 1990. Evolutionary relationships of Australian marsupials as assessed by albumin immunology. In Mammals from Pouches and Eggs. Eds JAM Graves, RM Hope and DW Cooper. pp. 131–145. CSIRO Publishing, Melbourne. Bear IJ, Thomas RG 1966. Genesis of petrichor. Geochemica et Cosmochemica Acta 30, 869–880. Bell BD 1981. Breeding condition of possums, Trichosurus vulpecula in the Orongorongo Valley, near Wellington, New Zealand, 1966–1975. In Marsupials in New Zealand. Ed. BD Bell. Zoological Publications from Victoria University of Wellington 74, 87–139. Bellingham PJ, Wiser SK, Hall GMH, Alley JC, Allen RB, Suisted PA 1999. Impacts of possum browsing on the long-term maintenance of forest biodiversity. Science for Conservation 103, 1–59. Bennett AF, Baxter BJ 1989. Diet of the long-nosed potoroo, Potorous tridactylus (Marsupialia: Potoroidae), in south-western Victoria. Australian Wildlife Research 16, 263–272. Bentley PJ, Shield JW 1962. Metabolism and kidney function in the pouch young of the macropod marsupial Setonix bracyurus. Journal of Physiology, London 164, 127–137. Bergallo HG, Cerqueira R 1994. Reproduction and growth of the opossum Monodelphis domestica (Mammalia: Didelphidae) in northeastern Brazil. Journal of Zoology, London 232, 551–563. Berger PJ 1966. Eleven-month ‘Embryonic diapause’ in a marsupial. Nature 211, 435–436. Berger PJ 1970. The reproductive biology of the tammar wallaby, Macropus eugenii, Desmarest. PhD Thesis, Tulane University. New Orleans, USA. Berry PF 1991. Ecological Survey of Abydos-Woodstock Reserve, Western Australia. 1. Introduction. Records of the Western Australian Museum 37, 2–8. Biggers JD 1967. Notes on reproduction of the woolly opossum (Caluromys derbianus) in Nicaragua. Journal of Mammalogy 48, 678–680. Birdsell JB 1993.Microevolutionary patterns in Aboriginal Australia. A gradient analysis of clines. p. 469. Oxford University Press, Oxford. References 387

Block M 1964. The blood forming tissues and blood of the newborn opossum (Didelphys virginiana). 1. Normal development through about the one hundredth day of life. Ergebnisse der Anatomie und Entwicklungsgeschichte 37, 237–366. Böer M 1998. Observations on reproduction in the common wombat Vombatus ursinus in captivity. In Wombats. Eds RT Wells and PA Pridmore. pp. 129–146. Surrey Beatty, Chipping Norton. Bradford MG, Harrington GN 1999. Aerial and ground survey of sap trees of the yellow-bellied glider (Petaurus australis) near Atherton, North Queensland. Wildlife Research 26, 723–729. Bradley AJ 1997. Reproduction and life history in the red-tailed phascogale Phascogale calura (Marsupialia: Dasyuridae): the adaptive-stress senescence hypothesis. Journal of Zoology 241, 739–755. Bradley AJ, Stoddart DM 1997 Plasma androgen, social position and response to GnRH in the marsupial sugar glider, Petaurus breviceps (Marsupialia: Petauridae). Journal of Zoology, London 241, 579–87. Bradshaw F, Everett L, Bradshaw D 2000. On the rearing of honey possums. The Western Australian Naturalist 22, 281–288. Bradshaw FJ, Bradshaw SD 2001. Maintenance nitrogen requirement of an obligate nectarivore, the honey possum, Tarsipe rostratus. Journal of Comparative Physiology B 171, 59–67. Bradshaw SD, Bradshaw FJ 1999. Field energetics and the estimation of pollen and nectar intake in the marsupial honey possum, Tarsipes rostratus, in heathlands of South-Western Australia. Journal of Comparative Physiology B 169, 569–580. Bradshaw SD, Bradshaw FJ 2002. Short-term movements and habitat use of the marsupial honey possum (Tarsipes rostratus). Journal of Zoology, London 258, 343–348. Braithwaite LW 1983. Studies on the arboreal marsupial fauna of eucalypt forests being harvested for woodpulp at Eden, NSW. I. The species and distribution of animals. Australian Wildlife Research 10, 219–229. Braithwaite LW 1984. The identification of conservation areas for possums and gliders within the Eden woodpulp concession district. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 501–508. Australian Mammal Society, Sydney. Braithwaite LW, Belbin L, Ive J, Austin M 1993. Land use allocation and biological conservation in the Batemans Bay forests of New South Wales. Australian Forestry 56, 4–21. Braithwaite LW, Dudzinski ML, Turner J 1983. Studies on the arboreal marsupial fauna of eucalypt forests being harvested for woodpulp at Eden, NSW. II. Relationship between the fauna density, richness and diversity, and measured variables of the habitat. Australian Wildlife Research 10, 231–247. Braithwaite RW 1979. Social dominance and habitat utilization in Antechinus stuartii (Marsupialia). Australian Journal of Zoology 27, 517–528. Brammall J, Archer M. 1999. Living and extinct Petaurids, Acrobatids, Tarsipedids and Burramyids (Marsupialia): relationships and diversity through time. Australian Mammalogy 21, 24–25. Breed WG 1994. How does sperm meet egg? – In a marsupial. Reproduction Fertility and Development 6, 485–506. Breed WG 1996. Egg maturation and fertilization in marsupials. Reproduction Fertility and Development 8, 617–643. Brocke RH 1970. The winter ecology and bioenergetics of the opossum, Didelphis marsupialis, as distribu- tional factors in Michigan. PhD Thesis. Michigan State University, East Lansing. Brockie R 1992. A Living New Zealand Forest. pp. 172. David Bateman, Auckland. Brockie RE, Bell BD, White AJ 1981. Age structure and mortality of possums Trichosurus vulpecula populations from New Zealand. In Marsupials in New Zealand. Ed. BD Bell. Zoological Publications from Victoria University of Wellington 74, 63–86. Brown GD 1984. Thermoregulation in the common wombat (Vombatus ursinus) in an alpine environ- ment. In Thermal Physiology. Ed JRS Hales. pp. 331–334. Raven Press, New York. Buffon GLL 1749. History of Quadrupeds [derived from vols IV–XV of L’Histoire naturelle, generale et particuliere]. Paris. Burbidge AA 1990. The value of Western Australian islands as biological reservoirs and the development of management priorities. In Australian and New Zealand Islands: nature conservation values and management. Ed. AA Burbidge. CALM Occasional Paper 2/89, 17–24. Burbidge AA, Johnson, KA, Fuller PJ, Southgate RI 1988. Aboriginal knowledge of the mammals of the Central Deserts of Australia. Australian Wildlife Research 15, 9–39. Burk A, Springer MS 2000. Intergeneric relationships among Macropodoidea (: Diprotodontia) and the chronicle of kangaroo evolution. Journal of Mammalian Evolution 7, 213–237. 388 Life of Marsupials

Burk A, Westerman M, Kao DJ, Kavanagh JR, Springer MS 1999. An analysis of marsupial interordinal relationships based on 12S rRNA, tRNA valine, 16S rRNA and cytochrome b sequences. Journal of Mammalian Evolution 6, 317–334. Burk A, Westerman M, Springer M 1998. The phylogenetic position of the musky rat-kangaroo and the evolution of bipedal hopping in kangaroos (Macropodidae: Diprotodontia). Systematic Biology 47, 457–474. Caceres NC 2002. Food habits and seed dispersal by the white-eared opossum, Didelphis albiventris, in southern Brazil. Studies on Neotropical Fauna and Environment 37, 97–104. Caceres NC, Monteiro-Filho ELA 2001. Food habits, home range and activity of Didelphis aurita (Mammalia, Marsupialia) in a forest fragment of southern Brazil. Studies on Neotropical Fauna and Environment 36, 85–92. Calaby JH 1965. Early European description of an Australian mammal. Nature 205, 516–517. Calaby JH 1971. Man, fauna, and climate in Aboriginal Australia. In Aboriginal Man and Environment in Australia. Eds DJ Mulvaney and J Golson. Australian National University Press, Canberra. Caldwell WH 1884. On the arrangement of the embryonic membranes in marsupial animals. Quarterly Journal of Microscopical Science 24, 655–658. Calvert SJ, Holland RAB, Gemmell RT 1994. Respiratory properties of the neonatal blood of the common brushtail possum (Trichosurus vulpecula). Physiological Zoology 67, 407–417. Campbell DJ 1990. Changes in structure and composition of a New Zealand lowland forest inhabited by brushtail possums. Pacific Science 44, 277–296. Campeau-Peloquin A, Kirsch JAW, Eldridge MDB, Lapointe F-J 2001. Phylogeny of the rock-wallabies, Petrogale (Marsupialia: Macropodiae) based on DNA/DNA hybridisation. Australian Journal of Zoology 49, 463–486. Cannon JR, Bakker HR, Bradshaw SD, McDonald IR 1976. Gravity the sole navigational aid to the newborn quokka. Nature 259, 42. Carthew SM 1994. Foraging behaviour of marsupial pollinators in a population of Banksia spinosa. Oikos 69, 133–139. Carthew SM, Goldingay RL, Funnell DL 1999. Feeding behaviour of the yellow-bellied glider (Petaurus australis) at the western edge of its range Wildlife Research 26, 199–208. Cartmill M 1974. Daubentonia, Dactylopsila, woodpeckers and klinorhynchy. In Prosimian Biology. Eds RD Martin, CA Doyle and AC Walker. pp. 655–670. University of Pittsburgh Press, Pittsburgh. Case JA 1985. Differences in prey utilization by Pleistocene marsupial carnivores, Thylacoleo carnifex (Thylacoleonidae) and Thylacinus cynocephalus (Thylacinidae). Australian Mammalogy 8, 45–52 Castro-Arellano I, Zarza H, Medellin RA 2000. Philander opossum. Mammalian Species 638, 1–8. Catling PC, Burt RJ 1995. Why are red foxes absent from some Eucalypt forests in eastern New South Wales? Wildlife Research 22, 535–546. Catling PC, Newsome AE 1981. Responses of the Australian vertebrate fauna to fire: an evolutionary approach. In Fire and the Australian Biota. Eds AM Gill, RH Groves and IR Noble. pp. 273–310. Australian Academy of Science, Canberra. Catling PC, Sutherland RL 1980. Effect of gonadectomy, season, and the presence of females on plasma testosterone, luteinising hormone, and follicle stimulating hormone levels in male tammar wallabies (Macropus eugenii). Journal of Endocrinology 86, 25–33. Catling PC, Vinson GP 1976. Adrenocortical hormones in the neonate and pouch young of the tammar wallaby, Macropus eugenii. Journal of Ednocrinology 69, 447–448. Caughley G 1983. The Deer Wars: the Story of Deer in New Zealand. p. 187. Heinemann, Auckland. Caughley G, Brown B, Dostine P, Grice D 1984. The grey kangaroo overlap zone. Australian Wildlife Research 11, 1–10. Caughley G, Grice D, Barker R, Brown B 1988. The edge of the range. Journal of Animal Ecology 57, 771–785. Caughley G, Shepherd N, Short J (Eds) 1987a. Kangaroos, their ecology and management in the sheep rangelands of Australia. p. 253. Cambridge University Press, Cambridge. Caughley G, Short J, Grigg GC, Nix H 1987b. Kangaroos and climate: an analysis of distribution. Journal of Animal Ecology 56, 751–761. Caughley GJ 1964. Density and dispersion of two species of kangaroo in relation to habitat. Australian Journal of Zoology 12, 238–249. References 389

Cerqueira R 1984. Reproduction de Didelphis albiventris dans le nord-est du Brésil (Polyprotodontia: Didelphidae). Mammalia 48, 95–104. Cerqueira R, Gentile R, Fernandez FAS, d’Andrea PS 1993. A five-year population study of an assemblage of small mammals in Southeastern Brazil. Mammalia 57, 507–517. Chappell J 1993. Late Pleistocene coasts and human migrations in the Austral region. In A community of culture.The people and prehistory of the Pacific.Eds M Spriggs, DE Yen, W Ambrose, R Jones, A Thorne, A Andrews, pp. 43–48.Department of Prehistory, Australian National University, Canberra. Charles-Dominique P, Hladik CM 1971. Le Lepilemur du sud de Madagascar: ecologie, alimentation et vie sociale. Terre et Vie 25, 3–66. Charles-Dominique P 1983. Ecology and social adaptations in didelphid marsupials: comparison with eutherians of similar ecology. In Advances in the Study of Mammalian Behavior. Eds JF Eisenberg and DG Kleiman, pp. 395–422. Philadeplphia, American Society of Mammalogists. Charles-Dominique P, Atramentowicz M, Charles-Dominique M, Gerard H, Hladik A, Hladik CM, Prevost MF 1981. Les mammiferes frugivores, arboricoles nocturnes d’une foret guyanaise: inter- relations plantes-animaux. Revue d’Ecologie (Terre et Vie) 35, 341–435. Chilcott MJ 1984. Coprophagy in the common ringtail possum, Pseudocheirus peregrinus, (Marsupialia: Petauridae). Australian Mammalogy 7, 107–110. Chilcott MJ, Hume ID 1984. Nitrogen and urea metabolism and nitrogen requirements of the common ringtail possum, Pseudocheirus peregrinus, fed Eucalyptus andrewsii foliage. Australian Journal of Zoology 32, 615–622. Chilcott MJ, Hume ID 1985. Coprophagy and selective retention of fluid digesta: their role in the nutri- tion of the common ringtail possum, Pseudocheirus peregrinus. Australian Journal of Zoology 33, 1–15. Chilcott MJ, Moore SA, Hume ID 1985. Effects of water restriction on nitrogen metabolism and urea recycling in the macropodid marsupials Macropus eugenii (tammar wallaby) and Thylogale thetis (red- necked pademelon). Journal of Comparative Physiology B 155, 759–767. Choquenot D, Bowman DMJS 1998. Marsupial megafauna, Aborigines and the overkill hypothesis: appli- cation of predator-prey models to the question of Pleistocene extinction in Australia. Global Ecology and Biogeography Letters 7, 167–180. Christian DP 1983. Water balance in Monodelphis domestica (Didelphidae) from the semiarid Caatinga of Brazil. Comparative Biochemistry and Physiology A 74, 665–669. Churcher CS 1985. Dental functional morphology in the marsupial sabre-tooth Thylacosmilus atrox (Thalacosmilidae) compared to that of felid sabre-tooths. Australian Mammalogy 8, 201–220. Cifelli RL, Davis BM 2003. Marsupial origins. Science 302, 1899–1900. Claridge AW, Cork SJ 1994. Nutritional value of hypogeal fungal sporocarps for the long-nosed potoroo (Potorous tridactylus), a forest-dwelling mycophagous marsupial. Australian Journal of Zoology 42, 701–710. Clark MJ 1968. Termination of embryonic diapause in the red kangaroo, Megaleia rufa, by injection of progesterone or oestrogen. Journal of Reproduction and Fertility 15, 347–355. Cockburn A 1990. Life history of the bandicoots: developmental rigidity and phenotypic plasticity. In Bandicoots and Bilbies. Eds JH Seebeck, RL Wallis and CM Kemper. pp. 285–292. Surrey Beatty, Sydney. Cockburn A 1992. The duration of lactation in Antecinus stuartii (Marsupialia: Dasyuridae). Australian Journal of Zoology 40, 195–204. Cockburn A 1994. Adaptive sex allocation by brood reduction in . Behavioural Ecology and Sociobiology 35, 53–62. Cockburn A, Lazenby-Cohen KA 1992. Use of nest trees by Antechinus stuartii, a semelparous lekking marsupial. Journal of Zoology, London 226, 657–680. Coleman, J. D., Gillman, A., and Green, W. Q. 1980. Forest patterns and possum densities within podocarp/ mixed hardwood forests on Mt Bryan O’Lynn, Westland. New Zealand Journal of Ecology, 3, 69–84. Coles RB, Guppy A 1986. Biophysical aspects of directional hearing in the tammar wallaby, Macropus eugenii. Journal of Experimental Biology 121, 371–394. Collet C, Joseph R 1994. The identification of nuclear and mitochondrial genes by sequencing randomly chosen clones from a marsupial mammary gland cDNA library. Biochemical Genetics 32, 181–190 Collins BG, Wooller RD, Richardson KC 1987. Torpor by the honey possum, Tarsipes rostratus (Marsupialia: Tarsipedidae), in response to food shortage and low environmental temperature. Australian Mammalogy 11, 51–57. 390 Life of Marsupials

Comport SS, Ward SJ, Foley WJ 1996. Home ranges, time budgets and food-tree use in a high-density tropical population of greater gliders, Petauroides volans minor (Pseudocheiridae: Marsupialia). Wildlife Research 23, 401–419. Cooke B, Kear B. 1999. Evolution and diversity of kangaroos (Macropodoidea, Marsupialia). Australian Mammalogy 21, 27–29. Cooke BD 1998. Did introduced European Rabbits Oryctolagus cuniculus (L.) displace common wombats Vombatus ursinus (Shaw) from part of their range in South Australia? In Wombats. Eds RT Wells and PA Pridmore. pp. 262–270. Surrey Beatty, Chipping Norton. Cooper DW 1993. The evolution of sex determination, sex chromosome dimorphism and X-inactivation in therian mammals: a comparison of metatherians (Marsupials) and eutherians (‘Placentals’). Pages 183–198. In Sex chromosomes and sex-determining genes. Eds KC Reed and JAM Graves. Harwood Academic Publishers, Chur, Switzerland. Copley PB, Read VT, Robinson AC, Watts CHS 1990. Preliminary studies of the Nuyts Archipelago bandi- coot, Isoodon obesulus nauticus on the Franklin Islands, South Australia. In Bandicoots and Bilbies, Eds JH Seebeck, RL Wallis and CM Kemper. pp. 345–356. Surrey Beatty, Sydney. Cork SJ 1986 Foliage of Eucalyptus punctata and the maintenance nitrogen requirements of koalas, Phascolarctos cinereus. Australian Journal of Zoology 34, 17–23. Cork SJ 1992. Polyphenols and the distribution of arboreal folivorous marsupials in eucalyptus forests of Australia. In Plant Polyphenols; Biogenesis, ChemicalProperties and Significance. Ed. R. W. Hemmingway. pp. 653–663. Plenum Press, New York. Cork SJ, Dove H 1989. Lactation in the tammar (Macropus eugenii). II. Intake of milk components and maternal allocation of energy. Journal of Zoology, London 219, 399–409. Cork SJ, Hume ID 1983. Microbial digestion in the koala (Phascolarctos cinereus, Marsupialia), an arboreal folivore. Journal of Comparative Physiology 152, 131–135. Cork SJ, Pahl L 1984. The possible influence of nutritional factors on diet and habitat selection by the ringtail possum (Pseudocheirus peregrinus). In Possums and Gliders. Eds AP Smith and ID Hume. pp. 269–276. Australian Mammal Society, Sydney. Cork SJ, Sanson GD 1990. Digestion and nutrition in the koala: a review. In Biology of the Koala. Eds AK Lee, KA Handasyde, GD Sanson. pp. 129–144. Surrey Beatty, Sydney. Cork SJ, Warner ACI 1983. The passage of digest markers through the gut of a folivorous marsupial, the koala (Phascolarctos cinereus). Journal of Comparative Physiology 152, 43–51. Cork SJ, Hume ID, Dawson TJ 1983. Digestion and metabolism of a natural foliar diet (Eucalyptus puncata) by an arboreal marsupial, the koala (Phascolarctos cinereus). Journal of Comparative Physiology 153, 181–190. Cowan PE 1989a. Changes in milk composition during lactation in the common brushtail possum, Trichosurus vulpecula (Marsupialia: Phalangeridae). Reproduction Fertility and Development 1, 325–335. Cowan PE 1989b. Denning habits of common brushtail possums, Trichosurus vulpecula, in New Zealand lowland forest. Australian Wildlife Research 16, 63–78. Cowan PE, Moeed A 1987. Invertebrates in the diet of brushtail possums, Trichosurus vulpecula, in lowland podocarp/broadleaf forest, Orongorongo Valley, Wellington, New Zealand. New Zealand Journal of Zoology 14, 163–177. Cowan PE, Chilvers BL, Efford MG, McElrea GJ 1997. Effects of possum browsing on northern rata, Orongorongo valley, Wellington, New Zealand. Journal of the Royal Society of New Zealand 27, 173–179. Craig SA 1985. Social organisation, reproduction and feeding behaviour of a population of yellow-bellied gliders, Petaurus australis (Marsupialia: Petauridae) in Victoria Australian Wildlife Research 12,1–18. Crawley MC 1973. A live-trapping study of Australian brush-tailed possums, Trichosurus vulpecula (Kerr), in the Orongorongo valley, Wellington, New Zealand. Australian Journal of Zoology 21, 75–90. Crisp EA, Messer M, VandeBerg JL 1989. Changes in milk carbohydrates during lactation in a didelphid marsupial, Monodelphis domestica. Physiological Zoology 62, 1117–1125. Croft DB 1981a. Social behaviour of the euro, Macropus robustus (Gould) in the Australian arid zone. Australian Wildlife Research 8, 13–49. Croft DB 1981b. Behaviour of red kangaroos, Macropus rufus (Desmarest, 1822) in northwestern New South Wales, Australia. Australian Mammalogy 4, 5–58. Croft DB 1991a. Home range of the red kangaroo Macropus rufus. Journal of Arid Environments 20, 83–98. References 391

Croft DB 1991b. Home range of the euro, Macropus robustus erubescens. Journal of Arid Environments 20, 99–111. Crossman DG, Johnson CN, Horsup AB 1994. Trends in the population of the northern hairy-nosed wombat Lasiorhinus krefftii in Epping Forest National Park, Central Queensland. Pacific Conservation Biology 1, 141–149. Cunha AA, Vieira MV 2002. Support diameter, incline, and vertical movements of four didephid marsu- pials in the Atlantic forest of Brazil. Journal of Zoology, London 258, 419–426. Cutts JH, Krause WJ, Leeson CR 1978. General observations on the growth and development of the pouch young opossum, Didelphis virginiana. Biology of the Neonate 33, 264–272. Dampier W 1939. Voyage to New Holland &c. in the Year 1699. Ed. JA Williamson. Argonaut Press, London. Darwin C 1859. The Origin of Species, 1st Edition, p. 101. John Murray, London. Davis DE 1945. The annual cycle of plants, mosquitoes, birds and mammals in two Brazilian forests. Ecological Monographs 15, 243–295. Dawson L, Flannery TF 1985. Taxonomic and phylogenetic status of living and fossil kangaroos and wallabies of the genus Macropus Shaw (Macropodidae: Marsupialia) with a new subgeneric name for the larger wallabies. Australian Journal of Zoology 33, 473–498. Dawson TJ 1973. Thermoregulatory responses of the arid zone kangaroos, Megaleia rufa and Macropus robustus. Comparative Biochemistry and PhysiologyA 46, 153–169. Dawson TJ 1995. Kangaroos. Biology of the largest Marsupials. pp. 162. University of New South Wales Press, Sydney. Dawson TJ, Brown GD 1970. A comparison of the insulative and reflective properties of the fur of desert kangaroos. Comparative Biochemistry and Physiology 37, 23–38. Dawson TJ, Degabriele R 1973. The cuscus (Phalanger maculatus) – a marsupial sloth? Journal of Comparative Physiology 83, 41–50. Dawson TJ, Denny MJS 1969a. A bioclmatological comparison of the summer day microenvironments of two species of arid-zone kangaroo. Ecology 50, 328–332. Dawson TJ, Denny MJS 1969b. Seasonal varaiton in the plasma and urine electrolute concentration of the arid zone kangaroos Megaleia rufa and Macropus robustus. Australian Journal of Zoology 17, 777–784. Dawson TJ, Denny MJS, Hulbert AJ 1969. Thermal balance of the macropodid marsupial Macropus eugenii Desmarest. Comparative Biochemistry and Physiology 31, 645–653. Dawson TJ, Hulbert AJ 1970. Standard metabolism, body temperature, and surface areas of Australian marsupials. American Journal of Physiology 218, 1233–1238. Dawson TJ, May EL 1984. Daily variation in brain and body temperatures of the sugar glider (Petaurus breviceps): some insights into the control of thermoregulation. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 357–383. Australian Mammal Society, Sydney. Dawson TJ, Taylor CR 1973. Energetic cost of locomotion in kangaroos. Nature 246, 313–314. de Blainville HMD 1816. Prodrome d’une nouvelle distribution systématique du règne animal.Bulletin des Sciences par la Société philomatique de Paris 105–125. De Leo AA, Guedlha N, Toder R, Voullaire L, Ferguson-Smith MA, O’Brien PCM, Graves JAM 1999. Comparative chromosome painting between marsupial orders: relationships with a 2n=14 ancestral marsupial karyotype. Chromosome Research 7, 509–517. Deane EM, Cooper DW 1984. Immunology of pouch young marsupials. I. Levels of immunoglobulin, transferrin and albumin in the blood and milk in euros and wallaroos (hill kangaroos: Macropus robustus. Marsupialia). Developmental and Comparative Immunology 8, 863–876. Deane EM, Cooper DW, Renfree MB 1990. Immunological development of pouch young marsupials II. Immunoglobulin G levels in embryos, extra-embryonic membranes, neonates and early milk of tammar wallabies Macropus eugenii. Reproduction, Fertility and Development 2, 369–337. Degabriele R, Dawson TJ 1979. Metabolism and heat balance in an arboreal marsupial, the koala (Phascolarctos cinereus). Journal of Comparative Physiology 134, 293–301. Degabriele R, Harrop CJF, Dawson TJ 1978. Water metabolism in the koala (Phascolarctos cinereus). In The Ecology of Arboreal Folivores. Ed. GG Montgomery. pp. 163–172. Smithsonian Institution Press, Washington DC. Delaney R 1997a. Population dynamics of the allied rock-wallaby. Petrogale assimilis: implications for conservation. Australian Mammalogy 19, 199–207. 392 Life of Marsupials

Delaney R 1997b. Reproductive ecology of the allied rock-wallaby. Petrogale assimilis. Australian Mammalogy 19, 209–218. Delaney R, Marsh H 1995. Estimating the age of wild rock-wallabies by dental radiography: a basis for quantifying the age structure of a discrete colony of Petrogale assimilis. Wildlife Research 22, 547–559. Dellow DW, Nolan JV, Hume ID 1983. Studies on the nutrition of macropodine marsupials. V. Microbial fermentation in the forestomach of Thylogale thetis and Macropus eugenii. Australian Journal of Zoology 31, 433–443. Denker H-W, Tyndale-Biscoe CH 1986. Embryo implantation and proteinase activities in a marsupial (Macropus eugenii). Histochemical patterns of proteinases in various gestational stages. Cell and Tissue Research 246, 279–291. Dennis AJ, Marsh H 1997. Seasonal reproduction in the musky rat kangaroo, Hypsiprymnodon moschatus: a response to changes in resource availability. Wildlife Research 24, 561–578. Diamond J 1999. Guns, Germs, and Steel: The Fate of Human Societies. p. 480. Norton, New York. Dickman CR 1982. Some ecological aspects of seasonal breeding in Antechinus (Dasyuridae, Marsupialia). In Carnivorous Marsupials. Ed. M. Archer. pp. 139–150. Royal Zoological Society of New South Wales, Sydney. Dickman CR, Braithwaite RW 1992. Postmating mortality of males in the dasyurid marsupials, Dasyurus and Parantechinus. Journal of Mammalogy 73, 143–147. Dickman CR, Pressey RL, Lim L, Parnaby HE 1993. Mammals of particular conservation concern in the western division of New South Wales. Biological Conservation 65, 219–248. Dove H, Cork SJ, Christian KR 1989. Recycling of water during lactation and its effects on the estimation of milk intake in a developing macropodid, the tammar wallaby (Macropus eugenii). pp. 231–237. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. Surrey Beatty & Sons, Sydney. Dunlop SA, Coleman L-A, Harman AM, Beazley LD 1988. Development of the marsupial primary visual pathway. Chapter 9 In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. pp. 117– 131. Springer-Verlag, Berlin. Dunnet GM 1964. A field study of local populations of the brushtailed possum Trichosurus vulpecula in Eastern Australia. Proceedings of the Zoological Society of London 142, 665–695. Dwiyahreni AA, Kinnaird MF, O’Brien TG, Supriatna J, Andayani N 1999. Diet and activity of the bear cuscus, Ailurops ursinus, in North Sulawesi, . Journal of Mammalogy 80, 905–912. Ealey EHM 1967a. Ecology of the euro, Macropus robustus (Gould) in north-western Australia. I. The environment and changes in euro and sheep populations. CSIRO Wildlife Research 12, 9–25. Ealey EHM 1967b. Ecology of the euro, Macropus robustus (Gould) in north-western Australia. II. Behaviour, movements, and drinking patterns. CSIRO Wildlife Research 12, 27–51. Ealey EHM, Bentley PJ, Main AR 1965. Studies on water metabolism of the hill kangaroo, Macropus robustus (Gould), in northwest Australia. Ecology 46, 473–479. Edwards D, Westerman M 1992. DNA-DNA hybridisation and the position of Leadbeater’s possum (Gymnobelideus leadbeateri McCoy) in the family Petauridae (Marsupialia: Diprotodontia). Australian Journal of Zoology 40, 563–571. Edwards D, Westerman M 1995. The molecular relationships of possum and glider families as revealed by DNA-DNA hybridisations. Australian Journal of Zoology 43, 231–240. Edwards GP 1989. The interaction between macropodids and sheep: a review. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 795–803. Surrey Beatty, Sydney. Efford M 1998. Demographic consequences of sex-biased dispersal in a population of brushtail possums. Journal of Animal Ecology 67, 503–517. Efford MG 1991. The ecology of an uninfected forest possum population. In Symposium on Tuberculosis. Ed. R Jackson. Veterinary Continuing Education, Palmerston North. Publication 132, 41–51. Efford MG 2000. Possum density, population structure, and dynamics. In The Brushtail Possum. Ed. TL Montague. Chapter 5, pp. 47–61. Manaaki Whenua Press, Lincoln, New Zealand. Eldridge MDB, Close RL 1992. of rock wallabies, Petrogale (Marsupialia: Macropodidae). I. A revision of the eastern Petrogale with the description of three new species. Australian Journal of Zoology 40, 605–625. Eldridge MDB, Close RL 1997. Chromosomes and evolution in rock wallabies, Petrogale (Marsupialia: Macropodidae). Australian Mammalogy 19, 123–135. References 393

Eldridge MDB, King JM, Loupis AK, Spencer PBS, Taylor AC, Pope LC, Hall GP 1999. Unprecedented low levels of genetic variation and inbreeding depression in an island population of the black-footed rock-wallaby. Conservation Biology 13, 531–541. Eldridge MDB, Kinnear JE, Onus ML 2001. Source population of dispersing rock-wallabies (Petrogale lateralis) identified by assignment tests on multilocus genotype data. Molecular Ecology 10, 2867– 2876. Ellis BA, Russell EM, Dawson TJ, Harrop CJF 1977. Seasonal changes in diet preferences of free-ranging red kangaroos, euros and sheep in Western New South Wales. Australian Wildlife Research 4, 127–144. Ellis WAH, Melzer A, Green B, Newgrain K, Hindell MA, Carrick FN 1995. Seasonal variation in water flux, field metabolic rate and food consumption of free-ranging koalas (Phascolarctos cinereus). Australian Journal of Zoology 43, 59–68. Elton C 1942. Voles, Mice and Lemmings. Clarendon Press, Oxford. Emmons LH 1990. Neotropical Rainforest Mammals. A Field Guide. p. 281. University of Chicago Press, Chicago. Enders RK 1935. Mammalian life histories from Barro Colarado Island, Panama. Bulletin of the Museum of Comparative Zoology, Harvard 78, 385–502. Enders RK 1937. Panniculus carnosus and formation of the pouch in Didelphids. Journal of Morphology 61. 1–26. Enders RK 1966. Attachment, nursing and survival of young in some didelphids. Symposia of the Zoological Society of London 15, 195–203. Ewen KR, Temple-Smith PD, Bowden DK, Marinopolous J, Renfree MB, Yan H 1993. DNA fingerprint- ing in relation to male dominance and paternity in a captive colony of tammar wallabies (Macropus eugenii). Journal of Reproduction and Fertility 99, 33–37. Eyre TJ, Goldingay RL 2003. Use of sap trees by the yellow-bellied glider near Maryborough in south-east Queensland. Wildlife Research 30, 229–236. Fanning FD 1980. Nests of the feathertail glider, Acrobates pygmaeus (Burramyidae: Marsupialia), from Sydney, New South Wales. Australian Mammalogy 3, 55–56. Farris EJ 1950. The opossum. In The Care and Breeding of Laboratory Animals. Ed. EJ Farris. pp. 256–267. John Wiley, New York. Field J, Fullagar R 2001. Archaeology and . Science 294, 7a. Fish FE 1993. Comparison of swimming kinematics between terrestrial and semiaquatic opossums. Journal of Mammalogy 74, 275–284. Fisher DO, Dickman CR 1993. Body-size prey-size relationships in insectivorous marsupials – tests of three hypotheses. Ecology 74, 1871–1883. Fitzgerald, AE 1976. Diet of the opossum Trichosurus vulpecula (Kerr) in the Orongorongo Valley, Wellington, New Zealand, in relation to food-plant availability. New Zealand Journal of Zoology 3, 399–419. Flannery TF 1983. Review of the subfamily Sthenurinae (Marsupialia) and the relationships of the species of Tropsodon and Lagostrophus. Australian Mammalogy 6, 15–28. Flannery TF 1989. Phylogeny of the Macropodoidea; a study in convergence. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman, I Hume. pp. 1–46. Surrey Beatty, Sydney. Flannery TF 1994. The Future Eaters. p. 423. Reed Books, Sydney Flannery TF 1995. Mammals of New Guinea. p. 568. Reed Books, Sydney. Flannery TF, Archer M, Rich THV, Jones R 1995. A new family of monotremes from the Cretaceous of Australia. Nature, London 377, 418–420 Flannery TF, Gott B 1984. The Springs Creek locality, southwestern Victoria, a late surviving megafaunal assemblage. The Australian Zoologist 21, 385–422. Flannery TF, Schouten P 1994. Possums of the World. A monograph of the . p. 240. Geo Productions, Sydney. Fleck DW, Harder JD 1995. Ecology of marsupials in two Amazonian rain forests in northeastern Peru. Journal of Mammalogy 76, 809–818. Fleming MR 1980. Thermoregulation and torpor in the sugar glider, Petaurus breviceps (Marsupialia: Petauridae). Australian Mammalogy 8, 79–90. Fleming MR 1985. The thermal physiology of the feathertail glider, Acrobates pygmaeus (Marsupialia: Burramyidae). Australian Journal of Zoology 33, 667–681. 394 Life of Marsupials

Fleming MR, Frey H 1984. Aspects of the natural history of feathertail gliders (Acrobates pygmaeus) in Victoria. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 403–408. Australian Mammal Society, Sydney. Fleming MW, Harder JD, Wukie JJ 1981. Reproductive energetics of the Viriginia opossum compared with some eutherians. Comparative Biochemistry and Physiology B 70, 645–648. Fleming TH 1973. The reproductive cycles of three species of opossums and other mammals in the Panama Canal Zone. Journal of Mammalogy, 54, 439–55. Fletcher T, Selwood L 2000. Possum reproduction and development. In The Brushtail Possum. Ed. TL Montague. Chapter 6, pp. 62–81. Manaaki Whenua Press, Lincoln, New Zealand. Fletcher TP 1989. Plasma progesterone and body weight in the pregnant and non-pregnant kowari, Dasyuroides byrnei (Marsupialia: Dasyuridae). Reproduction, Fertility and Development 1, 65–74. Fletcher TP, Shaw G, Renfree MB 1990. Effects of bromocriptine at parturition in the tammar wallaby, Macropus eugenii. Reproduction, Fertility and Development 2, 79–88. Flood J 1997. Rock Art of the Dreamtime. Images of Ancient Australia. Angus and Robertson, Sydney. Foley WJ 1987. Digestion and energy metabolism in a small arboreal marsupial, the greater glider, Petauroides volans fed high-terpene Eucalyptus foliage. Journal of Comparative Physiology 157, 355– 362. Foley WJ, Charles-Dominique P, Julien-Laferriere D 2000. Nitrogen requirements of the didelphid marsu- pial Caluromys philander. Journal of Comparative Physiology B 170, 345–350. Foley WJ, Hume ID 1987a. Digestion and metabolism of high-tannin Eucalyptus foliage by the brushtail possum (Trichosurus vulpecula) (Marsupialia: Phalangeridae). Journal of Comparative Physiology B 157, 67–76. Foley WJ, Hume ID 1987b. Passage of digesta markers in two species of arboreal folivorous marsupials – the greater glider (Petauroides volans) and the brushtail possum (Trichosurus vulpecula). Physiological Zoology 60, 103–113. Foley WJ, Hume ID 1987c. Nitrogen requirements and urea metabolism in two arboreal marsupials, the greater glider (Petauroides volans) and the brushtail possum (Trichosurus vulpecula), fed eucalyptus foliage. Physiological Zoology 60, 241–250. Foley WJ, Hume ID, Cork SJ 1989. Fermentation in the hind gut of the greater glider (Petauroides volans) and the brushtail possum (Trichosurus vulpecula) – two arboreal folivores. Physiological Zoology 62, 1126–1143. Foley WJ, Kehl JC, Nagy KA, Kaplan IR, Borsboom AC 1990. Energy and water metabolism in free-living greater gliders, Petauroides volans. Australian Journal of Zoology 38, 1–9. Fonseca GAB, Paglia AP, Sanderson J, Mittermeier RA 2003. Marsupials of the New World: status and conservation. In Predators with Pouches. Eds M Jones, C Dickman, M Archer. pp. 399–406. Melbourne, CSIRO Publishing. Frappell P, Baudinette R, MacFarlane P, Wiggins P, Shimmin G 2002. Ventilation and metabolism in a large semi-fossorial marsupial, the hairy nosed wombat: the effect of graded hypoxia and hypercapnia, Physiological and Biochemical Zoology 75, 77–82. Frederick H, Johnson CN 1996. Social organisation of the rufous bettong, Aepyprymnus rufescens. Australian Journal of Zoology 44, 9–17. Freeland WJ, Winter JW 1975. Evolutionary consequences of eating: Trichosurus vulpecula (Marsupialia) and the genus Eucalyptus. Journal of Chemical Ecology 1, 439–455. Freudenberger DO, Hume ID 1992. Ingestive and digestive responses to dietary fibre and nitrogen by two macropodid marsupials (Macropus robustus erubescens and M. r. robustus) and a ruminant (Capra hircus). Australian Journal of Zoology 40, 181–194. Freudenberger DO, Wallis IR, Hume ID 1989. Digestive adaptations of kangaroos, wallabies and rat- kangaroos. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 179–187. Surrey Beatty, Sydney. Frey H 1991. Energetic significance of torpor and other energy-conserving mechanisms in free-living Sminthopsis crassicaudata (Marsupialia: Dasyuridae). Australian Journal of Zoology 39, 689–708. Friend G 1990. Breeding and population dynamics of Isoodon macrourus (Marsupialia: Peramelidae): studies from the wet dry tropics of northern Australia. In Bandicoots and Bilbies, Eds JH Seebeck, RL Wallis and CM Kemper. pp. 357–365. Surrey Beatty, Sydney. Frith HJ 1964. Mobility of the red kangaroo, Megaleia rufa. CSIRO Wildlife Research 9, 1–19. J78 References 395

Frith HJ, Sharman GB 1964. Breeding in wild populations of the red kangaroo, Megaleia rufa. CSIRO Wildlife Research 9, 86–114. Galloway RW, Kemp EA 1981. Late Cainozoic environments in Australia. In Ecological Biogeography of Australia. Ed. A Keast. pp. 53–80. Junk, the Hague. Gardner AL 1973. The systematics of the genus Didelphis (Marsupialia: Didelphidae) in North and Middle America. Special Publication of the Museum, Texas Technical University 4, 1–81. Gardner AL 1982. Virginia opossum. In Wild mammals of North America: biology, economics, and manage- ment. Eds JA Chapman and GA Feldhamer. pp. 3–36. The John Hopkins University Press, Baltimore. Gaughwin MD, Breed WG, Wells RT 1998. Seasonal reproduction in a population of southern hairy- nosed wombats Lasiorhinus latifrons in the Blanchetown region of South Australia. In Wombats. Eds RT Wells and PA Pridmore. pp. 109–112. Surrey Beatty, Chipping Norton. Geiser F 1993. Hibernation in the eastern pygmy possum, Cercartetus nanus (Marsupialia: Burramyidae). Australian Journal of Zoology 41, 67–75. Geiser F 1994. Hibernation and daily torpor in marsupials: a review. Australian Journal of Zoology 42, 1–16 Geiser F, Sink HS, Stahl B, Mansergh IM, Broome LS 1990. Differences in the physiological response to cold in wild and laboratory-bred mountain pygmy-possums, Burramys parvus (Marsupialia). Australian Wildlife Research 17, 535–539. Gemmell RT 1981. The role of the corpus luteum of lactation in the bandicoot, Isoodon macrourus (Marsupialia: Peramelidae). General and Comparative Endocrinology 44, 13–19. Gemmell RT, Nelson J 1988. The ultrastructure of the pituitary and the adrenal gland of three newborn marsupials (Dasyurus hallucatus, Trichosurus vulpecula and Isoodon macrourus). Anatomy and Embryology 177, 395–402. Gemmell RT, Nelson J 1989a. The ultrastructure of the lung of two newborn marsupial species, the north- ern native cat, Dasyurus hallucatus, and the brushtail possum, Trichosurus vulpecula. Cell and Tissue Research 252, 683–685. Gemmell RT, Nelson J 1989b. Vestibular system of the newborn marsupial cat Dasyurus hallucatus. The Anatomical Record 225, 203–208. Gemmell RT, Peters B, Nelson J 1988. Ultrastructural identification of Merkel cells around the mouth of the newborn marsupial. Anatomy and Embryology 177, 403–408. Gemmell RT, Rose RW 1989. The senses involved in movement of some newborn Macropodoidea and other marsupials from cloaca to pouch. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman, I Hume. pp. 339–347. Surrey Beatty, Sydney. Gemmell RT, Sernia C 1992. The role of photoperiod on the initiation of the breeding season of the possum, Trichosurus vulpecula. Journal of Reproduction and Fertility 95, 701–708. Gemmell, RT, Jenkin G, Thorburn GD 1980. Plasma concentrations of progesterone and 13, 14-dihydro–

15-keto-prostaglandin F2_ at parturition in the bandicoot, Isoodon macrourus. Journal of Reproduction and Fertility 60, 253–256. Gemmell RT, Veitch C, Nelson J 2002. Birth in the northern brown bandicoot, Isoodon macrourus (Marsupialia: Peramelidae). Australian Journal of Zoology 47, 517–528. Gentile R, Cerqueira R 1995. Movement patterns of five species of small mammals in a Brazilian restinga. Journal of Tropical Ecology 11, 671–677. George GG, Maynes GM 1990. Status of New Guinea bandicoots. In Bandicoots and bilbies, Eds JH Seebeck, RL Wallis and CM Kemper. pp. 93–104. Surrey Beatty, Sydney. Gilmore DP 1969. Seasonal reproductive periodicity in the male Australian brush-tailed possum (Trichosurus vulpecula). Journal of Zoology, London 157, 75–98. Gipps JM, Sanson GD 1984. Mastication and digestion in Pseudocheirus. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 237–246. Australian Mammal Society, Sydney. Glas R, De Leo AA, Delbridge ML, Reid K, Ferguson-Smith MA, O’Brien PCM, Westerman M, Graves JAM 1999. Chromosome painting in marsupials: genome conservation in the kangaroo family. Chromosome Research 7, 167–176. Godfrey GK 1968. Body-temperatures and torpor in Sminthopsis crassicaudata and S. larapinta (Marsupialia: Dasyuridae). Journal of Zoology, London 156, 488–511. Godfrey GK 1969. Influence of increased photoperiod on reproduction in the dasyurid marsupial, Sminthopsis crassicaudata. Journal of Mammalogy 50, 132–133. 396 Life of Marsupials

Godthelp H, Archer M, Cifelli R, Hand SJ, Gilkeson CF 1992. Earliest known Australian Tertiary mammal fauna. Nature 356, 514–516. Godthelp H, Wroe S, Archer M 1999. A new marsupial from the Early Eocene Tingamarra Local fauna of Murgon, Southeastern Queensland: a prototypical Australian marsupial? Journal of Mammalian Evolution 6, 289–313. Goin FJ 2003. Early marsupial radiations in South America. In Predators with Pouches. Eds M Jones, C Dickman, M Archer. pp. 30–42. Melbourne, CSIRO Publishing. Goin FJ, Case JA, Woodburne MO, Vizcaino SF, Reguero MA 1999. New discoveries of “opossum-like” marsupials from Antarctica (Seymour Island, Medial Eocene). Journal of Mammalian Evolution 6, 335–365. Goldingay RL 1984. Photoperiod control of diel activity in the sugar glider (Petaurus breviceps). In Possums and Gliders. Eds AP Smith and ID Hume. pp. 385–391. Australian Mammal Society, Sydney. Goldingay RL 1989. Time budget and related aspects of the foraging behaviour of the yellow-bellied glider, Petaurus australis. Australian WildlifeRresearch 16, 105–112. Goldingay RL 1991. An evaluation of hypotheses to explain the pattern of sap feeding by the yellow- bellied glider, Petaurus australis. Australian Journal of Ecology 16, 491–500. Goldingay RL 1992. Socioecology of the yellow-bellied glider (Petaurus australis) in a coastal forest. Australian Journal of Zoology 40, 267–278. Goldingay RL 1994. Loud calls of the yellow-bellied glider, Petaurus australis: territorial behaviour by an arboreal marsupial? Australian Journal of Zoology 42, 279–293. Goldingay RL, Kavanagh RP 1990. Socioecology of the yellow-bellied glider, Petaurus australis, at Waratah Creek, NSW. Australian Journal of Zoology 38, 327–341. Goldingay RL, Kavanagh RP 1993. Home-range estimates and habitat of the yellow-bellied glider, Petaurus australis, at Waratah Creek, NSW. Wildlife Research 20, 387–404. Goldingay RL, Possingham H 1995. Area requirements for viable populations of the Australian gliding marsupial Petaurus australis. Biological Conservation 73, 161–167. Gordon G 1974. Movements and activity of the short-nosed bandicoot, Isoodon macrourus Gould (Marsupialia). Mammalia 38, 405–431. Gordon K, Fletcher TP, Renfree MB 1988. Reactivation of the quiescent corpus luteum and diapausing embryo after temporary removal of the sucking stimulus in the tammar wallaby. Journal of Reproduction and Fertility 83, 401–406. Gordon G, Brown AS, Pulsford T 1988. A koala (Phascolarctos cinereus) population crash during drought and heatwave conditions in south western Queensland. Australian Journal of Ecology 13, 451–461. Gould, J 1973. Kangaroos. New edition of The Mammals of Australia, Vol 2, first published in 1863. Ed JM Dixon. p. 152. Macmillan Australia, Melbourne. Gould J 1974. Australian Marsupials and Monotremes. New edition of The Mammals of Australia, Vol 1, first published in 1863. Ed JM Dixon. p. 143. Macmillan Australia, Melbourne. Gowen M 1978.Obituary of Mick Gowen. Tallaganda Times, Braidwood 20 December 1978, p. 5. Gowland PN 1973. Aspects of the Digestive Physiology of the Common Wombat Vombatus ursinus (Shaw 1800). BSc Hons Thesis, Australian National University, Canberra. Graves JAM 1996. Mammals that break the rules: genetics of marsupials and monotremes. Annual ReviewsinGenetics 30, 233–260. Green B 1997. Field energetics and water fluxes in marsupials. In Marsupial Biology: Recent research, new perspectives. Eds NR Saunders and LA Hinds. pp. 143–162. University of New South Wales Press, Sydney. Green B 1984. Composition of milk and energetics of growth in marsupials. Symposium of the Zoological Society of London 51, 369–387. Green B, Krause WJ, Newgrain K 1996. Milk composition in the North American opossum (Didelphis virginiana). Comparative Biochemistry and Physiology B 113, 619–623. Green B, Merchant J, Newgrain K 1997. Lactational energetics of a marsupial carnivore, the eastern quoll (Dasyurus viverrinus). Australian Journal of Zoology 45, 295–306. Green B, Merchant JC 1988. The composition of marsupial milk. Chapter 4 In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. pp. 41–54. Springer-Verlag, Berlin. Green B, Newgrain K, Catling P, Turner G 1991. Patterns of prey consumption and energy use in a small carnivorous marsupial, Antechinus stuartii. Australian Journal of Zoology 39, 539–547. References 397

Green B, Newgrain K, Merchant JC 1980. Changes in milk composition during lactation in the tammar wallaby (Macropus eugenii). Australian Journal of Biological Science 33, 35–42. Green SW, Renfree MB 1982. Changes in the milk proteins during lactation in the tammar wallaby, Macropus eugenii. Australian Journal of Biological Science 35, 145–152. Green WQ 1984. A review of ecological studies relevant to managemnt of the common brushtail possum. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 483–499. Surrey Beatty, Chipping Norton, NSW. Green WQ, Coleman JD 1987. Den sites of possums, Trichosurus vulpecula, and frequency of use in mixed hardwood forest in Westland, New Zealand. Australian Wildlife Research 14, 285–292. Griffiths M, Barker R 1966. The plants eaten by sheep and kangaroos grazing together in a paddock in south western Queensland. CSIRO Wildlife. Research 11, 145–167. Griffiths M, Barton AA 1966. The ontogeny of the stomach in pouch young of the red kangaroo.CSIRO Wildlife Research 11, 169–185. Griffiths M, McIntosh DL, Leckie RMC 1972. The mammary glands of the red kangaroo, with observa- tions on the fatty acid components of the milk triglycerides. Journal of Zoology, London 166, 265–275. Griffiths ME, Slater E 1988. The significance of striated muscle in the mammary glands of marsupials. Journal of Anatomy 156, 141–156. Griffiths M, Barker R, McLean L 1974. Further observations on the plants eaten by kangaroos and sheep grazing together in a paddock in south western Queensland. Australian Wildlife Research 1, 27–43. Grigg G 1988. Kangaroo harvesting and the conservation of the sheep rangelands. Australian Zoologist 24, 124–128. Groves CP, Flannery T 1990. Revision of the families and genera of bandicoots. In Bandicoots and bilbies, Eds JH Seebeck, RL Wallis and CM Kemper. pp. 1–11. Surrey Beatty, Sydney. Guilday JE 1958. The prehistoric distribution of the opossum. Journal of Mammalogy 39, 39–43. Guiler ER 1985. Thylacine: the Tragedy of the Tasmanian Tiger. Oxford University Press, Melbourne. Haight JR, Nelson JE 1987. A brain that doesn’t fit its skull: a comparative study of the brain and endoc- ranium of the Koala, Phascolarctos cinereus (Marsupialia: Phascolarctidae). In Possums and Opossums: Studies in Evolution. Ed. M Archer. pp. 331–352. Surrey Beatty and Royal Zoological Society of New South Wales, Sydney. Handasyde KA, Martin RW 1996. Field observations on the common striped possum (Dactylopsila trivir- gata) in north Queensland. Wildlife Research 23, 755–766. Harder JD, Fleck DW 1997. Reproductive ecology of new world marsupials. In Marsupial Biology. Recent research, new perspectives. N Saunders and L Hinds (Eds.) pp. 175–203. UNSW Press, Sydney. Harder JD, Hinds LA, Horn CA, Tyndale-Biscoe CH 1984. Oestradiol in follicular fluid and in utero- ovarian venous and peripheral plasma during parturition and post-partum oestrus in the tammar, Macropus eugenii. Journal of Reproduction and Fertility 72, 551–558. Harder JD, Hinds LA, Horn CA, Tyndale-Biscoe CH 1985. Effects of removal in late pregnancy of the corpus luteum, Graafian follicle or ovaries on plasma progesterone, oestradiol, LH, parturition and post-partum oestrus in the tammar, Macropus eugenii. Journal of Reproduction and Fertility 75, 449–459. Harder JD, Hsu MJ, Garton DW 1996. Metabolic rates and body temperature of the gray short-tailed opossum (Monodelphis domestica) during gestation and lactation. Physiological Zoology 69, 317–339. Harris PM, Dellow DW, Broadhurst RB 1985. Protein and energy requirements and deposition in the growing brushtail possum and rex rabbit. Australian Journal of Zoology 33, 425–436. Hartman CG 1920. Studies in Didelphis. I. The phenomenon of parturition. Anatomical Record 19, 251–262. Hayman DL 1990. Marsupial cytogenetics. In Mammals from Pouches and Eggs. Eds JAM Graves, RM Hope and DW Cooper. pp. 189–207. CSIRO Publishing, Melbourne. Hayward JS, Lisson PA 1992. Evolution of brown fat: its absence in marsupials and monotremes. Canadian Journal of Zoology 70, 171–179. Hearn JP 1974. The pituitary gland and implantation in the tammar wallaby, Macropus eugenii. Journal of Reproduction and Fertility 39, 235–241. Hearn JP 1975. The role of the pituitary in the reproduction of the male tammar wallaby, Macropus eugenii. Journal of Reproduction and Fertility 42, 399–402. Heinsohn G 1966. Ecology and reproduction of the Tasmanian bandicoots (Perameles gunnii and Isoodon obesulus). University of California Publications in Zoology 80, 1–107. V2 398 Life of Marsupials

Hemmi JM 1999. Dichromatic colour vision in an Australian marsupial, the tammar wallaby. Journal of Comparative Physiology A 185, 509–515. Hemmi JM, Grünert U 1999. Distribution of photoreceptor types in the retina of a marsupial, the tammar wallaby (Macropus eugenii). Visual Neuroscience 16, 291–302. Hemmi JM, Mark RF 1998. Visual acuity, contrast sensitivity and retinal magnification in a marsupial, the tammar wallaby (Macropus eugenii). Journal of Comparative Physiology A 183, 379–387. Henry SR 1984. Social organisation of the greater glider (Petauroides volans) in Victoria. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 221–228. Australian Mammal Society, Sydney. Henry SR, Craig SA 1984. Diet, ranging behaviour and social organisation of the yellow-bellied glider (Petaurus australis Shaw) in Victoria. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 331– 341. Australian Mammal Society, Sydney. Hershkovitz P 1972. The recent mammals of the neotropical region: a zoogeographic and ecological review. In Evolution, Mammals, and Southern Continents, ed. A Keast, FC Erk & B Glass. pp. 311–432. Albany: State University of New York Press. Hershkovitz P 1992. Ankle bones: the Chilean opossum Dromiciops gliroides Thomas, and marsupial phylogeny. Bonner Zoologische Beiträge 43, 181–213. Hill JP 1910. The early development of the Marsupialia with special reference to the native cat (Dasyurus viverrinus). Quarterly Journal of Microscopical Science 56, 1–134. Hill JP 1918. Some observations on the early development of Didelphys aurita. Quarterly Journal of Microscopical Science 63, 91–139. Hill JP, Hill WCO 1955. The growth-stages of the pouch-young of the native cat (Dasyurus viverrinus) together with observations on the anatomy of the new-born young. Transactions of the Zoological Society of London 28, 349–425. Hindell MA, Handasyde KA, Lee AK 1985. Tree species selection by free-ranging koala populations in Victoria. Australian Wildlife Research 12, 137–144. Hinds LA 1988. Hormonal control of lactation. Chapter 5 In The Developing Marsupial. Eds CH Tyndale- Biscoe and PA Janssens. pp. 55–67. Springer-Verlag, Berlin. Hinds LA, den Ottolander RC 1983. Effect of changing photoperiod on peripheral plasma prolactin and progesterone concentrations in the tammar wallaby (Macropus eugenii). Journal of Reproduction and Fertility 69, 631–639. Hinds LA, Loudon SI 1997. Mechanisms of seasonality in marsupials: a comparative view. In Marsupial Biology: recent research, new perspectives. Eds NR Saunders and LA Hinds. pp. 41–70. University of New South Wales Press, Sydney. Hinds LA, Tyndale-Biscoe CH 1985. Seasonal and circadian patterns of circulating prolactin during lacta- tion and seasonal quiescence in the tammar, Macropus eugenii. Journal of Reproduction and Fertility 74, 173–183. Hinds LA, Tyndale-Biscoe CH, Shaw G, Fletcher TP, Renfree MB 1990. Effects of prostaglandin and prolactin on luteolysis and parturient behaviour in the non-pregnant tammar, Macropus eugenii. Journal of Reproduction and Fertility 88, 323–333. Ho SM 1997. Rhythmic motor activity and interlimb co-ordination in the developing pouch young of a wallaby (Macropus eugenii). Journal of Physiology 501, 623–636. Ho SM, Stirling RV 1998. Development of muscle afferents in the spinal cord of the tammar wallaby. Developmental Brain Research 106, 79–91. Hocking GJ 1981. The population ecology of the brushtail possum, Trichosurus vulpecula (Kerr) in Tasmania. MSc Thesis, University of Tasmania. Holland RAB, Rimes AF, Comis A, Tyndale-Biscoe CH 1988. Oxygen carriage and carbonic anhydrase activity in the blood of a marsupial, the tammar wallaby (Macropus eugenii), during early develop- ment. Respiration Physiology 73, 69–86.

Holland RAB, Calvert SJ, Hope RM, Chesson CM 1994. Blood O2 transport in newborn and adult of a very small marsupial (Sminthopsis crassicaudata). Respiration Physiology 98, 69–81. Holloway JC, Geiser F 1996. Reproductive status and torpor of the marsupial Sminthopsis crassicaudata: effect of photoperiod. Journal of Thermal Biology 21, 373–380. Hope JH 1974. The biogeography of the mammals of the islands of Bass Strait. In. Biogeography and Ecology in Tasmania. Ed. WD Willliams. Chapter 14, pp. 397–415. W Junk, The Hague. References 399

Horn CA, Fletcher TP, Carpenter S 1985. Effects of oestradiol–17` on peripheral plasma concentrations of LH and FSH in ovariectomised tammars (Macropus eugenii). Journal of Reproduction and Fertility 73, 585–592. Horst CJ van der, Gillman J 1942. Pre-implantation phenomena in the uterus of Elephantulus. South African Journal of Medical Science 7, 47–71. Horsup AB 1994. Home range of the allied rock-wallaby, Petrogale assimilis. Wildlife Research 21, 65–84. Horsup AB, Marsh H 1992. The diet of the allied rock-wallaby (Petrogale assimilis), in the wet-dry tropics. Wildlife Research 19, 17–33. Horton DR 1984. Red kangaroos: last of the Australian megafauna. In Quartenary Extinctions, a Prehistoric Revolution. Eds PS Martin and RG Klein. pp. 639–680. University of Arizona Press, Tuscon. Hossler RJ, McAninch JB, Harder JD 1994. Maternal denning behavior and survival of juveniles in opossums in southeastern New York. Journal of Mammalogy 75, 60–70. How RA 1972. The ecology and management of Trichosurus species (Marsupialia) in New South Wales. PhD Thesis, University of New England How RA 1976. Reproduction, growth and survival of young in the mountain possum, Trichosurus caninus (Marsupialia). Australian Journal of Zoology 24, 189–199. How RA 1981. Population parameters of two congeneric possums, Trichosurus spp., in North-eastern New South Wales. Australian Journal of Zoology 29, 205–215. How RA, Barnett JL, Bradley AJ, Humphreys WF, Martin R 1984. The population biology of Pseudocheirus peregrinus in a Leptospermum laevigatum thicket. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 261–268. Australian Mammal Society, Sydney. Hsu M, Harder JD, Lustick SI 1988. Seasonal energetics of opossums (Didelphis virginiana) in Ohio. Comparative Biochemistry and Physiology A 90, 441–443. Huang C, Ward S, Lee AK 1987. Comparison of the diets of the feathertail glider, Acrobates pygmaeus, and the eastern pygmy-possum, Cercartetus nanus (Marsupialia: Burramyidae) in sympatry. Australian Mammalogy 10, 47–50. Hughes RD 1965. On the age composition of a small sample of individuals from a population of the banded hare wallaby, Lagostrophus faciatus (Peron & Lesueur). Australian Journal of Zoology 13, 75–95. Hughes RL 1965. Comparative morphology of spermatozoa from five marsupial families. Australian Journal of Zoology 13, 533–543. Hughes RL, Hall LS 1988. Structural adaptations of the newborn marsupial. Chapter 2. In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. pp. 8–27. Springer-Verlag, Berlin. Hughes RL, Hall LS, Tyndale-Biscoe CH, Hinds LA 1989. Evolutionary implication of macropod organo- genesis. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 377–405. Surrey Beatty & Sons, Sydney. Hulbert AJ 1988. Metabolism and the development of endothermy. Chapter 11 In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. pp. 148–161. Springer-Verlag, Berlin. Hulbert AJ, Else PL 1999. Membranes as possible pacemakers of metabolism. Journal of Theoretical Biology 199, 257–274. Hume ID 1977. Maintenance nitrogen requirements of the macropod marsupials Thylogale thetis, red- necked pademelon, and Macropus eugenii, tammar wallaby. Australian Journal of Zoology 25, 407–417. Hume ID 1999. Marsupial Nutrition. p. 434. Cambridge University Press, Cambridge. Hume ID, Barboza PS 1998. The gastrointestinal tract and digestive physiology of wombats. In Wombats. Eds RT Wells and PA Pridmore. pp. 67–74. Surrey Beatty, Chipping Norton. Hume ID, Dunning A 1979. Nitrogen and electrolyte balance in the wallabies Thylogale thetis and Macropus eugenii when given saline drinking water. Comparative Biochemistry and Physiology 63a, 135–139. Hume ID, Foley WJ, Chilcott MJ 1984. Physiolgical mechanisms of foliage digestion in the greater glider and ringtail possum (Marsupialia: Pseudoceiridae). In Possums and Gliders. Eds AP Smith and ID Hume. pp. 247–251. Australian Mammal Society, Sydney. Hume ID, Smith C, Woolley PA 2000. Anatomy and physiology of the Julia Creek dunnart, Sminthopsis douglasi (Marsupialia: Dasyuridae). Australian Journal of Zoology 48, 475–485. Hunsaker D II 1977. Ecology of New World marsupials. In The Biology of Marsupials. Ed. D Hunsaker II, pp. 95–156. Academic Press, New York. 400 Life of Marsupials

Huxley TH 1880.On the application of the laws of evolution to the arrangement of the Vertebrata, and more particularly of the Mammalia. Proceedings of the Zoological Society of London 43, 649–662. Ingram JN, Shaw G, Renfree MB 1999. Cortisol in fetal fluids and fetal adrenal at parturition in the tammar wallaby (Macropus eugenii). Biology of Reproduction 60, 651–655. Inns RW 1980. Ecology of the Kangaroo Island wallaby, Macropus eugenii (Desmarest), in Flinders Chase National Park, Kangaroo Island. PhD Thesis, University of Adelaide. Inns RW 1982. Seasonal changes in the accessory reproductive system and plasma testosterone levels of the male tammar wallaby, Macropus eugenii, in the wild. Journal of Reproduction and Fertility 66, 675–680. Jackson MA 2000a. Glide angle in the genus Petaurus and a review of gliding in mammals. Mammal Review 30, 9–30. Jackson MA 2000b. Population dynamics and life history of the mahogany glider, Petaurus gracilis, and the sugar glider, Petaurus breviceps, in north Queensland. Wildlife Research 27, 21–37. Jackson MA 2000c. Habitat relationships of the mahogany glider, Petaurus gracilis, and the sugar glider, Petaurus breviceps. Wildlife Research 27, 39–48. Jackson MA, Johnson CN 2002. Time allocation to foraging in the mahogany glider, Petaurus gracilis (Marsupialia, Petauridae) and a comparison of activity times in exudivorous and folivorous possums and gliders. Journal of Zoology, London 256, 271–277. Jacobs HTTM, SJ 1971. A Treatise on the Moluccas (c. 1544). Jesuit Historical Institute, Rome. James CD, Landsberg J, Morton SR 2000. Provision of watering points in the Australian arid zone: a review of effects on biota. Journal of Arid Environments 41, 87–121. Jansa SA, Voss RS 2000. Phylogenetic studies on didelphid marsupials I. Introduction and preliminary results from nuclear IRBP gene sequences. Journal of Mammalian Evolution 7, 43–77. Janssens PA, Hulbert AJ, Baudinette RV 1997. Development of the pouch young from birth to pouch vacation. In Marsupial Biology: Recent Research and New Perspectives. Eds NR Saunders and LA Hinds. Chapter 4, pp. 71–89. University of New South Wales Press, Sydney. Janssens PA, Messer M 1988. Changes in nutritional metabolism during weaning. Chapter 12. In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. pp. 162–175. Springer-Verlag, Berlin. Janssens PA, Rogers A 1989. Metabolic changes during pouch vacation and weaning in macropodoids. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 367–376. Surrey Beatty & Sons, Sydney. Jarman PJ 1983. Mating system and sexual dimorphism in large, terrestrial, mammalian herbivores. Biological Reviews 58, 485–520. Jarman PJ, Bayne P 1997. Behavioural ecology of Petrogale penicillata in relation to conservation. Australian Mammalogy 19, 219–228. Jarman PJ, Coulson G 1989. Dynamics and adaptiveness of grouping in macropods. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman, I Hume. pp. 527–547. Surrey Beatty, Sydney. Jernvall J, Keranen SVE, Thesleff I 2000. Evolutionary modification of development in mammalian teeth: quantifying gene expression patterns and topography. Proceedings of the National Academy of Science 97, 14 444–14 448. Ji W, Sarre SD, Aitken N, Hankin RKS, Clout MN 2001. Sex-biased dispersal and a density-independ- ent mating system in the Australian brushtail possum, as revealed by minisatellite DNA profiling. Molecular Ecology 10, 1527–1537. Johnson CN 1983. Variations in group size and composition in red and western grey kangaroos, Macropus rufus (Desmarest) and M. fuliginosus (Desmarest). Australian Wildlife Research 10, 25–31. Johnson CN 1991. Utilization of habitat by the northern hairy-nosed wombat Lasiorhinus krefftii. Journal of Zoology, London 225, 495–507. Johnson CN 1994. Nutritional ecology of a mycophagous marsupial in relation to production of hypogeous fungi. Ecology 75, 2015–2021. Johnson CN 1998. The evolutionary ecology of wombats. In Wombats. Eds RT Wells and PA Pridmore. pp. 34–41. Surrey Beatty, Chipping Norton. Johnson CN, Clinchy M, Taylor AC, Krebs CJ, Jarman PJ, Payne A, Ritchie EG 2001. Adjustment of offspring sex ratios in relation to the availability of resources for philopatric offspring in the common brushtail possum. Proceedings of the Royal Society of London B 268, 2001–2005. Johnson CN, Crossman DG 1991. Dispersal and social organization of the northern hairy-nosed wombat Lasiorhinus krefftii. Journal of Zoology, London 225, 605–613. References 401

Johnson CN, McIlwee AP 1997. Ecology of the northern bettong, Bettongia tropica, a tropical mycophagist. Wildlife Research 24, 549–559. Johnson JI 1977. Central nervous system of marsupials. In The Biology of Marsupials. Ed. D Hunsaker II. Chapter 4, pp. 157–278. Academic Press, New York. Johnson KA 1980. Spatial and temporal use of habitat by the red-necked pademelon, Thylogale thetis (Marsupialia: Macropodidae). Australian Wildlife Research 7, 157–166. Johnson MH, Selwood L 1996. Nomenclature of early development in mammals. Reproduction, Fertility and Development 8, 759–764. Johnson-Murray JL 1987. The comparative myology of the gliding membranes of Acrobates, Petauroides and Petaurus contrasted with the cutaneous myology of Hemibelideus and Pseudocheirus (Marsupialia: Phalangeridae) and with selected gliding rodents (Sciuridae and Anomaluridae). Australian Journal of Zoology 35, 101–113. Jones BA, How RA, Kitchener DJ 1994. A field study of Pseudocheirus occidentalis (Marsupialia: Petauridae). I. Population studies. Wildlife Research 21, 189–201. Jones CJ, Geiser F 1992. Prolonged daily torpor in the feathertail glider, Acrobates pygmaeus (Marsupialia: Acrobatidae). Journal of Zoology, London 227, 101–108. Jones FW 1924. The Mammals of South Australia. Government Printer, Adelaide. (1968 reprint.) Jones ME, Barmuta LA 1998. Diet overlap and relative abundance of sympatric dasyurid carnivores: a hypothesis of competition. Journal of Animal Ecology 67, 410–421. Jones ME, Stoddart DM 1998. Reconstruction of the predatory behaviour of the extinct marsupial thyla- cine (Thylocinus cynocephalus). Journal of Zoology, London 246, 239–246. Jones RC 1989. Reproduction in male Macropodidae. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P. Jarman and I Hume. pp. 287–305. Surrey Beatty, New South Wales. Jordan SM, Morgan EH 1968. The changes in the serum and milk whey proteins during lactation and suckling in the quokka (Setonix brachyurus). Comparative Biochemistry and Physiology 25, 271–283. Julien-Laferriere D 1991. Organisation du peuplement de marsupiaux en Guyane Francaise. Revue d’Ecologie (Terre et Vie) 46, 125–144. Julien-Laferriere D 1995. Use of space by the woolly opossum Caluromys philander (Marsupialia, Didelphidae) in French Guiana. Canadian Journal of Zoology 73, 1280–1289. Julien-Laferriere D 1999. Foraging strategies and food partitioning in the neotropical frugivorous mammals Caluromys philander and Potos flavus. Journal of Zoology, London 247, 71–80. Julien-Laferriere D, Atramentowicz M 1990. Feeding and reproduction of three didelphid marsupials in two neotropical forests (French Guiana). Biotropica 22, 404–415. Jurgilas PB, Neves-Ferreira AGC, Domont GB, Moussatche H, Perales J 1999. Detection of an antiboth- ropic fraction in opossum (Didelphis marsupialis) milk that neutralizes Bothrops jararaca venom. Toxicon 37, 167–172. Kaethner MM, Good BF 1975. Seasonal thyroid activity in the tammar wallaby, Macropus eugenii (Desmarest). Australian Journal of Zoology 23, 363–369. Kaldor I, Ezekiel E 1962. Iron content of mammalian breast milk: measurements in the rat and in a marsupial. Nature, London 196, 175. Kanowski J, Hopkins MS, Marsh H, Winter JW 2001. Ecological correlates of folivore abundance in north Queensland rainforests. Wildlife Research 28, 1–8. Kavanagh RP 1987. Forest phenology and its effect on foraging behaviour and selection of habitat by the yellow-bellied glider, Petaurus australis Shaw. Australian Wildlife Research 14, 371–384. Kavanagh RP 1988. The impact of predation by the powerful owl, Ninox strenua, on a population of the greater glider, Petauroides volans. Australian Journal of Ecology 13, 445–450. Kavanagh RP, Lambert MJ 1990. Food selection by the greater glider, Petauroides volans: is foliar nitrogen a determinant of habitat quality? Australian Wildlife Research 17, 285–299. Kavanagh RP, Rohan-Jones WG 1982. Calling behaviour of the Yellow-bellied glider, Petaurus australis Shaw (Marsupialia: Petauridae). Australian Mammalogy 5, 95–111. Kavanagh RP, Wheeler RJ 2001. Home range of the greater glider, Petauroides volans, in tall montane forest of south-eastern New South Wales, and changes following logging. Proceedings of 47th meeting of Australian Mammal Society. p. 59. Kehl J, Borsboom A 1984. Home range, den tree use and activity patterns in the greater glider, Petauroides volans. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 229–236. Australian Mammal Society, Sydney. 402 Life of Marsupials

Kemp EM 1981. Tertiary palaeogeography and the evolution of Australian climate. In Ecological Biogeography of Australia. Ed. A Keast. pp. 33–49. Junk, The Hague. Kerle JA 1984. Variations in the ecology of Trichosurus: its adaptive significance. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 115–128. Australian Mammal Society, Sydney. Kerle JA 2001. Possums. The Brushtails, Ringtails and Greater Glider. p. 128. University of New South Wales Press, Sydney. Kerle JA, McKay GM, Sharman GB 1991. A systematic analysis of the brushtail possum, Trichosurus vulpecula (Kerr 1792) (Marsupialia: Phalangeridae). Australian Journal of Zoology 39, 313–331. Kerr JB, Hedger MP 1983. Spontaneous spermatogenic failure in the marsupial mouse Antechinus stuartii Macleay (Dasyuridae: Marsupialia). Australian Journal of Zoology 31, 445–466. Kielan-Jaworowska Z, Cifelli RL, Luo X 1998. Alleged Cretaceous placental from down under. Lethaia 31, 267–268. Kielan-Jaworowska Z, Cifelli RL, Luo Z-X 2004. Mammals from the Age of Dinosaurs: Origins, Structure, and Evolution. Columbia University Press, New York. King DR, Oliver AJ, Mead RJ 1978. The adaptation of some Western Australian mammals to food plants containing fluoroacetate. Australian Journal of Zoology 26, 699–712. Kinnear J, Brown GD 1967. Minimum heart rates of marsupials. Nature 215, 1501. Kinnear JE, Onus ML, Sumner NR 1998. Fox control and rock-wallaby population dynamics. II. An update. Wildlife Research 25, 81–88. Kinnear JE, Purohit KG, Main AR 1968. The ability of the tammar wallaby (Macropus eugenii, Marsupialia) to drink sea water. Comparative Biochemistry and Physiology 25, 761–782. Kirk J, Smith GC, Agnew G 2000. Trial radio-tracking of feathertail gliders Acrobates pygmaeus. Australian Mammalogy 22, 129–131. Kirkpatrick TH 1965. Studies on Macropodidae in Queensland. 2. Age estimation in the grey kangaroo, the red kangaroo, the eastern wallaroo and the red-necked wallaby, with notes on dental abnormali- ties. Queensland Journal of Agriculture and Animal Sciences 22, 301–317. Kirsch JAW 1977. The comparative serology of Marsupialia, and a classsification of marsupials. Australian Journal of Zoology Supplementary Series 52, 1–152. Kirsch JAW, Dickerman AW, Reig OA, Springer MS 1991. DNA hybridization evidence for the Australasian affinity of the American marsupial Dromiciops australis. Proceedings of the National Academy of Science 88, 10 465–10 469. Kirsch JAW, Lapointe F-J, Foeste A 1995. Resolution of portions of the kangaroo phylogeny (Marsupialia: Macropodidae) using DNA hybridization. Biological Journal of the Linnean Society 55, 309–328. Kirsch JAW, Lapointe F-J, Springer MS 1997. DNA-hybridization studies of marsupials and their implica- tions for metatherian classification. Australian Journal of Zoology 45, 211–280. Kirsch JAW, Palma RE 1995. DNA/DNA hybridization studies of carnivorous marsupials. V. A further estimate of relationships among opossums (Marsupialia: Didelphidae). Mammalia 59, 403–425. Kirsch JAW, Poole WE 1972. Taxonomy and distribution of the grey kangaroos, Macropus giganteus Shaw and Macropus fuliginosus (Desmarest), and their subspecies (Marsupialia: Macropodidae). Australian Journal of Zoology 20, 315–339. Kirsch JAW, Waller PF 1979. Notes on the trapping and behaviour of the Caenolestidae (Marsupialia). Journal of Mammalogy 60, 390–395. Kitchener DJ, Chapman A, Muir BG 1980. The conservation value for mammals of reserves in the Western Australian wheatbelt. Biological Conservation 18, 179–207. Klettenheimer BS, Temple-Smith PD, Sofronidis G 1997. Father and son sugar gliders: more than a genetic coalition? Journal of Zoology, London 242, 741–750. Klima M 1987. Early development of the shoulder girdle and sternum in marsupials (Mammalia: Metatheria). Advances in Anatomy Embryology and Cell Biology 109, 1–91. Kojima T, Hinds LA, Muller WJ, O’Neill C, Tyndale-Biscoe CH 1993. Production and secretion of proges- terone in vitro and presence of platelet activating factor (PAF) in early pregnancy of the marsupial, Macropus eugenii. Reproduction, Fertility and Development 5, 15–25. Körtner G, Geiser F 2000. Torpor and activity patterns in free-ranging sugar gliders Petaurus breviceps (Marsupialia). Oecologia 123, 350–357. Kraaijeveld-Smit FJL, Ward SJ, Temple-Smith PD, Paetkus D 2002. Factors influencing paternity in Antechinus agilis: last-male sperm precedence, timing of mating and genetic compatibility. Journal of Evolutionary Biology 15, 100–107. References 403

Krajewski C, Blacket MJ, Westerman M 2000a. DNA sequence analysis of familial relationships among dasyuromorphian marsupials. Journal of Mammalian Evolution 7, 95–108. Krajewski C, Buckley L, Westerman M 1997. DNA phylogeny of the marsupial wolf resolved. Proceedings of the Royal Society B 264, 911–917 Krajewski C, Woolley PA, Westerman M 2000b. The evolution of reproductive strategies in dasyurid marsupials: implications of molecular phylogeny. Biological Journal of the Linnean Society of London 71, 417–435. Krajewski C, Wroe S, Westerman M 2000c. Molecular evidence for the pattern and timing of cladogenesis in dasyurid marsupials. Zoological Journal of the Linnean Society 130, 375–404. Krause WJ, Cutts JH, Leeson CR 1979. Morphological observations on the metanephros in the postnatal opossum, Didelphis virginiana. Journal of Anatomy 129, 459–477. Krefft G 1862. On the vertebrated animals of the lower Murray and Darling, their habits, economy and geographical distribution. Transactions of the Philosophical Society of NSW 1862–1865, 1–33 Krockenberger A 2003. Meeting the energy demands of reproduction in female koalas, Phascolarctos cinereus: evidence for energetic compensation. Journal of Comparative Physiology B 173, 531–540. Krockenberger AK 1993. Energetics and nutrition during lactation in the koala, Phascolarctos cinereus. PhD Thesis, University of Sydney. Krockenberger AK 1996. Composition of the milk of the koala, Phascolarctos cinereus, an arboreal folivore. Physiological Zoology 69, 701–718. Krockenberger AK, Hume ID, Cork SJ 1998. Production of milk and nutrition of the dependent young of free-ranging koalas (Phascolarctos cinereus). Physiological Zoology 71, 45–56. Lambert MJ, Turner J, Kelly J 1983. Nutrient relationships of tree species in a New South Wales subtropi- cal rainforest. Australian Forest Research 13, 91–102. Langworthy OR 1925. The development of progression and posture in young opossums. American Journal of Physiology 74, 1–13. Laplante ES, Burrell R, Watne AL, Taylor DL, Zimmermann B 1969. Skin allograft studies in the pouch young of the opossum. Transplantation 7, 67–72. Laurance WF 1990. Comparative responses of five arboreal marsupials to tropical forest fragmentation. Journal of Mammalogy 71, 641–653. Laurance WF 1997. A distributional survey and habitat model for the endangered northern bettong Bettongia tropica in tropical Queensland. Biological Conservation 82, 47–60.

Lawler IR, Foley WJ, Woodrow IE, Cork SJ 1997. The effects of elevated CO2 atmospheres on the nutritional quality of Eucalyptus foliage and its interaction with soil nutrient and light availability. Oecologia 109, 59–68. Lawler IR, Foley WJ, Eschler BM 2000. Foliar concentration of a single toxin creates habitat patchiness for a marsupial folivore. Ecology 81, 1327–1338. Lazenby-Cohen KA, Cockburn A 1988. Lek promiscuity in a semelparous mammal, Antechinus stuartii (Marsupialia: Dasyuridae)? Behavioural Ecology and Sociobiology 22, 195–202. Leatherland JF, Renfree MB 1983. Structure of the pars distalis in pouch-young tammar wallabies (Macropus eugenii). Cell and Tissue Research 230, 587–603. Lee AK, Cockburn A 1985. Evolutionary Ecology of Marsupials. Cambridge University Press, Cambridge. Lee CS, O’Shea JD 1977. Observations on the vasculature of the reproductive tract in some Australian marsupials. Journal of Morphology 154, 95–114. Lemos B, Marriog G, Cerqueira R 2001. Evolutionary rates and stabilizing selection in large-bodied opossum skulls (Didelphimorphia: Didelphidae). Journal of Zoology, London 255, 181–189. Lima M, Stenseth, Yoccoz NG, Jaksic FM 2001. Demography and population dynamics of the mouse opossum (Thalamys elegans) in semi-arid Chile: seasonality, feedback structure and climate. Proceedings of the Royal Society of London B 268, 2053–2064. Lincoln DW, Renfree MB 1981. Mammary gland growth and milk ejection in the agile wallaby, Macropus agilis, displaying concurrent asynchronous lactation. Journal of Reproduction and Fertility 63, 193–203. Lincoln GA 1978. Plasma testosterone profiles in male macropod marsupials. Journal of Endocrinology 77, 347–351. Lindenmayer D 1996. Wildlife and Woodchips: Leadbeater’s possum a test case for sustainable forestry. University of New South Wales Press, Sydney. Lindenmayer D 2002. Gliders of Australia. A Natural History. pp. 160. University of New South Wales Press, Sydney. 404 Life of Marsupials

Lindenmayer DB, Cunningham RB, Pope M, Donnelly CF 1999a. The Tumut fragmentation experiment in south-eastern Australia: the effects of landscape context and fragmentation on arboreal marsupials. Ecological Applications 9, 594–611. Lindenmayer DB, Cunningham RB, Pope ML 1999b. A large scale “experiment” to examine the effects of landscape context and habitat fragmentation on mammals. Biological Conservation 88, 387–403. Lindenmayer DB, Possingham HP 1995. Ranking conservation and timber management options for Leadbeater’s possum in southeastern Australia using population viability analysis. Conservation Biology 10, 235–251. Lindenmayer DB, Taylor A 2003. Gliding ghosts of a forest past. Nature Australia 27, 30–37. Lintern SM, Barker S 1969. Renal retention of urea in the Kangaroo Island wallaby, Protemnodon eugenii (Demarest). Australian Journal of Experimental Biology and Medical Science 47, 243–250. Lintern-Moore S, Moore GPM, Tyndale-Biscoe CH, Poole WE 1976. The growth of oocyte and follicle in the ovaries of monotremes and marsupials. The Anatomical Record 185, 325–332. Liu GB, Hill KG, Mark RF 1997. Development of auditory sensitivity in the inferior colliculus of the tammar wallaby Macropus eugenii. In Acoustical Signal Processing in the Cental Auditory System. Ed. Ed Syka. pp. 287–293. Plenum Press, New York. Lobert B, Lee AK 1990. Reproduction and life history of Isoodon obesulus in Victorian heathland. In Bandicoots and bilbies. Eds JH Seebeck, RL Wallis and CM Kemper. pp. 311–318. Surrey Beatty, Sydney. Long J, Archer M, Flannery T, Hand S 2002. Prehistoric Mammals of Australia and New Guinea. p. 244. University of New South Wales Press, Sydney Lorini ML, DeOliveira JA, Persson VG 1994. Annual age structure and reproductive patterns in Marmosa incanca (Lund, 1841) (Didelphidae, Marsupialia). Zeitschrift für Saugertierkunde 59, 65–73. Loudon A, Rothwell N, Stock M 1985. Brown fat, thermogenesis and physiological birth in a marsupial. Comparative Biochemistry and Physiology A 81, 815–819. Loupis AK, Eldridge MDB 2001. Restriction fragment length polymorphisms (RFLP) analysis of three nuclear genes in rock-wallabies (Petrogale: Marsupialia: Macropodidae): a search for genic markers to identify taxa within the Petrogale lateralis–penicillata group. Australian Journal of Zoology 49, 27–36. Luckett WP, Woolley PA 1996. Ontogeny and homology of the dentition in dasyurid marsupials: develop- ment in Sminthopsis virginiae. Journal of Mammalian Evolution 3, 327–364. Lunney D 1994. Royal Commission of 1901 on the western lands of New South Wales – an ecologist’s summary. In Future of Western New South Wales. Eds D Lunney, S Hand, P Reed, D Butcher. pp. 221– 240. Royal Zoological Society of NSW, Mosman. Lunney D, Leary T 1988. The impact on native mammals of land-use changes and exotic species in the Bega district, New South Wales, since settlement. Australian Journal of Ecology 13, 67–92. Luo Z-X, Ji Q, Wible JR, Yuan C-X 2003. An early Cretaceous tribosphenic mammal and metatherian evolution. Science 302, 1934–1940. Lustick S, Lustick DD 1972. Energetics in the opossum. Didelphis marsupialis virginiana. Comparative Biochemistry and Physiology 43A, 643–647. Lyne AG 1974. Gestation period and birth in the marsupial Isoodon macrourus. Australian Journal of Zoology 22, 303–309. Lyne AG, Hollis DE 1977. The early development of marsupials, with special reference to bandicoots. In Reproduction and Evolution. Eds JH Calaby and CH Tyndale-Biscoe. pp. 293–302. Australian Academy of Science, Canberra. Main AR 1961. The occurrence of Macropodidae on islands and its climatic and ecological implications. Journal of the Royal Society of Western Australia 44, 84–89. Main AR, Yadav M 1971. Conservation of macropods in reserves in Western Australia. Biological Conservation 3, 123–133. Mallick J, Stoddart DM, Jones I, Bradley AJ 1994. Behavioural and endocrinological correlates of social status in the male sugar glider (Petaurus breviceps Marsupialia: Petauridae). Physiology and Behaviour 55, 1131–1134. Mann G 1958. Reproduccion de Dromiciops australis (Marsupialia, Didelphidae). Investigaciones Zoologicas Chilenas 4, 209–213. Mark RF 1997. Hugh Tyndale-Biscoe and the development of neurobiology of vision at the Australian National University. Chapter 18 In Marsupial Biology: Recent Research and New Perspectives. Eds NR Saunders and LA Hinds. pp. 311–326. University of New South Wales Press, Sydney. References 405

Mark RF, Marotte LR 1992. Australian marsupials as models for the developing mammalian visual system. Trends in Neurosciences 15, 51–57. Marotte LR, Sheng XM 2000. Neurogenesis and identification of developing layers in the visual cortex of the wallaby (Macropus eugenii). Journal of Comparative Neurology 416, 131–142. Marshall LG 1977. Lestodelphys halli. Mammalian Species 81, 1–3. Marshall LG 1978a. Dromiciops australis. Mammalian Species 99, 1–5. Marshall LG 1978b. Chironectes minimus. Mammalian Species 109, 1–6. Marshall LG 1982. Evolution of South American Marsupialia. In Mammalian biology in South America. Eds MA Mares and HH Genoways. Special Publications Series, Pymatuning Laboratory of Ecology, University of Pittsburgh, Vol. 6, 251–272. Marshall LG, Butler RF, Drake RE, Curtis GH, Tedford RH 1979. Calibration of the great American inter- change. Science 204, 272–279. Marshall LG, Case JA, Woodburne MO 1990. Phylogenetic relationships of the families of marsupials. Current Mammalogy 2, 433–505. Marshall LG, Muizon C de 1988. The dawn of the age of mammals in South America. National Geographic Research 4, 23–55. Martin HA 1981. The Tertiary flora. In Ecological Biogeography of Australia. Ed. A Keast. pp. 393–426. Junk, The Hague. Martin JK 2000. The mating system of the bobuck, Trichosurus caninus Ogilby (Marsupialia: Phalangeridae). Honours Thesis in Zoology, University of Melbourne, Victoria. Martin PS, Klein RG (Eds) 1984. Quartenary Extinctions, a Prehistoric Revolution. p. 892. University of Arizona Press, Tuscon. Martin R 1997. Managing over-abundance in koala populations in south-eastern Australia – future options. Australian Biologist 10, 57–63. Martin R, Handasyde K 1999. The Koala. Natural History, Conservation and Management. University of New South Wales Press, Sydney. Martin RW 1985. Overbrowsing, and decline of a population of the koala, in Victoria. I. Food preference and food tree defoliation. Australian Wildlife Research 12, 355–365. Martin RW, Lee AK 1984. The koala, Phascolarctos cinereus, the largest marsupial folivore. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 463–467. Australian Mammal Society, Sydney. Mason R 1958. Foods of the Australian opossum (Trichosurus vulpecula Kerr) in New Zealand indigenous forest in the Orongorongo Valley, Wellington. New Zealand Journal of Science 7, 590–613. Maynes GM 1989. Zoogeography of the Macropodoidea. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 47–66. Surrey Beatty and Sons, Sydney. McAllan BM, Dickman CR 1986. The role of photoperiod in the timing of reproduction in the dasyurid marsupial Antechinus stuartii. Oecologia 68, 259–264 McCarron HCK, Dawson TJ 1989. Thermal relations of the Macropodoidea in hot environments. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 255–263. Surrey Beatty & Sons, Sydney. McCrady E 1938. The embryology of the opossum. American Anatomical Memoirs 16, 1–233. McIlroy JC 1973. Aspects of the ecology of the common wombat, Vombatus ursinus (Shaw, 1800). PhD Thesis, Australian National University, Canberra. McIlroy JC 1976. Aspects of the ecology of the common wombat, Vombatus ursinus. I. Capture, handling, marking and radio-tracking techniques. Australian Wildlife Research 3, 105–116. McIlroy JC 1977. Aspects of the ecology of the common wombat, Vombatus ursinus. II. Methods for estimating population numbers. Australian Wildlife Research 4, 223–228. McIlroy JC, Rishworth C 1998. The effect on common wombat, Vombatus ursinus, populations of replac- ing native eucalypt forests with plantations of Monterey pine Pinus radiata. In Wombats. Eds RT Wells and PA Pridmore. pp. 271–279. Surrey Beatty, Chipping Norton. McIlwee AP, Johnson CN 1998. The contributions of fungus to the diets of three mycophagous marsupi- als in Eucalyptus forests, revealed by stable isotope analysis. Functional Ecology 12, 223–231. McKay GM 1989. Family Petauridae. In Fauna of Australia. Mammalia. Eds DW Walton and BJ Richardson. pp. 665–678. Australian Government Publishing Service, Canberra. McKenzie LM, Cooper DW 1997. Hybridization between tammar wallaby (Macropus eugenii) populations from Western and South Australia. Journal of Heredity 88, 398–400. 406 Life of Marsupials

McNab BK 1986. Food habits, energetics, and the reproduction of marsupials. Journal of Zoology, London 208, 595–614. McNab BK 1988. Complications inherent in scaling the basal rate of metabolism in mammals. Quarterly Review of Biology 63, 25–54 Meads, MJ 1976. Effects of opossum browsing on northern rata trees in the Orongorongo Valley, Wellington, New Zealand. New Zealand Journal of Zoology 3, 127–139. Menkhorst PW, Weavers BW, Alexander JSA 1988. Distribution, habitat and conservation status of the squirrel glider, Petaurus norfolcensis (Petauridae: Marsupialia) in Victoria. Australian Wildlife Research 15, 59–71. Merchant J, Marsh H, De’ath G, Spencer PBS 1996. Milk composition and production in free-living allied rock-wallabies Petrogale assimilis. Australian Journal of Zoology 44, 659–674. Merchant JC 1979. The effect of pregnancy on the interval between one oestrus and the next in the tammar wallaby, Macropus eugenii. Journal of Reproduction and Fertility 56, 459–463. Merchant JC 1989. Lactation in macropodoid marsupials. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 355–366. Surrey Beatty & Sons, Sydney. Merchant JC 1990. Aspects of lactation in the Northern brown bandicoot Isoodon macrourus. In Bandicoots and bilbies. Eds JH Seebeck, RL Wallis and CM Kemper. pp. 219–228. Surrey Beatty, Sydney. Merchant JC, Calaby JH 1981. Reproductive biology of the red-necked wallaby (Macropus rufogriseus banksianus) and Bennett’s wallaby (Macropus r. rufogriseus) in captivity. Journal of Zoology, London 194, 203–217. Merchant JC, Libke JA 1988. Milk composition in the northern brown bandicoot, Isoodon macrourus (Gould) in captivity. Australian Journal of Biological Science 41, 495–505. Merchant JC, Libke JA, Newgrain K 1996. Milk consumption and estimates of growth energetics in pouch young of the northern brown bandicoot, Isoodon macrourus (Peramelidae, Marsupialia), in captivity. Journal of Zoology, London 238, 483–494. Merchant JC, Sharman GB 1966. Observations on the attachment of marsupial pouch young to the teats and on the rearing of pouch young by foster mothers of the same or different species. Australian Journal of Zoology 14, 593–609. Merrilees D 1984. Comings and goings of late Quarternary mammals in extreme Southwestern Australia. In Quartenary Extinctions, a Prehistoric Revolution. Eds PS Martin and RG Klein. pp. 629–638. University of Arizona Press, Tuscon. Messer M, Crisp EA, Czoli R 1989. Lactose digestion in suckling macropodids. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 217–221. Surrey Beatty and Sons, Sydney. Messer M, Green B 1979. Milk carbohydrates of marsupials. II. Quantitative and qualitative changes in milk carbohydrates during lactation in the tammar wallaby (Macropus eugenii). Australian Journal of Biological Science 32, 519–531. Milewska A 2003. Marsupial lion or pouch-raiding possum? Australasian Science 24, 14–17. Miller JFAP, Block M, Rowlands DT, Kind P 1965. Effect of thymectomy on hemopoietic organs of the opossum ‘embryo.’ Proceedings of the Society for Experimental Biology and Medicine 118, 916–921. Mitchell P 1990. The home ranges and social activity of koalas – a quantitative analysis. In Biology of the Koala. Eds AK Lee, K. Handasyde and GD Sanson. pp. 171–187. Surrey Beatty, Sydney. Montague TL (Ed.) 2000. The Brushtail Possum: Biology, Impact and Management of an Introduced Marsupial. p. 292. Manaaki Whenua Press, Lincoln, New Zealand. Moore BD, Wallis IR, Marsh KJ, Foley WJ 2004. The role of nutrition in the conservation of the marsupial folivores of eucalypt forests. In Conservation of Australia’s Forest Fauna. 2nd Edn. Ed. D Lunney. pp. 549–575. Royal Zoological Society of New South Wales, Mosman. Moore GPM 1978. Embryonic diapause in the marsupial Macropus eugenii. Stimulation of nuclear RNA polymerase activity in the blastocyst during resumption of development. Journal of Cellular Physiology 94, 31–36. Moore HDM 1996. Gamete biology of the new world marsupial, the grey short-tailed opossum, Monodelphis domestica. Reproduction Fertility and Development 8, 605–615. Moore SJ, Sanson GD 1995. A comparison of the molar efficiency of two insect-eating mammals. Journal of Zoology, London 235, 175–192. Morrison PR 1965. Body temperatures in some Australian mammals. 4. Dasyuridae. Australian Journal of Zoology 13, 173–187. References 407

Mortola JP, Frappell PB, Woolley PA 1999. Breathing through skin in a newborn marsupial. Nature London 397, 660. Morton SR 1978a. Torpor and nest-sharing in free-living Sminthopsis crassicaudata (Marsupialia) and Mus musculus (Rodentia). Journal of Mammalogy 59, 569–575. Morton SR 1978b. An ecological study of Sminthopsis crassicaudata (Marsupialia: Dasyuridae). II. Behaviour and social organisation. Australian Wildlife Research 5, 163–182. Morton SR 1978c. An ecological study of Sminthopsis crassicaudata (Marsupialia: Dasyuridae). III. Reproduction and Life History. Australian Wildlife Research 5, 183–211. Morton SR 1990. The impact of European settlement on the vertebrate animals of arid Australia: a conceptual model. Proceedings of the Ecological Society of Australia 16, 201–13. Morton SR, Denny MJS, Read DG 1983. Habitat preferences and diets of sympatric Sminthopsis crassicau- data and S. macroura (Marsupialia: Dasyuridae). Australian Mammalogy 6, 29–34 Muirhead J 1999. Bandicoot diversity and evolution (Peramelomorphia, Marsupialia): the fossil evidence. Australian Mammalogy 21, 11–13. Mulvaney J, Kamminga J 1999. Prehistory of Australia. p. 480. Allen and Unwin, Sydney. Munks SA, Green B 1995. Energy allocation for reproduction in a marsupial arboreal folivore, the common ringtail possum (Pseudocheirus peregrinus). Oecologia 101, 94–104. Munks SA, Green B 1997. Milk consumption and growth in a marsupial arboreal folivore, the common ringtail possum, Pseudocheirus peregrinus. Physiological Zoology 70, 691–700. Munks SA, Green B, Newgrain K, Messer M 1991. Milk composition in the common ringtail possum, Pseudocheirus peregrinus (Petauridae: Marsupialia). Australian Journal of Zoology 39, 403–416. Munn AJ, Dawson TJ 2001. Thermoregulation in juvenile red kangaroos (Macropus rufus) after pouch exit: higher metabolism and evaporative water requirements. Physiological and Biochemical Zoology 74, 917–927. Munn AJ, Dawson TJ 2003a. How important is milk for near-weaned red kangaroos (Macropus rufus) fed different forages? Journal of Comparative Physiology B 173, 141–148. Munn AJ, Dawson TJ 2003b. Energy requirements of the red kangaroo (Macropus rufus): impacts of age, growth and body size in a large desert-living herbivore. Journal of Comparative Physiology B 173, 575–582. Murray P 1991. The Pleistocene megafauna of Australia. Chapter 24. In Vertebrate Palaeontology of Australasia. Eds P Vickers-Rich, JM Monaghan, RF Bird, TH Rich. pp. 1071–1164. Monash University Publications, Melbourne. Murray PF 1998. Palaeontology and palaeobiology of wombats. In Wombats. Eds RT Wells and PA Pridmore. pp. 1–33. Surrey Beatty, Chipping Norton. Myers, K 1986. Introduced vertebrates in Australia, with emphasis on the mammals. In Ecology of Biological Invasions. Eds RH Groves and JJ Burdon. pp. 120–136. Cambridge University Press, Cambridge. Myers T, Archer M, Krikman A, Pledge N 1999. Diversity and evolutionary relationships of Ilariids, Wynyardids, Vombatids and related groups of marsupials. Australian Mammalogy 21, 18–19. Nagy KA, Lee AK, Martin RW, Fleming MR 1988. Field metabolic rate and food requirement of a small dasyurid marsupial, Sminthopsis crassicaudata. Australian Journal of Zoology 836, 293–299. Nagy KA, Martin RW 1985. Field metabolic rate, water flux, food consumption and time budget of koalas, Phascolarctos cinereus (Marsupialia: Phascolarctidae) in Victoria. Australian Journal of Zoology 33, 655–665. Nagy KA, Meienberger C, Bradshaw SD, Wooller RD 1995. Field metabolic rate of a small marsupial mammal, the honey possum (Tarsipes rostratus). Journal of Mammalogy 76, 862–866. Nagy KA, Suckling GC 1985. Field energetics and water balance of sugar gliders, Petaurus breviceps (Marsupialia: Petauridae). Australian Journal of Zoology 33, 683–691. Nardone RM, Wilber CG, Musacchia XJ 1955. Electrocardiogram of the opossum during exposure to cold. American Journal of Physiology 181, 352–356. Nelson JE, Gemmell RT 2003. Birth in the northern quoll, Dasyurus hallucatus (Marsupialia: Dasyuridae). Australian Journal of Zoology 51, 187–198. Newsome AE 1964. Oestrus in the lactating red kangaroo Megaleia rufa (Desmarest). Australian Journal of Zoology 12, 315–321. Newsome AE 1965. The distribution of red kangaroos, Megaleia rufa (Desmarest), about sources of persistent food and water in Central Australia. Australian Journal of Zoology 13, 289–299. 408 Life of Marsupials

Newsome AE 1971. Competition between wildlife and domestic stock. Australian Veterinary Journal 47, 577–586. Newsome AE 1975. An ecological comparison of the two arid-zone kangaroos of Australia, and their anomalous prosperity since the introduction of ruminant stock to their environment. The Quarterly Review of Biology 50, 389–424. Newsome AE 1980. The eco-mythology of the red kangaroo in central Australia. Mankind 12, 327–333. Newsome AE, McIlroy J, Catling P 1975. The effects of an extensive wildfire on populations of twenty ground vertebrates in South-East Australia. Proceedings of the Ecological Society of Australia 9, 107– 123. Nicholas KR 1988. Control of milk protein synthesis in the marsupial Macropus eugenii: a model system to study prolactin-dependent development. In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. Chapter 6, pp. 68–85. Springer-Verlag, Berlin. Nicholas KR, Tyndale-Biscoe CH 1985. Prolactin-dependent accumulation of _-lactalbumin in mammary gland explants from the pregnant tammar wallaby (Macropus eugenii). Journal of Endocrinology 106, 337–342. Nicholson PJ 1963. Wombats. Timbertop Magazine, Geelong Grammar School 8, 32–38. Noble JC 1993. Relict surface-soil features in semi-arid mulga (Acacia aneura) woodlands. Rangelands Journal 15, 48–70. Noble JC 1997. The Delicate and Noxious Scrub. p. 137. CSIRO, Canberra. Noble JC 1999. Fossil features of mulga Acacia aneura landscapes: possible imprinting by extinct Pleistocene fauna. Australian Zoologist 31, 396–402. Norbury GL, Coulson GM, Walters BL 1988. Aspects of the demography of the western grey kangaroo, Macropus fuliginosus melanops, in semiarid north-west Victoria. Australian Wildlife Research 15, 257–266. O’Brien TP, Lomdahl A, Sanson GD 1986. Preliminary microscopic investigations of the digesta derived from foliage of Eucalyptus ovata (Labill.) in the digestive tract of the common ringtail possum, Pseudocheirus peregrinus (Marsupialia). Australian Journal of Zoology 34, 157–176. O’Connell MA 1989. Population dynamics of neotropical small mammals in seasonal habitats. Journal of Mammalogy 70, 532–548. O’Neill RJW, Eldridge MDB, Graves JAM 2001. Chromosome heterozygosity and de novo chromosome rearrangements in mammalian interspecies hybrids. Mammalian Genome 12, 256–259. O’Neill RJW, Eldridge MDB, Toder R, Ferguson-Smith MA, O’Brien PC, Graves JAM 1999. Chromosome evolution in kangaroos (Marsupialia: Macropodidae): cross species chromosome painting between the tammar wallaby and rock wallaby spp. With the 2n=22 ancestral macropodid karyotype. Genome 42, 525–530. Oftedal OT 1984. Milk composition, milk yield and energy output at peak lactation: a comparative review. Symposia of the Zoological Society of London 51, 33–86. Old JM, Deane EM 2000. Development of the immune system and immunological protection in marsu- pial pouch young. Developmental and Comparative Immunology 24, 445–454. Oliveira ME, Santori RT 1999. Predatory behaviour of the opossum Didelphis albiventris on the pitviper Bothrops jararaca. Studies on Neotropical Fauna and Environment 34, 72–75. Oliver AJ 1986. Social organisation and dispersal in the red kangaroo. PhD Thesis, University of Western Australia. Oliver AJ, King DR, Mead RJ 1979. Fluoroacetate tolerance, a genetic marker in some Australian marsupi- als. Australian Journal of Zoology 27, 363–372. Opie A, Gullan P, Mansergh I 1990. Predictions of the geographic range and habitat preferences of Isoodon obesulus and Perameles nasuta in Gippsland. In Bandicoots and Bilbies. Eds JH Seebeck, RL Wallis and CM Kemper. pp. 327–334. Surrey Beatty, Sydney. Osawa R, Blanchard WH, Callaghan PG 1993. Microbiological studies of the intestinal microflora of the koala, Phascolarctos cinereus. II. Pap, a special maternal faeces consumed by juvenile koalas. Australian Journal of Zoology 41, 611–620. Osgood WH 1921. A monographic study of the American marsupial, Caenolestes. Field Museum of Natural History, Zoological Series 14, 1–162. Paddle R 2001. The Last Tasmanian Tiger. The History and Extinction of the Thylacine. p. 273. Cambridge University Press, Cambridge. References 409

Pahl LI, Hume ID 1990. Preferences for Eucalyptus species of the New England Tablelands and initial development of an artificial diet for koalas. In Biology of the Koala. Eds AK Lee, KA Handasyde and GD Sanson. pp. 123–128. Surrey Beatty, Sydney. Pallas PS 1766. Didelphis orientalis. Miscellanea Zoologica 12, 224. Palma RE 1997. Thylamys elegans. Mammalian Species 572, 1–4. Palma RE, Spotorno AE 1999. Molecular systematics of marsupials based on the rRNA 12S mitochondrial gene: The phylogeny of didelphimorphia and of the living fossil microbiotheriid Dromiciops gliroides Thomas. Journal of Molecular Phylogenetics and Evolution 13, 525–535 Parry LJ, Bathgate RAD, Shaw G, Renfree MB, Ivell R 1997a. Evidence for a local fetal influence on myometrial oxytocin receptors during pregnancy in the tammar wallaby (Macropus eugenii). Biology of Reproduction 56, 200–207. Parry LJ, Clark JM, Renfree MB 1997b. Ultrastructural localization of relaxin in the corpus luteum of the pregnant and early lactating tammar wallaby, Macropus eugenii. Cell and Tissue Research 290, 615–622. Parry LJ, Guymer FJ, Fletcher TP, Renfree MB 1996. Release of an oxytocic peptide at parturition in the marsupial, Macropus eugenii. Journal of Reproduction and Fertility 107, 191–198. Pass DM, Foley WJ, Bowden B 1998. Vertebrate herbivory on Eucalyptus – identification of specific feeding deterrents for common ringtail possums (Pseudocheirus peregrinus) by bioassay-guided fractionation of Eucalyptus ovata foliage. Journal of Chemical Ecology 24, 1513–1527. Patterson B, Pascual R 1972. The fossil mammal fauna of South America. In Evolution, Mammals, and Southern Continents. Eds A. Keast, FC Erk and B Glass. pp. 247–310. State University of New York Press, Albany. Patterson BD, Gallardo MH 1987. Rhyncholestes raphanurus. Mammalian Species 286, 1–5. Patton JL, Costa LP 2003. Molecular phylogeography and species limits in rainforest didelphid marsupials of South America. In Predators with Pouches. Eds M Jones, C Dickman, M Archer. pp. 63–82. CSIRO Publishing: Melbourne. Payton IJ, Forester L, Frampton CM, Thomas MD 1997. Response of selected tree species to culling of introduced Australian brushtail possums Trichosurus vulpecula at Waipoua forest, Northland, New Zealand. Biological Conservation 81, 247–255. Pearce AR, James AC, Mark RF 2000. Development of functional connections between thalamic fibres and the visual cortex of the wallaby revealed by current source density analysis in vivo. Journal of Comparative Neurology 418, 441–456. Phillips S, Zenger KR, Richardson BJ 2002. Are Sydney rabbits different? Australian Zoologist 32, 49–55. Pike IL 1981. Comparative studies of embryo metabolism in early pregnancy. Journal of Reproduction and Fertility, Supplement 29, 203–213. Pine RH, Dalby PL, Matson JO 1985. Ecology, postnatal development, morphometrics, and taxonomic status of the short-tailed opossum, Monodelphis dimidiata, an apparently semelparous annual marsu- pial. Annals of Carnegie Museum 54, 195–231. Piso W 1648. Historia naturalis Brasiliae. L Elzevirium, Amsterdam. Poole WE 1975. Reproduction in two species of grey kangaroos, Macropus giganteus Shaw and M. fuligi- nosus (Desmarest) II. Gestation, parturition and pouch life. Australian Journal of Zoology 23, 333–353. Poole WE, Wood JT, Simms NG 1991. Distribution of the tammar, Macropus eugenii, and the relation- ships of populations as determined by cranial morphometrics. Wildlife Research 18, 625–639. Pope LC, Sharp A, Moritz C 1996. Population structure of the yellow-footed rock-wallaby Petrogale xanthopus (Gray 1854) inferred from mtDNA sequences and microsatellite loci. Molecular Ecology 5, 629–640. Proske U 1980. Energy conservation by elastic storage in kangaroos. Endeavour 4, 147–154. Quin D, Goldingay R, Churchill S, Engel D 1996b. Feeding behaviour and food availability of the yellow- bellied glider in north Queensland. Wildlife Research 23, 637–646. Quin DG 1995. Population ecology of the squirrel glider (Petaurus norfolcensis) and the sugar glider (Petaurus breviceps) (Marsupialia: Petauridae) at Limeburners Creek, on the central north coast of New South Wales. Wildlife Research 22, 471–505. Quin DG, Smith AP, Norton TW 1996a. Eco-geographic variation in size and sexual dimorphism in sugar gliders and squirrel gliders (Marsupialia: Petauridae). Australian Journal of Zoology 44, 19–45. Ramsey D, Efford M, Cowan P, Coleman J 2002. Factors influencing annual variation in breeding by common brushtail possums (Trichosurus vulpecula) in New Zealand. Wildlife Research 29, 39–50. 410 Life of Marsupials

Randall D, Gannon B, Runciman S, Baudinette RV 1984. Gas transfer by the neonate in the pouch of the tammar wallaby, Macropus eugenii. In Respiration in Embryonic Vertebrates. Ed. R Seymour. Chapter 29, pp. 423–436. Junk, The Hague. Read DG 1984. Movements and home ranges of three sympatric dasyurids, Sminthopsis crassicaudata, Planigale gilesi and P. tenuirostris (Marsupialia), in semiarid western New South Wales. Australian Wildlife Research 11, 223–234. Recher HF, Lim L 1990. A review of current ideas of the extinction, conservation and management of Australia’s terrestrial vertebrate fauna. Proceedings of the Ecological Society of Australia 16, 287–301. Regidor HA, Gorostiague M 1996. Reproduction in the white-eared opossum (Didelphis albiventris) under temperate conditions in Argentina. Studies on Neotropical Fauna and Environment 31, 133–136. Reig OA, Kirsch JAW, Marshall LG 1987. Systematic relationships of the living and Neocenozoic American ‘opossum-like’ marsupials (Suborder Didelphimorphia), with comments on the classification of these and of the Cretaceous and Palaeogene New World and European metatherians. In Possums and Opossums: Studies in Evolution. Ed. M Archer. pp. 1–89. Surrey Beatty and Sons and the Royal Zoological Society of NSW, Chipping Norton, Australia. Renfree MB 1973a. Proteins in the uterine secretions of the marsupial Macropus eugenii. Developmental Biology 32, 41–49. Renfree MB 1973b. The composition of fetal fluids of the marsupial Macropus eugenii. Developmental Biology 32, 62–79. Renfree MB, Fletcher TP, Blanden DR, Lewis PR, Shaw G, Gordon K, Short RV, Parer-Cook E, Parer D 1989. Physiological and behavioural events around the time of birth in macropodid marsupials. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 323–337. Surrey Beatty, New South Wales. Renfree MB, Holt AB, Green SW, Carr J, Cheek DB 1982. Ontogeny of the brain in a marsupial, Macropus eugenii, throughout pouch life. I. Brain growth and structure. Brain, Behavior and Evolution 20, 57–71. Renfree MB, Lewis AM 1996. Cleavage in vitro in the marsupial Macropus eugenii. Reproduction Fertility and Development 8, 725–742. Renfree MB, Meier P, Teng C, Battaglia FC 1981. Relationship between amino acid intake and accretion in a marsupial, Macropus eugenii. Biology of the Neonate 40, 29–37. Renfree MB, O W-S, Short RV 1996a. Sexual differentiation of the urogenital system of the fetal and neonatal tammar wallaby, Macropus eugenii. Anatomy and Embryology 194, 111–134. Renfree MB, Parry LJ, Shaw G 1996b. Infusion with an oxytocin receptor antagonist delays parturition in a marsupial. Journal of Reproduction and Fertility 108, 131–138. Renfree MB, Shaw G 1996. Reproduction of a marsupial: from uterus to pouch. Animal Reproduction Science 42, 393–403. Renfree MB, Shaw G 1999. Embryo-maternal interactions after diapause in a marsupial. In Embryo Implantation: Molecular, Cellular and Clinical Aspects. Ed. DD Carson. pp. 54–66. Springer Verlag, New York. Renfree MB, Shaw G, Fletcher TP 1994. Evidence for the essential role of prostaglandins for parturition in a marsupial, Macropus eugenii. Journal of Reproduction and Fertility 102, 433–446. Rens W, O’Brien PCM, Yang F, Graves JAM, Ferguson-Smith MA 1999. Karyotype relationships between four distantly related marsupials revealed by reciprocal chromosome painting. Chromosome Research 7, 461–474. Rens W, O’Brien PCM, Yang F, Solanky N, Perelman P, Graphodatsky AS, Ferguson-Smith MWJ, Svartman M, De Leo AA, Graves JAM, Ferguson-Smith MA 2001. Karyotype relationships between distantly related marsupials from South America and Australia. Chromosome Research 9, 301–308. Rens W, O’Brien PCM, Fairclough H, Harman L, Graves JAM, Ferguson-Smith MA 2003. Reversal and convergence in marsupial chromosome evolution. Cytogenetic and Genome Research 102, 282–290. Retief JD, Krajewski C, Westerman M, Dixon GH 1995. The evolution of protamine P1 genes in dasyurid marsupials. Journal of Molecular Evolution 41, 549–555. Reynolds HC 1952. Studies on reproduction in the opossum (Didelphis virginiana virginiana). University of California Publications in Zoology 52, 223–283. Reynolds ML, Saunders NR 1988. Differentiation of the neocortex. In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. Chapter 8, pp. 101–116. Springer-Verlag, Berlin. References 411

Ribbons KA, Baudinette RV, McMurchie EJ 1989. The development of pulmonary surfactant lipids in a neonatal marsupial and the rat. Respiration Physiology 75, 1–10. Rich TH 1991. Monotremes, placentals, and marsupials: their record in Australia and its basis. In Vertebrate Palaeontology of Australasia. Eds P Vickers-Rich, JM Monaghan, RF Baird, TH Rich. Chapter 23, pp. 894 –1004. Monash University, Melbourne. Rich TH, Vickers-Rich P, Constantine A, Flannery TF, Kool L, van Klaveren N 1997. A tribosphenic mammal from the Mesozoic of Australia. Science 278, 1438–1442. Richardson KC, Wooller RD, Collins BG 1986. Adaptations to a diet of nectar and pollen in the marsupial Tarsipes rostratus (Marsupialia: Tarsipedidae). Journal of Zoology, London 208, 285–297. Ride WDL 1956. The affinities of Burramys parvus Broom a fossil phalangeroid marsupial. Proceedings of the Zoological Society of London 127, 413–429. Ride WDL 1959. Mastication and taxonomy in the macropodine skull. In Function and Taxonomic Importance. Ed. AJ Cain. pp. 33–59. Systematics Association Publication Number 3. Ride WDL, Pridmore PA, Barwick RE, Wells RT, Heady RD 1997. Towards a biology of Propleopus oscillans (Marsupialia: Propleopinae, Hypsiprymnodontidae). Proceedings of the Linnean Society of New South Wales 117, 243–328. Roberts RG, Flannery TF, Ayliffe LK, Yoshida H, Olley JM, Prideaux GJ, Laslett GM, Baynes A, Sith MA, Jones R, Smith BL 2001. New ages for the last Australian megafauna: continent-wide extinction about 46,000 years ago. Science 292, 1888–1892. Robinson AC, Canty P, Mooney T 1996. South Australia’s Offshore Islands. p. 531. Australian Heritage Commission, Canberra. Robinson AC, Smyth MFB 1976. The vertebrate fauna of Nuyts Archipelago, South Australia. Transactions of the Royal Society of South Australia 100, 171–176. Robinson KW, Morrison PR 1957. The reaction to hot atmospheres of various species of Australian marsupial and placental animals. Journal of Cellular and Comparative Physiology 49, 455–478. Rodger JC, Bedford JM 1982a. Induction of oestrus, recovery of gametes, and the timing of fertilization events in the opossum, Didelphis virginiana. Journal of Reproduction and Fertility 64, 159–169. Rodger JC, Bedford JM 1982b. Separation of sperm pairs and sperm-egg interaction in the opossum, Didelphis virginiana. Journal of Reproduction and Fertility 64, 171–179. Rodger JC, Fletcher TP, Tyndale-Biscoe CH 1985. Active anti-paternal immunization does not affect the success of marsupial pregnancy. Journal of Reproductive Immunology 8, 249–256. Rofe RH 1978. G-banded chromosomes and the evolution of the Macropodidae. Australian Mammalogy 2, 53–63. Romer AS 1966. Vertebrate Paleontology. 3rd Edn. Chicago University Press, Chicago. Rosenberg HI, Rose R 1999. Volar adhesive pads of the feathertail glider, Acrobates pygmaeus (Marsupialia: Acrobatidae). Canadian Journal of Zoology 77, 233–248. Rothery FM 1970. Atlas of Bundaleer Plains and Tatala. p. 55. ANU Press, Canberra. Rowlands D, La Via MF, Block MH 1964. The blood forming tissues and blood of the newborn opossum (Didelphys virginiana). 2. Ontogenesis of antibody formation to flagella of Salmonella typhi. Journal of Immunology 93, 157–164. Runciman SIC, Gannon BJ, Baudinette RV 1995. Central cardiovascular shunts in the perinatal marsupial. The Anatomical Record 243, 71–83. Russell EM 1973. Mother-young relations and early behavioural development in the marsupials Macropus eugenii and Megaleia rufa. Zeitschrift für Tierpsychologie 33, 163–203. Russell EM 1982. Patterns of parental care and parental investment in marsupials. Biological Reviews 57, 423–486. Russell EM, Renfree MB 1989. Tarsipedidae. In Fauna of Australia. Volume 1B. Mammalia. Eds DW Walton and BJ Richardson. Chapter 33, pp. 769–782. Australian Government Publishing Service, Canberra. Russell R 1984. Social behaviour of the yellow-bellied glider, Petaurus australis reginae, in North Queensland. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 343–353. Australian Mammal Society, Sydney. Ryser J 1992. The mating system and male mating success of the Virginia opossum (Didelphis virginiana) in Florida. Journal of Zoology, London 228, 127–139. 412 Life of Marsupials

Sadleir RMF, Warburton B 2001. Advances in New Zealand mammalogy 1990–2000: wallabies. Journal of the Royal Society of New Zealand 31, 7–14. Sadleir RMFS 1965. Reproduction in two species of kangaroo Macropus robustus and Megaleia rufa in the arid Pilbara region of Western Australia. Proceedings of the Zoological Society of London 145, 239–261. Sadleir RMFS, Tyndale-Biscoe CH 1977. Photoperiod and the termination of embryonic diapause in the marsupial Macropus eugenii. Biology of Reproduction 16, 605–608. Sadler LM, Ward SJ 1999. Coalitions in male sugar gliders: are they natural? Journal of Zoology, London 248, 91–96. Sander U, Short J, Turner B 1997. Social organisation and warren use of the burrowing bettong, Bettongia lesueur (Macropodoidea: Potoroidae). Wildlife Research 24, 143–157. Sanson GD 1980. The morphology and occlusion of the molariform teeth in some Macropodinae (Marsupialia: Macropodidae). Australian Journal of Zoology 28, 341–365. Sanson GD 1985. Functional dental morphology and diet selection in dasyurids (Marsupialia: Dasyuridae). Australian Mammalogy 8, 239–247. Sanson GD 1989. Morphological adaptations of teeth to diets and feeding in the Macropodoidea. In Kangaroos, Wallabies and Rat-Kangaroos. Eds G Grigg, P Jarman and I Hume. pp. 151–168. Surrey Beatty and Sons, Sydney. Santori RT, Astua de Moraes D, Grelle CEV, Cerqueira R 1997. Natural diet at a restinga forest and labora- tory food preferences of the opossum Philander frenata in Brazil. Studies on Neotropical Fauna and Environment 32, 12–16. Saunders DA 1994. The effects of habitat reduction and fragmentation on the mammals and birds of the Western Australian central wheatbelt: lessons for western New South Wales. In Future of the Fauna of Western New South Wales. Eds D Lunney, S Hand, P Reed and D Butcher. pp. 99–105. Royal Zoological Society of NSW, Sydney. Schmidt-Nielsen K 1964. Desert Animals. Physiological Problems of Heat and Water. p. 277. Clarendon Press, Oxford. Schubert I, Oud JL 1997. There is an upper limit of chromosome size for normal development of an organism. Cell 88, 515–520. Schultze-Westrum TG 1969. Social communicaton by chemical signals in flying (Petaurus breviceps papuansis). In Olfaction and Taste. Ed C Pfaffmann. pp. 268–277. Rockefeller University Press, New York. Scott LK, Hume ID, Dickman CR 1999. Ecology and population biology of long-nosed bandicoots (Perameles nasuta) at North Head, Sydney Harbour National Park. Wildlife Research 26, 805–821. Scott MP 1986. The timing and synchrony of seasonal breeding in the marsupial, Antechinus stuartii: interaction of environmental and social clues. Journal of Mammalogy 67, 551–560. Seebeck JH, Brown PR, Wallis RL, Kemper CM (Eds) 1990. Bandicoots and Bilbies. p. 392. Surrey Beatty & Sons, Sydney. Seebeck JH, Warneke RM, Baxter BJ 1984. Diet of the bobuck, Trichosurus caninus (Ogilby) (Marsupialia: Phalangeridae) in a mountain forest in Victoria. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 145–154. Australian Mammal Society, Sydney. Selwood L 1992. Mechanisms underlying the development of pattern in marsupial embryos. Current Topics in Developmental Biology 27, 175–233. Selwood L, Young GJ 1983. Cleavage in vivo and in culture in the dasyurid marsupial Antechinus stuartii (Macleay). Journal of Morphology 176, 43–60 Serena M (Ed.) 1994. Re-introduction biology of Australian and New Zealaand fauna. Surrey Beatty, Sydney. Serventy DL, Marshall AJ 1964. A natural history reconnaissance of Barrow and Montebello Islands, 1958. Division of Wildlife Research Technical Paper 6, 1–23. Setchell BP, Waites GMH 1969. Pulse attenuation and countercurrent heat exchange in the internal spermatic artery of some Australian marsupials. Journal of Reproduction and Fertility 20, 165–169. Setchell PJ 1974. The development of thermoregulation and thyroid function in the marsupial Macropus eugenii (Desmarest). Comparative Biochemistry and Physiology A 47, 1115–1121. Shang F, Ashwell KWS, Marotte LR, Waite PME 1997. Development of commissural neurons in the wallaby (Macropus eugenii). Journal of Comparative Neurology 387, 507–523. Sharman GB 1962. The initiation and maintenance of lactation in the marsupial, Trichosurus vulpecula. Journal of Endocrinology 25, 375–385. References 413

Sharman GB 1982. Karyotypic similarities between Dromiciops australis (Microbiotheridae, Marsupialia) and some Australian marsupials. In Carnivorous Marsupials. Ed. M. Archer. pp. 711–714. Royal Zoological Society of New South Wales, Sydney. Sharman GB, Berger PJ 1969. Embryonic diapause in marsupials. Advances in Reproductive Physiology 4, 211–240. U57 Sharman GB, Clark MJ 1967. Inhibition of ovulation by the corpus luteum in the red kangaroo, Megaleia rufa. Journal of Reproduction and Fertility 14, 129–137. Sharman GB, Hughes RL, Cooper DW 1990. The chromosomal basis of sex differentiation in marsupials. Australian Journal of Zoology 37, 451–466. Sharp A 1997. Insights into the dispersal patterns of yellow-footed rock-wallabies, Petrogale xanthopus. Australian Mammalogy 19, 229–238.

Shaw G 1983. Effect of PGF2_ on uterine activity, and concentration of 13, 14-dihydro–15-keto PGF2_ in peripheral plasma during parturition in the tammar wallaby (Macropus eugenii). Journal of Reproduction and Fertility 69, 429–436.

Shaw G 1990. Control of parturient behaviour by prostaglandin F2_ in the tammar, Macropus eugenii. Journal of Reproduction and Fertility 88, 335–342. Shaw G 1996. The uterine environment in early pregnancy in the tammar wallaby. Reproduction Fertility and Development 8, 811–818.

Shaw G, Gehring HM, Bell EC 1999. Production of prostaglandin F2_, and its metabolite by endometrium and yolk sac placenta in late gestation in the tammar wallaby, Macropus eugenii. Biology of Reproduction 60, 611–614. Shaw G, Renfree MB 1986. Uterine and embryonic metabolism after diapause in the tammar wallaby, Macropus eugenii. Journal of Reproduction and Fertility 76, 339–347. Shaw G, Renfree MB 2001. Fetal control of parturition in marsupials. Reproduction, Fertility and Development 13, 653–659. Shaw G, Renfree MB, Fletcher TP 1996. A role for glucocorticoids in parturition in a marsupial, Macropus eugenii. Biology of Reproduction 54, 728–733. Shield JW 1966. Oxygen consumption during pouch development of the macropod marsupial Setonix brachyurus. Journal of Physiology, London 187, 257–270. Shimmin GA 2001. Northern hairy-nosed wombat warren architecture. Report of findings. Unpublished report for Northern hairy-nosed wombat recovery team, Queensland. Shimmin GA, Skinner J, Baudinette RV 2002. The burrow architecture and environment of the southern hairy-nosed wombat (Lasiorhinus latifrons). Journal of Zoology, London 258, 469-477. Shimmin GA, Taggart DA, Temple-Smith PD 2000. Sperm competition and genetic diversity in the agile antechinus (Dasyuridae: Antechinus agilis). Journal of Zoology, London 252, 343–350. Shorey CD, Hughes RL 1973a. Cyclical changes in the uterine endometrium and peripheral plasma concentrations of progesterone in the marsupial Trichosurus vulpecula. Australian Journal of Zoology 21, 1–19. Shorey CD, Hughes RL 1973b. Development, function, and regression of the corpus luteum in the marsu- pial Trichosurus vulpecula. Australian Journal of Zoology 21, 477–489. Short J 1998. The extinction of rat-kangaroos (Marsupialia: Potoroidae) in New South Wales, Australia. Biological Conservation 86, 365–377. Short J 2000. Reintroducing endangered mammals: the Herrison Prong experience. In Biodiversity and the Re-introduction of Native Fauna at Uluru-Kata Tjuta National Park. Eds JS Gillen, R Hamilton, WA Low and C Creagh. pp. 75–82. Bureau of Rural Sciences, Canberra. Short J, Bradshaw SD, Giles J, Prince RIT, Wilson GR 1992. Reintroduction of macropods (Marsupialia: Macropodoidea) in Australia – a review. Biological Conservation 62, 189–204. Short J, Richards JD, Turner B 1998. Ecology of the western barred bandicoot (Perameles bougainville) (Marsupialia: Peramelidae) on Dorre and Bernier Islands, Western Australia. Wildlife Research 25, 567–586. Short J, Turner B 1991. Distribution and abundance of spectacled hare-wallabies and euros on Barrow Island, Western Australia. Wildlife Research 18, 421–429. Short J, Turner B 1992. The distribution and abundance of the banded and rufuous hare-wallabies, Lagostrophus fasciatus and Lagorchestes hirsutus. Biological Conservation 60, 157–166. Short J, Turner B 1993. The distribution and abundance of the burrowing bettong (Marsupialia: Macropodoidea). Wildlife Research 20, 525–534. 414 Life of Marsupials

Short J, Turner B 1999. Ecology of burrowing bettongs, Bettongia lesueur (Marsupialia: Macropodoidea), on Dorre and Bernier Islands, Western Australia. Wildlife Research 26, 651–669. Short J, Turner B, Majors C, Leone J 1997. The fluctuating abundance of endangered mammals on Bernier and Dorre Islands, Western Australia – conservation implications. Australian Mammalogy 20, 53–61. Short RV, Flint APF, Renfree MB 1985. Influence of passive immunization against GnRH on pregnancy and parturition in the tammar wallaby, Macropus eugenii. Journal of Reproduction and Fertility 75, 567–575. Shortridge GC 1910. An account of the geographical distribution of the marsupials and monotremes of south-west Australia, having special reference to the specimens collected during the Balston Expedition of 1904–1907. Proceedings of the Zoological Society of London 1909, 803–847. Sibley CG, Ahlquist JE 1986. Reconstructing bird phylogeny by comparing DNAs. Scientific American 254, 68–78. Smith A 1982a. Diet and feeding strategies of the sugar glider in temperate Australia. Journal of Animal Ecology 51, 149–166. Smith A 1982b. Is the striped possum (Dactylopsila trivirgata; Marsupialia, Petauridae) an arboreal anteater? Australian Mammalogy 5, 229–234. Smith A 1984a. Demographic consequences of reproduction, dispersal and social interaction in a popula- tion of Leadbeater’s possum. In Possums and Gliders. Eds A Smith and ID Hume. pp. 359–373. Australian Mammal Society, Sydney. Smith A 1984b. Diet of Leadbeater’s possum, Gymnobelideus leadbeateri (Marsupialia). Australian Wildlife Research 11, 265–273. Smith AP, Broome L 1992. The effects of season, sex and habitat on the diet of the mountain pygmy- possum (Burramys parvus). Wildlife Research 19, 755–768. Smith AP, Green SW 1987. Nitrogen requirements of the sugar glider (Petaurus breviceps), an omnivorous marsupial, on a honey-pollen diet. Physiological Zoology 60, 82–92. Smith AP, Lindenmayer D 1988. Tree hollow requirements of Leadbeater’s possum and other possums and gliders in timber production ash forests of the Victorian Central Highlands. Australian Wildlife Research 15, 347–362. Smith AP, Nagy KA, Fleming MR, Green B 1982. Energy requirements and water turnover in free-living Leadbeater’s possums. Gymnobelideus leadbeateri (Marsupialia: Petauridae). Australian Journal of Zoology 30, 737–749. Smith M 1982. Review of the thylacine (Marsupialia, thylacinidae). In Carnivorous Marsupials. Ed. M Archer. pp. 237–253. Royal Zoological Society of New South Wales, Sydney. Smith MJ, Brown BK, Frith HJ 1969. Breeding of the brush-tailed possum, Trichosurus vulpecula (Kerr), in New South Wales. CSIRO Wildlife Research 14, 181–193. Smith MJ, Bennett JH, Chesson CM 1978. Photoperiod and some other factors affecting reproduction in female Sminthospis crassicaudata (Gould) (Marsupialia: Dasyuridae) in captivity. Australian Journal of Zoology 26, 449–463. Smith RFC 1969. Studies on the marsupial glider, Schoinobates volans (Kerr). I. Reproduction. Australian Journal of Zoology 17, 625–636. Spencer PBS 1996. Coping with a naturally fragmented environment: a genetic and ecological study of the allied rock-wallaby, Petrogale assimilis. PhD Thesis, University of New England, Armidale, NSW. Spencer PBS, Adams M, Marsh H, Miller DJ, Eldridge MDB 1997. High levels of genetic variability in an isolated colony of rock-wallabies (Petrogale assimilis): evidence from three classes of molecular markers. Australian Journal of Zoology 45, 199–210. Spencer PBS, Horsup AB, Marsh H 1998. Enhancement of reproductive success through mate choice in a social rock-wallaby, Petrogale assimilis (Macropodidae) as revealed by microsatellite markers. Behavioral Ecology and Sociobiology 43, 1–9. Spindler RE, Renfree MB, Shaw G, Gardner DK 1998. Reactivating tammar wallaby blastocysts oxidize glucose. Biology of Reproduction 58, 1425–1431. Springer MS, Kirsch JAW, Case JA 1997. The chronicle of marsupial evolution. In Molecular Evolution and Adaptive Radiation. Eds TJ Givnish and KJ Sytsma. pp. 129–161. Cambridge University Press, Cambridge. Springer MS, Westerman M, Kavanagh JR 1998. The origin of the Australasian marsupial fauna and the phylogenetic affinities of the enigmatic monito del monte and marsupial mole. Proceedings of the Royal Society of London B 265, 2381–2386. References 415

Springer MS, Westerman M, Kirsch JAW 1994. Relationships among orders and families of marsupi- als based on 12S ribosomal DNA sequences and the timing of the marsupial radiation. Journal of Mammalian Evolution 2, 85–115 Spurr E 2000. Impacts of possum control on non-target species. In The Brushtail Possum. Ed. TL Montague. Chapter 16, pp. 175–186. Manaaki Whenua Press, Lincoln, New Zealand. Steele VR, Temple-Smith PD 1998. Physical structure of warrens of a small colony of southern hairy- nosed wombats Lasiorhinus latifrons. In Wombats. Eds RT Wells and PA Pridmore. pp. 113–124. Surrey Beatty, Chipping Norton. Stevens HN (Ed) 1930. New Light on the Discovery of Australia as revealed by the Journal of Captain Don DiegodePradoyTovar. Henry Stevens, Son and Stiles, London. Stewart F 1984. Mammogenesis and changing prolactin receptor concentrations in the mammary glands of the tammar wallaby (Macropus eugenii). Journal of Reproduction and Fertility 71, 141–148. Stodart E 1966. Management and behaviour of breeding groups of the marsupial Perameles nasuta Geoffroy in captivity. Australian Journal of Zoology 14, 611–623. Stoddart DM, Bradley AJ 1991. The frontal and gular dermal scent organs of the marsupial sugar glider (Petaurus breviceps). Journal of Zoology, London 225, 1–12. Stoddart DM, Braithwaite RW 1979. A strategy for utilization of regenerating heathland habitat by the brown bandicoot (Isoodon obesulus; Marsupialia, Peramelidae). Journal of Animal Ecology 48, 165–179. Storr GCC 1780. Prodromus Methodi Mammalium. Tubingen. Strahan R (Ed.) 1995. The Mammals of Australia. Revised edition. p. 756. Australian Museum/Reed New Holland, Sydney. Strehlow TGH 1947. Aranda Traditions. Melbourne University Press, Melbourne. Strehlow TGH 1971. Songs of Central Australia. p. 775. Angus & Robertson, Sydney. Streilein KE 1982a. The ecology of small mammals in the semi-arid Brazilian Caatinga. III. Reproductive biology and population ecology. Annals of the Carnegie Museum 51, 251–259. Streilein KE 1982b. Behavior, ecology, and distribution of South American marsupials. In Mammalian Biology in South America. Eds MA Mares and HH Genoways. Special Publications Series, Pymatuning Laboratory of Ecology, University of Pittsburgh, Vol. 6, 231–250. Suckling GC 1984. Population ecology of the sugar glider (Petaurus breviceps) in a system of fragmented habitats. Australian Wildlife Research 11, 49–75. Sunquist ME, Austad SN, Sunquist F 1987. Movement patterns and home range in the common opossum (Didelphis marsupialis). Journal of Mammalogy 68, 173–176. Szalay FS 1982. A new appraisal of marsupial phylogeny and classification. In Carnivorous Marsupials. Ed. M. Archer. pp. 621–640. Royal Zoological Society of New South Wales, Sydney. Taggart DA 1994. A comparison of sperm and embryo transport in the female reproductive tract of marsupial and eutherian mammals. Reproduction Fertility and Development 6, 451–472. Taggart DA, Steele VR, Schultz D, Dibben R, Dibben J, Temple-Smith PD 1998. Semen collection and cryopreservation in the southern hairy-nosed wombat Lasiorhinus latifrons: implications for conser- vation of the northern hairy-nosed wombat Lasiorhinus krefftii. In Wombats. Eds RT Wells and PA Pridmore. pp. 180–191. Surrey Beatty, Chipping Norton. Taggart DA, Temple-Smith PD 1991. Transport and storage of spermatozoa in the female reproductive tract of the brown marsupial mouse, Antechinus stuartii (Dasyuridae). Journal of Reproduction and Fertility 93, 97–110. Talamoni SA, Dias MM 1999. Population and community ecology of small mammals in southeastern Brazil. Mammalia 63, 167–181. Taylor AC, Cooper DW 1999. Microsatellites identify introduced New Zealand tammar wallabies (Macropus eugenii) as an ‘extinct’ taxon. Animal Conservation 2, 41–49. Taylor AC, Cowan PE, Fricke BL, Cooper DW 2000. Genetic analysis of the mating system of the common brushtail possum (Trichosurus vulpecula) in New Zealand farmland. Molecular Ecology 9, 869–879. Taylor GK 1998. A long-term population study of the southern hairy-nosed wombat Lasiorhinus latifrons at Moorunde Wildlife Reserve, South Australia. In Wombats. Eds RT Wells and PA Pridmore. pp. 198–205. Surrey Beatty, Chipping Norton, NSW. Taylor RJ 1982. Population composition of the wallaroo, Macropus robustus (Marsupialia: Macropodidae) in the New England tablelands of New South Wales. Australian Mammalogy 5, 221–223. Taylor RJ 1993. Observations on the behaviour and ecology of the common wombat Vombatus ursinus in northeast Tasmania. Australian Mammalogy 16, 1–7. 416 Life of Marsupials

Temple-Smith PD 1986. Marsupials: the Chilean connection. Monash Review 5/86, 1–2. Temple-Smith PD 1994. Comparative structure and function of marsupial spermatozoa. Reproduction Fertility and Development 6, 421–435. Thesleff I, Sharpe P 1997. Signalling networks regulating dental development. Mechanisms in Development 67, 111–123. Thompson HV, King CM (Eds) 1994. The European Rabbit: the History of a Successful Coloniser. p. 245. Oxford University Press, Oxford. Thomson JA, Owen WH 1964. A field study of the Australian ringtail possum, Pseudocheirus peregrinus (Marsupialia: Petauridae). Ecological Monographs 34, 27–52. Thorne A, Grun R, Mortimer G, Spooner NA, Simpson JJ, McCullock M, Taylor L, Curnoe D 1999. Australia’s oldest human remains: age of the Lake Mungo 3 skeleton. Journal of Human Evolution 36, 591–612. Tibben EA, Holland RAB, Tyndale-Biscoe CH 1991. Blood oxygen carriage in the marsupial, tammar wallaby (Macropus eugenii), at prenatal and neonatal stages. Respiration Physiology 84, 93–104. Tiver F, Andrew MH 1997. Relative effects of herbivory by sheep, rabbits, goats and kangaroos on recruit- ment and regeneration of shrubs and trees in eastern South Australia. Journal of Applied Ecology 34, 903–914. Toder R, O’Neill RJW, Weinberg J, O’Brien PCM, Voullaire L, Graves JAM 1997. Comparative chromo-

some painting between two marsupials: origin of an XX/XY1Y2 sex chromosome system. Mammalian Genome 8, 418–422. Towers PA, Shaw G, Renfree MB 1986. Urogenital vasculature and local steroid concentrations in the uterine branch of the ovarian vein of the female tammar wallaby (Macropus eugenii). Journal of Reproduction and Fertility 78, 37–47. Traill BJ, Lill A 1997. Use of tree hollows by two sympatric gliding possums, the squirrel glider, Petaurus norfolcensis and the sugar glider, P. breviceps. Australian Mammalogy 20, 79–88. Triggs B 1996. The Wombat. Common wombats in Australia. University of New South Wales Press, Sydney. Trott JF, Wilson MJ, Hovey RC, Shaw DC, Nicholas KR 2002. Expression of novel lipocalin-like protein gene is developmentally-regulated during lactation in the tammar wallaby, Macropus eugeni. Gene 283, 287–297. Troughton E 1957. Furred Animals of Australia. Angus & Robertson, Sydney. Tunbridge D 1991. The Story of the Flinders Ranges Mammals. p. 96. Kangaroo Press, Kenthurst, NSW. Turner V 1984a. Eucalyptus in the diet of the feathertail glider, Acrobates pygmaeus (Marsupialia: Burramyidae). Australian Wildlife Research 11, 77–81. Turner V 1984b. Banksia pollen as a source of protein in the diet of two Australian marsupials Cercartetus nanus and Tarsipes rostratus. Oikos 43, 53–61. Turner V, McKay GM 1989. Burramyidae. In Fauna of Australia. Volume 1B. Mammalia. Eds DW Walton and BJ Richardson. Chapter 27, pp. 652–664. Australian Government Publishing Service, Canberra. Tyndale-Biscoe CH 1965. The female urogential system and reproduction of the marsupial Lagostrophus fasciatus. Australian Journal of Zoology 13, 255–267. Tyndale-Biscoe CH 1973. Life of Marsupials. p. 254. Edward Arnold, London. Tyndale-Biscoe CH 1981. Evidence for relaxin in marsupials. In Relaxin. Eds G Bryant-Greenwood, H Niall and F Greenwood. pp. 225–232. North Holland/Elsevier, Amsterdam. Tyndale-Biscoe CH, Smith RFC 1969a. Studies on the marsupial glider Schoinobates volans (Kerr). II. Population structure and regulatory mechanisms. Journal of Animal Ecology, 38, 637–650. Tyndale-Biscoe CH, Smith RFC 1969b. Studies on the marsupial glider Schoinobates volans (Kerr). III. Response to habitat destruction. Journal of Animal Ecology, 38, 651–659. Tyndale-Biscoe CH, Hinds LA 1990. Control of seasonal breeding in the tammar and Bennett’s wallabies. In Progress in Comparative Endocrinology. Eds A Epple, CG Scanes and MH Stetson. pp. 659–667. Wiley Liss, New York. Tyndale-Biscoe CH, Hinds LA, Horn CA 1988. Fetal role in the control of parturition in the tammar, Macropus eugenii. Journal of Reproduction and Fertility 82, 419–428. Tyndale-Biscoe CH, Hinds LA, Horn CA, Jenkins G 1983. Hormonal changes at oestrus, parturition and post-partum oestrus in the tammar wallaby Macropus eugenii. Journal of Endocrinology 96, 155–161. Tyndale-Biscoe CH, Hinds LA, McConnell SJ 1986. Seasonal breeding in a marsupial: opportunities of a new species for an old problem. Recent Progress in Hormone Research 42, 471–512. References 417

Tyndale-Biscoe CH, Janssens PA (Eds) 1988. The Developing Marsupial: Models for Biomedical Research. Springer-Verlag, Heidelberg. Tyndale-Biscoe CH, MacKenzie RB 1976. Reproduction in Didelphis marsupialis and D. albiventris in Colombia. Journal of Mammalogy 57, 249–265. Tyndale-Biscoe CH, Renfree MB 1987. Reproductive Physiology of Marsupials. p. 476. Cambridge University Press, Cambridge. Tyndale-Biscoe CH, Rodger JC 1978. Differential transport of spermatozoa into the two sides of the genital tract of a monovular marsupial, the tammar wallaby (Macropus eugenii). Journal of Reproduction and Fertility 52, 37–43. Ullman SL, Shaw G, Alcorn GT, Renfree MB 1997. Migration of primordial germ cells to the developing gonadal ridges in the tammar wallaby Macropus eugenii. Journal of Reproduction and Fertility 110, 135–143. Ullrey DE, Robinson RT, Whetter PA 1981. Composition of preferred and rejected Eucalyptus browse offered to captive koalas, Phascolarctos cinereus (Marsupialia). Australian Journal of Zoology 29, 847–852. van der Ree R 2002. The population ecology of the squirrel glider (Petaurus norfolcensis) within a network of remnant linear habitats. Wildlife Research 29, 329–340. van der Ree R, Bennett AF 2003. Home range of the squirrel glider (Petaurus norfolcensis) in a network of remnant linear habitats. Journal of Zoology, London 259, 327–336. van Tets IG, Hulbert A 1999. A comparison of the nitrogen requirements of the eastern pygmy-possum, Cercartetus nanus, on a pollen and on a mealworm diet. Physiological and Biochemical Zoology 72, 127–137. VandeBerg JL 1983. The grey short-tailed opossum: A new laboratory animal. Institute of Laboratory Animal Resources News 26, 9–12. VandeBerg JL 1990. The gray short-tailed opossum (Monodelphis domestica) as a model didelphid species for genetic research. Australian Journal of Zoology 37, 235–247. VandeBerg JL, Robinson ES 1997. The laboratory opossum (Monodelphis domestica) in biomedical research. In Marsupial Biology. Recent Research, New Perspectives. Eds N Saunders and L Hinds. pp. 238–253. UNSW Press, Sydney. Veltman C 2000. Do native wildlife benefit from possum control? In The Brushtail Possum. Ed. TL Montague. Chapter 22, pp. 241–250. Manaaki Whenua Press Lincoln, New Zealand. Vernes K, Castellano M, Johnson CN 2001. Effects of season and fire on the diversity of hypogeous fungi consumed by a tropical mycophagous marsupial. Journal of Animal Ecology 70, 945–954. Vernes K, Marsh H, Winter J 1995. Home-range characteristics and movement patterns of the red-legged pademelon (Thylogale stigmatica) in fragmented tropical rainforest. Wildlife Research 22, 699–708. Vernes K, Pope LC 2001. Stability of nest range, home range and movement of the northern bettong (Bettongia tropica) following moderate-intensity fire in a tropical woodland, north-eastern Queensland. Wildlife Research 28, 141–150. Vernes K, Pope LC 2002. Fecundity, pouch young survivorship and breeding season of the northern bettong (Bettongia tropica) in the wild. Australian Mammalogy 23, 95–100. Viggers KL, Lindenmayer DB 2000. A populaton study of the mountain brushtail possum (Trichosurus caninus) in the central highlands of Victoria. Australian Journal of Zoology 48, 201–216. Waite PME, Marotte LR, Leamy CA, Mark RF 1998. Development of whisker-related patterns in marsupi- als: factors controlling timing. Trends in Neurosciences 21, 265–269. Waite PME, Weller WL 1997. Development of somatosensory pathways from the whiskers. In Marsupial Biology: Recent Research and New Perspectives. Eds NR Saunders and LA Hinds. Chapter 19, pp. 327– 344. University of New South Wales Press, Sydney. Walker KZ, Tyndale-Biscoe CH 1978. Immunological aspects of gestation in the tammar wallaby, Macropus eugenii. Australian Journal of Biological Science 31, 173–182. Wallis IR, Green B, Newgrain K 1997. Seasonal field energetics and water fluxes of the long-nosed potoroo (Potorous tridactylus) in southern Victoria. Australian Journal of Zoology 45, 1–11. Walsh GL, Morwood MJ 1999. Spear and spearthrower evolution in the Kimberley region, N. W. Australia: evidence from rock art. Archaeology in Oceania 34, 45–58. Warburton B 1990. Control of Bennett’s and tammar wallabies in New Zealand using compound 1080 gel on foliage baits. Australian Wildlife Research 17, 541–546. 418 Life of Marsupials

Ward JW 1954. The development of the cortico-spinal tract in the pouch young of the opossum. Journal of Comparative Neurology 101, 483–494. Ward SJ 1990. Life history of the eastern pygmy-possum, Cercartetus nanus (Marsupialia: Burramyidae), in South-eastern Australia. Australian Journal of Zoology 38, 287–304. Ward SJ, Renfree MB 1988a. Reproduction in females of the feathertail glider Acrobates pygmaeus (Marsupialia). Journal of Zoology, London 216, 225–239. Ward SJ, Renfree MB 1988b. Reproduction in males of the feathertail glider Acrobates pygmaeus (Marsupialia). Journal of Zoology, London 216, 241–251. Watt A 1997. Population ecology and reproductive seasonality in three species of Antechinus (Marsupialia: Dasyuridae) in the wet tropics of Queensland. Wildlife Research 24, 531–547. Webster KN, Dawson TJ 2003. Locomotion energetics and gait characteristics of a rat-kangaroo, Bettongia penicillata, have some kangaroo-like features. Journal of Comparative Physiology B 173, 549–557. Wellard GA, Hume ID 1981. Nitrogen metabolism and nitrogen requirements of the brushtail possum, Trichosurus vulpecula (Kerr). Australian Journal of Zoology 29, 147–157. Wells A, Yetton MD, Duncan RP, Stewart GH 1999. Prehistoric dates of the most recent Alpine fault earth- quakes, New Zealand. Geology 27, 995–998. Wells RT 1978. Thermoregulation and activity rhythms in the hairy-nosed wombat, Lasiorhinus latifrons (Owen), (Vombatidae). Australian Journal of Zoology 26, 639–651. Wells RT 1982. Personal Communication (Notebook 1982–83. Entry 26/11/1982 – Visiting Rod Wells at Flinders University). Wells RT 1989. Vombatidae. In Fauna of Australia. Mammalia. Vol 1B. Eds DW Walton and BJ Richardson. pp. 755–768. Australian Government Publishing Service: Canberra. Wells RT, Green B 1998. Aspects of water metabolism in the southern hairy-nosed wombat Lasiorhinus latifrons. In Wombats. Eds RT Wells and PA Pridmore. pp. 61–66. Surrey Beatty, Chipping Norton. Westerman M, Burk A, Amrine-Madsen HM, Prideaux GJ, Case JA, Springer MS 2002. Molecular evidence for the last survivor of an ancient kangaroo lineage. Journal of Mammalian Evolution 9, 209–223. Westerman M, Loke S, Springer MS 2004. Molecular phylogenetic relationships of two extinct potoroid marsupials, Potorous platyops and Caloprymnus campestris (Potoroidae: Marsupialia). Journal of Molecular Evolution 31, 476–485. Westerman M, Springer MS, Dixon J, Krajewski C 1999. Molecular relationships of the extinct pig-footed bandicoot Chaeropus ecaudatus (Marsupialia: Perameloidea) using 12S rRNA sequences. Journal of Mammalian Evolution 6, 271–288. Wilkes GE, Janssens PA 1988. The development of renal function. In The Developing Marsupial. Eds CH Tyndale-Biscoe and PA Janssens. Chapter 13, pp. 176–189. Springer-Verlag, Berlin. Williams CK 1982. Nutritional properties of some fruits eaten by the possum Trichosurus vulpecula in a New Zealand broadleaf-podocarp forest. New Zealand Journal of Ecology 5, 16–20. Williamson P, Fletcher TP, Renfree MB 1990. Testicular development and maturation of the hypotha- lamic-pituitary-testicular axis in the male tammar, Macropus eugenii. Journal of Reproduction and Fertility 88, 549–557. Wimborne BM, Mark RF, Ibbotson MR 1999. Distribution of retinogeniculate cells in the tammar wallaby in relation to decussation at the optic chiasm. Journal of Comparative Neurology 405, 128–140. Winter JW 1976. The behaviour and social organisation of the brush-tail possum (Trichosurus vulpecula Kerr). PhD Thesis, University of Queensland. Winter JW 1984. Conservation studies of tropical rainforest possums. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 469–481. Surrey Beatty, Chipping Norton, NSW. Wiseman GL, Hendrickson GO 1950. Notes on the life history and ecology of the opossum in south east Iowa. Journal of Mammalogy 31, 331–337. Withers PC, Richardson KC, Wooller RD 1990. Metabolic physiology of euthermic and torpid honey possums, Tarsipes rostratus. Australian Journal of Zoology 37, 685–693. Wodzicki KA 1950. Introduced mammals of New Zealand: an ecological and economic survey. DSIR Bulletin No. 98, 1–255. Government Printer, Wellington. Wood DH 1970. An ecological study of Antechinus stuartii (Marsupialia) in a south-east Queensland rain forest. Australian Journal of Zoology 19, 347–353. Woodburne MO, Case JA 1996. Dispersal, vicariance, and the Late Cretaceous to early Tertiary land mammal biogeography from South America to Australia. Journal of Mammalian Evolution 3, 121–161. References 419

Woods HA, Hellgren EC 2003. Seasonal changes in the physiology of male Virginia opossums (Didelphis virginiana): signs of the dasyurid semelparity syndrome? Physiological and Biochemical Zoology 76, 406–417. Wooller RD, Renfree MB, Russell EM, Dunning A, Green SW, Duncan P 1981. Seasonal changes in a population of the nectar-feeding marsupial Tarsipes spencerae (Marsupialia: Tarsipedidae). Journal of Zoology, London 195, 267–279. Wooller RD, Russell EM, Renfree MB 1984. Honey possums and their food plants. In Possums and Gliders. Eds AP Smith and ID Hume. pp. 439–443. Australian Mammal Society, Sydney. Wooller RD, Russell EM, Renfree MB, Towers PA 1983. A comparison of seasonal changes in the pollen loads of nectarivorous marsupials and birds. Australian Wildlife Research 10, 311–317. Woolley P 1966. Reproduction in Antechinus spp., and other dasyurid marsupials. Symposium of the Zoological Society of London 15, 281–294. Woolley PA 1975. The seminiferous tubules in dasyurid marsupials. Journal of Reproduction and Fertility 45, 255–261. Woolley PA 1982. Phallic morphology of the Australian species of Antechinus (Dasyuridae, Marsupialia): a new taxonomic tool? In Carnivorous Marsupials. Ed. M. Archer. pp. 767–781. Royal Zoological Society of New South Wales, Sydney. Woolley PA 1991. Reproduction in Dasykaluta rosamondae (Marsupialia: Dasyuridae): field and labora- tory observations. Australian Journal of Zoology 39, 549–568. Woolley PA 1992. New records of the Julia Creek Dunnart, Sminthopsis douglasi (Marsupialia: Dasyuridae). Wildlife Research 19, 779–783. Woolley PA 2002. Reproductive biology of dasyurid marsupials of New Guinea. In Predators with Pouches: the Biology of Carnivorous Marsupials. Eds M Jones, C Dickman and M Archer. pp. 169–182. CSIRO Publishing, Melbourne. Woolley PA, Watson MR 1984. Observations on a captive outdoor breeding colony of a small dasyurid marsupial, Sminthopsis crassicaudata. Australian Wildlife Research 11, 249–254. Woolley PA, Patterson MF, Stephenson GM, Stephenson DG 2002. The ilio-marsupialis muscle in the dasyurid marsupial Sminthopsis douglasi: form, function and fibre-type profiles in females with and without suckling young. Journal of Experimental Biology 205, 3775–3781. Woolnough AP, Steele VR 2001. The palaeoecology of the Vombatidae: did giant wombats burrow? Mammal Review 31, 33–45. Wright W, Sanson GD, McArthur C 1991. The diet of the extinct bandicoot, Chaeropus ecaudatus. In Vertebrate Palaeontology of Australia. Eds P Vicker-Rich, JM Monoghan, RF Baird and TH Rich. pp. 229–245. Pioneer Design, Melbourne. Wroe S, Brammall J, Cooke BN 1998. The skull of Ekaltadeta ima (Marsupialia, Hypsiprymnodontinae?): an analysis of some cranial features within Marsupialia and re-investigation of propleopine phylog- eny, with notes on the inference of carnivory among mammals. Journal of Palaeontology 72, 738–751. Wroe S, Field J 2001. Giant wombats and red herrings. Australasian Science 22, 18. Wroe S, Muirhead J 1999. Evolution of Australia’s Marsupicarnivores: Dasyuridae, Thylacinidae, Myrmecobiidae, Dasyuromorphia incertae sedis and Marsupialia incertae sedis. Australian Mammalogy 21, 10–11. Yadav M, Eadie M 1973. Passage of maternal immunoglobulins to the pouch young of a marsupial, Setonix brachyurus. Australian Journal of Zoology 21, 171–181. Yadav M, Stanley NF, Waring H 1972. The microbial flora of the gut of the pouch-young and the pouch of a marsupial, Setonix brachyurus. Journal of General Microbiology 70, 437–442. Young IR, Renfree MB 1979. The effects of corpus luteum removal during gestation on parturition in the tammar wallaby (Macropus eugenii). Journal of Reproduction and Fertility 56, 249–254. Yousef A, Selwood L 1996. Type and differentiation of cells in vitro from unilaminar and bilaminar blast- ocysts of two marsupials, Antechinus stuartii and Sminthopsis macroura. Reproduction, Fertility and Development 8, 743–752. Zeigler AC 1977. Evolution of New Guinea’s marsupial fauna in response to a forested environment. In Marsupial Biology. Eds B Stonehouse and D Gilmore. pp. 117–140. MacMillan Press, London. This page intentionally left blank Index 421

Index

Figure references in italics; Box , Table & Plate references in bold

Aardvark (aardwark), 271 Alphadon 16 Aboriginal people, American marsupials, 3, 19, 103–38 arrival in Australia, 367 diversity & abundance in tropical forests, 118–9 Dreamtime creation stories, 313–4, 367 genera listed, 115 effects on marsupials, 237 in forests of French Guiana, 119 fire regimes, 381 neotropical marsupials, 114–17 hunting efficiency, 373 neotropical rainforest environment, 118–30 information on pig-footed bandicoot, 171, 178 in restinga forest of Brazil, 119–20 predation on kangaroos & euros, 330 see also Central & North American marsupials; return of land to, 330 South American marsupials Abrolhos Islands, 4 amnion, development of, 60, 61 tammar population, 333, 334, 335 anaerobic fermentation, see fermentation see also Houtman’s Abrolhos archipelago anatomy, & the classification of marsupials, 14–20 Acacia sp., 32, 185, 208 ankle bones 18–19 Acacia aneura, 325 Antarctica, 31, 168 Acacia dealbata, 203 Antechinomys sp., 143, 144 Acrobates pygmaeus (feathertail glider), 28, 29, 183, Antechinomys laniger (kultarr), 143 187, 192, Plate 10, Antechinus sp.,143, 144, 145 behaviour & energy metabolism, 197, 199, 199, newborn, 7 200 habitats, 145 chromosome number, 192 oestrous cycle, 44 diet & daily activity, 195 reproductive patterns, 145, 155 distribution, 188, 197 Anthechinus (common names) feet pads, 194–5, 194 agile, see Antechinus agilis gliding membrane, 192, 193, 194 brown, see Antechinus stuartii hibernation & torpor, 198, 198, 199, 200 dusky, see Antechinus swainsonii home range, 196 fat-tailed false antechinus, see Pseudantechinus relationships with marsupial genera, 186, 196, macdonnellensis reproduction, 201 rusty, see Antechinus adustus embryonic diapause, 56 subtropical, see Antechinus subtropicus torpor & energy conservation, 199, 199, 202 swamp, see Antechinus minimus Acrobatidae, 22, 28, 188, 192 yellow-footed, see Antechinus flavipes acrosome, 53 Antechinus adustus (rusty antechinus), 145 adaptive radiations, 4–5 Antechinus agilis (agile antechinus), 145, 156, 156, in dasyurids, 145 157 in opossums, 137 Antechinus flavipes (yellow-footed antechinus), 145, adrenal glands in fetus, 64 147, 153, 156, Plate 7 adrenocorticotrophin hormone, ACTH, Antechinus macdonnellensis, see Pseudantechinus fetal, 66, 68 macdonnellensis Aepyprymnus sp., 28, 29, 359 Antechinus minimus (swamp antechinus), 145 Aepyprymnus rufescens (rufous bettong), 301 Antechinus stuartii (brown antechinus), 145, 150, 153 diet, 362 breeding control, 156–7, 157 distribution, 291, 293 distribution, 150 chromosome number, 21 energy requirements, 150 reproduction, 362 field metabolic rate, 154 Agrotis infusa, bogong moth, 197 male die off, 154 Ailurops sp., 186 numbers of, 153 Ailurops ursinus, bear cuscus, 196, 227, 229 reproduction, 40, 153, 154, albumins, 24 Antechinus subtropicus (subtropical anthechinus), Alectryon excelsa, (titoki), 261, 262 145, 153 allantoic placenta, 60, 75, 167, 174 Antechinus swainsonii (dusky antechinus), 145, 156 allantois, development of, 60, 61, Plate 3 birth posture, 70 Allen’s Rule, 132 day-length & reproduction, 40 422 Life of Marsupials

Aquila audax (wedge-tailed eagle), 349 hormonal control, 175, 176 Aranda people of Central Australia, 313–14 lactation, 174–5 Araucaria sp., 32 respiration in newborn, 75 arboreal folivores, special characters, 167–8 comparisons of diet, 232 status of, 179 energy metabolism, 233 teeth structure, 18, 167 fossil records, 227 bandicoots (common names) living representatives, 227 eastern barred, see Perameles gunnii relationships & past history, 227–9 golden, see Isoodon auratus reproductive data, 235 long-nosed, see Perameles nasuta size contraints, 224–6, 224 northern brown, see Isoodon macrourus see also forest folivores pig-footed, see Chaeropus ecaudatus arboreal marsupials, 27, rufus spiny, see Echymipera rufescens habitat of, 221, 249 southern brown, see Isoodon obesulus see also arboreal folivores; forest folivores western barred, see Perameles bougainville archeological sites in Australasia, 30, 369 Banksia sp., 185, 187, Argyrolagus sp., 108 nectar production, 189, 208 Aristida sp., 271 Banksia attenuata, 192 arms race, Banksia coccinea, 192 brushtail possums in New Zealand forests, Banksia nutans, 192 259–61 Banksiae, 32 see also folivore-eucalyptus arms races Barbourofelis sp. 106, 107 Arnhem Land, 369 Barrow Island, Western Australia, 344 Aru Island, south of New Guinea, 369, 370 marsupial populations, 354–5, 355 Astrebla pectinata (Mitchell grass), 325 rock wallaby population, 344, 355 auditory system, 90 Bassia diacantha (chenopod), 321 see also ears; hearing Beilschmiedia tawa (tawa), 261, 262 Ausktribosphenus, 27 Bergman’s Rule, 132, 206 Australian marsupials Bernier Island, Western Australia, 355–9 first European encounter, 3–4 Bettongia sp., 28, 300, 359 diet of, 359 bacteria, classification, 301 anerobic, 231 Bettongia gaimardi (Tasmanian bettong), 173 in kangaroo gut, 289, 303 bounty payments, 379 in koala colon and caecum, 231 diet of, 360 in wombat colon, 273 distribution of, 291, 292 for fermentation, 22, 225, 239 maternal investment, 174 for protein synthesis, 225 Bettongia lesueur (burrowing bettong), 27, Plate 14 transfer to young, 99, 235, 281, 329 Bernier & Dorre island populations, 355–9, 356 bacterial infections in the gut, 97 bounty payments, 379 Balbaridae, 302 distribution, 219, 293, 355, 355–9, 358 Bandicoota bengaliensis (large Indian rat), 167 social behaviour, 357–8 bandicoots, 165–81 skull & dentition, 292 antecedents, 168–9 Bettongia penicillata (brush-tailed bettong), as pests, 379 bounty payments, 379 Australian & New Guinean relationships, 169 diet of, 361 behaviour of, 167 distribution, 291, 292–3, 359 conservation of, 178–80 fertilisation, 52 diet of, 167, 172 locomotion & oxygen consumption, 309, 309 distribution of, 167 Bettongia tropica (northern bettong), fossils, 32, 169 diet of, 362 growth rate & longevity, 174, 179 distribution, 291, 359 metabolic rate, 173 habitat, 362, 363 on small islands, 180–1 reproduction, 362 physiology of, 172–3 social behaviour, 362 relic populations, 180 survival after low intensity fires, 362 reproduction of, 173–4, 178 bettongs (common names) breeding seasons, 176 brush-tailed, see Bettongia penicillata control of birth, 69 burrowing or boodie, see Bettongia lesueur Index 423

northern, see Bettongia tropica synchronised, 40 rufous, see Aepyprymnus rufescens see also big bang breeders Tasmanian, see Bettongia gaimardi brush-tailed phascogale, see Phascogale tapoatafa big bang breeders, 152, 153–4, 155–6 Bulungamayinae, 302 bilbies (common names) Burramyidae, 21, 22, 28, 188, 192, 195 greater or rabbit-eared, see Macrotis lagotis Burramys sp. 28, 29, 34, 186, 187, 188 lesser bilby, see Macrotis leucura Burramys parvus (mountain pygmy possum), 185, bilbies, 195, Plate 11 diet of, 172 discovery of, 195 distribution & behaviour, 171 diet of, 197 maternal investment, 174 distribution, 202 physiology of, 172 hibernation & torpor, 198 bipedalism, 302–3 reproduction, 201–2 see also hopping burrows, birds, standard metabolic rates, 9 of burrowing bettong, 357 birth, size of marsupial at, 3, 8 of bilby, 171 birth canal, 7. 63, 66, 68 of wombat, 271, 274–7, 275 birth posture, 69, 70, architecture of, 274–5 in males, 70 breathing in, 277 Black Rock, Queensland, 348, 349, 351, 352 environment of, 276–7 blastocyst, formation of, 55 investment of time & energy, 277–80 blood glucose levels, maintenance of, 97 patterns of, 278 blood serum protiens, 47 burrowing habit, 27, 171, 357 blood systems, vascular changes after birth, 75 blood vessels, Caenolestes sp., 28, 29, 115 of newborn, 75, Plate 3 Caenolestes fuliginosus (silky shrew opossum), 110– bobuck, see Trichosurus caninus 12, Plate 5 body mass, Caenolestidae, 18, 19, 21, 22, 26, 28, 34, 110, 115 comparison of mammals, 11–12 caenolestids, 24, 32, 110, 168 in relation to metabolism, 10–11, 149, Caloprymnus campestris (desert rat kangaroo), 219, in relation to breeding success 301, 302 & thermoregulation, 147 caluromid sp., spermatozoa, 20 body temperature, 9, 10 Caluromyidae, 28, 115, 117 maintenance of, 78, 93–5 Caluromyinae, 115 of dasyurids, 147, 148, 149, 159 Caluromys sp., 28, 29, 115, 119 of kangaroos, 318–19 Caluromys derbianus (woolly opossum), 130 of opossums, 100, 133 Caluromys philander (bare-tailed woolly opossum), of pygmy possums, 199 118, 123–6, 126, Plate 5 of suger gliders, 210 breeding success, 124–6, 125, of tammars, 335 diet of, 123, 123–4 bogong moth, see Agrotis infusa digestive physiology, 124 bores, distribution, 123 consequences of sinking, 330 habitat, 119 Borhyaenidae, 106 lactation, 124 bounty payments, 379–80, 380 rate of development, 125 Brachyglottis repanda, 262 reproduction, 40, 124, 125 Bradypodidae, 108 Bradypus sp. SMR values, 12 social behaviour & home range, 123 Bradypus tridactylus, 4 survival of, 137 brain, Caluromysiops sp., 115, 117 anatomy of, 19, 20 Caluromysiops irrupta (black-shouldered opossum), cerebellum of, 88 117 cerebral hemispheres of, 88 Camelus dromedarius (camel), 305 differentiation of, 88 Canis familiaris (dingo dog), 145, 370, 372, 373 optic chiasma, 89 footprints, 372 brian stem, 88 speed of in relation to kangaroos, 308–9 breeding strategies, 40 Capra hircus (goat), 352, 378 in macropods, 312 Capreolus capreolus (roe deer), 191 of male marsupial, 50–1 carbohydrate metabolism in macropods, 305–6 424 Life of Marsupials

carbohydrates circannual rhythm, 339 in milk, 83–4 classification of marsupials, in plant sap, 186 groups based on anatomical characteristics, 20 see also sugars cleavage & blastocyst formation, 55–6

carbon dioxide (CO2), 10, 75, 94 Clemensia, 16 in wombat burrows, 277 climate changes, 31, 368

carbon14 dating, 146, 371, 372 coati, see Nasua nasua Casuarina, 32 colon, fermentation in, 225, 273 cats, in Central Australia, 381 common echymipera, see Echymipera kalabu cattle & sheep, see sheep; stock Compositae, 32 caves, 318–20, 349, 352 condylarths, 34 Cavia sp. (guinea pig), 106 copulation in marsupials, 51–2 cellulose, breakdown in macropods, 303 corpus luteum, 43–4, 47, 79 Central America, 108 role of, 56, 63–4 Central Australia, 329, 330, corticosteroid binding globulin, CBG changes in, 381–2 cortisol, mammal species & body weight, 382 in brown antechinus, 154 Central & North American marsupials, 130–6 in fetus, 64, 76 Cercartetus sp., 28, 29, 186 cost of reproduction, 151 Cercartetus caudatus (long-tailed pygmy possum), 185 cranial nerves, 88 Cercartetus concinnus (western pygmy possum), 188, Cretaceous Period marsupials, 30, 33 194 Cuddie Springs, NSW, 369, 371, 372 Cercartetus lepidus (little pygmy possum), 188 cultural phases in prehistoric Australia, 371 Cercartetus nanus (eastern pygmy possum), 187 Cupressaceae, 32 activity patterns & home range, 196 cuscus, behaviour & energy metabolism, 197, 200 & environment change by humans, 383 diet of, 197 cuscus (common names), distribution, 188, 198 bear, see Ailurops ursinus embryonic diapause, 201 common spotted, see Spilocuscus maculatus hand pads, 194 ground, see Phalanger gymnotis heart rate, 198 northern common, see Phalanger orientalis hibernation in, 198, 198, 200, 200 southern common, see Phalanger intercastellanus reproduction, 201 cyanogens in plant leaves, 224 cerebral cortex, development of, 92–3 cytology, 20 Chaeropus ecaudatus (pig-footed bandicoot, extinct), 165, 168, 178 Dacridium, 32 chewing, Dactylopsila sp., 28, 29, 186 in kangaroos, 297 Dactylopsila megalura, (triok), 188, 202 in wombats, 272 Dactylopsila palpator (triok), 188, 202, Plate 11 Chironectes minimus (water opossum), 28, 29, 115, Dactylopsila trivirgata (striped possum), 188, 202, 117, 119, 127, 130, Plate 4 205, 248 Chlamydia sp., Danthonia sp., 271 genital fungus of koala, 237 Darwin, Charles, chromosome painting, 21, 22, 23, 294, 300, 345, & adaptive radiation, 34–5 Plate 1 Dasycercus sp., 144 chromosomes, Dasycercus byrnei (kowari), 48, 70, 335 dosage compensation in female mammals, 23–4 Dasycercus cristicauda (mulgara), G-banding, 22, 345 body temperature, 149 numbers, 20–1, 22, 41 metabolism, 147 in American marsupial genera, 115, 117 SMR values, 11 in bandicoots, 167 Dasykaluta sp., 144 in dasyurids, 143 Dasykaluta rosamondae (red kaluta), 154 in kangaroos, 300, 346 Dasypus novemcinctus (armadillo), 108, 271 relationships between marsupials, 20–4 Dasyuridae, 18, 21, 22, 26, 28, 141, 143, 145 sex, 8–9, 19, 23, 41 dasyurids, 11, 8, 20, 24, 32, 113, 161 paternal X inactivation, 23 & big bang breeding, 153, 155 random X inactivation, 23 birth weight, 69 sex chromosome mosaicism, 23 body mass, humidity & air temperature, 147 Chulpasia, 32 breeding season, control of, 152–4 Index 425

chromosome numbers, 143 in mammals, 14–18 dentition, 142–3 in marsupial carnivores, 142–3, 142 diet, 141, 147 in marsupials, 15, 20 embryonic development, 55 variation in, of rat kangaroos, wallabies & kanga- feet, 143 roos, 298 in New Guinea, 157 see also under individual species longevity of, 135–6 destruction of native mammals, 379–80 metabolism & metabolic rates, 147–8, 150 diapause, see embryonic diapause, of the semi-arid rangelands, 158–60 diastema, 18, 226, 272, 292–3, 297 origins, 146–7 dibbler see Parantechinus apicalis reproduction cost, 151 didactyly, 18 male die-off, 154–5 Didelphidae, 21, 22, 26, 28, 110, 114–17, 115 thermoregulation, 148–9 Didelphinae, 115, 115 Dasyurinae, 28, 29, 144, 145, 153 Didelphis sp., 20, 24, 106, 119, 135, 168, 28, 29, 115 Dasyuroides sp., 144 Didelphis albiventris (white-eared opossum), Plate 4 Dasyuroides byrnei, see Dasycercus byrnei abundance & home range, 128 Dasyurus sp., 144 breeding patterns, 128–30, 129 Dasyurus albopunctatus, Plate 7 distribution of, 115, 116 Dasyurus geoffroii (western quoll), 148, 161 habitat & diet, 119, 128 Dasyurus hallucatus (northern quoll), 69, 153 Didelphis aurita (big-eared opossum) Dasyurus maculatus (spotted-tailed or ), abundance & home range, 128 141, 161, 240 breeding patterns, 128–30, 129 Dasyurus viverrinus (eastern quoll), 141, 161 distribution of, 115, 116, dentition, 142 habitat & diet, 128 embryonic development, 55 Didelphis marsupialis (common opossum), 131, 126 fat metabolism, 151 abundance & home range, 128 fetal membranes, 60 breeding patterns, 128–30, 129 field energetics, 150 dentition of, 114–15, 114 field metabolic rate, 150 distribution of, 115, 116, lactation, 152 habitat & diet, 119, 128 maternal investment, 174 reproduction, 40 rate of food through gut, 147 survival of, 137 reproduction, 151–2 Didelphis orientalis see Phalanger orientalis newborn, 7 Didelphis virginiana (Virginia opposum), 130–3 Daubentonia madagascarensis (aye-aye), 4, 205 body temperature, 100, 133 day length, breeding seasons & litter size, 133–4 & reproductive cycles, 40, 122, 130, 156–57, 157, chromosomes in, 131 160, 192, 197, 252, 338–43 diet & dentition, 132 Dendrolagus sp., 28, 29, 227, 293, 344 distribution, 108, 116, 131–2 Dendrolagus bennettianus, 295 ecology of, 132–3 Dendrolagus dorianus, 295 environment of, 113 Dendrolagus goodfellowi, 295, Plate 14 evolution of, 131 Dendrolagus inustus, 295 fertilisation & male competition, 134 Dendrolagus lumholtzi, 295 heart rate, 133 Dendrolagus matschiei, 295 hind leg fat index, 134, 135 Dendrolagus mbaiso, 295 kidney function, 132 Dendrolagus scottae, 295 lack of adaption to a cold climate, 132 Dendrolagus spadix, 295 lactation, 134 Dendrolagus ursinus, 295 longevity, 135–6, 136 dentition, origin of, 131–2 classification of wallabies & kangaroos, 294–300 reproduction, 40 browser grade, 297–9 fertilisation & male competition, 52, 134 browser-grazer grade, 299 litter size, 134–5 grazer grade, 300 number of young, 71 cusps in molar teeth, 14, 17 spread to North America, 130, 367 evolution in mammals, 16–18, 16 standard metabolic rate, 10 in bandicoots, 167 survival in winter, 133 in Didelphidae, 114, 114 diet, see under individual species in koala, 230 diet changes for young marsupial, 99 426 Life of Marsupials

digestive tracts, in kangaroos & wallabies, 313 in bandicoots, 172 of newborn, 75–6 in dasyurids, 147 echidna, see Tachyglossus aculeatus in honey possum, 187 Echymipera sp., 28, 29, 171 in koala, 231 Echymipera clara, 169 in leaf eating marsupials, 225 Echymipera kalabu (common spiny bandicoot), 169, in macropods, 304–6, 304 170, Plate 9 in opossums, 124 Echymipera rufescens (rufus spiny bandicoot), 170, in rat kangaroos, 291 171, 179, 248, 344 in wombats, 273 egg coats, 59–61 dingo, see Canis familiaris Ekaltadeta, 290–1 Dinornis sp., 259 Elaeocarpus dentatus (hinau), 252 Diprotodon sp., 142, 370 Elephantulus myurus (elephant shrew), 55 Diprotodon optatum, 371 embryo, reproduction of, 374–6, 375 early development, 55–6 footprints of, 372, 372 embryonic diapause, 43, 44, 56–8 hunting of, 373–6 glucose metabolism, 58 Diprotodontia, 18, 19, 24, 26 in eastern pygmy possum, 56 diprotodonts, 269, 270 in feathertail glider, 56 Dirk Hartog Island, WA, 4, 355, 356, 357, 359 in honey possum, 56 distinctive features of marsupials, 5–14 in macropod reproduction, 311–12, 322–4 Distoechurus sp., 186 in tammar wallaby, 56–7, 338, 342 Distoechurus pennatus (feathertail possum), 28, 29, 192 in rat kangaroos, 291 distribution, see under individual species in red kangaroos, 322–4 Djiarthia murgonensis, 32 endangered species, DNA, microsatellite, 334 rehabilitation of, 383 DNA sequence analyses, energy requirements, & marsupial relationships, 25, 28, 187, 294, 346 during lactation, 100, 101 DNA/DNA hybridisation, 25–6, 26 (Box 1.2) of arboreal folivores, 231–3, 233, 245 & marsupial relationships, 26, 28, 29, 294 of dasyurids, 151, 160 & wallaby & kangaroo classification, 301, 346 of pygmy possums, 189, 197, 209–10 techniques, Plate 1, Plate 2 of Virginia opossum, 133 domestication of gray short-tailed opossums, 122 of wombats, 274 Dorcopsis sp., 28, 29, 293, 297, 300, 302, 353 Enneapogon avanaceus (bottlebrush), 271, 321 Dorcopsis atrata, 294 enzymes, 24, 97 Dorcopsis hageni, 294 proteolytic, 304, Dorcopsis luctuosa, 294 Eragrostis setifolia (kangaroo grass), 321 Dorcopsis muelleri, 294 Eremitalpa granti (golden mole), 4 Dorcopsoides, 298 Eremophila longifolia (berrigan), 325 Dorcopsulus sp., 28, 293, 297, 302 Eremophila sp. 32 Dorcopsulus macleayi, 294 essential oils, see terpenes, Dorcopsulus vanheurni (small dorcopsis), 294 Eucalyptus sp., 32, 34, 185 Dorre Island, Western Australia, 355–9 dependence on hypogeal fungi, 359–60 Dromiciops sp., 19, 20, 24, 28, 29, 32, 34, 115 Eucalyptus foliage, Dromiciops gliroides, 19, 113 as food, 221–7, 251 drought, constituents, 222, 239 & kangaroo reproduction, 45, 322, 323–4, 326 defensive chemicals, 224 effects on wombats, 274 foliage density & concentration of potassium, 223 dryolestids, 27 nectar production, 208 didactyly, 18 nutrients, 221 dunnarts (common names) secondary compounds, 223 fat-tailed, see Sminthopisis crassicaudata Eucalyptus andrewsii, 232, 245 Julia Creek, see Sminthopsis douglasi Eucalyptus camaldulensis, 234 red-cheeked, see Sminthopsis virginiae Eucalyptus cypellocarpa, 239 stripe-faced, see Sminthopsis macroura Eucalyptus dalrympleana, 221 dwarfism, 180, 354 Eucalyptus fastigata, 238 ears, Eucalyptus maculata, 216 development of, 90 Eucalyptus melliodora, 232, 251 Index 427

Eucalyptus obliqua, 239 role during birth, 66, 68 Eucalyptus punctata, 232 secretion of hormones, 64, 65, 66, 69 Eucalyptus radiata, 221, 232, 238, 239 field metabolic rates (FMR), 11–14, 11 Eucalyptus resinifera, 217 in dasyurids, 149, 150, 154, 158 Eucalyptus regnans, 203 in greater glider, 239 Eucalyptus sideroxylon, 223, Plate 13 in honey possum, 189, 191 Eucalyptus viminalis, 216, 221, 238 in koala, 233, 233, 235 euro, see Macropus robustus in Leadbeater’s possum, 204 European encounters with marsupials, 3 in ringtail possum, 245 European impact on Australian marsupials, 376–82 in sugar glider, 210 European settlement, 330 measurement of, 13(Box 1.1) Euryapterix sp., 259 foliage, as food for marsupials, 185, 222 , 5 see also folivores eyes, development of, 88–90 folivore-eucalyptus arms races, 226–7 Eyre Peninsula, South Australia, 332 folivores, extinction foraging times & body mass relationships, 217 & climate change, 367 living in trees, 230 of animals in Pleistocene epoch, 367 relationships between marsupials, 196 extinct species, 370–1 see also arboreal folivores exudivores, follicle stimulating hormone (FSH), 44, 45 foraging time & body mass, 217 forest folivores, 228, relationships between marsupials, 196 constituents of milk, 236 forests in New Zealand faecal pap, 100, 235, 281 & possum introduction, 259–61 faecal pellets in macropods, 305 forestomach fermentation, 289 families of marsupials, 28 fossil record, 32 Fagara rhoifolia, 124 fossils, 3, 27, 30 fasciculus aberrans, 19 of kangaroo, 289, 371 see also brain of wombats, 269 fat storage, records of bandicoot teeth, 169 caudal, 112, 113 records of carnivorous marsupials, 146 in antechinus, 151, 152 sites in South America, 109, 110, 113, 131 in bandicoot, 174, 175 sites in South Australia, 34, 138, 141 in burramys, 197 in Tiupampa, Bolivia, 30, 105 in Chilean shrew opossum, 113 mammalian, 32 in fat-tailed dunnart, 159 fox, European, see Vulpes vulpes in Virginia opossum, 132, 134, 135 Fuchsia excorticata (fuchsia), 261, 262 feet, fungi, 185 in marsupials, 17, 18, 20 hypogeal, 359–61 of bandicoots, 167 mycorrhizal, 172 of pygmy possums, 194–5, 194 odour from, 360 fermentation, see also Phytophthera anaerobic, 231 fur, 27 bacterial, 226, 239 developement of, 93, 95–6 foregut, 289, 291, 303–8 in brushtail possum, 257 hindgut, 226 in cuscus, 249 in colon & caecum, 172, 225 in desert kangaroos, 319 in gut, 99 in greater glider, 239 in wombats, 273 in koala, 232 prior to digestion, 99 in opossum, 5, 126–7, 132 fertilisation, 43, 52–5, 53 Virginia opposum, 52 gametes, production of, 41 brush-tailed bettong, 52 gametogenesis to fertilization, 41–52 in placental mammals, 54 gamma-globulins, in whey milk, 85 fertilisation to parturition, 52–63 Garden Island, Western Australia, fetus, Plate 3 tammar population, 334, 335, 354 development, 58–9 Gastrolobium sp., 161, 251, 317 membranes of, 59–63, 60 genera of living marsupials, 28 428 Life of Marsupials

gene pool of gray short-tailed opossums, 123 arid, 319, 321 Geniostoma ligustrifolia, 262 Atlantic forest in Brazil, 119, 137 genome studies on tammar wallaby, 332 cool temperate, 131 Genyornis (mihirung, extinct bird), 370, 371 in Central & South America, 137 germ cells, primordial, 41 neotropical rainforest, 118 gestation, 57 rainforest, 130, 137 gestation length, effect of fetal genotype on, 64 haemoglobin, effect of anti-paternal sensitisation on, 59 embryonic, 76 gliders (common names) hare wallabies, 299 feathertail, see Acrobates pygmaeus hare wallabies (common names) greater, see Petauroides volans banded, see Lagostrophus fasciatus mahogany, see Petaurus gracilis rufous, see Lagorchestes hirsatus squirrel, see Petaurus norfolcensis spectacled, see Lagorchestes conspicillatus sugar, see Petaurus breviceps hearing in kangaroos & wallabies, 312–13 yellow-bellied, see Petaurus australis heart rate, during hopping, 311 gliding membranes see patagium in pygmy possum, 198–200 Glironia sp., 28, 29, 115 in Virginian opossum, 133 Glironia venusta ( bushy-tailed opossum), 117 Hemibelideus sp., 28, 29 Glironinae, 115 Hemibelideus lemuroides, (lemuroid ringtail possum), gluconeogenesis, 97 228, 247, glucose concentration in fetal fluids, 62 diet of, 248, glucuronic acid, 223, 239 habitat of, 248–9 glycolysis during diapause, 58 Heirisson Prong, Western Australia, 359 goats, see Capra hircus Heteropogon sp., 271 golden mole, see Eremitalpa granti hibernation, 112, 113, 132 Golgi apparatus, 43, 48 hindgut fermenters, 226, 271 gonadotrophin releasing hormone (GnRH), 44, 45, history of marsupials, 30–4, 33 50, 207 in Central & North America, 130–1 Gondwana, 3, 27, 31 in South America, 105–8 Graafian follicle, 43, 47 see also Tertiary Period, Cretaceous Period in some marsupials, 43, 44 honeydew, 185 Gracilinanus sp., 115, 117 honey possum see Tarsipes rostratus Graminae, 32 hopping economies, 308–11 gravity, detection by newborn marsupial, 74 hormonal control of reproduction in male marsupial, grazing by stock, 363, 378, 381 50, 342 grey dorcopsis, see Dorcopis luctuosa hormones, 99 ground-dwelling moas, Dinornis, Euryapterix, adaptation to times of the year, 45 Pachyornis, 259 & parturition, 64, 65, 66, 66 gum, sap & insect feeders, 202–17 responses to photoperiod in tammar wallaby, gums, acacia, 185 340, 341–2 Gunguroo, 302 Houtman’s Abrolhos archipelago, 354 gut of leaf-eating marsupial, 225 humans, Gymnobelideus sp., 186 interactions with megafauna, 367–8, 371–3 Gymnobelideus leadbeateri, (Leadbeater’s possum), 28, effects on mammal species, 136–7 29, 202 human occupation, active metabolic rates, 204 in Australia & New Guidea, 367 discovery of, 203 & holocene changes, 367–76 diet of, 203–4 Hydrochoeris sp. (capybaras), 106, 284 distribution of, 188 Hypsiprymnodon moschatus, (musky rat kangaroo), energy budget, 211 290, 291, 302, 303 habitat & home range, 203 chromosome numbers, 300, 301 reproduction, 204 diet, 361 survival of, 204–5 habitat, 361 reproduction of, 361–2 habitat fragmentation, 378 stomach structure, 305 habitat use, 120 Hypsiprymnodontidae, 22, 290–1 habitats, distribution, 291 aquatic, 117, 127 Hystrix (porcupines), 271 Index 429

immune system fossil history, 290 development of, 86 giant (extinct), 142, 270, 370 immunity, passive, 85 grazers, 300 immunoglobulins, hopping, provided by the mother, 87 bipedal, 289 see also gamma-globulin economies of, 308–11 immunological competence, acquisition of, 87–8 speed & oxygen consumption, 308–9 immunological tolerance, 88 large, 315–25 infections, bacterial, 97 reproduction, 311–2, 322–4, 326, 328–9 inorganic elements in marsupial milk, 85 relationships, distribution & origins, 289–303 insect, gum & sap feeders, 202–17 sensory attributes, 312–13 insects extration of proteins from leaves, 185 success of, 303 invertebrates, food for marsupials, 185 see also Macropodoidea; macropods iron in marsupial milk, 85 kangaroos (common names) island populations of wallabies, 353–9 eastern grey, see Macropus giganteus Isoodon sp., 28, 29 euro, see Macropus robustus erubescens Isoodon auratus (golden bandicoot), 169, 170, 171, giant, see Sthenurus 179, 355 red kangaroo, see Macropus rufus Isoodon macrourus (northern brown bandicoot), 169, western grey, see Macropus fuliginosus 171, 179 kangaroos & wallabies, 293–303 birth posture, 70 karyotype, see chromosome number distribution of, 170, 180 Kawau Island, New Zealand, 333, 334 reproduction, 176 kidney, hormonal profiles, 176 development of, 98–9, 98 newborn, 7 uro-genital systems, 6 lactation, 175 function, 97–8 oestrous cycle, 44 maturation of, 93 Isoodon obesulus (southern brown bandicoot), 169, Kinchega National Park, 328, 329 171, 179, Plate 8 kinkajou see Potos flavus breeding seasons, 176, 177 kino, 185 distribution of, 170, koala, see Phascolarctos cinereus molar teeth, 168 Koala Rescue Strategy (1997), 237 survival of, 177 Kochia sp., (round-leaf chenopod), 321 Kollikodon sp. 27 jaw muscles, kowari, see Dasycercus byrnei, in wombats, 273 kultarr, see Antechinomys laniger in pademelon & red-necked wallaby, 296, 297 in carnivores, 15 lactation, 63, 102 late stage, 81–2 kaluta, see Dasykaluta rosamondae length of, 40, 101 kamahi, see Weinmannia racemosa marsupial, 76–9 Kangaroo Island, South Australia, 237–8, 329, 333, phases of, 79 334, 353 lactogenesis, 79–81 tammar population, 332–5 lactose, in marsupial milk, 83–4 kangaroo management, 331 Lagorchestes sp., 293 kangaroo origins, 302–3 Lagorchestes asomatus, 294 kangaroos, 287–364 Lagorchestes conspicillatus (spectacled hare-wallaby), browser grade, 297 293, 294, 354, 355 classification by chromosomes, 300–1 Lagorchestes hirsutus (mala), 27, 294, 355–9, 358 classification by teeth, 294–7 on Bernier & Dorre island populations, 355–9, competition with stock, 314–15, 329–31 356 commercial use, 331 Lagorchestes leporides, 294, 354 dentition, 294, 298 Lagostorphus sp., 28 digestive tracts, 303, 304–6, 304 Lagostrophus fasciatus (banded hare-wallaby), 4, 293, embryionic diapause, 56, 311–2, 322–4 358 evolution of, 289–90 on Bernier & Dorre island populations, 355–9, foregut fermentation, 289, 303–8 356 fossils, 289 relationships, 301 430 Life of Marsupials

Lake Menindee, NSW, 371, 372, 373 Loranthaceae, 32 Lake Mungo, NSW, 369 lowland pitpit, see Saccharum edule land clearance, 376, 377–9 Lucilia cuprina, sheep blow fly, 330 landowners, threat to habitats, 243 luteal phase of the uterus, 47–8 Lasiorhinus sp., 20, 28, 29, 186, 270 luteinising hormone (LH), 44, 45, 50 Lasiorhinus krefftii (northern hairy-nosed wombat), in oestrous cycle, 64 breeding strategies, 281 Lutreolina sp., 28, 29, 106, 115 distribution of, 270, 271 Lutreolina crassicaudata (comadreja or marsupial diet of, 271 weasel), 117 pattern of burrows, 278 warren complexes, 276 Macropodidae, 19, 21, 22, 28 Lasiorhinus latifrons (southern hairy-nosed wombat), success attributes, 303 269 Macropodoidea, 289, adaptations for feeding on bulky herbage, 273 relationships, distribution & origin, 289–303 breeding strategies, 280–1 relationships of genera, 301 burrow of, 275, Macropodinae, 22, 293, 332 burrow environment, 276–7, 277–80 classification according to chromosomes, 300–2 dentition of, 272, 272 classification according to dentition, 294–300 diet of, 271 macropods, digestive system, 273–4 large, 315–331 distribution of, 270, 271 absorption of water, 305 energetics of, 274 caecum & colon structure & function, 305 faecal pellets, 305 embryonic diapause in reproduction, 311–2 maintenance nitrogen requirement, 232, 273 hopping economies, 308–11 metabolic rates, 274, 277 stomach structure & function, 303–5, 304 recycling of water, 273–4 small, 332–63 lateral geniculate nucleus ( LGN), 88, 89, 90 Macropus group, 299 Laurasia, 27 Macropus sp., 28, 29, 141, 142, 293 leaf-eating marsupials, 221, 226 birth canal, 7 evolution of, 226 chromosome number, 301 see also folivores Macropus agilis (agile wallaby), 293, 299, 316, 332 leaves see foliage Macropus antilopinus (antilopine wallaroo), 299 Lepilemur leucopus, (leaf-eating lemur of distribution, 299, 300, 316, 317 Madagascar), 244 Macropus bernardus (black wallaroo), 299, 300, 317 Leptospermum sp., (tea-tree), 262 Macropus dorsalis (black-striped wallaby), 299, 332, Lepus, hares, 244, 379 333 Lestodelphys sp., 115 Macropus eugenii (tammar wallaby), Plate 15 Lestodelphys halli (Patagonian opossum), 28, 29, 111, adaptations to island life, 335–7 112–13, 117 chromosome painting, Plate 1 Lestoros inca, 110, 111, 115 development & growth, 78, Plate 3 LGN see lateral geniculate nucleus diet, 335 Libocedrus bidwillii, (conifer), 260 distribution, 299, 316, 332 licences, export, 243 ears, 313 lipids in milk, 83, 84–5, 134, 175, 350 hopping & oxygen consumption, 309, 309 litter size, 78, 100 island populations, 333, 334 in bandicoots, 171–3, 174, 176, 178, 180 genome project, 332 in dasyurids, 152, 160 lactation, 78, 79, 337 in opossums, 113, 122, 124, 126, 127–8, 129, 134, nutrition & nitrogen balance, 336 135 origins of, 334–5 in pygmy possums, 191, 201, 204, 212 population dynamics, 336–7 in ringtail possums, 248, 249 population on Kangaroo Island, 335, 336, 336 short-chain fatty acid production, 306 little red kaluta, see Dasykaluta rosamondae reproduction, 6, 7, 9 liver funtion, 96–7 & photoperiod signal, 340–2 LLP (late lactation protein), see under proteins anatomy of newborn, 72, Plate 3 locomotion of young marsupial, 71–3 annual cycle, 337 locomotor generator, 73 behaviour & birth, 69, 70 Lomandra sp., (mat-rush), 271 birth & pouch exit, 39, 40 Index 431

birth posture, 70 distribution, 299, 316–17, 316 breeding, 334, 337–40 home range, 319 control of birth time, 69 living in an arid environment, 319–21 early development, 54 nutrition, 320–1 embryo, 55, Plate 3 standard metabolic rate, 10 embryonic diapause, 56–8, 57 Macropus robustus erubescens (euro), 300 in males, 342, 343 Macropus robustus isobellanus (dwarf euro), 354, 355 newborn, 7, Plate 3 Macropus rufogriseus (red-necked wallaby), 332, oestrous cycle, 44, 45–6 Plate 15 onset of breeding in Louisiana, North distribution of, 299, 316 America, 338, 338 jaw muscles, 296, 297 oogenesis, 41 newborn, 7 placental transport, 60–1 Macropus rufogriseus rufogriseus (Bennett’s wallaby), pregnancy of, 64 breeding season behaviour, 343 protein components of secretions, 47 Macropus rufus, (red kangaroo), weight of young, 100 & change of environment by humans, 383 standard metabolic rate, 10 Ara, of Central Australia, 313–4 temperature regulation & water turnover, 335 behaviour & birth, 69, 70 urea retention in kidneys, 307 chromosome number, 300 vision, 312, 313 competition & introduced stock, 314–5, 315 Macropus ferragus, 300 coping with heat, 317–19, Macropus fuliginosus, (western grey kangaroo), 271 distribution of, 299, 315–6, 316 adaptation to woodland & forest, 325–7 food preferences, 321 birth & pouch exit, 39 home range, 321 distribution of, 299, 316, 317 living in an arid environment, 321–2 & introduced stock, 314–5, 315 locomotion, 309–9, 309 with eastern grey kangaroo, 327–8 use a treadmill, 308 numbers of (1981), 314 metabolic rates of young, 323 plant preferences, 326 numbers of (1981), 314 pregnancy of 64 standard metabolic rate of young, 323 reproduction reproduction, 314, 322–4 breeding & drought, 326 breeding strategy, 323 seasonal breeders, 39, 328–9 & drought, 326 social behaviour, 329 newborn, 7 speciation, 325 response to rainfall, 321–2 Macropus giganteus (eastern grey kangaroo), Plate 15 size at birth, 8 adaptation to woodland & forest, 325–7 skull of, 292 anatomy of stomach, 304 social behaviour, 322 birth posture, 70 Macropus titan, 300, 370 competition Macrotis sp., 28, 29 & introduced stock, 314–5, 315 Macrotis lagotis (greater bilby or rabbit-eared), 27, with western grey kangaroo, 327–8 167, 168, 169, 171, 178, Plate 9 distribution of, 293, 299, 316, 317 Macrotis leucura (lesser bilby) 171, 179 numbers of (1981), 314 magnesium in milk, 86 plant preferences, 325 maintenance nitrogen requirement (MNR), 232 pregnancy of 64 in brushtail possums, 252 reproduction in greater glider, 239 seasonal breeders, 328–9 in macropods, 308 embryonic development, 55 in ringtail possum, 232, 245 size at birth, 8 in tammar wallaby, 336 speciation, 325 in woolly opossum, 124 social behaviour, 329 in wombat, 273 Macropus greyi (toolache wallaby), 299, 332, 380 Maireana pyramidata (black bottlebrush), 321 Macropus irma (western brush wallaby), 299, 332 male die-off, Macropus parma (parma wallaby), 299, 332, 333, 336 in Antechinus, 153 Macropus parryi (whiptail wallaby), 299, 332 in Phascogale, 153 Macropus robustus (euro or common wallaroo), physiology of, 154 coping with heat, 318–9 ecological significance of, 154–5 432 Life of Marsupials

mala, see Lagorchestes hirsutus maternal & newborn body weights, 7, 8 mammal extinctions, 136, 163, 179–80, 179, 367, 370, physiological differences with other mammals, 376, 384 9–14 mammal species, predatory, 32, 106–8, 139–63 Huxley’s terminology, 5 relationships based on teeth & feet, 17 mammals, relationships based on protein analysis, 24 deliberate destruction of native species, 379–80 relationships based on DNA, 25–7 history in Australasia, 32–4 relationships within marsupials, 14–27 metabolic rates, 9 terminology of, 5 placental, Tertiary history of, 30–4, 105–6 reproductive organs, 6, 6 see also South American marsupials sexual differentiation, 8 Marsupialia, 5 radiation of, in South America, 32, 105–6 megafauna, 290, 370–3 mammary glands, 78 classes of, based on body size, 373, growth & suckling, 81 extinction isoclines, 374 muscle penetration, 74 Melaleuca sp. 32 preparation of, 79 melatonin, profile in tammar wallaby, 341 prolactin receptors, 80, 134 Meles meles (badger), 191, 271 mammogenesis, 79 Melicytus ramiflorus (mahoe), 262 Mammoth Cave, Western Australia, 372 Merkel cells, 73 Manis (pangolin ), 271 Mesophelia sp., 361 manna, as food for marsupials, 185 Mesopheliaceae, 360 Marmosa sp. 22, 28, 29, 106, 115, 117, 119 mesotocin, Marmosa canescens (mouse opossum), 130 & parturition, 64, 65, 66, 67(Box 2.1) Marmosa incana, 137 & lactation, 82 Marmosa mexicana (mouse opossum), 130 metabolic rates, see field metabolic rate; standard Marmosa murina, 118, 119 metabolic rate Marmosa robinsoni, 44, 130 metabolism of dasyurids, 147 Marmosinae, 115, 117 Metachirus sp., 28, 29, 115, 119 Marmosops sp. 28, 115, 117, 119 Metachirus nudicaudatus (brown four-eyed opossum), Marmosops impavidus (common mouse opossum), 117, 119, 130 130, Plate 5 metacone, 16, 168 Marmosops incanus, 120 metaconule, 16, 18, 168 Marmosops invictus (slender mouse opossum), 130 Metatheria, 5 Marmosops noctivagus, 119 Metrosideros fulgens (rata vine), 262 Marmosops parvidens, 118, 119 Metrosideros robusta (northern rata), 261, 262 Marmota monax (woodchuck), 133 Metrosideros umbellifera (southern rata), 260 marsupial carnivores, 32, 141–3, 163 Micoureus sp. 28, 29, 115, 117, 119 anatomy of, 142–3 Micoureus alstoni (mouse opossum), 130 body size, 143 Micoureus demerarae (woolly mouse opossum), 118, conservation, 160–1 119 diet, 147 Microbiotheriidae, 21, 22, 26, 28, 110, 115 fossil records, 146 Microperoryctes longicauda, 169, 170, 171 in South America, 106–8, 107 Microperoryctes murina, 171 relationships of, 144 Microperoryctes papuensis, 170, 171 marsupial families, 28 Microtis rossiameridonalis (vole), 15 marsupial leopard, see Thylacoleo carnifex milk, marsupial, marsupial mole, see Notoryctes typhlops composition of, 41 marsupial tapir, see Palorchestes in common brush tail possum, 253 marsupial weasel, see Lutreolina crassicaudata in forest folivores, 236 marsupials, in koala, 235 Australasian & South American convergence of in short-tailed opossum, 122 species, 137 in tammar, 82–6, 83 distinctive features of, 5–14 in Virginia opossum, 134 distribution in time & space, 3, 27–30 milk, placental, divergence of, 31 composition of, 82 European impact on, 376–82 Miocene fossil history, 34 in South America, 105 moas, 259 Index 433

monito del monte, see Dromiciops gliroides New Zealand, Monocalyptus group of eucalypts, 224, 238 arms race & forests, 259–61, 263–4 Monodelphis sp., 22, 28, 29, 106, 115, 117, 119 introduction of wallabies, 332 Monodelphis brevicaudata, 10, 118,, 119, 120, Plate 5 studies on brushtail possum, 250–64, 253, 255, Monodelphis dimidiata, 122 257, 260, 262 Monodelphis domestica (gray short-tailed opossum), tammar wallaby population, 333–4 120–3 newborn marsupial, breeding, development & longevity, 121–2, 121 adaptations for reaching teat, 71–3 diet, 120 anatomy of, 71, 72, Plate 3 distribution of, 116, 120 locomotion, 71 domestication of, 122 see also young marsupials in biomedical research, 123 Ningaui sp., 141, 143, 144, 187 reproduction & development, 52, 53, 122, 173 Ninox sp. 203 social behaviour, 121 Ninox strenua, powerful owl, 240, 241 monogamy, 216, 240, 250, 346, 350–1 nitrogen from urea, 303 Monotremata, 5 nitrogen balance, 124, 154, 320 Montebello Islands, Western Australia, 355 nitrogen conservation, in macropods, 307–8 mouse, house, see Mus musculus nitrogen metabolism, 232 mouse opossums, nitrogenous waste, excretion of, 99, 307 survival of, 137 non-gliding possums (common names) see also Marmosa sp.; Thylamys sp. non-gliding striped, see Dactylopsila trivirgata MNR see maintenance nitrogen requirement trioks, see Dactylopsila megalura; Dactylopsila mulgara, see Dasycercus cristicauda palpator Murexia sp., 144 noolbenger, see Tarsipes rostratus Murexia habbema, 145 North American marsupials, Murgon fossil site, Queensland, 30, 32, 34, 113, 169 see Central & North American marsupials Mus musculus, wild house mouse, 158 Nothofagus (southern beech), 32, 34, 254 muscles, Notomacropus subgenus, 299, 299, 300, 302, 332 cremaster, 76 Notomys (hopping mice), 143 ilio-marsupialis, 74, 76, 77 Notoryctes sp., 144 in teats, 76 Notoryctes typhlops (marsupial mole), 4, 27, 28, 29, involved in jumping, 310–11, 310 49, 141 jaw, 15, 273, 296, 297 Notoryctidae, 21, 22, 26, 28 panniculus carnosus, 76 numbat, see Myrmecobius fasciatus myelinisation of nerve fibres, 93 Myoictis sp., 144, 145 Ochotona, pikas 244 Myrmecobius sp., 144 odour, Myrmecobius fasciatus (numbat), 4, 141, 146, Plate 7 from hypogeal fungi, 360–1 Myrmecobiidae, 21, 22 in grey kangaroos, 329 Myrtaceae, 32, 221 in sugar gliders, 207 in wombats, 279 nabarlek, see Petrogale concinna in yellow-bellied gliders, 216 nailtail wallabies, as grazers, 300 oestrogen, nailtail wallabies (common names) in oestrous cycle, 45–6, 64 northern, see Onychogalea unguifera role in birth, 63–4 bridled, see Onychogalea fraenata oestrous cycle, 41, 42, 44–8 Nasua nasua (coati of South America), 106, 108 hormonal control of, 42 nectar, as a food for marsupials, 185, 189 in brushtail possum, 42 nectar & pollen feeders, 187–202 in tammar wallaby, 45–6 Neophascogale sp., 144, 145 in Virginia opossum, 134 nerve cells see neurones uterine changes, 47–8 nervous system, 89 (Box 2.5) vaginal complex, 46 development of, 88–93 olfactory knobs, 73 neural control of newborn, 71–3 oogenesis, 41–4 neurones, Onychogalea sp., 293 structure & function of, 89 (Box 2.5) Onychogalea fraenata (bridled nailtail wallaby), 294 New Guinea, Onychogalea lunata, 294 wallabies, hare wallabies, 294 Onychogalea unguifera (northern nailtail wallaby), forest wallabies, chromosome number, 300 294 434 Life of Marsupials

oocytes, 41 Pachyornis sp., 259 opossums (common names) pademelons, bare-tailed woolly, see Caluromys philander as pests, 379 big-eared, see Didelphis aurita chromosome numbers, 300, 344 black-shouldered, see Caluromysiops irrupta distribution, 295, 344 brown four-eyed, see Metachirus nudicaudatus hopping economies, 309 bushy-tailed, see Glironia venusta stomach structure & function, 305 Chilean mouse, see Thylamys elegans reproduction, 344 Chilean shrew, see Rhyncholestes raphanurus pademelons (common names) comadreja, see Lutreolina crassicaudata dusky, see Thylogale brunii common, see Didelphis marsupialis red-bellied or Tasmanian, see Thylogale billardierii common mouse, see Marmosops impavidus red-legged, see Thylogale stigmatica dusky mouse, see Marmosa fuscata red-necked, see Thylogale thetis fat-tailed, see Thylamys elegans palaeoclimates and human arrival, 368–70 gray four-eyed, see Philander opossum palaeolithic culture, 371 gray short-tailed, see Monodelphis domestica Palorchestes sp. (extinct marsupial tapir), 371, Plate 1 mouse, see Micoureus alstoni pancreas, development of, 74 monito del monte, see Dromiciops gliroides panniculus carnosus muscle, 78 murine mouse, see Marmosa robinsoni panting, 319 Patagonian, see Lestodelphys halli paracone cusps, 16, 168 silky shrew, see Caenolestes fuliginosus Parantechinus sp., 144 short-tailed, see Monodelphis Parantechinus apicalis (dibbler), 153, 161 slender mouse, see Marmosops invictus parturition, 63–76 Virginia opossum, see Didelphis virginiana hormone changes at, 63, 65 water opossum, see Chironectes minimus role of fetus during birth, 68 (Box 2.4) white-bellied slender mouse, see Marmosops maternal behaviour, 69–71, 70 noctivagus pastoral industry in Australia, 314–5, 315, 329–31, white-eared, see Didelphis albiventris 376, 377 woolly, see Caluromys derbianus patagium, 27, 193, 194 woolly mouse, see Micoureus demerarae penis, opossums of the Americas, 103–138 glans penis, 7–9 see also South American marsupials structure of, 19, 20, 21, 143 Orongorongo, New Zealand, Perameles sp., 28, 29, 170 brushtail possum studies, 252, 253, 253, 254, 257, Perameles bougainville, (western barred bandicoot), 258, 170, 170, 178, 179, 358, Plate 8 long term responses, 261–4 Perameles eremiana (desert bandicoot), 170, 170, 179 response of plant species, 263 Perameles gunnii (eastern barred bandicoot), 170, Ornithorhyncus anatinus (platypus), 5, 55 170, 174, 179 Oryctolagus cuniculus (wild rabbit), Perameles nasuta (long-nosed bandicoot), 169, 170, grazing pressures & wombats, 283 179 in Australia, 380–1 distribution, 170 in Central Australia, 381, 382 metabolic rate, 10 maternal investment, 174 reproduction spread of, 376, 378–9, 380, 381 fetal membranes, 60 Osphranter subgenus, 299, 300, 302 Peramelidae, 19, 21, 22, 28, 34, 167, 170, 171 Ovis aries, see sheep peramelids, 24, 26 ovulation, 43 embryonic development, 55 hormone control of, 44–5 Peramelomorphia, 18, 170 oxygen, Peroryctes sp., 28, 29 affinity for in newborns, 76 Peroryctes broadbenti, 167, 170, 171 oxygen consumption, 10 Peroryctes raffrayana, 169, 170, 171 of dunnarts, 159, 160 Peroryctidae, 28, 71 of honey possums, 191 petaurid sperm, 20 of kangaroos, 308, 309 Petauridae, 21, 22, 22, 28, 185, 202, 206 of pygmy possums, 198, 200, 200 non-gliding, 205 of sugar gliders, 208–9, 209 Petauroides sp., 28, 186 of young marsupial, 94–6, 94 Petauroides volans (greater glider), 27, 238–43, Plate Oxylobium, 161 12 oxytocin, 82 dependency on forest habitat, 242–3 Index 435

dentition, 239 Petrogale brachyotis (short-eared rock wallaby), 295, diet of, 124, 238–9, 248 345, 346 digestability & nitrogen requirement, 124, 232, species complex, 347 239 Petrogale burbidgei (monjon rock wallaby), 295, 345, distribution, 228, 229, 240 346 ecology & social behaviour, 240–1 Petrogale coenensis, 295, 345, 346 embryonic development, 55 Petrogale concinna (nabarlek), 295, 299, 345, 346, 347 energy metabolism, 233, 239–40 Petrogale godmani, 295, 345, 346 eucalypt leaf selection, 224 Petrogale herberti, 295, 345, 346 gliding membrane, 193 Petrogale inornata (unadorned rock wallaby), 295, habitat & habitat loss, 241–2, 249 345, 346 metabolic rates, 239 Petrogale lateralis (black-footed rock wallaby), 295, population rates, 240 300, 344, 345, 346, 352–3, 354, 355 predators, 240–1, 241 species complex, 347, 348 reproduction in, 240 Petrogale mareeba, 295, 345, 346 reproductive data, 235 Petrogale penicillata (brush-tailed rock wallaby), 295, survival in small patches, 242 345, 346, 351, 352 water metabolism, 232, 239 species complex, 347–8 Petaurus sp., 28, 29, 186 Petrogale persephone (Proserpine rock wallaby), 295, Petaurus abidi, 188, 212 345, 346, 346, 348 Petaurus australis (yellow-bellied glider), 213–8, Petrogale purpureicollis (purple-necked rock wallaby), Plate 11 295, 345, 346, 347 conservation, 217–8 Petrogale rothschildi (Rothschild’s rock wallaby), 295, diet, 213–4, 226 345, 346, 346 distribution, 188, 213 Petrogale sharmani, 295, 345, 346 home range, 214–6, 215 Petrogale xanthopus (yellow-footed rock wallaby), 295, indicator species, 217 345, 346, 351, Plate 14 reproduction, 216–7 species complex, 346, 348 scent glands, 216 study of, 352 social behaviour, 213, 214, 216 Petropseudes dahli (rock ringtail possum), 228, 247, vocalisation, 214–5 248 Petaurus breviceps (sugar glider), 158, 206, Plate 11 Phalanger sp. 27, 28, 29, 186 dentition, 208 Phalanger gymnotis (ground cuscus), 230 diet of, 207–9 Phalanger intercastellanus (southern common cuscus), distribution of, 188 228, 229, 247, 344 effects of competition, 212 Phalanger orientalis (northern common cuscus), 3, 4, energy conservation, 209–11, 209, 211, 228, 229 frontal & sternal gland, 207 phalangerid sperm, 20 gliding membrane & gliding, 193, 211 Phalangeridae, 21, 22, 22, 28, 228, 229 nitrogen requirements, 124, 209 Phascogale sp., 143, 144, 145, 155, 153 reproduction & life history, 211–2 Phascogale calura (red-tailed phasocogale), 145, 161 social behaviour, 206–7, 212 Phascogale tapoatafa (brush-tailed phascogale), 141 standard metabolic rate, 210–11, 211 Phascogalinae, 28, 29, 143, 144, 145 testosterone levels, 207 Phascolarctidae, 21, 22, 28, 227 Petaurus gracilis (mahogany glider), 188, 212 Phascolarctos sp., 186 Petaurus norfolcensis (squirrel glider), 188, 212–3 Phascolarctos cinereus (koala), 28, 29, 230–8, Plate 12 Petrogale sp., 28, 29, 293, 344 bacteria in caecum & colon, 231 diet & dentition, 298–9 behaviour & ecology, 234 species complexes, 346–8, 346 birth posture, 70 Petrogale assimilis (allied rock wallaby) brain, 23, 233, 234 diet of, 348–9 breeding strategy, 50 case study, Black Rock, Queensland, 348–52 dentition, 230, 231 distribution, 295, 345 dependancy on habitat, 242 genetic diversity, 346, 351 diet selection, 231 home range, 349, 349 diet, 230 reproduction, 350 digestion, 231 & rainfall, 351 distribution, 227 social behaviour, 349–50 energy metabolism, 232–4 436 Life of Marsupials

eucalypt leaf selection, 224 platypus, see Ornithorhyncus anatinus food processing, 231 Platyrhinni (New World monkeys), 106 lactation, 234, 236 Pleistocene epoch, metabolic rates, 234 vegetation changes, 367 nitrogen metabolism, 232, 232 Poa sp., 271, 324 on Kangaroo Island, 237–8 Podocarpus sp., 32 passage of food through gut, 147 Podocarpus totara (totara), 260 population decline, 236–7 pollen, 185 reproduction, 60, 234–7, 235 digestion by pygmy possums, 187, 189 sperm heads, 20 pollen & nectar feeders, 187–202 water balance, 231–2 Polyprotodontia, 17, 18, 19, 24 Phascolomys medius, 269, 270 Portulaca olevacea, 325 Phascolonus gigas, 269, 270, 284, 371 possums, Phascolosorex, 145 control in New Zealand, 254 phenolics, in eucalypt leaf, 222–3, 224, 239 relationships based on DNA hybridisation Philander sp., 28, 29, 106, 115, 117, 119 criteria, 186 Philander andersoni, 119, 120 possums & ringtail possums (common names), Philander opossum (gray four-eyed opossum), 126–7, brushtail, see Trichosurus vulpecula 126, Plate 4 common ringtail, see Pseudocheirus peregrinus density of, 126 coppery brushtail, see Trichosurus vulpecula distribution, 126 johnstoni diet, 126 feather-tail, see Distoechurus sp. habitat, 119, 126 green ringtail, see Pseudochirops archeri home range, 126 Herbert River ringtail, see Pseudochirulus reproduction, 127 herbertensis survival of, 137 honey, see Tarsipes rostratus phosphorous Leadbeater’s, see Gymnobelideus leadbeateri in milk, 86 lemuroid ringtail see Hemibelideus lemuroides photoperiods & breeding, 40, 130, 176, 177, 192, 252, lowland ringtail, see Pseudochirulus canescens 328, 343 mountain brushtail see Trichosurus caninus in dasyurids, 156–7, 157, 160 pygmy ringtail, see Pseudochirulus mayeri in tammar wallaby, 337–42, 337, 338, 339, 340 rock ringtail, see Petropseudes dahli Phyllocladus, 32 scaly-tailed, see Wyulda squamicaudata physiological functions striped, see Dactylopsila trivirgata & independence, 93–99 western ringtail possum, see Pseudocheirus & postnatal development, 86 peregrinus occidentalis Phytophthera sp., 363 see also pygmy possums Pilbara, Western Australia, 329, 330 postnatal development, 86–100 pituitary gland, 44, 45, 79 potassium ions in blood plasma, 99 in fetus, 64, 66, 68 Potoroinae, 22, 291, 294 in red kangaroo, 314, 324 dentition, 292, 294 in tammar, 341–2 distribution, 291 placental transport, 59–63 potoroos (common names), Placentalia, 5 long-footed, see Potorous longipes planigales (common names) long-nosed, see Potorous tridactylus Giles’, see Planigale gilesi Potorous sp., 302, 359 narrow-nosed, see Planigale tenuirostris diet of, 359, 361 long-tailed, see Planigale ingrami Potorous longipes, 291, 359 Planigale sp., 141, 143, 144 Potorous platyops, 291 Planigale gilesi (Giles’ planigale), 158 Potorous tridactylus (long-nosed potoroo), 291 Planigale ingrami (long-tailed), 10, 12, 149 chromosomes, 23 Planigale tenuirostris (narrow-nosed planigale), 158, end of diapause, 57 Plate 6 locomotion & oxygen consumption, 309, 309 plant food, newborn, 7 eaten by marsupials, 185, 185 range of, 292 preferences by red & grey kangaroos, 321, 325 stomach of, 304 see also nectar; foliage; fungi; pollen weight of, 359 Index 437

Potos flavus (kinkajou), 108, 123 Pseudantechinus sp., 144 pouch, 3, 4, 76–8 Pseudantechinus macdonnellensis, 354, 355 anatomy of, 77, 78, Pseudocheiridae, 21, 24, 28, 227–9 pouch cleaning, 69 Pseudocheirus sp., 186 pouch life phases, Pseudocheirus peregrinus occidentalis (western ringtail acquiring immune competence, 87–8 possum), 228, 228, 247, 248 becoming physiological independent, 93–9 Pseudocheirus peregrinus (common ringtail possum), getting wired up, 88–93 243–7, Plate 12 growing up and leaving, 99–100 nitrogen requirements, 124 predators, caecum, 244 extinction of marsupial predators, 163 costs of lactation, 245–6 introduced, 376 dentition of, 230, 243 predatory marsupials, see marsupial carnivores dependency on forest habitat, 242 pregnancy, diet & toxicity of leaves, 223, 224 in small marsupials, 8, 187 diet of, 223, 243, 248 development of embryo, 41 digestability & nitrogen requirement, 232 hormonal control, 44–5 distribution of, 21, 227–8, 228, 249 Procoptodon sp., 371 energy metabolism, 233, 245 procyonids, 108 maintenance nitrogen requirement of, 232, 245 progesterone, metabolic rates, 245 in corpus luteum, 43, 44 nitrogen metabolism, 245 in macropod reproductive cycle, 312 reingestion of caecal pellets, 244, 244, 245, 248 in oestrous cycle, 45–6, 46 reproduction, 217, 246, 246–7 role in parturition, 63 reproductive data, 235 progesterone levels, social behaviour, 246 at parturition, 64, 65 Pseudochirops sp., 28, 29, 186, 229 & photoperiod in tammar wallaby, 341 Pseudochirops alberti, 248 prolactin, Pseudochirops archeri (green ringtail possum), 228, & parturition, 64, 65, 66, 68 (Box 2.3) 247, 248, 248, 249 & milk synthesis, 79 Pseudochirops corinnae, 248 profile in tammar wallaby, 341–2 Pseudochirops coronatus, 248 prolactin receptors on mammary glands, 80, 134 Pseudochirops cupreus, 247, 248 Propleopinae, 290, 290 Pseudochirulus sp., 28, 29, 186, 229 Propleopus oscillans, 290, 303, 370 Pseudochirulus canescens (lowland ringtail possum), prostaglandins, 247, 248 & parturition, 64, 65, 66, 67 (Box 3.1) Pseudochirulus caroli, 248 & the birth position, 69, 70 Pseudochirulus cinereus (Daintree River ringtail effect on birth behaviour, 64, 70 possum), 228, 247, 248

PGF2a, 67, 69 Pseudochirulus forbesi, 248 PGFM, 67 (Box 2.1), 69 Pseudochirulus herbertensis (Herbert River possum), prostate gland, 46, 50 228, 247, 248, 248, 249 weight change, 49 Pseudochirulus mayeri (pygmy ringtail possum), 247, Protea sp., 185 248 Proteaceae, 32 Pseudochirulus schlegeli, 248 protein analysis, & marsupial relationships, 24, 187 Pseudopanax arboreus, 262 protein nitrogen, 185, 186, 226 pygmy possums, 187–92 proteins, 20, 187 pygmy possums (common names) in milk, 83, 84, 134 eastern, see Cercartetus nanus concentration in fetal fluids, 47, 62 little, see Cercartetus lepidus late lactation protein (LLP), 84, 85 long-tailed, see Cercartetus caudatus maternal immunoglobins, 87 mountain, see Burramys parvus uterine secretion, 47, 48 western, see Cercartetus concinnus Protemnodon genus, 298 protocone, 16, 168 Quarternary period Prototheria, 5 history of marsupials, 33 protozoa, quokka, see Setonix brachyurus for carbohydrate fermentation, 99 quolls (common names) in stomach of kangaroos, 305, 329 eastern, see Dasyurus viverrinus, 438 Life of Marsupials

northern, see Dasyurus hallucatus rock art, 365, 367, 372, Plate 16 spotted-tailed or tiger, see Dasyurus maculatus rock wallabies, western, see Dasyurus geoffroii chromosome numbers, 300 conservation strategies, 352–3 rabbits, wild, see Oryctolagus cuniculus diet & dentition, 299 radiation of placental mammals, 32 distribution, 295, 344, 345 Ramsayia sp., 270 habitat, 344 rangelands, island populations, 344 pasture growth, 326–8, 326 spread across Australia, 348 plant selection by kangaroos & sheep, 321, 325 taxonomic relationships of, 345–6, 346 plant selection by wombats, 271, 282, 283 species complexes, 346–8 rat kangaroos, 185, 291–3, 359–61 see also hare wallabies; nailtail wallabies; wallabies dentition, 291, 298 rock wallabies (common names) destruction of, 377, 379 allied, see Petrogale assimilis diet, 291, 359 black-footed, see Petrogale lateralis embryonic diapuse, 56, 291 brush-tailed, see Petrogale penicillata in North Queensland, 361–3 little or nabarlek, see Petrogale concinna response to conditions after fire, 361 monjon, see Petrogale burbidgei stomach structure & function, 305 Proserpine, see Petrogale persephone see also Aepyprymnus; Bettongia purple-necked, see Petrogale purpureicollis rat kangaroos (common names), Rothschild’s, see Petrogale rothschildi desert, see Caloprymnus campestris Sharman’s, see Petrogale sharmani musky, see Hypsiprymnodon moschatus short-eared, see Petrogale brachyotis rufous, see Aepyprymnus rufescens unadorned, see Petrogale inornata Rattus exulans (Pacific rat), 259 yellow-footed, see Petrogale xanthopus red-tailed phascogale, see Phascogale calura Rottnest Island, Western Australia, marsupial popula- reingestion of caecal pellets, 244, 244, 247, 248 tions, 354 relationships within marsupials, 14–27 Royal Commission of 1901 into the Western Lands relaxin, 42, 43, 63, 68 Division of New South Wales, 377 reproduction of marsupials rufous bettong, see Aepyprymnus rufescens control of sexual differentiation, 8–9 & day-length, 40, 130 Saccharum edule (lowland pitpit), 208 energetics of, 101–2, 101 saliva of kangaroo, 305, 307 manner of, 35 speading on skin, 310 maternal investment in 5 species, 174 salts in marsupial milk, 85–6 size at birth, 7, 8 sap, see also big bang breeders; male die-off as food, 185 see also under individual species composition of, 185 reproduction & development, 39–102 sap, gum & insect feeders, 202–17 reproductive organs, Sarcophilus harrisii (Tasmanian devil), 141, 147, 148 anatomy of, 5–7 field energetics, 150 birth canal, 7 field metabolic rate, 150 differences between marsupials & placentals, 6, 6 food rates through gut, 147 marsupial male & female, 6 relationship with carnivores, 144 placental male & female, 6 reproduction in, reproductive process & body size, 100–2 follicle production, 43 reptiles, metabolic rates, 9 SC see superior colliculus research animals, 123 SCFA see short chain fatty acids respiration, scent glands, cutaneous, in newborn marsupials, 74–5 in brushtail possum, 256 pulmonary, 75 in sugar glider, 207 Rhipogonum scandens (supplejack), 261, 262 in yellow-bellied glider, 216 Rhizophascolonus, 269 scrotum, in marsupials, 8–9, 49–50, 70, 76, 121, 127, Rhynchomeles prattorum, 170, 171 Plate 10 Rhyncholestes raphanurus. 110, 111, 112 sea level changes, 368, 368 Rhyncholestes sp., 28, 29, 115 & human arrival in Australia, 369–70, 369 ringtail possums, 243–9 sense organs of neonatus, 73 see also possums & ringtail possums serology, 24 Riversleigh fossil site, Queensland, 30, 34, 289, 302 Sertoli cell of testis, in tammar, 48 Index 439

Setonix brachyurus (quokka), 28, 29, 57 somatosensory system, 90 browsers, 298 somites of early embryo, 59, Plate 3 birth posture, 70 South American marsupials, 110–30 chromosome numbers, 300 distribution of, 111 distribution, 293, 294 high altitude species, 110–12 fetal membranes, 60 high latitude species, 112–13 locomotion & oxygen consumption, 309 history of, 105–8 newborn, 7 natural regions of South America, 108–10, 109 sex hormone binding globulin, SHBG, 43 neotropical marsupials, 114–30 sexual differentiation in marsupials, 8–9 see also American opossums; opossums sexual dimorphism, South Australia, island marsupial populations, 353–4 in bandicoots, 180 spermatogenesis, 49, 50, 51, in Didelphis, 114, 115 & hot weather, 314 in honey possum, 191 single cycle, 153 in kangaroos, 100, 294, 295, 299, 322 spermatogonia, 48 in Leadbeater’s possum, 203 spermatozoa, 48–50 in possums and gliders, 196, 206 competition, 46, 52, 155 sexual maturity, 86, 100 conjugated pairs in American marsupials, 168 sheep, 18, 101, 102, 304, 306, 320–1, 339 heads, 19 competition with, 161, 178, 281–3, 315, 325, 327, in female tract, 51–2 329–32, 339, 347, 356–7, 363, 378–9 morphology of 19–20, 20 numbers in Australia, 315, 330, 377, 377 production of, 51 see also stock Spilocuscus sp., 186 short-beaked echidna, see Tachyglossus aculeatus Spilocuscus maculates (common spotted cuscus), 228, short-day breeders, 40 229, 247, 249, 344, Plate 13 short chain fatty acids (SCFA), 225, 231, 303, 306, 306 spinifex, see Triodia pungens; Triodia mitchelli short-faced kangaroos, see Sthenurinae standard metabolic rates (SMR), 9, 10, 11, 9–14 sideroxylonal, 223, 224 in arboreal marsupials, 233, 232–3, 239, 245, 252 Smilodon sp. 106 in bandicoots, 10, 173 Sminthopsinae, 28, 143, 144 in dasyurids, 10, 149, 158, 198 Sminthopsis sp., 11, 143, 144, 217 in kangaroos, 10, 100, 317, 323 Sminthopsis crassicaudata (fat-tailed dunnart), 17, in small possums and gliders, 189, 198, 199, 204, 147, 148, 158–60 210 breeding, 160 in wombats, 10, 274, 277 diet, 158 starvation, 11, 237, 322, 328, 330 distribution, 158 Steropodon sp., 27 energy conservation, 158 Sthenuridae, 28 metabolic rates, 14, 158 Sthenurinae, 22 oxygen consumption, 160 distribution, 291, 303 reproduction, fertilisation, 52 Sthenurus (giant kangaroo), 142, 370 social behaviour, 158 Stipa nitida, 271 standard metabolic rate, 158 stock, introduction of, 330 torpor, 159, 159, 198 competition with kangaroos, 314–15, 315, 329–31 Sminthopsis douglasi (Julia Creek dunnart), 75, 147, see also grazing; sheep 161, Plate 6 stylar shelf of molars, 16, 17, 18, 168 Sminthopsis macroura (stripe-faced dunnart), 148, sucking, 149, Plate 1 adaptations for, 74 Sminthopsis virginiae(red-cheeked dunnart), 70 stimulus and control of lactation, 80–1 SMR see standard metabolic rate sugars, in nectar, 185 sodium fluoroacetate in native plants, 161, 251, 317 summer solstice, relation to season of birth, 338–40, sodium fluoroacetate (1080), 161, 292–3 337, 338, 339, 342, 343 as control agent, 264, 292–3, 352, 383 Suncus etruscus (Etruscan shrew), 143 lethal dose, superior colliculus (SC), 88, 90 in brushtail possums, 251 surfactant, in lungs of newborn marsupial, 76 in foxes, 161 swallowing in newborn, 74 in tammar wallabies, 334 sweat glands, 329 in western grey kangaroos, 334 sweating, 319 tolerance to, 161 swimming speeds, 127 sodium ions in blood plasma, 99 Sylvilagus, 244 440 Life of Marsupials

symmetrodonts, 27 thylacine, see Thylacinus cyanocephalus Symphyomyrtus group of eucalypts, 224 Thylacinidae, 21, 22, 28, 145–6 syndactyly, 24, 167 Thylacinus cyanocephalus (Tasmanian tiger), 28, 29, 106, 141, 145–6, 370, 380 Tachyglossus aculeatus, 5, 55 remaking of, 161–2 tannins in Eucalyptus leaf, 222 Thylacoleo carnifex (marsupial leopard), 141, 142, Tarsipedidae, 21, 22, 28, 186,, 187 163, 370 Tarsipes sp., 28, 29 Thylacomyidae, 22, 28, 171 Tarsipes rostratus, (honey possum or noolbenger), Thylacosmilus atrox, 106, 107, 107 185, 185, Plate 10 Thylacotinga bartholomaii, 32 chromosome number, 186 Thylamyinae, 115, 117 diet & feeding, 186, 187, 189, 190, 191 Thylamys elegans, 111, 112 distribution, 187, 188 Thylamys sp. 28, 29, 106, 115, 117 metabolic rates, 189, 191 Thyogale sp., 28, 141, 293, 297 reproduction, 7, 7, 191–2 Thylogale billardierii (red-bellied or Tasmanian embryonic development, 55 pademelon), 141, 293, 295, 353 embryonic diapause, 56, 191 Thylogale browni, 295 lactation, 191 Thylogale brunii (dusky pademelon), 4, 295 sperm size & forms, 19, 20 Thylogale calabyi, 295 torpor, 191 Thylogale stigmatica (red-legged pademelon), 295 weight, 69, 187 Thylogale thetis (red-necked pademelon), Tasmania, 353 chromosome numbers, 300, Plate 1 bounty payments, 379 distribution of, 295 human occupation, 370 jaw muscles, 296 islands, 359 stomach of, 304 marsupial populations, 354, 359 thymus, 87 Tasmanian bettong, see Bettongia gaimardii thyroid gland, 93, Tasmanian devil, see Sarcophilus harrisii development of, 95–6, 96 Tasmanian tiger, see Thylacinus cyanocephalus thyroxine, 95 teats Tingamara porterorum, 34, 138 arrangements of, 77 torpor, 159–6, 198 ilio-marsupialis muscle, 76, 77 in pygmy possums, 187, 191, 198–200, 199, 210 numbers of, 78 in opossums, 112, 113 teeth & relationships of marsupials, 14–18 in small dasyurids, 159–60, 159 see also dentition tree-kangaroos, temperature see thermoregulation as browsers, 297 terpenes in Eucalyptus leaf, 224, 239 chromosome numbers, 301 Tertiary Period distribution of, 295 Australasia, 32–4 stomach structure & function, 305 climate changes, 31 tree-kangaroos (common names), continental movements, 30–1 Bennett’s, see Dendrolagus bennettianus vegetation change, 32 Goodfellow’s, see Dendrolagus goodfellowi Tertiary Period marsupials, Huon, see Dendrolagus matschiei history of, 30 Lumholtz’s, see Dendrolagus lumholtzii mammal history of Australasia, 32–4, 33 tree living marsupials, 4, 221–265, 293, Plates 11–14 of Australasia, 32, tree sloths, see Bradypus tridactylus of South America, 32 Trichosurus sp., 22, 28, 29, 141, 186, testes, 8–9, 191 Trichosurus caninus (mountain brushtail possum, testosterone, bobuck), 203, 228, 229, 235 & hot weather in red kangaroo, 314 dependency on forest habitat, 242 & male die off, 154 diet of, 249–50 Themeda sp., (tropical grass), 271 reproduction, 250 thermoregulation, 93, 95, 96 Trichosurus vulpecula (brushtail possum), 229, 250–8, development of, 78, 93–5 Plate 13 in dasyurids, 148–9 arms race in New Zealand forests, 259–61 in Virginia opossum & the quokka, 94 body mass & breeding success, 253 major organs involved, 95 breeding biology, 252–4 see also body temperature density variations, 254–6, 255 Index 441

dependency on forest habitat, 242, 249 vision, 88–90 diet of, 251 in kangaroos, 312–13 digestability & nitrogen requirement, 232 vocalisation, distribution, 228, 229, 250–1 in sugar glider, 207 energy metabolism, 233, 252 in yellow-bellied glider, 214–6 eucalypt leaf selection, 224 volatile fatty acids, see short-chain fatty acids developmental stages, Plate 3 Vombatidae, 19, 21, 22, 28 home ranges, 256 origin of burrowing, 284–5 impact on forests in New Zealand, 259–61 Vombatus sp., 28, 29, 186, 270 density & tree mortality, 260 Vombatus hacketti (Hacketti’s wombat), 270, 270 long-term responses, 261–4 Vombatus ursinus (common wombat), 141, 267, juvenile dispersal & mortality, 256–8, 257 brain, 233 life expectancy, 258 breeding strategies, 281–2 metabolic rates, 252 burrowing & burrows, 271, 274–80, 284 New Zealand studies, 252–8 competition with stock, 282–3 nitrogen metabolism, 252 dentition, 271 on Barrow Island, 355 diet, 271 predators of, 258 distribution of, 270, 270 reproduction, effects of forestry activites, 283–4 birth & pouch exit, 39, 40, 48 energetics, 274 birth posture, 70 feeding adaptations, 273 control of birth, 69 movements of, 279 embryonic development, 55 reproduction, 281–2 fertilisation, 52 standard metabolic rate, 10, 274 lactation, 236, 253 Vulpes vulpes (European red fox), 141, 160, 290, 363, mating patterns, 256 376, 377, 380 newborn, 7, 72 in Central Australia, 381 oestrous cycle, 42, 44 predation on rock wallabies, 352–3 reproductive data, 235 spermatozoa, 48–50 wallabies, Plate 14, Plate 15 social behaviour & ecology, 254 classification by teeth, 294–7, 298 tolerance to sodium fluoroacetate, 251 classification by chromosomes, 300–1 Trichosurus vulpecula johnstoni (coppery brushtail classification on DNA criteria, 301–2 possum), 247, 248, 249 island wallabies, 353–9 tri-iodothyronine, 95 New Guinea wallaby browers, 297 Triodia sp., 326 Notomacropus group, 332 Triodia mitchelli (spinifex), 235 hare wallabies, 299 Triodia pungens (spinifex), 320, 321 rock wallabies, 344–53 trioks, 205, 206 wallabies (common names) see also Dactylopsia megalura; Dactylopsia agile, see Macropus agilis palpator Bennett’s, see Macropus rufogriseus rufogriseus black-striped, see Macropus dorsalis urea, parma, see Macropus parma & nitrogen, 303 red-necked, see Macropus rufogriseus excretion of, 98, 99, 307–8, 307 swamp, see Wallabia bicolor in wombats, 273 tammar, see Macropus eugenii recycling of, 273, 307–8, 336 whiptail, see Macropus parryi urine, see also hare wallabies; nailtail wallabies; rock concentration of, 93, 98, 99 wallabies in Virginia opossum, 132 Wallabia bicolor (swamp wallaby), 21, 28, 293 in pouch young, 97 chromosome number, 301, Plate 1 urogenital tract, newborn, 7 in placental & marsupial mammals, 6, 6 skull & dentition, 292, 298 uterine changes during oestrous cycle, 47–8, 47 Wallace’s Line, 228 wallaroos (common names) vaginal complex, changes during oestrous cycle, 46 antilopine, see Macropus antilopinus vasopressin, 99 black, see Macropus bernardus Virginia opossum, see Didelphis virginiana common, eastern or euro, see Macropus robustus 442 Life of Marsupials

Wanburoo, 302 dentition, 272, 272 Warendja wakefieldi, 269, 270 diet of, 271–4 warrens, 358 home ranges, 271 water conservation and balance, fossil, 269 in dasyurids, 149, 150, 151, 152 wombats (common names), in greater glider, 232, 239 common, see Vombatus ursinus in koala, 231–3, 232, 238 Hackett’s, see Vombatus hacketti in macropods, 100, 305, 307–8, 319–3, 335, 349, northern hairy-nosed, see Lasiorhinus krefftii 358, 361 southern hairy-nosed, see Lasiorhinus latifrons in small possums and gliders, 89, 204, 209 woodchuck see Marmota monax in Virginia opossum, 120, 124, 132 Wyulda sp., 186 in wombats, 232, 273–4, 276–7 Wyulda squamicaudata (scaly-tailed possum), 229 in young marsupials, 93, 95, 97 water economy & kidney function, 97–9 Xanthorrhoea, 190 water for stock, 314, 330 weaning, 100–1, 235 yellow-footed antechinus, see Antechinus flavipes weight of young marsupial, 100 yolk sac, Weinmannia racemosa (kamahi), 261, 262 Western Australia, marsupial island populations, 354 development of, 61–2 Western Australian Museum, 330 yolk sac placenta, 62–3 whiptail wallaby, see Macropus parryi young marsupials whiskers, development of, 90–2 from birth to independence, 76–86 surface pattern of, 91 growing up & leaving the pouch, 99–100 winter solstice & reproduction, independence of, 86–100 in bandicoots, 176 see also neonatus or newborn marsupial; in dasyurids, 40, 156, 157, 160 postnatal development in tammar wallabies, 337, 338, 342 Yucatan Peninsula, 30 wombats, 185 antecedents, 269–71 zona pellucida, 53, 59, 60 burrowing habit, 271–3 Zygomaturus sp., 269, 371