The Biology of African Savannahs THE BIOLOGY OF HABITATS SERIES
This attractive series of concise, affordable texts provides an integrated overview of the design, physiology, and ecology of the biota in a given habitat, set in the context of the physical environ- ment. Each book describes practical aspects of working within the habitat, detailing the sorts of studies which are possible. Management and conservation issues are also included. The series is intended for naturalists, students studying biological or environmental science, those begin- ning independent research, and professional biologists embarking on research in a new habitat.
The Biology of Streams and Rivers Paul S. Giller and Björn Malmqvist The Biology of Soft Shores and Estuaries Colin Little The Biology of the Deep Ocean Peter Herring The Biology of Lakes and Ponds, 2nd Edition Christer Brönmark and Lars-Anders Hansson The Biology of Soil Richard D. Bardgett The Biology of Mangroves and Seagrasses, 2nd Edition Peter J. Hogarth The Biology of Polar Regions, 2nd Edition David N. Thomas et al. The Biology of Deserts David Ward The Biology of Caves and Other Subterranean Habitats David C. Culver and Tanja Pipan The Biology of Alpine Habitats Laszlo Nagy and Georg Grabherr The Biology of Rocky Shores, 2nd Edition Colin Little, Gray A. Williams and Cynthia D. Trowbridge The Biology of Coral Reefs Charles R.C. Sheppard, Simon K. Davy, and Graham M. Pilling The Biology of Disturbed Habitats Lawrence R. Walker The Biology of Freshwater Wetlands, 2nd Edition Arnold G. van der Valk The Biology of Peatlands, 2nd Edition Håkan Rydin and John K. Jeglum The Biology of African Savannahs, 2nd Edition Bryan Shorrocks and William Bates The Biology of African Savannahs
SECOND EDITION
Bryan Shorrocks Ph.D. Emeritus Professor University of Leeds and Honorary Professor University of York Environment Department University of York Heslington, York United Kingdom and William (Bill) Bates, Ph.D. Wildlife Bioacoustics Research Professor Emeritus of Zoology University of British Columbia Kelowna, British Columbia Canada
1
The Biology of African Savannahs. Second Edition. Bryan Shorrocks & William Bates. © Bryan Shorrocks & William Bates 2015. Published 2015 by Oxford University Press. 1 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Bryan Shorrocks & William Bates 2015 The moral rights of the authors have been asserted First Edition published in 2007 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2014946890 ISBN 978–0–19–870270–2 (hbk.) ISBN 978–0–19–870271–9 (pbk.) Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work. Bryan and William would like to dedicate this book to Jo and Jeannie who also love Africa and have supported them in all their adventures. Preface to the First Edition
This book is an introduction to African savannahs and their biology, concen- trating on large mammal ecology, behaviour, and conservation. Although it is a book on savannahs I hope that it sets out its ideas within a general frame- work of ecological and behavioural ideas. In my 40 years as an ecologist I have worked on both temperate and tropic- al terrestrial ecosystems. I have carried out field surveys, field experiments and built mathematical and computer models. I believe all these elements are essential for understanding Nature, and they all appear in this book. You cannot simply watch wildlife populations and communities in order to understand how they function. You have to employ field manipulations, either natural or artificial, to tease out the possibilities. You have to describe what you see not just in words but in mathematical models, to allow your less precise verbal suspicions to be tested. And savannah ecology must fit comfortably within the framework of general ecological ideas. There can be no special pleading or special mechanisms that other ecologists just don’t understand. I have tried to present such an encompassing view of savannah ecology and behaviour. African savannahs are magical places. The brightness of the light, the inten- sity of the colours, and the beauty and excitement of the wildlife, and of course the people. These savannahs are worth preserving for future gener- ations and I hope, in a small way, this book will generate an interest that helps them survive. I would like to thank all the people that have helped in my savannah work in East Africa, in particular the people at the Mpala Research Centre, in Laiki- pia, Kenya. Their help and insights have been invaluable during my frequent stays there. Finally my thanks to Paul Ward who, over several beers, over several nights, while running an undergraduate field course in Derbyshire reawakened my latent interest in Africa and its wildlife. Bryan Shorrocks June 2007 Preface to the Second Edition
Once again this book is an introduction to African savannahs and their biol- ogy, concentrating on large mammal ecology, behaviour, and conservation. However, this second edition contains new material and many updates. These include hundreds of new references, new technique boxes, updated distribu- tion maps for the mammals and some birds, and an expanded focus on the conservation of African savannahs. Also new to the second edition is a com- panion website,
Contents
1 Savannahs 1 Distribution world-wide 2 African savannahs 9
2 The Vegetation 27 Rainfall, plant biomass and grass–tree mixture 27 Morphology and life history 30 Grasses 38 Trees 41 Local vegetation patterns: the Serengeti–Mara ecosystem 57
3 The Animals 65 The insects 65 The birds 69 The mammals 74
4 Single-species Populations 137 Estimating numbers 138 What changes numbers? 153 Population models 163 Other species in other areas 168
5 Species Interactions 182 Predator–prey type interactions (+ −) 183 Competitive interactions (− −) 210 Mutualistic interactions (+ +) 227
6 The Savannah Community and Its Conservation 237 Energy flow and food webs 238 Assembly rules 252 x CONTENTS
Island biogeography 257 Conserving savannah ecosystems 264
References 279 Index 309 1 Savannahs
Savannahs constitute one of the largest biomes of the world, comprising about 20% of the land surface. Stated simply, they are tropical and subtropical grasslands, with scattered bushes and trees. Most savannah occurs in Africa, with a smaller amount in South America, India and Australia. In Africa the trees tend to be deciduous, while in South America and Australia they tend to be evergreen. Savannahs occur around the equator, (between the Tropic of Capricorn and the Tropic of Cancer), where it is warm, but relatively dry. Most experience seasonal drought and the vegetation is influenced by rain- fall, soil type, grazing, browsing and fire. The word savannah, or savanna, is probably derived from a 16th century Spanish word zavanna, meaning ‘treeless plain’. It was recorded in 1535, by the Spanish historian Gonzalo Fernandez de Oviedo, as coming originally from Carib, a language spoken in northern South America and the Caribbean. Savannah therefore shares its origin with other well known words such as barbecue, cannibal and papaya. Like many plants that are found in warmer and drier regions, the dominant
grasses in savannahs tend to be C4 plants. Only in some very wet environ-
ments do C3 grass species become abundant. The terms 3C and C4 refer to
the type of CO2 trapping mechanism (photosynthesis) used by the plant. C4
plants have evolved a secondary carbon fixation pathway. CO2 is first com- bined into a 4-carbon compound, in the mesophyll cells of the leaf, and then
passed to the cells around the leaf veins where the CO2 is released at high concentrations. It then enters the usual photosynthetic carbon reduction
(PCR) pathway, or Calvin-Benson cycle, used by C3 plants. C4 plants are
capable of utilising higher light intensities than C3 plants, have greater max-
imum photosynthesis, and use less water in the process. Because of these C4 grasses, the photosynthetic efficiency of many savannahs is very high. How-
ever, C4 plants are extremely poor quality food for most herbivores, verte- brate or invertebrate, (Caswell et al. 1973), unless the animal can break down cellulose. Intriguingly, both ungulates and termites, which are common in many savannah systems, have a symbiotic gut flora that produces an enzyme, cellulase, that can digest this plant cell-wall constituent.
The Biology of African Savannahs. Second Edition. Bryan Shorrocks & William Bates. © Bryan Shorrocks & William Bates 2015. Published 2015 by Oxford University Press. 2 THE BIOLOGY OF AFRICAN SAVANNAHS
Distribution world-wide
Figure 1.1 shows how global temperatures and precipitation have an influ- ence on the major biomes, and Figure 1.2 shows a map of the world with the major areas of savannah indicated. In total they occupy some 23 million km2 (Cole 1986). Mean annual rainfall typically varies between about 20 and 150 cm, with 60 to 90% of the year’s rain falling in a short period of a few months. Although climatic factors, such as the annual temperature and the annual amount of rain, are not the only determinants of the savannah biome, they do potentially have a major effect. The computer model BIOME 3 pre- dicts a distribution of savannahs, based solely on climatic factors, almost identical to that of Figure 1.2 (Haxeltine and Prentice 1996). On a World scale, therefore, savannahs tend to occupy a climatic region between deserts and tropical forests, a picture that is seen very clearly in Afri- ca. Here the huge savannah area surrounds the tropical forests of the Congo basin and to the north is bordered by the Sahara desert and to the south by the Namib and Kalahari deserts. This climatic position is seen clearly in Fig- ure 1.1. Because savannahs are defined as grasslands with varying amounts of tree cover, they also sit between grasslands and tropical seasonal forests in Figure 1.1. However, the boundaries between these biomes are never very clear cut, and grassland and seasonal forest frequently merge into savannah.
–15
–10 Tundra
C) 5 ° – Northern 0 coniferous forest 5 Temperate forest 10 Temperate rainforest 15 D e Grassland s Mean annual temperature ( 20 e r Tropical Savannah Tropical 25 t seasonal forest forest 30 050 100 150 200 250 300 350 400 450 Mean annual precipitation (cm)
Figure 1.1 Distribution of the major terrestrial biomes of the world with respect to mean annual temperature and mean annual precipitation (after Whittaker 1975). SAVANNAHS 3
Equator
Figure 1.2 World distribution of the savannah biome. It makes up about 20% of the land surface, but on this familiar Mercator world projection it appears less, because those land areas near the poles (North America, Greenland, Europe, and northern Asia) are artificially enlarged.
Savannah areas usually have a positive water regime (rainfall greater than evaporation) during the wet season and a negative balance during the dry season. When the water regime becomes positive during the dry season, savannahs are replaced by forest. Mistry (2000) gives an excellent overview of world savannahs.
South America In South America there are several areas of savannah (Figure 1.3). South of the equator is the extensive cerrado (area 1 on Figure 1.3), comprising an area of 1.55 million km2, about 20% of the area of Brazil. In terms of plant species it is the richest neotropical savannah, with about 430 species of trees and shrubs, about 300 herbaceous species and over 100 grass species (Sarm- iento 1996). Large mammals such as the giant anteater (Myrmecophaga tri- dactyla), yellow armadillo (Euphractus sexcinctus), jaguar (Panthera onca) and maned wolf (Chryocyon brachyurus) compete with the rapidly expand- ing Brazilian agricultural industry, which focuses primarily on soybean, maize and rice. In the southwest of Brazil is a giant flooded area known as the Pantanal. Here the vegetation is a mosaic of forest with grassland and savannah vegetation more typical of the cerrado. Capybara (Hydrochaeris hydrochaeris), several species of monkeys (the howler monkey, Alouatta caraya and the brown cap- uchin, Cebus apella are the most common) thrive in this mosaic of wetland 4 THE BIOLOGY OF AFRICAN SAVANNAHS
8
3 4 9 5 7
1 2
6
Figure 1.3 The major South American savannahs: 1, cerrado; 2, llanos de Moxos; 3, llanos del Orinoco; 4, Gran Sabana; 5, savannahs of the Rio Branco-Rupununi; 6, chaco; 7, Amazonian campos; 8, llanos of the Magdalena; 9, coastal savannah of the Guayanas (after Solbrig 1996, The diversity of savanna ecosystems, in O.T. Solbrig, E. Medina and J.F. Silva, ed. Biodiversity and savanna ecosystem processes. A global perspective, p. 5, with kind permission of Springer Science and Business Media).
and savannah, along with the occasional maned wolf, jaguar, several small forest cats (of the genus Felis), giant anteaters and giant river otters (Pter- onura brasiliensis). Another area of wet and frequently flooded savannah is the llanos de Moxos (area 2 on Figure 1.3) and south of this area is the chaco (area 6), a region of woodland and savannah. It occupies an area of approxi- mately 1,000,000 km2. The southern part of the chaco, like certain savannahs in southern Africa, is unusual in that it can have winter frost. The chaco has many of the same mammals as the cerrado, and like the cerrado it is threat- ened by the rapid expansion of soybean plantations. North of the equator is an area called the llanos del Orinoco (area 3 on Fig- ure 1.3), situated around the Apure-Orinoco river system in Colombia and Venezuela and dominating an area of approximately 500,000 km2. Most of the 12 million cattle in Venezuela are bred here, on 30% of the country’s land area, but with only 13% of the population. Parts of the llanos del Orinoco are flooded in winter (May to October) by the Orinoco river and in these flooded areas the capybara and marsh deer (Blastocerus dichotomus) have adapted themselves to a semi-aquatic life, wading in the silt-bearing floodwater by day and sleeping on higher ground at night. Other smaller areas of savannah SAVANNAHS 5
north of the equator are the Gran Sabana (area 4) in Venezuela, characterised by a unique flora, the coastal savannahs of the Guayanas (area 9) and the savannahs of the Rio Branco-Rupunini (area 5) in Brazil (Figure 1.3). In South America a distinct savannah fauna is not well developed; most South American mammals and birds are not restricted to savannah habitats. Most large herbivores of grasslands are ungulates (mammals with hooves), but although the continent has 21 species of ungulates, only three are savann- ah species: the pampas deer (Ozotoceros bezoarticus), the marsh deer (Blas- tocerus dichotomus) and the white-tailed deer (Odocoileus virginianus). As already noted, there is a fourth neotropical herbivore, the capybara, asso- ciated with savannah, particularly the flooded llanos. It is the largest living rodent (average weight 49 kg, although a maximum of 91 kg has been record- ed). When the first European naturalists visited South America they called these giant rodents ‘water pigs’ or ‘Orinoco hogs’. Capybaras are exclusively herbivorous, feeding mainly on grasses that grow in or near water. Unfortu- nately, a major constituent of these grasses is cellulose, which no mammal can break down. However, the capybara has a fermentation chamber, the caecum (of which the human appendix is a vestige), containing symbiotic microorganisms which can digest cellulose. However, because the caecum is located after the small intestine, most of the products cannot be absorbed. The capybara therefore resorts to coprophagy (reingestion of faeces) in order to take full advantage of the symbionts’ efforts (see also Chapter 3).
Africa In Africa, savannahs are very widespread (Figure 1.2; see also Figure 1.9). They extend in a broad semicircle, sandwiched between rain forest and des- ert, from the west coast to East Africa and then round to Angola and Namib- ia. However, this more or less continuous band of savannah vegetation differs considerably in its detailed composition. In the north, the band of savann- ah changes from more wooded to more open grassland as you move from the wetter Congo basin to the dryer Sahara. South of the Congo basin the same gradation occurs within the miombo woodland savannah, as you move south towards the dryer regions of Angola, Namibia, and the South Afri- can veldt. Joining these two regions are the classical savannahs of Kenya and Tanzania. The second half of this chapter looks at these African savannahs in more detail.
Australia Savannahs are widespread in the north of Australia (Figure 1.4). However, these savannahs are not uniform. In the east, temperatures are lower, rain- fall higher and the dry season shorter; in the north, the climate is warmer with less rainfall and a longer dry season. In addition, there is a gradient of 6 THE BIOLOGY OF AFRICAN SAVANNAHS
Monsoon tallgrass Tropical tallgrass Subtropical tallgrass Tussoc Midgrass Midgrass on clay
Figure 1.4 Savannahs of northern Australia (after McKeon et al. 1991 and Mott et al. 1985).
rainfall from the wetter coast to the drier interior. These combinations of temperature and rainfall produce a variety of savannahs, from the so called ‘tallgrass savannahs’ of the north and east coast, to the ‘midgrass savannahs’ of the interior. The fertility of Australian soils is generally low (Nix 1981) and many native trees such as Eucalyptus have an array of nutrient-conserving mechanisms, such as the ability to extract almost all nutrients from dying leaves. Conse- quently there is a low decomposition rate of litter. These Australian savannahs are dominated by invertebrates, particularly by species of grasshoppers, ants and termites (Table 1.1). These invertebrates are better able to cope with the infertile soils and harsh conditions. Ter- mites can fix nitrogen using symbiotic microrganisms in their hindgut and are able to store grass in their large mounds or terminaria. Because of the low fertility of Australian tropical soils, and the consequent low nutritional value of their grasses, large grazing mammals occur at relatively low den- sities. In fact the major grazing mammals are often feral ungulates, such as water buffalo, donkeys, cattle and horses (Freeland 1990). Interestingly, these introduced feral species are frequently found at a higher density than in their native habitat (Freeland 1991), probably because of a combination of reduced competition, lack of predators and pathogens, and reduced plant defences. The indigenous grazing mammals are macropods, of which there are some 19 species in the northern savannah regions. These include five species of kangaroo, including the large red kangaroo (Macropus rufus) weighing up to 95 kg, eight species of wallaby and six species of rock wallaby. Macropod digestion is aided by a fermentation chamber in the enlarged fore stomach, a modification similar to that found in many eutherian ruminants (Chapter 3). SAVANNAHS 7
Table 1.1 Numbers of various animal species found in savannah habitat in Kakadu National Park, in the north of the Northern Territory, Australia
Taxon Birds Bats Other Snakes Lizards Frogs Termites Grasshoppers mammals Number of 93 14 16 10 36 18 36 47 species
Data from Braithwaite (1991).
The Australian savannahs are similar in many ways to those of Africa, but with Eucalypus species and kangaroos replacing Acacia species and wil- debeest. However, as mentioned above, the main animal biomass is insect rather than large mammal. Of course the large numbers of insect herbivores in the Australian savannahs means that there is lots of food for insectivores, such as lizards. This may account for the fact that Australia has the highest diversity of lizard species in the world (Table 1.1).
India Savannahs in Indian are widespread in the north and east (Figure 1.2), but are thought to be derived from woodland systems through deforesta- tion, abandoned cultivation, and burning (Misra 1983; Gadgil and Meher Homji 1985). These savannahs are prevented from returning to woodland by repeated grazing and burning, two factors that are not unimportant in other savannah systems. However, some of the anthropic savannahs of the north and central region of Rajasthan may have been derived from natural savannah. Although several types of Indian savannah have been described by Dabadghao and Shankarnarayan (1973), they are all rather similar in form and appearance and are defined by their major grass species. These derived Indian savannahs contain several species of mammal, although ranges are frequently restricted. There are Bengal tigers Panthera( tigris tigris), although these are not only found in savannah habitats. Other mammals include the Indian rhinoceros (Rhinoceros unicornis), Asian elephant (Elephas maximus indicus), wild water buffalo (Bubalus arnee), Manipur brow antler deer (Cervus eldii eldii), Reeve’s muntjac or barking deer (Muntiacus reevesi), sambar (Cervus unicolor) and chital or spotted deer (Axis axis). Interestingly, the water buffalo, sambar and chital have been introduced to Australia. Biomass data for these Indian species is unfortunately lacking, but it is known that some areas of Indian grass savannah can support 3.5 cattle per hectare (Yadava 1991).
A comparison of the major savannahs The savannahs of South America, Africa and Australia are produced by simi- lar climatic conditions of temperature and rainfall distribution. They are all 8 THE BIOLOGY OF AFRICAN SAVANNAHS
grasslands with trees. However, the savannahs of Africa could be portrayed as vast fertile grasslands with Acacia trees and large mammals, the Austral- ian savannahs as eucalypt open woodlands with marsupials and the South American savannahs as attenuated rainforests with large rodent herbivores. Of course this is a simplistic picture and we shall see in the next section, on Africa, that savannah structure within a continent, or even within a local area, can be quite different. Nonetheless, a rather generalised view of some important differences between the savannahs of the three continents is sum- marised in Table 1.2. The fauna and flora of savannahs on different continents share very few of their species. The introduction of water buffalo, sambar and chital into Aus- tralia and the invasion of South America and Australia by African grasses is a recent event, for which humans are responsible. In fact savannahs from different continents usually show more similarities with adjacent, different, biomes than with each other. For example, the flora of the Brazilian cerrado shows more affinities with the flora of Amazonia than with savannahs in Africa and Australia. Savannahs on different continents often ‘look simi- lar’, but the individual species are quite different. The grasses are different species, the trees are different species and the herbivores and carnivores are different species. The rodent capybara in the llanos, the ungulate wil- debeest in the Serengeti and the marsupial kangaroo in Northern Australia are all large, herbivorous, mammals but they are taxonomically quite dif- ferent. The similarities are due to the fact that the biotas of savannahs, in different areas of the world, are the product of convergent evolution. Dif- ferent species function in similar ways when they have to deal with similar environments.
Table 1.2 A generalized comparison of the major ecological features of savannahs on three continents
Ecological feature Africa South America Australia
Primary production High Intermediate Low Nutrient acquisition by plants Easy Intermediate Difficult Decomposition rate of litter Rapid Intermediate Slow Plant dispersal by Large vertebrates Smaller vertebrates Invertebrates Fire tolerance of vegetation Medium Low High Type of pollination Specialized Very specialized Unspecialized Main herbivores Large vertebrates Smaller vertebrates Invertebrates Main predation by Large vertebrates Medium vertebrates Small vertebrates Mutualism Uncommon Common Common Animal migration Major Intermediate Minor Animal productivity High Medium Low Competition Widespread Patchy Patchy SAVANNAHS 9
African savannahs
Climatic patterns As we saw in Figure 1.1, temperature and rainfall are major determinants of savannahs worldwide. The same is true in Africa, although in most of tropical Africa temperature does not have the same limiting effect on plant growth as it does, for example, in temperate Europe or North America. How- ever, the average air temperature falls by about 0.6 °C per 100 m increase in altitude and therefore African montane areas have a temperature regime not suitable for savannah vegetation to develop. Rainfall is very variable across the continent, and its amount largely determines the location of the three major African biomes: tropical forest, savannah and desert, and the different types of African savannah (Figure 1.5). It is therefore important to under- stand the climate patterns of Africa, particularly for rainfall. It is not just the amount of rain falling in an area that determines the vegetation, but also its seasonal distribution throughout the year. Whether there is one or two rainy seasons, and the length and severity of the dry season are also important to the structure of the vegetation, particularly the development and establish- ment of trees. During the course of a year, as the Earth orbits the sun with its axis tilt- ed relative to the plane of its orbit, the sun changes its position in the sky and produces the changing seasons. The limits of this annual movement of the sun occur in June, when the sun is overhead at the Tropic of Cancer (23.45°N) and December when the sun is overhead at the Tropic of Cap- ricorn (23.45°S). During March (spring or vernal equinox) and September (autumnal equinox) the sun is overhead at the Equator. In effect, the sun’s position moves north and south across the Equator twice a year. A conse- quence of this seasonal change in the sun’s position is a change in the pattern
40
20 Height (m)
0 0 200 500 700 1200 1500 Rainfall (mm) Desert Subdesert scrub Grass tree Broad-leaved deciduous Forest with low bushy savannah woodland with tall grass savannah acacias and few shrubs mosaic
Figure 1.5 Idealized relationship between rainfall and African savannah vegetation. 10 THE BIOLOGY OF AFRICAN SAVANNAHS
of air movement over the continent. To illustrate this, a simple climate model that influences Africa is shown in Figure 1.6. The atmosphere above Africa can be seen as two circulating air systems, ris- ing in the centre, and moving outwards towards the poles. During March, air is heated over the Equator (Figure 1.6a) and therefore rises, causing a low pressure area called the Intertropical Convergence Zone (ITCZ). The ascending air is cooled by expansion resulting from reduced pressure, caus- ing saturation, condensation, cloud formation and rain. This drier air moves towards the poles, cools, and descends, forming a subtropical high pressure region. This descending ‘dry’ air is warmed by compression, reducing its rela- tive humidity even further. This results in two arid belts, the Sahara desert in the north and the Kalahari and Namib deserts in the south. Winds from the northern and southern high pressure zones blow (converge) towards the low pressure ITCZ. These winds always blow from the east because of the direc- tion of the Earth’s spin. Following the sun’s movement by about one month, this system of circulating air, with its attendant low pressure, wet, centre (ITCZ) and high pressure, arid, margins, moves north and by June–August
(a) North pole (b) North pole
Tropic of cancer
Spain Med Tropic of cancer High Equator SpainMed Equator ITCZ ITCZ Low Tropic of Low Capricorn High Tropic of Capricorn High Low
South pole South pole
(c) North pole
Med Spain High
Tropic of cancer ITCZ Low Equator High Tropic of Capricorn
South pole
Figure 1.6 Simple climate model showing how the intertropical convergence zone (ITCZ) moves with the seasons. (a) Vernal and autumnal equinox, rain near the Equator to about 4°N–S. (b) June–August, sun over Tropic of Cancer, rain in northern tropics, dry in southern tropics. (c) November–February, sun over Tropic of Capricorn, rain in southern tropics. ←, winds converging at ITCZ. SAVANNAHS 11 is centred over the Tropic of Cancer (Figure 1.6b); it then moves south again and by September is centred over the Equator once more (Figure 1.6a). By November–February it has moved further south and is centred over the Tropic of Capricorn (Figure 1.6c), eventually returning north to be centred over the Equator again by March. In the days of sailing ships the winds that converge on the ITCZ were called the ‘trade winds’, a term that comes not from any reference to commerce but from the expression ‘to blow trade’, meaning to blow regularly. Under the ITCZ there is no wind, an area known as the ‘doldrums’. Because of these age-old wind patterns, ancient mariners leaving Europe to sail round the African Cape had to sail across the Atlantic on the northeast trade winds, down the South American coast and then back across the Atlantic on the southeast trades. As already noted, one consequence of this circulation of air above Africa is that there is a central ‘wet’ zone around the equator, and two ‘arid’ zones to the north and south. However because the circulation of air moves north and south during the year there is, in general, a gradation from wet, to moist, to dry, to arid as we move away from the Equator. However, the ITCZ does not move north and south, over Africa, as a straight-line zone, running east west. It bends (Figure 1.7). This bending of the ITCZ distorts the moist-arid, north-south, gradient so that East Africa is relatively dry. This is because in January, when the ITCZ is ‘south’, the northeast trade winds, blowing across east Africa, come across relatively dry land areas. At the same time the ITCZ is still over parts of cen- tral and west Africa. In July, when the ITCZ is ‘north’ the southeast trade winds come across the wet Indian Ocean. This creates the pattern of rainfall seen in Figure 1.8 and the consequent pattern of biomes seen in Figure 1.9. In the area with most rainfall, around the equator, there is tropical forest. This grades into forest savannah mosaic, woodland savannah, tree and bush savannah, grass and shrub savannah, and finally desert (Figure 1.5). How- ever in the east, the dry savannah extends down from Sudan, through Kenya, into northern Tanzania. Also, because the ITCZ passes over the central equatorial area twice, there are two rainy seasons here (bimodal rains), one from late March to mid-June and one from late October to mid-December. These short rainy seasons are separated by short dry seasons. As you move away from the equator one of the rainy seasons tends to be longer and more reliable (the ‘long rains’), and the other shorter and less reliable (the ‘short rains’). As you go further north, or south, the rain pattern becomes unimod- al. In northern areas rain falls in one long wet season from April to October, with a long dry season from November to March. In the drier, north-eastern parts, the rainy season tends to be shorter, and unreliable, and the interven- ing period extremely dry. In southern areas the unimodal rainfall regime is the opposite of that in the north with rain falling mainly from November to April. However, these latitudinal rainfall patterns can be moderated by close- ness to oceans and large water bodies such as Lake Victoria. There are also 12 THE BIOLOGY OF AFRICAN SAVANNAHS
ITCZ
ITCZ
January July
Figure 1.7 Seasonal movement of the ITCZ and its associated trade winds (indicated by the arrows).
Turkana
Meru
Kainji
Serengeti
Etosha Above 3,200 1,600–3,200 Luangwa 800–1,600 400–800 200–400 Kruger 100–200 Below 100
Rainfall (mm)
Figure 1.8 Average rainfall in Africa with seven climograms. Both temperature (the rather level line) and rainfall are plotted on the climograms. The ordinate shows both temperature and rainfall, with one division = 20 mm of rain and 10 °C. The abscissa shows months from January to December. A relative dry period occurs when the rainfall curve falls below the temperature curve (stippled area) and a relative humid period occurs when the rainfall curve is above the temperature curve (dark grey area) (map from Kingdon 1989). SAVANNAHS 13
Desert
Desert
Rainforest Equator
Grass and shrub savannah
Tree and shrub savannah
Woodland savannah
Forest–savannah mosaic Desert
Figure 1.9 The principal African savannahs. The black areas are montane.
large rainfall variations due to the modifying influences of montane areas. For example, the Laikipia plateau and Amboseli, in East Africa, receive lower rainfall, being in the shadow of Mt Kenya and Mt Kilimanjaro, respectively. Even the ‘El Niño–Southern Oscillation’, (variations in the temperature of the surface of the tropical eastern Pacific Ocean) is thought to affect the tim- ing of droughts and floods in the Mara-Serengeti system (Ogutu et al. 2007). Figure 1.8 shows a selection of climate diagrams, or climograms, for different savannah sites in Africa. These climograms are superimposed upon a map of Africa showing mean annual rainfall. They are a convenient way to allow an immediate visual assessment of climates within savannah regions. Of course, the amount of water available to a plant is a product not only of the amount of rainfall but how that rainfall is converted into available soil moisture. In practice, the idealised relationship between vegetation and total rainfall, seen in Figure 1.5, is affected by the seasonality of that rainfall. An area with two short rainy seasons and two short dry seasons may pro- duce woodland savannah, while another area with the same amount of rain, but restricted to one short wet season followed by a long dry season, might have no trees, only low Acacia scrub. Additionally, the long dry season might 14 THE BIOLOGY OF AFRICAN SAVANNAHS
make the second area more prone to fires, that would also reduce tree cover. The influence of total annual rainfall can also be modified by topography, drainage and soil type. For example, sandy or free-draining loamy soils read- ily absorb rainfall, whereas the surface of heavy clay soils quickly becomes sealed and further rain simply runs off and is lost to the vegetation.
Vegetation patterns There have been many attempts to classify the vegetation, or biotic, zones of Africa (Happold and Lock 2013). At one extreme, some ‘safari guides’ talk only about tropical forest, savannah, desert, montane and perhaps ‘Mediterranean’ vegetation. The latter is a type of dwarf shrubland found on the extreme north, Mediterranean coast (called maquis) and southern Cape (called fynbos). At the other extreme we have the detailed UNESCO/ AETFAT/UNSO Vegetation Map of Africa published in 1983 (White 1983) after more than 15 years of collaboration between UNESCO and AETFAT (Association pour l’Etude Taxonomique de la Flore de l’Afrique Tropicale). It was based on physiognomy and floristic composition, but not climate, and it included a total of 80 vegetation types and mosaics. Frank White (1983) recognised 17 major vegetation types, based on the cover and height of the vegetation, as well as the species involved (see Happold and Lock (2013) for a very nice coloured map of White’s vegetation zones). For example, forests were defined as all vegetation comprising a continuous stand of trees at least 10 m tall with interlocking crowns. Woodland was defined as open stands of trees at least 8 m tall, with a canopy cover of 40% or more, and with a ground layer usually dominated by grasses and other herbs. This physiog- nomic approach to vegetation classification (looking at form or appearance) has largely been replaced by the phytosociological method which looks at species associations. However, in the context of African it is still frequently used because the ‘physiognomy’ of a plant is a useful, indirect, measure of the environment, particularly temperature, rainfall and soil. We will there- fore adopt a simplified version of White’s 1983 classification, concentrating on savannahs, but making mention of species associations within the broad physiognomic classification. The map of ‘the biotic zones of Africa’ in Hap- pold and Lock (2013), is a similar attempt at combining the detailed zones of White. For example their ‘Mediterranean Coastal Biotic Zone’ is a combina- tion of White’s ‘Mediterranean sclerophyllous forest and sub-Mediterranean semi-desert grassland and shrubland’. The MODIS IGBP Land Cover Classi- fication detailed by Hill, Román and Schaaf (2011) is a similar simplification and Kingdon’s (1989) map of African vegetation is also similiar, although he uses different names to us. His ‘drier savanna types, bush and thicket’ corre- spond to our ‘grassland and shrub plus part of our tree and shrub savannah’. His ‘woodlands’ is similar to our ‘woodland savannah’ and his ‘moist savanna types’ correspond to our ‘forest–savannah mosaic’. SAVANNAHS 15
In Africa, different types of savannahs make up approximately 50% of the land area (Figure 1.9). Of course the sharp demarcation between the veg- etation zones shown on such maps of Africa is artificial, since the environ- mental conditions that produce the biomes (e.g. rainfall) follow gradients (Figure 1.5). This gradation is particularly true of the ‘grass and shrub’ (sometimes called tree and shrub steppes) and ‘tree and shrub’ savannahs. Local conditions can also blur the edges of these biomes. For example, rain forest can penetrate deeply into the savannah biome, as gallery forest along river banks. A range of savannah profiles in shown in Figure 1.10 and some photographs are shown in Figure 1.11.
Grass and shrub savannah The northern border of this type of savannah is called the Sahel and the pre- dominant tree genus is Acacia. It stretches across Africa from northern Sen- egal and Mauritania on the Atlantic coast, to Sudan on the Red Sea. The word ‘sahel’ means ‘shore’ in Arabic and is a reference to this region being a transition zone between the more wooded savannahs of the south and the Sahara Desert to the north. This arid, ‘grass and shrub’ savannah continues
(a) 15 m
(b) 12 m
(c) 24 m
Figure 1.10 Savannah profiles: (a) Brachystegia, Terminalia woodland savannah; (b) Acacia tree and shrub savannah; (c) Combretum, Acacia, Borossus tree and shrub savannah (modified from Kingdon 1997). 16 THE BIOLOGY OF AFRICAN SAVANNAHS
(a) (b)
(c) (d)
Figure 1.11 African savannahs: (a) Laikipia, central Kenya; (b) Laikipia, central Kenya; (c) Amboseli, southern Kenya; (d) Serengeti, northern Tanzania.
into the Acacia–Commiphora savannah of the Horn of Africa, to the east of the Ethiopian highlands, and down into East Africa as the Somali–Masai dry savannah. In the northern areas, typical woody species include Acacia tortilis, A. laeta, Commiphora africana, Balanites aegyptiaca and Boscia senegalensis. Grass cover is continuous, with annual species such as Cenchrus biflorus, Schoe- nefeldia gracilis, and Aristida stipoides. In some areas Acacia and Commi- phora species are joined by Euphorbia and Aloe species, as well as grasses such as Dactyloctenium aegyptium and Panicum turgidium. Two species of tree in this region are characterised by aromatic resins used in biblical times for incense and perfumes. Frankincense comes from the resin of Boswellia carteri, found in Somalia, and myrrh is extracted from Commiphora abys- sinica and other species found in Ethiopia. In the southern part of this region predominant plants again include species of Acacia and Commiphora, along with Crotalaria and the grasses Themeda triandra, Setaria incrassata, Pani- cum coloratum, Aristida adscensionis, Andropogon species and Eragrostis species. In northern Tanzania, the region is bisected by patches of grassland on volcanic soil (southern Serengeti) and patches of montane forest. The vol- canic grassland has no trees because of the nature of the soil but is none the less an integral part of the Serengeti savannah ecosystem. The Sahel region is not especially rich in mammal species although it does possess at least four species of endemic gerbil (Gerbillus bottai, G. muricu- lus, G. nancillus and G. stigmonyx). The scimitar oryx Oryx( dammah)(now SAVANNAHS 17 regarded as extinct in the wild), dama gazelle (Nanger dama), dorcas gazelle (Gazella dorcas) and red-fronted gazelle (Eudorcas rufifrons), all now rare, were formerly abundant and widespread (East 1999). Predators such as wild dog (Lycaon pictus), cheetah (Acinonyx jubatus) and lion (Panthera leo) were once common, but have now been exterminated over most of the region. The Horn of Africa has a number of unique antelopes, such as the dibatag (Ammodorcas clarkei), beira (Dorcatragus megalotis), and Speke’s gazelle (G.spekei). The African wild ass (Equus africanus), desert warthog (Phaco- choerus aethiopicus) and Ethiopian wolf (Canis simensis) are also unique to this region. Several species are found in reduced numbers in the Horn and adjacent regions, but are more abundant further south. These include Grevy’s zebra (Equus grevyi), beisa oryx (Oryx beisa beisa), gerenuk (Litocranius walleri), greater kudu (Tragelaphus strepsiceros), lesser kudu (T. imberbis), elephant (Loxodonta africana) and African buffalo (Syncerus caffer). Lion, leopard (Panthera pardus), cheetah, striped hyaena (Hyaena hyaena) and spotted hyaena (Crocuta crocuta) are the main large carnivores in this region. The endangered wild dog is found in Ethiopia, in Mago and Omo National Parks (Woodroffe et al. 1997). The central Laikipia region of Kenya, between Mount Kenya and the Rift Valley, probably hosts the only intact savannah mammal community outside a Kenyan national park. At least 75 mammal species and 400 bird species can be found in the region. Elephant, eland (Tau- rotragus oryx), common zebra (Equus quagga), Grevy’s zebra, beisa oryx, and reticulated giraffe Giraffa( camelopardalis reticulata) are seasonally abun- dant, migrating long distances across the region, depending on rainfall and forage availability. Defassa waterbuck (Kobus ellipsiprymnus defassa), impala (Aepyceros melampus), Grant’s gazelle (Nanger granti), the hybrid Jackson’s hartebeest (Alcelaphus buselaphus), gerenuk, lesser kudu and bushbuck (Tragelaphus scriptus) are resident. Lion, leopard, cheetah, spotted hyaena, black-backed jackal (Canis mesomeles) and bat-eared fox (Otocyon megalo- tis) are regularly seen, and striped hyaena, aardwolf (Proteles cristatus) and wild dog are present. Of course it is the southern Kenya–northern Tanzania part of this region that is most well-known for its outstanding concentra- tions of large mammals. The Serengeti–Mara migration of approximately 1.3 million wildebeest (Connochaetes taurinus), 200,000 common zebra, and 400,000 Thomson’s gazelle (Eudorcas thomsoni) is the most spectacular mass movement of terrestrial mammals anywhere in the world. Not surprisingly the area also supports one of the highest concentrations of large predators, with approximately 7,500 spotted hyaena and 2,800 lion, along with leop- ard and cheetah. Both Tarangire and Serengeti National Parks have approxi- mately 350 to 400 recorded bird species. In southern Africa, the equivalent ‘grass and shrub’ savannah is an arid tran- sition zone between the northern mopane savannah and the southern desert. Vegetation varies from a dense, short, shrub savannah (bushveld) to an open tree savannah and reflects the diverse topography, soil and microclimate. In 18 THE BIOLOGY OF AFRICAN SAVANNAHS
the north of the area is Euphorbia guerichiana, a shrub or small tree with con- spicuous, shiny, brownish-yellow, papery bark growing to a height of 5 m. Also common are Cyphostemma species with succulent stems, Adenolobus species, the quiver tree (Aloe dichotoma), and Moringa ovalifolia. Two spe- cies of Acacia are confined to this vegetation type; these are the Brandberg acacia (Acacia montis-ustii) and A. robynsiana. Acacia senegal and A. tortilis are found along ephemeral rivers. To the south, the vegetation becomes more open and is dominated by karoo shrubs (Rhigozum trichotomum is charac- teristic) and grasses. Parkinsonia africana, Acacia nebrownii, Boscia foetida, B. albitrunca, and Catophractes alexandri as well as smaller karoo bushes such as Pentzia species and Eriocephalus species are also typical. Tufted grasses, mainly Stipagrostis species, are found scattered between the woody plants. On rocky ridges, the conspicuous quiver tree becomes very abundant. Endemic and near-endemic mammals are mainly bats, rodents, and small carnivores. The only endemic large mammal is the mountain zebra (Equus zebra), which is found to the west of this region, and is the only large mam- mal endemic to Namibia. This western edge is also well-known for its desert- dwelling populations of elephant and black rhinoceros (Diceros bicornis). The black rhino population in this area is one of the few unfenced populations of black rhinos in the world, and it is estimated to number more than 100 individuals. Other large mammals found within this southern arid region are greater kudu, springbok (Antidorcas marsupialis), gemsbok (Oryx gazella gazella), Kirk’s dik-dik (Madoqua kirkii), and black-faced impala (Aepyceros melampus petersi). Predators include lion, leopard, cheetah, bat-eared fox, and Cape fox (Vulpes chama). Protected areas within the ‘grass and shrub’ savannah region include many of the famous East African safari parks. These include Samburu National Reserve (165 km2) in central Kenya, Nairobi National Park (117 km2) in southern Kenya, Serengeti National Park (14,763 km2) in northern Tanzania, and the adjacent Masai Mara Game Reserve (1,510 km2) in southern Kenya, Ruaha National Park (13,000 km2) in central Tanzania, Amboseli National Park (392 km2) and Tsavo (East and West) National Parks (11,747 km2 and 9,065 km2) in southern Kenya, and Tarangire National Park (2,600 km2) and Mkomazi Game Reserve (1,000 km2) in northern Tanzania.
Tree and shrub savannah Like the ‘grass and shrub’ savannahs, the ‘tree and shrub’ savannahs form two separated blocks of vegetation, lying north and south of the rainforest and miombo woodland savannahs of central Africa. In the north the climate is tropical and strongly seasonal and the vegetation is composed mainly of Combretum and Terminalia shrub and tree species and tall elephant grass (Pennisetum purpureum). This region lies south of the SAVANNAHS 19
Sahel and stretches from the west coast of Africa to Sudan, and also extends south into northwestern Uganda. Typical trees include Anogeissus leiocarpus, Boswellia papyrifera, Balanites aegyptiaca, Lannea schimperi, Stereospermum kunthianum, Kigelia aethi- opica, Acacia seyal, Commiphora africana, Prosopis africana, Tamarindus indica, Ziziphus mucronata and, of course, species of Combretum and Ter- minalia. Dominant grasses include tall species of Hyparrhenia, Cymbopogon, Echinochloa, Sorghum, and Pennisetum. Mammals include elephant, African buffalo, oribi (Ourebia ourebi), western hartebeest (Alcelaphus buselaphus major), wild dog, cheetah, leopard and lion. The giant eland (Taurotragus derbianus) still survives in parts of the eastern section of this savannah. The roan antelope (Hippotragus equinus) is widespread throughout this northern region, but not in large numbers. Remnant populations of the west African giraffe (Giraffa camelopardalis peralta) numbering about 250 are still found in Niger. To the south of the central African block of miombo ‘woodland savannah’, described in the next section, are ‘tree and shrub’ savannahs characterised by the dominance of the mopane tree Colophospermum mopane. Mopane is a single-stemmed tree or shrub with distinctive, butterfly-shaped leaves. Two regions of mopane can be identified, the southeast ‘Zambezian’ region and the southwest ‘Angolan’ region. The Zambezian region extends into South Africa, Mozambique, Botswana, Zambia, Zimbabwe, Swaziland, Namibia and Malawi. Here, the mopane tree is frequently the sole canopy species but can be associated with other prominent trees and shrubs. These include Kirkia acuminata, African blackwood (Dalbergia melanoxylon), baobab (Adansonia digitata), Combretum apiculatum, C. imberbe, Acacia nigrescens, Cissus cornifolia, and Commiphora species. These Zambezian mopane com- munities show considerable variation in height and density. Trees in dense woodland or in more open savannah woodland may reach heights of 10 m to 15 m on deep alluvial soils, and even attain 25 m in the so-called ‘cathedral mopane’ of Zambia. In contrast, mopane tends to be stunted and shrubby (1 to 3 m) where it occurs on impermeable alkaline soils. These two struc- tural forms often occur together in a mosaic depending on microclimatic factors and soil conditions (White 1983, Smith 1998). The ground layer can also vary markedly. For example, dense grass swards are found beneath gaps in the mopane canopy on favourable soils, while grasses are almost com- pletely absent in shrubby mopane communities on heavy, impermeable alkaline clays. Typical grasses include Aristida species, Eragrostis species, Digitaria eriantha, Brachiaria deflexa, Echinochloa colona, Cenchrus ciliaris, Enneapogon cenchroides, Pogonarthria squarrosa, Schmidtia pappophoroides, Stipagrostis uniplumis, and Urochloa species. The Angolan mopane woodland- savannah is located in Namibia and Ango- la, completely surrounding the large Etosha salt pan in northern Namibia. 20 THE BIOLOGY OF AFRICAN SAVANNAHS
Again, mopane trees dominate the vegetation. They often form a dense, single-species stand under which grass is virtually absent. Hence fire damage is minimal, even though the mopane tree itself is resinous and flammable. However, if the canopy is opened up (for example by browsing elephants), grasses invade and the fire frequency and intensity is increased. Browsing elephants frequently push down mopane trees, which then regrow into low (0.3 to 1.6 m), multi-stemmed, shrubs. The ‘typical’ mopane woodland is then converted into a tall grassland, with the grass frequently as tall as the mopane coppice (White 1983). In Angola, mopane grows over vast areas in a low, thorny bushveld. It is then associated with Acacia kirkii, A. nilotica subalata, A. hebeclada tristis, A. erubescens, Balanites angloensis, Combretum apiculatum, Commiphora species, Dichanthium papillosum, Dichrostachys cinerea, Grewia villosa, Indigofera schimperi, Jatropha campestris, Melanthera marlothiana, Peltophorum africanum, Rhigozum brevispinosum, R. virgatum, Securinega virosa, Spirostachys africana, Terminalia prunoides, T. sericea, Ximenia americana, and X. caffra. Because the vegetation in the mopane region is more nutritive than in miom- bo woodland-savannah, coupled with the extensive and well maintained system of protected areas, the mopane region supports large concentrations of ungulates. These includes elephant, black rhinoceros, white rhinoceros (Ceratotherium simum), hippopotamus (Hippopotamus amphibius), Afri- can buffalo, wildebeest, mountain zebra, kudu, nyala (Tragelaphus angasii), gemsbok, eland, greater kudu, roan, steenbok (Raphicerus campestris), dik- dik and the near-endemic black-faced impala. Large predators include lion, leopard, cheetah, spotted hyaena, brown hyaena (Hyaena brunnea), and wild dog. Smaller predators include the black-backed jackal and the bat-eared fox. In addition to biomass differences for many mammals, there are differ- ences in the species assemblages between miombo and mopane. Side-striped jackal (Canis adustus), sable antelope (Hippotragus niger), roan antelope, and Lichtenstein’s hartebeest (Alcelaphus lichtensteinii) are associated with miombo, while black-backed jackal, greater kudu, and impala (Aepyceros melampus) are identified with mopane (Huntley 1978). This region has a diverse avian fauna, with over 375 species recorded.
On the southern borders of the mopane region are Baikiaea woodlands, a mosaic of dry deciduous Baikiaea plurijuga dominated forest, thicket and grassland. It forms a belt along the Angola–Namibia border, and extends southwest into Botswana, Zimbabwe and the northern province of South Africa. In well-developed Baikiaea communities, species of Brachystegia, Julbernardia and Colophospermum mopane (species typical of miombo and mopane woodlands) are totally absent. Baikiaea plurijuga forms a fairly dense, dry, semi-deciduous forest with trees up to 20 m in height. There is a dense and shrubby lower story of Combretum engleri, Pteleopsis anisop- tera, Pterocarpus antunesii, Guibourtea coleosperma, Dialium engleranum, SAVANNAHS 21
Strychnos species, Parinari curatellifolia, Ochna pulchra, Baphia massaiensis obovata, Diplorhynchus condylocarpon, Terminalia brachystemma, Burkea africana, Copaifera baumiana and Bauhinia petersiana serpae. This type of savannah has a diverse bird life, with 468 species recorded. Protected areas within the ‘tree and shrub’ savannah region include Kru- ger National Park (19,633 km2) in northern South Africa with its wild dogs, Chobe National Park (10,570 km2) in Botswana with its elephants (estimat- ed 50,000), and Etosha (Great White Place) National Park (22,912 km2) in northern Namibia with its gigantic ‘white’ salt pan. This salt pan covers an area of 4,731 km2, and is believed to have once been a great inland lake fed by a large river (probably the Kunene), which over time changed its route, causing the lake to dry out and form a salt desert. Etosha has over 340 bird species recorded and the park is particularly rich in raptors, with 46 species recorded, including all the vultures found in Namibia.
Woodland savannah There are two areas of ‘woodland savannah’, a huge area called miombo, in central/south Africa, and a reduced area called doka in the north. Miombo covers an estimated 3 million km2 of Zimbabwe, Zambia, Mozam- bique, Angola, Malawi, Katanga province of the Democratic Republic of the Congo, and southern Tanzania. It is the largest vegetation unit in the Zambe- zian centre of endemism, or Zambezian phytochorian. A phytochorian is a plant-geographic area, recognised by its species composition rather than its structure or physiognomy. Miombo woodland savannah takes its name from the miombo (Muuyombo) tree (Brachystegia boehmii), the dominant tree genus over this entire area. In the centre, the miombo area experiences a sea- sonal tropical climate, with most of its rainfall is concentrated during the hot summer months of November to March/April, followed by a pronounced winter drought, which can last up to 7 months in some areas. Typically, mature miombo trees are 15 to 20 m tall, with a shrub and grass understorey. In Angola, the canopy height is lower (5 to10 m), with little or no shrub layer. Dominant tree species include Brachystegia spiciformis, B. boehmii, B. allenii, B. glaberrima, B. taxifolia, B. utilis, Marquesia mac- roura, Julbernardia globiflora, J. paniculata and Copaifera baumiana. Brachy- stegia floribunda, B. gossweilerii, B. wangermeeana, B. longifolia, B. bakerana, Guibourtea coleosperma and Isoberlinia angolensis are locally dominant. In the southern part of the miombo region, other common tree species include Uapaca kirkiana, Monotes glaber, Faurea saligna, F. speciosa, Combretum molle, Albizia antunesiana, Strychnos spinosa, S. cocculoides, Flacourtia indi- ca, and Vangueria infausta. Most of the miombo tree and shrub species shed their leaves in the late dry season, and the miombo woodland is bare for 22 THE BIOLOGY OF AFRICAN SAVANNAHS
two or three months. A few weeks before the rainy season starts, the trees produce their new, predominantly bright reddish foliage. Fire is an impor- tant ecological factor in miombo woodland. The strong seasonality in rain- fall leaves the vegetation dry for several months of the year, and fires, either natural or man-made, can be frequent. Miombo does not support large mammals in high densities, although due to the vast size of the region its overall importance for mammal species is quite high. The low, large mammal, density is probably caused by the poor soils, which generally support vegetation of low nutritional value. Condi- tions are made worse by the harsh dry season and long droughts. Of course many of the ‘miombo large mammals’ are found in other savannah regions and include elephant, black rhinoceros, and African buffalo. These are able to survive on poor quality forage by consuming it in large quantities. Specialised grazers are also common. They selectively feed on grass, and include sable antelope, Lichtenstein’s hartebeest, and southern reedbuck (Redunca arundinum), all species largely restricted to the miombo belt, as well as the more widespread roan antelope. Browsers such as eland and mixed feeders such as greater kudu are also present. Many species make use of the wooded margins or open areas of the numerous grassy flood plains scattered through the miombo region. These include lechwe Kobus( leche), puku (K. vardonii), tsessebe (Damaliscus lunatus), oribi, wildebeest (Con- nochaetes taurinus) and sitatunga (Tragelaphus spekii). Common waterbuck (K. ellipsiprymnus ellipsiprymnus) and bushbuck are mostly found in more wooded areas close to permanent water. Other large ungulates include com- mon zebra and the restricted, but abundant, Thornicroft’s giraffe (Giraffa camelopardalis thornicrofti). Hippopotamus are relatively common near water. Due to annual droughts and frequent fires, many species are sea- sonally dependent on non-miombo vegetation, within or adjacent to the region, to provide food, water, or shelter. For example, sable antelope are largely confined to the miombo belt but move onto more open grassy areas during the dry season (Kingdon 1997). There are many carnivores, although most are not confined to miombo woodland savannah. These include lion, leopard, cheetah, spotted hyaena, striped hyaena, African wild dog, side-striped jackal, wild cat (Felis sylves- tris), serval (Leptailurus serval), caracal (Caracal caracal), miombo genet (Genetta angolensis), Selous’s mongoose (Paracynictis selousi), and bushy- tailed mongoose (Bdeogale crassicauda). Two large insectivores, distributed over much of the African savannah, Temminck’s ground pangolin (Smutsia temminckii) and the aardvark (Orycteropus afer), feed on the numerous ants and termites found in the miombo. The miombo area supports a number of primate species, mostly on its north- western borders, next to areas of ‘forest–savannah mosaic (see next sec- tion)’ and rainforest, and where it grades into submontane forest habitats SAVANNAHS 23
in Uganda. The Gombe Stream National Park (52 km2) in western Tanzania, although largely covered by evergreen forests, includes substantial miombo habitat on its lower slopes. The area is well-known as the site of Jane Goodall’s long-term study of the chimpanzee (Pan troglodytes) (Goodall 1988). This area also supports several species of red colobus (Procolobus species), black and white colobus (Colobus angolensis), blue monkey (Cercopithecus mitis) and red-tailed monkey (C. ascanius). More widespread primates that are typ- ical of many African savannahs are vervet monkey (Chlorocebus pygerythrus) and savannah baboons (Papio species). Miombo bird life is rich in species. However, most are not restricted to this type of savannah. Birds breeding in miombo woodlands generally have relatively short breeding seasons and start nesting before or during the early rains. As well as the Gombe Stream National Park, protected areas within the miombo woodland savannah region include South Luangwa National Park (9,050 km2) in Zambia, with its Thornicroft’s giraffes and Selous Game Reserve (54,000 km2) in central Tanzania, with a sizeable wild dog population. Like miombo and mopane savannahs, the northern doka savannah takes its name from a dominant tree, Isoberlinia doka. Here, where human pop- ulation remains sparse, patches of dense dry forest remain, dominated by Isoberlinia doka with Afzelia africana, Burkea africana, Anogneissus leiocar- pus, Terminalia species and Borassus aethiopum. In many ways these north- ern doka woodland savannahs are similar to the forest–savannah mosaics described below.
Forest–savannah mosaic Encircling the tropical rainforest of the Congo basin, the ‘forest–savannah mosaic’ forms the edge of the ‘true’ savannah. This encircling, sometimes narrow, transition zone can be conveniently divided into three regions. In the north, and to the west of the Cameroon highlands, is the Guinean forest–savannah mosaic of West Africa running through Guinea, Ivory Coast, Ghana, Togo, Benin and Nigeria. The interlacing forest, savannah and grass- land habitats are highly dynamic, and the proportion of forest versus other habitat components has varied greatly over time. The protected areas in this region are underfunded and cover only 2% of the area. In the north, and to the east of the Cameroon highlands, is the northern Congolian forest–savannah mosaic. This narrow transition zone marks an abrupt habitat discontinu- ity between the extensive southern rainforests and the dryer savannahs to the north and east. It extends east through the Central African Repub- lic, north-eastern Democratic Republic of Congo and into south-western Sudan and a sliver of north-western Uganda. To the south and west of the Congo basin is the Zambezian forest–savannah mosaic. Because all these mosaics are edge, or ecotonal, regions they frequently have high species richness. 24 THE BIOLOGY OF AFRICAN SAVANNAHS
The vegetation of this region is an interesting mosaic of rainforest and savan- nah elements. Gallery forests, extending along rivers, interdigitate with drier, semi-evergreen, rainforest, which in turn grades into grassland and wooded grassland. Widespread gallery species include Berlinia grandiflora, Cola lau- rifolia, Cynometra vogelii, Diospyros elliotii, Parinari congensis, and Pterocar- pus santalinoides. Species restricted to the drier forests, and widespread across this region, include Afzelia africana, Aningeria altissima, Chrysophyllum per- pulchrum, Cola gigantea, Combretum collinum, Morus mesozygia, and Khaya grandifoliola. Trees found in the wooded grasslands include Annona senega- lensis, Afzelia africana, Burkea africana, Butyrospermum paradoxum, Daniel- lia oliveri, Hymenocardia acida, Maranthes polyandra, Parinari curatelifolia, Parkia biglobosa, Piliostigma thonningii, Pseudocedrela kotschyi, Pterocarpus erinaceus, Stereospermum kunthianum, Strychnos species, Terminalia species and Vitex species. Common grasses, many growing taller than 2 m, include Andropogon species, Hyparrhenia species and Loudetia species. The southern forest–savannah mosaics are similar in structure to the north- ern mosaics except that the drier southeast boundary merges more gradually into the adjacent miombo woodland savannah already described. Covering a broad area of the southern Democratic Republic of Congo, these south- ern forest–savannah mosaics are a blend of forest, woodland, shrubland and grassland habitats. Some mammals of these forest–savannah mosaics include the forest subspe- cies of elephant (Loxodonta africana cyclotis), giant eland and in the eastern sector, bongo (Tragelaphus eurycerus). Predators are lion, leopard and the Nile crocodile (Crocodylus niloticus), which is present in northern waterways across the region. Protected areas within the forest–savannah mosaic include Queen Elizabeth National Park (1,978 km2) and Murchison Falls National Park (3,840 km2), both in Uganda. Both still have good numbers of Uganda kob (Kobus kob thomasi), and a very wide range of bird life. Murchison Falls National Park forms part of the Murchison Falls Conservation Area (5,308 km2).
Species richness patterns So far we have looked at climatic patterns in Africa and seen how these influence and determine patterns in the structure (physiognomy) of savannah vegetation. But do these patterns of climate and vegetation physiognomy influence species richness and biodiversity? A detailed look at these two topics will be delayed until later chapters, but it is appropriate to end this chapter with a sneak preview. Species richness varies with area (Gaston 1996) and to make meaningful comparisons it is necessary to compare areas of a similar size. For vegetation, Cailleux (1953) considered a reference area of 10,000 km2 to be adequate and this is the data (areal plant richness) shown in Figure 1.12. The floristic SAVANNAHS 25
richness of the African savannahs stands out clearly. The average areal plant richness of African savannahs is about 1,750 species, not much lower than that of rainforests (c.2020 species). The miombo savannahs of East Africa have an even greater (>3000) areal plant richness than the rainforest. Notice also that Figure 1.12 is remarkably similar to Figure 1.8 and Figure 1.9. Rain- fall (Figure 1.8) appears to be a major determinant of savannah physiogno- mic types (Figure 1.9) and floristic richness (Figure 1.12). There is, therefore, a great similarity between the distribution of the areas of comparable floris- tic richness and that of the major physiognomic categories of African veg- etation. A detailed study of bioclimatic variables and plant species diversity was carried out by Gwitira et al. (2014) in Zimbabwe. They used over 54,000 records of georeferenced plant species in this southern African ecosystem to calculate species richness, and investigated the predictive strength of 19 bioclimatic variables, using regression analysis. Their results showed that just two variables, precipitation of the warmest quarter of the year (R2 = 0.92, P < 0.001) and temperature of the warmest month (R2 = 0.67, P < 0.001), significantly explain variation in plant species richness. Sadly, bioclimatic
0–500 500–1000 1000–1500 1500–2000 2000–3000 >3000
Figure 1.12 Spatial variation in the floral richness of Africa, expressed as the number of species per 10,000 km2 (from Menaut 1983; redrawn from Lebrum 1960). 26 THE BIOLOGY OF AFRICAN SAVANNAHS
Ungulates Carnivores
Number of Species <5 6 to 10 11 to 15 16 to 20 21 to 25 26 to 30
Figure 1.13 Spatial variation in the mammal richness of Africa.
modelling, using future climatic change projections, showed a reduction in those areas presently supporting high species richness. Figure 1.13 shows an equivalent pair of maps for two groups of African mammals. These maps have been constructed from the individual species distribution maps in Stuart and Stuart (1997). Each of these species maps was enlarged and the information transferred onto the gridded maps of Africa shown in Figure 1.13. These gridded maps therefore accumulated the num- ber of species present in each square (equivalent to about 46,500 km2). Two maps were produced, one for 94 ungulate species and one for 67 carnivore species. Like the areal vegetation map (Figure 1.12), these mammal maps show high species richness in savannah areas, particularly down the eastern side of Africa, from Kenya to South Africa. Much of this area of high species richness is miombo woodland savannah, but northern areas are more arid grass and shrub savannah. Savannah biomes are therefore extremely rich in both plant and animal spe- cies, and consequently have an extremely rich and fascinating network of species interactions. In Chapter 2 we explore the plants, and in Chapter 3 the animals, in more detail. In the later chapters we then explore the ecological interactions that take place between them. 2 The Vegetation
This chapter examines the biology and ecology of the plants that are found in African savannahs. Of course not all species of plants can be mentioned individually, so we examine two groups of plants—grasses and trees—and describe some of the more important savannah species. These two types of plant dominate African savannah vegetation (Chapter 1), and this vegeta- tion is an expression of the interactions of climate, soils, herbivores, fire and human activities. Climate was examined briefly in Chapter 1, and soils are mentioned in the case study at the end of this chapter. The important interac- tion between fire, elephants and trees is examined in Chapter 5, and human activities are dealt with in Chapter 6. To start this chapter, we look briefly at the effect of rainfall on plant biomass, and the grass–tree mixture. As dis- cussed in Chapter 4, the relationship between rainfall and plant biomass is important for the animal populations that graze and browse on savannah vegetation. In the final section of this chapter we examine vegetation at a ‘local scale’ using the Serengeti–Mara ecosystem, of northern Tanzania and southern Kenya, as an example.
Rainfall, plant biomass and grass–tree mixture
In African savannahs plant biomass is closely associated with the annual amount of rain falling in an area, and this has an important effect on the biomass of herbivores and carnivores (Chapter 4). One African study that demonstrates this relationship clearly is that of Deshmukh (1984), who com- piled published information on above ground herbage production and rain- fall for several eastern and southern African sites. All these studies harvested the herb layer (grasses, forbs and dwarf shrubs), away from tree and shrub canopy effects. The resulting data are shown in Figure 2.1, along with the lin- ear regression line describing the relationship between rainfall and biomass. This significant regression (P <0.001) predicts approximately 800 kg ha−1 herb layer biomass for every 100 mm of rainfall.
The Biology of African Savannahs. Second Edition. Bryan Shorrocks & William Bates. © Bryan Shorrocks & William Bates 2015. Published 2015 by Oxford University Press. 28 THE BIOLOGY OF AFRICAN SAVANNAHS
8000
6000 ) – 1
4000 Peak biomass (kg ha 2000
0 200 400 600 800 Precipitation (mm)
Figure 2.1 The relationship between rainfall and above-ground herb biomass, for sites in East and southern Africa: ●, Marsabit, northern Kenya; Δ, Kaputei, southern Kenya; □, Nairobi National Park, southern Kenya; ▽, Namib desert, western Namibia; ▼, Namibian grasslands, western Namibia; ○, Serengeti National Park, northern Tanzania; ■, Mkomazi Game Reserve, northern Tanzania; ▲, Mweya peninsula, Queen Elizabeth National Park, Uganda. Also shown are the regression line and 95% confidence limits (from Deshmukh 1984).
Several of the studies that produced the data points for Figure 2.1 also calcu- lated regression lines for their local data (Marsabit, Mweya, Namib desert and Serengeti). They all fit within the 95% confidence limits of Figure 2.1 and have very similar slopes. This suggests that although herb layer biomass production may be influenced by edaphic factors that affect soil water-holding capacity and water runoff, the major determinant of the amount of grazing material available to herbivores is rainfall. The other potential modifying influence is that of the grass species present in the herb layer. Not all grass species respond to rain in quite the same way. O’Connor, Haines and Snyman (2001) looked at this same relationship for three conditions of southern African savannahs. They termed these ‘good’ (dominated by the tufted perennial grass,Themeda triandra), ‘medium’ (dominated by the tufted perennial grass, Eragrostis lehmanniana) and ‘poor’ (dominated by the stoloniferous perennial grass, Tragus koeleriodes), reflecting the quality of forage. Ground cover decreased from ‘good’ to ‘poor’, and the medium and poor sites were maintained artificially. Their results are shown in Figure 2.2. Clearly species composition, and perhaps more important- ly ground cover, affect the precise nature of this plant biomass–rainfall relation- ship. However, the overriding influence of the amount of rainfall still remains. TTHE VEGETATION 29
300
250 ) – 2 200
150
100 Phytomass (g m 50
0 300 420 540 660 780 900 1020 Precipitation (mm)
Figure 2.2 Relationship between plant biomass and rainfall for three ‘grass composition’ sites in southern Africa: ●, good; ○, medium; □, poor (from O’Connor et al. 2001).
When we turn to the other major component of savannah vegetation, trees, this type of detailed information, about the effect of rainfall on biomass, is more difficult to find. However, Birkett and Stevens-Woods (2005), working at a savannah site in Laikipia, Kenya found that tree growth is also influenced by the amount of rain. When rainfall was high (109 mm per month), Acacia drepanolobium grew by 4.8 cm per month, but when rainfall was low (45 mm per month) it grew by only 1.4 cm per month. Perhaps more importantly, rain- fall, or mean annual precipitation (MAP), has been shown to limit the amount of tree cover in African savannahs (Sankaran et al. 2005). They compiled data from 854 African savannah sites, for which annual rainfall, fire incidence and tree cover were known (Figure 2.3). Notice that tree cover ranges from 0 to 90% across these sites, and that the ‘maximum’ tree cover increases linearly with rainfall. However, the actual tree cover for any one site can of course be lower than this maximum. For arid and semi-arid savannah sites (MAP 100 80 60 40 Tree cover (% ) 20 0 0 200 400 600 800 1,000 1,200 Rainfall (mm) Figure 2.3 Tree cover as a function of mean annual precipitation (MAP), as related to the fire regimes of 854 African savannah sites. The line representing maximum tree cover was obtained from a regression analysis that used the 99th quantile of y, rather than the mean, as in conventional regression (Cade and Noon 2003). Filled circles represent savannahs with low fire frequencies (average fire return times of >3 years), and open circles represent sites with high fire frequencies (fire return intervals of ≤3 years). Fires typically tend to be more frequent in sites with greater rainfall. For sites receiving <650 mm MAP, the upper bound on tree cover exists despite low fire frequencies (some of the sites had not burned for >50 years) (from Sankaran et al. 2005, Determinants of woody cover in African Savannahs. Nature. Reprinted by permission of Macmillan Publishers Ltd.). rainfall and both tree cover and tree height, in Namibia. This topic is dis- cussed in more detail in Chapter 5, and further comments on the interaction between rainfall, trees and grasses (and fire) are left until that chapter. To see how satellites are being used to study changes in vegetation see Box 3.1. Morphology and life history In African savannahs, both woody and herbaceous plant species have to overcome two major environmental problems: seasonal drought and/or flood and periodic burning. They do this in various ways. However, rather than describe these morphological and life history adaptations individually for each plant species, we examine and discuss these features within a sim- ple classification. The term morphology, or life form, is used to encompass all those traits of external morphology and overall organisation shown by the plant. The termlife history, or phenology, is used to encompass all those traits of sequential development of plant structures during an annual cycle. TTHE VEGETATION 31 Life forms The Danish botanist Christen Raunkier proposed a system of life forms (Raunkier 1934) which, with various later improvements, has been wide- ly employed to compare floras from different biomes, in different parts of the world. Insight into the relationship between life form and climate leads to the classification of biomes, which is largely based on the dominance of plant life forms (Rutherford 1997, Oldeland, Schmiedel and Jürgens 2010). The system classifies plants into a series of categories (Figure 2.4) primarily according to the position of their buds during the unfavourable season of the year (Barbour et al. 1999, Vandvik and Birks 2002). The Raunkier categories most encountered in African savannahs are: • phanerophytes: woody or herbaceous perennials, taller than 20 cm, whose shoots do not die back (trees and tall shrubs) • chamaephytes: woody or herbaceous perennials 10–20 cm tall, whose shoots die back periodically (small shrubs and herbs) • hemicryptophytes: perennial (or biennial) herbaceous plants in which the stems die back to a remnant shoot system that lies on the ground (herba- ceous plants with runners along the ground, and perennial grasses) • geophytes: perennial (or biennial) herbaceous plants in which the stems die back to a remnant shoot system with storage organs (bulbs, corms, rhizomes, tubers) that are embedded in the soil (rhizomatous grasses and bulb-forming herbs) • therophytes: annuals, i.e. plants that die after seed production and com- plete their entire life cycle within one year. (annual grasses). Some authors (e.g. Aubréville 1963) consider this Raunkier system to be of limited use for savannah plants since the unfavourable season was originally Phanerophyte Hemicryptophyte Geophyte Chamaephyte Figure 2.4 The life forms of Raunkier most relevant to savannahs. Therophytes are not shown. 32 THE BIOLOGY OF AFRICAN SAVANNAHS assumed to be a low temperature winter. In most African savannahs low win- ter temperature does not represent a serious limiting factor. Here the main environmental limitations are an extended drought, and periodic burning. None the less, this system of life forms is widely used by African botanists (Menaut 1983) and, as an example, Table 2.1 shows the distribution of the rel- evant Raunkier life forms across some western African savannahs of differ- ing humidity. Notice that the proportion of phanerophytes does not change much along the humidity gradient, although it diminishes slightly in the dryer savannahs. More humid savannahs appear to be especially rich in geo- phytes and hemicryptophytes, forms that are particularly resistant to fierce bush fires fed by an abundant grass layer. Chamaeophytes seem to thrive in mesic savannahs and theophytes are largely dominant in arid savannahs. However, one of the problems with using the Raunkier system in this com- parative way is that different botanists frequently assign the same species to different categories. Also the same species may appear to have individuals that fall into different categories (e.g. phanerophyte vs chamaephyte) depending upon the local conditions of soil, grazing and fire. Another problem, which certainly distorts the ‘community structure’ of savannah plant assemblages, is that botanists have usually compared the percentage of species that fall into the Raunkier categories (Table 2.1). This gives a wrong impression of the ‘importance’ of annual grasses (therophytes) in savannah systems. For exam- ple, using species, the grass layer in one savannah site (Hopkins 1962) gave: 32% hemicryptophytes, 29% geophytes, and 39% therophytes. Using the per- centage of individual plants produced a different result: hemicryptophytes Table 2.1 Percentages of species in the various Raunkier life forms, in nine west African savannahs; there is a moisture gradient from the dry Sahelian savannahs to the more humid Guinean savannahs Phanerophytes Chamaephytes Hemicryptophytes Geophytes Therophytes Guinean savannahs Ivory Coast 28 — 42 6 24 Congo 14 25 26 16 19 Nigeria 30 — 23 21 25 Sudan–Zambezian savannahs Central African 13 24 8 11 40 Republic Rwanda 29 30 12 7 20 DRC 38 44 9 4 5 Sahelian savannahs Southern Mauritania 24 7 6 2 61 Northern Senegal 19 2 2 2 75 Northern Chad 14 2 12 5 67 For the original references see Menaut (1983). TTHE VEGETATION 33 71%, geophytes 14% and therophytes 15%. Not withstanding these issues, Raunkiaer’s life-form concept has been successfully employed by Ruther- ford and Westfall (1994), Irish (1994), and Oldeland, Schmiedel and Jürgens (2010) in their biome classifications in southern Africa. The data of Old- eland et al.confirm the prevalence of phanerophytes and hemicryptophytes for savannah, in terms of relative cover. However, therophytic species also play an important role in the richness and cover of their ‘woodland and the thornbush savannah’ biome. Grazing had a significant effect. At a site with moderate grazing pressure, relative cover of hemicryptophytes was about 35%, whereas at an adjacent site with high grazing pressure, it dropped to almost 1%. Here, we combine many of the Raunkier categories and use a simpler clas- sification of life forms (Sarmiento and Monasterio (1983). This has only three categories: perennials with permanent above ground woody struc- tures (phanerophytes and chamaephytes), perennials whose above ground structures are seasonal (hemicryptophytes and geophytes), and annuals (therophytes). Perennials with above ground woody structures (trees and shrubs) The main architectural types within this group are trees, often looking like shrubs, while palms, woody vines, succulents and so forth are much less frequent. In the Raunkiaer system this group would include mainly phan- erophytes and chamaephytes. One characteristic feature of savannah trees is their relatively modest aerial development. In many savannah communities an overwhelming proportion of the trees have a mean height ranging between 2 and 6 m, although the tallest specimens may reach 12 m. An exception are species such as Brachystegia which occur in what we have called the miombo woodland savannah (Chapter 1), where the taller trees may reach a height of 20 or 25 m. Another obvious exception is the familiar fever tree (Acacia xanthophloea), found along the rivers that penetrate many savannahs, that may reach a height of 30 m. Menaut (1971) sampled the woody populations in the Lamto savannahs of West Africa and found that all trees were lower than 10 m, except the palm Borassus aethiopum which could reach 20 m. Associated with this feature, stem girth rarely exceeded 150 cm, with most individuals in the range 20–40 cm. Another feature of many savannah trees is that they are low-branched and often ramify from the base. This is frequently a response to fire damage, or some other mechanical injury such as grazing. New shoots sprout from the damaged stump, or sucker from the lateral roots. The thickness of the bark of many savannah trees has been interpreted as a protection against repeated bush fires. This might be the case in some species, but the lack of a thick bark does not prevent many other species from surviving in savannahs that are regularly burned. Many savannah trees (e.g. Acacia spp., Balanites aegyptiaca, and Euphorbia spp.) have spines, which may prevent, or 34 THE BIOLOGY OF AFRICAN SAVANNAHS interfere with, browsing by some species of herbivore. Many savannah trees, such as Acacia drepanolobium, certainly produce more spines if they are heav- ily grazed. However, spines do not prevent browsing by some species such as giraffe and black rhinoceros. The leaves of savannah trees are generally moderate or small in size, and except for the palms (e.g. Borassus aethiopum), plants with large leaves are rarely found. Many species (e.g. thorn trees or Aca- cia) have moderate sized compound leaves but they are divided into many, quite small, leaflets. The leaves of many savannah species live for about one year, with leaf fall proceeding the development of new leaves. They therefore appear to be evergreen. However, some savannah trees are deciduous, or semi- deciduous, with the alternation of leaf and leafless conditions corresponding with the wet and dry seasons. This phenology can vary between species even within the same genus, and also between geographical areas (see the three Acacia species in Figure 2.6). Leaf fall is one method of adaptation to drought, particularly for those species with soft, mesomorphic leaves. Another method is the production of xeromorphic leaves that can survive the dry season and have adaptations that reduce water loss through transpiration. These adapta- tions include hard leaves with a thick cuticle and cuticular layers, stomata (the leaf’s breathing pores) placed at the bottom of deep depressions, and small leaves reduced to scales. Some trees have enhanced water storage facilities in their trunk, such as the cactus-like Euphorbia species and the baobab Adan- sonia digitata. A final characteristic of savannah trees is their extensive root system that allows them to exploit the water and mineral resources of a great volume of soil. Many savannah trees are reputed to have a downward growing tap-root, from which a number of lateral, horizontal, roots develop (Hopkins 1962, Sarmiento and Monasterio 1983). However, Walter (1973) has shown this is often not the case. Particularly in arid areas, root systems tend to flatten out in order to provide the best opportunity to absorb water from the upper soil layers after relatively light rain. Perennials with above ground seasonal vegetation In the Raunkiaer system this group would include mainly geophytes and hemicryptophytes. It includes two types: perennial species with woody underground storage organs and perennial species with non-lignified stor- age organs, such as fleshy rhizomes. Dominant among this latter group are the tussock grasses, a very obvious and dominant part of the savannah flora. The stem of grasses, bearing the leaves and flower-head, is called the culm. It is cylindrical and hollow except at the nodes, or joints, which are of solid tis- sue. The hollow sections between the nodes are called the internodes, and the basal nodes may be swollen. The nodes are where the leaves are attached, in two rows on opposite sides of the stem. The upper, expanded, part of the leaf is the blade or lamina, and the basal part, surrounding the stem, is the sheath (Figure 2.5). At the point where the blade forms the sheath there is usually a membranous outgrowth called the ligule. TTHE VEGETATION 35 (a) (b) Blade Ligule Node Blade Sheath Spikelets Figure 2.5 (a) Diagram showing the basic structure of a grass stem and (b) a Themeda inflorescence. The normal method of branching in grasses involves the production of new stems from the axils of older leaves, and is called tillering. If the new stem grows upwards, within the sheath, dense tufts of grass are produced. If the new stem breaks through the sheath near its point of origin, then loose or open tufts result. The thin and fibrous roots are adventitious (meaning they arise from ‘abnormal’ positions; in this case they arise from the lower node or nodes of the stem). In some perennial species (e.g. Chloris gayana) the roots are produced at every node of a surface creeping stem, or stolon. In other perennial grasses (e.g. Pennisetum purpureum) they arise from the nodes of underground creeping stems, or rhizomes. The ‘flower head’, or inflores- cence, is usually at the end of the stem and consists of a much-branched struc- ture called the panicle. The panicle branches bear structures called spikelets (Figure 2.5). These spikelets are composed of one or more flowers (usually called florets in grasses) with enveloping scales that conceal the flowers from view until flowering time. The florets of perennial grasses are often self-sterile and achieve cross-pollination using wind. The fruit, or grain, of grasses is called a caryopsis, although the term ‘grass seed’ is often applied to the entire spikelet. The growth habit of perennial grasses has made them ideal primary pro- ducers for grazing ecosystems. The production of new shoots by tillering provides a rapid means of recovery from the grazing pressure imposed by herbivores, and new growth takes place chiefly at the base of the leaves where it is least likely to be damaged by grazing mammals. The root system also improves the condition of the underlying soil. The roots bind the soil particles together, forming a sod, and bring to the surface layers nutrients that have been leached into the subsoil by heavy rain. A morphological characteristic 36 THE BIOLOGY OF AFRICAN SAVANNAHS of many perennial savannah grasses is the protection of the apical bud from fire and desiccation by a thick tunic formed by the old leaf sheaths. These spe- cies, having their buds protected at ground level, are able to regrow rapidly after fire. Some geophytic species may have their buds protected in the soil at depths of 10 cm or more. Although the tussock grasses are the dominant growth form, both in species and biomass, among plants with non-woody underground storage organs, there is a rich diversity of other herbaceous, bulbous, geophytes. These are represented by numerous species of Amaryllidaceae (daffodils), Iridaceae (irises), Liliaceae (lilies), and Orchidaceae (orchids). Perennial plants with woody underground storage organs, but with all shoots annual, form another characteristic feature of African savannahs. The annual shoots dry out completely during the dry season. They have been variously called subshrubs, geofrutescent, geoxyles, or hemixyles. Some of these growth forms are permanent geoxyles, conserving this habit under all circumstances, while others are traumatic geoxyles, in which this habit results from external injury. These traumatic geoxyles revert to a normal tree habit when circumstances allow. Many African savannahs have a rich flora of these species. In the Lake Edward plain, subshrubs such as Vigna friesiorum and Cissus mildbraedii can form up to 35% of the ‘grassland’ flora (Sarmiento and Monasterio 1983). This growth form is also characteristic of the grasslands of the high plateaus covered by the Kalahari sands. Annuals The contribution of annuals to the savannah herb layer is generally incon- spicuous. In the Raunkiaer system this group would include the thero- phytes. Although many of these annuals are grasses, they are not very common in savannahs, compared to perennial grasses. Annuals are more characteristic of savannahs that have been modified by overstocking of cattle, at least in mesic and arid savannahs inhabited by pastoral tribes. Bush fires can also contribute to the elimination of perennials by favour- ing annuals whose seeds ripen before burning takes place, and survive in the ground. When burning is repeated, or is started very late in the dry season, its influence is very much the same as that of overgrazing. The basic above ground structure of annual grasses is similar to that of per- ennial grasses, however, in contrast to perennial grasses, the flowers of annual grasses are usually self-fertile and do not rely on wind for pollina- tion. They also tend to have thin spreading root systems compared to the deeper root systems of perennial grasses. Sillans (1958) considered that perennial grasses make up the stable, basic component of the savannah grass layer and that annuals constitute only a fleeting component, filling gaps when the opportunity arises. TTHE VEGETATION 37 Phenology As we have already said, variations in plant structure and function associat- ed with the annual cycle of day length, temperature, and rainfall are called phenology. These phenological events include the germination of seeds, the appearance of leaves (the time between leaf bud opening and leaf senescence), flowering, fruit maturation, and the growth of vegetative parts such as stems and roots. The overriding influence on the phenology of savannah plants is the alternation of wet and dry seasons. The deciduous nature of many trees, as a response to this, has already been mentioned. Three Acacia examples are shown in Figure 2.6 and since the phenology varies in different parts of Afri- ca, these three examples are specifically for Tanzania. For example, in Kenya, for Acacia sieberiana, the no-flowers period would extend from March to October. Leaf expansion in some trees may precede the rains, especially if the previous rainy season was exceptionally wet, or prolonged, leaving resid- ual water in the soil. Early leaves in some species (e.g. red-leaved rock fig Ficus ingens) are bright red because anthocyanin pigments predominate over chlorophyll. One important aspect of phenology is the timing of pollination. A lot of work on pollination biology has considered how plants might avoid competition for pollinators. Plants may differ in the pollinators they recruit, but many East African Acacia flowers (Stoneet al. 1998) are visited by a wide diversity of species, at least some of which are shared by more than one Aca- cia species. Another solution is to use different populations of pollinators over time, and certainly in parts of East Africa this appears to be the case. Table 2.2 shows the seasonal flowering patterns of Acacia species present in the Mkomazi reserve, Tanzania (Stone et al. 1998). There is still some overlap between Acacia species, but there is also phenological separation. One interesting correlate of plant phenology is that the annual cycle of grass- es, and to a lesser extent trees, is associated with changes in their palatability Jan Feb Mar Apr May JunJul Aug Sep Oct Nov Dec Leaves Acacia mellifera Flowers Pods Leaves Acacia Flowers senegal Pods Leaves Acacia Flowers sieberiana Pods Figure 2.6 Phenology of three Acacia species in Tanzania. The varying degrees of shading indicate different intensities of leaves, flowers, and pods, in different months, with dark grey indicating prolific, and white none (modified from Dharani 2006). 38 THE BIOLOGY OF AFRICAN SAVANNAHS Table 2.2 Seasonal flowering patterns of Mkomazi Acacia species Acacia species Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec A. brevispica + * * + + A. bussei + * * + A. etbaica + * * A. reficiens + + * * + + A. thomasii * * + + A. drepanolobium * + + + + * A. nilotica * + + + * A. senegal * + + + + * A. tortilis * + + + * A. zanzibarica * + + + * An asterisk (*) indicates mass flowering, and a cross (+) slight, scattered flowering. Data from Stone et al. (1998). (see Table 2.4 later). For example, in many plants the nitrogen content of young tissue is high and the fibre content low. As the plant grows, and the dry season approaches, the ratio of leaf nitrogen to fibre tends to fall. Scholes et al. (2003) comment that as a consequence of this, on nitrogen-poor soils in Kruger National Park (granite uplands and sandstones), the nitrogen content may drop below the threshold for ruminant digestion (see Chapter 3). These areas are called sourveld in southern Africa. On soils with a high nitrogen- supplying capacity (granitic bottomlands and soils derived from basalt sedi- ments), this does not happen and grazers can be sustained throughout the year. Such areas are called sweetveld in southern Africa. This difference in nutritive phenology is thought to underlie the difference in herbivore bio- mass and composition between the granite and basalt landscapes in Kru- ger (Scholes et al. 2003). A more detailed example, with buffalo, of how this annual change in the nutritive value of vegetation can greatly affect some herbivores, is given in Chapter 4. Grasses Savannahs are essentially tropical grasslands with trees. African savannahs are dominated by herbivores, many of which eat grasses, either as grazers or mixed feeders (Chapter 3). It is estimated that there are about 10,000 species belonging to the family of grasses (Poaceae) in the world. However, the num- ber of grass species in any one savannah is usually only between 30 and 60, with 6 to 10 dominant species. Some widespread African savannah species are listed in Table 2.3. Many of these are species mentioned in other chapters. TTHE VEGETATION 39 Table 2.3 Some principal grass species of African savannahs Scientific name Common name Type Height Savannah (cm) type Aristida adscensionis Common needle grass Annual or perennial 10–100 g Aristida stipoides Needle grass Tufted annual 60–100 t Andropogon gayanus Bluestem grass Tufted perennial up to 300 t w m Andropogon greenwayi Bluestem grass Tufted perennial 45–60 g Brachiaria brizantha Common signal grass Erect perennial up to 120 w Brachiaria deflexa Signal grass Annual up to 45 t w Cenchrus biflorus African foxtail Tufted annual up to 90 t Cenchrus ciliaris Common African foxtail Rhizomatous perennial 20–110 g t Chloris gayana Rhodes grass Stoloniferous perennial 30–120 g m Chrysopogon aucheri Aucher’s grass Tufted perennial up to 50 t Ctenium newtonii Sickle grass Tufted wiry perennial up to 100 g Cymbopogon afronardus Blue citronella Tufted perennial 200 g w Cymbopogon plurinodis Citronella Tufted perennial 40–100 g Cynodon dactylon Bermuda grass Stoloniferous perennial 8–180 g t Dactyloctenium aegyptium Common crowfoot Annual up to 60 g Digitaria eriantha Pangola grass Tufted perennial 50–250 g Digitaria macroblephora Woolly finger grass Tufted perennial 40–100 g t Echinochloa colonum Barnyard grass Tufted annual up to 60 g Echinochloa pyramidalis Antelope grass Reed like perennial up to 500 g m Enneapogon cenchroides Grey head grass Tufted annual 30–60 g Enteropogon macrostachyus Bush rye Tufted perennial 90 g t Eragrostis lehmanniana Lehmann’s lovegrass Tufted perennial 45–61 g t Eragrostis patens Desert lovegrass Tufted annual 15–30 g t Eragrostis superba Masai lovegrass Tufted perennial 30–90 g t w Heteropogon contortus Spear grass Tufted perennial up to 90 g t w Hyparrhenia cymbaria Coloured hood grass Tall perennial 180—600 g w Hyparrhenia diplandra Sword hood grass Tufted perennial 150—300 g Hyparrhenia dissoluta Yellow hood grass Tufted perennial up to 300 g w m Hyparrhenia filipendula Fine hood grass Slender perennial up to 300 w m Hyparrhenia rufa Brown hood grass Variable perennial up to 300 w m Imperata cylindrica Cotton grass Perennial up to 120 g w Loudetia kagerensis Sour russet grass Tufted perennial up to 110 g Loudetia simplex Common russet grass Tufted perennial up to 150 g m Monocymbium ceresiiforme Wild oatgrass Tufted perennial up to 120 g Oryza barthii Wild rice Tall annual 60–200 g w Panicum coloratum Blue guinea grass Tufted perennial 8–100 g Panicum maximum Slender guinea grass Tufted perennial 70–150 g w Panicum turgidium Taman guinea grass Tufted perennial up to 100 g Pennisetum clandestinum Kikuyu grass Stoloniferous perennial up to 46 g 40 THE BIOLOGY OF AFRICAN SAVANNAHS Table 2.3 (Continued) Scientific name Common name Type Height Savannah (cm) type Pennisetum mezianum Bamboo grass Tufted perennial 40–160 g Pennisetum purpureum Elephant grass Tufted perennial up to 450 g Pennisetum schimperi Wire grass Tufted perennial up to 120 g Pogonarthria squarrosa Herringbone grass Tufted perennial up to 120 w m Schizachyrium semiberbe Crimson bluestem Tufted perennial 60–150 g w Schmidtia bulbosa Zand kweek grass Tufted perennial up to 80 g t w Schoenefeldia gracilis Spiky grass Tufted annual 70–120 g t Setaria incrassata Setaria Tufted perennial 60–100 g Setaria sphacelata Common setaria Tufted perennial 45–180 g w Sporobolus pellucidus Dropseed Tufted perennial up to 60 g Sporobolus pyramidalis Pyramid dropseed Tufted perennial 45–150 g t w Sporobolus robustus reed dropseed Tufted perennial 100–200 g Sporobolus spicatus Spike dropseed Stoloniferous perennial 9–30 g Stipagrostis uniplumis Silky bushman grass Short-lived perennial up to 75 g t Themeda triandra Red oat grass Tufted perennial 30–200 g t w Tragus koelerioides Creeping carrot grass Stoloniferous perennial 12–95 g g, open grass and shrub savannah; m, miombo; t, tree and scrub savannah; w, woodland savannah. Some species, such as Ctenium newtonii, are representative of a widespread group of very similar species. For example, C. newtoni is found in open scrub savannah across western African, from Senegal to Angola, extending east- wards into the Sudan. In East Africa it intergrades with a similar species, C. somalense, which extends down to Zambia. A third, southern African, species C. concinnum, then intergrades with C. somalense. This species com- plex therefore extends across scrub savannahs from western Africa, through East Africa, to southern Africa. Many grass genera, for example Cymbopo- gon, are notorious for their considerable variation within species and the weak separation between them. Consequently their taxonomy is still in a very fluid state. Sometimes even different genera, for exampleCymbopogon , Andropogon, and Hyparrhenia, are quite difficult to separate. The grass family is divided into five subfamilies (Bambusoideae, Arun- dinoideae, Pooideae, Chloridoideae and Panicoideae). The most primi- tive grasses are thought to be the bamboos, but Africa is relatively poor in these species. In fact, tropical African grasses are predominantly of the subfamily Panicoideae, with the two tribes Andropogoneae (e.g. Andropo- gon, Cenchrus, Chrysopogon, Cymbopogon, Heteropogon, Hyparrhenia, Imperata, Monocymbium, Pennisetum, Schizachyrium, Themeda) and Pan- iceae (e.g. Brachiaria, Digitaria, Echinochloa, Panicum, Setaria) accounting TTHE VEGETATION 41 (a) (b) (c) (d) (e) (f) Figure 2.7 Some African grasses: (a) Echinochloa pyramidalis, (b) Heteropogon contortus, (c) Mono cymbium ceresiforms, (d) Panicum maximus, (e) Schmidtia bulbosa, (f) Sporobolus nitens (drawings by Cythna Letty, from J.W. Bews, The world's grasses: their differentiation, distribution, economics and ecology, Longman, London, 1929). for the majority of species. Some examples of African grasses are shown in Figure 2.7. Notice that the inflorescence (the panicle and its component spikelets) is quite variable. In grass taxonomy it tends to take the place of the flower in flowering plant (angiosperm) taxonomy. Trees Some of the more important families of savannah trees are described below, and within each family some representative and important species are described. The information on trees is taken from Coates Palgrave (1983), Dharani (2006), Noad and Birnie (1989) and my own observations. Most of the families with tree forms are dicotyledons. However, the monocotyledons (in addition to the grasses) produce a few savannah forms that have tree-like growth forms. The palms and the aloes are first briefly mentioned. 42 THE BIOLOGY OF AFRICAN SAVANNAHS Palm family (Arecaceae) Palms are widespread throughout the tropics and subtropics and are remark- ably constant in their growth pattern. They have a slender, smooth, stem with a crown of large leaves. There are over 2,700 species of palms. The fruits are often rich in food and at least three species have economic importance. These are the coconut palm (Cocos nucifera), the date palm (Phoenix dactylifera) and the oil palm (Elaeis guineensis). Theborassus palm (Borassus aethiopum) is a wide- spread African species, being found from East Africa to southern Africa. Lily family (Liliaceae) The main savannah species within this family are the aloes. These plants are more or less succulent, and some are shrubby to tree-like. The leaves are long and fleshy, the base clasping the stem. There are some 250 species of aloes, the majority of which are indigenous to the drier parts of eastern, central, and southern Africa. The medicinal properties of some species has been recog- nised in Europe for many centuries. Aloe vera, from North Africa, is widely cultivated and an extract from the pulp is used to sooth and heal the skin. The quiver tree (A. dichotoma) reaches 3–5 m, occasionally 7 m, in height. It is a thickset and massive aloe, with a stem up to 1 m in diameter at ground level. The San people (Bushmen) of southern Africa make quivers for their arrows from the soft branches of this tree, hence its common name. Birds and locusts are attracted to the plentiful nectar, and baboons tear the flowers apart to get at the sweet liquid. Grevillea family (Proteaceae) This is one of the most notable families of the southern hemisphere, with related species in South America, southern Africa and Australasia. The trees are suited to dry conditions with leathery, evergreen, foliage. The beechwood (Faurea saligna) is a small to medium-sized (7–10 m) deciduous shrub, widespread in many highland savannahs in East Africa, and the woodland (miombo) savannahs of southern Africa. In the miombo it is frequently asso- ciated with the broad-leaved beechwood (F. speciosa), a small (4–7 m) leafy tree. Both Faurea species have flowers in spikes and fruits that are small nuts. The genus Proteus dominates the Cape flora in South Africa but most are montane rather than savannah species. The highveld protea (Protea caffra) is found in highland savannahs. Sour plum family (Olaceae) The sour plum family contains the genus Ximenia. The small sour plum (X. americana) and the large sour plum (X. caffra) are both found in the mopane ‘woodland’ savannahs of southern Africa. They are small to medium TTHE VEGETATION 43 bushes (4 to 6 m in height), often associated with termite mounds. They have grey bark, oblong elliptic leaves (5 to 2.5 cm) and, like many savannah trees, spines. The flowers are small and white, and the fruit red and sour. Caper family (Capparaceae) This is a medium-sized, mainly tropical family of shrubby trees and herbs. They are common in dry areas and recognised by the conspicuous spreading stamens of the flowers. The caper, commonly used as a pickle, is the bud of Capparis spinosa, a native of the Mediterranean region. In Africa, the genus Boscia is frequently encountered in dry grass and shrub savannahs, often on rocky outcrops. Boscia senegalensis is common along the Sahel, and in East Africa, and the shepherd’s tree (B. albitrunca) and the smelly boscia (B. foetida) are common in southern Africa. They are stocky trees (3–7 m height), stiffly branched, often associated with termite mounds. Mobola family (Chrysobalanaceae) Themobola plum (Parinari curatellifolia) is found in the wooded grasslands associated with the forest–savannah mosaic, and also the tree and shrub (baikian) savannah and wooded savannahs of central and southern Africa. It is a large, evergreen, spreading, mushroom-shaped, tree reaching up to 13 m in height, often forming a conspicuous feature of the landscape. When David Livingstone died on 1 May 1873 at Chitambo, in central Zambia, a commemorative inscription was carved on the trunk of a fine specimen of P. curatellifolia (Coates Palgrave 1983). Pod-bearing family (Leguminosae) The Leguminosae is the third largest family of flowering plants, containing some 17,000 species worldwide. It contains three major subfamilies: the Cas- sia subfamily (Caessalpiniodes), the Acacia subfamily (Mimosoideae), and the pea subfamily (Papilionoideae). However, they all share two features. First, the fruit is always a one-chambered pod, and second, they have root nodules containing nitrogen-fixing bacteria. The latter enables legumes to grow in relatively poor soil with little available nitrogen. Undoubtedly, one of the most characteristic features of savannahs are the thorn trees of the genus Acacia, and we therefore describe some of these species in some detail. All African species bear spines (Vachellia spp.) (Figure 2.8b) or thorns (Senegalia spp.) (Figure 2.8a), and get their name from the Greek word akis, meaning a sharp point. The spine is a hardened modified stipule and may be either straight or recurved. They have com- pound leaves (Figure 2.8a), subdivided into pinnae, which in turn are usu- ally subdivided into many, long, feathery leaflets. Occasionally, the leaflets 44 THE BIOLOGY OF AFRICAN SAVANNAHS (a) (b) Petiole Hook (c) Leaet Pinna Figure 2.8 The structure of an Acacia: (a) a leaf and hooks of A. senegal, (b) flowers and spines of A. tortilis, (c) pods of A. senegal subsp. leiorhachis (from Dharani 2006). are fewer and more rounded (e.g. A. mellifera and A. nigrescens). The flow- ers are small, but they are grouped into a flower-head that is either a spike (bottle-brush or spicate inflorescence) or a sphere (capitate inflorescence) (Figure 2.8b). In the subgenus Aculeiferum (e.g. A. senegal) the flowers are usually in spikes and produce nectar, while in the subgenus Acacia (e.g. A. nilotica) they are usually in a sphere and lack nectar. The flowers of both subgenera contain both male and female parts, and last for only a day. The flowers are white, cream or yellow. In many savannah areas the small leaves of these trees provide the only green browse in the dry season. The stem sap is also eaten by vervet monkeys and baboons, and the stringy bark by ele- phants and giraffe. Bees and butterflies utilise the nectar, and bruchid beetles the seeds. The pods of some Acacia species, such as A. tortilis and A. nilotica, are also utilised extensively by both wild mammalian species and domestic livestock. The pods vary in shape, size, texture and colour and are important in field identification. They may be thick and woody (e.g. A. nilotica) or thin and papery (e.g. A. mellifera). The indigenous African acacias are frequently described as taxonomically difficult, and some species can be very variable in their growth form. Since the 17th International Botanical Congress (2005), all 142 species of Afri- can Acacia have been controversially reclassified into two newly constituted genera: Vachellia (73 species) and Senegalia (69 species). This renaming was carried out despite the fact that an African Acacia, the Egyptian thorn (A. nilotica), was the original type specimen for the genus. Many African bota- nists, and end-users, feel quite aggrieved about this change and intend to TTHE VEGETATION 45 carry on using the name Acacia for the African species. We adopt this strat- egy here, although the new genera are indicated below for each species. See Kyalangalilwa et al. (2013) for a phylogenetic discussion and analysis. The prickly acacia (Acacia brevispica; Senegalia brevispica), also called the wait-a-bit-thorn, is a shrub or small tree, 1–8 m high, with long, thin, ram- bling branches covered with small scattered thorns. The leaves are fairly large and the flowers are in white balls. The pods are thin, flat, and broad and leath- ery. The young pods are used as fodder for goats and cattle, although sheep refused diets containing A. brevispica, perhaps because of its proanthocya- nidin content (Woodward and Reed 1995). Both this species and A. drepa- nolobium appear to be key indicators of rangelands that have been subjected to high grazing and browsing pressure over long periods of time (Solomon et al. 2007). It is widespread in Africa, being found in the Sudan, Ethiopia, Somalia, Kenya, the Democratic Republic of the Congo, Angola, Natal and Cape Province. The whistling thorn (A. drepanolobium)(Vachellia drepanolobium), is a small shrub or tree (1–7.5 m), usually either short and robust or tall and slender, with a rounded canopy. It is abundant on poorly drained black- cotton and clay soils in East Africa. Its most noticeable feature are the swol- len, purplish or black, galls at the base of the pair of long, whitish, spines (1.5–4.5 cm). These galls contain various species of ant (Chapter 5). The pods (4–7 × 0.5 cm) are brown and slightly spiralled. Young galls, pods and leaves are all eaten by wildlife, and the pods are much liked by giraffe. Whistling thorn is a rather variable species, with soil conditions, browsing and fire all contribut- ing to its final shape. The blue thorn (A. erubescens) (Senegalia erube- scens) can either be a multistemmed Acacia drepanolobium shrub, or a 10 m high tree. The name erubescens is Latin for ‘becom- ing red’ and refers to the colour of the young flowers. It is found in dry tree and shrub savannah and is quite characteristic of Angolan mopane woodland (Chapter 1). The bark is yellowish to greyish-brown, and scaly or papery. The thorns are in pairs (no spines), strongly hooked and up to 7 cm long. It has classic Acacia compound leaves, with 3–7 pairs of pinnae, each with 10–25 pairs of leaflets. The flowers are in short, squat, spiked inflorescences, and are, eventually, yellowish-white. The pods are light brown to dark brown, leathery, and 3–13 × 1–1.8 cm in size. It is found southwards from Tanzania, to Namibia, Botswana and South Africa. 46 THE BIOLOGY OF AFRICAN SAVANNAHS Thegrey-haired acacia, or Gerrard’s thorn (A. gerrardii)(Vachellia gerrardii) is a widespread species, occurring from Nigeria to Sudan in the north, and south- wards to Botswana and South Africa. Along with A. drepanolobium, A. robusta and A. tortilis it is one of the dominant acacias in the Serengeti ecosystem. The bark is dark grey or reddish and the stipular spines, in pairs, are greyish. In East Africa they can occasionally have galls like A. drepanolobium. The leaves have 2–12 pairs of pinnae, each with 8–23 pairs of leaflets, and the flowers are white, tinged with pink, in a capitate inflorescence. The clustered pods (5–16 × 0.6–1.7 cm) are dark brown, curved, and covered with grey hairs. The tree is browsed (bark and leaves by black rhino and the foliage and seed by giraffe, duik- er, kudu and steen- bok), improves soil quality, and is exten- sively used by bees. The bark is used as a traditional remedy for coughs and sore throats. Acacia gerrardii The candle-pod aca- cia (A. hebeclada)(V. hebeclada) is a shrub or small tree up to 7 m in height, often branching near the ground, often forming thickets. Along with A. kirkii, A. nilotica and A. erube- scens it is a typical species of Angolan mopane savannah woodland (Chapter 1). The bark is dark grey, fissured and often flaking. The stipular spines are quite variable, sometimes short and sometimes long (up to 3.5 cm). The leaves have 2–9 pairs of pinnae, each with 7–16 pairs of leaflets. The flowers are creamy white, in balls, and the pod is hard and woody (4–15 × 1.4–4 cm) covered in yel- low-grey hairs. In southern Africa it is known as the ‘house of the lion’, because the low, spreading habit provides shelter and shade for lion and Bushmen alike. Kirk’s acacia (A. kirkii)(V. kirkii), also known as the flood-plain acacia, is a handsome, flat topped, shrub or tree which can reach a height of 2–15 m. Like A. ebeclada it is a typical species of Angolan mopane savannah woodland (Chapter 1). It is found in Uganda, Acacia kirkii Kenya, and Tanza- nia, extending south- wards to Zambia, Botswana and south-west Africa. The bark is brown, fissured, and rough with TTHE VEGETATION 47 each leaf composed of 3–14 pairs of pinnae, each with 9–18 pairs of leaflets. The flowers are cream or white in a capitate inflorescence, and the short pods (2–10 × 0.8–2 cm) are reddish-brown. The leaves and pods are used as fodder and the Masai make ‘tea’ from the bark. The black-hooked thorn (A. laeta)(S. laeta) is a shrub or small tree up to 6 m high, with a greyish-green bark looking blackish from a distance. The leaves have 2–3 pairs of pinnae, each with 3–5 pairs of leaflets. The thorns, or hooks, are paired (sometimes in threes), and the flowers are creamish white or pink- ish, in 5 cm long spicate inflorescences. The mature pods are pale brown (3.5–8 × 1.7–2.8 cm). It is found in grass and shrub savannah along the whole Sahel, from Niger to Egypt, Sudan, Ethiopia, and Somalia and south into Tanzania and Kenya. Many butterfly and beetle species feed on the flowers. It may be a hybrid between A. mellifera and A. senegal. The wait-a-bit-thorn or hook thorn (A. mellifera)(S. mellifera) is a tall, rounded shrub or small tree with a round crown, reaching occasionally 9 m in height in the southern limit of its distribution. The branches are covered with very sharp recurved thorns which attach themselves to clothing and pull the wearer back; hence the common name. The bark is smooth grey, with white lenticels on the young branches and shoots. The leaves have 1–2 pairs of pin- nae with 1–2 pairs of asymmetrical leaflets. It simultaneously develops leaves and flowers in the early rainy season. There are no spines, just the recurved thorns, in pairs. The fragrant flowers are white to cream-coloured, 3.5 cm long, and gathered in short, dense, hanging spikes. The pods are flat, oblong, 3–8 × 1.5–2.5 cm, with slightly con- stricted margins between the seeds, papery, and gener- ally containing three Acacia mellifera seeds. The pods, young twigs, leaves and flowers are all very nutritious and eagerly consumed by wildlife. It is a widespread species, occurring in the Sahel and in eastern and southern Africa. The nob-thorn (A. nigrescens)(S. nigrescens) is usually a low tree but can grow up to 8–10 m. The bark is dark brown and covered with large, conspicu- ous, knobby prickles. There are no spines. Each leaf has only 2–4 pinnae, each with one or two round leaflets. The inflorescence is a spike (1–10 cm long), with many yellowish-white flowers. The pods are green when young, becoming dark brown when mature. The species is both fire and drought resistant. It is widespread in tree and shrub savannah, being found from Tan- zania southwards into Namibia, Botswana and South Africa. The powder from the dry root is used as a traditional treatment for snake bite. 48 THE BIOLOGY OF AFRICAN SAVANNAHS TheEgyptian thorn (A. nilotica)(V. nilotica) is a small to medium-sized tree (5–20 m high), with a dense spherical crown. The stems and branches are usually dark to black coloured, with fissured bark. The thin, straight, light grey spines are in pairs, 5–7.5 cm long in young trees. Mature trees are com- monly without thorns. The leaves are bipinnate, with 3–6 pairs of pinnae and 10–30 pairs of leaflets each. The flowers are in round heads, 1.2–1.5 cm in diameter, and bright yellow. The pods are green when young, becom- ing darker with age and eventually drying black. They are velvety, straight or slightly curved, 5–15 cm long × 0.5 to 1.2 cm wide, with constrictions between the seeds giving a necklace appearance. It is widespread and has been divided into seven subspecies which vary in shape, size and hairiness of the pods. Acacia nilotica has a wealth of medicinal uses. It is used for stomach upset and pain, the bark is chewed to protect against scurvy, and an infusion is taken for dysentery and diarrhoea. Thesplendid acacia (A. robusta)(V. robusta) has a very variable growth form from a short, multibranched, shrub to a tall (up to 25 m) flat-topped tree. The latter height is attained on wet soils. The bark is often deeply fissured, and the straight spines are in pairs (up to 14 cm long). However, old branches fre- quently have short spines, up to 1.2 cm, thickened at the base. The leaves are bipinnate, with 1–10 pairs of pinnae and 6–27 pairs of leaflets each. The flow- ers are white, in a round (capitate) inflorescence, and the flattened, curved, pods (10–19 × 0.7–1.5 cm) are red-brown to black. The pods and leaves are occasionally browsed, but the pods are suspected of causing prussic acid poi- soning in livestock. Theumbrella thorn (A. tortilis)(V. tortilis), the quintessential savannah Aca- cia, is widespread throughout African savannahs. It is a wide-spreading, flat- topped or umbrella-shaped tree, up to 4 m high, with mixed spines, some white, straight, and slender, up to 7.5 cm long, and others grey with black or brown tips, sharply curved, and very small. The pinnae are in 3–10 pairs, and leaflets in 7–15 pairs. It is almost always ever- green. The flower heads are white to cream, and the pods Acacia tortilis are yellow-brown. The pods are spirally twisted, sometimes even curled into rings, slightly constricted between the seeds, circular in cross-section, and 7.15–15 cm long × 0.6–0.8 cm thick. TTHE VEGETATION 49 It occurs in the drier areas of Africa, in the Sudan, Kenya and Tanzania, in southern Africa and Namibia. Analysis of 14C content and anatomy of wood (Andersen and Krzywinski 2007) suggests that trees can be 200–650 years old. Thethree-thorned or gum arabic acacia (A. senegal)(S. senegal) is a bush or small tree, usually 2 to 8 m high, occasionally reaching 10 m under optimal conditions, frequently forming thickets. It has a short stem, is usually low branched with many upright twigs, with the crown eventually flattened and umbrella-shaped. The prickles are in threes, up to 0.5 cm long, with the cen- tre one sharply curved downwards and the other two curved upwards. The leaves are bipinnate, small, greenish-grey, with 3–6 pairs of pinnulae hav- ing 10–20 pairs of leaf- lets each. The leaflets are grey-green, 3–8 x 1–2 mm. The flow- ers are very fragrant, creamy white, usu- ally appearing before the leaves in spikes, Acacia senegal 3–10 cm long, either solitary or two to three together. The pods are 7–10 cm long by 2 cm wide, flat and thin, papery, attenuated at both ends, yellowish to brown. The leaves and pods are protein-rich and are eaten by wildlife such as giraffe, black rhinoceros and gerenuk. It is found in the Sahel from the Atlantic to Ethiopia, extending southwards into eastern and south- ern Africa. The white thorn (Acacia seyal)(V. seyal) is a small, slender tree, reaching 6 to15 m in height, which can develop a characteristic umbrella-shaped canopy when mature. The twigs have paired thorns, up to 7 cm long, which are narrow, straight, sharp-ended, and grey in colour. The leaves are dark green, with 4–12 pairs of pin- nae, each with 10–22 pairs of leaflets. The Acacia seyal flowers are clusters of 2–3, with bright yellow globose heads about 1.5 cm in diameter, on peduncles about 3 cm long. The pods are hang- ing, slightly curved, dehiscent, light brown when mature, 10–15 cm 50 THE BIOLOGY OF AFRICAN SAVANNAHS long by 1 cm wide, containing 6–10 seeds each. A. seyal is found in the Sahel, from the Atlantic Ocean to the Red Sea, and south to East and South Africa. In East Africa there is a variety called fistula, with myr- mecophilous, swollen-based thorns, or ‘ant galls’ similar to those of A. drepanolobium. The paperbark acacia (A. sieberiana)(V. sieberiana) is another, very com- mon, flat-topped acacia. Like A. tortilis, it is also called the umbrella thorn. It can reach 15–25 m in height, with strongly fissured bark, which is yellow to cream-coloured in young trees and twigs, scaly in old trees. The thorns are greyish white, strong and straight, in pairs up to 6–10 cm long. The leaves are bipinnate, with 10–25 pairs of pinnulae having 15–50 pairs of leaflets each. The flowers are fragrant, cream-coloured or light yellow, in round heads. The pods are straight or slightly curved, woody and without hairs, yellowish to reddish-brown, and varnished and shining when mature. It is widely distrib- uted in Africa from the central Sudan to the southern Sudan, all of the central and southern Sahel, Kenya, Zambia, Zimbabwe, Angola, Botswana and the Transvaal. Although most Acacia species are utilised to some degree by browsers, the actual frequency of differentAcacia species in the diet varies with the area (containing differentAcacia species, in different frequencies), and with the preference of the browsing species. For example, Pellow (1981) looking at giraffe in the Serengeti found that six principalAcacia spe- cies were browsed, but that their selection rating varied with the season (Table 2.4). In the dry season the giraffe positively selected for other tree species such as Balanites aegyptiaca and Albizia harveii. The latter species, known as the sickle-leaved albizia, is also a tree in the subfamily Mimosoidea and is closely related to the Acacia species. The genusAlbizia has over 100 spe- cies, and is named after an Italian nobleman called Filipo degli Albizia who Table 2.4 Seasonal selection of the six principal Acacia species browsed by giraffe in the Serengeti National Park Acacia species Wet season Dry season A. robusta − − − − A. gerrardii + + 0 A. hockii 0 − A. senegal + − A. tortilis 0 − A. xanthophloea − + The symbols represent degrees of selection from + + (very positive), through +, 0, and −, to − − (very negative). Data from Pellow (1981). TTHE VEGETATION 51 introduced them to cultivation 200 years ago (Noad and Birnie 1989). The genus is widespread throughout the tropics. In many ways they resemble aca- cias, with their flattened pods, and bipinnate compound leaves. However, they do not have spines, or thorns, and the flowers are born in an inflores- cence that is a half-spherical head, never a spike. The purple-leaved albizia (A. antunesiana) is another species commonly found in woodland (miombo) savannahs. Both these albizias are small to medium trees, 5–11 m in height. Other members of the Mimosoidea occurring in tree and shrub savannah are the sickle bush (Dichrostachys cinerea) a small, spiny acacia-like tree to 6 m, and mesquites (Prosopis spp.), thorny shrubs/trees that grow to about 3 m, but can reach 15 m. Prosopis africana occurs from Senegal in the west, to Sudan and Kenya in the east. The subfamily Caesalpinioideae includes some spectacular flowering trees as well as climbers. Most are forest species, but several genera (Baikiaea, Bauhinia, Brachystegia, Burkeae, Colophospermum, Guibourtia, Julberlandia) are important species in many wooded savannahs (Chapter 1). The camel’s foot trees (genus Bauhinia), contains over 300 species throughout the trop- ics. However, two species, the yellow tree bauhinia (B. tomentosa) and the white bauhinia (B. petersiana) are low shrubs, or small trees up to 5 m, often found in East African tree and shrub savannahs. Another species found in this type of savannah, further south, is the Rhodesian teak (Baikiaea pluri- juga), dominating the so-called baikiaea open woodlands (Chapter 1). This is a medium to large tree (8–16 m), with a large, dense, spreading crown. It has very attractive, large, golden-brown flowers and flattened, woody, pods. The wild green-hair tree (Parkinsonia africana) is widespread in arid savannahs. The African mahogany (Afzelia africana) is a large tree with a spreading crown. Its height varies from 10 to 20 m and in west Burkina Faso mean average height was found to be 10 m for a tree with a 36 cm diameter. It has scaly grey-brown bark, with a pink or red slash. Twigs and leaves are hairless (glabrous), the latter up to 30 cm long, with 7–17 pairs of leaflets (5–15 × 3–9 cm) with an obtuse tip. The flowers have one single white petal, with red stripes, 1.5 cm long, set in terminal panicles up to 20 cm long. The pods are flattened, 12–17 × 5–8 × 3.5 cm, glabrous, black, woody, persistent, burst- ing open at maturity to spread the seeds. The seeds are poisonous, with a sweet edible covering. The tree is associated with ectomycorrhizal fungi and is found from Senegal to Uganda. The sturdy, spreading, miombo tree or Prince-of-Wales’ feathers (Brachy- stegia boehimii) (5–16 m in height) gives its name to the miombo wood- land savannahs of central African described in Chapter 1. This large genus is also part of the subfamily of legumes. Brachystegia have pinnate leaves, small inconspicuous flowers, and flat woody pods that split explosively. Apart from B. boehimii, miombo woodlands are dominated by two other species of Brachystegia. B. spiciformis, also known as miombo or msasa, is a 52 THE BIOLOGY OF AFRICAN SAVANNAHS medium to large tree (8–15 m in height) and in Zimbabwe and Mozambique it is the most widespread species of Brachystegia. It hybridises readily with B. glaucescens. Themountain acacia (B. glaucascens) is a spreading, beautiful tree (up to 15 m in height) that often develops a flat top, hence its common name. Coexisting in miombo woodlands are the closely related munondo (Julberlandia globiflora) and large-leaved munondo (J. paniculata). Julber- landia also have pinnate compound leaves, inconspicuous, small, white/ cream flowers and dark brown, velvety, pods. Theburke (Burkeae africana) is a widely distributed medium sized tree (8–10 m). It is found in both the wooded grasslands of the forest–savannah mosaic and the tree and shrub savannahs. Once again, the leaves are pinnate and the fruit a thin, flat, pod. Emphasising the similarity of many of these leguminous savannah trees, the burke is often confused with Albi- zia; a tree in another subfamily. The mopane (Colophospermum mopane) is a medium to large tree (4–18 m) dominant over many areas south of the miombo woodland savannahs. Thelarge false mopane (Guibourtia cole- osperma) is a medium to large, evergreen, tree (6 to 20 m) with a rounded crown. The small false mopane (G. conjugata) is a small to medium tree (7–9 m). Brachystegia spiciformis Desert date family (Balanitaceae) A small tropical fam- ily with only one genus, Balanites, of which there are sev- eral African species. They are shrubs or trees, with spines that are simple or forked, and a fleshy fruit. A species frequently found in dry savan- nah grassland is the Balanites egyptiaca desert date (Balan- ites aegyptiaca,), a slow-growing evergreen tree usually 3–6 m in height, occasionally growing to 10 m. It is armed with stout green or yellow spines, up to 8 cm long. The date-like fruit is long (5 × 2.5 cm) and yellowish-red when mature. The fruits are even produced in very dry years. Both fruit and foliage are browsed by TTHE VEGETATION 53 goats and camels and also by game, especially giraffe. It is widely distributed in both grass and shrub and tree and shrub savannahs, from northern and eastern Africa to southern areas. Myrrh family (Burseraceae) This tropical family, notable for their aromatic resins since ancient times, are found in Africa, Malaysia and Central America. Even in Biblical times these where used for incense and perfumes. Frankincense comes from the resin of Boswellia car- teri and other species found in Somalia and Ethiopia, and myrrh is extracted from Com- miphora abyssinica. The larg- est African genus in this family, with about 30 species, is Com- miphora. These are shrubs or trees, often with spines, and are leafless for most of the year. Commiphora africana The angular-stemmed commi- phora (Commiphora karibensis) is a small to medium tree (up to 7 m in height) often associated with the soft-leaved commiphora Commiphora( mollis) (8 m in height) in tree and shrub savannah. The poison-grub commiphora (Commiphora africana) is a deciduous thorny shrub or small tree, usually about 5 m high, with a grey- green bark that peels in papery scrolls. Although it is widely distributed in dry savannahs, its leaves, although they look a good source of browse, are rarely eaten because of the bitter tannins they contain. The larva of the beetle Diamphidia, from which the Bushmen make their arrow poison, feeds exclu- sively on this tree. Euphorbia family (Euphorbiaceae) The euphorbias comprise a very large family of herbs, shrubs and trees, with their distribution centred in the tropics. There are approximately some 5,000 species, in 300 genera. Economically important species include the cassa- va (Manishot esculenta). A widespread savannah species is the snowberry tree (Securinega virosa), a many-stemmed, bushy shrub 2–3 m in height. In Tanzania the roots and fruits are believed to be an effective snakebite remedy (Coates Palgrave 1983). Theheart-fruit (Hymenocardia acida) is a shrub or small tree (up to 6 m), often found with Brachystegia species and 54 THE BIOLOGY OF AFRICAN SAVANNAHS Uapaca species, in woodland miombo savannah. Uapaca kirkiana, known as mahobohobo in southern Africa, and the narrow-leaved mahobohobo (U. nitida), are just two of a number of species. They are medium-sized trees of about 10 m, with leathery leaves, creamy-green inconspicuous flowers, and an edible fruit. One of the best-known savannah species is the cande- labra tree (Euphorbia candelabrum). It grows up to 15 m, with a character- istic crown of massive ascending branches that make it look like a complex candelabra. The spined green stems are succulent and rather cactus-like. It is widely distributed in savannah areas from East Africa to southern Africa. Although a characteristic savannah species, it is not used by animals because its stems produce an extremely toxic latex. Some euphorbias, such as the non- succulent western woody euphorbia (Euphorbia guerichiana) are woody shrubs that have fleshy leaves, rather than succulent stems. Mango family (Anacardiaceae) A worldwide, but mainly tropical family, of about 600 trees, shrubs and vines. It includes a number of commercially important species, such as the cashew, pistachio, and mango. Savannah species including the marula (Sclerocarya birrea), a medium-sized tree up to 10 m in height found in tree and shrub savannah. The fruits are fleshy, up to 3.5 cm in diameter, and yellow when mature. There are very old stories that elephants, which eat them eagerly, become intoxicated after eating the over-ripe, fermenting, fruits lying on the ground. However, elephants show a clear preference for marula fruit still on the tree and it is unlikely that a 3 tonne elephant, gorging itself quickly on nothing but marula fruit, would be able to ingest enough ethanol to reach a blood alcohol content indicative of inebriation (Morris, Humphreys and Reynolds 2006). Observations on baboons, which also like these fruits, sug- gest that they also prefer fresh marula fruit, and because the pulp is digest- ed and the seeds passed within a 24 hour period, internal fermentation is impossible. It is regarded as a keystone large tree species in southern Afri- ca, but it is declining at an unprecedented rate in the Kruger National Park (Helm and Witkowski 2012), where loss of adult trees over the last decade exceeded 25%. Baobab family (Bombacaceae) This is a small family of tropical flowering trees, whose best-known species is the baobab (Adansonia digitata) with its massive bottle-shaped trunk, up to 15 m in height. Like many trees in this family, the swollen trunk of the bao- bab stores water, allowing it to survive in very dry savannahs. It can be bare of leaves for as much as 9 months of the year, giving it an odd ‘upside-down’ appearance. Elephants are quite fond of its soft pithy wood, and baobabs may not maintain viable populations in the presence of elephants. Edkins et al. (2007) found that comparing trees in the Limpopo National Park in TTHE VEGETATION 55 Mozambique (few elephants) and the Kruger National Park in South Africa (many elephants), the baobab populations in the Limpopo National Park showed a J-shaped size class distribution with many small baobabs, while Kruger National Park showed an absence of this size class, suggesting a lack of recruitment. In Kruger National Park, elephant impact (% bark stripped to a height of 3 m) decreased with increasing rockiness and slope steepness, sug- gestive of a possible baobab refuge in such areas. The baobab is one of the long- est-lived trees in the world. Radiocarbon dating has estimated that trees 5 m in diameter might be 1,000 years old and the very largest trees as much as 3,000 years old. These extraordinary trees are surround- ed by a wealth of African myth and legend, including the story that God planted the tree upside down and that a lion will devour anyone that picks a flower from baobab tree the baobab, because the flowers are inhabited by spirits. Terminalia family (Combretaceae) A mainly tropical family of trees, shrubs and climbers, all with simple leaves and flowers in clusters, often producing abundant nectar. The main African genus is Combretum, with over 30 species. They are shrubs, climbers or trees. They have woody, four or five winged, fruit. In contrast the related genus,Ter - minalia, with 12 indigenous species, has woody fruit with only two wings, often reduced to ridges. They are small to medium sized trees. In both genera the flowers have no petals and it is the stamens which are conspicuous. They are mainly found in the open ‘tree and shrub’ savannahs, north and south of the miombo ‘woodland’ savannah. Thered bushwillow (Combretum apiculatum) is a small to medium tree (3–10 m in height), occa- sionally shrub-like. The flowers are yellow to creamy-green and heavily scented. The wood is very resistant to termites. The lead- wood (C. imberbe) is sometimes a shrub, but more frequently a small to large tree (7–15 m in height). The bark is dark grey Combretum molle and rough, with characteristic deep longitudinal fissures. Again 56 THE BIOLOGY OF AFRICAN SAVANNAHS the flowers are heavily scented. The velvet-leaved combretum (C. molle) is a small to medium tree (up to 10 m in height). The flowers are heavily scented, attracting insects. The variable combretum (C. collinum) is a spreading, deciduous, small to medium-sized tree (4–12 m in height). The bark is light grey and rough, and the flowers are cream to yellow and sweetly scented. The genusTerminalia Terminalia sericea gets it name from the Latin termi- nus, referring to the fact that the leaves appear at the very tips of the shoots. The silver terminalia (T. sericea) is a small to medium-sized tree (4–6 m in height), with silvery silky leaves. The small cream flowers are unpleas- antly scented. In the drier wooded savannahs of East Africa, the muhutu (T. brownii), is a widely distributed deciduous tree 4–5 m in height (in Kenya it is the most common Terminalia). It is both drought and termite resistant, with a heavy leaf fall. The purple-pod terminalia (T. prunoides) is a shrub or medium-sized tree (3–7 m in height). They have dark green leaves and the cream flowers that have a strong, unpleasant smell. TheKalahari sand terminalia (T. brachystemma) is a small bushy tree (3–5 m in height) with the branches often horizontal. It has leathery green leaves and pale yellow flowers. It hybridises with T. sericea and the hybrids are widespread. The large-leaved terminalia (T. mollis) (10 m in height) has large leaves (up to 32 × 13 cm) and large greenish-white flowers. Ebony family (Ebenaceae) This is a largely tropical and subtropical family named after theebony tree (Diospyros ebenum) which is native to the rainforests of South-East Asia. However, several species of Euclea and Diospyros are found in African savan- nahs. Both the large-leaved euclea (E. natalensis) and E. pseudebenus, also confusingly called the ebony tree, are widespread southern African species. The giant diospyros (Diospyros abyssinica) (30–40 m in height) is a wide- spread African species, but really a forest tree, rather than a savannah spe- cies. It is, however, the codominant species, with the mutanga (Elaeodendron buchanii), in zone VI of the Serengeti (see Figure 2.14). Jacaranda family (Bignoniaceae) The jacaranda family is a tropical group of trees and shrubs with showy flowers, such as the striking mauve-blue of the introduced jacaranda tree TTHE VEGETATION 57 (Jacaranda mimosifolia). Thesausage tree (Kigelia africana) is a medium to large tree (up to 18 m in height), noticeable for its unusual sausage-shaped, greyish-brown hanging fruit, up to 1 m × 18 cm in size and much loved by elephants. Each fruit can weigh up to 10 kg. Under this tree, David Living- stone pitched camp just before he saw the Victoria Falls for the first time (Coates Palgrave 1983). Plants of the genus Rhigozum are small, spiny, shrubs or small trees. Thewestern rhigozum (R. brevispinosum) (2–4 m in height) is widespread in southern Africa in the mopane tree and shrub savannah. The flowers are golden-yellow and sweetly scented. Local vegetation patterns: the Serengeti–Mara ecosystem In Chapter 1 we looked at the distribution of savannah vegetation on a world scale and, for Africa, on a continental scale. Rainfall, its yearly amount and annual distribution, was important in determining this continental pattern. However, the large areas of savannah types shown in Figure 1.9 are not uni- form stands of vegetation. There is widespread local variation in species and the different types of savannah often form a mosaic at a local scale. It would be impossible to talk about this local scale for the whole of Figure 1.9 and so to illustrate this we will now examine the vegetation of one local geographical area of savannah that has been very well studied. This is the Serengeti–Mara ecosystem, in northern Tanzania and southern Kenya. Here we will see the local interplay of rain, topography, soil, and fire producing local changes in vegetation. Rainfall and green biomass The Serengeti–Mara ecosystem is an area of approximately 25,000 km2 on the border of Tanzania and Kenya (Figure 2.9), and essentially defined by the movement of the migratory wildebeest (see Chapter 5). The system has a long history as a protected area (Sinclair 2012). In 1929 an area of 2,286 km2 was established as a game reserve, in what is now southern and eastern Serengeti. In 1959 the northern extension to the Kenyan border was added and the eastern area was removed as the Ngorongoro Conservation Area. In 1965 the northern ‘Lamai wedge’ was added, and the Mara National Reserve was formed. The mean maximum monthly temperature is relatively con- stant between 27 °C and 28 °C at Seronera. The minimum temperature varies between 13 °C and 16 °C. However, the most important climatic variable is rainfall. Rainfall is seasonal (Figure 2.10a) because of the movement of the intertropical convergence zone (see Chapter 1), modified by local influences. The rainfall pattern is influenced by the Crater highlands, which rise to over 3,000 m and form a rain shadow immediately to the north-west. This pro- duces a gradient of increasing rainfall from the south-eastern grass plains, 58 THE BIOLOGY OF AFRICAN SAVANNAHS Kenya Lake Victoria Tanzania Loita Plains Loita Hills Northern Extension Speke Gulf Western Corridor Lake Natron Seronera Gol Mtns N Serengeti Plains Ngorongoro 0 30 km Crater Highlands Figure 2.9 The Serengeti–Mara ecosystem defined as the area used by the migratory wildebeest (broken line). The Serengeti National Park is outlined by a solid line. Hills are shaded and lakes are horizontal hatched (simplified from Sinclair 1979, with permission from University of Chicago Press). to the north-western woodlands on the Mara river. This gradient is present in both the wet and the dry seasons, and Figure 2.10 shows the rainfall con- tours (isohyets) for both seasons. Lake Victoria also modifies the rainfall pat- tern. The lake affects northern and western Serengeti by producing its own convergence zone and increasing rainfall between June and November. One consequence of this gradient in rainfall is a corresponding gradient in green grass biomass (there is also brown, or dry, grass biomass which of course is related to the biomass of new green grass). The widespread relationship between rainfall and green grass biomass, shown earlier in Figure 2.1, is also found more locally within the area of the Serengeti. Figure 2.11 shows the results of Braun (1973) for 69 different sites, over short grass (II), interme- diate grass (III), long grass (IV) and savannah woodland (V and VI) (see vegetation zones in Figure 2.10b and below). Aerial surveys of green bio- mass during 1974 and 1975 (McNaughton 1979a, 1979b) give a remarkable insight into how Figure 2.10 and Figure 2.11 combine to produce the local spatial distribution of available grazing vegetation, that must be typical of TTHE VEGETATION 59 (a) (b) 300 1,000 700 150 Kogatende VI 100 900 50 250 VI I VI 200 800 V 800 Handajega VIII 200 700 V 700 100 600 800 IV 100 II Simiyu 500 III Olduvai 400 150 I V 100 50 III 30 mm 30 50 km km ND JFMAMJ JAS O Figure 2.10 Mean rainfall isohyets (mm) for the Serengeti: (a) Dry season, plus four annual profiles, (b) wet season, with vegetation zones (see text) (from Sinclair 1979, with permission from University of Chicago Press.). 12,000 10,000 8,000 6,000 4,000 Green biomass (kg/ha) 2,000 0 200 400 600 800 1,000 Rainfall (mm) Figure 2.11 Serengeti grass yield with increasing rainfall. Data are for short grassland (●), inter mediate grassland (▲ Andropogon greenwayi stands, ○ mixed stands), long grassland (■), and savannah woodland (□) (from Braun 1973). 60 THE BIOLOGY OF AFRICAN SAVANNAHS (a) (b) 60 30 80 60 100 60 20 90 120 150 30 100 150 60 80 60 120 40 30 60 90 00 40 30 20 60 150 90 120 90 60 30 120 20 20 30 120 150 0 120 60 20 0 90 Figure 2.12 Isoclines of green biomass for the Serengeti: (a) 6 September 1974, in the middle of the dry season, (b) March 1975, in the middle of the wet season (from McNaughton 1979b, with permission from University of Chicago Press.). many ‘uniform’ savannah areas (Figure 2.12). During the dry season (Fig- ure 2.12a) the southern grass savannahs have little or no available green for- age. This is a major factor in the northern migration of wildebeest, plains zebra and Thomson’s gazelle. With the arrival of the rains these ungulates can migrate back to the southern grasslands which not only now have a good supply of green grass, but are also rich in minerals (Murray 1995) and lower in predators (Fryxell 1995). These three attributes of the southern grasslands, in the wet season, combine to produce ideal conditions for the production of young, which takes place at the beginning of the wet season. However the important message here is that the grass in savannahs is not a uniform ‘lawn’. Green biomass can vary quite markedly over relatively short distanc- es, and grazing animals must respond to this spatial variation. In fact this ‘local’ variation is not simply confined to green grass, most of the vegetation responds to rainfall, soil and fire (Sinclair 1979). More recently, Anderson et al. (2007a) have shown that both climate and topographic variation explain plant species richness in Serengeti grasslands, but specific patterns depend upon grain (sample area), extent (distance between samples), and, to a less- er degree, focus (geographical region). Consistent with the ideas of Shmida and Wilson (1985), determinants of plant species richness shifted from niche relations to habitat heterogeneity between spatial grains of 1 and 103 m2. TTHE VEGETATION 61 Soils, fire and vegetation In the south-eastern plains, the soils of the Serengeti are highly saline and alkaline as a result of their recent volcanic origin in the Crater highlands (Fig- ure 2.9). These highlands are volcanoes of Pleistocene age that produced aeri- ally discharged material which was blown westward and formed the Serengeti plains. The youngest soils, in zone I (Figure 2.10b), are highly porous, coarse- grained sands, which combined with the low rainfall of this area results in vir- tual desert conditions that are on the margin of true savannah. Stable dunes support deep-rooted grasses like Sporobolus consimilis. West of these dunes are the ‘short grasslands’ (zone II) characterised by dwarf growth forms of grasses such as Digitaria macroblephora and Sporobolus marginatus, and the grass-like sedges of the genus Kyllinga. The short growth forms are produced by the peculiarities of the soil. Rain leaches salts (calcium) out of the highly porous sandy top layers and redeposits them as a calcium carbonate hardpan about 1 m below the surface. This impermeable hardpan restricts root growth and the soil above it is highly saline and alkaline. Only plants that can with- stand these conditions and have short roots are able to survive here. Grass cover is sparse, ranging from 10% to 30%, but typically nearer 10%. The top layers of soil are susceptible to erosion, which can be made worse by grazing. West and south of the short grasslands the rainfall is higher and the soils become deeper. Because they are further from their volcanic origins, the soils are finer, hold water longer (therefore less leaching, and a less continuous hard- pan), and are less alkaline and saline. They therefore present a milder environ- ment for the vegetation. These are the intermediate grasslands (zone III). Taller growth forms of grasses found in zone II are found here and some patches of Andgropogon greenwayi can have 100% grass cover, although 30% is more typical. In both zone II and zone III there are extensive areas of the herbs Indigofera basiflora and Solanum incanum. The long grass plains (zone IV) have deep soils (about 2 m) of fine volcanic ash mixed with local material to form silty clay with low alkalinity and salin- ity, with no hardpan. Here there are grasses such as Themeda triandra and Pennisetum mezianum. Basal cover is now about 50%. Herbs are infrequent and there are no shrubs or trees. One of the features of both the interme- diate grasslands and the tall grasslands is the mosaic of Cynodon dactylon and D. macroblephora patches in a surrounding matrix of A. greenwayi and T. triandra. This mosaic is caused by termites (Sinclair 1979). When the ter- mites build their mounds, they tend to use weatherable subsoil material of high salinity and alkalinity. These termite mounds produce patches about 3 m across on which salt- and alkali- tolerant species such as C. dactylon and D. macroblephora can grow. In the tree and shrub savannahs of the north and west (zones V, VI and VII), the soils tend to be formed from the parent granite or quartzite rocks, although towards the east there is increasing volcanic ash content. Salinity 62 THE BIOLOGY OF AFRICAN SAVANNAHS and alkalinity are low and there is no hardpan, allowing trees to grow. Grass cover varies from 20% to 60%, with an average of 45%. The typical topogra- phy of these areas is a series of undulations which, together with the associ- ated soil and vegetation, is known as a catena or catenary sequence. Rain falling on an undulation gravitates from the top, down the slopes, to the bot- tom, or sump. Here it usually drains away along drainage lines or occasional- ly, if drainage is impeded, it accumulates. In its passage, the water carries with it soluble material and small soil particles. The top of the catena therefore becomes progressively leached of organic matter, and the finer soil particles, and comes to consist of shallow, coarse, sandy soil resembling the underly- ing parent rock. In contrast, the lower levels come to consist of deep clay, with high organic content and a high capacity for water retention. The slopes show a gradient between these two extremes. Tree species, in particular Aca- cia, prefer particular soil conditions and locations on the catena (Mduma, Sinclair and Turkingon 2007). Grass species also are adapted to the different conditions, so that small perennials and annuals like D. macroblephora and Chloris pycnothrix predominate on the sandy soils. On the slopes, the deeper soils and good drainage provide optimal conditions for Themeda triandra, which is one of the dominant grasses of these shrub and woodland savan- nahs. As Figure 2.13 shows, these western and northern areas are particularly 75 to 100% 50 to 74% 25 to 49% 1 to 24% 0% Figure 2.13 Distribution of the incidence of fire in the Serengeti between 1963 and 1972 (modified from Sinclair 1975). TTHE VEGETATION 63 TANZANIAKENYA Mara R. 0 20 km2 Mara R. eti R. rum G Ora ngo R. Mbalageti R. Seronera Duma R. Grassland lduvai R. O Acacia gerrardii Acacia tortilis, Acacia xanthophloea Commiphora trothae, Acacia tortilis, Acacia drepanolobium Acacia hockii Semi-decidous bushland, Hill woodland, Acacia clavigera, Terminalia mollis, Combretum molle Acacia drepanoloblum Commiphora trothae, Acacia xanthophloea Grassland, Acacia clavigera, Acacia drepanolobium Acacia nilotica, Acacia clavigera, Acacia seyal, Commiphora trothae, Acacia drepanolobium Acacia drepanolobium Grassland, Acacia drepanolobium Commiphora trothae, Acacia seyal, Acacia clavigera Balanites aegyptiaca, Acacia gerrardii, upland evergreens Acacia clavigera, Acacia tortilis, Acacia drepanolobium, Commiphora trothae Figure 2.14 Vegetation types within the Serengeti–Mara ecosystem (data from Herlocker 1975). prone to bush fires, and Hassan et al. (2007) have shown that fires can signifi- cantly change the sward structure. Using a series of experimental plots in the western corridor, they showed that total standing phytomass was consistent- ly lower on burned plots through the entire 1 year period after a fire. These savannahs are frequently regarded as fire-induced (see Figure 2.3). In some of the drainage areas, coarser grasses like Pennesetum mezianum predomi- nate, and in areas that remain flooded for long periods very tall grasses such as Panicum maximum grow. At the end of the western corridor there are soils derived from old beds of Lake Victoria (zone VIII) that support floodplain grasslands with few trees. Herlocker (1975) described the woody vegetation of these ‘tree and shrub’ savannahs of the north and west Serengeti (zones V, VI and VII). Two broad 64 THE BIOLOGY OF AFRICAN SAVANNAHS zones can be recognised. A thorn-tree zone (V) is dominated by 38 species of Acacia, but with 10 species accounting for 45% of the woody vegetation. The most extensive is the whistling thorn (A. drepanolobium), with A. clavigera and A. tortilis also common. Commiphora trothae is the only other common tree that occurs in this Acacia-dominated zone. North of the Grumeti River there is open woodland savannah (zones VI and VII). The dryer zone VI is essentially an evergreen thicket of Croton, Techlea and Euclea, surrounded by grassland with A. clavigera and A. gerrardii. Zone VII is open woodland savannah with the broad-leaved trees Combretum molle and Terminalia mol- lis. A more detailed view of the distribution of the woody vegetation, which is a simplified version of Herlocker’s map, is shown in Figure 2.14. The map of African savannahs in Figure 1.9 therefore shows broad areas where a particular savannah type predominates. Locally there is a patchwork of plant species assemblages and areas of other types of savannah. The ani- mals, particularly the large mammals so characteristic of African savannahs, are frequently more wide-ranging. We consider these next. 3 The Animals This chapter will take a brief look at the ecology and behaviour of a selection of animals that play a major role in the dynamics of the African savannah ecosystem. Of course not all species of animal can be mentioned. Even just listing all the animals that live in African savannahs would require a volume many times larger than this entire book. Most of the species discussed in this chapter will therefore be species mentioned in the studies and topics detailed in later chapters. Many of these are very prominent, often common, and con- tribute in a major way to the dynamics of the savannah system. Several spe- cies that are prominent and have important interactions with other species can be treated as a group, rather than individual species (e.g. grasshoppers and rodents). Many species are visually prominent, but as individual species do not greatly influence the savannah ecosystem (e.g. most birds). Below we describe very briefly the diversity of the savannah avifauna, and mention two quintessential savannah species, and two groups: vultures and weavers. Rep- tiles, although important savannah predators of small mammals and birds, are not included. They do not figure in the later chapters even though species such as the leopard tortoise (Geochelone pardalis), the Nile crocodile (Croco- dylus niloticus), and snakes such as the boomslang (Dispholidus typhus), the black mamba (Dendroaspis polylepis), the puff adder (Bitis arietans), and the Egyptian cobra (Naja haje) are familiar, common and widespread. Of course what makes African savannahs so unique and exciting is the wealth of ‘big game’: their mammalian megafauna. Many of these charismatic ani- mals have been extensively studied and feature prominently in the examples in subsequent chapters. These ‘large’ mammals are given extensive coverage. The insects There are thousands of insect species living in African savannahs, anden masse they can play an important role in the ecology of this biome. For The Biology of African Savannahs. Second Edition. Bryan Shorrocks & William Bates. © Bryan Shorrocks & William Bates 2015. Published 2015 by Oxford University Press. 66 THE BIOLOGY OF AFRICAN SAVANNAHS example, they remove the tonnes of dung that would otherwise completely cover large areas of savannah ecosystems. Anderson and Coe (1974) counted 16,000 dung beetles arriving at a 1.5 kg heap of elephant dung in East Africa, which was eaten, buried and rolled away in 2 hours. In West African savan- nahs, Cambefort (1984) estimated that beetles bury one metric tonne of dung per hectare per year. Another important role for insects in savannahs is that of pollinators. Of course this is an insect role that is not confined to these eco- systems. Globally, some 40,000 plant species have been recorded as impor- tant food resources for honeybees of the genus Apis, and races of the African honeybee are undoubtedly important pollinators for a wide variety of plants in savannahs systems. In the Mkomazi Game Reserve of northern Tanza- nia, Stone et al. (1998) recorded over 94 insect species from just 6 species of Acacia, including the widespread Apis mellifera. Termites are instrumen- tal in removing the dry, old, grass and therefore are important detritivores. Because they remove the dry combustible grass they influence the intensity of fires and therefore indirectly influence tree recruitment (see Chapter 2 and Chapter 5). Ants are frequently involved in mutualistic associations (Chapter 5) with Acacia trees, helping in their defence against herbivory. Grasshoppers (Orthoptera) are by far the most abundant insect herbivores in many savannah systems. As an example we can look at the work of Sin- clair (1975). He estimated the total annual grass production in three sepa- rate parts of the Serengeti ecosystem. These were long grass (characterised by Themeda triandra, Pennisetum mezianum and Sporobolus pyramidalis), short grass (characterised by Digitaria macroblephora, Andropogon green- wayi, Sporobolus pellucidus and S. spictus) and kopjes, or inselbergs, with grasses from the surrounding long grass area. He also estimated the total annual grass production consumed by each of four animal groups, and also destroyed by fire. His results are shown in Table 3.1. A total of 38 species of grasshopper were recorded in the Serengeti grass- lands. However, Mesopsis abbreviatus, Coryphosima stenoptera, Afrohippus Table 3.1 Total annual production and off-take in different parts of the Serengeti ecosystem Offtake by Long grassland Short grassland Kopjes kg h−1 yr−1 % total kg h−1 yr−1 % total kg h−1 yr−1 % total Annual grass production 5,978 4,703 5,978 Ungulates 1,122 18.8 1,597 34.0 122 2.0 Small mammals 69 1.2 4 0.1 259 4.3 Grasshoppers 456 7.6 194 4.1 484 8.1 Detritivores (Termites) 1,146 19.2 2,322 49.5 1,683 28.2 Burning 3,185 53.3 586 12.5 3,430 57.4 Data from Sinclair (1975). THE ANIMALS 67 taylori, Rhaphotitha species and Acrida species comprised 80% of the long grass population and A. taylori, Acrotylus elgonensis and A. patrrualis were the dominant species in the short grass areas. In total these grasshoppers removed about 1,134 kg of grass per hectare per year, amounting to about 7% of the total grass production, and approximately 26% of the green grass consumption. Perhaps one of the most profound effects of insects, on savannahs, involves their interaction with disease, humans and land clearance. Several mosquito genera and the tsetse fly Glossina( spp.) have played essential roles in the con- servation of African savannahs because these insects are vectors for malaria and trypanosomiasis, respectively. These two small insect vectors helped to protect savannahs from agricultural activities that spread across Africa during the colonial era. Settlers cleared vast areas of savannah vegetation in order to raise livestock and grow crops, and elephants, lions, wild dogs and many other species were hunted in order to protect these investments. In a relatively short time, the diverse ecology of the savannahs, that had evolved for millions of years, was reduced to rangelands for grazing livestock and fields for growing crops. In addition, disturbance of natural habitats by agri- culture is thought to have triggered the explosive expansion of mosquito populations across sub-Saharan Africa, causing a dramatic rise in malaria (Waters et al. 1991, Joy et al. 2003). In many moist, low-lying areas of Africa—for example in the lowveld regions of Mpumalanga and Limpopo Provinces in South Africa where Kru- ger National Park would eventually be established—Anopheles, the principle vector for the transmission of the malarial pathogen Plasmodium, helped to restrict agriculture (Mabunda et al. 2003). Paradoxically, high mortality rates in humans were sustained by the belief that it was an acacia tree, Acacia xanthophloea, and not Anopheles, that caused malaria. The belief that these ‘fever trees’ emitted toxic ‘malarial vapours’ helped to delay the implemen- tation of mosquito control methods, such as DDT spraying and sleeping under nets, at a critical time during the colonial expansion in Africa (Bengis et al. 2003). The success of mosquito disease vectors in limiting habitat destruction by humans is largely due to the complex digenetic lifecycle of Plasmo- dium spp. living within cold-blooded and warm-blooded hosts, refined by host jumping and coevolution over millions of years. (Rich et al. 2009, Waters et al. 1991). When a female Anopheles takes a blood meal, sporo- zoites leave the salivary glands, enter the host’s circulatory system and then rapidly enter liver cells to evade the host’s immune response, causing death and morbidity. This intracellular strategy has prevented vaccine development and so malaria continues to play an important role in pro- tecting savannahs, because of its devastating effects on human popula- tions (Murray et al. 2012). Given that hundreds of different pathogens 68 THE BIOLOGY OF AFRICAN SAVANNAHS cause malaria in mammals, birds and reptiles (Rich and Xu 2011) and these pathogens can jump from one host species to another, it seems unlikely that the proposed release of genetically modified mosquitoes that lack Plasmodium symbionts (Wang et al. 2012) will control the infec- tion of 300–500 million people annually that results in an annual global death rate of about 1.2 million people. Early settlers, hunters and poachers who invaded savannahs also encoun- tered the tsetse fly (Diptera: Glossinidae), the only genus that transmits one of at least 22 different species of the flagellated protozoan parasite Trypa- nosoma to humans, cattle and wildlife. Human sleeping sickness or human African trypanosomiasis (HAT) is caused by Trypanosoma brucei gambi- ense in West Africa and by Trypanosoma brucei rhodesiense in East Africa (Grab et al. 2009). Glossina palpalis is the major vector for the transmis- sion of Trypanosoma brucei gambiense. Several species of Glossina transmit trypanosome species that have varying degrees of virulence to livestock and cause a non-human trypanosomiasis called nagana. In striking contrast to humans and domestic livestock, wild populations usually have immunolog- ical resistance to the lethal effects of nagana, and function as reservoirs for the spreading HAT and nagana (Auty et al. 2012). Tsetse flies are highly suc- cessful because they exhibit pseudo-placental unilarviparity (females carry and nourish their larva for the entire developmental cycle), a remarkable type of viviparous development that enhances offspring fitness, because the female deposits a fully developed third instar larva that rapidly burrows into the soil to escape predation. Nourishment for the larva is provided by so-called milk proteins, GmmMGP (Attardo et al. 2008), sphingomyelinase (Benoit et al. 2012) and PGRP-LB (Wang and Aksoy 2012) that are made by the female milk gland and transferred to the single intrauterine larva during pregnancy. Intriguingly, phylogenetic analysis using the PAUP algo- rithm (Attardo et al. 2008) showed that GmmMGP is similar to marsupial lipocalin milk protein. The fly’s reproductive strategy therefore influences fly density, which in turn plays a pivotal role in determining trypanosomia- sis epidemiology. Fly densities have been studied using remote sensing and GIS of various types of habitats and the results indicate there is an inverse relationship between habitat fragmentation and fly density (Ducheyneet al. 2009). Another reason that HAT continues to be a major cause of human mortality in Africa is because trypanosomes evade the host immune sys- tem by using variable surface glycoproteins (VSGs), enabling the parasites eventually to enter the brain and cause sleep disruptions and eventual coma (Grab et al. 2009). The trypanosome ‘cloaking device’ that is used to evade the host’s immune system is very sophisticated at the molecular genetic level, and nearly half (806 out of a total of 1,700 genes) of the entire trypano- some genome encodes for these cell surface proteins (Berriman et al. 2005). Trypanosome cell surface proteins change during its developmental life- cycle. When trypanosomes are in the dividing (slender) and non-dividing THE ANIMALS 69 (stumpy) developmental stages they are found in the mammalian blood- stream and express VSG surface proteins. By contrast, dividing (procyclic) and (epimastigote) developmental stages live in the tsetse midgut and sali- vary glands and express different sets of proteins called GREET/EPs and BARPs. Non-dividing (metacyclic) stages of trypanosomes switch back to displaying VSG coat proteins in anticipation of entering the mammalian bloodstream (Savage et al. 2012). This anticipation of adaptive changes required for living in either a cold (tsetse) or warm (mammal) microhabitat is an excellent example of what is called a ‘predictive adaptive response’ that benefits pathogen fitness. Proof that this highly evolved mechanism is successful is clearly evident in the fact that at present, tsetse flies carry- ing trypanosomes are placing an estimated 60 million people in 36 African countries at risk of HAT (Savage et al. 2012). In conclusion, mosquitoes and tsetse flies have functioned as important guardians of African savannahs. The last remaining wild patches of African savannah may not exist today if not for these insects and their associated pathogens. The birds The avifauna of Africa is large and diverse, and although most birds have little impact on savannah systems as individual species, en masse they can be important, and certainly intriguing. Many species consume vast quanti- ties of plant seeds and insects, many frugivores are important dispersers of seeds, and many scavengers, such as vultures, disperse essential nutrients and important diseases. There is a rich ensemble of avian predators that con- sume an array of reptiles, birds and small mammals; and there are idiosyn- cratic individuals, like the ostrich and secretary bird, that are quintessential savannah sights. There are about 1,850 species of birds in Africa, although only about 1,450 are resident south of the Sahara. Two orders are endemic to sub-Saharan Africa, the Coliiformes (containing six species of mousebirds) and the Struthioni- formes containing the ostrich (Struthio camelus). Thirteen other taxonomic groups are also endemic to sub-Saharan Africa. These are the hammerkop (Scopus umbretta), shoebill (Balaeniceps rex), secretary bird (Sagittarius serpentarius), guinea fowls (7 species), turacos (18), wood-hoopoes (6), bush-shrikes (39), helmet-shrikes (9), rockfowls (2), sugarbirds (2), buffalo- weavers (2), parasitic weavers (9) and oxpeckers (2). Although not specific to Africa, several other groups are especially well represented. These include the bustards (16 species in Africa out of 22), honeyguides (11/13), larks (47/69), shrikes (55/74), Cisticola grass warblers (36/37) and ploceine weav- ers (101/113). 70 THE BIOLOGY OF AFRICAN SAVANNAHS Table 3.2 Distribution of African bird populations Habitat Total area (km2) Density (km2) Numbers (millions) 1 Desert 8,547,000 25 214 2 Sub-desert scrub savannah 2,530,000 125 316 3 Acacia grass savannah 4,671,000 1,500 7,007 4a Dry woodland savannah 3,869,000 2,500 9,673 4b Isoberlinia–Brachystegia woodland 3,979,000 3,500 13,927 4c Moist woodland savannah 1,304,000 4,500 5,868 5 Forest–savannah mosaic 1,502,000 6,000 9,012 6 Lowland rainforest 2,515,000 8,000 20,120 7 Other habitats, montane, etc. 1,082,000 2,000 2,164 8 Lakes, rivers, swamps 301,000 7,000 2,107 Modified after Brown, Urban and Newman (1982). Table 3.2, from Brown, Urban and Newman (1982), shows their estimates of the distribution of bird populations over various African habitats. Rainfor- ests are clearly the most productive for birds at an estimated 8,000 km−2 indi- viduals, followed by freshwater habitats at 7,000 km−2. However, savannahs are clearly not devoid of birds. Densities range from 125 km−2 individuals for the very arid scrub savannah, through 1,500 for the dry Acacia savannah, to 6,000 for the rich forest–savannah mosaic. Adding together all the habitats that in this book we call savannah (types 2–5 in Table 3.2), results in a figure of 2,565 km−2. Below we describe two characteristic savannah species. The ostrich (Struthio camelus) (height up to 2.4 m) is a fairly common bird in dry, open grass savannah, often at densities of 0.2–20 km−2. Four races/ species have been rec- ognised (Brown, Urban and Newman 1982, Zim- merman et al. 1996). It is mainly diurnal and roosts squatting on the ground at night. It eats both herbs and grasses, but prefers the former. Outside the breeding season it lives in rather open groups of 2–5 individuals. In the breed- ing season both males and females are usually soli- tary, and groups are made up of immature birds. ostrich Although in moister areas THE ANIMALS 71 it may be almost sedentary, in more arid areas it is more or less nomadic. There are two kinds of female. Major hens that mate with a territorial male and incubate the eggs in a nest, and minor hens that wander through many male territories and lay eggs in several nests. A clutch of 5–11 eggs is laid into a ground nest, by a major hen, the nest site having been chosen by the male. The female lays on alternate days, and incubation begins about 16 days after the first egg is laid. The major hens clutch may comprise less than 50% of the nest contents. Incubation period is 45–46 days. Broods from several nests may join together to form one large crèche, guarded by several adults. The secretary bird (Sagittarius serpentarius) (height up to 1.3 m) is an unmistakable resident of almost all African savannahs. It roosts and nests in trees. When hunting during the day it walks steadily through grass- land and catches its prey with its beak, but some- times it kills with its feet. It eats insects, amphib- ians, lizards, snakes, young birds and small mammals. Although it is often viewed as a snake killer, it probably feeds mainly on grasshoppers and small rodents. In the breeding season, pairs secretary bird defend large home ranges (5000–6000 ha). The nest is a large flat structure of twigs and 1–3 pale bluish or white, eggs are laid at daily intervals. The incu- bation period is 43–44 days. The young leave the nest at 75–85 days, without parental stimulus. Although decreasing in many areas, due to habitat loss, it is still locally common. Vultures Although there are about 97 bird of prey species in Africa (Zimmerman et al. 1996), one of the most obvious and recognisable groups in savannahs is the vultures, with 10 African species, 8 of which are sub-Saharan. They are large, open country birds, adapted to carrion-feeding and to extended peri- ods soaring on long, usually broad, wings. Except for the lammergeier, feath- ering is reduced or absent on the face, neck and legs. In many areas, most of these species coexist and are frequently found together at a carcass, where they compete for food (Kruuk 1972). However, their different body sizes and 72 THE BIOLOGY OF AFRICAN SAVANNAHS beak structures allow them to partition the carcasses they locate. The small, black and white, Egyptian vulture (Neophron percnopterus) (body weight 1.8 kg), inhabits the arid savannahs of the Sahel, down the east of Africa to northern Tanzania. It also occurs in northern Namibia and Angola. It feeds on carrion and meat scraps of any kind, and bird’s eggs. It cannot compete at carcasses with the larger vultures and waits on the outskirts to snatch scraps. It nests on cliffs. The other small species, thehooded vulture (Necrosyrtes monachus) (body weight 1.9 kg) is very widespread, occurring in most of the sub-Saharan savannahs. It also feeds on all kinds of carrion, and is a scavenger in towns. Several pairs nest together, in trees. Both these small species have long, thin beaks that are relatively weak. In addition to scraps of meat from carcasses, they take termites, beetles, snakes, and lizards. They usually fly at low altitudes and are non-territorial. The three, large, griffon vulturesGyps ( species) are all widespread and specialise on the soft flesh and intestines of large dead mammals. Their beaks are long, with a sharp cutting edge. The tongue is barbed for gripping soft tissue. They are non-territorial, fly at high altitudes, rely on soaring, and fly long distances each day. They frequently collect in large numbers at suitable carcasses. The African white-backed griffin (G. africanus) (body weight 5.3 kg) is the commonest large vulture, found in all sub-Saharan savannahs from the Sahel to southern Africa. Like the other two species it is gregarious at roosts, and at carrion, but more soli- tary when breeding. The other twoGyps species have a more restricted range, but are still common. They are both colonial cliff breeders. Ruppell’s griffon vulture (G. rupellii) (body weight 7.4 kg) is found south of the Sahara from Senegal to Sudan, Kenya and Tanzania and roosts at night on cliff faces. This species has the high- est credible altitude record for any bird, one having been killed by a jet aircraft at 11,300 m. The Cape griffin vulture (G. coprotheres) is restricted to south- ern Africa. These last three species have large beaks, enabling them to tear the meat from large carcasses. Although mainly eating flesh, they can also eat sinews, bones and skin, parts normally left by the vulture 1 griffin vultures. They THE ANIMALS 73 all appear to be ter- ritorial. The fleshy pink head and neck E with large fleshy lappets make the lappet-faced vulture (Torgus tracheliotus) B (body weight 6.2 kg) one of the easiest to R recognise. It nests singly in trees. The widespread white- headed vulture (Trigonoceps occipi- C talis) eats mostly carrion, but is also suspected of killing vulture 2 hares and gazelle calves and raids fla- mingo colonies. It nests singly in trees such as Acacia. Recent GPS tracking experiments have shown that lappet-faced vultures find food more easily than white-backed vultures despite the fact that white-backed vultures are more common than lappet-faced vultures (Spiegel et al. 2013). Thelammergeier or bearded vul- ture (Gypaetus barbatus meridionalis) is a graceful, long-winged, soaring bird that will consume almost every scrap of meat, skin, or bones from a carcass. Its tongue is adapted for removing marrow from long bones, and its most remark- able feeding habit is dropping bones on rocks to split them. The distribution of these eight vultures are shown in the two adjacent maps. Vulture 1 shows the distributions of the hooded vulture, lappet-faced vulture, white-backed griffin vulture, white-headed vulture. They have distributions that cover virtually the same areas, although while the hooded vulture and white-backed grif- fin vulture are essentially continuous in their distribution, the lappet-faced vulture and white-headed vulture are rather patchy in theirs. Vulture 2 shows the distribution of the lammergeier or bearded vulture (B), Cape griffin vul- ture (C), Ruppell’s griffin vulture (R), and Egyptian vulture (E). Weavers Another large group of birds that are often viewed as quintessential savannah species is the weavers (family Ploceidae). They are small passerine birds related to the finches, with a thick bill and strong legs. Plumage is very variable and there are seasonal changes in some species. Their diet includes seeds of grasses and herbs, buds and insects. They are in fact not confined to savannahs, with their habitats varying from dry bush to marshes and forest. As their name implies, 74 THE BIOLOGY OF AFRICAN SAVANNAHS these birds weave remarkable nests from vegetation (e.g. green grasses and reeds) and one of the classic pictures of savannahs is the sight of an acacia tree with its dozens, or even hundreds, of little straw nest bags hanging from the branches. In fact their breeding habits are quite diverse, with some species being essentially solitary (e.g. the vitelline masked weaver Plocus velatus uluensis) and others being highly colonial (e.g. the black-headed or village weaver Plocus cuc- ullatus). But all construct covered nests. The ploceine weaver family is primarily African, with 101 of the world’s 113 species found in Africa. The savannah red- billed quelea (Quelea quelea) is reputed to be the world’s most abundant wild bird species, with an estimated adult breeding population of 10 billion (Fry and Keith 2004). The mammals In this section we look at the ecology and behaviour of several individual species of large mammals—ungulates, carnivores, and primates—and con- sider rodents as a group of species. Modern molecular opinion (Macdonald 2001; Murphy et al. 2007), based on both nuclear and mitochondrial genes, now place placental mammals into four groups, or clades: the Euarchontoglires, Laurasiatheria, Xenarthra and Afrotheria. The Euarchontoglires include the primates, insectivores, and rodents. Members of the Xenarthra are found entirely in the New World and include such mammals as anteaters, sloths and armadillos. The Afrotheria are interesting. They are entirely African and contain a very diverse group of species which include elephants, hyraxes (conies), the aardvark, golden moles and elephant shrews. Within the Laurasiatheria are the carnivores (e.g. cats, dogs, badgers, and otters), ungulates (e.g. giraffe, antelopes, horses, and rhinoceroses), whales and dolphins, bats and the enigmatic pangolins. The Laurasiatheria, Xenartha and Afrotheria almost certainly share a common ancestor (they are monophyletic), but it remains uncertain if the Euarchor- data and Glires have a common, or separate, ancestral root. The term ‘ungulate’ is used for those groups of mammals that have hooves. A hoof consists of a hard or rubbery sole and a hard wall formed by a thick horny nail (keratin) rolled around the tip of the toe. The primitive mam- malian limb ends in five digits, all of which are placed on the ground during locomotion. Ungulates have reduced this basic number of digits and also lengthened and compressed the metapodial bones (the long bones in human hands and feet). Ungulates therefore walk on tiptoe, rather like a ballerina. In addition there is restricted movement in the joint surfaces so that the main movement of the limbs is forwards and backwards, rather than sideways. This suite of ungulate limb characteristics appears to be associated with a terrestrial, herbivorous lifestyle, often in open savannah habitats, in which THE ANIMALS 75 both fast and sustained galloping locomotion are useful. Ungulates were tra- ditionally classified into two different groups, the perissodactyls and artio- dactyls, based on foot morphology. However, more recent analysis of two types of genetic mobile elements called SINEs (short interspersed elements) and LINEs (long interspersed elements) that are inserted into DNA over evo- lutionary time suggested that whales have an origin within the artiodactyls, and in particular are closely related to hippopotamus species (Nikaido et al. 1999). Furthermore, in 2001, a crucial bit of evidence came to light that con- firmed this molecular finding. One of the most diagnostic features found in the skeleton of artiodactyls is the shape of the astragalus bone in the ankle, which somewhat resembles that of a double pulley, and gives the foot greater flexibility. Exactly this type of bone was found to be part of the skeleton of two archaeocetes (primitive whales). With both molecular and morphologi- cal evidence now supporting a cetacean–artiodactyl relationship, the tradi- tional order Cetacea was merged with Artiodactyla, thus forming the now widely accepted order Cetartiodactyla. However, the older way to group ungulates, based on foot morphology, is still a useful and sensible group- ing for a terrestrial ecosystem such as African savannahs, and is still widely employed in field guides and also in the recentmagnus opus, Mammals of Africa (Kingdon et al. 2013). Ungulate feet have either an odd (Greek: perisso-) or even (Greek: artio-) number of toes (Greek: dactyla). In artiodactyls (hippoptamus, giraffe, ante- lopes, buffalo) the axis of the foot passes between digits 3 and 4, while in perissodactyls (zebras and rhinoceroses) it passes through the third middle digit (Figure 3.1). Early ungulates had four hooves (digits 2–5), a condition that persists in pigs and hippos. The number of hooves has been reduced to one pair in the horses although the vestiges of digits 2 and 5 are still present as the false hooves. Ungulates have also evolved two rather different systems for converting cel- lulose, a major constituent of all plant tissues, into digestible carbohydrates. These two systems are called hindgut fermentation and rumination. In both systems, the conversion from cellulose to digestible carbohydrate is not directly brought about by the ungulate itself, but by gut microrganisms that achieve the breakdown by a process of fermentation. The two systems are closely associated with the two groups already mentioned, perissodactyls and cetartiodactyles. The perissodactyles (e.g. zebra and rhino) are hindgut fermenters, whereas most of the cetartiodactyles (e.g. giraffe, buffalo, ante- lopes and gazelles) are ruminants. Only pigs, hippos, camels and cetaceans in the cetartiodactyles are not ruminants. Figure 3.2 shows a simplified comparison of the two systems. In the hindgut fermenters, who have a sim- ple stomach like ourselves, food is chewed once and plant cell contents are digested in the stomach. The food is then passed to the caecum (our appen- dix is a reduced vestige of this), and colon, where the cellulose of the plant cell wall is fermented by microrganisms. Ruminants, apart from the four 76 THE BIOLOGY OF AFRICAN SAVANNAHS (a) Calcaneum Astragalus Tarsus Navicula Cuboid Metapodials Phalanges 5 1 4 2 Digit 3 (b) (c) (d) (e) (f) 5 2 5 2 4 2 4 3 3 3 4 3 43 Figure 3.1 Ungulate feet showing modification from the primitive mammalian limb. (a) primitive mammalian foot, (b) rhinoceros, (c) horse, (d) hippopotamus, (e) pig, (f) buffalo (modified from MacDonald 2001). species of primitive ruminants called chevrotains, have a more complex four- chambered stomach. Vegetation is swallowed and initially enters the first chamber or rumen. The coarsest plant particles (the cud) float to the top of the rumen, are regurgitated by its contractions, and rechewed in the mouth (rumination). The rechewed food is then returned to the rhythmically con- tracting second stomach, or reticulum, for ‘sorting’. Food particles are even- tually pumped into the third stomach, or omasum (also known as the ‘book organ’ because of the leaf-like plates that line it), and filtered. The digestive process is completed in the fourth or ‘true stomach’ (abomasum) and small intestine, and additional fermentation and absorption takes place in the cae- cum. In ruminants therefore cellulose digestion takes place before normal digestion, in the rumen. In hindgut fermenters, cellulose digestion takes place after normal digestion, in the caecum. Ruminants are therefore able to extract more nutrients from their plant food (cellulose utilisation in a cow is about 80%), but pass it through their digestive system very slowly (rate of passage in a cow is about 80 hours). They require high-quality food. Hindgut THE ANIMALS 77 Hindgut fermentation Small intestine Cecum Food chewed once Stomach Colon Fermentation of Absorption of cellulose sugars fermentation products Rumination Reticulum Omasum Abomasum Colon Food chewed twice Rumen Small intestine Cecum Figure 3.2 Stylised comparison of ‘hindgut fermentation’ (zebra) and ‘rumination’ (gazelle) digestion (modified after MacDonald 2001). fermenters are less efficient (cellulose utilisation in a horse is about 50%), but pass food through their digestive system faster (rate of passage in a horse is about 48 hours). They can therefore use lower-quality food. Much more fibre is left undigested in the hindgut fermenter, as can be observed in the coarse dung of zebra, rhino, elephant, and of course horse. In comparison, the dung of ruminants, such as the buffalo, wildebeest, and domestic cow, is fine grained. Ruminants can also recycle urea (normally a waste product excreted in the urine) and can therefore conserve water. In the ruminant stomach, the pro- tein in their plant food is converted to ammonia, and transported in the blood to the liver. Here it is converted to urea and transported back to the stomach where the microrganisms use it as food. In addition to their role in cellu- lose fermentation, these microrganisms excrete fatty acids which are used by their host. They also provide their hosts with protein, as they pass down the ruminant’s digestive system, mixed with the vegetable food (Demment and 78 THE BIOLOGY OF AFRICAN SAVANNAHS Van Soest 1985). However, although ruminants have a more efficient system for dealing with the cellulose component of plant food, and gain protein and fatty acids from the microrganisms themselves, their one major disadvan- tage is that they cannot process food quickly. The more fibrous the food, the longer the process can take. When the protein content of their plant food falls below about 6%, ruminants cannot process their food fast enough to maintain their weight and condition (McNaughton and Georgiadis 1986). This presents a particular problem in the savannah dry season, when the percentage of crude protein in many grasses drops below this value (Table 3.3). It is therefore important for many ruminants to be selective in their feeding (see Chapter 5). Another feeding distinction that has arisen in ungulates is that between grazers (roughage eaters), browsers (selective eaters) and mixed feeders (Hofmann 1968, Estes 1991). Grazing ungulates feed on monocotyledon- ous plants (essentially grasses, but also reeds and rushes), while browsers feed on dicotyledonous plants (mainly tree foliage, but also fruits and herbs). Mixed feeders (e.g. elephant and impala) are both grazers and browsers, often switching their foraging activity between the wet season (grass) and the dry season (foliage). Some small browsers, such as klipspringer and bush- buck, are extremely selective feeders, choosing food with the highest protein content. They tend to feed in short bouts and rarely fill their rumen more than half full before stopping to ruminate. Larger browsers such as giraffe, gerenuk and lesser kudu need to take in more food and cannot afford to be as choosy. They often have larger and more subdivided stomachs designed to slow down the passage of food and allow more time for fermentation. In contrast to these selective browsers, bulk grazers (Estes 1991) such as water- buck, wildebeest and buffalo, tend to keep eating until the rumen is full, often regardless of grass quality. The different stomach chambers are more Table 3.3 Percentage crude protein in two grass species in the Serengeti, at different growth stages Growth stage and grass species Green leaf Dry leaf Stem and sheath Young green growth Pennisetum mezianum 20.0 – 15.0 Digitaria macroblephora 16.6 – 10.5 Mature Pennisetum mezianum 9.8 4.3 4.4 Digitaria macroblephora 10.0 4.0 3.8 Dry season Pennisetum mezianum 6.8 3.9 3.1 Digitaria macroblephora 11.1 4.9 2.5 Data from Duncan (1975). THE ANIMALS 79 muscular and subdivided in order to churn, pump, sieve and filter the ingest- ed food. The omasum, which in the selective browsers is simply a straining chamber, is highly developed. So-called roughage grazers (Estes 1991), such as hartebeest and topi, are frequently more selective in the parts of the grass they eat and can consequently often survive on poorer grass swards. These dietary differences may aid the coexistence of herbivores (see Chapter 5). Other parts of the ungulate body that show adaptation to their herbivorous way of life are the teeth and jaw muscles. The generalised placental mammal has four types of teeth. The position and number of these teeth can be rep- 2.1.2.3 resented by a dental formula, which in humans is 2.1.2.3 . The upper numbers represent the teeth on one side of the upper jaw, and the lower numbers the teeth on one side of the lower jaw. The four successive numbers indicate the number of incisors, canines, premolars and molars. Typical formulae for a 0.0.2.3 0.0.3.3 ruminant would be 3.1.2.3 or 3.1.3.3 . The upper incisors and canines have been lost and the lower canines look like incisors, forming a row of eight chisel- shaped teeth. The cheek teeth (premolars and molars) are adapted for grind- ing vegetable matter and their grinding surface has complex folds and sharp ridges. There is typically a gap, called the diastema, between the cutting inci- sors/canines and the grinding cheek teeth (Figure 3.3, giraffe and bushbuck). In hindgut fermenters the presence and absence of incisors and canines is more variable, but all have the grinding cheek teeth. In male zebra the dental 3.1.3.3 formula is 3.1.3.3 (Figure 3.3), but in females the canines are rudimentary or absent. In both black and white rhinos, all incisors and canines are absent and 0.0.3.3 the formula is 0.0.3.3 (Figure 3.3). In the hippopotamus the canines and inci- sors are continually growing and the lower canines are enlarged as tusks (Fig- 2.1.3.3 ure 3.3). The adult dental formula is 2.1.3.3 but while the cheek teeth have the usual grinding function, the protruding position of the incisors and canines renders them useless for feeding. They are used as weapons only, and the cropping of grass is taken over by the hard edges of the lips. In the pigs, as represented by the warthog (Figure 3.3), the cheek teeth have rather flat uncomplicated surfaces. The number of teeth is more variable in this species 1.1.3.3 than in any other ungulate. A full complement for adults is 23− .1.2.3 , but in old adults it may reduce to 0.1.0.1 . The canines are enlarged as persistently grow- 0.1.0.1 ing tusks. As mentioned earlier, the cheek teeth of most ungulates (apart from the pig family) have complex grinding surfaces, with many folds and sharp ridg- es. Vegetation in general, and grass in particular, is very tough material to grind and over a lifetime of continual abrasion these cheek teeth are very likely to be worn down to the gum. To compensate for this, some ungulates (mainly grazers) have evolved high-crowned (hypsodont) cheek teeth. A simple way to produce a high-crowned tooth would be by simple elongation of the tooth—a higher tooth would give longer grinding life. Some early fossil mammals actually took this path but it is relatively unsuccessful at 80 THE BIOLOGY OF AFRICAN SAVANNAHS Male giraffe bushbuck common zebra black rhinoceros hippopotamus warthog Figure 3.3 Skulls of six ungulates. Giraffe and bushbuck are ruminants, the other four are non- ruminants (from Skinner and Smithers 1990). prolonging the grinding life of the tooth. Once the hard enamel surface of the tooth is worn away, most of the wear comes on the soft and easily abrad- ed internal dentine, with the resistant enamel remaining only as a thin rim. In the modern hypsodont tooth the height is attained by a growth of each cusp or ridge on the tooth. These hard peaks are fused together by a growth of hard cement over the entire tooth surface, producing a tall tooth very resistant to grinding. THE ANIMALS 81 This problem of extreme tooth wear in grazing animals is solved in an entirely different way by the elephant. The infamous elephant’s tusks, fre- quently responsible for the animal’s premature death at the hand of hunters, are upper incisors that grow continuously throughout life. At 60 years they can average 61 kg in a bull and 9.2 kg in a cow. Like many other grazers, elephants have six cheek teeth (all molars) in each quarter of the mouth, but they appear in succession throughout the life of the elephant. Only two molars normally occur together on each side of the upper and lower jaw, and are in use at the same time. As each successive tooth comes into use, it moves forward in the jaw, becoming worn and breaking down towards its forward edge, the roots being reabsorbed. The next one then moves for- ward from the back of the jaw to replace it. Laws (1966) shows that the six molars develop and erupt at the following approximate ages: molar 1 (1 year), molar 2 (2 years), molar 3 (6 years), molar 4 (15 years), molar 5 (28 years) and molar 6 (47 years). By the age of about 60 only the last molar remains. Once this last grinding tooth breaks down, the elephant dies of starvation. Ungulates dominate terrestrial communities, making up about 80% of mam- mal species over 50 kg. The great majority of these ungulate species are artio- dactyls (about 92%). This percentage is reflected in the African fauna, with only three zebra and two rhinoceros species that are perissodactyls. In the next two sections we look at some of these successful ungulates that inhabit African savannah ecosystems. However, a word should be said about distri- bution maps. In all field guides (e.g. Estes 1991, Kingdon 1997, 2004, Stuart and Stuart 1997) the distribution of an African species is shown as a shaded area on a map of Africa. Intriguingly, these shaded areas often differs in detail in different Field Guides. The probable reason for this is that a cross-section through a species distribution is not likely to be rectangular in shape. Spe- cies are not usually equally abundant across their distribution range, with a sudden drop to zero at the edge. Cross-sections across the shaded area of these maps are more likely to look like a bell-shaped curve (often skewed), with high abundance towards the centre of the range and a gradual decrease towards zero at the edge (unless the edge is the sea). Some authors give a distribution map that includes all the range, some authors (we suspect) give only that area in which you are quite likely to see the animal. In other words, they miss out the edge of the species distribution. The distribution maps in the species accounts below are a compilation from Estes (1991), Stuart and Stuart (1997), Kingdon (1997), Kingdon et al. (2013) and for the ‘antelopes’, East (1999). In the species accounts, the conservation status is also stated. These are IUCN categories indicating the threat to a species. The categories are: • extinct—no doubt that the last individual has died • extinct in the wild—only known in captivity 82 THE BIOLOGY OF AFRICAN SAVANNAHS • critically endangered—a species faces an extremely high risk of extinction in the immediate future • endangered—a species faces a high risk of extinction in the near future • vulnerable—a species faces a high risk of extinction in the medium-term future, near threatened—near to qualifying for a threatened category in the near future • least concern or not threatened—none of the above. Ungulates: non-ruminants Zebras (Family Equidae) There are three spe- cies of zebra in Afri- ca but only two are detailed here. The common, or plains, zebra (Equus quag- ga) (Figure 3.4a) (shoulder height 1.3 m, weight 175– 340 kg) is Africa’s most abundant and widespread wild horse. The species includes several subspecies, which in older books are fre- quently called spe- cies. These include common zebra Grant’s, Crawshay’s, Chapman’s and the Damara zebra. They can be found in a wide range of savannah habitats, from short or tall grassland, to open woodland. They are adaptable grazers, taking the grasses that are most frequent. They also occasionally browse. In the great migrations of the Serengeti they are the first species to move, often graz- ing the tall flowering grasses ahead of the wildebeest and Thomson’s gazelle. They live in small family groups (harems) of four to six animals. These com- prise a male, with several females and their young. Males that have not got a family group join together, in bachelor herds. They are not territorial, and rather nomadic. Females produce a single foal (30–35 kg), after a gestation period of about 375 days. Birth usually occurs at the start of the rainy season when green grass is available. It is not threatened, and the population trend is stable. THE ANIMALS 83 Africa’s largest equid is Grevy’s zebra (Equus grevyi) (Figure 3.4b) (shoulder height 1.5 m, weight 350–430 kg). It was the ‘hippotigris’ paraded in Roman circuses, in 211–217ad, pulling carts. Less water dependent than the com- mon zebra, it occupies dry savannahs in northern Kenya. It is predominantly a grazer, often mak- ing use of a tough grass Pennisetum schimperi, that is not fully used by other grazers. Up to 30% of the diet can be browse. This zebra is territorial and migratory. Mature males (over 6 years) establish core territo- ries as large as 12 km2 in the wet season, which they actively defend against other males. Ten or fewer mares, with their young, makeup Grevy’s zebra nursery herds which are allowed to pass through the male territories unchallenged. During the dry season Grevy’s zebra are often forced to roam over vast home ranges, up to 10,000 km2. Birth occurs at any time of the year, but peaks at the start of the wet season. After an exceptionally long gestation period of about 400 days, a single foal is produced. Its conservation status is endangered, and it is estimated to have declined by more than 50% over the past 18 years. Rhinoceroses (Family Rhinocerotidae) The hook-lipped orblack rhinoceros (Diceros bicornis) (Figure 3.4c) (shoul- der height 1.6 m, weight 800–1,100 kg) is not black but does have a hook-lip which is used to grasp leaves and twigs. The two horns are made of matted hair-like filaments and are attached to the skin rather than the bone. Once very common in savannah habitats, it is now found only in protected areas. It requires savannahs with shrubs and trees to a height of 4 m, to provide both food and shade. It is a browser, taking a wide range of leguminous herbs and shrubs. Some 200 species from 50 families have been recorded in their diet. A female and her young, is the basic social unit. The males are usu- ally solitary but it is not clear if they have territories or simply home rang- es. In Natal, radio-tracked bulls had ‘territories’ of 3.9–4.7 km2, and female (a) (b) (c) (d) (e) (f) (g) (h) Figure 3.4 Savannah mammals. (a) common or plains zebra, (b) Grevey’s zebra, (c) black rhinoceros, (d) warthog, (e) hippopotamus, (f) reticulated giraffe, (g) African buffalo, (h) common eland (photgraphs a, b, d, e, f, g, h, by Bryan & Jo Shorrocks, photograph c by Tory Bennett). THE ANIMALS 85 ranges were 5.8–7.7 km2 (Hitchins 1968, 1969). Bulls and cows only come together briefly for mating. After a gestation period of some 450 days one calf (40 kg) is dropped. The calf may stay with the mother for between two and four years. Black rhinos are critically endangered, having undergone cata- strophic declines in numbers over the past few decades. Only about 3,000 remain in protected areas. The square-lipped or white rhinoc- eros (Ceratotherium simum) (shoulder height 1.8 m, weight ♂ 2000–2300 kg, ♀ 1400–1600 kg) is not white; it is much the white rhinoceros same colour as the black rhinoceros. blach rhinoceros 86 THE BIOLOGY OF AFRICAN SAVANNAHS Figure 3.5 Head of square-lipped or white rhinoceros, a grazer (left), and the hook-lipped or black rhinoceros, a browser (right), showing the different shape of the lip (from Dorst and Dandelot 1972). The name is derived from the Dutch ‘weit’ (wide) referring to the mouth (Figure 3.5). It is twice the weight of the black rhinoceros and the second- largest land mammal after the elephant. It prefers areas of short-grass savan- nah with access to trees for shade. It is a selective grazer with a preference for short grasses such as Cynodon, Digitaria, Heteropogon, and Chloris spe- cies. Females and their offspring occupy large (4–60 km2), overlapping home ranges and sometimes associate in larger groups. About one-third of males defend territories that can be quite small (3 km2) but their size depends on the productivity of the habitat and population density. The remaining males live as satellites on the territories. After a gestation period of some 480 days a single calf (40 kg) is dropped. A calf stays with the mother for between two and three years. The conservation status of the southern African race is lower risk. The northern race isnear threatened, although the IUCN population trend is increasing. Pigs (Family Suidae) Africa has six species of pig, only two of which are mentioned here. The common warthog (Phacochoerus africanus) (Figure 3.4d) (shoulder height: 60–70 cm, weight: ♂ 60–105 kg, ♀ 45–70 kg) is the most abundant member of the family, and is distributed widely throughout the savannahs of Africa. In the wet season it grazes on grass, particularly species such as Sporobolus, Cynodon, Panicum and Brachiaria. In the dry season it eats underground rhizomes of perennial grasses and sedges, and bulbs and tubers. Character- istically it frequently ‘kneels’ when grazing. They are heavily preyed upon by carnivores such as lion, leopard and cheetah. When threatened by predators they retreat to burrows, previously excavated either by them or by another species such as the aardvark. These burrows are used at night. The basic social THE ANIMALS 87 unit, or sounder, consists of a female with her young, sometimes with a second female. Boars only accompany sounders containing oestrus females, oth- erwise adult males are usually solitary. Both sexes tend to remain in their natal area, so that clans of related sounders develop, often using the same network of burrows. These clan areas average warthog about 4 km2. The young are born in a burrow, after a gestation period of 160–170 days. Litters average two or three, but can go up to eight. Weaning takes between 9 weeks and 5 months. The desert warthog (P. aethiopicus) is slightly smaller than the common warthog, but otherwise indistinguishable in the field. It is confined to the arid savannahs of Somalia and north-east Kenya. It also eats grass, and roots. Both warthog species are least concern on the IUCN list. Hippopotamuses (Family Hippopotamidae) Africa has two members of this family. Thepygmy hippopotamus (Hexapro- todon liberiensis) is not a savannah species and has a restricted range in West Africa. It will not be detailed here. The more familiar, and much larger, common hippopotamus (Hippopotamus amphibius) (Figure 3.4e) (shoul- der height 1.5 m, weight ♂ 1,000–3,000 kg ♀ 1,000–2,500 kg) is commonly observed in, or near, rivers that run through savannah areas. Their skin is quite unique. They have a thin epidermis with no sweat glands, and lose water at several times the rate of other mammals. Because of this they spend most of the day in water, however, their skin produces a natural pink ‘sunscreen’ that helps to protect their skin. At night they emerge and commute to their feeding grounds, which may be a few hundred metres, to several kilometres away. Here they graze on grasses (c.40–60 kg per animal per night), frequent- ly reducing the grass in such areas to ‘grazed lawns’. At high density, hippos can have a devastating effect on the vegetation and soils bordering savan- nah rivers. Both creeping and tussock grasses are eaten, including Cynodon, Panicum, Brachiara, Themeda, Chlois and Setaria species. They usually live in 88 THE BIOLOGY OF AFRICAN SAVANNAHS groups of 5–15 indi- viduals. These herds are usually domi- nated by an adult male who holds tenure over a num- ber of females and their young. Males vigorously defend their group against intruding males and serious fights can take place. Females conceive at about 9 years, and calve at 2 year intervals. Mat- ing takes place in the water. A single calf hippopotamus (30 kg) is born after an 8 month gesta- tion period, usually in the rainy season. They have declined over the last 50 years, although estimates of numbers are in excess of 150,000. Nonetheless, the common hippopotamus is listed as vulnerable. Zambia is thought to have the largest national herd, with an estimated 20,000 along the Luangwa river. Ungulates: ruminants Giraffes (Family Giraffidae) Thegiraffe (Giraffa camel- opardalis) (Figure 3.4f) (shoulder height 2.6– 3.5 m, height to top of head 3.9–5.2 m, weight W 970–1,400 kg) is the tallest K animal. The name prob- N Re able comes from the Ara- Ro bic zarāfa, which in turn M comes from the Ethiopic zarat meaning ‘slender’. Considerable uncertainty T surrounds the validity A and geographical limits S of the described subspe- cies; there are between giraffe THE ANIMALS 89 six and nine, depending on the author. Following Dagg (2014), we indicate nine subspecies: the West African giraffe (W) G.( c. peralta), the Kordofan giraffe (K)(G. c. antiquorum), the Nubian giraffe (N) G.( c. camelopardalis), the Rothschild’s giraffe (Ro) (G. c. rothschildi), the reticulated giraffe (Re) (G. c. reticulatus), the Masai giraffe (M) (G. c. tippleskirchi), Thornicroft’s giraffe (T) (G. c. thornicrofti), the southern, or Cape, giraffe (S) (G. c. giraffa), and the Anglolan, or smoky, giraffe (A) (G. c. angolensis). They like dry savan- nahs where Acacia, Commiphora and Terminalia are abundant trees. They are browsers, and have been recorded feeding from over 100 species of plant. However, Acacia and Combretum form the bulk of the diet. Most of the diet is made up of leaves, flowers, shoots, and some pods. These are stripped from the tree by the powerful 45 cm long tongue and by the lips. Giraffes form loose, open, herds of usually between 4 and 30 individuals, called fis- sion–fusion societies (Shorrocks and Croft 2009, Bercovitch and Berry 2013, Carter et al. 2013a, Carter et al. 2013b; see Box 3.1). They do not defend ter- ritories, but occupy large home ranges of 5 to 650 km2, depending on habitat quality and giraffe density. However, like many savannah herbivores they disperse widely during the rains and concentrate along watercourses in the dry season, where high-quality food continues to be available. A single calf (100 kg) is born after a long gestation period of 15 months. The conserva- tion status of the different subspecies varies. The Masai (<37,000) and the two southern giraffes, Cape (<12,000) and Angolan (<20,000) giraffe are not endangered. Thornicroft’s giraffe (<1,000) has low, but stable, numbers. The west African (<300), Kordofan (<3,000), Nubian (<650), Rothschild’s (<1,100), and reticulated (<4,700) are all declining and regarded as endan- gered. Population estimates are taken from Tutchings et al. (2013). Buffalo (Family Bovidae, Tribe Bovini) There are two subspecies of African buffalo, the largersavannah buffalo (Syncerus caffer caffer) (Figure 3.4g) (shoulder height 1.4 m, weight ♂ 320 kg, ♀ 550 kg) and the smaller forest, or red, buffalo (S. c. nanus), found in the rainforests of the Congo basin. Only the former subspecies is detailed here. The savannah buffalo prefers open wooded savannah, with abundant drinking-water, but also inhabits montane forest in areas such as the Aberd- ares and Mount Kenya in East Africa. They are predominantly (95%) bulk grazers, taking a wide selection of grass species. They are found in herds of 20–2,000, depending on habitat quality, although the larger herds frequent- ly split into smaller family clans for part of the time. These clans comprise several related cows and their offspring, to which are attached a number of adult and subadult bulls. Within these family groups, both adult males and females establish dominance hierarchies. Old bulls are often solitary, and rather sedentary (3–4 km2). Herds occupy clearly defined home ranges which rarely overlap. Most calves (40 kg) are born in the wet season, after a gestation period of about 11 months. Birth intervals of 2 years are normal. 90 THE BIOLOGY OF AFRICAN SAVANNAHS Box 3.1 Network analysis Social network analysis has been widely used in the social sciences for many years but has only been used by researchers of animal behaviour over the past decade (reviewed by Krause, Croft, and James 2007, Croft, James, and Krause 2008, Wey et al. 2008). However, such analysis has now allowed the detection of subtle interactions in the social organization of many animal species that had not previously been described, and the range of network measures now available make it a very powerful analytical tool. In savannah species there has been revealing work on Grevy’s zebra (Equus grevyi; Sunda- resan et al. 2007), reticulated giraffe (G. c. reticulatus; Shorrocks and Croft 2009), the Angolan giraffe (G. c. angolensis; Carter et al. 2013a, 2013b), Thornicroft’s giraffe (G. c. thornicrofti; Bercovitch and Berry 2013), and elephants (de Silva and Wittemyer 2012). An excellent introduction to the subject can be found in Croft, James, and Krause (2008), with important comments on hypothesis testing in networks in Croft et al. (2011). The essential elements of a network are nodes and edges. In biology, the nodes are indi- vidual animals and the edges are observed interactions between two individuals (a dyad). In practice the interaction is often inferred, rather than observed. If two individuals are seen in the same group, for example, then they are presumed to interact. As an illustra- tion, Figure 3.1.1 shows the social network for 80 giraffes sampled in Kenya in 2006. Each dot represents one individual (identified by a number) and the lines indicate that two individuals were seen in the same group at least once. The network was produced using 8 103 106 11 105 82 85 13 87 107 84 104 117 128 83 95 42 111 129 73 86 112 108 18 127 81 130 71 115 114 109 113 110 123 75 72 126 74 124 132 98 131 35 22 122 49 5 96 125 97 101 76 51 48 102 70 80 99 6 100 139 27 94 59 77 41 121 88 78 118 90 79 120 89 119 91 92 Figure 3.1.1. The social network for 80 giraffes sampled in 2006, in Kenya (from Shorrocks and Croft 2009). THE ANIMALS 91 Box 3.1 (Continued ) the computer program Netdraw (using the spring-embedding function), and is deter- mined by the links between individuals. When some edges are stronger (more observa- tions), the network can be adjusted accordingly. In the figure, black dots are individuals first seen in 2005, and grey dots are individuals first seen in 2006. Quantitative analysis is carried out not on these graphical representations of networks, but on a matrix of values that conveys the same information. Many of these matrices only contain simple binary data. For example, in the giraffe matrix for the above network, all the individuals were listed along both top and side, and in each cell (apart from the diagonal) of the matrix a value was written. If two individual giraffe were never seen together this was a 0, and if two giraffe were seen in the same group it was a 1. The 1s translate into the edges in the network graph. If sufficient data is available, the 1 can be replaced by a larger number, indicating the frequency of that edge. These are edge weights and, if available, should always be used. If there are not enough data to use the edge weights then probably there is not enough data to allow a meaningful analysis of the social network. Alternatively, a measure of association can be used as the edge value. For example, Carter et al. (2013a, 2013b) used the half-weight index (HWI), which is the most appropriate association index when both members of the pair are located infrequently in each sampling period (sight- ing) (Cairns and Schwager 1987). The index is defined as HWI/=+yyab (0ab yy0 ++.5 aby ), where yab is the number of times both giraffe a and giraffe b are sighted in the same group; ya is the number of times giraffe a is observed but not giraffe b; yb is the number of times giraffe b is observed but not giraffe a; and y0 is the number of times giraffe a and giraffe b are seen in different groups. The HWI produces a value between 0 (a pair of animals are never seen together) and 1 (a pair of animals are always seen associating). Analysis of the network matrix can be carried out using Ucinet (Borgatti, Everett, and Free- man 2002). This is a comprehensive computer package for the analysis of social networks and can deal with very large data sets. The current version is 6.508 (10 March 2014). Integrated with Ucinet is Netdraw, written by Steve Borgatti. Another useful tool is Socprog (Whitehead 2009). This is a series of Matlab programs for analysing data on social struc- ture, population structure, and movements of identified individuals. For those not run- ning Matlab on their computer, there is a compiled version that will run without Matlab. Using the above software, analysis of networks can be extensive. However, simply stat- ed, detailed examination tends to focus on two things: (1) who are the important indi- viduals in the network, and (2) what is its structure? Question (2) can partly be answered by looking at the network graph, but more sophisticated tools are available that search for components and sub-structures. One parameter is the average clustering coefficient (C), which is a measure of the local cliquishness of the network and is calculated as the mean fraction of all possible connections that occur in a local neighbourhood (Newman 2003). C lies in the range 0 to 1. It measures the probability that two of one’s network Continued 92 THE BIOLOGY OF AFRICAN SAVANNAHS Box 3.1 (Continued ) neighbours are themselves neighbours. For the network above, the mean clustering coefficient C( ) is 0.89. Knowledge of C may contribute to our understanding of a popu- lation’s susceptibility to epidemics. For example, high levels of clustering may decrease the spread of a disease in a population, because the infection will be contained among a few individuals that are highly interconnected. Question (1) focuses on individuals and their connection to other individuals. For example, the degree of an individual is the number of other individuals joined to it. The mean degree (k) of the network shown above is 5.47. The mean path length (L), is calculated as the mean number of connec- tions in the shortest path between two individuals in the network. This is calculated among ‘reachable’ pairs of nodes or, in other words, pairs of nodes in the same com- ponent. The observed mean path length for the giraffe population was 3.06. The mean degree of the networks, suggests that on average a given giraffe is connected to approx- imately five or six others. Knowledge of the degree of an individual could be important in predicting the probability that a disease will spread through a population. If the infected individual is highly connected, then the probability that the infection will spread in the network will be greater than if the individual has a low degree. The mean path length of reachable pairs in the network suggests that any two giraffes can be connected on average via 1–3 others. L is a global network measure and can be used to predict how quickly information and disease will spread through populations, with information and disease spreading more quickly between any two individuals with smaller values of L. Finally, a network approach may allow us to identify the role of individuals in maintain- ing social structure. In the current study, certain individuals may provide important links in the social network. For example in the network shown, individuals 71 (adult female), 73 (adult female), 75 (juvenile), and 76 (young male) were important in joining smaller groups into one central social network. Temporal dynamics are also becoming a key issue for many ecological and evolutionary questions. These dynamics include both changes in network topology and the ‘infor- mation’ flow through the network. Network analyses that ignore or do not adequately account for these temporal dynamics can result in inappropriate inferences (Blonder et al. 2012). This paper presents an overview of time-ordered networks, and reviews available methods and software. The calf can stand within a few hours. Interactions between environmental and genetic factors have been shown to affect African buffalo fecundity and population structure (Hooft et al. 2007, Ryan et al. 2007, Hooft et al. 2010). Ryan et al. (2007) examined interactions between African buffalo and the nutritional landscape of Klaserie Private Nature Reserve bordering Kruger National Park (KNP) using the Normalised Difference Vegetation Index (NDVI) (see Box 3.2). NDVI data is obtained from orbiting NASA weather satellites and has been shown to be a more direct proxy for measuring sea- sonal changes in vegetation quality than rainfall (Rasmussen et al. 2006). THE ANIMALS 93 Box 3.2 Normalised Difference Vegetation Index (NDVI) Seasonal changes in the nutritional quality of vegetation play a fundamental role in regu- lating herbivore population dynamics. Satellites are now routinely used to study changes in vegetation quality at the landscape-ecosystem level (Tucker et al. 1985). Multispectral radiometers carried by orbiting satellites detect specific wavelengths, and intensities of the sun’s energy, that are reflected back to satellite detectors by plants. These spectral wavelengths and intensities are then used to construct digital satellite images of plant primary productivity (PPP) in which each pixel corresponds to a specific wavelength band and intensity reflected by plants. For example, the Advanced Very High Resolution Radi- ometer (AVHRR) carries six different detectors. Two of the detectors, sensitive to reflected visible (RED) light (0.55–0.68 μm) and near-infrared (NIR) wavelengths (0.73–1.0 μm), are used to construct detailed maps of savannah PPP. Each pixel represents a given intensity of NIR or RED light. Decades of free NIR and RED spectral data for any region of Africa are available online at NASA and NOAA websites. PPP is directly related to RED and NIR reflectance because the energy required for pho- tosynthesis is obtained by the absorption of RED light by chlorophyll to make protein. Furthermore, the amount of leaf biomass is related to NIR reflectance because non- chlorophyll plant cell structures strongly reflect NIR. Therefore, a high NIR pixel value and a low RED pixel value represent a habitat with high primary productivity. Using NIR and RED data obtained from satellites, one can calculate NDVI values according to the following algebraic formula (Tucker 1979): NDVI =−()NIRRED /N()IR + RED NDVI values range from −1 (water, rock, and sand) to +1 (dense leafy forests). A hypothetical example of how NDVI values might change from wet to dry season might look like this. During the wet season the development of leaf biomass is maximal, there- fore ‘greenness’ or chlorophyll content is at a high level. Say that the percent reflected NIR is 60% and the percentage of reflected RED is only 6%, because most of the RED light is absorbed by chlorophyll. Using the above equation, the wet-season NDVI value is 0.82. By contrast, during the dry season leaf biomass quality and quantity is reduced due to less rainfall and chlorophyll concentrations are relatively low. Assume that the percentage reflected NIR is reduced to 40% due to leaf loss and death and the percentage reflected RED is increased to 35% because RED light absorption is reduced due to a decrease in chlorophyll. The dry-season NDVI value is now reduced to 0.07. Free online NDVI data sets can be obtained from the U.S. Geological Survey FEWS NET (Famine Early Warning System Network) Africa Data Portal at Continued 94 THE BIOLOGY OF AFRICAN SAVANNAHS Box 3.2 (Continued ) NASA has also developed a new vegetation index called the Enhanced Vegetation Index (EVI). This corrects for particles in the air and ground-cover reflections that can affect the NDVI calculations. EVI is more NIR sensitive than NDVI, and more responsive to leaf area index and canopy cover, so it is preferred for studying changes in dense forest PPP. By contrast, NDVI is more chlorophyll sensitive than EVI and is widely used to study changes in savannahs. Using remote sensing methods one can easily compare the current NDVI, for any region of Africa, with a 20 year average to determine if today’s PPP is typical, increased, or decreased (e.g. drought) compared to the past two decades. Many researchers want to create and analyse their own NDVI maps. This can be done using a free online algorithm called MultiSpec developed by the Purdue University Research Foundation, which can be downloaded from This is because soil quality (basalt vs granite), topology, root anatomy and many other factors affect the response of plants to rainfall. Ryan 2007 used NDVI data to test three basic ideas pertaining to how environmental factors regulate the birth timing in African buffalo: predator swamping (Estes 1976), optimisation of social interactions (Brown 1985) and synchronisation of birth with optimal environmental conditions (Rutberg 1987). Examination of 786 monthly calving events from 1993 to 2001 indicated that calving events are correlated with NDVI. This suggests that conception coincides with the peak in plant protein content 1 month after the peak NDVI, when cows are in their best condition. Furthermore, parturition occurred within 1 month of peak nutrition, about 1 year later. January was the peak month for birthing in this study. Optimal foraging by cows ensured the best conditions for lacta- tion and provided their offspring with highly nutritious food. These studies THE ANIMALS 95 clearly indicate that environmental fac- tors play a significant role in recruitment in African buffalo populations. Hooft et al. (2007, 2010) also tested the idea that variations in seasonal rainfall cause sex distor- tion in buffalo. They compared micro- satellite markers (repeating sequences of 2–6 base pairs of DNA) in two differ- ent populations: one buffalo large, socially unsta- ble population living in KNP with a smaller more socially stable population living in Hluhluwe- iMfolozi (HiP), South Africa. Given that sex is determined by genes associated with the X and Y chromosomes, they asked the intriguing question: are there seasonal variations in male:female sex ratios that are correlated with the wet and dry seasons? They found rainfall-correlated changes in the frequencies of Y chromosome halotypes in both populations (Hooftet al. 2007). Later, (Hooft et al. 2010) they showed that seasonal changes in the frequencies of specific Y chromosome haplotypes (557 and 112) showed significant correlations with foetal sex in the KNP population. Haplotype 557 was high after dry seasons while 112 was high after wet seasons, and these patterns were correlated with sex-ratio variations. For several years following the severe drought of 1992, the frequency of haplotype 557 increased fivefold compared to wet season averages. Rainfall and vegetation therefore affect buffalo life history traits, reproductive timing and body condition, and may have cascading effects on buffalo population dynamics. The African buffalo was very badly hit by the rinderpest epidemic that started in 1890. Some estimates say that 10,000 died for every one that survived. Although their numbers have been greatly reduced in recent years by hunting for meat, neither subspecies is threatened and buffalo are listed by IUCN as least concern. Spiral-horned antelopes (Family Bovidae, Tribe Strepsicerotini) Thecommon eland (Taurotragus oryx) (Figure 3.4h) (shoulder height ♂ 1.7 m, ♀ 1.5 m, weight ♂ 700–900 kg, ♀ 450 kg) is the largest living antelope (along with the giant or Derby’s eland). It is very adaptable, being found in 96 THE BIOLOGY OF AFRICAN SAVANNAHS Box 3.3 Microsatellite analysis of population dynamics Microsatellites are short tandem repeats (STRs) of DNA that are scattered throughout the genome in non-coding regions. Repeats are 2–5 nucleotides in length and the num- ber of repeating units for a given microsatellite can be highly variable for individuals. For example, the common dinucleotide CA for a specific microsatellite location may be CACACA (three repeats) for individual A but CACACACACA (five repeats) for individual B. Each population will maintain a specific variety of microsatellites that are different from the microsatellites in other populations, assuming that reproductive barriers pre- vent gene flow between populations. Microsatellite analysis is widely used to study the relatedness of individuals and is therefore an excellent method to explore various aspects of population structure, including gene flow between populations, inbreeding, and kin selection. How are microsatellites detected? Researchers often use DNA extracted from faeces as a source of genetic material for microsatellite analysis. This non-invasive source relies on the fact that epidermal cells from the digestive tract are present in faeces. Next, poly- merase chain reactions (PCR) amplify regions of the genome containing microsatellites in order to have sufficient quantities of DNA for electrophoresis. This is accomplished by designing PCR primers (using software programs) that specifically bind to either side of a unique microsatellite region. Source DNA, primers, nucleotides, and polymerase (an enzyme that makes new DNA from nucleotides) are loaded into a PCR machine which then automatically makes millions of copies of the microsatellite region. These PCR prod- ucts are then separated on a polyacrylamide or agarose gel. GENOTYPER software (Perkin Elmer–Applied Biosystems) is a powerful algorithm for detailed analysis of microsatel- lites located on both alleles. To study fine-scale population structure researchers often use STRUCTURE software (Pritchard, Stephens, and Donnelly 2000). There are hundreds of excellent journal articles describing the use of microsatellite analysis to study African savannah populations. For example, elephants, like many other social animals, live in fission–fusion societies in which individuals move in and out of genetically related groups over time spans ranging from hours to weeks. Archie et al. (2008) obtained 33 years of behavioural and genetic data for 545 elephants living in Amboseli National Park in Kenya. They discovered that the genetic structure of this population is not random, due to non- random mating, and is strongly influenced by male dispersal behaviour. This is a classic example of what is called ‘fine-scale population structure’ based on the fact that while genetically related females live together in core groups, males do not form stable relation- ships, but instead roam the savannah searching for sexually active females. Furthermore, their results suggest that the altruistic behaviour of elephant core groups is most likely the result of kin selection. Kin selection is a central evolutionary mechanism in which repro- ductive success or ‘fitness’ is driven by social relationships between related individuals. Finally, this study shows that poaching can destroy this fine-scale population structure and this can result in population extinctions. Microsatellite methods are therefore an impor- tant tool in conservation biology. Two additional genetic methods are used to study populations. Y-DNA haplogroup analy- sis is used to study patrilineal inheritance; that is, the transmission of genes that affect fitness from father to sons. Haplogroups are groups of similar haplotypes derived from THE ANIMALS 97 Box 3.3 (Continued ) the same ancestor (where haplotypes are defined as groups of linked alleles that are inherited together as a ‘package’ of DNA). The analysis of mitochondrial DNA (mtDNA) haplogroups is used to study matrilineal inheritance. In this case, genetic information that affects fitness is passed from the mother to the offspring of both sexes because an individual’s organelles are transmitted only through the egg. This method has been used to study mother–calf interactions in elephants to determine the extent of ‘allo-mothering’ when younger females in a core group assume the role of mother after the mother is killed by poaching (see Archie, Moss, and Alberts 2006 for more details on this topic). all types of savannah from subdesert to miombo woodland. They are predomi- nantly browsers of foliage and herbs, but also graze green grass (50–80% of wet season diet), fruits, pods, seeds and tubers. Food is gathered by the lips, not the tongue. Although elands are less desert adapted than the oryx, or some of the gazelles, common eland they can go indefi- nitely without drink- ing. They can allow their body temperature to rise as much as 7 °C during the day and let the night temperature cool them down. This would be equivalent to using about 5 litres of water per day in evaporative cooling. Elands form open groups of between 1 and 500 individuals, the latter usually in the rainy season. The only regular associations are between mother and calf. They are one of the most mobile antelopes, with home ranges varying between 200 and 1,500 km2. A single calf (22–36 kg) is born after a gestation period of 9 months. The calf remains hidden in bush cover for the first 2 weeks. The eland’s conservation status is least concern. Africa has two species of kudu. The large and elegantgreater kudu (Trage- laphus strepsiceros)(shoulder height 1.4–1.6 m, weight ♂ 250 kg, ♀ 180 kg) 98 THE BIOLOGY OF AFRICAN SAVANNAHS is one of few large antelopes that has extended its range in recent years, mainly the southern African subspecies. This is largely because of its secretive nature and its abil- ity to survive in settled areas with sufficient cover. It pre- fers wooded savannahs and is predominantly a browser. However, a wide range of foliage, herbs, vines, flow- ers, fruits, succulents, and a little new grass, is taken. It can survive without water if greater kudu browse is sufficiently moist. It is non-territorial, but lives in small, nursery, herds of 2–15 individuals: cows, their young and sometimes an associated male. Males may be solitary, or form temporary bachelor herds of 2–10 individuals. The home range of nursery herds varies from 1 to 25 km2 depending on the quality of the habitat, but males may have home ranges of 50 km2. In the wet season a single calf (16 kg) is born after a gestation period of 9 months. The young lie up for about 3 weeks, but are weaned and inde- pendent by about 6 months. Overall the greater kudu is not endangered but has declined in the northern part of its range (Somalia, Chad, Uganda, and Kenya) where it may be endangered or vulnerable. The smaller lesser kudu (Tragelaphus imberbis) (shoulder height 1 m, weight ♂ 100 kg, ♀ 62 kg) is restricted to north-eastern savannahs where it inhab- its arid Acacia–Commiphora woodland and scrub. It is a browser, with a wide and varied diet of leaves, flow- ers, fruits, and seed pods. In the rainy season a little green grass is eaten. They are water independent, and eat succu- lents such as wild sisal in the dry season. The basic social unit is 1–3 females with their offspring. These female asso- ciations tend to be exclusive, and relatively long lasting (three known females stayed together for 4–5 years). Adult lesser kudu THE ANIMALS 99 males spend much of their time alone, only associating with females for reproduction. They are not territorial and quite sedentary. Most births occur in the rainy season, when a single calf (7 kg) is born after a gestation period of 7 months. The lesser kudu suffered a substantial decrease as a result of the 1990 rinderpest outbreak, and although it appeared to be recovering and estimated numbers were high (about 22,000, according to East 1999), it is now listed as near threatened by IUCN. The bushbuck (Tragelaphus scriptus) (Figure 3.6b) (shoulder height ♂ 80 cm, ♀ 70 cm, weight ♂ 45 kg, ♀ 30 kg) is a very variable species, both in colour and in markings, with over 29 subspecies described. They like wood- land and scrub savannah near water (riverine habitats). It only leaves cover, for more open areas, to find choice food items, water and other bushbuck. It eats tender new grass, but is predominantly a browser on herbs and shrubby legu- minous plants. In some areas they cause some damage to agricultural crops (fruits and flowers) and are consid- ered a pest. They are solitary, non-territorial and rather sedentary. The only regu- larly associated individuals are a female and her latest offspring. A single calf (4 kg) is produced after a gesta- tion period of 6 months. The bushbuck young lie up in cover for up to 4 months, before mov- ing around with the mother. Some montane subspecies (Mount Elgon and northern Uganda) are vulner- able, but overall the bushbuck is widespread and abundant, and is listed as least concern. It occurs in a larger number of African countries (40) than any other antelope species. Horse antelopes (Family Bovidae, Tribe Hippotragini) The oryx (Oryx gazella) (Figure 3.6c) (shoulder height 1.2 m, weight ♂ 240 kg, ♀ 210 kg) has five subspecies. These include thebeisa oryx (O. g. beisa), the fringe-eared oryx (O. g. callotis) and the gemsbok (O. g. gazella). The fringe-eared oryx, of southern Kenya and Tanzania, inhabits Acacia– Commiphora savannah that is less open, and less arid than the scrub savan- nahs occupied by the other two subspecies. One of the best desert-adapted large mammals, it prefers green grass, but also grazes dry grass, and some- times browses on wild fruits and tree pods (Acacia). It will dig up roots, (a) (b) (c) (d) (e) (f) (g) (h) Figure 3.6 Savannah mammals: (a) sable antelope, (b) bushbuck, (c) beisa oryx, (d) defassa waterbuck, (e) topi, (f) common wildebeest, (g) impala, (h) Grant’s gazelle (all photgraphs by Bryan & Jo Shorrocks). THE ANIMALS 101 bulbs and tubers for moisture. Like other desert mammals, when deprived of drinking water it employs various methods to minimise its water needs. These include allowing the body temperature to rise during the day (from a normal 35.7 °C to 45 °C) before beginning evaporative cooling, and con- centrating the urine. It is nomadic, gregarious (mixed herds of up to 30) and the males are probably ter- ritorial. A single calf is born after a gestation period of about 9 months. Like other beisa species that hide their new- born calves, oryx cows isolate themselves from their herd just before calving. The young remain hidden for 3–6 weeks, fringe-eared after which they join the herd. The northern subspecies, the beisa oryx, is listed as near gemsbok threatened, with the gems- bok least concern. The fringe- eared oryx has not recently oryx been assessed by IUCN but in 2008 was listed as vulnerable. The horse-like roan antelope (Hippotragus equinus) (shoulder height 1.1– 1.5 m, weight 220–300 kg) was historically one of the most wide-ranging antelopes in sub-Saharan Africa. Its present distribution is now largely divid- ed between the northern and southern savannahs. It inhab- its open or slightly wooded savannah with long grass. It is a grazer. Herds of between 6 and 20 individuals (adult cows and their young), led by an adult bull, are formed. These groups share a tradi- tional and exclusive home range, which in the Kruger National Park is 60–120 km2. In Tanzania, a herd of 14 ani- mals was observed to use an area of 12 km2 over a period of 17 years. Ranges are often roan antelope smaller, and herds larger, in the dry season. Herd males appear to simply defend an area around the moving herd. Between the ages 102 THE BIOLOGY OF AFRICAN SAVANNAHS of 3 and 6 years young males form bachelor herds. A single calf is born (16– 18 kg) after a gestation period of about 9 months. The female leaves her herd before giving birth and the young roan remains hidden for about 2 weeks, before joining the herd. Numbers have been severely reduced in recent years, but 1998 estimates suggest about 39,000 roan in sub-Saharan Africa. It is classified as least concern, but conservation dependent. Like the roan antelope, the sable antelope (Hippotragus niger) (Figure 3.6a) (shoulder height 1.2–1.5 m, weight ♂ 200–270 kg, ♀ 190–230 kg) is a large, strongly built animal with magnificent curved horns. There are four subspe- cies. The giant sable,H. n. varianii, noted for its large horns and confined to Angola; H. n. roosevelti, now confined to Shimba Hills National Park on the East African coast; H. n. niger, the black sable found south of the Zambezi river and H. n. kirkii, centred in the miombo woodlands. They inhabit open woodland with medium to tall grass. They are mainly grazers of medium height grasses, but forbs and foliage make up 20% of their diet. Herds of 15–25 females and young are common, but larger groups of 30–75 are not uncommon, especially in the dry season. Females tend to remain in their birth area so that local ‘clans’ tend to develop, and knowledge of the area is sable antelope passed on from generation to generation. A single calf (13– 18 kg) is produced after a ges- tation period of 9 months. Females leave the herd to give birth, and the calf remains hidden for several weeks. Overall IUCN consider the sable as least concern, since estimated numbers are c.750,000. However, the giant sable is classed as critically endangered. Reedbucks, kob and waterbuck (Family Bovidae, Tribe Reduncini) The large, shaggy, and robust waterbuck Kobus( ellipsiprymus) (shoulder height 1.3 m, weight ♂ 200–260 kg, ♀ 160–215 kg) is one of the heaviest antelopes. There are two subspecies, thecommon waterbuck (K. e. ellip- siprymus), with a white ring encircling its rump, and the defassa waterbuck (K. e. defassa) (Figure 3.6d), with a white patch on its rump. It is possibly the most water dependent of all antelopes and consequently occupies savan- nah habitats near to water. It is a grazer of short to medium-length grass but THE ANIMALS 103 acquires additional protein by browsing on herbage such as acacias and Cap- arris for up to 21% of its feeding time. In Benin, Kassa et al. (2008) found that in the dry season perennial grasses (>40%) and tree forage (35%) were the dominant components of the diet. In comparison, annual grasses (>50%) were domi- nant during the rainy season. When about 6 years old, males establish territories (0.5–2.8 km2) which they hold for an average of 1.5–2 years (the range is a few months to sev- eral years). Females associate in small herds of 5–10 indi- defassa viduals. The home ranges of these female ‘nursery herds’ may encompass the territo- ries of several males. Young waterbuck common males live in small bachelor herds after being chased out of the nursery herd by terri- torial males. A single calf is produced (13 kg) after a gestation period of about 9 months. The waterbuck is not endangered and is classified as least concern. Thekob (Kobus kob) (shoulder height ♂ 92 cm, ♀ 78 cm, weight ♂ up to 120 kg, ♀ 60 kg) is rather similar to the impala, but more heavily built. There are three main subspecies: the white-ear kob (K. k. leucotis) in Sudan, the Ugandan kob (K. k. thomasi) in Uganda and the Demo- cratic Repubic of the Congo, white-eared kob and Buffon’s kob (K. k. kob) in the northern savannah. It is a grazer that inhabits grassy floodplains. At low density, kobs have conventional ter- Buffon’s kob ritorial behaviour with males spaced at least 100–200 m Ugandan puku kob apart. Females with their calves live in small, loosely structured, herds (5–15 indi- viduals). Young males join bachelor herds. However, kob and puku studies on the Ugandan kob (Leuthold 1966) have shown 104 THE BIOLOGY OF AFRICAN SAVANNAHS that at high densities one-third of males cluster on traditional breeding are- nas, or leks, where territories are extremely small (15–30 m). The continued close presence of males and transitory females removes most of the grass cover from such arenas, leaving nothing substantial to eat. However, since 80–90% of females visit these leks, it is worth the strongest males defend- ing them, for an obvious reproductive advantage. A single young (4–5 kg) is produced after a gestation period of 7 months. The kob’s status islower risk, but conservation dependent. The puku (Kobus vardonii) is a slightly smaller, but very similar, species that replaces the kob ecologically in the central and southern savannahs. They are classified as separate species because there are no hybrids between them, a reflection of their geographical isolation. The conservation status of the puku is least concern, but conservation dependent. The three reedbuck species are all rather similar. The bohor reedbuck (Redunca redunca) (shoulder height ♂ 70–90 cm, ♀ 65–80 cm, weight 45–55 kg) and common, or southern, reedbuck (R. arundinum) (shoulder height ♂ 96 cm, ♀ 80 cm, weight 43–51 kg) frequent grasslands, with tall grass. The mountain reedbuck (R. fulvorufula) (shoulder height 72 cm, weight 30 kg) inhabits hill country with scattered trees. All three are grazers, although the common reedbuck will also take browse. Studies of their digestive system suggest that reedbucks can take grasses that are unpalatable to most other antelopes. Reedbucks bridge the gap between soli- tary and gregarious territorial systems. The common reed- buck lives in monogamous pairs while the mountain and the bohor reedbuck live in small herds of 3–8 females and young. Males defend ter- common reedbuck ritories in all three species. bohor reedbuck Reedbucks produce a single mountain reedbuck young (3–5 kg) after a gesta- tion period of 7–8 months. The status of all three species is least concern, but conservation dependent. Hartebeests, topi, wildebeests and impala (Family Bovidae, Tribe Alcelaphini) The common hartebeest (Alcelaphus buselaphus) (shoulder height 1.25 m, weight ♂ 150 kg, ♀ 120 kg) is a typical open-savannah antelope. There are several subspecies, the more easily recognised being western harte- beest (A. b. major) (Senegal to western Chad), lelwel (A. b. lelwel) (south THE ANIMALS 105 Sudan, Ethiopia, Uganda and Kenya), tora (A. b. tora) (east Sudan, Ethiopia), Swayne’s hartebeest (A. b. swaynei) (Ethiopia and Somalia), Jackson’s hartebeest (A. b. jacksoni) (central Kenya), Coke’s hartebeest (A. b. common cokii) (Kenya, Tanzania), and Cape or red hartebeest (A. b. caama) (southern Africa). Lichenstein’s hartebeest (A. (Sigmoceros) lichensteinii) (miombo woodland savan- Lichtenstein’s hartebeest nah of central East Africa) is sometimes regarded as another subspecies, but more often as a separate species. Hartebeest prefer open savannah but will enter open woodland and bush savannah more readily than wildebeest or topi. They are almost entirely grazers, often associated with medium grasslands dominated by red oat grass (Themeda triandra). In Nairobi National Park in the 1960s and 1970s, female herds (6–15 ♀♀ + young) had home ranges of 3.7–5.5 km2 and wandered through male territories. Prime males (4–7.5 years) held small territories in the best habitats. A single calf (15 kg) is born after a gestation period of 8 months. The calf usually remains hidden until it is strong enough to follow the herd. Swayne’s hartebeest (only about 200 individuals) and lelwel are regarded as endangered, the western hartebeest is near threatened, and the tora hartebeest (numbers unknown, but decreasing) is regarded as critically endangered. The other subspecies are lower risk, but conservation dependent. Damaliscus lunatus (shoulder height 1.2 m, weight ♂ 140 kg, ♀ 126 kg) has several subspecies, each known by a unique common name. In north- west Africa it is the korrigum (D. l. korrigum); in north-east Africa it is the tiang (D. l. tiang); in East Africa and north-eastern Congolian ecore- gion it is the topi (D. l. jimela) or coastal topi (D. l. topi) (Figure 3.6e); and in southern Africa it is the tsessebe (D. l. lunatus). However, Cotter- ill (2003) recognised tsessebe, in the southern Bangweulu flats of north- eastern Zambia, as a new species, Damaliscus superstes, based on differences in cranial morphology and pelage. They all resemble a hartebeest, but are smaller and darker. They inhabit open woodland and scrub savannah and eat grass. The long narrow muzzle and mobile lips enable them to select the more tender green blades. Like many other grazers they like the new green flush after rain, but cannot use short grass as efficiently as bulk feeders such as wildebeest, waterbuck and zebra. Social and reproductive organisation is 106 THE BIOLOGY OF AFRICAN SAVANNAHS more variable than any other antelope. In grassland patches in more wooded savannahs where population density may be low, males hold small territories (0.25–4 km). Each territory holds the resources sufficient for one herd of 2—6 females (rarely over 10) and their young. The resident male has exclusive rights to the females. Where Damalis- cus lunatus live at higher den- sities, on more open extensive grasslands (e.g. Queen Eliza- topi, tsessebe, tiang and korrigum beth National Park in Ugan- da), temporary territorial networks can be established wherever the circulating mass of animals settle for a few hours. In Akagara National Park, in Rwanda, leks are established by the males on traditional arenas. Females and young circulate in large herds (200–300), and females enter the lek arena singly or in small groups. A single young (10–12 kg) is born after a gestation period of 8 months. Calves are either hiders (remain hidden after birth for several days) or followers (keep up with the herd from birth) depending on the habitat and social organisation. The coastal topi is listed asnear threatened, the korrigum subspecies (about 2,500 individuals) is vulnerable, while the other subspecies are lower risk, but conservation dependent. The common wildebeest (white-bearded wildebeest, blue wildebeest, brin- dled gnu) (Connochaetes tau- rinus) (Figure 3.6f) (shoulder height ♂ 1.5 m, ♀ 1.3 m, weight ♂ 250 kg, ♀ 180 kg) has five subspecies. The most widespread and abundant (East 1999) are the white- bearded wildebeest (C. t. mearnsi) in East Africa and the blue wildebeest (C. t. taurinus) in southern Africa. This species dominates some open short-grass and shrub savannahs. In the Serengeti ecosystem of northern Tan- zania, the annual migration common wildebeest THE ANIMALS 107 of 1.5 million white-bearded wildebeest from the southern grass savannahs to the northern Masai Mara of Kenya is renowned. The animals move off the grasslands as the dry season begins and return with the advent of the rainy season. However, large migratory populations of blue wildebeest also occur, or used to occur, on the extensive short-grass plains and Acacia savannahs of Botswana, Namibia, southern Angola and south-western Zambia. Some populations, for example those resident in the northern part of the Serengeti ecosystem, and those in Ngorongoro crater, do not migrate. They graze a wide variety of grasses that form short swards. They are bulk feeders and their broad muzzle, with its wide incisor row, enables them to rapidly close- crop the short grass. Resident populations form small segregated herds of females that are relatively sedentary and overlap the smaller territories of several males. When on the move, migratory males simply defend a zone around their cows. A single calf (14 kg) is born after a gestation period of about 8 months. The calving season varies from area to area. They are not endangered, and classified as least concern but conservation dependent. However, the future prospect of some of the subspecies, particularly the east- ern white-bearded wildebeest, has undergone a precipitous decline in recent years. In southern Africa there is also a black wildebeest (C. gnou) which is slightly smaller, but also a grazer. Regarded as the quintessential antelope, the impala (Aepyceros melam- pus) (Figure 3.6g) (shoulder height 90 cm, weight ♂ 50 kg, ♀ 40 kg) is difficult to classify taxonomically. Sometimes it is placed near gazelles, and sometimes near hartebeests. There is one easily recognised subspecies, the black-faced impala of south-west Africa. It inhabits open and lightly wooded savannah with access to water. It is a mixed feeder. During the wet season it eats green grass and during the dry season it browses on foliage, impala forbs, shoots, and seed pods. For example, in the north- west of Zimbabwe their diet was observed to change from 94% grass in the wet season to 68% browse in the dry season. Mature males alternate between bachelor and territorial sta- tus and rarely hold a territory for more than a few months. In the Serengeti only about one-third of adult males were territorial at any one time, how- ever, in East Africa territorial behaviour is seen most of the year. In southern 108 THE BIOLOGY OF AFRICAN SAVANNAHS Africa vigorous territorial behaviour is limited to only a few months, during the annual rut. During the rest of the year males and females can associate in mixed herds. Male territory size varies with population density, and habitat quality. Groups of females, and their offspring, associate in herds of anything from 6 to 100 individuals. These herds, or clans, occupy traditional home ranges of 80–120 ha. A single calf (5 kg) is born at the beginning of the rainy season after a gestation period of 6.5 months. The common impala is one of the most abundant antelopes in Africa. It is not endangered. The black-faced impala is vulnerable, mainly due to the risk of hybridisation with the com- mon impala. Gazelles and dwarf antelopes (Family Bovidae, Tribes Antilopini and Neotragini) The giraffe-necked antelope or gerenuk (Litocranius wal- leri) (shoulder height 95–100 cm, weight 30–50 kg) is one of the most easily identifiable antelopes. It is an inhabitant of the arid grass and shrub savannah of the Horn of Africa. It is a strict browser of trees, not known to eat grass or herbs. Although 87 different types of shrubs and trees have been recorded in its diet, within a tree it feeds quite selectively, using its very gerenuk narrow muzzle to take choice foliage from even thorny branches. When browsing, it frequently stands on its hind legs and uses its long neck to reach foliage inac- cessible to other browsers except the giraffe. Although they can be found in small groups of 2–12 individuals, they are frequently seen alone (in Tsavo National Park, 42% of those sighted were single individuals). Males become permanently territorial when mature and rarely move off their territory, which may be up to 4 km2. Most gerenuks are local and very sedentary. A single young (3 kg) is produced after a gestation period of 7 months. Its pop- ulation is decreasing, and it is regarded as near threatened. Gazelles The pale Grant’s gazelle (Nanger granti) (Figure 3.6h) (shoulder height ♂ 85–95 cm, ♀ 80–85 cm, weight ♂ 55–80 kg, ♀ 35–50 kg) is found over a wide range of savannahs from semi-desert, through arid scrub, to open woodland savannah. It is a mixed feeder, but takes grass only when it is green. Herbs and foliage are preferred during the later wet and dry seasons. THE ANIMALS 109 The fruits of bothBalan - ites and Solanum have been recorded in the diet. Grant’s gazelle is gregarious, territo- rial and migratory. A study in the Serengeti suggests that the size and composition of social groups varies with habitat. On the Serengeti grasslands, nearly half the herds were mixed, with the most com- mon group size being about 50 animals. In woodland savannah, only 12% were Grant’s gazelle mixed groups, most being harems of one male with sev- eral females and their young. Here the most common herd size was only about 10 individuals. When males reach 3 years they tend to leave their natal herd and become territorial. After a gestation period of 6.5 months, a single young (2 kg) is dropped. Although most populations are in decline, this gazelle remains widespread in East Afri- ca and is not presently endangered. Its conservation status is least concern, but the population trend is decreasing. The ‘tommy’ orThomson’ s gazelle (Eudorcas thomsonii) (shoulder height 55–65 cm, weight 15–25 kg) is an abun- dant species (about 550,000) with a restricted East Afri- can range. The subspecies in Sudan (G. t. albonotata), is known as the Mongalla gazelle. Thomson’s gazelle inhabits open grassland shrub savannah, although it prefers areas with short grass. Unlike Thomson’s gazelle Grant’s gazelle, this species is predominantly a grazer of green grass, requiring regular access to drinking-water. Some herbs and seeds are taken in the dry season. Like the Grant’s gazelle this species is gregarious, territorial and migratory. Social groupings are rather loose and open, especially during long migra- tions, with females leaving and entering female herds. During the mating 110 THE BIOLOGY OF AFRICAN SAVANNAHS season, males defend small territories (100–300 m) and attempt to mate with any females that enters the area. Towards the end of the rainy season, a single young (2–3 kg) is born, after a gestation period of about 6 months. It is not endangered, although its conservation status is near threatened. The springbok (Antidor- cus marsupialis) (shoul- der height 75 cm, weight 26–41 kg) is the only gazelle in southern Africa, where it is abundant (estimated 670,000). When alarmed the springbok, like many ante- lopes, employs a distinctive bouncing gait called stot- ting. The springbok’s stot- ting is particularly dramatic, reaching a height of 2 m, and gives the species its common name. It inhabits open arid grass and shrub savannah, springbok but has considerable habitat tolerance. It can be found in savannahs with rainfall of up to 750 mm to the Namib desert with only 0–100 mm. It is a mixed grazer/browser that takes young and tender grasses, and a variety of shrubs and succulents. In the Transvaal reserve, springboks fed on 68 different species, of which 9 grasses and 11 shrubs formed the major part of the diet. Its social structure, and territorial behaviour, is very similar to that of Thomson’s gazelle. In the past, mass migrations of springbok have been recorded. Four such migra- tions, involving upwards of 15,000 individuals, were recorded between 1946 and 1959, usually at the end of the dry season in October/November. Springbok in the Kalahari moved south-west, into Cape province, and were shot in their thousands. These journeys (trekbokkens) still occur in Botswa- na from time to time, but on a much smaller scale. A single young (about 4 kg) is born after a gestation period of about 5.5 months. The springbok is not endangered, and its conservation status is least concern. The population trend is increasing. Dwarf antelopes The dwarf antelopes are small, delicate, antelopes with the female slightly larger than the male. They are largely crepuscular or nocturnal and live singly or in small family groups. There are 13 species but only the four dik-diks, steenbok, oribi and two grysboks are savan- nah species. The klipspringer and beira inhabit rocky areas, often within savannah ecosystems. Only three species of dwarf antelope are outlined here. THE ANIMALS 111 The steenbok or steinbok (Raphicerus campestris) (shoul- der height 50 cm, weight 11 kg) occurs widely in dry savan- nahs, in two separate regions either side of the miombo woodland savannah. It inhab- its open savannah grasslands, with bush and light woodland which can be used for cover. The steenbok is a selective mixed feeder on high-quality browse, and tender new grass. It browses on the leaves and steenbok shoots of low shrubs and trees, forbs, seeds and seed pods, ber- ries and fruit. In the Kalahari it also digs for roots and tubers. It is usually seen singly or in loosely maintained pairs. A single young (900 g) is produced after a gestation period of 5.5 months. The young remain hidden for the first few weeks after birth. The steenbok is abundant, and its conservation status is least concern, with an estimated total population of 663,000 (East 1999). Kirk’s dik dik (Madoqua kirki) (Figure 3.7a) (shoulder height 38 cm, weight 5 kg) and Guenther’s dik-dik (M. guentheri) (shoulder height 34–38, weight 3.7–5.5 kg) are virtually impossible to distinguish in the field, and are treated here as a single entity. The two species overlap in the Laikipia region of Kenya and may well interbreed. The other two species (M. piacentinii and M. salti- ana) are also very similar, and confined to the dry savannahs of the Horn of Africa. Like the steenbok, Kirk’s dik-dik has a disjunct distribution. They inhabit relatively dry shrub savannah and are browsers. Because of Guenther’s their relatively large surface dik dik area, these small antelopes have water-loss problems in arid environments—they Kirk’s cannot afford to sweat. Evap- dik dik orative cooling of the blood is achieved through nasal pant- ing. To develop this mecha- Kirk’s dik dik nism, the nose has become swollen and elongated, a characteristic feature of all 112 THE BIOLOGY OF AFRICAN SAVANNAHS dik-diks. They live in closely associated pairs, in territories defended by the male. There is some evidence that they mate for life. This territory varies in size depending upon the quality of the habitat. A single young (600–750 g) is born after a gestation period of about 5.5 months. Both species of dik-diks are common and classed as least concern. Subungulates: mammals out of Africa Molecular biology and anatomy have shown that elephants (3 species), hyraxes (11 species), and the aardvark, although looking very different, may well be closely related. They appear to have evolved in Africa during the con- tinent’s isolation in the Eocene period. They have short, nail-like hooves, and are collectively called the subungulates. Only the two African elephants are described here. African elephant (Family Elephantidae) Elephants are sometimes referred to as ‘the ambassadors of Africa’ as they are iconic animals of the savan- nah. Furthermore they are important ‘ecological engineers’ that help to maintain the tree:grassland ratios characteristic of savannahs (Chapter 5). There are two species of Afri- can elephants, the savannah elephant and the forest ele- phant. They were originally regarded as subspecies, but recent molecular work (using biopsy samples collected by dart from 195 free-ranging elephants, in 11 African countries) (Roca et al. 2001) suggests that they should be treated as separate species. The savannah populations in southern, eastern and north central Africa, although widely separated, were genet- African elephant ically indistinguishable, while the forest animals showed more genetic diversity. The paucity of gene introgression (exchange) between forest and savannah populations, even near regions of potential physical contact, suggests that hybridisation in nature is rare. Most ecological and behavioural studies have probably been carried out on the savannah species. Thesavannah elephant (Loxodonta africana) (Figure 3.7b) (shoulder height ♂ 3.2–4 m, ♀ 2.5–3.4 m, weight ♂ 5,000–6,300 kg, ♀ 2,800–3,500 kg) is the largest land mammal. Its range of habitats is extremely wide, from semi-desert through all types of THE ANIMALS 113 (a) (b) (c) (d) (e) (f) (g) (h) Figure 3.7 Savannah mammals. (a) Kirk’s dik-dik, (b) elephant, (c) lion, (d) cheetah, (e) serval, (f) spotted hyaena, (g) wild dog, (h) black-backed jackal (photographs by Bryan & Jo Shorrocks). 114 THE BIOLOGY OF AFRICAN SAVANNAHS savannah to forests and swamps. It is a mixed feeder, taking a wide variety of plants. In the wet season they eat green grass, and in the dry season they browse on woody and herbaceous species. The life history strategy of the African elephant is unique in several ways (but similar to that of humans). Elephants have large brains and complex sociality; they are highly intelli- gent, show empathy, and have complex social bonds, a long gestation period (656 days) and long-term care of offspring (spanning 2 years). Maternal investment for elephants is very high, with females investing a great deal of energy producing milk and teaching their young how to survive. Long post- productive survival is rare in mammals, but is found in both elephants and humans, presumably because leadership and experience play an important role in social organisation. Female elephants form matriarchal herds (2–24 individuals, 9–11 being typical) of related individuals (McComb et al. 2001). The matriarch of a family (i.e. the oldest female) transmits social knowl- edge to other members of the family that is critical for the survival of the entire group. Within the family group the matriarch usually has the biggest tusks, and is therefore the prime target for poachers. The death of the matri- arch therefore puts the rest of the family group at risk. Direction and rate of movement are set by the matriarch. Mature males form separate herds, or are single. Many populations are nomadic/migratory, with very large home ranges. Females come into oestrus for 2–6 days, after a period of between 3 and 9 years. Males that are in ‘musth’ (in a sexually active or rutting phase) search out these rare females by responding to their calls. A single calf (120 kg) is born after a gestation period of 22 months. Research using satellite-based NDVI analysis (see Box 3.2) has provided us with new insights into elephant–habitat interactions. Wittemyer et al. (2007a, 2007b) studied breeding phenology in relation to NDVI in the Samburu Nature Reserve in northern Kenya. Their results show that parturition rates are synchronised with the average seasonal peak in plant development. This shows that the triggering of female reproduction is related to environmental conditions, so that birthing occurs during the time of maximum plant devel- opment when plant protein is the highest. They also showed that increased rainfall associated with El Niño caused a significant increase in NDVI values that in turn affected conception rates of Samburu elephant. NDVI values were related to hormone levels in elephants, with 5-alpha-pregnane-3 levels (measured from dung) fluctuating with NDVI values. In another study (Cer- ling et al. 2009) dietary histories of elephants were studied using isotope- ratio mass spectrometry (IRMS) (see Box 3.4). From a single tail hair they could determine dietary histories spanning 6 years for each elephant. The ingestion of C3 versus C4 plants, plus isotope signatures from all water sourc- es utilised by elephants, were recorded along the length of a single tail hair. Furthermore they were able to correlate IRMS data with NDVI data in order to get a detailed picture of how the environment regulates elephant feeding behaviour. THE ANIMALS 115 Box 3.4 Analysis of dietary histories using isotope-ratio mass spectrometry Isotope-ratio mass spectrometry (IRMS) is a powerful new method to study seasonal changes in herbivore diets. As herbivores feed on plants, natural isotopes contained within the plants are digested and then incorporated into the cells of the animal. For example, Cerling et al. (2006, 2009) discovered that IRMS can be used to study how elephants utilise plant resources by examining 13C/12C isotopic ratios contained in tail hair samples. They found that growing hair is able to record when individuals were feeding on grasses and when they were feeding on trees. This is based on the fact that there are 13 12 significant differences between C/ C ratios in grasses that utilize the C3 photosynthetic pathway compared to trees that utilise the C4 photosynthetic pathway (Chapter 1). The youngest isotopic signatures are incorporated into the hair shaft located closest to the hair follicle, embedded in the skin, and the oldest isotopic signatures are located at the hair tip. Hair growth rates were determined to be 0.55 ± 0.11 mm day−1 and a single 500 mm length of hair provided migration/diet data for more than a year. Briefly, the IRMS method involves the combustion of samples prior to mass spectrometry analysis. Isotopic ratios are expressed in the per mil (‰) notation. For example, in the study by Cerling et al. (2006), C3 plants for Samburu have an average of −27‰ whereas for C4 plants the aver- age is −13‰. IRMS results showed that elephants moved into the highlands during the dry season (with relatively lower 15N values) and in the wet season elephants were feeding in the lowlands (with relatively higher 15N values). These IRMS data were correlated with NDVI values (see Box 3.1). In Cerling et al. (2009), isotopic data is presented for 6 years of dietary history for a single elephant family living in the Samburu-Buffalo Springs National Reserves in Kenya. 13C/12C, 15N/14N, 34S/32S, D/H, and 18O/16O ratios recorded both dietary and environmental features in hair samples and IRMS was able to determine sources of water utilised by elephants. Furthermore, they found that the Ewaso N’giro river showed seasonal differences in deu- 18 13 terium (D) and O values. C4 plants (i.e. most tropical grasses) have C values between 13 −11 and 14‰, in contrast to C3 plants (most trees, shrubs, and forbs have C values between −25 and −29‰. Given that elephants prefer to graze on grasses in the wet sea- son and browse on C4 plants during the dry season, isotopic signatures reflecting these dietary changes are incorporated into growing hair. Their results demonstrate that IRMS combined with NDVI methods can be used to monitor the effects of climate change on herbivore diets. IRMS and NDVI were also used to predict when elephants were most likely to breed (about 3 weeks after peak grazing and about 5 weeks after the peak in NDVI). Since the gestation period for elephants is 22 months, the peak time for births would occur at the beginning of the rainy season when both water and protein-rich young grass is in abundance. Given that significant changes in vegetation caused by global climate change and resource competition between wildlife and humans will likely intensify in future, obtaining data on dietary histories for a wild range of species living in various habitats is of great interest to conservation biologists. For example, climate change may cause dietary changes that further intensify seasonal reproduction and this in turn could lead to changes in popula- tion dynamics. 116 THE BIOLOGY OF AFRICAN SAVANNAHS The study of elephant brains supports the idea that elephants are highly intel- ligent (Hart et al. 2008). Elephants are capable of empathy directed towards helping sick family members and grieving the death of family members (McComb et al. 2006). The neurological basis of these highly evolved behav- iours is now being studied by many scientists. Gene expression in elephant brains has revealed a striking similarity to that in humans (Goodman et al. 2009). Genes involved in aerobic energy metabolism were similar in ele- phants and humans and are correlated with large brain size and the high oxygen consumption of brain cells. Elephant brains are the largest of any land animal (6.5 kg) consuming four times more oxygen than human brains. Elephant brains contain large quantities of white matter, which is involved in rapid transmission of information. Dendritic branching is extremely high and the neocortex has an extremely high number of glial cells that function to help transport energy substrates to neurons that play direct roles in pro- ducing complex social skills, compassion (Douglas-Hamilton et al. 2006), and self-awareness. Vocalisations play a critical role in elephant fission–fusion societies. Calls help to determine intra- and inter-group associations. Individuals rec- ognise each other’s calls. Spectrographic analysis (see Box 3.5) can be used to understand acoustic signatures produced by individuals. Infra- sonic communication and seismic communication play important roles in elephant behaviour (Garstang 2004, O’Connell-Rodwell 2007). Ele- phants are one of the few animal species capable of vocal learning (Poole et al. 2005). Based on spectrograms, six call types were identified (Poole et al. 1988): four laryngeal (rumble, bark, grunt, and roar) and two trunk calls (snort and trumpet), although there may be more, as we are just begin- ning to understand elephant vocal communication. The forest elephant (L. cyclotis) (shoulder height ♂ 2.35 m, ♀ 2.1 m, weight ♂ 2,800–3,200 kg, ♀ 1,800–2,500 kg) is smaller and darker than its savan- nah relative and has characteristically rounded ears. Forest elephants appear to live in much smaller groups, of between 5 and 8 individuals. The IUNC have not distinguised between the two species/subspecies and class African elephants, in total, as vulnerable. Carnivores Although very variable in outward appearance, most African carnivores share certain features. The body is typically long, supple and agile, and the legs are well muscled, with movement in all directions. They are usually equipped with strong claws, which in the case of the cats (except the chee- tah), and some genets and civets, are retractable. Probably associated with the advantage of a long stride when running, is the presence of a reduced and unattached collar bone. Like the herbivores, they have special teeth that help them with their diet. The last upper premolar and first lower molar have THE ANIMALS 117 Box 3.5 Bioacoustics The African savannah is a mosaic of diverse acoustic habitats spread over a vast landscape. The sounds of the African savannah are an essential part of the ecosystem, facilitating social interactions within and between different species. From the chirps and chirrups of stridulation (rubbing body parts together) produced by bush-crickets, grasshoppers, ants, and dung beetles to the more complex trumpeting of elephants and roaring of lions, sonic signals have many diverse roles in animal communication. These include territorial defence, sexual competition, family recognition, predator alarm calls, and mate choice. Sounds are pressure waves that move through air, water, or earth at specific frequencies that are often measured in hertz (wave cycles per second). Sound waves also exert pres- sure that is measured in pascals (Pa; 1 Pa equals the force of 1 Newton/metre squared). Microphones respond to this pressure and then Fast Fourier Transform (FTT) algorithms are used to transform the electric analogue of the primary signal into a digital image of the sonic energy called a spectrogram. Software programs can analyse many features of the spectrographic representation of the acoustic energy, including frequency, intensity (amplitude), duration, entropy, and acoustic envelope. The acoustic envelope determines the shape and texture of a sound as it travels in time through the environment. The most common type of spectrogram shows how the frequency of a signal changes over time, with frequency being plotted on the vertical axis and time plotted on the horizontal axis with the amplitude of frequency frequently represented by a colour. For an excellent introduction to bioacoustics see Bradbury and Vehrencamp (1998). Several free online algorithms are widely used by researchers to analyse animal vocalisa- tions: for example, Raven software (Bioacoustics 2014, Continued 118 THE BIOLOGY OF AFRICAN SAVANNAHS Box 3.5 (Continued ) Most animals emit acoustic signals within the frequency range of the human ear (from 20 to 20 000 Hz). However, some animals communicate using infrasound (e.g. elephants at 15 Hz), or high-frequency sounds up to 100 kHz (e.g. bats, rodents, and some crickets), that humans cannot hear. Computer software allows researchers to change these fre- quencies so that we can ‘hear’ these vocalisations. The source–filter theory provides the best understanding of how avian and mammalian vocalisations are produced (Fant 1960). This theory is based on a two-step biomechani- cal mechanism for sound production. The ‘source’ corresponds to vibrations generated by vocal folds (vocal cords or voice reeds) and the spaces between the vocals (collectively called the glottis). In birds, their song is produced by vibrations of the walls of the syrinx, not the vocal folds. Air pressure generated by the lungs results in the vibration of the vocal folds and this vibration produces sound. The rapid opening and closing of the glottis produces what is called the fundamental frequency (F0) of the vocalisation. Fo is related to pitch but it is not identical to it, because pitch is the psychoacoustic interpretation of F0. What determines F0 are the length and mass of the vocal folds. Because the vocal folds of individuals are not identical, it is possible to identify individual callers based on their F0. The supralaryngeal tract (including the mouth) functions like a series of bandpass fil- ters that either dampen or enhance specific frequencies, so that as sound energy travels towards the mouth, the anatomical features of the supralaryngeal tract modify it. The major resonating chambers of the mammalian supralaryngeal tract are the pharynx, nasal cavities, and mouth, and the filtered resonating energy emitted from the mouth appears on spectrograms as formants. In humans, the vocal folds vibrate at the same periodic rate; however, many non-human animals produce functionally important non-periodic ener- gies (called non-linear phenomena, NLP). NLP include subharmonics (harmonics below F0), biphonation (two independent F0s), and deterministic chaos (broadband with no har- monics). All of these types of vocalisation are used as social cues by African wild dogs, for example. There have been many studies on vocalisation in elephants and lions. Elephants produce a wide range of vocalisations including F0s less than 20 Hz. These low-frequency vocali- sations or rumbles function in long-distance communication between individuals. Low- frequency rumbles are ideally suited for this because they are not as easily absorbed by the atmosphere, and they are less altered by the environment than are higher-frequency signals. Research has shown that elephants emit rumbles that can travel 10 km when thermal inversions of the atmosphere at sunset are optimal for signalling (Garstang et al. 2004). In fact, many other African species vocalise when the atmospheric conditions are optimum for long-distance communication (e.g. hyaenas, frogs, and nightjars). Elephant rumbles play an important role as a contact call between individuals in their complex and highly evolved fission–fusion societies. While searching for food, members of an ele- phant family are often out of visual contact. Rumbles convey an individual’s identity and reproductive and emotional state, thus facilitating social bonding (Soltis 2009). Research has shown that rumbles are produced using a unique myoelastic-aerodynamic (MEAD) Continued THE ANIMALS 119 Box 3.5 (Continued ) mechanism that does not require neural input (Herbst et al. 2012). Playback experi- ments in Etosha National Park in Namibia have demonstrated that elephants can produce and detect seismic information transmitted through the ground. This seis- mic information is used to discriminate between family and non-family alarm calls (O’Connell-Rodwell et al. 2007). Furthermore, elephants can select habitats that facili- tate seismic transmission. The roar of the lion is another iconic sound of the savannah. Roaring lions help to create what ecologists call ‘the landscape of fear’. Figure 3.5.1 shows a spectrogram of a Masai Mara lion produced using SAP software. Lions have a uniquely evolved vocal folds for roaring (Klemuk et al. 2011) and have evolved a unique hyoid apparatus that surrounds the larynx, tongue, and upper vocal tract (Weissengruber et al. 2002). One component of the hyoid, the elastic epihyoid, 15102 16612 18122 1963221142 15102 16612 18122 1963221142 Figure 3.5.1 Visualization of lion full-throated roars using the Sound Analysis Pro (SAP) spectral derivatives method (top panel) compared to using a typical spectrographic method (bottom panel). Four full-throated roars located approximately in the middle region of a roaring bout are shown. The resolution of vocal energy patterns is more clearly defined using the spectral derivatives algorithm. Note that the first two roars are comprised of both a harmonic tonal segment (as indicated by evenly spaced horizontal lines) and a noisy segment (appears fuzzy). The combination of harmonic and noisy segments into a single syllable is termed segmental concatenation. The fourth roar in this series is primarily noisy in structure. Variations in patterns of vocal energy convey different information that are important in many aspects of animal behavior. The roars shown here were recorded in the Masai Mara by W. Bates and analyzed using SAP software by W. Bates. 120 THE BIOLOGY OF AFRICAN SAVANNAHS Box 3.5 (Continued) appears to be important for lower formant frequencies that are characteristic of lion roars. Roars are delivered as bouts lasting about 90 s, reaching sound pressures of about 114 dB at 1 m and travelling a maximum of 7 km depending on the topology of the land- scape, vegetation type, and other factors. Roars are long-distance contact calls between pride members, attract mates, prevent potentially lethal contests, prevent infanticide by invading males, and repel predators. Playback experiments show that roars convey an individual’s sex, but not male condition (Pfefferle et al. 2007). Bioacoustics can be used to monitor the health of any environment. The book Silent Spring, written by Rachel Carson in 1962, profoundly demonstrated how nature’s sounds are intimately linked to the health of the environment. Carson’s book proved to be prophetic, as the dawn chorus sung by many species of birds is diminishing world- wide. In a recent study conducted in Tanzania, automated acoustic monitoring showed that the morning chorus for Tanzanian birds showed lower acoustic diversity for dis- turbed habitats (Sueur et al. 2008). The development of automated recorders makes it possible to obtain acoustic data for large spatial and temporal scales, and one thing is now clear: the ‘great animal orchestra’ (Krause 2012) of the natural world is being dis- rupted by human activities. This new field of ecology, in which the collective sounds of habitats comprised of abiotic (geophony: wind, rain, thunder, etc.), biotic (biophony), and human-generated sounds (anthrophony) is called soundscape ecology (Pijanowski et al. 2011). The term ‘soundscape’ was first used by the composer R. Murray Schafer, who pioneered the study of the collective sounds of different habitats beginning with Stanley Park in Vancouver, Canada. Natural soundscapes are spatially unique from the landscape scale down to the microhabitat scale and temporally dynamic, changing over the course of the day and with the seasons. The phenology of natural soundscapes is related to wet and dry season changes in vegetation that absorb or reflect acoustic energy, local climate changes (temperature inversions, rain, thunder), animal migra- tions (wildebeests, quelas, locusts, cuckoo birds, and many other species). Recently, the analysis of a savannah soundscape in Zimbabwe led Kraus to postulate the acous tic niche hypothesis. When these soundscapes were analysed using spectrograms, he observed that the acoustic energy emitted by each species formed discrete patterns of species-specific acoustic bandwidths within the same vocal territory occupied by many animals, in such a manner that individual signals are not masked. However, when habitats are disturbed these acoustic niches become less distinct and there is signal interference. Given that ‘acoustic garbage’ (Berendt 1992) is increasing in many regions of Africa due to mass tourism and many other forms of anthrophony (logging, mining, road development), the unique mosaic of African soundscapes must be protected from human disturbance as soundscapes have an important role in the survival of African savannahs. For additional information on bioacoustics and soundscape ecology, see the companion website for this book at lion spotted hyaena C C C C wild dog black-backed jackal C C large-spotted savannah genet baboon C C Figure 3.8 Skulls of five carnivores, and one primate. The carnassial shearing teeth are indicated by a letter c, on the carnivores. (from Skinner and Smithers 1990). and also to collectively repel such larger predators (e.g. some mongooses). This latter ‘antipredator’ explanation can also be extended to defending a car- cass. Several spotted hyaenas can protect their kill against a lion, as can sever- al wild dogs against a spotted hyaena. In fact the necessary ratio of defenders to ‘robbers’ (in both examples) appears to be about 4:1. However, some spe- cies of carnivore live in groups but hunt and travel as single individuals. For THE ANIMALS 123 Table 3.4 Percentage of the diet obtained by scavenging, and measured percentage success rate once hunting has been initiated Cheetah Leopard Lion Spotted hyaena Wild dog Scavenged (%) 0 5–10 10–15 33 3 Success (%) 37–70 5 15–30 35 50–70 Data from Bertram (1979). example, brown hyaenas, which hunt at night and are therefore not frequent- ly observed, were thought to be asocial. However, radio tracking of individu- als has found that they often share roughly the same home range. They occur together in the same area more often than would be expected by chance. Of course, they may simply be using the same best habitats, but they may also be said to live in ‘spatial groups’ (MacDonald 2001). In some carnivores it is not just the parents that look after the young, and a third explanation, or benefit, of group living may be shared parenthood. In the black-backed jackal, for example, an older female offspring may delay reproduction and stay with her parents to help rear another, younger, set of pups. In lions the females of a pride (who are sisters) help to rear all the cubs in the pride. In a wild dog pack, one pair are reproductively active and subordinate dogs help to maintain the offspring of this dominant pair. This behaviour may appear dis- advantageous to the helpers who are delaying their own reproduction. They may, however, benefit by learning parenthood and therefore being more suc- cessful when they have young of their own. Also, if the breeding pair are monogamous, the helper may in fact be just as closely related to the group’s young as they would be to their own young. From an evolutionary point of view they are helping to ensure the survival of their own genes. The same comments apply to the carnivore distribution maps as applied to the herbivore maps. However, with carnivores our estimates of numbers are often more unreliable than for the larger, more visible herbivores (see Chapter 4). The distribution maps in the species accounts are based on Estes (1991), Stuart and Stuart (1997), Kingdon (1997), Nowell and Jackson (1996) (cats), Mills and Hofer (1998) (hyaenas) and Woodroffeet al. (1997) (wild dog). Cats (Family Felidae) The lion (Panthera leo leo) (Figure 3.7c and Figure 3.8) (shoulder height ♂ 1.2 m, ♀ 1 m, weight ♂ 150–240 kg, ♀ 122–182 kg) is not just an African spe- cies although its Asian distribution is now sadly restricted to the Gir Forest Sanctuary in India. It is the largest of Africa’s three ‘big cats’. Lion-like cats first appeared in East Africa during the Pliocene (Turner and Anton 1997). They use a very wide range of habitats, from semi-desert, through grassland, to wooded savannahs. Their prey are medium to large mammals, such as buffalo, giraffe, zebra, warthog and antelopes. Some prides learn to specialise on certain prey. For example, a pride that foraged along Namibia’s Skeleton 124 THE BIOLOGY OF AFRICAN SAVANNAHS Coast learned to prey upon Cape fur seals. They usu- ally (but not always) hunt at night. Prides consist of a stable core of related females (sisters and daughters) and coalitions of males that are unrelated to the females. The males in a coalition are relat- ed, having been born in the same pride. Despite maternal defence, infanticide is com- mon when males take over a pride. Pride home ranges usually vary from 26 to 220 lion km2 but can go up to 2,000 km2. One to four cubs (1.5 kg) are born after a gestation period of 3.5–4 months. Birth can take place at any time of the year, although adult females in a pride often conceive at the same time so that cubs can suckle from any female that is lactating. Most young females join their natal pride. Males tend to leave their natal pride at 2–4 years. Charles Darwin (1859) asked why male lions devel- op a mane. He suggested that the mane functioned as a defensive shield. However, extensive research on this topic suggests that the mane is used to attract females. Lion manes show a high degree of variation and it is estimated that about 50% of this variation is caused by climate (Patterson et al. 2006). Long, dark manes that cover the male’s chest are highly attrac- tive to females. However, the costly trade-off is that dark manes absorb more solar energy, causing an increased body temperature that lowers sperm count (West and Packer 2002). Vocal signals play important roles in lion behaviour. Males establish and maintain their territories using roars, and these complex patterns of acoustic energy are used to attract mates. Indi- vidual members of a pride are often spread out over considerable distances, and vocal signals function to maintain social contacts between pride mem- bers (Schaller 1972, McComb et al. 1994, Grinnell and McComb 2001). Depending on landscape features and atmospheric conditions, a lion’s roar can be transmitted 7 km. This long-range communication is due to the evo- lution of a unique vocal tract (Klemuk et al. 2011). There are no reliable estimates of the total number of lions in Africa. Guesstimates range from 30,000 to 100,000 (Nowell and Jackson 1996). Their scavenging behaviour makes them vulnerable to poisoned carcasses put out to eliminate predators. Its IUCN status is vulnerable, with populations decreasing. Many recent local extinctions have occurred in West and Central Africa due to habitat fragmentation. Riggio et al. (2013) suggest that free-ranging lions now live in only about 25% of savannah area, and that this area will significantly decrease over the next two decades. THE ANIMALS 125 The leopard (Panthera pardus) (shoulder height 70–80 cm, weight ♂ 20–91 kg, ♀ 17–60 kg) is the quintessential cat, the most catlike of all cats. It has the widest distribution of all felines, and has been found in virtually all African habitats, from desert to savannah and from rainforest to high mountains (a carcass was found on the rim of Mount Kilimanjaro’s Kibo crater in 1926). It also takes a huge variety of prey, from dung beetles to adult male eland. Bailey (1993) found that at least 92 prey species have been documented in the leopard’s diet from sub- Saharan Africa. It escapes kleptoparasitism (stealing of its prey) by stashing it up a tree. It is solitary and ter- ritorial. Adults associate only long enough to mate, and males defend territories by calling and demarcating with urine, dung and tree- scratching. One to four cubs (500 g) are born, at any time of year, after a gestation peri- leopard od of 3.5 months. The conser- vation status of the leopard is classed as near threatened. The English name of thecheetah (Acinonyx jubatus) (Figure 3.7d) (shoulder height 80 cm, weight 30–72 kg) is derived from the Hindi chita, meaning ‘spotted one’. The genus nameAcinonyx refers to its non-retractile claws. It is the fastest land animal, attaining speeds of 95 km h−1 (60 mph). Because a final, fast, sprint is an integral part of its hunting technique, it prefers to live in relatively open savan- nah, with some cover. It hunts by day and takes mainly medium-sized antelopes such as Thomson’s gazelle, puku, springbok and impa- la, although it will also eat smaller prey such as hares and birds. It escapes some kleptoparasitism by avoiding areas with lions. Females are usually solitary, while males can form coalitions, or also be solitary. Where their main prey is migratory, they can cheetah have very large home ranges. 126 THE BIOLOGY OF AFRICAN SAVANNAHS For example, in the Serengeti (main prey Thomson’s gazelle) ranges aver- aged about 1,000 km2. In other areas they may be only 12–36 km2. One to five cubs (250–300 g) are born, at any time of year, after a gestation period of about 3 months. For the first 6 weeks they are usually hidden in dense cover. For a large carnivore, cheetahs have a high mortality rate. In the Serengeti, 95% of young cheetahs never reach independence, largely because lions kill the cubs. It is classed as vulnerable throughout its range. Between 5,000 and 15,000 cheetahs may remain in sub-Saharan Africa. Southern Africa is the cheetah’s regional stronghold, with a roughly estimated population of at least 4,500 adults. Seven species of smaller cat are found in Africa, although only five are sub- Saharan and only four are found in savannahs. (Felis nigripes) is found in South Africa, Botswana, and Namibia where it inhabits semi-arid scrub and grassland savannah. The African wild cat (Felis silvestris), looking like our domestic cat, is found everywhere in sub-Saharan Africa, except the rainfor- ests of the Congo basin. The serval (Leptailurus serval) (Figure 3.7e) (shoulder height 60 cm, weight 8–13 kg) is found in all types of savannah, especially with long grass. The serval locates prey in tall grass mainly by hearing and takes a variety of small rodents, hares, birds, reptiles, inverte- brates and probably the young of small antelopes. The final pounce is a high leap, which may span 1–4 m and be over 1 m high. Servals are solitary serval and territorial, although they may well fall into the ‘spatial groups’ category mentioned in the carnivore introduction. In other words, they share an area but avoid con- tact. In a study carried out in the Ngorongoro crater of northern Tanzania, up to seven servals occupied an area of 7.5 × 4 km over a 2.5 year period (Geertse- ma 1981). One to three, rarely five, young (200 g) are produced, mainly during the wet season, after a gestation period of 2.5 months. It is common over most of its range and listed as least concern by IUCN. They appear very tolerant of agricultural development which encourages increased rodent densities. The caracal (Caracal caracal) (shoulder height 40–45 cm, weight 7–19 kg) is the largest of the small cats. It inhabits all types of savannah, although to some extent the serval and caracal replace each other. Servals tend to be found in wetter habitats, caracals in more arid. They take a wide range of THE ANIMALS 127 prey including rock hyraxes, reedbuck, springbok, steen- bok, hares, rabbits, small rodents and birds. It uses the same killing technique as the big cats—a suffocating bite to the neck. Like the serval, the caracal is solitary and territo- rial. One to four young (250 g) are produced after a ges- tation period of 2.5 months. The conservation status of the caracal is least concern in sub-Saharan Africa. It is most caracal abundant in South Africa and southern Namibia. Hyaenas (Family Hyaenidae) Hyaenas are dog-like carnivores which are, however, more closely related to cats. There are four species, in four genera. Three species are ‘typical’ hyae- nas, while the fourth, the aardwolf, is rather different. The aardwolf (Proteles cristatus) (shoulder height 50 cm, weight 6–11 kg) eats mainly harvester termites and has reduced, peg-like, teeth, rather than the massive cheek teeth of the three hyaenas that are used for crushing bone. It prefers relatively open savannahs. It is a nocturnal, solitary forager and probably monogamous. Evidence from observations of scent marking suggests that male and female aard- wolves share a 1–2 km2 terri- tory together with their most recent offspring. Between one and four pups (500 g) are pro- duced after a gestation period brown hyaena of 2 months. Its conserva- aardwolf tion status is least concern although data is poor for the north-eastern populations. The brown hyaena (Hyaena (Parahyaena) brunnea) (shoulder height 80 cm, weight 45 kg) is the dominant large carnivore in the south-west arid 128 THE BIOLOGY OF AFRICAN SAVANNAHS savannahs and even penetrates the Namib desert. It is a nocturnal scavenger, killing only 6% of its food. It will eat almost anything and over 58 different kinds of food have been identified in its droppings. Since most of the large mammals that inhabit the arid savannahs of the south-west are migratory, the composition of the diet of brown hyaenas changes with the season. In the dry season the amount of large mammal material drops to 17% and fruit and vegetables increase. They cache surplus food. Brown hyaenas live in clans of 1–4 adults, 0–5 subadults, and 0–4 cubs, but forage alone. Clan members mark and defend territories (Mills 1990). Two to three cubs are born after a gestation period of 3 months, in a birthing den separate from the commu- nal den of the clan. After about 3 months they are brought to the clan den where they are reared communally. Its status is near threatened, and poison- ing, trapping and hunting have had a detrimental effect on populations. Esti- mated total population is between 5,000 and 8,000 (Mills and Hofer 1998). The striped hyaena (Hyaena hyaena) (shoulder height 72 cm, weight 40–55 kg) has a northern African distribution, spreading across the Mid- dle East into India. There are at least five subspecies. It inhabits arid scrub savannah. Its diet is mainly mammalian carrion, with some invertebrates, eggs, wild fruit and organic human waste. They probably have a similar clan structure to that of brown hyaenas, although they are relatively unstudied. They hunt alone, at night. A large study in Laikipia District, central Kenya, estimated the minimum regional density at 0.03 adults km−2 (Wagner 2006). Two to four cubs are born, at any time of year, after a gestation period of 3 months. Its conservation status is near threatened. The estimated African population is between 5,000 and 14,000, with the greater part being in Egypt and Kenya (about 60%) (Mills and Hofer 1998). The large, dog-like,spotted hyaena (Crocuta crocuta) (Figure 3.7f and Figure 3.8) (shoulder height 85 cm, weight 60–80 kg) is a predator rather than a scavenger. In fact, it is one of the few predators that may influence the numbers of its prey. It is found in semi-desert, arid scrub savannah, wood- land savannah, and even mountainous forest up to 4,000 m. It usually hunts in small groups, running down its prey in a prolonged chase. These prey include small, medium, and large antelopes, buffalo, zebra, warthog and the young of giraffe. However, the composition of the diet varies depending upon the area. In Kruger National Park the most important prey are wildebeest, buffalo, zebra, greater kudu and impala. In the southern Kalahari they take gemsbok, wildebeest and springbok and in the Masai Mara, 80% of the prey are topi and Thomson’s gazelle. Unlike most carnivores that waste up to 40% of their kills, the spotted hyaena consumes almost everything, including the bones. It is highly social, living in fission–fusion clans of up to 80 individu- als, and using complex vocal communication to maintain a strict dominance hierarchy (Smith et al. 2008, Smith et al. 2011). Females have high levels of testosterone, are dominant over males, and have unusual masculinised exter- nal genitalia (Glickman et al. 2006). Hyaenas have primate-like intelligence THE ANIMALS 129 and field experiments have shown that they have the abil- ity to count (Benson-Amram et al. 2011). When researchers challenged a clan of hyaenas with playback recordings of non-clan contact calls from one, two or three intrud- ing hyaenas, clan members showed increased vigilance (by looking at the loudspeak- ers) as the number of differ- ent intruders increased. The striped hyaena ability of hyaenas to count spotted hyaena potential aggressors supports the game theory prediction that animals assess risk before engaging in potentially deadly interactions (Maynard Smith and Price 1973). Furthermore, these studies showed that individual call discrimination plays an important role in hyaena societies. Sex, age, dominance and individual identity are encoded in the hyaena’s call (Mathevon et al. 2010). Territory size is variable from less than 40 km2 in the Ngorongoro crater to over 1,000 km2 in the Kalahari. One to two cubs (1.5 kg) are born after a gestation period of about 3.5 months. Its conservation status is least concern, and estimated numbers are between 27,800 and 48,200 (Mills and Hofer 1998), although the population trend is decreasing. Genets, civets and mongooses (Family Viverridae and Herpestidae) These are small to medium-sized carnivores, with an elongated body and face, a pointed muzzle, short legs (apart from the African civet), and general- ly a long and furry tail. They are thought to be related to the cats and hyaenas. Between 10 and 12 species of African genets (shoulder height 18–25 cm, weight 1.2–3.5 kg) are recognised, depending on the author (Stuart and Stu- art 1997, Macdonald 2001, Duff and Lawson 2004 ). They all look rather similar and most people would have some difficulty separating them in the field. Only three species could be said to use savannah type habitats; the rest are forest species. These are thecommon , or small-spotted, genet (Genet- ta genetta) (including the feline genet, G. felina), the large-spotted genet (G. tigrina) (Figure 3.8) (including the forest genet, G. maculata) and the Hausa, or Villier’s, genet (G. thierryi). The last species is restricted to the drier savannahs of West Africa, while the small-spotted and large- spotted genets occur over most of sub-Saharan Africa. They eat insects, small rodents, reptiles, birds and fruit. They are nocturnal foragers. Between two 130 THE BIOLOGY OF AFRICAN SAVANNAHS and five young (50–80 g) are born after a gestation period of about 2 months. None of the genets are listed as endangered (least concern). Two species of civets can be found in savannahs. The African civet (Civet- tictis civetta) (shoulder height 40 cm, weight 9–15 kg) and the African palm civet (Nandinia binotata) (shoulder height 22 cm, weight 1.5–3 kg). Both prefer forested savannah and forage at night. Their diet is similar to that of genets. They produce two to four young after a gestation period of about 2 months. Both civets are common and listed as least concern. At least 22 species of mongoose live in Africa, of which 14 frequent savan- nahs. Of these, only nine are common (Table 3.5), although the Egyptian mongoose is uncommon but very widely distributed. All nine are diurnal foragers except the white-tailed mongoose, which is nocturnal, and the Egyptian mongoose, which can be both nocturnal and diurnal. All mon- gooses depend on scent for communication and to mark territories. None of the mongooses that inhabit savannahs are listed as endangered. Dogs and jackals (Family Canidae) Dogs are more closely related to bears, seals and mustelids than to the cats and their allies mentioned above. There are 11 species, ranging in size from Table 3.5 Common African savannah mongooses Name Scientific name Weight (kg) Main diet Behaviour Distribution Egyptian Herpestes ichneumon 2.4–4.0 Small rodents, s, f North, sub-Saharan, invertebrates, fruit Nile valley slender H. sanguineus 0.5–0.7 Invertebrates, small s Sub-Saharan vertebrates, fruit banded M. mungo 1.0–2.0 Insects, birds, reptiles c Sub-Saharan rodents white-tailed Ichneumia albicauda 3.0–4.0 Invertebrates, rodents, s, f Sub-Saharan fruit dwarf Helogale parvula 0.2–0.4 Insects, small reptiles, c East and central birds desert dwarf H. hirtula 0.2–0.4 Insects, small reptiles, c Horn of Africa birds yellow Cyntis penicillata 0.3–0.8 Invertebrates, vertebrates, c South carrion cape grey H. pulverulenta 0.5–1.0 Small rodents, s, f South invertebrates meerkat Suricata suricata 0.6–0.9 Insects, other c South invertebrates Social structure: s, solitary; f, family groups; c, colonial. Distribution: sub-Saharan implies everywhere except rainforest of Congo basin and horn = the Horn of Africa. THE ANIMALS 131 the fennec fox (1.0–1.5 kg) to the African wild dog (17–36 kg). Four of these are not savannah species. Thered fox (Vulpes vulpes) is confined to the northern coast and Nile basin, the fennec fox (V. zerda) and Ruppell’s fox (V. ruppelli) are found in the Sahara desert, and the Ethiopian wolf (Canis simensis) is confined to the alpine moorlands of the Ethiopian highlands. The other seven species are found in savannah habitats. The wild dog (Lycaon pictus) (Figure 3.7g and Figure 3.8) (shoulder height 65–80 cm, weight 17–36 kg) was once found throughout the savannahs of sub-Saharan Africa but its numbers are now greatly reduced and its distribu- tion very fragmented. It will inhabit a range of habitats from semi-desert to miombo woodland savannah. It hunts a wide range of medium-sized ungu- lates, which vary with the region. In Kruger National Park it takes mainly impala, in the Kalahari it takes springbok, and in the Serengeti ecosystem its main prey are Thomson’s gazelle and wildebeest. Wild dogs have the same evil folklore associated with them as do wolves in Europe, and they have been hunted systematically by humans. They are diurnal, cooperative hunters. Packs consist of a breeding pair and several non-breeding adults (often 20 or more dogs) that assist in provisioning the lactating mother and pups. Young males, when they mature, stay with the natal pack while females emigrate. Hunting ranges are often huge, with estimates in the Serengeti of 1500–2000 km2. Two to 19 pups are born after a gestation period of 2–2.5 months. The young are born in dens, often the abandoned burrows of other animals, and remain close to this den for about 3 months. Woodroffeet al. (1997) esti- mate that there are between 3,000 and 5,500 wild dogs, in perhaps 600–1,000 packs, remaining in Africa. The major causes of their historic decline are persecution by humans, habitat fragmentation, loss of kills (kleptoparasit- ism), and road deaths. Disease from domestic dogs is another major threat. They are listed as endangered. The three savannah jackals are all rather similar in size (shoulder height 30–48 cm, weight 6–15 kg). Theblack- backed jackal (Canis mesome- les) (Figure 3.7h and Figure 3.8) is the most commonly seen jackal of East Africa and southern Africa. Theside- striped jackal (C. adustus) is rather uncommon while the golden jackal (C. aureus) is the most desert adapted and restricted to Africa north of the side-striped. They have wild dog a wide tolerance of habitats 132 THE BIOLOGY OF AFRICAN SAVANNAHS from quite arid semi-desert to more open wooded savan- nah. They are opportunistic feeders and take a wide vari- ety of food items, from small mammals to carrion and fruit. The black-backed jackal is perhaps more carnivo- rous, taking young antelopes, rodents, birds and reptiles. They are both nocturnal and diurnal. All three species side-striped jackal are thought to have similar both species social structure, although the side-striped jackal is largely black-backed jackal unstudied. They are monoga- mous and territorial, with yearling offspring helping to rear the next set of young. Three to six pups (occasionally 1–9) are born after a gestation period of 2–2.5 months. None of the jackal species are listed as endangered. Of the three savannah foxes, the Cape fox (Vulpes chama) is confined to the grassland and arid scrub savannahs of southern Africa while the pale fox (V. pallida) is found only in the dry Sahel, from the Atlantic to the Red Sea, bordering the Sahara desert. Thebat-eared fox (Otocyon megalotis) (shoul- der height 30–40 cm, weight 3–5 kg) is very much a savan- nah species, being found in both southern and eastern areas. It likes open grassland and light Acacia woodland. Its food is mainly insects, particularly the harvester ter- mite (Hodotermes mossam- bicus). It will occasionally eat small vertebrates and fruit. In a sample of 72 bat-eared foxes from Botswana, 88% of stomachs contained mainly insects. The most common bat-eared fox food items were termites, followed by beetles, grass- hoppers, scorpions, rodents and reptiles. They live as monogamous, non- territorial pairs, sometimes with female helpers that are daughters. One to six pups are produced after a gestation period of 2–2.5 months. None of the three savannah foxes is listed as endangered, although there is no detailed information on the abundance of the pale fox. THE ANIMALS 133 Badgers, weasels, polecats and otters (Family Mustelidae) Mustelids are small to medium-sized carnivores, with plantigrade feet; that is, they walk flat-footed like elephants, hyraxes, bears and humans. They include the weasels, badgers and otters. There are 10 African species, 3 of which—the Eurasian otter (Lutra lutra), weasel (Mustela nivalis) and polecat (M. putorius)—are only found along the North African coast. Three otters (Lutra maculicollis, Aonyx capensis and A. congica) are found in rivers that sometimes run through savannahs, but they are not really savannah animals, and the North African banded weasel (Poecilictis libyca) is largely confined to the Sahara desert and North African coast. Of the three truly sub-Saharan savannah species, the African striped weasel (P. libyca) is restricted to parts of central and southern Africa, and is uncommon. The two common savan- nah species are the honey badger or ratel (Mellivora capensis) (shoulder height 30 cm, weight 8–14 kg) and the striped polecat or zorilla (Ictonyx striatus) (shoulder height 10–15 cm, weight 0.6–1.4 kg). Both have an exten- sive range, from the Sahel through dry scrub savannah to miombo wooded savannah. The honey badger eats a wide range of items, including inverte- brates, rodents, reptiles, birds, carrion and wild fruit. It will break into bee colonies and eat the honey and larvae, a behaviour that gives it one of its common names and its scientific name of Mellivora. It is mainly nocturnal and crepuscular. It appears to have a mutualistic relationship with a bird, the greater honey guide (Indicator indicator), which it follows to beehives (Chap- ter 5). Its social behaviour is largely unstudied, but it is probably monoga- mous. One to four young are produced after a gestation period of about 6 months. The honey badger isnot endangered. The zorilla eats mainly insects, but takes some rodents and other small mam- mals. It is nocturnal and solitary. Two to three young (15 g) are produced after a gestation period of 36 days. The zorilla is not endangered. Primates This order of mammals, to which we belong, has approximately 59 African species, although whether some of these are species or subspecies is debat- ed. However, in general they are not inhabitants of savannah ecosystems, except perhaps for the forest–savannah mosaics that surround the Congo basin of central Africa. They range in size from the gorilla (Gorilla gorilla) (♂ 140–180 kg, ♀ 70–100 kg) to the Demidoff’s galago Galagoides( demidoff) (60 g). Primate characteristics include a rather generalised, highly mobile, skeleton; five fingers and toes equipped with flat nails instead of claws; a large collar bone; eyes placed frontally giving binocular vision; and, in many spe- cies, a large and complex brain. The nails protect sensitive fingertips, and in most species one digit (e.g. the thumb or big toe) is opposable, enabling the hand or foot to grasp objects. Clearly many of these primate features are associated with life in trees, grasping branches and judging distances. 134 THE BIOLOGY OF AFRICAN SAVANNAHS Two species commonly seen in savannahs are detailed below. A third species, the patas monkey (Erythrocebus patas), inhabits the dry savannahs of West Africa, from Senegal eastwards to Ethiopia, but is less frequently seen. It is diurnal and typically lives in groups of about 15. An home range of 52 km2 was recorded for one troop of 31 indiviuals. It is thought to be closely related to the vervet monkey and is probably the most terrestrial of primates, apart from the baboons. It feeds on Acacia fruit, galls, leaves, insects and tree gum. It is not endangered. The vervet, savannah, or green monkey (Chlorocebus aethiops) (weight ♂4.3 kg, ♀ 3 kg) is common and widespread in sub-Saharan Africa. Several geographical races (subspecies) have been described, which some people give specific status. These includeC. aethipos (grivet monkey), C. pygery- thus (vervet monkey), C. sabaeus (green monkey), C. tanatalus (tantalus monkey) and C. cynosuros (malbrouk monkey) (Grubb et al. 2003, Groves 2005). It is by far the most numerous and widespread monkey in Africa, and in many urban and agricul- tural areas is regarded as a pest. It is found in all types of savannah, but usually never far from water and tall trees, in which it spends the night. Vervets are opportun- istic omnivores and take a wide range of fruits, flowers, leaves, gum, seeds and even insects. It will forage on the ground, turning over small logs to search for beetles and other invertebrates. These vervet monkey monkeys live in troops with a social structure rather similar to that of baboons. Long-term successive studies in Amboseli National Park suggest these troops can vary in size from 8 to 50 individuals, with anything from 1 to 8 males. The troops are in fact a hierarchy of families whose mem- bers sleep, forage, and rest together. In Amboseli, troops lived in a mosaic of stable territories that varied in size from about 18 to 76 ha. A single young (300–400 g) is born after a gestation period of about 7 months. Vervet monkeys are not endangered. The savannah baboon (Figure 3.8) (shoulder height 40–60 cm, weight♂ 16.9– 25.1 kg, ♀ 9.9–13.3 kg) occurs in a number of races (most recently regarded as species) that have always been given their own name. In the lowlands of East and Central Africa lives the yellow baboon (Papio cyanocephalus) with, THE ANIMALS 135 as its name suggests, yellowish fur. In the highlands of East Africa, with olive- greenish fur, is the olive baboon (P. anubi s ). The southern African form has dark grey fur and is known as the chacma baboon (P. ursinus); finally, the form found on the extreme west Guinea baboon olive baboon coast is known as the Guinea baboon (P. papio) and has brown fur. They are found in most savannahs, limited only by the availability of water and secure sleeping sites, such as tall trees and cliffs. Like the vervet they are omni- vores taking a wide range of yellow baboon plants and animal food. These include flowers, seeds, fruits, chacma baboon resin, leaves, roots, bulbs, invertebrates and even young savannah baboon antelopes, rodents, birds and reptiles. In agricultural areas they will raid crops. Baboons live in groups (troops), usually of 30–40 individuals, although the observed range is 8–300 individuals. Adult females tend to remain their whole lives in one troop and therefore make up the stable core of the group (comprising about 30% of individuals) with immature classes making up about 50%. Males emigrate during adolescence (after about age 4) and often transfer repeatedly between troops, even when mature. There is a strict hierarchy between adult females, with the young taking on the rank of their mother. A lower-ranking adult female will therefore defer to the infant of a higher-ranking female. Home ranges vary between 400 ha and 4000 ha, dependent on the quality of the habitat, and not the size of the troop. A single young is produced after a gestation period of about 6 months. The IUCN classification of the savannah baboon is least concern. Rodents Of all the mammal orders, rodents contain the largest number of species worldwide. In Africa, they include the ground squirrels (5 species), the rope squirrels (9 species), the bush squirrels (10 species), the giant and sun squirrels (9 species), the spring hare, the gundis (4 species), the dormice (2 species), the blesmols (5 species), the porcupines (3 species), the cane-rats (3 species), the jerboas (3 species), and the rat-like murids (1,330 species). One of the main characteristics of rodents is their continuously growing, gnawing incisors, and chewing molars, separated by a gap (called the diastema) left by the absence of the canine and premolar teeth. They consume a great vari- ety of plant material and have a relatively large caecum containing bacteria that facilitate the digestion of cellulose. Many, of course, have a prodigious 136 THE BIOLOGY OF AFRICAN SAVANNAHS capacity for reproduction. Most species are pregnant for just 19–21 days and mate again within 2 days of giving birth; the young can begin breeding when 6 weeks old. Theoretically, a single breeding pair of mice could produce 500 mice in 21 weeks. Although many of these rodents, such as the ground and bush squirrels, porcupines, cane-rats, and the murid rats and mice, are abun- dant in savannahs, there are simply too many species to pick out any for individual description. However, their combined effect upon savannah eco- systems can be noticeable. They are both consumed and consumers. Many of the smaller carnivores described above take rodents as part of their diet. In the Serengeti study mentioned earlier, Sinclair (1975) (Table 3.1) estimated that rodents consumed about 69 kg ha−1 yr−1 in the long grasslands, about 4 kg ha−1 yr−1 in the short grasslands, and about 259 kg ha−1 yr−1 in the kopjes. While this is much less than the amount consumed by the total grasshopper and termite populations, and considerably less than that consumed by the total mammalian herbivore population, it is nonetheless significant. 4 Single-species Populations The next three chapters look at the numbers of savannah organisms, mainly large mammals, at the single-species, two-species, and multiple-species level. This division is obviously artificial, because if you are trying to determine what influences the numbers of, say, wildebeest, one of the factors is grass and you immediately have a two-species situation. Another major influ- ence would be the disease organism rinderpest, and buffalo may compete for food, and so on. However, this artificial division is not just convenient, and traditional in ecology, but necessary. The reason is simply one of detail. If you are looking at a specific animal population in order to try and under- stand its numerical changes, and possibly predict future numbers, you need to know many details. You need good estimates of population size, for several consecutive years. This in itself will be very time consuming. You need details of birth rates and death rates, and how they change over time. You need to know what agents are responsible for these rate changes, such as food short- age, disease, and predation. The final description may well involve a detailed mathematical model or computer simulation of the population. Only a few single-species populations of savannah animals have received this amount of detailed attention. For single species populations we can therefore talk about population estimation, birth rates, death rates and such things as k factor analysis. We discuss these topics in this chapter. Once you have two interacting species the amount of detail, and work required, more than doubles and even fewer detailed studies are available. At this level of detail, ecologists talk about types of interactions between spe- cies, such as competition, predation and mutualism. Many of the descrip- tions, or models, of these interactions are more general, with less specific detail and less predictive power. Ecologists look more for evidence of inter- actions, rather than attempting detailed description, and speculate on their past or future effects. This approach to savannah organisms is dealt with in Chapter 5. With more than two species, the approach is frequently even less analytical and more descriptive. Ecologists now talk about species richness (numbers), The Biology of African Savannahs. Second Edition. Bryan Shorrocks & William Bates. © Bryan Shorrocks & William Bates 2015. Published 2015 by Oxford University Press. 138 THE BIOLOGY OF AFRICAN SAVANNAHS species diversity (numbers plus frequency), body size–abundance relation- ships and species–area relationships, and try to formulate ‘assembly rules’ (Diamond 1975, Prins and Olff 1998). All these approaches are an attempt to reduce the complexity of analysing many species, and to search for repeated natural patterns that will make sense of nature. We do this for savannah eco- systems in Chapter 6. This chapter , on savannah populations, is divided into four sections: how population size is estimated; what regulates populations, using two detailed studies; what regulates populations, using two population models; and a sur- vey of the possible regulatory factors for other species. Estimating numbers The aim of this section is to give readers an appreciation of some of the com- mon methods that have been used to estimate animal and plant numbers in savannah ecosystems. Many of these numbers will inform our discussions in this and later chapters, and it is useful to have an appreciation of the effort required and the accuracy of the estimates obtained. An appreciation of the techniques is the aim of this section, not a detailed introduction that would allow the reader to estimate population size. For a more detailed coverage of these techniques, the reader should look at Norton-Griffiths (1978), Wilson et al. (1996), McCallum (2000), Buckland et al. (2001), Williams et al. (2002), and Buckland (2004). Population estimates are conventionally divided into absolute population estimates (e.g.10 lion km−2) or relative population estimates (e.g. more buf- falo in that area than in this area). Absolute estimates may be shown as either population density (numbers per area), or population size (total numbers in an area). Since population size divided by area equals population density, it may seem that the two measures are essentially the same. In most cases they are. However, with some conservation issues it is important to appreciate that they are not always equal. For example in Figure 4.1, if each circle rep- resents a single individual (animal or plant) and its spatial position, then a curious dilemma may occur. In A the population size is 48 and in B the popu- lation size is 25, but A occupies a much larger area than B. The population size is greater in A than B, but density is clearly greater in B than A. If A and B were two species, which is the one more in need of conservation measures? Which is rarer? In most cases the population sizes, or densities, obtained are estimates. It is very rarely that all animals or plants can be counted. Of course if individuals can be recognised repeatedly, when observed on subsequent occasions, then very good population data can be obtained. With static organisms, like trees SINGLE-SPECIES POPULATIONS 139 (a) (b) Figure 4.1 Population size versus population density. A dilemma in measurement? for example, this usually involves simply tagging the individual tree with a marker and recording its position with a global positioning system (GPS). Birkett (2002) carried out this type of census on individual whistling thorn (Acacia drepanolobium) trees in Sweetwater Game Reserve, in the Laikipia area of Kenya. In this particular study it was the growth rate of the trees, and how this was affected by browsing, that was of interest, not the population size or density of the species. Trees were selected to give a height-stratified sample. Random trees were chosen, within all the available height categories. Some trees were in an area protected from browsers (elephant, giraffe and black rhinoceros) and some were not. Birkett found that in the protected area, the 78 selected trees had a mean annual growth rate of 19.1 ± 2.1 cm, while the 879 selected trees in the unprotected areas had a mean annual growth rate of only 7.5 ± 0.5 cm. With animals, that are mobile, identifying individuals can be more difficult although not impossible. Schaller (1972), in his study of Serengeti lions, found that he could recognize about 60 lions by their individual natural markings such as a missing ear, tail or other deform- ity. However, many more lions were marked by placing coloured metal tags in their ears. In a study of cheetahs, Caro (1994) used the pattern of spots to identify individual cheetahs. He took black and white photographs of each side of the face and tail, which he kept in his vehicle, so that animals could be matched in the field. In Kruger National Park, Maputlaet al. (2013) used the unique spot patterns on the flanks, legs and face of leopards to iden- tify 14 idividuals. In Mole National Park, in northern Ghana, bushbuck were uniquely identified using a combination of their spot and line body markings (Dankwa-Wiredu and Euler 2002). Some animals, such as zebra, have the equivalent of barcodes on their body and these have been used to identify individuals (Briand Peterson 1972). One of us has developed a similar coded system for identifying reticulated giraffe in Laikipia, Kenya (Shorrocks and Croft 2006, 2009) (Figure 4.2). 140 THE BIOLOGY OF AFRICAN SAVANNAHS Head end Acute angle = A Mane Right angle = R Obtuse angle = O Figure 4.2 Right-hand side of a reticulated giraffe showing neck pattern, and diagram illustrating the three types of line recorded. The giraffe in the photograph has the code ROAARAAROO (from Shorrocks and Croft 2009). The ‘neck pattern’ is converted to a ‘neck code’, written for each individual. The neck pattern that is coded is the section that runs along the back of the neck, adjacent to the mane. The yellow lines, of the giraffe’s pattern, that reach the mane are coded according to the angle they make with the mane, starting at the head end. Three types of line are recognized: right-angled lines (R), acute-angled lines (A) and obtuse-angled lines (O). Obviously, each giraffe has two codes, one for the right-hand side of the neck, the other for the left. The right-hand code for the giraffe in Figure 4.2 is ROAARAAROO. With three types of line (R, A and O) and 10 positions on the neck there are 310 (= 59,049) permutations. If both sides of the neck are observed there would be 320 or 3,486,784,401 permutations. The advantage of this type of linear code is that it can be stored in a computer database, and new sightings checked against the stored codes each day. This computerised searching is far more efficient than looking through photographs and allows hundreds of individuals to be recognised. Photo-identification software such as Wild ID ( in the nest; small mammals are caught in live traps; and large mammals are usually shot with an anaesthetic dart. One way of permanently marking large mammals (marks must be permanent otherwise you think there are more animals than there really are) is to cut notches in their ears. In Nairobi National Park, the Kenya Wildlife Service (Torchio and Manconi 2004) used a coded system of 6 notches (Figure 4.3) which allowed them to uniquely mark up to 65 black rhinoceros. Kruuk (1972), studying spotted hyaenas in the Ngorongoro crater, Tanzania, used a similar system of ear notches to rec- ognize up to 51 hyaenas. The Kenya Wildlife Service black rhino work was also involved in translo- cating surplus rhinos from Nairobi National Park to a rhino sanctuary in northern Laikipia. These translocated animals were marked using a more hi- tech, but much more expensive, method that also allowed them to be easily located once released. After a rhino had been anaesthetised using a dart, the tip of the horn was sawn off. A broad cavity was drilled into the base of the horn and a narrow hole drilled from the top of the horn into this basal cavity. A small radio transmitter (8 × 8 × 2 cm) was placed into this basal cavity, with its antenna projecting up through the narrow vertical hole. The openings in the horn were then sealed with resin. As a wholly insensitive appendage, 20 10 12 40 4 Rhino 34 Rhino 61 (20+10+4) (40+20+1) Figure 4.3 Top: diagram of black rhinoceros head showing the position of of the six notches, and their assigned number, used by the Kenya Wildlife Service. Bottom: two diagrams showing notch combinations used in two individuals (after Torchio and Manconi 2004). 142 THE BIOLOGY OF AFRICAN SAVANNAHS a rhino’s horn is the perfect receptacle for a sophisticated location-transmitting device. The embedded transmitter was preprogrammed to emit a unique signal, at specified time intervals that could be picked up within a radius of 7 km. More usually, radio transmitters are attached to collars which fit around the animal’s neck. Although older radio collars were quite big, and only used on large mammals, modern radio collars can be quite small. In Comoé National Park, Ivory Coast, kob antelopes were fitted with nylon radio collars (total weight about 200 g) and located twice a day by triangulation (Fischer and Linsenmair 2001). In Queen Elizabeth National Park, Uganda, Hoffmann and Klingel (2001) were able to fit 2.5 g radio collars to rodents Lemniscomys( striatus), equivalent to only 5–7% of their body weight, and tiny radio ear- tags have been used successfully on bushbuck (Dankwa-Wiredu and Euler 2002). The most hi-tech development is the attachment of satellite trans- mitters to radio collars. In north-western Namibia, Lindeque and Lindeque (1991) attached 12 kg collars, containing both traditional VHF transmitters and UHF satellite transmitters, to seven elephants. A single individual with- in a family unit was collared, so that in effect seven groups of elephants were tracked. Home ranges were found to vary between 2302 km2 and 10738 km2. Hensman et al. (2014a, 2014b) fitted collars to sable antelope in the Oka- vango delta, Namibia. These collars had a GPS unit, UHF and VHF trans- mitters, and batteries designed to last a year. A single female, in each of three sable herds, was collared. Data stored in the collars was downloaded peri- odically, using a UHF modem with VHF R-1000 telemetry receiver that was used to get within 500 m of the herd. When animals are marked, released, and recaptured (or otherwise re- examined) at a later date, an estimation of their population size/density can be made. If (1) marked and unmarked animals behave in the same way (marks do not affect capture) and (2) the population is closed (no input or loss), then the ratio of marked individuals to the total population size is equal to the ratio of marked recaptures to the total size of a second sample. This can be written as P = n2 n1 m where P is the unknown population size, n1 is the size of sample 1, n2 is the size of sample 2, and m is the number marked in sample 2. This is known as the Lincoln–Peterson estimate (Seber 1982). The main problem of this method of population estimation involves the sec- ond condition above; that the population remain closed between the two sampling events. This closure includes not only migration of individuals, but birth and death. One solution to this would be to take the second sample immediately after release of the first, marked, sample. However, this would SINGLE-SPECIES POPULATIONS 143 inevitably mean violating condition (1), because the marked individuals would be more likely to be recaptured. For this reason, the two sampling events (marking and recapturing) are usually at least two consecutive days apart. To overcome the problem of input and output, the rate of gain or rate of loss must be estimated. If either of these can be done, then a suitable adjust- ment can be made to provide a correct estimate of population size. To do this, at least one more sampling of the population must be done, and individuals caught on different days must be marked in a distinctive way. However, this is only a ‘day’ mark and individuals need not be identifiable. Details of these multiple mark–release–recapture methods can be found in Seber (1982). Intriguingly, one of the first people to use these multiple recapture meth- ods was Jackson (1933) who employed it for estimating populations of tsetse flies in East Africa. Most uses of this technique in savannah systems have employed the simple Lincoln–Peterson estimate, and the results must therefore be viewed with some caution. Kruuk (1972) marked 51 spotted hyaenas in the relatively small Ngorongoro crater, northern Tanzania. By subsequently resighting these marked individuals (recaptures) he estimated that there were about 385 adult hyaenas in the crater. In the much larger area of the Serengeti ecosystem he marked 200 hyaenas, but obtained a much less precise estimate of about 3000. Thompson (1989) estimated the popu- lation of lovebirds (Agapornis sp.) living around Lake Naivasha, in Kenya. Mist-nets were used to capture the birds, which were marked with a leg ring. Because the marking of the lovebirds was quite time consuming, the two sampling occasions were quite separated, the first (n1) between August and December 1985, and the second (n2) between February and June 1986. Numbers caught and ringed on the first occasion n( 1) = 412, while n2 = 255 and m =15. This gave a population estimate of 7004 birds. Maputla et al. (2013) estimated that a leopard population in Kruger National Park con- sisted of 19 individuals, based on mark–recapture photographic data from motion-sensitive camera traps. Another method of estimating population size that uses marked, or natu- rally identifiable, animals is the removal method. Animals are not, in prac- tice, removed (because they would be replaced by new animals moving into the area), but simply identified. On each successive sampling occasion the number of ‘new’ unmarked (unknown) animals will decrease until all the animals have been marked. For example, Wittemyer (2001) observed ele- phants in the Samburu and Buffalo Springs National Reserves, Kenya. Each elephant within the study area was identified using sex, age and features unique to the individual. Photographs and drawings of these features were used to compile an identification dossier. Over a period of 21 consecutive months (from November 1997) population numbers were recorded (Figure 4.4a). New individuals were recorded each month until the last 3 months when the numbers of identified elephants levelled out at 744—the estimated population of elephants using these two reserves. Of course once again, the 144 THE BIOLOGY OF AFRICAN SAVANNAHS (a) (b) 800 250 600 200 150 400 100 200 New animals 50 Number of elephants 0 0 1998 1999 0 200 400 600 800 Consecutive months Total previously seen Figure 4.4 Removal sampling of elephants in Samburu National Reserve in Kenya. (a) Original data: black, newly identified; grey, previous total (from Wittemyer 2001). (b) Population estimation using the regression method on data points 2–11. method only works if the population is confined to the area in question. Otherwise you are estimating the population of an unknown area. Addi- tionally, it is not usually necessary to continue observing the population until every animal has been recorded. Since the number of ‘new’ animals drops progressively, it is possible to extrapolate to ‘no new’ animals. Figure 4.4b shows this extrapolation, using only the trend from the 2nd to the 11th month’s data and a simple regression analysis. It predicts a total population of about 740 elephants, not very different from the figure obtained after a further year’s counting. Most population surveys in savannah habitats involve some type of transect. A line is travelled at a constant pace, by the observer, and all animals seen are counted. Sometimes a straight line transect is walked on foot, as was done by Dangerfield (1990) estimating termite (Cubitermes sankurensis) numbers in a miombo woodland savannah in Zimbabwe. More frequently a road transect is driven in a vehicle. This is often more convenient, and safer. The fact that the road transect is not a straight line doesn’t really matter; the transect method is the same. In effect, a strip on either side of the road is sampled. If the sample strip is a random sample of the study area, and if all animals within the strip are seen, then the estimation of density, or numbers, is relatively easy. For example, imagine that observers in a game reserve drive along its network of roads (20 km) observing all buffalo within 100 m either side of the road. They see 32 buffalo. The density (D) of the population would be D = n/(2aL), where n = the number of animals (32), a is half the strip width (0.1 km), and L is the length of the transect (20 km). This would equal D = 32/(0.2 × 20) = 8.0 buffalo km−2. In practice, it is never so simple. The width of the strip will rarely be constant. For example, if the transect runs through dense vegetation the observed strip will be narrower. If the transect runs through open grass savannah, the observed strip might be wider. One answer to this problem is to record the sighting distance to each animal, SINGLE-SPECIES POPULATIONS 145 and use this to estimate an average value for a. Population density is then estimated by n 11 D = ∑ 2Ln ri where n and L are as before, and ri is the observed distance to each sighted animal i. For example, Table 4.1 shows part of the data from a transect of white-eared kob (Kobus kob leucotis) in the Sudan. The transect length was 10 km. The estimated density is therefore 12 1 1 1 1 1 2 D = +++ + 3= .80kob perkm. 2(10)12 0.15 0.20 0.16 0.20 Although this estimate is one way to allow for variable strip width, it still assumes that all animals within the strip are seen. It assumes what is called a rectangular detection function (Figure 4.5, line a). In practice the probability of detection falls off the further an animal is from the line transect baseline. Line b in Figure 4.5 is the simplest such relationship. With this detection function, only half the animals within your transect strip would be seen, and in effect a is half the value you think it is. The constant a is therefore more than just the strip width. It estimates how wide the strip would be if every organism was seen and none were missed, and there are numerous ways of estimating it (Buckland et al. 2001). The shaded area in Figure 4.5 encloses the general zone for detection functions for wildlife populations. Eltringham et al. (1999) used a road transect (393 km) to estimate numbers of large mam- mals in the Mkomazi Game Reserve (2,850 km2), northern Tanzania. Four surveys, each lasting 3–4 days, were undertaken during the 1996 dry season when the animals were more easily seen because of the sparse vegetation. They were able to estimate the numbers shown in Table 4.2, which should be regarded as minimum estimates. Shorrocks, Cristescu and Magane (2008) used distance sampling to estimate densities of zebra, impala and dik-dik in Laikipia, Kenya. Dunham and Du Toit (2013) sampled 36 transects, total- ling 94.15 km, four times over 2 days. Using DISTANCE analysis (Buckland et al. 2001) they were able to estimate detection functions for nine species, and therefore their densities. Table 4.1 Sighting distances for white-eared kob, from a line transect in the Sudan Animal number 1 2 3 4 5 6 7 8 9 10 11 12 Sighting distance ri 150 200 160 200 250 130 150 130 200 100 140 200 Modified from Krebs (1999). 146 THE BIOLOGY OF AFRICAN SAVANNAHS 1.0 0.8 Line a 0.6 0.4 Line b Probability of detectio n 0.2 0.0 0 2 4 6 8 10 Distance from line (m) Figure 4.5 Detection functions of a line transect survey (modified from Krebs 1999). The type of transect most associated with savannah is probably the aerial survey (Norton-Griffiths 1978). Rather than walking or driving the sampled strip, it is flown over, in a light aircraft, and the animals counted from the air. Sometimes, particularly if the animals are in groups, they are photo- graphed and counted later. The technique was pioneered by Bernhard and Michael Grzimek (Grzimek and Grzimek 1960a, 1960b), who used it in the Serengeti ecosystem in the late 1950s. The aerial method remains essentially Table 4.2 Minimum estimates of the numbers of some large mammals in Mkomazi Game Reserve in the 1996 dry season Species Numbers Eland 473 Elephant 314 Gerenuk 933 Giraffe 979 Grant’s gazelle 306 Impala 3,564 Dik-dik 55,978 Kongoni 840 Lesser kudu 5,739 Steinbuck 554 Warthog 1,460 Common zebra 1,438 From Eltringham et al. (1999). SINGLE-SPECIES POPULATIONS 147 unchanged today (Foguekem et al. 2010), although the statistical analysis surrounding it has become something of an industry. Only aircraft with high wings give an unobstructed downwards view. An observer sits in the back of the aircraft and, using markers placed on the window and wing struts to delimit a transect strip on the ground, counts the animals visible. Figure 4.6 illustrates how these markers are positioned while the aircraft is on the ground. If W is the required strip width during the survey and H is the flying height chosen, thenw (the strip width on the ground) is obtained from wW=×hH/. This width,w , can then be marked on the hanger floor and the position of a′ and b′ (on the window) and a and b (on the wing strut) can be adjusted until they are in line with the margins of this width (A and B). The pilot simply has to fly at height H for the observer to ‘see’ a transect of the required width on the ground below. In East Africa such aerial surveys normally travel at 160 km/h or simply use the still accepted km/h, at a height of 100 m, giving a transect width of 150 m. In open terrain, and/or for conspicuous animals, the height can be greater; in more wooded terrain, and/or for less visible animals, the height can be lower. Two observers can be employed, one on each side of the aircraft. Cameras can also be fitted to the underside of the aircraft and photographs taken of a transect strip directly below, rather than to each side. Of course, not all the savannah area being studied is usually flown over. This would require many hours of expensive flying time. The study area is sampled and the total population, or density, is estimated (Parker et al. 2011). This is a similar procedure to that employed in the road transect estimates above. A suitable baseline is usually fixed through the area to be sampled and random Eye b' a' Wing strut b Window a h A B W Figure 4.6 How the positions of window and strut markers are fixed while the aircraft is on the ground (after Norton-Griffiths 1978). 148 THE BIOLOGY OF AFRICAN SAVANNAHS flight transects chosen at right angles to this baseline. This procedure is well illustrated by the counting of the Serengeti migratory wildebeest, carried out by Norton-Griffiths (1973). Figure 4.7 shows the migrating wildebeest herds (shaded areas), the survey baseline through these herds, and the random- ly chosen transect flight paths (1–31) at right angles. These random flight transects were chosen before the flight, with the aid of a large-scale map, and in the event transects 32, 33, and 34 were not required because there were no herds under these transects. In this particular example very large numbers of animals were involved and the observer operated a 35 mm camera mounted on the door sill of the aircraft. Along each transect, photographs were taken every 10 seconds of flying time so that two levels of sampling were in fact employed. In all the 31 transects flown, the estimated number of wildebeest was 88,884, and the total population estimate (all the wildebeest in the shad- ed areas in Figure 4.7) was 754,028, with 95% confidence limits of ±8.5%. A slightly different type of aerial survey has been carried out in the Addo Elephant National Park, South Africa (Whitehouse, Hall-Martin and Knight 2001). Annual helicopter-based total counts have been conducted since 1978. Seronera 4 3 N 1 2 5 6 7 8 9 National Park 10 11 13 boundary 12 14 15 16 20 17 21 24 22 19 23 18 base line 26 25 27 28 29 30 31 Moru Kopjes 32 33 34 Not counted Naabi Hill 0510 km Figure 4.7 Orientation of the baseline and of the random transects in the May 1971 sample count of the migrating Serengeti wildebeest (from Norton-Griffiths 1973). SINGLE-SPECIES POPULATIONS 149 Parallel strips are flown at a ground speed of approximately 80–120 km h−1, with strip width and flying height varying with vegetation and weather conditions. Unlike the light aircraft surveys, herds were circled to facilitate counting. Helicopters were also used to estimate former and curreent breed- ing territories of the bearded vulture (Guptaetus barbatus meridionalis) in southern Africa (Krüger et al. 2014). These aerial surveys were conducted by flying approximately 100–500 m from the cliff edge (the vultures’ nest- ing sites) at an average flying speed of 70–80 km h−1 with a minimum of two observers looking out of the port side of the aircraft. Aerial surveys usually underestimate the populations on the ground (Caugh- ley 1974, Craig 2004). The proportion seen declines with flight speed and observer inexperience. In the Addo elephant count, above, it was estimated that about 7% of adults were missed by the helicopter flights. However, for calves it was much higher, with about 48% remaining undetected. Declining numbers and distributions of many wildlife populations call for data at ever increasing levels of spatial resolution, prompting questions about the precision of sample survey estimates. While sample surveys are less costly, total surveys are thought to be more reliable. Georgiadis et al. (2011) made a direct comparison between these two methods for wildlife in Laikipia district, Kenya, discussing their strengths and weaknesses. They found that neither total surveys nor sample surveys yield accurate popula- tion estimates. Both methods in fact tend to underestimate numbers, with the bias varying among species, observers, and habitats (see also Caughley 1974, East, 1999, and Redfern et al. 2002). Undercounting appeared to be slightly greater in the total count than in the sample survey, but, on average, differences were not large. However, comparison factors, such as cost and the scope of information provided, revealed huge differences in favour of sample surveys. Although aerial surveys with light aircraft remain the most commonly used technique for counting large mammals in African savannahs, there are drawbacks. These include safety issues (Wattset al. 2010) and logisti- cal issues, namely the lack of airfields or even appropriate aircraft in some areas of Africa (Vermeulen et al. 2013). Moreover, these operations are quite expensive for most African wildlife managers, and it is therefore dif- ficult to plan long-term and regular surveys. As a result, in many African protected areas, the interval between two successive surveys can often be as great as 10–25 years (Dunham 2012). This makes it difficult to quan- tify accurately any short-term change in wildlife populations (Ferreira and Aarde 2009). As a result, some populations may collapse between two sur- veys because no appropriate action had been taken (Bouché et al. 2011, 2012). Unmanned aerial systems (UAS), using small drones, may be more cost-effective, safer, and capable of reaching more remote areas (Lisein et al. 2013). These authors tested the accuracy of drone counts in Nazinga Game 150 THE BIOLOGY OF AFRICAN SAVANNAHS Ranch (940 km2) in Burkina Faso, using a small fixed-wing UAS, equipped with an autopilot and digital camera, and a ground control station com- prising a tablet computer. Transect surface area estimation was found to be good, although no actual animals were counted in this trial. However, the use of such systems in big mammal surveys opens up interesting prospects for the future. Another hi-tech development in ‘counting’ mammals (also used in behav- ioural studies), is the use of camera traps (Swann, Kawanishi and Palmer 2010, O’Connell, Nichols and Ullas 2011, WWF 2011, Abu Baker and Brown 2014). These are placed in fixed transects or grids and are remotely activated by a motion sensor, or an infrared sensor, or a light beam. They are usu- ally baited to attract the mammal in question. For example, in the duiker example (Abu Baker and Brown 2014) the bait was alfalfa-based pellets, and camera photos were each given an ID: date, time, species, number of indi- viduals, sex, posture and specific individual whenever possible (marks such as chipped horns). So far, estimating animal populations has involved counting animals. However, ecologists frequently count the signs left by animals rather than the animals themselves, such as footprints, feeding signs and dung. In African savannahs, and particularly with elephants, counting dung is a frequently employed method of population estimation. An example is the survey of Jachmann (1991) carried out in the Nazinga Game Ranch, Burkina Faso, in western Africa. The vegetation is a tall-grass tree and shrub savannah with Vitellaria paradoxa, Terminalia spp., Combretum spp., Acacia spp. and Detarium microcarpum as the dominant trees. Jach- mann estimated the elephant population by four methods, unfortunately not all in the same year: aerial counts of elephants (both total and sample), line transect foot surveys of elephants, vehicle road surveys of elephants and a dung count using quadrats. The ‘dung’ quadrats were positioned along the line transect used for the foot survey. Of course, simply count- ing the number of dung piles is not enough to estimate the elephants producing them. Jachmann had to estimate two other things: the rate of dung production and the rate of dung decay. Defecation rate was esti- mated by following elephants on foot for several hours. Dry-season def- ecation rate was estimated at 14.1 droppings per elephant per day, while the wet-season rate was estimated at 27.2 droppings per elephant per day. By watching 31 droppings, over several weeks, the decomposition rate was estimated to be 0.59% per day. Table 4.3 shows the estimates of popula- tion size obtained by the various methods. Using the total aerial survey as the best estimate, we might believe that the dung count gives a very good estimate of the elephant population at Nazinga. A study by Olivier, Fer- reira and van Aarde (2009) used dung surveys to estimate population size, and extracted an age structure from bolus diameters for the elephants liv- ing in the Maputo Elephant Reserve, Mozambique. Their estimates were SINGLE-SPECIES POPULATIONS 151 Table 4.3 Elephant population estimates in the Nazinga Game Ranch, Burkina Faso Survey method Year Population estimate Aerial (100% coverage) 1989 366 Aerial (6.1% coverage) 1988 610 Foot transect 1987 487 Foot transect 1988 306 Vehicle road transect 1988 293 Dung count 1987 353–396a a Two different assumptions about dung dynamics. From Jachmann (1991). based on published defecation rates, dung decay rates, distance sampling techniques and 1,672 dung piles encountered on 204 line-transects. Com- parison with a 1995 dung survey (De Boer et al. 2000) suggested that the elephant population was stable. In the next section we will look at what factors change the numbers of animals and plants in savannah ecosystems. However, before doing that it is appropriate to pause and ask what these numbers, estimated by the techniques described above, can be used for. We think there are two main uses. First, these numbers inform population models such as those described in the third section of this chapter. These models are an attempt to describe, and understand, the ecological processes that determine population numbers. Without the estimated numbers of real populations we would have nothing to compare our predictions to. Secondly, they alert managers, ecologists, and conservationists, to changes in popula- tions (trends) that we should know about. We will therefore end this first section, by describing a study that looked for such population trends, in savannah animals. The Masai Mara ecosystem comprises the Masai Mara National Reserve and the adjoining group ranches (Figure 4.8). The National Reserve is a formal conservation area owned by the Government of Kenya, while the group ranches are privately owned. The National Reserve and the Mara plains are mainly Themeda grassland, the Loita plains are mainly dwarf shrub and Acacia drepanolobium grassland and the Siana plains are main- ly Croton bush and other woody species interspersed with grassland. The main land uses are pastoralism, tourism and agriculture. Since 1977 the Department of Resource Surveys and Remote Sensing has carried out sys- tematic survey flights to estimate the wildlife populations. Ottichilo et al. (2000) used census data spanning the years 1977–1997 to look for popula- tion trends in several species. These data are shown in Figure 4.9. Those graphs with log linear regression lines have significant trends (P < 0.05), all of them downwards. 152 THE BIOLOGY OF AFRICAN SAVANNAHS Kenya Masai Mara Ecosystem Mara Plains Loita Plains Siria Escarpment Masai Mara National Reserve Loita Hills N Kenya–T W E anzania Border Siana Plains S 0 100 200300 400 km Figure 4.8 Location of the Masai Mara ecosystem (from Ottichilo et al. 2000). Declines for individual species were: warthog 88%, buffalo 82%, giraffe 79%, eland 76%, waterbuck 76%, topi 73%, Coke’s hartebeest 66%, Thom- son’s gazelle 62% and Grant’s gazelle 52%. There were no significant downward trends in elephant, impala, and ostrich. Livestock populations (cattle, sheep, goats, and donkeys), which were also examined, did not show any significant downward trend. The decline of most wildlife in the Masai Mara ecosystem, seen in Figure 4.9, could be due to a number of factors, acting individually or in combination. These include count- ing errors and biases, climatic effects, habitat changes, competition for forage, and poaching. The authors concluded that counting errors and climate can be discounted, although this is difficult to assess. Counting errors tend to increase the variation around the population estimate rath- er than introducing a systematic effect such as that observed here. Also, the method of data collection did not change over the 20 years. There were severe droughts in this part of Africa during 1984, 1986, and 1993 which would have affected grass growth (Chapter 2) and potentially herbivore numbers, but no effect could be detected in the data. A combination of habitat changes, competition, and poaching, all of which are known to be present, could be responsible. These possibilities are revisited in later chapters. Sadly, Ogutu et al. (2011) report that the Mara situation has not improved. They estimated losses at 95% for giraffes, 80% for warthogs, 76% for hartebeest, and 67% for impala. SINGLE-SPECIES POPULATIONS 153 45,000 8,000 5,000 7,000 36,000 4,000 6,000 5,000 27,000 3,000 4,000 18,000 3,000 2,000 Numer of eland Number of buffalo 2,000 9,000 Number of elephant 1,000 1,000 0 0 0 76 82 88 94 76 82 88 94 76 82 88 94 Year Year Year 25,000 1,60,000 7,000 6,000 20,000 1,20,000 5,000 15,000 4,000 80,000 10,000 3,000 2,000 40,000 Number of giraffe 5,000 Number of Grant's gazelle 1,000 Number of Thomson's gazell e 0 0 0 76 82 88 94 76 82 88 94 76 82 88 94 Year Year Year 1,20,000 11,000 1,500 1,00,000 8,800 80,000 1,000 6,600 60,000 4,400 40,000 500 Number of impala Number of ostrich Number of hartebeest 2,200 20,000 0 0 0 76 82 88 94 76 82 88 94 76 82 88 94 Year Year Year 51,000 8,100 3,000 7,200 42,500 6,300 34,000 5,400 2,000 4,500 25,500 3,600 17,000 2,700 1,000 Number of topi Number of warthog 1,800 Number of waterbuck 8,500 900 0 0 0 76 82 88 94 76 82 88 94 76 82 88 94 Year Year Year Figure 4.9 Population trend for 12 wildlife species in the Masai Mara ecosystem between 1977 and 1997. Those graphs with trend lines are statistically significant (modified from Ottichilo et al. 2000). What changes numbers? The simple answer to this question is of course birth, death, immigration and emigration. The movement of savannah animals can often cause dra- matic local, and seasonal, changes in numbers. We look at some examples 154 THE BIOLOGY OF AFRICAN SAVANNAHS of these animal movements in Chapter 5. However, population ecologists have often concentrated on birth and death, and we will examine these here. Birth and death are frequently age-specific, that is, the chance of an animal dying, or of a female giving birth, changes with its age. Such information can be summarised in a life-table, or its graphical counterpart (e.g. Figure 4.10 and Figure 4.12b). Unfortunately, with savannah animals, such detailed information is rarely available and the subsequent discussion of what influences population size is consequently less precise. For example, with mortality, the important dis- tinction is often between what simply ‘kills’ and what ‘regulates’. In the case of most savannah animals we only have lists of the former, which although very interesting do not illuminate the causes of population regulation. We will therefore concentrate initially, on two studies where quite detailed information is available. We believe these build a solid framework for the less precise lists of mortality agents seen in the literature. These two stud- ies involve African buffalo Syncerus( caffer) and wildebeest (Connochaetes taurinus), and both studies have been carried out by Tony Sinclair and his colleagues in the Serengeti ecosystem of East Africa. However, before look- ing at these two studies let me briefly return to this population problem of what kills and what regulates. The distinction is between ‘density inde- pendent’ and ‘density dependent’ mortality factors. Both types of factor can change the numbers of animals in a population, but only density depend- ent factors will regulate numbers. With density independent mortality the 3.5 1.0 3.0 0.8 2.5 0.6 2.0 Fecundity 1.5 0.4 Survivorship 1.0 0.2 0.5 0.0 0.0 0 24 6810 12 Age class (years) Figure 4.10 Age-specific fecundity and survival in the African wild dog Lycaon( pictus) in the Selous Game Reserve, Tanzania (from Creel and Creel 2002). SINGLE-SPECIES POPULATIONS 155 probability of an individual dying remains the same irrespective of density. With density dependent mortality the probability of death changes with den- sity, usually being higher in dense populations. Notice that this statement involves the probability, or chance, of dying, not the numbers dying. If 10 buffalo are killed by lions in a population of 100 buffalo, and 100 are killed in a population of 1000, this is not an example of density dependence. For any individual, the chance of being eaten by a lion is 10%, and the same in both populations. The difference can be visualised by thinking of two different ways of changing the temperature of a room. If the room has a heating system controlled by a thermostat, this can be set at the required temperature. When the temperature of the room drops below this point the thermostat switches on the heating system, and the room temperature increases. When the tem- perature of the room rises above this point the thermostat switches off the heating system, and the room temperature decreases. The room temperature is regulated by this negative feedback mechanism, and maintains a more or less constant temperature. Alternatively, someone could periodically open a window to cool down the room if it got too hot (assuming the room is in a cool climate) but this would not be so precise, and if done without any con- sideration of the actual room temperature would produce a very erratic level of heating. Because many savannah populations show long-term stability, it is often assumed that they are regulated by density dependent factors. African buffalo Between 1965 and 1972, Sinclair (1974, 1977) studied buffalo populations in the Serengeti National Park, East Africa. During this period, census informa- tion suggests that buffalo were increasing in numbers, probably as a result of the removal of rinderpest (see Chapter 5) (Figure 4.11a). This, as we shall see later, was even more pronounced in the Serengeti wildebeest population (see Figure 4.15). There is also some suggestion that buffalo numbers started to level off in the early 1970s. What caused this? In Chapter 2 we saw that there is a close relationship between annual rainfall and the biomass of green grass produced in East African savannah ecosystems (Figures 2.1 and 2.11). In East Africa, for post-rinderpest populations, there is a similar close relationship between buffalo density in an area, and mean annual rainfall (Figure 4.11b). Variation about the regression line in Figure 4.11b is probably caused by the extent of permanent water in the dry season. This is particularly noticeable in the case of Lake Manyara National Park which contains a number of per- manently flowing springs as well as extensive swamps and lakeshore alkaline grasslands which increase the food supply to buffalo beyond that produced by rainfall. Figure 4.11b therefore suggests that buffalo are ‘regulated’, in most places, by rainfall acting through their grass food supply. Sinclair measured recruitment as the number of yearlings (animals between 1and 2 years old) per 100 adult females. During the period of study it 156 THE BIOLOGY OF AFRICAN SAVANNAHS (a) 70 (b) 25 3 60 20 50 2 15 40 10 No/Km No. Buffalo´x10 30 5 20 1960 1965 1970 500 1000 1500 2000 Rainfall (mm) Figure 4.11 Buffalo in East Africa. (a) (●) Population increase in the Serengeti between 1958 and ■ 1973; ( ) population model of Serengeti buffalo using constant kf and kj, and density- dependent ka. Before 1964 an additional constant is added to juvenile mortality, representing mortality from rinderpest. (b) Relationship between buffalo density and annual rainfall; regression line excludes the Lake Manyara data point (▲) (both graphs modified from Sinclair 1977, with permission of Chicago University Press). fluctuated around an average of about 38 (32–43), suggesting that the observed increase in population was due to a change in mortality, rather than ‘birth’. Mortality in adults was density dependent (the regression line in Figure 4.12a) and age-specific (Figure 4.12b). In the latter case, notice that it is juveniles and ‘old’ buffalo that have higher mortality rates. It is also reveal- ing that the pattern of mortality in adults is rather similar between the sexes, and between the two study areas. This suggests similar mortality factors even though these two areas are over 100 km apart. As stated at the beginning of this section, changes in population numbers are due to birth, death, immigration and emigration. Understanding popu- lation change is therefore a matter of understanding the difference between potential fecundity and actual birth, and the difference between numbers at consecutive times in the life history (mortality). Within any generation, this can be visualised (Figure 4.13) as a series of consecutive, declining numbers (N), with the differences between them either representing a reduction in fecundity (N0 – N1), or a reduction in individuals after birth (mortality). Reduction can be recorded in different ways. For example, mortality 1 could simply be written as the number of deaths. Simply using numbers has the advantage that total deaths can be simply obtained by summing m1, m2, m3, etc. The disadvantage is that 10 deaths in a population of 100 is more intense than 10 deaths in 1000. Simple numbers are no guide to intensity, which is crucial in revealing the presence of density dependence. Mortality can also be represented by a mortality rate, which does reveal intensity (Figure 4.12). SINGLE-SPECIES POPULATIONS 157 (a) (b) 25 1.0 20 15 0.5 10 % Mortality Mortality rate 5 50 60 70 51015 Number´x103 Age in years Figure 4.12 Mortality in buffalo in the Serengeti. (a) Mortality as a percentage of population size immediately preceding it in June. The regression line is for mortality of adults (closed circles). Open circles are juvenile mortality. (b) Age-specific mortality rate for female buffalo (circles), northern males (squares) and southern males (triangles). The northern study area was around Kogatendi, the southern study area around the Moru Kopjes (modified from Sinclair 1977, with permission of Chicago University Press). N N0 N1 N2 3 etc. Reduction Mortality 1 Mortality 2 Mortality 3 of fecundity k value 1 k value 2 k value 3 k value 4 Figure 4.13 Schematic diagram showing the reduction in population numbers over one generation. This is simply the number dying out of the number available to die. Rates, however, cannot be simply added together to get an idea of the total mortal- ity rate over the lifetime of the animal. A third way of representing mortality, or reduction, combines both the good qualities of these last two measure- ments. It is called ‘killing power’ (or k value) and is simply log10Nx – log10Nx−1 (Haldane 1949, Varley and Gradwell 1970). It can be summed to give total mortality and it is a rate (on a logarithmic scale). Note that k values can also be calculated if consecutive numbers change because of migration. In his buffalo study, Sinclair was able to calculatek values for three reductions in population each year. These were a decrease in potential fecundity (birth), 158 THE BIOLOGY OF AFRICAN SAVANNAHS kf ; mortality of calves under 1 year old (juvenile mortality), kj; and mortal- ity of animals over 1 year old (adult mortality), ka. The first of these was the difference between the number of calves born if all females reproduced and the observed number of pregnancies. By plotting these k values for each year, against log numbers before the reduction, it is possible to detect the pres- ence of density dependence and therefore the potential causes of popula- tion regulation (Varley and Gradwell 1970). Sinclair (1974, 1977) did this for Serengeti buffalo (Figure 4.14a). This analysis helps to focus the search for the causes of regulation. It does not, however, specifically identify the agents of reduction. For Serengeti buffalo, the only reduction that showed a significant positive slope with density was adult mortality, with a regression slope of b = 0.2357 (P < 0.05). The points on the juvenile mortality graph showed so much scatter that the positive slope of the regression line (b = 0.1575) was not statistically significant. The effect of this juvenile variation is transmitted to the graph of combined juvenile and adult mortality, which also has a non-significant regression slope. The (a) (b) 0.16 Fertility loss Adult + juvenile 0.12 mortality 0.04 0.12 –0.04 Neonatal mortality 0.8 0.08 0.4 Adult mortality 0 0.04 Dry season calf mortality 0.8 0.4 k value Juvenile mortality k value 0.08 0 Yearling mortality 0.8 0.04 0.4 Fertility loss 0 0.04 0.15 Adult mortality 0.05 0 0 4.6 4.7 4.8 4.9 5.4 5.6 5.8 6.0 6.2 Log population size Log population size Figure 4.14 k factor analysis for Serengeti buffalo (a) and wildebeest (b) (modified from Sinclair 1974 and Mduma, Sinclair, and Hilborn 1999). SINGLE-SPECIES POPULATIONS 159 regression line for loss of fertility (b = 0.0449) is very close to zero, suggest- ing that this cause of reduction remains relatively constant with density. This analysis suggests, therefore, that adult mortality alone regulates Serengeti buffalo. To examine this possibility further, Sinclair used the results of thek factor analysis to construct a simple predictive model of population growth. Starting with the estimated buffalo population from the census of 1965 he extrapolated population size forwards to 1976 and backwards to 1958. He used the density dependent regression line, in Figure 4.14a for adult mortal- ity, and simply used the average k for juvenile mortality and fertility loss. This k factor model gave a very good fit to the forward data, predicting a buffalo population of 60,770 in 1974. The estimated observed population was 61,134. Although there are population losses due to reduced fertility and juvenile mortality, it is the density dependent losses of adult mortality that regulates the buffalo population to a predicted equilibrium of around 63,500. Of course, young animals are very sensitive to a variety of causes of mortality and therefore, although juvenile mortality does not regulate the buffalo population, it does produce fluctuations in population size from year to year (average kj is in fact greater than average ka). A factor that reduces populations in this way is called a ‘key factor’, and is defined as the factor that contributes most to total population reduction (Royama 1996). In Serengeti buffalo, total reduction K( ) = kf + kj + ka. One way to identify a key factor is to plot successive values of K against each of the separate reductions; when Sin- clair did this, juvenile mortality (kj) was the most important contributor to K. A simply way to express what is happening in these buffalo populations, from 1965 onwards, is to say that juvenile mortality causes fluctuations and adult mortality compensates for these disturbances in a density dependent way. The backwards fit of the k factor model, to 1958, was not very good. The model predicted a much faster rate of increase than actually happened between 1958 and 1965. Sinclair surmised that this might be due to the occurrence of rinderpest in these early years. The data used in thek factor analysis and in the model was collected in the years when rinderpest was no longer present in the Serengeti. We know that this introduced, exotic, disease caused repeat- ed mortality in juvenile buffalo and wildebeest. There is no estimate of this mortality for buffalo but in wildebeest 1.73x more yearlings died when it was present. Sinclair extrapolated this wildebeest figure to buffalo and added it to the mean juvenile mortality. This new value of kj was then used in the k factor model for the years before 1964. The new model’s predictions are shown in Figure 4.11a. The new model fits the real census data quite well, suggesting that fluctuating, but sometimes quite high, juvenile mortality and density dependent adult mortality are the major determinants of buffalo population growth. But what causes this mortality, particularly the important density dependent adult mortality? In Serengeti buffalo, as with most animals, the three main causes of mortality are predation, disease and under nutrition. The major predators of buffalo 160 THE BIOLOGY OF AFRICAN SAVANNAHS in the Serengeti are lions (Schaller 1972), and this also appears to be the case for other areas such as Lake Manyara, Tanzania (Schaller 1972, Prins 1996), Kafue Park, Zambia (Mitchell, Skenton and Uye 1965) and Kruger Park, South Africa (Pienaar 1969). However, although the act of lion preda- tion is spectacular, it is probably not an important source of mortality so far as population regulation is concerned. Schaller (1972) estimated that in the Serengeti, lions kill 2,468–2,961 buffalo each year. This represents only about 23–28% of total adult mortality. Because of the high numbers of scav- engers, calf mortality could not be estimated in the Serengeti. However, at Lake Manyara, where there are fewer scavengers, Sinclair (1977) estimated it at 11% of calf kills. Bearing in mind that many predated buffalo are either old, diseased or undernourished, it seems unlikely that predation is a major cause of mortality in the Serengeti. What about parasites? About 20 trematode worms (flukes), 6 cestodes (tape- worms) and 28 species of nematode worm have been recorded from buf- falo, along with 3 species of pentastomids (Sinclair 1977), although not all have been recorded from one area. Prevalence (percentage of the population infected) varies between both parasites and areas. For example, in Uganda one population of buffalo had 73% of individuals infected by Ashworthius lerouxi (a nematode), while another population had zero prevalence (Wood- ford and Sachs 1973). In addition to these endoparasites there are several common ectoparasites. At least 10 species of ticks have been found on buf- falo in the Serengeti, along with several species of biting flies, such as tsetse (Glossina) which is the vector for Trypanosoma. At least 15 diseases have been reported for Serengeti buffalo including rinderpest, theileriases (a blood parasite that causes East Coast fever in cattle), babesiasis, anaplas- mosis and Allerton-type herpes virus (Sinclair 1974). Although these lists of parasites and diseases are interesting, their role in significant mortality is equivocal. The only exception to this statement is the exotic rinderpest. Most healthy individuals can acquire immunity to endemic diseases, allow- ing them to ‘fight’ these parasites effectively. Young (in Sinclair 1974) found that in two captive animals, one in poor condition due to a lion injury, had a high infection of Trypanosoma while the other, in good condition, had a low infection. The major cause of poor condition in buffalo, and most other animals, is nutritional state and Scrimshaw, Taylor and Gordon (1968) have reviewed the interaction between disease and nutrition. Their conclusions are clear. In a majority of cases, the nutritional state of the individual is cru- cial in determining death or survival from a disease. Sinclair (1974, 1977) therefore concluded that both predators and parasites, although causing the death of some individuals, were secondary causes of mortality. The main fac- tor was under nutrition, as we suspected might be the case from Figure 4.11b. Sinclair (1974, 1977) measured the percentages of crude protein in the diet of Serengeti buffalo at different times of the year. The chemical analysis used SINGLE-SPECIES POPULATIONS 161 the Kjeldahl method for nitrogen, multiplied by 6.25 to obtain the crude pro- tein percentage. As the dry season progressed (July to October) the percent- age of crude protein in the diet declined, for two reasons. The quality of the most nutritious component, grass leaf, changed from 12.7% crude protein in March to 4.0% in October, while the proportion of leaf (as opposed to the less nutritious stem) in the diet declined from 56% to 13%. Since the activity of the rumen’s microflora is inhibited at crude protein levels below about 5% (Chalmers 1961), this figure is approximately the minimum level of crude protein needed to maintain body weight. Below this level animals lose weight because they are using their own body reserves to compensate for the nutri- tional shortage. During the dry season, the diet of the Serengeti buffalo was below this minimum level. At the end of the dry season the buffalo’s diet consisted of only 2.18% crude protein. These trends were present in all age groups and both sexes, suggesting a real shortage of good-quality grass leaf, affecting all sections of the population. Sinclair also examined the condition of some buffalo. When buffalo are undernourished, they use their body fat first and their bone-marrow fat second. Consequently, a decline in bone- marrow fat reflects a relatively severe state of undernourishment. Using this measure of condition, buffalo between 2 and 10 years old were in good con- dition during the dry season while buffalo less than 2 years old or more than 10 years old showed depletion of bone-marrow fat during the dry season. For the well-studied case of Serengeti buffalo, therefore, we can conclude that the density independent juvenile mortality is mainly due to disease (rinderpest before 1964) and periodic undernourishment. The regulating density dependent mortality of adults is probably due to undernourish- ment in the dry season. This conclusion is supported by the k factor analysis (Figure 4.14a), the k factor model (Figure 4.11a), crude protein levels of grass, bone-marrow fat content, the relationship between rainfall and buf- falo numbers (Figure 4.11b) and the pattern of age-specific mortality (Figure 4.12b). But what about other ungulates? Wildebeest After the removal of rinderpest, the population of Serengeti migratory wilde- beest, like that of buffalo, increased dramatically (Figure 4.15). There appear to have been three main phases of growth over the last 40 years. Between 1960 and 1977 the population increased from about 0.25 million to about 1.3 mil- lion. In the second phase, covering the 16 years from 1977 to 1992, the popu- lation entered an approximately stationary period with fluctuations between 1.1 million and 1.4 million. Between 1993 and 1994 the population declined to 0.9 million following severe dry-season mortality in the drought of 1993, and it has remained at that level. Figure 4.14b shows the results of the k factor analysis carried out by Mduma, Sinclair and Hilborn (1999). The life stages for which they had data were fertility loss, k1; newborn (neonatal) calf mortality 162 THE BIOLOGY OF AFRICAN SAVANNAHS 1,800 1,600 1,400 1,200 1,000 800 600 400 Population size´ x (1,000) 200 0 1950 1960 1970 1980 1990 2000 Figure 4.15 Wildebeest population numbers in the Serengeti ecosystem from 1958 to 1998 (from Mduma, Sinclair, and Hilborn 1999). occurring in the wet season (March–June, ages 0–4 months), k2; dry-season calf mortality (July–December, 5–10 months), k3; yearling mortality (ages 11–24 months), k4; and adult mortality (>24 months) k5. Calves, yearlings and adults all tended to show higher mortality during the dry season. The key factors were dry-season calf mortality (k3) and adult mortality (k5). Significant density dependent mortality was found in adult mortality (k5) and in fertility loss (k1). Adult mortality showed a curvilinear density dependence. Between 1962 and 1972 (when the population was increasing), mortality was low and actually inversely density dependent (b = −0.073, P = 0.0004). From 1975 to 1994, when the population was stable or declining, the linear slope was positive. In fact during this latter period there is clear evidence of delayed density dependence, since an even better relationship exists between k5 and density the previous year (b = 0.336, P = 0.086). Just as with buffalo, wildebeest appeared to die from a number of causes, judged by examination of carcasses. Predator deaths (26– 30%) were attributed to lion (33.8%), spotted hyaena (30.4%), cheetah (3.3%) and unidentified (33.3%). Lions tended to kill middle-aged animals and hyaenas young and old ones. However, Mduma and his colleagues believe that ‘predation played only a minor role in limiting the wildebeest population’. Only about 3% of adult females died from predation, so its effects on the number of births was small. The life stage that showed the most statistically significant density dependent response was adult mortality, most of which occurred in the dry season. Adult mortality (k5) was significantly negatively related to per-capita food supply during the dry season (b = −0.053, P = 0.0005). What is more, per-capita food supply during the dry season was also significantly negatively related to dry-season calf mortality (k3) (b = −0.322, P = 0.002) and not significantly SINGLE-SPECIES POPULATIONS 163 related, but still showing a negative regression slope, to yearling mortality (k4) (b = −0.410, P = 0.063), neonatal mortality (k2) (b = −0.017, P = 0.617), and fertility loss (k1) (b = −0.100, P = 0.34). The conclusion seems clear. For wil- debeest, per-capita food supply is crucial. Individuals die because they don’t get enough food in the dry season and, more importantly, they die in a den- sity dependent way. Predators kill wildebeest, but many of these individuals would have died from undernourishment anyway. If there were no predators it would simply take longer for the undernourished individuals to die. The conclusion, therefore, from these two well-studied cases (buffalo and wildebeest) is that rainfall, acting through dry-season food shortage, is the most important factor affecting population regulation. An addendum to these buffalo and wildebeest studies is that both popula- tions were badly affected by the great drought of 1993, which reduced the food supply and caused a 30% drop in both populations within the ecosys- tem (Sinclair et al. 2008). Both species populations subsequently rebounded to their previous levels, apart from the buffalo in the Masai Mara, which remained low. The Mara is that area of the ecosystem where buffalo and wil- debeest overlap during the dry season, and Sinclair et al. (2008) believe that competition by wildebeest has kept buffalo populations low in this area. Population models In the previous section we looked at what regulates populations and affects their numbers. If the required detail is available, k factor analysis allows insight into these population processes. The analysis allows us to describe, in words, what we think might be happening—a verbal description, or ver- bal model. Another approach is to produce a mathematical description, or mathematical model, to help understand the processes affecting a popula- tion. These mathematical models are often more precise, and can be used to predict future numbers under various ecological scenarios. Also, despite what many field biologists may say, they are no more removed from reality than the verbal descriptions they replace. In fact we have already looked at one type of model when we examined buffalo populations—a k factor model. However, models can be built in different ways, and in this section we look at two quite different population models, one for the migrating wildebeest population of the Serengeti, and one for the plains zebra population of Lai- kipia, Kenya. Wildebeest model Mduma, Hilborn and Sinclair (1998) used data on the Serengeti wildebeest population collected between 1961 and 1994. The data were (1) dry-season 164 THE BIOLOGY OF AFRICAN SAVANNAHS rainfall measured from numerous stations throughout the Serengeti over the entire study (Figure 2.10); (2) 15 censuses of total wildebeest population size (Figure 4.15); (3) 24 estimates of yearling/adult ratio (Figure 4.17b); (4) 22 estimates of calf mortality (Figure 4.16a); (5) eight estimates of dry-season adult mortality (Figure 4.16b). Pregnancy rate was estimated from autopsy and analysis of hormone levels in faecal samples (Mduma 1996), and esti- mated to be 0.85, with very little variation between years. In any year, the relationship between rainfall and dry-season grass production (Sinclair 1979) is described by the equation GR=1.25 equation (1) where G is the amount of grass produced measured in kilograms per hectare per month (kg ha−1 month−1), and R is the dry-season (July–October) rainfall measured in millimetres and averaged over the northern Serengeti region. Remember, from the last section, that green food supply in the dry season is considered to be the limiting resource for wildebeest. The amount of food per animal (F) is the total grass produced per month per hectare (from equation 1), times the number of hectares utilized in the dry season by the wildebeest (500,000), divided by the total number of wildebeest (T): 500000G F = equation (2) T (a) (b) 1.00 1.00 0.80 0.60 0.90 0.40 Calf survival Adult survival 0.20 0.00 0.80 050 100 150 200 250 050 100 150 200250 Food per animal (kg/month) Food per animal (kg/month) Figure 4.16 Survival of (a) calves and (b) adults plotted against estimated food per animal per month. The solid squares are the observed values and the lines are the rectangular hyperbolas (equation 3) fitted by the model. The shapes of the fitted curves are different because the values of the equation constants, a and b, are different for calves and adults (from Mduma, Hilborn, and Sinclair 1998). SINGLE-SPECIES POPULATIONS 165 Mduma and his colleagues divided the wildebeest population into three groups. Calves (up to 1 year), who have their own survival rate, yearlings (up to 2 years) who do not reproduce, and adults (over 2 years) who have a pregnancy rate of 0.85 and, along with the yearlings, an adult survival rate. However, survival rates are not constant. They are influenced by the per-capita amount of food (F). The relationship betweenF and survival (s) was assumed to be similar for both calves and adults and described by a rectangular hyperbola, sometimes referred to as Holling’s ‘disc equation’ (Holling 1959), and frequently used to described the relationship between food and herbivore response (Beddington et al. 1975, Crawley 1983). It is the simplest equation that gives zero survival when there is no food, and is at a maximum when food is unlimited (Figure 4.16). The general form of this equation is: aF s = equation (3) bF+ The values of the constants a and b varied between calves and adults, so that the precise shape of the rectangular hyperbola differed between the two (Figure 4.16). The initial conditions for the simulated population were those of the 1961 census. That is, 263,000 wildebeest, 10% yearlings, 10% (a) (b) 0.3 2,000 0.2 earling (%) 0.1 1,500 Y 0 1965 1975 1985 1995 (c) 1,000 6 Wildebeest numbers 4 500 2 0 Adult monthly mortality 0 1965 1975 1985 1995 1965 1975 1985 1995 Figure 4.17 Three sources of wildebeest data and the best-fit population model. The solid squares are observations and the lines are the fitted model. (a) Total wildebeest population size; (b) percentage of the population that is yearling; (c) adult mortality rate per month of the dry season (from Mduma, Hilborn and Sinclair 1998). 166 THE BIOLOGY OF AFRICAN SAVANNAHS 2 year-olds, and 80% 3 years or older. Each year from the number of adult (≥3 years) individuals divided by 2 (assuming an equal sex ratio), × 0.85, we obtain the number of calves born. Using the rainfall for that year to calcu- late total grass and the number of wildebeest to get the food per individual, the number of calves surviving can be calculated. Using the adult survivor- ship curve, yearlings become 3 year-olds, and 3 year-olds become 4 year- olds, and so on. The only unknown quantities are the four constants in the two survivorship curves. Mduma and his colleagues used repeated computer simulations to fit different values of these four constants until the best fit to three sources of data was obtained. These were (1) censused numbers of wildebeest, (2) the percentage of yearlings, and (3) the estimates of adult dry- season mortality (Figure 4.17). The survivorship curves for both calves and adults that give this best fit are shown in Figure 4.16. This best-fit model, using the relatively simple equa- tions 1–3, is a very good description of the real events unfolding in the Serengeti wildebeest population after 1961. It suggests that per-capita food in the dry season, acting through mortality, is the dominant factor determining wildebeest numbers. Of course animals also die in the wet season, and ani- mals will be killed by predators, but these other mortality factors, interesting though they may be, only add ‘noise’ to the population trajectory. Zebra model The Laikipia district of central Kenya (9,666 km2) consists mostly of semi- arid savannah, divided into a mosaic of privately, publicly, or communally owned properties. Before the 1960s wildlife had been almost eradicated from Laikipia, largely by shooting for food. In 1977, Kenya introduced a national ban on the use of wildlife for food, and subsequently most of the indigenous wildlife, including plains zebra (Equus burchelli), has returned. Georgiadis, Hack and Turpin (2003) and Georgiadis et al. (2011) developed a simulation model of the dynamics of the plains zebra population in Laikipia. Monitor- ing of this population began in 1985, and has continued at intervals since that date (Figure 4.18, black dots). Zebras in Laikipia do not migrate seasonally, although they do shift habitats in response to patchy rainfall. This annual rainfall, recorded, at five stations throughout the zebra range, is also shown in Figure 4.18 (open diamonds). For the purposes of the model, the zebra population was divided into three age classes identifiable in the field: foals (<1 year old), subadults (1–3.4 years), and adults (≥3.4 years). The division between subadult and adult is taken as 3.4 years because this is the youngest age of first conception in plains zebra. Sex ratio was assumed to be unity in foals and the sexes were treated sepa- rately in the subadult and adult categories. Successive population numbers were generated by using five difference equations. For example, foal numbers in year x were calculated as equal to the number of surviving adult females SINGLE-SPECIES POPULATIONS 167 40,000 1,600 30,000 1,200 20,000 800 Rainfall (mm) Number of Zebras 10,000 400 0 0 1984 1992 2000 2008 Years Figure 4.18 Laikipia plains zebra population. Observed estimated population (black dots, vertical bars are standard errors). Best-fit simulation model: rainfall-mediated density-dependent model (RMDD) (black line), rainfall-dependent model (RD) (grey line). Fluctuations in annual rainfall (open circles) (from Georgiadis et al. 2003). in year x − 1 that did not give birth in year x − 2 (females rarely give birth 2 years in succession), multiplied by the birth rate. Calculation of subadults (♂ and ♀) and adults (♂ and ♀) used similar equations, only using a survival rate rather than a birth rate, to link the two generations. Since Georgiadis and his colleagues lacked independent information about zebra carrying capacity (K) in Laikipia (equivalent to ‘grass’ in the wildebeest model) it was calculated using a power equation (K = hRi) that allowed carrying capacity to increase, with rainfall, at a greater than linear rate (see Chapter 2). Both h and i were obtained each generation by an iterative best-fit process. The five vital rates (birth, survival of subadult males and females, and survival of adult males and females) were linked to carrying capacity, and density dependence, in a novel way. Each vital rate was made equal to a density inde- pendent component plus a density dependent component. Survival of adult males, for example, was sm + δm, with s being determined as the minimal field observation, and δ linked to K. This of course fits neatly with what we know about real populations (see the k factor analyses discussed earlier); that they are influenced by both density-independent and density dependent factors. The density dependent component of the vital rate (δ in the survival exam- ple) was linked to carrying capacity in two different ways. One, called rainfall dependence (RD), was δ = (c/K), while the other, called rainfall-mediated density dependence (RMDD), was δ = (cN/K). In both equations, c is a con- stant defining the sensitivity to rainfall. Both models, using total annual rainfall, were able to simulate the observed zebra dynamics (Figure 4.18), and generate total vital rates consistent with field observations. However, the RMDD model gave the best fit to the observed data. 168 THE BIOLOGY OF AFRICAN SAVANNAHS Other species in other areas The conclusions from the two sections above seem clear. Savannah ungulates are regulated through their food supply (bottom-up regulation), determined by rainfall. However, these detailed studies involved only three species— wildebeest, buffalo and plains zebra—all in East Africa. What about other species, in other areas? One of the important pieces of evidence for the role of food in regulating ungulate populations is the presence of a correlation between rainfall and ungulate biomass. Figure 4.11b for buffalo was a good example. What is more, this type of association usually involves data points from many differ- ent savannah areas, suggesting that the relationship is not restricted to these three species, the Serengeti ecosystem, or even the East African region. Coe et al. (1976) first showed that the total biomass of large savannah herbivores was positively correlated to mean annual rainfall. They used data from 20 wildlife areas of southern and eastern Africa, in which rainfall varied from less than 200 mm to more than 1100 mm per year. Since primary production is also positively related to mean annual rainfall in African savannah systems (see Chapter 2), Coe et al. (1976) considered that the herbivore biomass/ rainfall relationship reflected the effects of water availability on the herbi- vores’ food supply. This, of course, is the mechanism implicated in the wilde- beest, buffalo, and zebra case studies above. However, the analysis reported by Coe et al. (1976) referred to large herbivore communities as a whole rather than to particular species. What is more, it was confined to what are sometimes called arid/eutrophic savannahs (Huntley 1982), characterised by high soil nutrients, relatively low rainfall and a low biomass of high-quality vegetation supporting a high total biomass of large herbivores (Chapter 2). Bell (1982) suggested that the herbivore biomass– rainfall relationship could be modified by soil nutrient status and that the positive correlation observed by Coe et al. (1976) only applied to savannahs with high soil nutrient status (e.g. soils of volcanic origin). Bell (1982) sug- gested that for savannahs with low nutrient soils the relationship might even reverse at high rainfall. Such savannahs would include the miombo wood- lands of south-central Africa (Chapter 1) whose Precambrian rocks fre- quently weather into highly leached soils with low supplies of nutrients. The biomass–rainfall relationship was therefore re-examined by East (1984) for individual species, with soil nutrient status being taken into consideration. These relationships are shown in Figure 4.19. Three classes of soil nutrient status were distinguished following the broad classification used by Bell (1982): high (volcanics), medium (Rift Valley sediments), and low (basement and basin sediments, granitic shields and Kalahari sands). Areas of low soil nutrient status were further subdivid- ed into those with mean annual rainfalls of less than or equal to 700 mm, 4 10 elephant zebra hippopotamus 103 102 10 1 104 buffalo wildebeest giraffe 103 102 10 1 black rhinoceros waterbuck hartebeest 103 102 10 1 3 10 kob eland roan 102 ) –2 10 1 3 10 impala springbok warthog 102 Biomass (kg km 10 1 topi and tessebe bohor reedbuck southern reedbuck 102 10 1 greater kudu bushbuck sable 102 10 1 oryx Grant's gazelle 102 Thomson's gazelle 10 1 100 250 500 1,000 100 250 500 1,000 102 oribi 10 1 100 250 500 1000 Figure 4.19 Biomass/rainfall relationships for 25 herbivore species. Symbols indicate different nutrient status: high (▲), medium (▼), low with annual rainfall ≤700 mm (●), and low with annual rainfall >700 mm in southern and eastern Africa (Ο) and West Africa (Δ). Closed symbols are arid/eutrophic savannahs and open circles are moist/dystrophic savannahs. Statistically significant regressions (P < 0.05) are indicated for arid/eutrophic savannahs as solid lines and for all savannahs in areas of low nutrient status as dotted lines (from East 1984). 170 THE BIOLOGY OF AFRICAN SAVANNAHS and those greater than 700 mm. Least squares regression analysis of log- transformed data revealed significant relationships (P ≤ 0.05) in arid/eutroph- ic savannahs for 19 out of 23 species for which sufficient data was available. Of the four remaining species, roan antelope had few data, topi/tsessebe had P = 0.075 and oryx and Grant’s gazelle showed little association with rainfall. Notice that on soils of low nutrient status, some species (elephant, zebra, buffalo, giraffe, black rhinoceros and eland) do show a decline in biomass at rainfall greater than about 700 mm (open symbols in Figure 4.19), although the regression for arid/eutrophic savannahs is still significant (solid line in Figure 4.19). Significant relationships between biomass and rainfall are also seen in several species on poor nutrient soils (hartebeest, roan, warthog, bushbuck and oribi) (Figure 4.19, dotted lines). Although this survey of East (1984) is not a detailed examination of popula- tion processes, as were the studies on wildebeest, buffalo and zebra, nonethe- less the evidence for a bottom-up regulation of savannah herbivores seems quite clear. What is more, there are other field studies, not as detailed as those of wildebeest and buffalo in the Serengeti, that reinforce this view obtained from the relationships in Figure 4.19. Macandza, Owen-Smith and Le Roux 2014) conclude that although faecal nutrient levels for sable appeared only marginally limiting under the conditions that prevailed during the study, malnutrition could have contributed to the sable’s population decline dur- ing a persistently low-rainfall period. Studies on greater kudu populations in Kruger National Park over the period 1974 to 1984 (Owen-Smith 1990) also reveal the controlling influence of annual rainfall on population dynamics, mainly through calf survival but also through yearling and adult survival. This is particularly interesting because the greater kudu is predominantly a browser, rather than a grazer like wildebeest and buffalo. Mills et al. (1995) confirmed the strong positive influence of annual rainfall on changes in the abundance of kudu, as well as buffalo and waterbuck, in the central region of Kruger National Park. Rainfall appears to determine primary production, and this in turn regulates population size. This is a useful general statement about the population regulation of savannah herbivores. It appears to apply to many African species, in many African savannahs. Of course there are interesting and instructive exceptions. In the Serengeti ecosystem, popula- tions of wildebeest and buffalo before 1960 were undoubtedly affected by rinderpest, an exotic disease. In many other parts of Africa the interaction between the rinderpest virus and wild and domestic populations of rumi- nants was the principal factor that kept these populations at reduced levels for many years (Sinclair 1979). In an outbreak in Lake Manyara in 1959, 250 buffalo (78% of the population) died (Prins 1996), and in Kruger National Park the buffalo population was only 20 individuals in 1902, while it had been common before the arrival of rinderpest (Stevenson Hamilton 1911). Of course rinderpest may even be exceptional as a disease. It was exotic, not endemic, and may have had more devastating results as a consequence. SINGLE-SPECIES POPULATIONS 171 Many disease outbreaks seem to occur at the end of the dry season. This has been noted for rinderpest in the Serengeti (Sinclair 1979), rinderpest in Manyara (Prins 1996) and anthrax in Kruger National Park (Pienaar 1961, 1967). As previously mentioned this may imply that disease is often remov- ing individuals that are already dying through starvation. In Lake Manyara National Park, both buffalo numbers (Δ in Figure 4.11b) and lion numbers are exceptionally high. Lion predation on buffalo is also quite significant, amounting to some 89% (Prins 1996, Table 5.4). Many other authors have reported high predation rates in savannah ungulates (Table 4.4), but it is difficult to know how to interpret these. These data come from carcasses that are found and identified as predator kills. Predator kills are identified by seeing predators at a kill, seeing claw marks on kills without predators, or even by hearing ‘killing’ noises the previous night. However, we know that even lions can obtain up to 15% of their diet by scavenging (Schaller 1972) so that many of these so-called predator kills might be deaths from other causes. Predators also pick out individuals in poor condition, and this implies that predators are often removing individu- als that are already dying through starvation and/or disease. In fact many Table 4.4 Reported predation rates on large herbivores of African savannah ecosystems. Serengeti and Ngorongoro are in central East Africa; Kalahari and Timbavati are in southern Africa Prey Main predator % Prey mortality Study area Reference Buffalo Lion 23–28 Serengeti Sinclair and Norton-Griffiths 1982 Wildebeest Lion 38–93 Kalahari Mills 1984 Wildebeest Lion 96 Timbavati Hirst 1969 Wildebeest Lion 13–21 Serengeti Sinclair and Norton-Griffiths 1982 Wildebeest Lion and hyaena 15 Serengeti Elliott et al. 1977 Wildebeest Lion 13 Kruger Mills and Shenk 1992 Plains zebra Lion 90 Timbavati Hirst 1969 Plains zebra Lion and hyaena 59–74 Serengeti Sinclair and Norton-Griffiths 1982 Plains zebra Lion 11 Ngorongoro Elliott et al. 1977 Plains zebra Lion 14 Kruger Mills and Shenk 1992a Springbok Leopard 87 Kalahari Mills 1984 Gemsbok Several 84 Kalahari Mills 1984 Th. gazelle Lion 8 Ngorongoro Elliott et al. 1977 Waterbuck Lion 82 Timbavati Hirst 1969 Kudu Lion 63 Timbavati Hirst 1969 Impala Leopard 55 Timbavati Hirst 1969 Giraffe Lion 34 Timbavati Hirst 1969 Warthog Lion 29 Timbavati Hirst 1969 a Mills and Shenk (1992) present some evidence for the ‘impact’ of lion predation. 172 THE BIOLOGY OF AFRICAN SAVANNAHS authors, including some in Table 4.4, have questioned the role of predators in regulating savannah herbivores (Hirst 1969, Pienaar 1969, Schaller 1972, Eloff 1984, Mills 1984). Equivocal support for this view is provided by a removal exercise carried out in Kruger National Park. In the late 1970s it was speculated that lion and spotted hyaena were responsible for the local decline of wildebeest and zebra populations (Smuts 1978). As a result the largest systematic culling operation in Kruger’s history took place. Over a 5 year period, 445 lions and 375 spotted hyaena were killed, but the reduction had no detectable influence on the population trends of the wildebeest and zebra (Whyte 1985). Unfortunately the cull had no detectable influence on the lion and hyaena population either, so the proposed regulatory effect of these two predators was still unclear. Mills and Shenk (1992) present some evidence for the possible impact of lion predation on wildebeest and zebra populations in the south-east of Kruger National Park. Using intensive observations over a 4 year period, they estimated wildebeest, zebra and lion population size. They also estimated the number of ‘killing’ lions (females) and their individual annual kill rate (7 wildebeest and 8 zebra). They used these two parameters to estimate the death rate of both prey species and, coupled with an estimate of prey fecundity, used these parameters in a simulation model. The model was then used to predict the predator numbers necessary to ‘stabilise’ the popula- tion, such that it showed little change over a 4 year period (like the wild prey populations). Unfortunately there is no regulation in this model (no density dependence) and the delicate balance observed relies on the model-makers. If fecundity, or kill rate, rose by as little as 10% the population of prey sim- ply went into exponential increase (or into exponential decline if they fell by 10%). The model demonstrates that predation, if very high, can have an impact on prey, but not a regulating effect. There is an important difference between predation being an ‘important source of mortality’ and predation ‘regulating populations’ of savannah herbivores. Quite recently, two excellent analyses of long-term population counts for savannah ungulates have claimed to show that predators influence popula- tion numbers. One of these was in the Kruger ecosystem of South Africa (Owen-Smith and Mason 2005, Owen-Smith et al. 2005) and the other was in the Serengeti–Mara ecosystem of East Africa (Sinclair et al. 2003). In both cases there is certainly something very interesting happening demographi- cally, but whether this is caused by predation is debatable. In the Serengeti– Mara, Sinclair et al. suggest that a ‘natural experiment’ (Diamond 1986) has taken place. They suggest that in an area of the northern Serengeti, poach- ing and indiscriminate poisoning removed the majority of carnivores (lion, spotted hyaena and jackals) for an 8 year period (1980–1987), citing Sin- clair (1995b). However, there are no predator population numbers to con- firm this, and what the Sinclair (1995b) paper actually says is ‘there is some circumstantial evidence that lions, which are the main predators in the area disappeared: roar counts were considerably lower in the study area in SINGLE-SPECIES POPULATIONS 173 1990 compared to other parts of the Serengeti (Packer 1990)’. Notice that only lions are mentioned and the comparison is with ‘other parts’ at this time, not the northern area prior to this date. In the contiguous Masai Mara Reserve predator populations are thought to have remained intact (Broten and Said 1995). During the period of ‘predator removal’, five smaller-bodied species (<150 kg), oribi (Ourebia ourebi), Thomson’s gazelle Gazella( thom- sonii), warthog (Phacochoerus africanus), impala (Aepyceros melampus) and topi (Damaliscus lunatus jumela) increased markedly in density in the north- ern Serengeti (no predators) relative to their population in the Mara (preda- tors present) (Figure 4.20). These northern Serengeti populations declined once predators returned to the area, after 1987. In contrast, the abundance of a large ungulate, the giraffe (Giraffa camelopardalis), did not increase during the predator removal peri- od (Figure 4.20). Clearly there is reason to think that something happened in the northern Serengeti during the period 1981 to 1987 that did not happen in the Mara. Was this due to a relaxation of predation? We think the evidence is equivocal. The evidence for reduced predators (perhaps only lions) is circum- stantial. However, we know that buffalo numbers in this northern area of the Serengeti were seriously reduced due to poaching about this time. Sometime between 1976 and 1982 approximately 85% of buffalo (about 9,000 animals) in northern Serengeti were removed by poachers (Sinclair 1977, Dublin et al. 1990a). This did not happen in the Mara Reserve. So would this have affected all the animals in Figure 4.20, except giraffe, through competition for graz- ing? We really don’t know, but it’s a possible scenario. Alternatively, the pre- dation explanation for Figure 4.20 might be reinforced by Figure 4.21, also from Sinclair et al. (2003), which suggests that indeed only small species are subjected to high predation rates, while larger species are not. This may be because smaller prey have more predators (Chapter 5). Notice, however, that this mortality is for adults only and that we have already seen that mortality from food shortage often targets the young. This issue of young versus adult mortality leads on to the second study in Kruger National Park (Owen-Smith and Mason 2005, Owen-Smith et al. 2005). Between 1965 and 1976, ungulate population numbers were estimat- ed using irregular aerial surveys, supported by ground counts. From 1977 to 1997 these surveys were carried out each year and from 1983 to 1996 popula- tion structure was also recorded. Where possible individuals were classified as adult (> 2 years, male or female), yearling (1–2 years, male or female), and juvenile (<1 year). There are also extensive rainfall records for the period 1960 to 1997. About 80% of the annual rainfall falls during the wet season (October–March) and only 20% during the 6 months of the dry season. Figure 4.22 shows both rainfall and estimated numbers of some major ungu- lates over the period 1965 to 1996. Notice that for both rainfall and popula- tion numbers, figures are shown as smoothed ‘running means’. For rainfall this is a 5 year running average, while for numbers it is a 3 year weighted 174 THE BIOLOGY OF AFRICAN SAVANNAHS 10 oribi (18 kg) 25 Thomson's gazelle (20 kg) 8 20 6 15 2 10 4 5 0 0 40 2.0 impala (50 kg) warthog (60 kg) ) 2 30 1.5 20 1.0 10 0.5 Density (number per km 0 0.0 1.2 7 topi (120 kg) giraffe (800 kg) 6 1.0 5 0.8 4 0.6 3 0.4 2 1 0.2 0 0.0 Removal Control Removal Control Figure 4.20 Densities of six ungulates in a predator removal area in northern Serengeti compared with an adjacent control in the Mara Reserve. The period before predator removal (open bar) is 1967–1980, the period of low predator numbers due to poaching (shaded bar) is 1981–1987, and the period after predators returned (solid bar) included years from 1989 onwards. There is no control for oribi because they do not occur in the Mara. Error bars are one standard error (from Sinclair et al. 2003). running average, N′t = 0.25Nt−1 + 0.5Nt + 0.25Nt+1. At the start of this period average rainfall was low. In the 1970s to early 1980s it was high, but from about 1981 it started to drop and remained low during the 1980s and 1990s. A feature of this latter period of low rainfall was very low dry-season rain- fall. Following the 1981 drop in annual rainfall there appears to have been a SINGLE-SPECIES POPULATIONS 175 100 O I T 90 80 W 70 Z 60 50 40 Predation (%) 30 20 B 10 G R H E 0 1234 Log [herbivore weight (kg)] Figure 4.21 The proportion of adult mortality accounted for by predation in 10 non-migratory ungulate populations in the Serengeti ecosystem. Error bars are 95% confidence limits. O, oribi; I, impala; T, topi; W, non-migratory wildebeest; Z, common zebra; B, African buffalo; G, Masai giraffe; R, black rhinoceros; H, hippopotamus; E, elephant (from Sinclair et al. 2003). change in population numbers across what Owen-Smith and Mason (2005) called a common ‘break point’, pivotal around 1987. While some species increased before this date and then persisted at high abundance afterwards (common zebra, wildebeest, impala and giraffe), others (buffalo, greater kudu, common waterbuck, warthog, sable antelope, tsessebe, eland and roan antelope) declined after this point. In other words, some species (stabilising species) did not appear to respond to the low rainfall after 1981, while others (declining species) did. Notice that there appears to have been a 5 year lag in this response. Owen-Smith and Mason (2005) estimated annual survival rates for juve- niles, yearlings and adults of all four stabilising species and five of the declin- ing species (buffalo, roan and eland were not included). For three species (giraffe, sable and tsessebe) they could not estimate yearling survival. They estimated survival separately for the period before 1987, and for the period after, and tested for any indication of a significant difference between the two periods. All survival rates after 1987 (low rainfall) are lower than the survival rates before 1987 (high rainfall) except for yearling wildebeest (significant, P = 0.019) and yearling waterbuck (not significant, P = 0.978), which are higher. Therefore in 22 out of 24 comparisons, the period after 1987 (lower rainfall) is associated with reduced survival. However, not all these reduc- tions are statistically significant and Table 4.5 summarizes the patterns that emerged. Owen-Smith and Mason (2005) draw attention to the fact that 176 THE BIOLOGY OF AFRICAN SAVANNAHS 1. 4 1. 3 1. 2 1. 1 1. 0 0. 9 0. 8 0. 7 0. 6 1. 4 1. 3 1. 2 1. 1 1. 0 0. 9 0. 8 0. 7 0. 6 1995 1995 kudu giraffe impala/20 1990 1990 1985 1985 1980 1980 Year Year 1975 1975 1970 1970 roan tsessebe 1965 eland 1965 0 0 800 600 400 200 8,000 6,000 4,000 2,000 1,200 1,000 12,000 10,000 1. 4 1. 3 1. 2 1. 1 1. 0 0. 9 0. 8 0. 7 0. 6 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 t 1995 1995 buffalo zebr a wildebees 1990 1990 1985 1985 1980 Year 1980 Year 1975 1975 1970 1970 warthog each graph). All numbers are smoothed running means (see text). Impala numbers are divided by 20 (modified from Owen-Smith and Ogutu divided by 20 (modified from smoothed running means (see text). Impala numbers are each graph). All numbers are 2003). Changes in population numbers of the major ungulates of Kruger National Park since 1965, in relation to mean rainfall (fine dotted line on Changes in population numbers of the major ungulates Kruger National Park since 1965, relation 1965 waterbuck 1965 sable 0 0 500 5,000 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 35,000 30,000 25,000 20,000 15,000 10,000 Figure 4.22 Figure SINGLE-SPECIES POPULATIONS 177 Table 4.5 Number (out of total) of statistically significant differences between pre-1987 survival rates and lower post-1987 survival rates. Notice that not all yearling survivals are available for estimation Juveniles Yearlings Adults Stabilising species 0/4 0/3 2/4 Declining species 1/5 1/2 4/5 From Owen-Smith and Mason (2003). most statistically significant reductions in survival are associated with declin- ing species and adult life stages. The first observation can be interpreted as the cause of the decline in these species. With reduced rainfall all species showed lower survival but some, the declining species, showed much more significant reduction. So far, this explanation fits with the one that we have been suggesting in this chapter: that is, reduced rainfall means reduced food, which in turn means reduced survival. However, Owen-Smith and Mason interpret the second observation (lower survival associated with adults) as indicating that it is predation, rather than malnutrition, that is the cause of the decline seen in so many of these Kruger populations after 1987. They suggest lions, whose kills are known to concentrate on the adult segment of medium-sized ungulates (Pienaar 1969, Mills and Biggs 1993). After 1987 when rainfall, particularly dry-season rainfall, was very low, the resultant lack of sufficiently good forage during the dry season could have reduced survival rates in adults either directly through malnutrition, or indirectly by making these adults more vulnerable to predation. With this in mind Owen-Smith et al. (2005) continued this analysis, and examined the relationship between survival and population density, rain- fall and predation using linear multiple regression, fitted by standard least squares. They used the same nine species analysed in Owen-Smith and Mason (2005), plus roan antelope. Because density, through resource avail- ability, will operate on adults and juveniles over an extended period, den- sity dependence was assessed using a 3 year weighted average of the census totals, with the weighting centred on the year preceding that in which the survival estimate was made. For rainfall, both the total annual rainfall and that for the wet and dry season alone, preceding the count, were used. To assess the effect of what they called ‘prior rainfall’, which might affect the vegetation cover and its composition, a 4 year running average prior to the survival estimate was used. No direct information on predator abundance was available, because these species are not readily visible from the air during a census. They therefore devised a ‘proxy’ index to represent this measure. All 10 ungulate species, plus buffalo, are taken by lions (Pienaar 1969) and all these were regarded as potential prey. The annual total counts for each 178 THE BIOLOGY OF AFRICAN SAVANNAHS prey species were transformed into a ‘carcass biomass’ through multiplying by their estimated mortality (to get the number of carcasses), and then by the species biomass (to get the estimated amount of food available for lions). Since uneaten carcasses are rarely encountered in Kruger, this annual carcass biomass was regarded as the total food available for lions and therefore a proxy for lion numbers. The initial regression analysis included numbers of animals and annual rainfall as the two potential predictors. Once the rela- tionship between survival rate and these predictors (if any) were established, either predation (carcass biomass) or prior rainfall was added to the regres- sion model. Table 4.6 summarizes the best-fit predictors. Even if you believe that ‘carcass biomass’ is equal to predation, the evi- dence for this being an important predictor of survival rate (and therefore Table 4.6 Best-fit predictors of stage-specific survival rates, from alternative linear regression models Species Stage One predictor Two predictors Stabilizing species Zebra Adults Abundance Yearlings Annual rain Juveniles Prior rain Wildebeest Adults Prior rain Juveniles Prior rain Impala Adults + yearlings Abundance Juveniles Dry-season rain Giraffe Adults Abundance Yearlings + juveniles Dry-season rain + prior rain Declining species Kudu Adults Annual rain Yearlings Annual rain Annual rain + prior rain Juvenile Dry + wet-season rain Waterbuck Adults + yearlings Dry-season rain + abundance Juvenile Dry-season rain Warthog Adult + yearling Predation Juvenile Annual rain Sable Adults + yearlings Predation + dry-season rain Juveniles Annual rain Tsessebe Adults + yearlings Annual rain + predation Juveniles Predation + dry-season rain Roan Adults + yearlings Predation Juveniles Annual rain From Owen-Smith and Oguto (2003). SINGLE-SPECIES POPULATIONS 179 population change) is not great, although it is interesting that the five cases where predation appears in Table 4.6 are all in the declining species, and four of them affect adults. However, some measure of rainfall seems to dominate these best-fit predictors, suggesting that vegetation is once again influential in determining ungulate survival. Both the Serengeti and Kruger analyses of population change are important attempts to look for patterns that could be interpreted as predation. At the moment the evidence is intriguing, but circumstantial. We clearly require more detailed studies of the factors influ- encing the population dynamics of savannah ungulates. Finally, we turn to predator populations, and what regulates them. The prob- lem is that we have few good estimates of predator population numbers over time. Even in the Serengeti and Kruger National Parks, where most of the long-term surveys of ungulate populations have been carried out, we have no detailed series of estimates for carnivores (as seen in the last two analyses). In the Serengeti, for example, we have occasional estimates that usually coin- cide with individuals doing short-term research on carnivores, and these are shown in Table 4.7. These estimates allow us to view the changes in predator composition that have occurred over this 26 year period (Figure 4.23), and this is helpful, but they do not allow any detailed understanding of how these predator populations are controlled. Spotted hyaena numbers have clearly increased since the 1960s, and lions may also have increased. This may be a response to the increase in wildebeest numbers, after the removal of rinder- pest. There is also a suspicion that spotted hyaena have interacted unfavour- ably with wild dogs, a topic we look at in Chapter 5. Table 4.7 Periodic estimates of the total population size for the five major carnivores of the Serengeti ecosystem, over a 25-year period from the late 1960s Species Year Numbers Spotted hyaena 1967 2,207 ± 120 1977 3,306 ± 432 1986 5,214 ± 828 1991 9,500 Lion Mid 1970s 2,000–2,400 1991 2,800 Leopard Mid 1970s 800–1,200 1991 840 Cheetah Mid 1970s 220–500 1991 600 Wild dog Mid 1970s 150–300 1991 101 180 THE BIOLOGY OF AFRICAN SAVANNAHS leopard leopard cheetah lion cheetah wild dog wild dog lion spotted spotted hyaena hyaena 1972–77 1991 Figure 4.23 Pie diagrams showing the change in predator composition in the Serengeti National Park over a 25 year period from the mid 1960s. lion cheetah 40 3 2 20 1 ) –2 spotted hyaena 60 leopard 40 3 Predator biomass (kg km 2 20 1 4,000 8,000 12,000 200 400 600 Prey biomass (kg km–2) Figure 4.24 Relationship between the biomass of large carnivore species and the biomass of their preferred size class of prey (lion and spotted hyaena: topi and hartebeest to buffalo) (cheetah and leopard: Thomson’s gazelle to impala and kob). All relationships are statistically significant as judged by linear regression (solid lines) (leopard P = 0.05, rest P = 0.01) (from East 1984). SINGLE-SPECIES POPULATIONS 181 rain grass herbivores carnivores Figure 4.25 Positive relationships in population control. East (1984), in his survey of biomass–rainfall relationships for herbivores, also looked at the relationship between the biomass of five large savannah carnivores (lion, spotted hyaena, cheetah, leopard and wild dog), and the biomass of their preferred size class of prey. There was a significant positive relationship for four of the carnivores, and these are shown in Figure 4.24. Like the Serengeti lions and hyaenas, this implies that more prey means more carnivores, suggesting again a bottom-up regulation of savannah animals. The scenario that emerges for population control in savannah mammals is therefore one of a series of positive relations (see Figure 4.25). That is, more rain produces more grass (and other vegetation), which produces more her- bivores, which produces more carnivores. Endemic disease may periodically cause episodes of mortality, in both herbivores and carnivores, but probably mainly kills individuals in poor nutritional condition at the end of the dry season. Predation may sometimes cause high mortality, but this may also be borne mainly by nutritionally weak or old individuals. Perhaps there is top-down regulation in Kruger and perhaps there is top-down regulation of small prey in Serengeti. We have indicated this by a dotted arrow. However, at the moment the evidence is not compelling. Notice that in Figure 4.24 there is no graph for wild dog (Lycaon pictus). This is because there was no significant relationship between dog biomass and their prey biomass. The populations dynamics of this predator are, however, quite interesting and well studied. They involve the special problems of small populations and their vulnerability and competitive interactions with other carnivores. This leads us to consider populations of two interacting species more directly, and is the subject of the next chapter. However, one final com- ment on single-species populations. This chapter has dealt with natural fac- tors that influence birth and death. There are also factors imposed by humans that act via tourism, hunting, poaching, habitat removal and fire. These are mostly dealt with in Chapter 6, although fire appears in Chapter 5. 5 Species Interactions Interactions between species have been given many different, often confus- ing, names. However, a simple and unambiguous method of classification, is to use the ‘effect’ that individuals of one species have upon the population growth of another species and vice versa. We ask the question: in the pres- ence of species A (+), does species B (1) increase its numbers (+), (2) not change its numbers (0), or (3) decrease its numbers (−) relative to when spe- cies A is absent (−)? The same question is asked of species A in the presence of species B. The answers can be conveniently summarised in a 3 × 3 table (Table 5.1). Because of the symmetry in the table, there are actually only six types of interaction. These are frequently called neutralism (0 0); commens- alism (+ 0); predator–prey, parasite–host, or herbivore–plant interactions (+ −); amensalism (0 −); competition (− −); and mutualism (+ +). Neutral interactions may well be very common in many ecosystems, but do not con- tribute to the dynamics of the system, and would be regarded by most people as uninteresting. Commensalism could be regarded as an extreme example of mutualism in which the beneficial effect of one species on the other is rather weak and undetectable. Amensalism could also be regarded as an extreme example of competition, in which the adverse effect of one species on the other is much greater than the reciprocal effect. This chapter therefore examines three types of interaction between savan- nah species: predator–prey (+ −), competition (− −), and mutualism (+ +). However, before proceeding it is worth drawing attention to a rather confus- ing interaction that is sometimes observed. An ecological situation can arise where two species appear to show reciprocal negative (− −) effects associated with interspecific competition, but this is in fact the result of predation by a third species. This is called apparent competition. Consider the situation shown in Figure 5.1. A single species of predator attacks two species of prey. The predator–prey interactions are of a − + type, and therefore both species are adversely affected by the predator, and the predator is positively affected by both species of prey. This means that the positive effect that prey 1 has on the predator will, in turn, increase the negative effect upon prey 2, and vice The Biology of African Savannahs. Second Edition. Bryan Shorrocks & William Bates. © Bryan Shorrocks & William Bates 2015. Published 2015 by Oxford University Press. Species Interactions 183 Table 5.1 A simple classification of species interactions Effect of A on B + 0 − Effect of B on A + + + + 0 + − 0 0 + 0 0 0 − − − + − 0 − − Predator – +–+ – Prey 1 – Prey 2 Figure 5.1 Apparent competition between two prey species. The solid lines indicate direct interactions between prey and predator, while the dotted lines show indirect (apparent) interactions between the two prey species. versa. The overall consequence of this is that there will appear to be a − − interaction between prey 1 and prey 2, even if they are not competitors for any essential and limiting resource. Predator–prey type interactions (+ −) This type of interaction actually encompasses three interactions that may appear rather different. They certainly tend to be studied by different ecolo- gists. They all have in common the fact that one species (carnivore, parasite or herbivore) consumes the body, or part of the body, of another species (prey, host or plant). In the case of the predator, the prey is killed. With the para- site the host is sometimes killed, but may recover from the infection. With herbivores, the whole plant is usually not killed or consumed (frequently just leaves and fruit are eaten) and the ‘prey’ regrows those parts that have been consumed. Exceptions to this general rule might be elephants who can some- times kill trees, and rodents that consume whole plants in the form of seeds. We will look at each of these + −interactions, and detail some of their special features that have relevance to savannah ecosystems. 184 THE BIOLOGY OF AFRICAN SAVANNAHS Herbivores and plants Interaction between grasses and bovids (antelopes and buffalo), in African savannahs, has probably taken place since at least the late Pliocene, some 5 million years ago (Gradstein et al. 2004). Grazer and grass are inextricably linked, and the adaptive radiation of both groups occurred simultaneously in the ensuing Miocene, Pliocene and Pleistocene geological periods. Bovids evolved grinding teeth and complex digestive systems to deal with grasses (Chapter 3), and grasses evolved defences to combat grazing, such as the depo- sition of silica. Grasses also have a growth form that allows continuous grazing (Chapter 2), and show compensatory growth as a response to grazing pressure (McNaughton 1979b). This compensating growth, stimulated by grazing, is the product of several mechanisms, including enhanced photosynthetic capacity, more efficient use of light (due to reduced leaf shading), reduced leaf ageing, nutrient recycling, and the stimulating effects of herbivore saliva (McNaughton 1979a). This leads to the idea that moderate grazing will actually stimulate above-ground net primary productivity above that of ungrazed areas, although heavy overgrazing would still depress productivity (McNaughton 1979b). This model of grazing is illustrated in Figure 5.2. This predicts that there will be an optimum level of defoliation at which there is a balance between residual leaf area and photosynthesis per unit of leaf area, which maximises net productivity. A quantification of this predicted pattern was obtained in the northern Serengeti and Kenya Mara during August and September 1974 (dry season), and January 1975 (wet season) (McNaughton 1979b). During the periods that measurements were made, wildebeest were the principal grazers, and the results are shown in Figure 5.3. During the dry season (Figure 5.3a) the above-ground green biomass declined in control stands where grazing was Stimulation Ungrazed + production Primary production – Intensity of grazing Overgrazing Figure 5.2 McNaughton’s grazing model, predicting the level of primary production of green grass versus the intensity of grazing. The level of primary production in ungrazed areas is shown as a dashed line, which coincides with zero grazing (modified from McNaughton 1979b). Species Interactions 185 3 4 2 1 2 0 Grazing 40 80 120 –1 Production 0 Grazing –2 Production 5 10 15 –3 –2 Figure 5.3 Relationship between grazing and net above-ground primary production in the Serengeti– Mara ecosystem, East Africa. (a) Dry season: production in g m−2 day−1 subsequent to the grazing period, grazing in g m−2 initial green biomass eaten by wildebeest. (b) Wet season: production in g m−2 day−1, grazing in g m−2 day−1 green biomass consumed (modified from McNaughton 1979b). prevented. This was because ungrazed grasses dried out subsequent to flow- ering, despite the fact that this dry season was unusually wet. Intermedi- ate levels of grazing resulted in more primary production, and overgrazing resulted in negative values. The wet-season grazing system (Figure 5.3b) appears slightly different, although a similar optimisation curve is pro- duced. As in the dry season, moderate grazing promotes productivity. The fact that the curve is asymmetrical, and skewed to the right, suggests that in the wet season savannahs are stimulated more by light grazing, and are more resistant to overgrazing, than they are in the dry season. In addition to stimulating growth, wet-season grazing also suppresses flowering, because the grass diverts nutrients from flower production to leaf production. This model of grazing dynamics leads to the intriguing idea that herbivores manipulate their food supply to enhance the quality of food available on later visits. McNaughton’s contention (1985) is that the herbivores (in this specific example, wildebeest) can produce ‘grazing lawns’ that are kept in a state of high above-ground productivity. This was an idea originally put forward by Vesey-Fitzgerald (1974) when he looked at the interaction between buffalo and various grasses in Arusha National Park, Tanzania. In Cynodon dactylon grassland he noted that optimum utilisation resulted in a short grass lawn being maintained by the buffalo. More recently, by monitoring exclosed vegetation plots over a 5 year period, when wildebeest numbers were high, Belsky (1992) showed that grazing was essential for the grazing system to persist. In the absence of grazing, the grazing lawns were rapidly transformed into rough pasture with dense foliage, in which the short-statured species favoured by the grazers disappeared. In fact McNaughton (1979b) had also noticed this effect on species composition. He re-examined areas in southern Serengeti that had originally been fenced over 10 years before, and Table 5.2 shows how the prevention of grazing had changed the frequency of four spe- cies of grass. Inside the exclosures grass species diversity had declined due 186 THE BIOLOGY OF AFRICAN SAVANNAHS Table 5.2 Frequency (%) of four grass species inside and outside exclosures in the southern Serengeti Inside exclosure Outside exclosure Andropogon greenwayi 0 56 Sporobolus marginatus 0 20 Pennisetum ultramineum 72 5 P. mezianum 26 3 Data from McNaughton (1979b). to the dominance of a few taller-growing species that invest heavily in stems (such as Pennisetum). Inside the exclosures, grass height was much greater, there was an accumulation of dry foliage, and a doubling of the grass biomass invested in stems. Not only does grazing appear to maintain the grass in a ‘young’ leafy state, with species preferred by grazers, but it also helps to reduce the risk of fire. Herbivores are therefore confronted by a shifting mosaic of available for- age. After grazing on one patch for a while they will sequentially move on to new patches, only to return to the original grazing lawn when it has regrown new leaf. This proposed sequence has been called ‘cyclic grazing’ (Drent and van der Wal 1999). McNaughton (1979a) noted that Serengeti wildebeest tended to revisit grassland patches at regular intervals, harvesting the high- quality fresh growth resulting from a previous grazing bout. In Lake Man- yara National Park, northern Tanzania, Prins (1996) collected a large data set on the visitation interval of buffalo to patches of the grass Cynodon dactylon (Figure 5.4). Patches (black areas in Figure 5.4) are, on average, about 20 ha in area, and are more or less uniform stands of this grass. In Manyara, buf- falo rely to a large extent on C. dactylon for their nutritional needs. Over a 6 month period buffalo returned, on average, after 4.5 days (histogram in Figure 5.4). Clipping experiments, followed by subsequent regrowth, sug- gested that the ideal return time would be about 5 days (Prins 1996). How- ever, the median may be a better indicator of the typical return time for a buffalo, and this is only 3 days. Prins suggests that buffalo are forced to return earlier than ideal because elephants would otherwise take the new grass. Herbivory is not the sole determinant of nutrient availability in savannah ecosystems. For example, fire can also affect the nutritive quality of savannah vegetation directly, independent of changes in the plant community, or indi- rectly, via effects on the plant community. Indirect effects can be further sub- divided into those occurring because of changes in plant species composition or plant abundance (i.e. quality versus quantity). Anderson et al. (2007b) studied relationships between fire, herbivory, rainfall, soil fertility, and leaf nitrogen, phosphorus, and sodium at 30 sites inside and outside the Serengeti Species Interactions 187 Marere river F W F W G Mbulu plateau Mkindu river W 1600 m asl F F 1 F W F F W Mchanga river F W Msasa river F W 3 2 4 G W G W 5 R W Swamp G R R N’dilana river 6 W Lake Manyara R G Mud flat G Chem chem river W 60 R Ndala river 50 W G 40 7 30 8 20 Bagayo river Mud W flat 9 Frequency of patches revisited 10 R G 0 0 1 234567891011–21 21–30 W Interval of patch visit (days) Figure 5.4 Visitation intervals of buffalo utilising patches of Cynodon dactylon in Lake Manyara National Park, Tanzania. Cynodon patches are indicated in black. The grey area is escarpment. Vegetation codes: R, riverine bush; F, forest; W, woodland savannah; G, grassland savannah (modified from Drent and van der Wal 1999, after Prins 1996). National Park. Using structural equation modelling, they asked whether fire and herbivory influences were largely direct or indirect and how their signs and strengths differed within the context of natural savannah processes. Her- bivory was associated with enhanced leaf nitrogen and phosphorus through changes in plant biomass and community composition. Fire was associated 188 THE BIOLOGY OF AFRICAN SAVANNAHS with reduced leaf nutrient concentrations through changes in plant commu- nity composition. Additionally, fire had direct positive effects on sodium and non-linear direct effects on phosphorus that partially mitigated the indirect negative effects. Key mechanisms by which fire reduced plant nutritive qual- ity were through reductions of sodium-rich grasses and increased abundance of Themeda triandra, which had below-average leaf nutrients. Not all herbivore–plant interactions in savannahs involve grazers: a smaller, but still significant, number involve browsers. However, woody plants pro- duce defensive compounds of far greater diversity and toxicity than grasses. In savannahs, condensed tannins in particular have a deterrent effect against browsing ruminants. Figure 5.5 shows how this affected browse preference for three herbivore species in Kruger National Park, in southern Africa (Du Toit 2003). In fact Du Toit points out that avoidance of condensed tannin has a stronger effect on feeding preference than selection for leaf nitrogen or phosphorus. What is more, ruminants can accurately regulate their intake of specific toxins (Du Toit et al. 1991) to match their detoxification abili- ties. This could explain why many browsers take many small meals from a wide range of plants, and plant parts. Steenbok in central Kruger, for exam- ple, feed on the bright yellow and toxic fruits of Solanum panduraeforme but rarely take more than one fruit from a single bush, even if it is laden with fruit (Du Toit 2003). Browsers also affect the trees they browse, particu- larly young trees. An example from Laikipia, in central Kenya, is shown in Figure 5.6. Acacia drepanolobium trees were tagged (Chapter 4), and the 1 0 –1 Feeding preference –2 0 123 Leaf condensed tannin content Figure 5.5 Feeding preference of kudu (■), impala (●), and steenbok (▲) in central Kruger (Figure modified and reprinted fromThe Kruger Experience by Johan T. DuToit et al. (ed.), © 2003 Island Press. Reproduced by permission of Island Press, Washington, DC.). Species Interactions 189 (a) 60 elephant rhino 50 Drought 40 % 30 20 10 0 0.5–1 1–2 2–3 3–4 4–6 6+ Height class (m) (b) 30 elephant 25 rhino Both 20 % 15 10 5 0 0.5–1 1–2 2–3 3–4 4–6 Height class (m) Figure 5.6 Herbivore damage on trees. The percentage of Acacia drepanolobium (a) killed and (b) reversed in each height-class over a 3 year period (from Birkett and Stevens-Wood 2005). damage inflicted by browsers recorded over a 3 year period. The princi- pal browsers were elephant, black rhinoceros, and reticulated giraffe, and the damage inflicted by each species could be recognised. Elephants either pushed trees over, or broke the main stems and left the bark hanging in strips; rhinos made a clean cut of the main stem; and giraffe ate the leaves and grow- ing tips. Damage was allocated to one of two categories: tree killed, or tree ‘reversed’ into a lower height category. What is clear is that browsing damag- es this Acacia to quite an extent. Giraffe had little effect, but black rhinoceros and elephants kill or reverse many trees, particularly young saplings. 190 THE BIOLOGY OF AFRICAN SAVANNAHS Of course, savannah ecologists and wildlife managers have frequently been worried by the ‘elephant problem’ (e.g. Croze 1974a, 1974b), coupled with an often observed trend in savannah vegetation from more wooded to more open grassland savannah (Tanzania, Vesey-Fitzgerald 1973; Kenya, Leuthold and Sale 1973; Uganda, Laws 1970; Zambia, Caughley 1976). However, the problem is more complex than it seems at first sight. Elephants actually pre- fer to eat green grass, fire may play a pivotal role, and the woodland to grass- land savannah trend may in fact be cyclical and quite natural. The interaction of these three factors (elephants, trees and fire) has been particularly well described for the Serengeti–Mara ecosystem (Norton-Griffiths 1979, Dublin et al. 1990b, Dublin 1995, Sharam, Sinclair and Turkington 2006). The reconstruction of the vegetation dynamics of the Serengeti–Mara eco- system involved the collection of historical information as well as more recent observations, and the construction of a simple mathematical model. In the early 1900s, Swahili slave traders and European explorers, hunters and naturalists described the Serengeti–Mara as open grassland with occa- sional Acacia trees, rather like much of the Mara Reserve today. The recently introduced rinderpest had reduced ruminant numbers to low levels, and the elephant population had suffered from heavy ivory poaching during the previous decade (Spinage 1973). Human numbers were also low, and conse- quently bush fires were infrequent. Tree recruitment was probably good and during the next 30–50 years dense woodland savannahs became established and heavy infestations of the tsetse fly became a serious problem (Ford and Clifford 1968, Ford 1971). It was at this time that the colonial administration started to introduce measures to protect these ‘pristine’ woodland savan- nahs, not realising that less than 50 years earlier they had been open grass- land savannahs. By the late 1950s and early 1960s, these woodland savannahs had started to decline. Ungulate populations had not yet recovered from the rinderpest epidemic and, as a consequence, a large standing crop of dry grass resulted in widespread bush fires. These fires were started by Masai to improve grazing pasture and clear tsetse-infested bush, by hunters, and by the Park authori- ties. The fires cleared large areas of bush and attracted ungulates to grazing lawns. Elephant numbers also increased, not so much because of population growth but because increasing human activity around the Park area forced elephants into the Serengeti–Mara, increasing the local density. Both fire and elephants had a detrimental effect on tree recruitment. An analysis of aerial photographs (Dublin 1995) from the Masai Mara shows a steady loss of Aca- cia cover (Figure 5.7). Between the 1960s and the 1980s two important changes occurred in the Serengeti–Mara ecosystem. First, with the eradication of rinderpest, the numbers of wildebeest increased dramatically (Chapter 4). More of the green grass was removed and consequently there was less dry grass available Species Interactions 191 30 25 20 15 10 Per cent cover 5 0 1940 1950 1960 1970 1980 1990 Year Figure 5.7 Mean percentage cover in Acacia woodland savannah in the Masai Mara Reserve. Percentage cover values are derived from five sets of aerial photographs taken between 1950 and 1982 (from Dublin 1995). to fuel fires. The incidence of fires dropped dramatically and by the 1980s it was as low as 5% (area burned) in the Mara. Second, in April 1977, the border between Tanzania and Kenya was closed, and remained closed until 1986. This closed the main tourist route between the Mara and the northern Serengeti and tourist numbers to the Serengeti dropped from 70,000 in 1976 to 10,000 in 1977. One consequence of this was a drop in the operating budg- et of the Serengeti, which was linked to income from visitors. As a result, anti-poaching patrols dropped to 60% of the number prior to the border closure. Two animals immediately suffered: black rhinoceros and elephant. Rhinos were effectively removed from the Serengeti part of the ecosystem (Figure 6.19a), and are still absent (although occasional animals visit from the Ngorongoro crater). The Serengeti elephant population declined by 81%, from 2,460 in 1970 to 467 in 1986. Of these, over 1,500 were killed by poach- ers, while the remaining 400–500 sought safe refuge in the Masai Mara, where tourist numbers and anti-poaching patrols were still high. To help understand this intriguing interaction between trees, elephants and fire, Dublin et al. (1990b) constructed a ‘tree population’ model to describe and visualise these events. Animals and fire have three different effects on tree seedlings, which these authors describe as ‘killing’ (complete removal), ‘reversing’ (removal to ground level with the possibility of resprouting), and ‘inhibiting’ (reduced, and kept in the height-class below 1 m). Trees were placed into five height-classes, the last being the adult class. The model com- putes the tree recruitment rate as the ratio of trees entering the adult height- class to the number of adults dying each year. It uses 14 simple equations, informed by field data, to describe the transfer of individual trees between height-classes (the dynamics of the tree population). These transfers are 192 THE BIOLOGY OF AFRICAN SAVANNAHS affected by elephant damage, fire damage, the effect of wildebeest trampling and the browsing of other antelopes. Our depiction of their results is shown in Figure 5.8 along with our interpretation of events in the Serengeti–Mara ecosystem. Tree recruitment is shown on the vertical axis and the important point is that all outcomes above zero (the black surface) are positive for tree recruitment, while all outcomes below are negative. It is also important to remember that this graph describes recruitment, not how the system will look at a particular time. The ‘look’ is a snapshot that will follow the pattern of recruitment by many years. For example, around 1900 tree recruitment was high but there were few trees. Over the next few decades tree cover increased. One interesting feature of these predictions is that an ecosystem can move to two divergent equilibria, as represented by the Mara and the Serengeti post 1980. With the reduction of fires and fewer elephants, trees have recovered in the northern Serengeti, while across the Tanzanian–Kenyan border the higher incidence of elephants has kept the grassland savannahs open. As a final comment on the ‘elephant–tree story’ it should be noted that medium-sized herbivores may also play a pivotal role in tree dynamics by feeding on the young saplings, and therefore affecting recovery of the tree population. Prins and Van Der Jeugd (1993), working in Lake Manyara National Park, Tanzania, convincingly related even-aged stands of Aca- cia woodland to collapses in local impala populations, caused by episodic anthrax and rinderpest epidemics. Using exclosure experiments in the ripar- ian woodlands of Chobe, northern Botswana, Moe et al. (2009) concluded Serengeti 1980 Serengeti Mara 1890? Serengeti 1990 Masai Mara 1980 4 2 0 0 10 Serengeti –2 1960 20 –4 30 –6 40 Burning rate (%) In (tree recruitment rate) –8 0255075 100 Elephants (% of 1980 population) Figure 5.8 Tree recruitment as a function of elephant numbers and fire incidence, as predicted by the population model of Dublin et al. (1990b). Species Interactions 193 that impala control woodland regeneration ‘. . . after elephants have killed the big trees’. O’Kane et al. (2014) describe a simple browse–browser model, parametrised from an extensive review of the literature and their own data, including quantitative assessment of impala impact, from a study site in iMfolozi Park, South Africa. They conclude that impala, via their impact on the seedlings of woody species, exert a major long-term effect onAcacia woodlands in iMfolozi Park. In fact, any smaller browser, at sufficent den- sity, may heavily impact on tree recruitment (Augustine and McNaughton 2004, dik-dik; Barnes 2001, kudu). In the exclosure experiments of Augus- tine and McNaughton (2004) the role that native ungulates play in regulat- ing woody plant dynamics on commercial rangeland in central Kenya were examined. The most common native browsing and mixed-feeding ungulates (in this area of Laikipia) are impala, dik-dik and elephant. One additional factor affecting shrub distribution in many African savannahs is the pres- ence of nutrient-rich patches (typically c.1 ha in size) that are created by the abandonment of overnight cattle enclosures (Augustine 2002). In Laikipia, these nutrient-rich patches occupy only c.1% of the total landscape, but they are dominated by short, nutrient- rich grasses and are important concentra- tion areas for medium-sized ungulates such as eland and impala (Augus- tine 2002). The absence of shrubs in these nutrient-rich glades is surprising, because abandoned livestock enclosures in northern Kenya are focal points for recruitment and rapid growth of Acacia tortilis (Reid and Ellis 1995). The suggestion is that mixed-feeding ungulates such as impala suppress shrub invasion into glades by consuming shrub seedlings and preventing their establishment. In other words, the nutrient-rich soils of these glades may indirectly suppress shrub invasion and growth by attracting ungulate herbi- vores. Augustine and McNaughton established two paired 0.5 ha (70 × 70 m) exclusion plots in each of three glade and three bushland sites at the Mpala Research Centre, Laikipia, excluding browsers for 3 years. Shrub density at glade sites was low and unaffected by exclusion of large herbivores. Moreo- ver, very few shrubs less than 0.5 m tall were observed in the glades (either inside or outside exclosures) throughout the 3 year study. In the bushland sites, at the scale of individual Acacia twigs, browsers significantly reduced leaf density, leaf biomass and growth rates of twigs less than 50 cm above the ground (within the foraging height of dik-diks), but browsers had no effects on twig leaf density or leaf biomass at a height of 1.0–1.25 m above ground. Reductions in the growth rate of twigs within the foraging height of dik- diks was associated with a sixfold reduction in the rate at which shrubs less than 0.5 m tall were recruited into the 0.5–1.5 m height-class. This reduced recruitment combined with measured rates of shrub mortality in larger height-classes shows that browsers reduced the rate of increase in shrub density nearly to zero (7.1 ± 10.2 shrubs ha−1year−1) compared to the rapid rate of increase in the absence of browsers (136.9 ± 13.6 shrubs ha−1year−1). Damage to shrub canopies by elephants caused large, significant reductions in cover of A. mellifera and Grewia tenax, but lesser reductions in cover of 194 THE BIOLOGY OF AFRICAN SAVANNAHS A. etbaica. For Acacias, elephant damage was focused on shrubs in excess of 2.5 m tall, such that Acacias in intermediate height-classes (0.5–2.5 m) experienced minimal browser impacts. In this dry savannah, elephants influenced shrubland dynamics by altering shrub height-class distributions, shifting species composition from broad-leavedGrewia tenax to fine-leaved Acacia species, and suppressing woody biomass accumulation; but elephants had little influence on changes in shrub density. In Chapter 1 we saw how savannah biomes in different parts of the world, although showing similarities, contained rather different ecosystems. Within African savannahs a variety of broad types exist, sandwiched between deserts and tropical forests. In the latter part of Chapter 2 we saw how within one of these broad types (the Serengeti ecosystem) there were areas containing different soils, rainfall and vegetation. Even the southern grasslands were not uniform, with short, intermediate, and long grass types having their own growth forms, and species. Now we find that even with- in a localised area of savannah there can be yet other vegetation mosaics, produced by plant defences, grazing and fire. Savannah plants, and their associated animals, clearly inhabit a mosaic environment, heterogeneous on many different scales. Predators and prey Just like plants and herbivores, bovids and their carnivores have probably been interacting in African savannahs for at least 5 million years. Carni- vores have developed physical attributes (claws) and behavioural techniques (hunting) that facilitate capture, and prey have developed physical attributes (long thin legs) and behavioural patterns (herding) for escaping capture. In this section we look at some of the characteristics of these ecological interac- tions. One of the predictions of ecological theory is that predator and prey will fluctuate numerically in a regular manner. In the original Lotka–Volt- erra model these predator–prey cycles are rather fragile, with neutral stabil- ity, but they have been replaced with models that have stable limit cycles as part of their behaviour (see Taylor 1984, Chapter 10). Possibly one of the best known, and most debated, examples in vertebrates is the 10 year cycle of snowshoe hares (Lepus americanus) and lynx (Felis lynx) in Canada, for which more than 100 years of population estimates are available from fur- trapping records (Elton and Nicholson 1942, Krebs et al. 1995). The original explanation is that predator numbers drive numbers of prey, but it is just as feasible that predator numbers simply follow prey, who in turn follow their plant food, which in turn follows some climatic variable such as rainfall. In this case, if a plant–herbivore–predator system is driven by a climatic vari- able that doesn’t cycle, or fluctuate in any regular manner, there may be no cycles at all. For most African savannahs we do not have the extensive popu- lation data that would allow us to search for this ‘characteristic’ feature of Species Interactions 195 predator–prey interactions. Long runs of herbivore population data are lim- ited and long runs of predator population data are non-existent (Chapter 4). However, large herbivore populations in Kruger National Park, South Africa, have reportedly shown quasi-cyclic fluctuations extending back to the 1920s (Stevenson Hamilton 1947, Owen-Smith and Ogutu 2003) and these have been analysed by Ogutu and Owen-Smith (2005). The background to these data is detailed in Chapter 4, and here we look only at the reported oscil- lations. Previous authors (Tyson 1986, Owen-Smith and Ogutu 2003) have reported an approximately 18 year cycle in rainfall and therefore Ogutu and Owen-Smith analysed this data also. They analysed the population data for 12 ungulate species (those in Figure 4.22), and 4 of these (buffalo, eland, giraffe and common zebra) are shown in Figure 5.9. They used two kinds of correlation. Autocorrelations within each species (Figure 5.9a), and cross- correlation between each species and rainfall (Figure 5.9b). If populations show regular cycles of abundance, the autocorrelation between peaks will produce positive autocorrelations (with the lag indicating the cycle length), and those between troughs negative autocorrelations (with the lag indicating half the cycle length). The half-cycle lengths indicated for those species in Figure 5.9a are approximately 12–14 years for buffalo, 11–12 years for eland, 16–18 years for giraffe and 16–18 years for zebra. The range of such half-cycle lengths was from 16–18 years (giraffe and zebra) to 8 years (warthog), and there is some indication that species with longer generation times had longer population cycles. Notice, however, that all the suggested cycle lengths are less than the observed cycle in rainfall. There is some effect of rainfall seen (a) buffalo eland giraffe zebra 2 2 2 2 6 6 6 6 10 10 10 10 14 14 14 14 18 18 18 18 –1.0 0.01.0 –1.0 0.01.0 –1.0 0.01.0 –1.0 0.01.0 (b) Lag in years –2 –2 –2 –2 –8 –8 –8 –8 –14 –14 –14 –14 –20 –20 –20 –20 –1.0 0.01.0 –1.00.0 1.0 –1.00.0 1.0 –1.00.0 1.0 Correlation Figure 5.9 Correlograms for four Kruger ungulate species. (a) Autocorrelations with previous population estimates. (b) Cross-correlations between population estimates and rainfall. Lines are 95% tolerance limits (modified from Ogutu and Owen-Smith 2005). 196 THE BIOLOGY OF AFRICAN SAVANNAHS in the cross-correlations (Figure 5.9b), with most species showing a posi- tive correlation with the rainfall of 16–20 years earlier. It is therefore diffi- cult to see these fluctuations in ungulate population size as a simple, direct, response to a cycling environment (rainfall) and the link with generation time may well indicate an influence of predators. Of course, predators in African savannahs do not only catch one type of prey, as in classical predator–prey models such as the Canadian lynx–snowshoe hare system. They take a variety of prey items, as indicated for the larger Kruger carnivores in Figure 5.10. Notice that the two largest prey categories (100–350 kg and >350 kg) are only taken by spotted hyaena and lion, the largest predators. The interaction between predator and prey is constrained by their physical size, and small predators usually catch smaller prey than large predators (of course small predators can catch the smaller young of larger prey). This means that larger prey may escape the predation pressures imposed on smaller prey. Indeed, this was an explanation put forward by Sinclair et al. (2003) to explain the results of their northern Serengeti–Mara analysis, outlined in Chapter 4 (Figure 4.21). In this ecosystem, 28 species of ungulate are prey for 10 species of large carnivore. As in the Kruger ecosystem, each predator species shows a characteristic pat- tern in the size of its prey, and the diet range of smaller carnivores is nested within that of larger carnivore species (Figure 5.11). Of course, within the larger range each predator has its preferred choice of prey. Thus the lion’s preferred prey (38% of animals consumed) is the wildebeest and common zebra (170–250 kg), but 44% of the diet is made up of smaller prey (2–45 kg). n = 214 n = 122 n = 63 n = 61 n = 52 100 Invertebrates 80 Fish Reptiles/amphibians 60 Birds/eggs % Very small mammals < 1 kg 40 Small mammals 1–25 kg Medium sized mammals 25–100 kg 20 Large mammals 100–350 kg Very large mammals > 350 kg 0 spotted lion leopard cheetah wild dog hyaena Figure 5.10 Percentage occurrence of different prey categories in the diets of large carnivores in Kruger National Park (n = number of food items eaten). The white boxes on the side of a species’ histogram shows the proportion of items scavenged (from Mills and Funston 2003). Species Interactions 197 golden jackal (8 kg) black-backed jackal (10 kg) wild cat (4 kg) caracal (16 kg) serval (15 kg) cheetah (8 kg) leopard (60 kg) wild dog (30 kg) spotted hyaena (60 kg) lion (150 kg) 0 100 200 300 400500 Prey weight range (kg) Figure 5.11 The range of weights of mammal prey consumed by carnivores of different sizes in the Serengeti ecosystem (from Sinclair et al. 2003). Buffalo and giraffe comprise 5–15%. Notice however, that in terms of bio- mass consumed, rather than individuals, these three categories are 63%, 14% and 15–24% respectively (Sinclair et al. 2003). Each carnivore therefore includes prey outside its preferred size range, but is inefficient at catching it. Lions are less efficient at catching gazelles than catching wildebeest, but they take gazelles if the opportunity arises. Therefore, despite their smaller size, gazelles are less vulnerable to lion predation than are wildebeest. Because of the ranges shown in Figure 5.11, ungulates with smaller bodies suffer preda- tion from many more predators than do larger ungulates (Figure 5.12). Small species of antelope, such as oribi, are prey to five species of cats, two canids, 8 7 6 5 4 3 2 Number of predators 1 0 01234 Log [herbivore weight (kg)] Figure 5.12 The range of weights of mammal prey consumed by carnivores of different sizes in the Serengeti ecosystem on a logarithmic scale (from Sinclair et al. 2003). 198 THE BIOLOGY OF AFRICAN SAVANNAHS hyaena, and many smaller carnivores in the Serengeti. Medium-sized ante- lopes like wildebeest are prey to only three cats, wild dog and hyaena, while larger ungulates like buffalo and giraffe are predated by lion only. Of course this does not necessarily mean that oribi suffer more predation than wil- debeest. Lion may be very efficient at catching wildebeest. How, and when, do the large carnivores of the African savannah catch their prey? Table 5.3 compares the details for the five largest predators. Recently there has been a renewed interest in the importance of so-called mesocarnivores in ecosystems (Prugh et al. 2009, Ritchie and Johnson 2009, Roemer, Gompper and Van Valkenburgh 2009). A dominant mesocarnivore in Africa may be the black-backed jackal (Canis mesomeles). Kamler, Sten- kewitz, and MacDonald (2013) showed that, in southern Africa, jackals had varying affects on two small carnivores, cape foxes (Vulpes chama) and bat- eared foxes (Otocyon megalotis). Comparing sites with and without jackals, less than 5 km apart, they found that in the absence of jackals, densities of cape foxes increased 64% despite similar food and habitat resources between sites, suggesting that jackals suppressed cape fox populations through a com- bination of predation and competition. In contrast, jackals did not suppress populations of bat-eared foxes. For both fox species, the absence of jackals also resulted in smaller home ranges and non-selective use of habitats for den sites. Another study (Blaum, Tietjen and Rossmanith, 2009) found that numbers of black-backed jackals were inversely related to numbers of cape foxes, bat-eared foxes, and small-spotted genets (Genetta genetta) across 22 sites in South Africa, associated with intensity of jackal control. Bagniewska and Kamler (2014) found a similar negative relationship between jackals and two prey species, South African ground squirrel (Xerus inauris) and hares (Lepus capensis and L. saxatilis), and between jackals and a mongoose (Cyn- ictis penicillata). This relationship is not surprising given that the diet of jack- als in this area comprises 19–36% ground squirrels, 7–27% hares and 0–12% small carnivores, primarily yellow mongooses (Klare et al. 2010). In addition to the restrictions of body size, already mentioned, predators clearly take different species in different areas. This is often because predators simply take those prey that are most abundant in the local area. However, local habitat can also restrict predation. In Kafue National Park, in western Zambia, puku (Kobus vardonii) and lechwe (Kobus leche) are abundant, but they are hunted with great difficulty by lions, who avoid the swampy habitats used by these prey. Seasonal habitat changes can also influence predation. In this same area, buffalo form 38.4% of lion kills in the dry season, when vegetation density is reduced by bush fires, while in the wet season buffalo form only 18.1% of lion kills. Probably one of the most intriguing instances of habitat structure influencing predation is that observed by Prins (1996). He studied buffalo predation by lions in Manyara National Park, Tanzania (see Figure 5.4 for a map of the area). By noting where buffalo carcasses, that Table 5.3 Predation data for the large African savannah carnivores cheetah leopard lion spotted hyaena wild dog Hunting time of day Day Mainly night Mainly night Night and dawn Day Number of animals 1 1 1–5 1–3 for gnu and Whole pack 2–19 hunting gazelles 4–20 for zebra Method of hunting Stalk then long Stalk then Stalk then short Long-distance pursuit Long-distance fast sprint short sprint sprint short pursuit Distance from prey 10–70 m 5–20 m 10–50 m 20–100 m 50–200 m when chase starts Speed of pursuit Up to 87 kph Up to 60 kph 50–60 kph Up to 65 kph Up to 70 kph Distance of pursuit 200–300 m Up to 50 m Up to 200 m 0.2–3.0 km 0.5–2.5 km Hunting success rate 37–70% 5% 15–30% 35% 50–70% Commonest prey Thomson’s and Grant’s impala, Thomson’s zebra, wildebeest, wildebeest, Thomson’s Thomson’s gazelle, species in East Africa gazelle, impala gazelle, dik-dik, buffalo, Thomson’s gazelle, zebra wildebeest, zebra, reedbuck gazelle, warthog Grant’s gazelle Commonest prey springbok, kudu impala, bushbuck, zebra, wildebeest, wildebeest, buffalo, impala, kudu, species in southern (calves), warthog, waterbuck, warthog, buffalo, gemsbok, impala, springbok reedbuck, roan, duiker Africa impala, puku nyala springbok kudu, zebra, giraffe % kills lost to other 10–12% 5–10% ≈ 0% Serengeti to 5% Serengeti 50% carnivores 20% Savuti 20% Ngorongoro % of diet obtained ≈ 0% 5–10% 17–47% 33% 3% by scavenging Data from Bertram (1979), Bothma and Walker (1999). 200 THE BIOLOGY OF AFRICAN SAVANNAHS were the result of lion kills were found (kills are not moved) he was able to estimate how ‘risky’ different habitat types were for buffalo. Sporobolus spi- catus grasslands were risky, bare mudflats were relatively safe. However, to understand the relationship between lions and buffalo it is necessary to look at combinations of vegetation type rather than individual vegetation patches in isolation. Lions try to stalk their prey unobserved. They therefore require vegetation cover. In Manyara, buffalo spend most of the day in the open grassland, where it is difficult for predators to approach. The best strategy for lions is to observe their potential prey from cover, in the ecotone between grasslands and woodlands. Here they can look for the best position to start their attack when buffalo move in their direction. The frequency distribution of kills in Figure 5.13 shows that this border between closed and open vegeta- tion (an ecotone) is a killing zone for buffalo. Ecotone Closed vegetation Open vegetation 16 14 12 10 Riverine grassland 8 Number of kills (%) 6 Foot slope grassland 4 2 0 800 600 400 200 0 200 400 600 Distance from ecotone (m) Figure 5.13 Distribution of buffalo probably and certainly killed by lion (solid bars) and those possibly killed by lions (open bars) in relation to the distance from the ecotone between open grasslands and woodlands. The small peak on the left indicates a second, minor, ecotone between foot-slope grasslands and woodlands. The peak at the right indicates the riverine delta grasslands where lions can easily hide in the many gullies (from Prins 1996). Species Interactions 201 Of course one of the reason that buffalo are relatively safe in the open grass- lands and mudflats is that they can be vigilant, and this leads us to consider the defences against predation that prey employ. Many savannah ungulates live in groups. There are benefits and costs to such behaviour (see Krause and Ruxton 2002 for a complete review); here we briefly examine some of the stated anti-predator benefits. These basically fall into three categories: increased vigilance (‘many eyes’ theory), predator confusion, and dilution of risk. Increased vigilance, with increased group size, is a consequence of more eyes looking for predators coupled with the transmission of information about detected predators throughout the group. As a result, an individual in a group doesn’t need to detect a predator itself in order for it to be aware of the potential attack. The ‘many eyes’ idea suggests therefore that as group size increases an individuals ‘share’ of the group vigilance can decline, leav- ing more time for other activities such as foraging. Figure 5.14 shows some data for impala in Nairobi National Park (Shorrocks and Cokayne 2005). Selected individuals within both breeding herds and bachelor herds were observed, and the amount of time devoted to eight behaviours noted. These were foraging, vigilance (head up), moving, grooming, allogrooming, lying down and urinating. Some of these behaviours are not mutually exclusive (e.g. an individual can move and be vigilant). Figure 5.14a shows the data for randomly selected females within breeding herds. Selected individuals spent less time being vigilant in larger groups, and overall individuals spent more time foraging when they were less vigilant (Figure 5.14b). In southern Africa, Smith and Cain (2008) found that vigilance in impala decreased with density rather than numbers. Interestingly, it may be unsafe for individuals to ‘cheat’ and spend all their time foraging. FitzGibbon (1989) found that cheetahs hunting Thomson’s gazelle preferentially attacked low-vigilance individuals. The second anti-predator effect of living in groups, predator confusion, is thought to arise because of an inability of the predator to single out and attack individual prey within a group. They are distracted by too many targets. It’s an appealing idea, but empirical support for the confusion effect is limited. The dilution effect basically says that there is safety in numbers. Included within this category of group benefits are two separate probabilities: a lower risk of detection, and a lower risk of capture when detected. The hypothetical example in Figure 5.15 shows the probability of detection during the search phase of the predator, and the probability of being selected as a target by the predator during the attack phase. The combination of the two probabilities shows how the overall probability of being captured by a predator could be reduced by being in a group. However, for this to work, the predator must only take one prey during an attack, and the probabilities must be ‘similar’ to those in Figure 5.15. For example, if we changed the risk of detection for the 1 group to 2 , perhaps because it was more easily seen from a distance (the risk 202 THE BIOLOGY OF AFRICAN SAVANNAHS (a) 30 20 Vigilance (%) 10 0 0204060 Group size (b) 100 80 60 40 Percentage foraging 20 0 0204060 Vigilance (%) Figure 5.14 Vigilance in impala. (a) Relationship between group size and percentage time vigilant for breeding females (fitted line is y = 19.15 − 0.2x, F = 4.74, P = 0.05). (b) Correlation between percentage time vigilant and percentage time foraging for 38 individuals in both breeding and bachelor herds (from Shorrocks and Cokayne 2005). 1 for individuals would now be 6 ), the overall group probability would now 1 1 1 1 1 rise to 2 ×=3 6 , the same as the overall individual probability ()6 ×=1 6 and the ‘safety in numbers’ would disappear. In fact, the only data on this kind of risk assessment, in savannah ungulates, suggest that the situation is less straightforward than the simple theoretical example in Figure 5.15. Creel and Creel (2002) calculated the risk of death for an individual wil- debeest in the Selous Game Reserve, Tanzania, when the predator was the Species Interactions 203 Risk of detection Risk of attack Overall risk 1/4 1 1/4 1/4 1 1/4 × 1 = 1/4 Single prey 1 1/4 1/4 Predators 1/3 1/4 1/4 × 1/3 = 1/12 Grouped prey 1/3 1/3 Figure 5.15 The suggested anti-predator dilution effect of being in a group. In the search phase (left- hand box) there are four targets each with a detection probability of 1/4. In the attack phase (middle box) the risk of attack is 1/3 for the three individuals in the discovered group, while the risk of attack is 1 if you are a discovered lone individual. The right-hand box gives the overall probabilities of being attacked for an individual in a group or alone (modified after Krause and Ruxton 2002). wild dog (Lycaon pictus). This measure of overall risk took into account all the processes by which herd size might affect an individual’s vulnerability to predation. These included risk of detection, the decision to hunt or not, the success of hunts, and the final risk of attack. Notice that not all these are depicted in Figure 5.15. Their results are shown in Figure 5.16, and it is important to remember that these are probabilities obtained from observa- tions made in the field, not a theoretical assessment. The risk of death does in fact change with group size. However, the lowest individual risk of death is obtained when herd size is intermediate (around 40), not at the largest group size actually observed in Selous Game Reserve. Larger groups of wil- debeest are more easily detected than smaller groups and this cancelled out any benefit of reduced risk of attack. This advantage for individuals in groups of intermediate size leads us to consider the distribution of group sizes natu- rally observed in savannah ungulates. The distributions of group (herd) size have been recorded for several African ungulates (Sinclair 1977, Wirtz and Lörscher 1983, Creel and Creel 2002). They usually show a right-skewed frequency distribution (Figure 5.17a and b). Small groups are very common, medium-sized and large groups less so. However, if we re-plot such distribution data in terms of the group size experienced by each individual, we find that most individuals in fact chose to 204 THE BIOLOGY OF AFRICAN SAVANNAHS 0.026 0.022 0.018 -corrected risk 0.014 0.010 0.006 0.002 Diluted encounter 100 18 80 16 wildebeest herd size 60 14 12 40 10 8 20 6 Pack size 4 0 2 Figure 5.16 The risk of death for an individual wildebeest as a function of wildebeest group size and wild dog pack size (from Creel and Creel 2002). live in groups of intermediate size (Figure 5.17b and d). In other words, large groups are uncommon; small groups are common but comprise a relatively small number of individuals. Being in a larger group might confer better anti-predator advantages (higher vigilance, lower risk of attack), but being in too large a group has additional costs, such as increased competition for resources and, from the wildebeest example, higher risk of detection. There appears to be a medium group size that maximises the benefits and mini- mises the costs. Of course these benefits and costs will vary from species to species, so we might expect group size to vary for savannah ungulates with different ecologies. In a paper that has influenced African ecologists for 40 years, Jarman (1974) compiled information on 75 species of ungulates and compared their body size, diet, group size, habitat preference, and anti-predator behaviour. His conclusions were based entirely on narrative description. Using more quan- titative statistical methods this data has been re-examined by Brashares et al. (2000) and their conclusions, relevant to group size and anti-predator behaviour, are shown in Figure 5.18. Jarman’s (1974) conclusion that group size and body size varied predictably with feeding style was confirmed (Figure 5.18a). Non-selective roughage-feeding antelopes are larger, and occur in larger groups, than selective feeders. Antelopes that flee when Species Interactions 205 (a) (b) 50 0.6 0.5 40 0.4 30 0.3 20 0.2 Number of herds Proportion of herds 10 0.1 0 0 0 200 400 600 800 1,000 1,200 1 to 5 6 to 10 11 to 20 21 to 50 >50 Group size impala herd size (c) (d) 10000 0.45 9000 0.4 8000 0.35 7000 0.3 6000 0.25 5000 0.2 4000 3000 0.15 Number of individuals 0.1 2000 Proportion of individuals 1000 0.05 0 0 0 200 400 600 800 1,000 1,200 1 to 5 6 to 10 11 to 20 21 to 50 >50 Group size impala herd size Figure 5.17 Group (herd) size in two African savannah ungulates. (a) Frequency of group size in the African buffalo, in Serengeti National Park; (c) same data but showing the group size experienced by individuals (data from Sinclair 1977). (b) Frequency of group size in impala, in Selous Game Reserve, (d) same data but showing the group size experienced by individuals (data from Creel and Creel 2002). faced with a predator are more likely to occur in larger groups than those that avoid detection by freezing or hiding (Figure 5.18b). The original nar- rative explanation (Jarman 1974) is still the ‘best story’ to explain these interrelated ungulate traits (body mass, group size, diet and anti-predator behaviour). Small-bodied species such as steenbok and dik-dik have small mouths and narrow muzzles that facilitate their specialisation on the most nutritious parts of plants (diets A and B in Figure 5.18). The clumped dispersion and limited availability of these high-quality plant parts, coupled with the rela- tively higher metabolic demands of small-bodied mammals (see Figure 5.25), results in competition for food and adoption of territorial spacing behaviour. These species are therefore likely to live alone or in pairs, in closed habitats, in which freezing or hiding is the best predatory defence. At the other end 206 THE BIOLOGY OF AFRICAN SAVANNAHS (a) 100 Diet A Non-selective feeders 50 Diet B Diet C Diet D Diet E 10 5 Group size 1 Selective feeders 51050 100 500 1,000 Body mass (kg) (b) 100 50 Flee 10 5 Group size Hide 1 51050 100 500 1,000 Body mass (kg) Figure 5.18 (a) Group size versus body mass (log10 axes) for 75 species, or subspecies, of African antelope separated by diet (Jarman 1974). A, selective browsers; B, feed selectively on grass or browse; C, feed on a range of grasses and browse; D, feed unselectively on grasses; E, feed non-selectively on a wide range of grasses and browse. Lines are least-square regressions. (b) Group size versus body mass (log10 axes) for 75 species, or subspecies, of African antelope separated by anti-predator response (Jarman 1974). Anti-predator behaviour is divided into two general categories, antelope that flee to avoid predation (●) and those that hide (○). Lines are least-square regressions (from Brashares et al. 2000). Species Interactions 207 of the size spectrum, large-bodied species (such as eland and buffalo) feed less selectively on coarse grasses (diets D and E in Figure 5.18). In part this is because they lack the morphology and dexterity to feed selectively. The widespread supply of coarse grasses results in little competition for food (but see later in this chapter) and therefore little selection for spacing behav- iour. Because of the low nutritive value and seasonal availability of coarse grasses, there would be little selection for spacing behaviour. These species would have to forage over much greater distances, in more open habitats. This would favour formation of large groups, and the anti-predator tactics of group vigilance, and fleeing when attacked. Parasites and hosts We know from serological studies that many African mammals contract dis- eases of various kinds. In other words, when tested, they are found to contain antibodies to various pathogens. For example, at least 11 viruses, 5 bacteria and 4 protozoa are recorded from free-ranging populations of wild dogs (Lycaon pictus) (Woodroffe, Ginsberg and MacDonald 1997). Adult spotted hyaena (Crocuta crocuta) from the Serengeti have antibodies to at least 10 pathogens, 4 of which (rabies, canine herpes, canine parvovirus and canine adenovirus) are identical to those listed for wild dog (Mills and Hofer 1998). With the cats there is recorded exposure of cheetah to anthrax and feline coronavirus, of lions to canine distemper virus, and of lions and leopards to feline immunodeficiency type viruses. Many of these diseases also affect domestic animals, which appear to serve as a reservoir for the pathogens, from which they can be transmitted to wildlife populations. However, this type of serological data does not tell us how severe the effects of the pathogens might be. A high seroprevalence (many individuals having antibodies) may simply indicate that most animals become infected early in life, but that the resulting disease is mild and most animals recover and become immune. Alternatively, the same high seroprevalence might indicate that the population has recently been exposed to a highly virulent disease, and that only those individuals that survived (and are therefore seroposi- tive) are now present. However, although we do not know how severely these pathogens affect most savannah animals, we do know that some pathogens cause high mortality in some species. Rabies (a rhabdovirus), for example, is an endemic, multispecies disease in many areas of sub-Saharan Africa, with sporadic epidemic cycles. It has been diagnosed in 33 carnivore species and 23 herbivore species. In 1989, a pack of wild dogs living at Aitong outside the Masai Mara Reserve, southern Kenya, was decimated by rabies. The follow- ing year at least one wild dog died of rabies in the nearby Serengeti National Park, and in 1991 all the wild dog packs under study in the Serengeti eco- system disappeared (Figure 5.28). Rabies was suspected, and is also known to have killed wild dogs in Namibia and Zimbabwe. An epidemic of rabies is 208 THE BIOLOGY OF AFRICAN SAVANNAHS known to have halved a population of Ethiopian wolves in the Bale moun- tains (Sillero-Zubiri et al. 1996). In the Aitong pack the disease spread rap- idly (rabies is transmitted mainly by biting). The time from the recording of the first infected dog to the death of the last of the 21 dogs that died was less than 2 months (Kat et al. 1995). Animals such as wild dogs, living at low density in groups, may therefore be at particular risk, since disease seems to have caused local population declines in other areas. For example, sightings of wild dogs declined dramatically after an outbreak of anthrax (which is known to kill wild dogs) in ungulates in the Luangwa Valley, Zambia, and population declines of wild dogs in Zimbabwe in the 1980s coincided with an epidemic of rabies in jackals. One of the most dramatic disease outbreaks in a carnivore population occurred in the Serengeti National Park in the early 1990s. By July 1994, an epidemic of canine distemper virus (CDV) had affected 20–30% of the 3,000 lions in the Park, and in a monitored popu- lation of 250 lions, 87 died or disappeared. However, the lion population recovered rapidly (Cleaveland et al. 2008). CDV was originally thought to be mainly a disease of canids. However, blood samples from Serengeti lions in 1985 suggest that the population may have been initially exposed to CDV around 1980 and that the virus has been increasing its pathogenicity in felids since that time. In addition, during this lion epidemic, several spotted hyae- na cubs below the age of 6 months died from CDV, and a molecular analysis of the virus isolated from hyaenas and lions indicated that they were more closely related to each other than to CDV found in the nearest domestic dog population (Mills and Hofer 1998). Cases of hyaenas killed by CDV have also been reported from South Africa, Namibia, Zambia, Malawi and Ethiopia. Of course the most dramatic African wildlife epidemic did not directly affect carnivores, although it almost certainly affected them indirectly (see Chapter 4). This was the rinderpest epidemic that spread throughout sub-Saharan Africa from about 1890. Rinderpest is a highly contagious viral disease of cattle, domestic buffalo, and some species of wildlife. The name is German, meaning cattle-plague. It is characterised by fever and high mortality. Rin- derpest virus is a single-stranded RNA virus in the family Paramyxoviridae, genus Morbillivirus. It is immunologically related to CDV, human measles virus, peste des petits ruminants virus, and marine mammal morbilliviruses. Most of the historical spread of rinderpest is likely to have occurred in the Indian subcontinent, the Near East, and Europe, following the domestication of ungulates. Since ungulate species are at low density in the Sahara desert they would probably not have supported a continuous infection of the disease. The desert would therefore have acted as a barrier to the spread of rinderpest into Africa. The pandemic that occurred at the end of the 19th century was initiated by the accidental introduction of a few infected cattle into the Horn of Africa in 1890. This caused a major pandemic which spread mainly via the movement of cattle along trade routes. It moved westwards and southwards and reached southern Africa in 1896. This initial ‘1890 epidemic’ is thought Species Interactions 209 Table 5.4 The main recorded epidemics of rinderpest in East Africa and the species of wildlife affected Year Species affected 1890 Most ungulate species 1897 hartebeest and kudu 1913–1921 eland and giraffe, then buffalo, bushbuck and reedbuck 1929 buffalo, bushbuck, and warthog, then eland and waterbuck 1931 buffalo, giraffe and wildebeest 1937–1941 buffalo, eland and giraffe, then buffalo, eland and kudu 1949 eland, and then wildebeest 1960 eland, kudu, warthog, buffalo, bushbuck, giraffe, impala and oryx Modified from Simon (1962). to have killed 80–90% of all cattle in sub-Saharan Africa. Unfortunately, wild ungulates are infected by contact with cattle, e.g. by drinking at the same water source, and Table 5.4 shows the main recorded epidemics of rinder- pest in East Africa and the wildlife species affected. More recently, another rinderpest outbreak that raged across much of Africa in 1982–84 is estimated to have cost at least US$500 million in domestic livestock losses. The impact that rinderpest had on wildlife populations is seen in the dramatic sixfold increase in the wildebeest population of the Serengeti ecosystem following the disappearance of rinderpest after a cattle vaccination campaign in the early 1960s (Chapter 4). This was followed by an increase in predator popula- tions, particularly the spotted hyaena. Another disease that has had severe effects on wildlife populations is anthrax (Bacillus anthracis). After a susceptible host ingests spores the bacilli enter the bloodstream, replicate exponentially, and produce fatal septicaemia. In the Kruger ecosystem in southern Africa, epidemics occur in episodes of 6–20 years. These episodes tend to occur during dry periods, and in seasons when overabundant populations of herbivores congregate around stagnant surface water. Although most mammal species in Kruger are susceptible to anthrax, kudu appear particularly vulnerable, with blowfly maggots playing a pivotal role in the epidemic. Blowflies feed on infected carcasses. Infected droplets produced by these blowflies contaminate leaves that are browsed by kudu, which are infected and die, so repeating the cycle. Vultures and hyae- nas also act as disseminators of the disease. Although lions in Kruger have been known to develop acute septicaemia, carnivores are generally more resistant to anthrax, only developing a localised buccal form of the disease in the membranes of the head. For example, during a serious outbreak in Etosha National Park, Namibia, lions, spotted hyaenas, brown hyaenas and black-backed jackals all fed from infected carcasses but showed no signs of the disease themselves. Similarly, during an epidemic in the Luangwa val- ley in 1987, one area of only 80 km2 yielded the carcasses of 101 hippos, 60 210 THE BIOLOGY OF AFRICAN SAVANNAHS buffalo, and 20 elephants, along with puku, kudu and other ungulates, but only one spotted hyaena and two leopard carcasses. Wafula, Atimnedi and Tumwesigye (2007) discuss the anthrax outbreaks of 1959, 1962, 1991 and 2004/5 among hippos in Queen Elizabeth National Park, Uganda. During the latter outbreak, 306 hippos died, representing 12% of the population. However, they comment that it seems to be a mechanism for ‘weeding out the aged’ and that the population recovered after this dip in numbers. One of the complications in assessing how severe the effects of these patho- gens might be, is that their effects depend upon the condition of the host. Animals that are stressed may be more susceptible to disease, and the subse- quent effects of a disease might be more severe. For example, canine distem- per appears to be more pathogenic to wild dogs kept in captivity than it is in free-ranging populations. Another example is provided by Sinclair (1977). Sarcoptic mange, in the Serengeti, affects both buffalo and wildebeest calves. However, while it was commonly seen in the severe dry season of 1968 and 1969, it was not observed in the mild seasons of 1971 and 1972. In fact Sin- clair (1974) suggests that since disease (and predators) kill mainly animals in poor condition, it is agents such as malnutrition that kill many of these animals, rather than disease. Pathogens and predators may simply hasten the process (Chapter 4). Climatic effects also influence disease. In the Kruger ecosystem, anthrax and epidemic foot and mouth disease are typically asso- ciated with the dry season, when animals congregate at water holes, provid- ing ideal conditions for disease transmission. Rift Valley fever, African horse sickness, theileriosis and other tick-borne diseases are associated with the wet season when these disease vectors (carriers) are more abundant. Competitive interactions (− −) First, some definitions and observations. Competitive interactions can be intraspecific (between individuals of the same species; see Chapter 4) or interspecific (between individuals of different species). This section of Chap- ter 5 looks at interspecific competition between African savannah species. An important distinction can be made between interference competition and exploitation competition, although elements of both can often be found in the competitive interactions between species. With exploitation compe- tition, individuals interact indirectly, via a common resource that is limit- ing, or in short supply (e.g. the grazing ungulates considered later). With interference competition, individuals interact directly, damaging each other in the process of obtaining a resource that might, or might not, be in short supply (e.g. the kleptoparasitism of carcasses by predators, also considered later). However, the essence of interspecific competition is that individuals of one species suffer a reduction in fecundity, survival or growth as a result of resource exploitation, or interference, by individuals of another species. Species Interactions 211 Interspecific competition is frequently highly asymmetric. The consequenc- es are often much greater for one of the species involved (e.g. wild dog in the kleptoparasitism example below). Finally, in discussions of interspecific competition one frequently encounters the term ‘ecological niche’. This term was first used by Joseph Grinnell in 1917 to simply mean the range of physical conditions under which a species could survive and reproduce. It thus had only physical dimensions. Now, however, we include resource dimensions, such as seed size, grass species, grass height, prey size, foraging height, and so on. Only part of each niche dimension is used by a species, not simply all that between two limits. Most ecologists visualise the use as being described by a ‘utilisation function’, often represented in models as a normal curve. For example, a seed-eating bird might take all seeds between small and large, but prefer (and therefore use more) seeds of intermediate size. Figure 5.19 shows a visualisation of this idea. Usually a species has a larger ecological niche (fundamental niche) in the absence of competitors than it has in their presence (realised niche). When we discussed predator–prey interactions earlier we stated how impor- tant the Lotka–Volterra model had been in forming ecologists’ views on those interactions and their population consequences. This is even more so for the Lotka–Volterra model of two-species competition. From this theo- retical model, and early laboratory studies, arose the widely held view that interspecific competition led either to competitive exclusion of one species, Rainfall Optimum part of niche gerenuk giraffe Niche overlap Browsing height Figure 5.19 A visualisation of the ecological niche. Two niches are shown, each with two dimensions. Because both types can be important in real niches, one physical dimension (rainfall) and one resource dimension (foraging height) are illustrated. To help place this idea into a savannah context, gerenuk and giraffe have been used as plausible examples. Gerenuk prefer more arid savannahs, and browse lower down in trees and shrubs, than giraffe. The diagram is modified from Shorrocks (1978). 212 THE BIOLOGY OF AFRICAN SAVANNAHS or coexistence. The Lotka–Volterra model predicts that coexistence will only occur if intraspecific competition is greater than interspecific competition, for both species. This condition is achieved if the two species occupy differ- ent niches. Then, individuals of each species will meet individuals of their own species more frequently than they meet individuals of the other species. One of the weaknesses of this ‘solution’ to the coexistence problem is that the Lotka–Volterra model does not predict how dissimilar the two niches have to be. Later models, exemplified by MacArthur and Levins (1967) and May (1973), attempted to rectify this and define the limits to similarity that would allow coexistence. These models, using normally distributed resource utilisation curves, and competition between three species along one niche dimension, suggested the simple solution that d/w > 1 would lead to sta- ble coexistence, where d is the distance between adjacent resource utilisa- tion curve peaks and w is the standard deviation of these curves. However, Abrams (1983) found that alternative resource utilisation curves, and com- petition in several dimensions, would often lead to lower values of d/w being compatible with coexistence. There may be no general, precise, solution to this problem, and savannah ecologists may have to measure specific resource utilisation curves and dimensions. In theory, interspecific competition is obviously important—excluding some species and determining which species coexist with others. But how impor- tant are these theoretical effects in the real world? There is no doubt that competition sometimes affects community structure. Equally, no one now believes that it is always of overriding importance in each and every case. If other agents (e.g. lions, rinderpest and rainfall) keep numbers low, so that competition is negligible, then competition cannot be a potent force in struc- turing communities. One way of answering this question is by examining the results of field manip- ulation experiments, in which one species is removed from, or added to, an ecosystem and the response of the other species observed. Schoener (1983) looked at 164 studies in the ecological literature. Unfortunately, published field manipulations are not representative of the world’s ecosystems. In the terrestrial studies, for example, most of those found by Schoener were carried out on temperate systems. Conclusions may therefore be biased. Neverthe- less, he found that 89% (terrestrial), 91% (freshwater) and 94% (marine) of worldwide studies demonstrated the existence of interspecific competition. When he looked at single species, rather than single studies, he found that 76% showed competitive effects sometimes, and 57% showed competitive effects always. Connell (1983) reviewed just 72 studies (215 species and 527 different experiments) of field manipulations. He found evidence of compe- tition in all the studies, in more than 50% of the species, and in about 40% of the experiments. These two literature surveys suggest that interspecific com- petition might be rather important in structuring all ecosystems. But what Species Interactions 213 about tropical savannahs? There is no indication from these two surveys that different ecosystems behave very differently, so we might assume that savan- nah ecosystems have competitive effects of this order of magnitude. What is more, these surveys suggest that vertebrate studies showed even higher values than the mean percentages given above, while many insect studies gave lower values. Competitive interactions may therefore be particularly important in the savannah’s large-mammal community. How have savannah ecologists attempted to study interspecific competition? There have been some ‘natural’ experiments with savannah ungulates, and we examine some of these below. However, savannah ecologists (along with many other ecologists) have more frequently looked at community patterns that might result from interspecific competition. Uncovering niche differ- ences is one such method that we examine here. Other community patterns are examined in Chapter 6. However, if several species coexist, and use the habitat in slightly different ways (i.e. have slightly different niches), is this evi- dence of competition? Many ecologists think it is, and believe these patterns are the result of past competitive interactions (the ‘ghost of competition past’, Connell 1980). However, as Begon et al. (1996) say ‘The trouble is . . . there is no proof’. The danger with using pattern to infer process (and this includes the statistical techniques of correlation and regression) is that usually a pat- tern can be caused by more than one causative agent. A competition pattern could be present as a result of non-competitive influences. Giraffe and kudu may well forage at different heights (see Figure 5.27) as a result of past com- petition. They may also do this simply because they are different sizes: a result of past phylogenetic pathways that have nothing to do with competition. We have to distinguish between species traits being the specific result of compe- tition (and therefore indicating past competition), and simply adaptations to a general mode of life. Resource partitioning and niche studies provide a start to examining interspecific competition, but savannah ecologists need to col- lect additional data to strengthen any competitive interpretation. This could include, for example, evidence of present resource limitation, observed niche shifts when one species is absent (a natural experiment), and looking for alternative explanations. It is against this general competition background that you should interpret the savannah examples below. The structure of savannahs: competition between trees and grass A central question in savannah ecology concerns the mechanism allowing the long-term coexistence of trees and grasses (Hanan and Lehmann 2011, van Langevelde et al. 2011). We have already touched upon this topic in Chapter 2, when the central role of rainfall on tree–grass dynamics was brief- ly examined (see Figure 2.2). From this graph it is clear that mean annual 214 THE BIOLOGY OF AFRICAN SAVANNAHS precipitation (MAP) sets an upper limit to tree cover. However, within this rainfall limit there are other limiting factors at work. These are fire (also indi- cated in Figure 2.2) and herbivory, and we have also looked at how these can limit tree recruitment in this chapter (see Figure 5.8). Notice that rainfall, fire and herbivory all appear to ‘limit’ trees, rather than ‘promote’ grass. The implication in Figure 2.2 is that the regression line shows the limit imposed on tree cover by MAP, but that for any particular MAP, tree cover can be lower because of fire or herbivory. But what role does interspecific competi- tion play in this? Traditionally, many savannah ecologists believed that trees and grasses competed for water in the soil. Following from the Lotka–Vol- terra model, this led to the line of argument that if trees and grasses com- pete for essential resources, yet they coexist over a wide range of savannahs, they must in some way have separate ecological niches. For example, Walter (1971) suggested that there was a separation of ‘root niches’ with trees having sole access to deeper soil water and grasses having superior access to surface soil water. Notice that this implies that it is small saplings, with surface root systems, that will be most affected by interspecific (actually inter-lifeform) competition. In fact, this effect of grasses on Acacia karroo saplings was observed by Mopipi et al. (2009). Of course, young trees will also be more vulnerable to browsing and fire damage. Biologists working in different Afri- can savannahs have claimed field evidence both for (Helsa et al. 1985, Knoop and Walker 1985), and against this root niche proposal (Belsky 1990, Le Roux et al. 1995, Seghieri 1995, Mordelet et al. 1997), although those working in eastern and southern Africa seem to favour the root niche idea, while those working in the humid savannahs of western Africa seem to find positive evi- dence lacking. There may be a difference between the two types of regional savannah. However, there may also be a difference of interpretation. Seghieri (1995) studied three sites in a savannah in northern Cameroon. Root profiles consistent with the root niche idea were found and yet the author failed to be convinced of the usefulness of the root niche idea. Figure 5.20 shows three 1 metre 1 metre Acacia seyal 02040 Lannea humilis 02040 Acacia hockii 02040 0 0 0 Humidity (%) Root depth of 40 40 40 annual grass 80 80 80 120 120 120 Depth (cm) Depth (cm) Depth (cm) Figure 5.20 Root profiles of three woody species from northern Cameroon. The humidity profile, to the right of each root profile, shows the moisture content of the soil (modified from Seghieri 1995). Species Interactions 215 root profiles, for three woody species, in three different soils. Importantly, previous study had shown that the two deeper-rooted woody species (Acacia seyal and A. hockii) actively grew over a longer period than the rainy season while Lannea humilis grew only during the rainy season. Of course coexistence between trees and grasses, resulting from a long- term stable equilibrium of the classic Lotka–Volterra kind, may be an illu- sion. We saw earlier in this chapter, with the elephant–fire–tree dynamics of the Serengeti–Mara system, that tree cover may well go up and down over time, because of long-term changes in tree recruitment. These demo- graphic events can lead to ‘non-equilibrium’ grass–tree coexistence through climatic variation and/or disturbance which limits successful tree seedling germination, establishment and transition to mature size classes. Sankaran et al. (2005) reviewed both the classic equilibrium (competition) and non- equilibrium (demographic) proposals, and attempted to integrate the two views (Figure 5.21). Competition-based proposals include the root niche separa- tion already mentioned, phenological niche separation, and balanced com- petition. Niche separation by phenology is based on the fact that savannah trees are able to store water and nutrients, and therefore achieve full leaf expansion quickly with the onset of rain (Scholes and Archer 1997). They also tend to retain leaves for several weeks at the end of the growing wet sea- son, after grasses have started to senesce. The idea is that trees have exclusive access to resources early and late in the growing season. They have niche sep- aration in time. With balanced competition, coexistence is achieved because the superior competitor (trees) is limited by intraspecific competition before it can exclude the inferior competitor, grasses (Scholes and Archer 1997). TREE Germination bottlenecks Seeds Seedlings Grazing and re Storage effects Saplings GRASS re browsing Balanced Adults Root and phenological competition niche separation Figure 5.21 Framework showing the demographic and competitive explanations for grass–tree coexistence in savannahs. Solid arrows show transitions between tree life-history components and open arrows represent competitive effects, both intra- and interspecific (modified from Sankaranet al. 2005). 216 THE BIOLOGY OF AFRICAN SAVANNAHS In contrast, proponents of the demographic view maintain that the critical problem for savannah trees is demographic and not competitive in nature. Tree recruitment, and transition into the adult life stage, can be limited by fire, drought, and browsing. For example, variation in tree seedling estab- lishment over time caused by random variation in rainfall, coupled with low adult tree mortality, leads to tree recruitment being pulsed in time. In effect the recruitment potential of trees is stored (sensu Warner and Chesson 1985) in the adult population and released in good years. In arid savannahs, tree recruitment can also be enhanced by the localised deposition of tree seeds in herbivore dung. Understandably, tests of these various explanations have usually been site specific (Sankaran et al. 2005), and inevitably all are supported in some sites and none are supported in all. Figure 5.21 is prob- ably an accurate reflection of what happens between grasses and trees, but the detectable strengths of the arrows in Figure 5.21 will vary considerably. What we have tried to show throughout this book is that although savan- nahs are a recognisable biome, local variation in soils, rainfall, fire, herbivory, and predation impose subtle variations on the savannah theme. This why in Figure 2.2 there is a clear upper limit to tree cover, imposed by MAP, but an amazing range of site-specific variations beneath it. The structure of savannahs: the effect of invasive Prosopsis Exotic plant invasions are known to reduce native plant species richness and alter ecosystem functioning (Bethune, Griffin and Joubert 2004). Inter- estingly, comparatively fewer studies have assessed the impacts of invasive plants on native fauna (Pyšek et al. 2012). Invasive plants have a bottom- up effect on higher trophic levels (Vilaet al. 2011)(see Chapter 5) and can impact wildlife communities by altering the characteristics of their envi- ronment such as vegetation structure, resource quality and habitat diversity (Mills and Retief 1984, Ferdinands, Beggs and Whitehead 2005). Plants of the genus Prosopis L. are invasive species of global concern because of their aggressive nature in introduced ranges, the effects their dense stands have on water resources, and the detrimental impact their introduction has had on open grasslands and woodland ecosystems (Global Invasive Species Data- base, et al. (2014) examined a 93 km stretch of the Swakop river, in the Namib- Naukluft National Park, in the central Namib desert. Thirteen camera traps were placed at 7 km intervals along the dry riverbed and presence/absence data was analysed using the occupancy software PRESENCE (MacKenzie 2012). Eight species had good enough occupancy estimates to allow compar- ison between sites with high vs low abundance of Prosopsis. Of these, three were not affected by Prosopis: greater kudu (Tragelaphus strepsiceros), black- backed jackal (Canis mesomelas) and klipspringer (Oreotragus oreotragus). Five species showed significant sensitivity to Prosopis abundance: common duiker (Sylvicapra grimmia), springbok (Antidorcus marsupialis), Cape hare (Lepus capensis), African wildcat (Felis silvestris lybica) and an unidentified species of the Muridae family. The first three species showed a decline in occupancy with increasing Prosopis abundance, while the last two species were higher with increasing Prosopis abundance. Since Prosopis pods and seeds constitute a significant part of small rodent diet, and small murids are eaten by wild cats, the increase of the last two species is probably related. So this ‘invasive system’ may elicit both − − and + − interections. Resource competition: niche differentiation among herbivores Eltringham (1974) described an event in Queen Elizabeth National Park, Uganda, that could be viewed as a removal experiment indicating interspe- cific competition. In 1957 a high density of hippopotamus was artificially reduced to very low numbers and then maintained at this level until 1967. Buffalo, which were originally low in numbers, increased sixfold by 1968. Eltringham believed that this was because vegetation cover, and therefore buffalo food, increased threefold when the hippopotamus were removed. Other areas of the Park, unaffected by hippopotamus, did not show such increases in vegetation or buffalo numbers. This type of field manipulation experiment, which might throw light on the importance of interspecific competition in structuring savannah communities, is rare. Instead, savan- nah ecologists have looked for natural experiments (Diamond 1986) of the kind mentioned in Chapter 4. Several reintroductions into protected areas have been made that can be used to suggest the presence, or absence, of inter- specific competition between savannah ungulates. In Kruger National Park, white rhinoceros (a grazer) was successfully reintroduced (Pienaar 1963, Penzhorn 1971, Novellie and Knight 1994) even though it is approximately the same weight as hippopotamus, also a grazer, suggesting they don’t com- pete for their grass resource. On the other hand, the oribi reintroduction in 1962 was not successful (Pienaar 1963), perhaps because of competition with klipspringer and steenbok, which are a similar size. However, the diets of these two species contain more browse than that of oribi (Chapter 3). In the Bontebok National Park, in southern Africa, common reedbuck did not 218 THE BIOLOGY OF AFRICAN SAVANNAHS establish itself (Novellie and Knight 1994), possibly because of interspecific competition from the similarly sized blesbok (both are predominantly graz- ers that take the occasional browse). Unfortunately these reintroductions, and others mentioned by Prins and Olff (1998), can fail for a variety of rea- sons that may have nothing to do with interspecific competition. We usually do not have enough information to be sure. The migrating wildebeest of the Serengeti ecosystem provide two natural experiments, where one species increases its numbers, allowing savannah ecologists to observe the effects on other potential competitors. The first such experiment (Sinclair and Norton-Griffiths 1982) was a singular event. During the 1960s, following the disappearance of rinderpest, the Serengeti wildebeest population underwent a dramatic increase in numbers (Chap- ter 4). The second experiment (Sinclair 1985) occurs repeatedly every year as huge numbers of wildebeest arrive in the northern part of the Serengeti in Tanzania, and the Masai Mara in Kenya. Since both experiments involve the migrating wildebeest, and its two fellow travellers, it is convenient to pause and briefly describe this amazing seasonal event. Observations on the migrations of the Serengeti wildebeest date back to the mid-1950s (Pearsall 1957, Swynnerton 1958, Grzimek and Grzimek 1960b, Talbot and Talbot 1963) with more detailed data analysis provided by Pen- nycuick (1975) (Figure 5.22). Wildebeest spend the wet season (December– April) on the grass plains in the south-east of the ecosystem. At the start of the dry season they move north-west, into the western corridor (May– July), finally moving north as the dry season advances (Aug–November). They finally return south as the new wet season approaches. The timing of the movement to and form the plains depends upon the rainfall. In dry years they move north earlier, and move further north. There has also been a long-term change in the pattern of migration (Figure 5.22). After 1969 the wildebeest migration penetrated further north into the Masai Mara. This is thought to be a consequence of the increase in wildebeest numbers in the 1960s. Zebra and Thomson’s gazelle show a similar pattern of movement, although Thomson’s gazelle tend to remain longer on the plains and never move beyond the western corridor. At the end of the wet season some zebra move west, but many move directly north (Figure 5.22). The three migrating species have similar but slightly different diets (overlapping but not identical niches) and in looking for evidence of present-day competition this should be born in mind. Wildebeest and Thomson’s gazelle are ruminants, zebra are not (Chapter 3). Wildebeest take green grass, Thomson’s gazelle green grass and herbs, and zebra can take dry grass (Chapter 3). We now return to the two natural experiments. Sinclair and Norton-Griffiths (1982) examined the results of the 1960s wil- debeest population increase on the other two species of migrating ungu- late: common zebra and Thomson’s gazelle. Since there is evidence for Species Interactions 219 wildebeest 1960–1969 1969–1973 Dec–Apr May–Jul Aug–Nov zebra Thomson's gazelle Dec–Apr May–Jul Aug–Nov Figure 5.22 Summaries of the ungulate migrations in the Serengeti–Mara ecosystem. The top maps show the wildebeest migration for two separate time periods, 1960–1969 and 1969–1973. The bottom graphs show the zebra and Thomson’s gazelle migration between August 1969 and August 1972 (from Maddock 1979, based on data in Pennycuik 1975). intraspecific competition for green grass in Serengeti wildebeest (Chapter 4), it seems reasonable to assume there may also be interspecific competition for green grass between wildebeest and other grazing ungulates. If this was so, the dramatic increase in wildebeest during the 1960s should have had a demonstrable, long-term, effect on the species that take part in the seasonal migration. Figure 5.23 shows the numbers of all three species over the rel- evant time period. There was no significant change in the zebra population over the 20 years. Perhaps this is not surprising since they have a food niche that is probably sufficiently different from wildebeest for them not to -com pete. Being hindgut fermenters, zebra can process large amounts of dry grass quickly, an option not available to the slower, but more efficient, ruminant digestive system (Chapter 3). However, there has been a decline in Thomson’s 220 THE BIOLOGY OF AFRICAN SAVANNAHS 1.6 1.4 1.2 1.0 0.8 0.6 Numbers´ 1,000,000 0.4 0.2 1960 1970 1980 Figure 5.23 Population estimates of wildebeest (●), Thomson’s gazelle (×), and zebra (■) in the Serengeti. Vertical lines are one standard error (modified after Sinclair and Norton- Griffiths 1982). gazelle. There is a significant differenceP ( < 0.02) between the estimates of 1971 and 1980, although not between 1971–1978 and 1978–1980. Since there is some overlap in the diet of wildebeest and gazelle this decline may be viewed as a consequence of interspecific competition. However, interpreta- tion of Figure 5.23 is made more complicated by the possible presence of a mutualistic interaction between the three species, which we outline in the final section of this chapter. In a second paper, Sinclair (1985) examined the effect that wildebeest have on other ungulates when they arrive in the northern Serengeti and Masai Mara. Observations were made from a vehicle travelling through the Mara Reserve, and all animals within 250 m either side of the vehicle were record- ed (Chapter 4). Numbers counted for wildebeest, and the eight potential competitors, are shown in Table 5.5. The increase in wildebeest and zebra in the July and August dry-season counts, is clearly seen. The other migrating species, Thomson’s gazelle, does not get this far north during the seasonal movement (Figure 5.22) but there is a resident population in the Masai Mara. Interestingly, all seven resident species show a decrease in observed numbers once the wildebeest arrive. This could reflect a movement of individuals in response to the arrival of vast numbers of migrating grazers. If present-day interspecific competition is present, we might expect some displacement of resident ungulate species from certain areas when the very large numbers of wildebeest arrive. During the vehicle transects, Sin- clair looked at this possibility by recording each ungulate’s use of (1) habi- tats (grassland savannah, open woodland savannah, thickets or riverine), (2) grass species (Themeda triandra, Eragrostis tenuifolia, Hyparrhenia Species Interactions 221 Table 5.5 Numbers of individuals counted during the Mara transects reported in Sinclair 1985. In the text all but the first two species are referred to as ‘residents’ Species June 1982 July 1982 August 1983 wildebeest 443 55,003 224,936 zebra 2,208 23,173 28,191 Thomson’s gazelle 8,290 4,867 9,629 Grant’s gazelle 695 455 443 topi 5,184 4,173 6,370 coke’s hartebeest 1,339 1,030 447 impala 3,709 3,177 3,971 waterbuck 135 53 292 warthog 299 120 237 filipendula, Pennisetum mezianum or Cynodon dactylon), (3) grass greenness (green, half green, green trace or dry), and (4) grass height (10, 25, 50, 75 or 100 cm). Of course there is some overlap in these categories. For exam- ple, grass species are correlated with habitat. The preferred habitat of wil- debeest was short, green, Themeda grassland savannah. In June before the wildebeest arrived, all species (with the exception of impala who preferred open woodland savannah) had 30% of their population in this ‘wildebeest habitat’. Surprisingly, this proportion remained similar in July when 50% of the newly arrived wildebeest population were using this habitat. Only at the height of the dry season (August) did use of this habitat drop markedly for zebra, topi and Coke’s hartebeest, although it actually increased for Thom- son’s gazelle. Sinclair also calculated the percentage overlap between pairs of species. Overlap with wildebeest was high and did not generally decrease in the dry season. However, there was a decrease in overlap in habitat use (17 pairs) and grass height use (18 pairs). Overlap for grass species was the same as that for habitats, and overlap for grass greenness remained high (average 96%) and constant for all species. However, there was a high degree of over- lap between all species. So is competition not important in structuring these ungulate assemblages? Certainly the evidence for present-day competition is equivocal. But what about its ghost? These ungulates might not be compet- ing now because past competition has produced niche separation that allows them to coexist now. There is a long history of looking for niche separation in savannah sys- tems (Fryxell et al. 2008). These include separation byhabitat (Lam- prey 1963, Bell 1970, Jarman 1972, Ferrar and Walker 1974, Sinclair 1977, Ben-Shahar and Skinner 1988, Ben-Shahar 1990, Macandza, Owen-Smith and Cain 2012a, 2012b), by plant food species (Field 1968, Jarman 1971, McNaughton and Georgiadis 1986, Prins et al. 2006), and by plants parts eaten (Gwynne and Bell 1968, Bell 1970, Sinclair 1977). 222 THE BIOLOGY OF AFRICAN SAVANNAHS A case can certainly be made for niche separation. Some ungulates are rumi- nants (wildebeest) and some are not (zebra), thus allowing both the efficient processing of green and dry grass (Chapter 3). Some species are browsers on trees (giraffe), some are browsers on herbs (Grant’s gazelle), some are grazers on grass (wildebeest), and others are mixed feeders (impala). With- in grazers, some are large bodied with wide mouths (buffalo) while some species are small bodied with narrow mouths (Thomson’s gazelle), allow- ing them to be more selective feeders. Figure 5.24 shows how this dissec- tion of resources can be used to visualise niche separation between potential 0 1 Grant’s gazelle % Browse % Herbs Thomson’s gazelle 1 0 giraffe wildebeest 0 % Grass 1 zebra impala topi buffalo 0 1 topi wildebeest % Leaf % Stem buffalo impala 1 0 0 zebra 1 % Sheath Thomson’s Gazelle Figure 5.24 Triangular graphs showing the percentage composition of ungulate diet in the Serengeti system. The top graph shows separation into grass, herbs, and browse (trees and shrubs). The bottom graph shows separation of the grazers into three grass components (see Chapter 2) (data from Sinclair and Norton-Griffiths 1979). Species Interactions 223 ungulate competitors in the Serengeti ecosystem. In the top graph there is a separation into woody browse, herbs and grass, while in the bottom graph the separation is a finer one into ‘parts’ of the grass resource (Chapter 2). The resources used by the major herbivores do therefore appear to differ. This is partly related to body size. In general, among related species with simi- lar digestive capabilities, smaller species require better-quality food because they have a relatively higher metabolic rate, but larger species require larger absolute quantities of food (Figure 5.25a). The relationship between energy turnover and body size is energy turnover =+KW0.75 where K is a constant and W = body weight. The relationship between body weight and protein turnover is very similar: proteinturnover =+KW0.74 The significance of this is that the maintenance requirement of protein per unit of body weight increases with decreasing body size, in the same way as the requirement for energy increases (Figure 5.25b). Given a series of animals, such as ruminants, with comparable digestive sys- tems, the smaller animals require a diet with a higher proportion of pro- tein and soluble carbohydrates, at the expense of fibre. These differences have tended to produce, at one end of the spectrum, small ungulates with small narrow mouths adapted for carefully selecting discrete, high-quality food items. At the other end of the spectrum are large species with mouths adapted for rapid ingestion of large quantities of undifferentiated items, pos- sibly of low quality (Figure 5.26). Notice that this size/specialist ranking also corresponds to the order of movement in the Serengeti migration, as the dry season approaches. In the Serengeti it also corresponds to the order of (a) (b) 50 ) 0.03 –1 d ) –1 40 –1 0.02 30 animal –1 20 0.01 10 (kg d Daily food intake 0 body mass (kg W 0.00 0 1,000 2,000 3,000 4,000 Daily food intake per uni t 041,0002,000 3,000 ,000 Body mass (kg) Body mass (kg) Figure 5.25 The relationship between body mass and (a) daily food intake per animal and (b) daily food intake per unit body mass for six African grazers. The species are, smallest to largest, Thomson’s gazelle, impala, wildebeest, buffalo, hippopotamus, and elephant (data from Delany and Happold 1979, graphs from Prins and Olff 1998). 224 THE BIOLOGY OF AFRICAN SAVANNAHS Thomson’s Grant’s Weight (kg) gazelle gazelle 0 20 57 100 200 300 400 500600 700 45 133 215 315 625 impala topi wildebeest zebra buffalo Figure 5.26 Scale of body weights for a selection of African savannah ungulates. The three species involved in the seasonal Serengeti migration are shaded. the lesser movement down the catena (slopes) as the dry season approaches, including buffalo, topi and impala in the correct order. This type of niche separation is not just confined to grazers. There is also evidence that browsers may also have different niches, thus avoiding inter- specific competition. Figure 5.27a shows the popular natural history view of this resource separation. Du Toit (1990) observed the height at which four species of ungulate were browsing in Kruger National Park, South Africa (Figure 5.27b). Giraffe allocated almost 90% of feeding time to feeding above the height ranges of kudu, impala and steenbok. Kudu allocated 33% of their feeding time to the height range 1.2–1.7 m, which was little used by giraffe and impala, and beyond the reach of steenbok. Among kudu, impala and steenbok there was a common pattern of increased mean feeding height dur- ing the dry season. The implication from Figure 5.27b is that feeding height stratification separates these species on an important niche dimension, thus promoting coexistence. It should be remembered, however, that mature trees (a) (b) 6 5 5 giraffe 4 4 3 m 3 kudu 2 Height (m) impala 2 1 steenbok 1 0 0 0.75 0 0.75 0 0.75 0 0.75 Proportion of feeding time Figure 5.27 Niche separation of savannah browsers. (a) the popular view (from Orbis Publications 1977), (b) data from Kruger National Park, South Africa (from Du Toit 1990). Species Interactions 225 grow from seedlings and saplings. Excessive browsing by the smaller spe- cies on these young trees might have long-term consequences for the mature tree population. Browsing height stratification between giraffe and kudu + impala + steenbok, does not exclude the possibility of resource competition. The search for niche differences has more recently led to more complex forms of multivariate statistical analysis, such as correspondence analysis. Examples for 10 ungulates from the northern Transvaal (including impala, sable, roan, and waterbuck) can be found in Ben-Shahar and Skinner (1988), and for 3 small ungulates from southern Mozambique (including suni) in Prins et al. (2006). However, these are simply more sophisticated detection and visualization techniques, and are still subject to all the caveats outlined in the introduction to this section on interspecific competition. Kleptoparasitism: African wild dogs and spotted hyaenas As we have seen, many field observations of competition have relied heav- ily on observing changing numbers over time. One species arrives in an area, or a resident species increases its numbers, while another species declines in abundance. An interesting example with carnivores is provid- ed by the increase in abundance of the spotted hyaena (Crocuta crocuta) in the Serengeti–Mara ecosystem, and the subsequent decline of the Afri- can wild dog (Lycaon pictus). As we saw in Chapter 4, spotted hyaena is the most abundant large predator in the Serengeti–Mara system. Between 1969 and 1976 the spotted hyaena population increased by 50%, probably as a response to the herbivore increase, following the disappearance of the rin- derpest virus. Some estimates for the subsequent population of hyaena in the Serengeti were shown in Table 4.7. Over this period the wild dog has shown a decline in numbers, with all remaining dog packs finally disappearing from the Serengeti system in the early 1990s (Figure 5.28). During the early 1970s part of this decline in wild dog numbers was certainly due to hunting (wild dogs were regarded as vicious vermin to be exterminated). However, the rea- son for the later slow decline is less certain, but interference competition with spotted hyaena may have been important. Both spotted hyaenas and wild dogs hunt in packs and run down their prey (mainly Thomson’s gazelle, wildebeest, and zebra; Table 5.3). Spotted hyae- nas frequently follow dog packs when they go hunting from a den (during the breeding season), and steal their kill (Lawick and Lawick-Goodall 1970, Lawick 1973). Dogs can defend a kill, but about four dogs are required to keep off one hyaena. Although food still on the hoof is probably not a limiting resource for large carnivores in the Serengeti, food in the form of kills may be. Particularly during the breeding season, dog packs are hunting not only for themselves, but also for the dominant bitch and her pups back at the den. Losing kills to competitors such as hyaenas may have a serious effect on pup survival. Because the wild dogs’ ability to defend kills depends on the ratio of 226 THE BIOLOGY OF AFRICAN SAVANNAHS 80 60 40 20 Number of dogs 0 70 72 74 76 78 80 82 84 86 88 90 92 Year Figure 5.28 Number of wild dogs in the Serengeti ecosystem. dogs to hyaenas, this effect may have become gradually more serious over the last 30 years. Additional evidence supporting this explanation comes from the correlation (r = −0.92, P = 0.01) between hyaena density and dog density over a series of ecosystems in eastern and southern Africa (Creel and Creel 1996; Figure 5.29a). Aggressive interactions at wild dog kills confirm that the frequency, and impact, of interference competition increases where hyaena density is high and where visibility is good (Creel and Creel 1996). Wild dogs also appear to suffer competition (and predation) from lions, and there is a similar negative correlation (r = −0.91, P = 0.03) between wild dog density and lion density in a number of savannah systems (Figure 5.29b). Kleptoparasitism between spotted hyaenas and lions also occurs. In the absence of adult male lions, hyaenas can drive female and subadult lions off their kills, provided they outnumber the lions by a factor of four. Cooper (a) (b) 0.04 Selous, Tanzania 0.04 Selous, Tanzania Hluhluwe, South Africa 0.03 0.03 0.02 0.02 Kruger, South Africa Kruger, South Africa Serengeti (1967–1979), Tanzania Serengeti (1967–1979), Tanzania 0.01 Serengeti (1985–1991), 0.01 Serengeti (1985–1991), Tanzania Tanzania Ngorongoro, Ngorongoro, Wild dog population density Tanzania Wild dog population density Tanzania 0 0 0 0.5 1 1.5 0 0.05 0.1 0.15 0.2 0.25 Hyaena population density Lion population density Figure 5.29 Correlation between population density of wild dogs and (a) spotted hyaena and (b) lion, over a number of African ecosystems (from Creel and Creel 1996). Species Interactions 227 (1991) found that in Chobe National Park, Botswana, where there was a shortage of adult male lions, the groups of female and subadult lions lost almost 20% of their food to hyaenas. Losses were most frequent for those lions living in small groups. In contrast, Trinkel and Kastberger (2005) found that in Etosha National Park, Namibia, spotted hyaenas were unable to pre- vent kleptoparasitism by lions, and failed to acquire kills from lions. This appeared to be due to the low ratio of hyaenas to lions, and the presence of male lions. Another predator that may suffer interference from hyaenas and lions is cheetah. Looking at distributional data, Durant (1998) suggested that, in the Serengeti National Park, cheetah avoid areas where hyaena and lions are common, even though these areas have lots of prey. This behaviour is strongest when cheetah are most likely to suffer loss of prey to these two larger predators—that is, when hunting. Mutualistic interactions (+ +) In savannah systems, mutualisms are frequently mentioned but much more rarely studied. In truth this is a general comment on most ecological systems (Vandermeer 1984). Mutualistic interactions in African savannahs include nodulated legumes (e.g. between root bacteria and Acacia species), ruminant digestion (between cellulose digesting bacteria and ruminants), pollination (e.g. between honeybees and Acacia species, fig wasps andFicus species), seed dispersal (e.g. between indehiscent Acacia and many herbivores), protection (between ant species and Acacia species), guiding behaviour (between the honey guide Indicator indicator and the honey badger), and scavengers (e.g. between vultures and large carnivores). Typical of a mutualism lacking detailed scientific study, but none the less described repeatedly in savannah texts, is that of the greater honey guide (Indicator indicator) and the ratel, or honey badger (Mellivora capensis) (see Chapter 3). Nice accounts can be found in Kingdon (1977) and Estes (1991). The greater honey guide has been observed to solicit help from honey badgers, humans and baboons. When it sees a potential follower it approaches within 5–15 m, chirring (a sound similar to shaking a box of matches rapidly) and fanning its tail to expose the white outer feathers. The guiding bird may lead its follower anything from a few metres to 2 km, and the journey may take up to half an hour. Once a beehive has been located the bird sits quietly in a nearby tree until its fellow honey-hunter has eaten from the hive and retired. It then approaches the broken hive and feeds on beeswax, grubs and eggs. The honey badger is reputed to have a most unusual method of dealing with the bees before it breaks open the hive with its powerful claws (Kingdon 1977): according to African honey- hunters the honey badger backs up to the hive, and uses its anal glands to 228 THE BIOLOGY OF AFRICAN SAVANNAHS fumigate the bees. The bees either flee or become inactive. This mutualism is facultative rather than obligate, since ratels are known to find beehives without the aid of Indicator, and non-guiding birds are found with bees- wax in their crops. An important example of mutualism is that between Acacia and herbivores. Acacia pods are of two types: (1) dehiscent, in which the pod splits and the seeds are dispersed by wind, and (2) indehiscent, in which the pods remain on the tree until removed by browsers, or as in A. tortilis the pods drop to the floor but do not split. The seed pods of indehiscentAcacia are eaten by a wide variety of wildlife including elephant, white rhinoceros, giraffe, eland, kudu, impala, steenbok, several rodents and ostrich. The seeds from inde- hiscent pods have a very hard coat which enables some of them to travel through the gut of herbivores and pass out unharmed in the faeces. However, many are destroyed by chewing and digestion. The herbivores benefit by eat- ing the nutritious pods, but the Acacia species benefit by having their seeds dispersed and fertilised. For the Acacia this association is more or less obli- gate. Because their seeds have a very tough outer coat they do not germinate unless they pass through the gut of a herbivore. Miller (1995) found that the germination rate for undigested seeds of A. tortilis was 7%, while for seeds retrieved from the stomach of kudu it was 48% and for seeds from kudu dung it was 60%. Feeding facilitation and the grazing succession The possibility of feeding facilitation through increased access to resources was first suggested by Vesey-Fitzgerald (1960). His observations were made in tall floodplain grasslands in the Rukwa valley, Tanzania, where trampling and feeding by elephants exposed medium-height grasses to buffalo. Buf- falo in turn generated shorter grass that was eaten by topi. Further evidence to support a grazing succession was provided by Gwynne and Bell (1968) and Bell (1970, 1971) who examined the movement of buffalo, zebra, wil- debeest, topi and Thomson’s gazelle down the catenas (see Chapter 2) of the western Serengeti. In the wet season these species concentrate on the higher ground where short grasses provide the best grazing. As the dry sea- son approaches they move down the catena, in order of decreasing body size, to feed on the taller but poorer-quality grasses on the lower ground. However, perhaps the most well-known example of a grazing succession is that provided by the Serengeti seasonal migrations already mentioned (Bell 1971). First to migrate are zebra, followed by wildebeest and finally Thomson’s gazelle. Again the movement is in decreasing order of body size. The successive nature of this movement, at least in the Kirawa area of the western Serengeti ecosystem, is shown in Figure 5.30, using data from Bell (1969). Peak numbers of zebra moved through a 300 m transect in April, peak numbers of wildebeest in July, and peak numbers of Thomson’s gazelle Species Interactions 229 300 200 gazelle 100 Numbers wildebeest zebra 0 JFMAMJ J ASOND Month Figure 5.30 Average daily numbers of zebra, wildebeest, and Thomson’s gazelle recorded in a 300 m transect at Kirawira, in the western Serengeti. Note that the numbers of wildebeest are divided by 10 to accommodate them on the same graph (data from Bell 1969). in September. The original idea was that zebra (a hindgut fermenter) lead the succession, opening up the herb layer by trampling and increasing the relative frequency of grass leaf by consuming stems. They therefore increase the suitability of the vegetation structure for wildebeest (a large ruminant). These in turn reduce the quantity of grass leaf, facilitating the use of herbs by Thomson’s gazelle (a small ruminant). One way to see this idea is as a multispecies extension of the ‘grazing lawn’ idea mentioned earlier in this chapter, and indeed the possibility of self-facilitation has been raised (Owen-Smith 1988): in effect, herbivores manipulating their plant food to their own advantage, or in this case that of other species. It’s a good story, but is it true? Of course we know that such facilitation could only be facultative, since not all the wildebeest follow the zebra (Sinclair and Norton-Griffiths 1982, Sinclair 1985, see below) and most of the migrating Thomson’s gazelle do not follow the zebra and wildebeest into the northern Serengeti and Masai Mara (Figure 5.22). In addition, only the two following species could gain an advantage. Zebra cannot gain any mutualistic advantage since they are at the front, leading the succession. Zebra may, however, gain an advantage in another way—by avoiding competition (Sinclair 1985). In June 1980, a systematic aerial survey of the complete Serengeti zebra population (using a grid of 1 km2) recorded 5,555 zebra of which 71% were within 1 km of the wildebeest herds. The great majority of these zebra had a well-defined posi- tion on the leading edge of the herds (Figure 5.31), confirming that they do indeed precede the wildebeest in the migration. Of course, not all the zebra were found with the wildebeest herds during this wet season; 29% were scat- tered over a large area many kilometres away. In the Masai Mara, during the dry season, this proportion increased to 55% in July and 50% in August, suggesting a change from significant association with wildebeest in the wet season to significant avoidance in the dry season. Sinclair maintains that this change in zebra behaviour is brought about by a continuous conflict between 230 THE BIOLOGY OF AFRICAN SAVANNAHS 120 wildebeest 80 Numbers 40 zebra 024 6810 Distance from herd edge (km) Figure 5.31 Location of zebra (●) along transects through the large wildebeest herds (□), on the Serengeti Plains in the wet season. Numbers are averages of eight transects. All wildebeest and zebra within a 200 m radius of the vehicle were counted (from Sinclair 1985). two other interactions—predation and interspecific competition. Wilde- beest are the preferred prey of all the large carnivore species in the Serengeti ecosystem (Kruuk 1972, Schaller 1972). Therefore zebra (and other ungulate species) could reduce the risk of predation on themselves by staying close to wildebeest. However, zebra may also suffer by being too close to wilde- beest because of disturbance and interspecific competition. Sinclair suggests therefore that when food is not limiting (in the wet season) the zebra move close to the wildebeest for protection, but nonetheless stay ahead of the main herd to reach the available food first. During the dry season when grass is limiting (Chapter 4) pressure from interspecific competition may outweigh that from predation, and the zebra move further away from the wildebeest herds. Sinclair is suggesting that what we are seeing in the Serengeti migra- tions (at least between zebra and wildebeest) is a response to predation and interspecific competition, rather than facilitation. McNaughton (1976) found evidence that Thomson’s gazelle were attract- ed to areas on the Serengeti plains where previously grazing wildebeest had improved the quality of the grass sward. What is more, the Thom- son’s gazelle maintained this preference for up to 6 months after the pas- sage of the wildebeest. Unfortunately, this local facilitation does not appear to have translated into a population response. There has been no detectable increase in Thomson’s gazelle following the eruption of the wil- debeest population in the 1960s (Figure 5.23); in fact, quite the opposite. An excellent review on this topic is provided by Arsenault and Owen- Smith (2002). Species Interactions 231 Acacia ants In sub-Saharan Africa 1,686 species of ants have been recorded (Bolton 1995), but at least three times this number probably remain to be described. Many plants provide food for ants in the form of extrafloral nectaries. These sites of sugary secretions are well away from any flower, often in the leaf peti- oles or blades. The ants are attracted to the plant and their presence results in some protection from herbivores for the host plant. For example, in Mexi- can species Janzen (1966) showed that 2.7% of shoots with ants had her- bivorous insects, while 38.5% of shoots without ants had insect herbivores. This type of plant–insect interaction is particularly well developed in savan- nahs, with several species of Acacia tree. In about 10% of spiny Acacia spe- cies, the base of the spines becomes greatly expanded, forming a pseudogall (Figure 5.32a). These are not true galls because they are not induced by the (a) (b) Figure 5.32 Ants and Acacia. (a) Whistling thorn, Acacia drepanolobium; (b) the four Acacia drepanolobium ants. (Head colour/thorax colour/abdomen colour, R = red and B = black), left to right: Crematogaster mimosae (RRB), C. nigriceps (BBR), C. sjostedti (BRR), and Tetraponera penzigi (BBB) ((photograph a by Jo Shorrocks, drawing b by Dino Martins). 232 THE BIOLOGY OF AFRICAN SAVANNAHS insect. The ants chew a hole in the new, soft, green pseudogall and hollow out the inside as a site for a mini-colony. As the gall grows older it darkens, and becomes harder and more protective. The colony of ants established on an Acacia tree aggressively removes any insect herbivore, frequently consuming them, and may even attack mammalian herbivores, such as giraffe. One of the problems that the tree has with this relationship, however, is that it needs insects to visit in order to pollinate its flowers. Intriguingly, work on Acacia zanzibarica, in Mkomazi Game Reserve in northern Tanzania, suggests that the ants are mainly active at dawn and dusk, while most pollinating bees and butterflies are generally active during the middle part of the day (Willmer, Stone and Mafunde1998). Because of this diurnal rhythm, bees and butter- flies almost never meet an ant. A particularly detailed study of these ant–Acacia mutualistic systems has been carried out in the Laikipia district of central Kenya (Young, Stubblesfield and Isbell 1997, 2002, Goheen and Palmer 2010, Palmer et al. 2013). In this semi- arid bush savannah Acacia drepanolobium, the whistling thorn, often makes up 95% of the trees (about 2000 trees h−1). Since virtually all whistling thorns over 0.5 m tall are occupied by ants it has been estimated that Acacia ants may account for up to 25% of the animal biomass in this ecosystem. In most plants there is a pair of scale-like appendage at the base of the leaf-stalk, or petiole. In whistling thorn these stipules are developed into a formidable pair of thorns, up to 7 cm long (Chapter 2). Between 5% and 40% of these paired thorns share a hollow inflated base that is 1.5–3.5 cm in diameter (Figure 5.32a). Ants produce winged reproductive forms that disperse to found new colonies. These newly mated winged females explore the exterior surfaces of whistling thorn looking for new, unoccupied galls. Having found a suitable swollen thorn, the female removes her wings, initiates a conversion of her flight muscles into food and chews an entry hole into the swollen thorn. She then plugs up the hole from the inside and remains sealed within the swollen thorn until her first workers are produced. Acting together, ants and sharp thorns form a quite effective defence against browsing by small to medium- sized herbivores. Madden and Young (1992) reported a negative correlation between the number of swarming ants and the duration of feeding bouts for giraffe calves r( = −0.91, P < 0.05). Adult giraffe feeding seemed less sensitive with a positive, but non-significant, correlation with ant swarming r( = 0.41, P = 0.10). Figure 5.33 shows the results of a nice manipulative field experi- ment in which four different types ofAcacia drepanolobium branch were offered to goats (Stapley 1998). Branch type 0 had thorns and ants removed, type 1 had ants but no thorns, type 2 had thorns but no ants, and type 3 had both ants and thorns (the natural condition). Both ants and thorns appear to have an effect upon the mean amount of vegetation eaten, and the mean number of bites taken, by the goat. Campbell et al. (2013) found a similar defence of Acacia erioloba in Namibia. Experimental exclusion of ants led to greater levels of damage from hemipteran and coleopteran herbivores. In Species Interactions 233 5.00 4.00 3.00 2.00 (g/30 secs) 1.00 Mean amount eaten 0 012 3 20 16 12 8 4 Mean number of bites 0 012 3 0.40 0.30 0.20 per bite (g) 0.10 Mean amount eaten 0 0123 Branch defence type Figure 5.33 The effect of ants and thorns on browsing in Acacia drepanolobium. 0 = no ants or thorns, 1 = ants but no thorns, 2 = thorns but no ants, 3 = both ants and thorns (from Stapley 1998 Oecologica, 115, pp. 401–405. With kind permission of Springer Science and Business Media.) the Laikipia district, four species of ants are the main occupants of these Aca- cia pseudogalls (Figure 5.32b). Tetraponera penzigi, Crematogaster nigriceps, and C. mimosae depend entirely on the interiors of swollen thorns for nesting space, and specialise on using A. drepanolobium. Crematogaster sjostedti also occurs on a less common swollen-thorn acacia called A. seyal variety fistula (Chapter 2). This last acacia-ant nests mainly within dead Acacia tissue, but 234 THE BIOLOGY OF AFRICAN SAVANNAHS its workers may also occupy swollen thorns. With the three Crematogaster species, a single colony of ants usually spans several trees, while with T. pen- zigi a colony occupies a single whistling thorn. The four ant species are very intolerant of each other and fight over the possession of trees. Observations suggest that a hierarchy exists with the order C. sjostedti > C. mimosae > C. nigriceps > Tretraponera penzigi. Of course, the Acacia–insect system is in fact even more complex than just trees and ants. Martins (2013) examined the influence of three ant species, bruchid beetle seed predators (mainly Bruchidtus sp.), and their parasitoid wasp (Dinarmus magnus), on the production and germination success of Acacia drepanolobium seeds. Martins examined almost monospecific stands of this Acacia in south-central Kenya, and in this region C. sjostedti is absent. Across the 5 years of the study, trees occupied by the three ant species differed significantly in fruit set: 94.8% (C. mimosae), 82.2% (T. penzigi) and 25.6% (C. nigriceps). However, for trees that did set fruit, there were no significant differences in seed produced, or germination success, between trees with dif- ferent ant species. Levels of bruchid beetle infestation were high (c.50% of seeds), but did not differ between trees with different ant species. Neither did numbers of parasitoids emerging. What did emerge from this study is that there is a significant mutualistic affect of the parasitoids on the trees. Only 6.2% of bruchid-infested seeds germinated versus 78.6% of normal seed, but 18.4% of seeds with parasitoids germinated, showing that the parasitoids exhert a small ‘rescue’ effect on bruchid-infested seeds. If this rescue effect is widespread it could have an important effect on recruitment at a commu- nity level. The abundance of bruchids and parasitoids suggests that, given the dominance of Acacias in East African savannah ecosystems, their role in the establishment of new seedlings may be as great as that of browsing mammals on the demography and survival of Acacia trees Scavenging vultures: mammalian carnivores facilitating birds? There are basically three ways in which a carnivore feeding on large ungu- lates can obtain food. It can hunt, steal kills of other hunters (kleptoparasit- ism), or scavenge from the remains of carcasses of animals that have been killed by other predators, or have died from other causes. In the Serengeti ecosystem Houston (1979) estimated that approximately 40,094,000 kg of ungulates die each year, mainly (70%) from starvation and disease. He also estimated that the five major predators (lion, spotted hyaena, cheetah, leop- ard and wild dog) consumed about 14,362,750 kg and that kleptoparasitism was of minor importance (but see earlier). This implies that, each year, there is approximately 26 million kg of dead animals not eaten by mammalian carnivores, available for avian scavengers. In the Serengeti, almost all these are vultures, the only other common scavenger species being the marabou Species Interactions 235 stork (Leptoptilos crumeniferus), with a body weight of 5.5 kg, but the stork is unable to tear flesh from a carcass and relies on other species to do this. It also has alternative food, taking fish, frogs and other small vertebrates. The figure of 26 million kg of carrion is a huge resource. Even if some 15% of this is consumed by fly larvae and bacteria, it leaves enough food to eas- ily explain the huge numbers of vultures seen in the Serengeti, and pre- sumably all other savannah ecosystems. Houston (1979) estimated that it would require an average of 25,000 griffin vultures to consume this carrion, a figure consistent with estimated vulture numbers in the Serengeti. This relationship between vultures and large carnivores may be truly mutual- istic. Not only do vultures gain from the kills of carnivores but there is anecdotal evidence that lions frequently home in on areas where they have seen vultures descending. Oxpeckers and herbivore: mutualism or parasitism? The two species of oxpeckers Buphagus( erythrorhynchus and B. a. africanus) are aberrant starlings endemic to Africa. They spend most of the day actively clambering about on ungulates, and the association is usually described as an example of an interspecific cleaning relationship, and therefore mutualis- tic (Dickman 1992). Cleaners benefit from the consumption of ectoparasites (Attwell 1966) plus ear wax, blood and other edible material gathered from the host (Bezuidenhout and Stutterheim 1980). Hosts benefit from the atten- tions of oxpeckers in several ways. For example, morbidity and mortality caused by ectoparasites are reduced (Lightfoot and Norval 1981, Norval et al. 1988), oxpeckers’ alarm calls warn of approaching predators (Attwell 1966), and hosts spend less time self-grooming (Mooring and Mundy 1996). Also, in a mutualistic association, hosts might be expected to solicit the attentions of oxpeckers, as is actually observed in impala (Aepyceros melampus) (Moor- ing and Mundy 1996), Burchell’s zebra (Equus quagga burchelli) and desert warthog (Phacochoerus aethiopicus) (Breitwisch 1992). However, studies in captive settings suggest that the oxpecker–ungulate association may not be consistently beneficial to the host. In a situation where alternative host species are absent, oxpeckers open up old wounds, increasing the time taken for them to heal (Weeks 2000) and create fresh wounds (McElligott et al. 2004). The observation that oxpeckers feed on blood from open wounds, and sometimes prefer host tissue to ectopara- sites (Weeks 1999, Plantan et al. 2013), raises the possibility of parasitism rather than mutualism within the oxpecker–host relationship. However, wound feeding by oxpeckers on wild hosts appears to be rare, since Plan- tan (2009) found it constituted only 3.1% of feeding events. This could be due to either the abundance of alternative wild hosts with ectoparasites, or resistance behaviour by the host to diminish parasitism. The study of Bis hop and Bishop (2014) suggests that the latter plays a part. They examined 236 THE BIOLOGY OF AFRICAN SAVANNAHS ungulate–oxpecker interactions in Nakuru National Park, Kenya and found that both waterbuck (78% of individuals exhibiting the behaviour) and impala (57%) performed resistance behaviours that prevented the mutualism becoming parasitism. However, only 6.6% of buffalo showed such behaviour. This suggests that in the wild this interaction is almost always mutualistic. 6 The Savannah Community and Its Conservation This chapter deals with a number of community-level topics for which there is savannah data. We have already looked at individuals (Chapters 2 and 3), populations (Chapter 4), and simple interactions between species (Chap- ter 5). However, in much the same way that a population is more than a sum of its constituent individuals, a community is more than a sum of its constituent populations. It is the sum plus the interaction between these pop- ulations. We will therefore revisit competition, herbivory, predation, para- sitism, mutualism and scavenging. However, in previous chapters the focus was on interactions between species that occupied the same trophic level (interspecific competition), or between members of adjacent trophic (feed- ing) levels (herbivore–plant, carnivore–herbivore, parasite–host). These are direct effects, as opposed to the cascading, indirect, unexpected, effects that become apparent when communities of interacting species are examined. Sih et al. (1985) looked at 100 experimental studies of predation and found that approximately one-third of them showed unexpected effects. In addi- tion, when we examine communities, new, emergent, properties and topics arise such as food web dynamics, species diversity and the ‘rules’ for assem- bling communities. Ecology at the community level is difficult because the database involved is enormous and can become unmanageable. Ecologists have attempted to circumvent this problem by studying energy flow through trophic groups, by using summaries of species diversity such as rank abun- dance models, and by concentrating their studies on only a subgroup of the community, such as the herbivore community. Of course, communities have no scale except for the one that we humans impose upon them. The ‘local savannah community’ looks quite different to a rinderpest virus, a rumen bacterium, an acacia-ant, a migrating wildebeest, or an elephant. We humans tend to see savannahs as large areas of grassland with Acacia trees, and large herds of mammalian herbivores and carnivores. It is mostly at this scale that we view savannahs in this chapter. An exception The Biology of African Savannahs. Second Edition. Bryan Shorrocks & William Bates. © Bryan Shorrocks & William Bates 2015. Published 2015 by Oxford University Press. 238 THE BIOLOGY OF AFRICAN SAVANNAHS is the first topic, energy flow, because here microbial decomposers are most important. Surprisingly, most of the energy passing through a savannah community passes through the detritivore system, not the herbivore system. One final point. Many ecologists use the term ‘community’ only for the biotic components of a system, using the term ‘ecosystem’ for the biotic plus abiotic components. However, we agree with Begon et al. (1996) that in practice the two terms are interchangeable, since it is impossible to ignore abiotic aspects when considering species interactions—e.g. rain, grass and herbivores. Energy flow and food webs In an attempt to find simple ways to describe and study complex communi- ties, many early ecologists adopted a dynamic trophic approach. Linderman (1942) attempted to quantify the idea of food chains and food webs by con- sidering the transfer of energy between trophic levels. This theme continues to the present day, with ecosystem productivity forming a major catalyst for the International Biological Programme (IBP) and the Global Change and Terrestrial Ecosystems (GCTE) component of the Geosphere—Biosphere Programme (IGBP) (Steffan et al. 1992). At the bottom of the food chain are the primary producers: green plants and bacteria. These autotrophs alone are responsible for primary production, by capturing the incident solar radiation from the sun. This primary productivi- ty of a community is the rate at which biomass is produced by plants, per unit area. It is conventionally expressed either as energy (e.g. J m−2 day−1) or dry organic matter (e.g. kg ha−1 year−1). The total fixation of energy by photosyn- thesis is called gross primary production (GPP) and GPP minus the energy lost through respiration (R) is known as net primary production (NPP). So NPP = GPP − R. Savannahs have a range of NPP from 20 to 2000 g m−2 per year, with a mean of about 900 g m−2 per year. Table 6.1 shows NPP for a range of biomes (along with their total world NPP), illustrating that savan- nahs have an intermediate productivity. Notice however, that below-ground productivity (roots, etc.) is almost always underestimated. Also notice that the productivity to biomass ratio is low in forests because a large part of the biomass is dead. Several factors limit primary production. First of all, terrestrial communi- ties use solar radiation inefficiently. Between 0 and 5 J of solar energy strikes each 1 m2 of the Earth’s surface every minute. If all this radiation was used by green plants there would be a prodigious production of plant material. This is not the case because, first, only about 44% of the incident short-wave radia- tion occurs at wavelengths suitable for photosynthesis. Secondly, even this suitable radiation is not used efficiently. Photosynthetic efficiency is defined The Savannah Community and Its Conservation 239 Table 6.1 Annual net primary production (NPP) and biomass estimates for a selection of the world’s biomes. The high forest biomass is a product of accumulated dead biomass (wood) Biome NPP (g m−2) World NPP (109 t) Biomass (kg m−2) Tropical forest 2200 37.4 45 Temperate deciduous forest 1200 8.4 30 Savannah 900 13.5 4 Boreal forest 800 9.6 20 Cultivated land 650 9.1 1 Temperate grassland 600 5.4 1.6 Desert and semi-desert 90 1.6 0.7 From Whittaker (1975). as the percentage of incoming photosynthetically active radiation (PAR) incorporated into above-ground NPP. It varies from 0.01% to 3% (Cooper 1975). Figure 6.1 shows a flow diagram depicting the path of solar radiation at Nylsvley, a woodland (Burkea/Combretum) savannah in southern Africa (Scholes and Walker 1993). Notice that only a very small amount of the solar radiation is available for photosynthesis. Because of cloud cover, Nylsvley receives, on average, only 75% of its potential annual total of 4,371 sunshine hours. The total annual solar radiation at canopy level is about 7,316 MJ −2m , which is 61% of the radiation received above the atmosphere. In addition to this inefficient use of solar radiation, other factors limit NPP. Foremost among these is water. In arid regions, which include many African 11915 Reflected, Above the absorbed and atmosphere reradiated Long-wave Evaporation radiation 914 7316 1427 2270 Reflected by Contacting plants and soil the ecosystem 6400 Absorbed by 4964 plants and soil Heat 2694 Convection 11 Photosynthesis Figure 6.1 Flow diagram showing the fate of annual solar radiation in a woodland savannah in southern Africa. Numbers are MJ m−2 (after Scholes and Walker 1993). 240 THE BIOLOGY OF AFRICAN SAVANNAHS savannahs, there is an approximately linear increase in NPP with increase in precipitation. Figures 2.1 and 2.2 are an obvious example from African savannahs. A further limiting factor is potential evaporation (PET). This is an index of the theoretical maximum rate at which water might evaporate into the atmosphere (mm per year) given the prevailing radiation, aver- age vapour pressure deficit in the air, wind speed and temperature. The latter factors all affect transpiration of water from the leaf. Therefore (PET − precipitation) provides a crude index of ‘drought’ or how far the water avail- ability for plant growth falls below what might be transpired by actively growing vegetation. Its relationship with above-ground NPP is shown in Figure 6.2 for a variety of community types in temperate North America, with African savannahs added for comparison. Water shortage affects plant productivity, leading to less dense vegetation. This in turn exposes more bare ground, leading to a wastage of incoming solar radiation. In fact, this wast- age of solar energy striking bare ground is one of the major causes of low productivity per area in arid areas (Table 6.1). This becomes clear when pro- ductivity per unit of leaf biomass, rather than productivity per unit of area, is calculated. For deciduous forest this is 2.22 g g−1 year−1, compared to an equal productivity of 2.33 g g−1 year−1 for deserts. Grassland have a productivity per unit of leaf equal to 1.21 g g−1 year−1. 1,400 1,200 1,000 per year) –2 800 600 400 Above ground net primary productivity (g m 200 0 –2,000 –1,000 0 1,000 2,000 PET - PPT (mm per year) Increasing water deficit Figure 6.2 Relationship between above-ground productivity and PET-precipitation for a variety of community types. Circles are for North American sites: black, woodlands; white, desert; grey, grassland. Squares denotes African savannahs: black, two Burkea sites in South Africa, white, three Loudetia sites in Lampto, West Africa; grey, two sites in Kruger, South Africa (data from Webb et al. 1983, Hartley and Walker 1982, and Du Toit et al. 2003). The Savannah Community and Its Conservation 241 Energy flow Primary production is only the start of the flow of energy through the com- munity. After the producers (green plants) come consumers and decom- posers. The former are further divided into primary consumers (e.g. grasshoppers, baboons, wildebeests), and secondary consumers (baboons, spotted hyaenas, lions). In fact these food chains can originate in either live plants (green grass, wildebeest, lion) or dead plants and animals (wood, ter- mite, bat-eared fox; carcass, hyaena; carcass, vulture, serval). A simplified view of this further complication is shown in Figure 6.3 for a generalised African savannah system, along with the energy lost through respiration and the part played by microrganisms in this trophic network. Notice that most of the primary production does not go through the herbivore system but through the scavenger and decomposer system (right-hand side of Fig- ure 6.3). This is because the faeces and dead bodies are lost to the grazer sys- tem (and enter the decomposer system), while the faeces and dead bodies of the decomposer system are simply returned to the dead organic matter box at the base (Figure 6.3). Figure 6.4 shows some actual data for the Nylsvley ecosystem in southern Africa (Scholes and Walker 1993). Figure 6.4a again shows that most above- ground primary production is consumed by decomposers and fire, with only a very small component finding its way into herbivores. Figure 6.4b shows R R C2 C2 R R C1 C1 Bodies and faeces Bodies and faeces Dead organic NPP matter Figure 6.3 A generalised model of savannah trophic structure. NNP = net primary production, R = respiration, C1 = consumer level 1, C2 = consumer level 2. On the NPP side of the diagram, C1 could be grasshopper, baboon, or wildebeest, and C2 could be baboon again (they are omnivorous), spotted hyaena, or lion. On the dead organic matter side of the diagram, C1 could be termite, spotted hyaena again (because they make kills and scavenge), vultures, or microrganisms, and C2 could be aardwolf, bat-eared fox (both take termites), serval (which is known to take vultures from time to time), or microrganisms again. 242 THE BIOLOGY OF AFRICAN SAVANNAHS (a) Grazers 18–30 Browsers c. 6 Herbivory –2 –1 Grasshoppers c. 13 46–61 g m y Caterpillars 9–12 Above ground primary production c. 470 g DM m–2 y–1 Decomposers and fires c. 410 g m–2 y–1 (b) 56.6 R 26.0 Grasshoppers Impala R A 6.0 A 100 138 100 I P I 2.4 P E E 41.0 68.0 38 Wastage Figure 6.4 Secondary consumption in Nylsvley, a wooded savannah in southern Africa. (a) Pathways showing the disappearance of plant material. (b) Partitioning of energy consumption by two herbivores into intake (I), excretion (E), production (P), and respiration (R). The fraction not excreted is assimilated (A) by the digestive system (after Scholes and Walker 1993). the fate of this consumed energy for an invertebrate and a vertebrate herbi- vore. Notice that most of the energy taken in (I) is channelled into excretion (E) or respiration (R). Only a small proportion goes into production (P). In the case of the grasshopper, a significant proportion of the material removed from the plant is not ingested, but wasted, because it falls to the ground dur- ing the feeding process. As organic matter is consumed, energy is transferred to the next trophic level. However, this transfer is inefficient, leading to the familiar pyramid of numbers within a community. The biomass of herbivores is much less than that of the vegetation that supports it, and the biomass of carnivores even less than that of the herbivores. Indeed, this is one of the reasons frequently put forward to explain the length of food chains. There simply are not enough lions to support another predator level above them (if you exclude disease). The Savannah Community and Its Conservation 243 There are in fact three transfer efficiencies between primary production and the secondary consumers above. These are (1) consumption efficien- cy, (2) assimilation efficiency, and (3) production efficiency. Consumption efficiency is the percentage of the total available productivity consumed (ingested) by the next trophic level. For example, the percentage of NPP finding its way into the guts of herbivores. In general this is low, between about 0.5–20%. Assimilation efficiency is the percentage of food taken into the gut of consumers that is assimilated across the gut wall. In general this is about 20–50% for herbivores and 80% for carnivores. Production effi- ciency is the percentage of assimilated energy that is incorporated into new biomass. For invertebrates this is about 30–40%, for ectotherm vertebrates (reptiles and amphibians) it is about 10% and for endotherm vertebrates (birds and mammals) about 1–2%. The much lower rate for endotherms is because they expend energy in regulating their body temperature. This sug- gests that the average efficiency of a mammalian herbivore, such as a wilde- beest, is approximately 0.1 × 0.35 × 0.015 = 0.05 = 5% of grass production. By the time this 5% of NPP has become incorporated into new lion biomass it will only be about 0.006% of NPP. No wonder the pyramid of numbers decreases so rapidly. Most terrestrial communities described so far typically have food chains of length three or four (Begon et al. 1996), and this is also true of African savannahs. Almost all the mammalian food chains are three links in length (e.g. grass, wildebeest, lion or grass, termite, aardwolf). Many avian food chains have four links (grass, grasshopper, insect-eating bird, raptor). Of course energy flow graphs like those of Figure 6.4 are annual averages. Incoming solar energy varies throughout the year (even in the tropics) because of cloud cover, and NPP varies because of factors such as seasonal water availability. Not only does production vary, but consumption does also. Owen-Smith (1982) suggested that grazing large mammals face a period of protein shortage during the dry season (see also Chapter 4) while browsers have a period of energy shortage at this time. Figure 6.5 shows that kudu at Nylsvley (Scholes and Walker 1993) are able to satisfy their protein require- ments at all times of the year, but that they are unable to meet their energy requirements for a 2 month period in late winter. This is because the forage consumed by browsers has a high protein content relative to grass, even in the dry season. However, because many browse species are deciduous, forage is scarce at this time. Food webs Since the early 1970s there has been an increasing interest among commu- nity ecologists in the dynamics of food webs (May 1972, 1973, Pimm 1991). This interest has included several topics, including the relationship between complexity and stability, the importance of ‘connectance’ (the fraction of 244 THE BIOLOGY OF AFRICAN SAVANNAHS Kudu Cattle 1.6 ) –1 1.2 d –0.75 0.8 0.4 (MJ Kg Daily energy intake 1.6 ) 1.2 –1 d 8 –0.75 (g Kg 4 Daily protein intake DJ FMAAM JJ SOND DJF M AAM JJ SOND Figure 6.5 Seasonal intake of protein and energy in a browser (kudu) and a grazer (cattle) (from Scholes and Walker 1993). all possible pairs of species that interact directly), the length of food chains, and the possibility that food webs are compartmentalised. Most of these topics involve the idea that food webs (community structure) are in some way constrained to be the way they are. We have already referred to the con- straint of energy transfer efficiency between trophic levels, which leads to the pyramid of numbers and constrains the length of food chains. These dis- cussions depend critically on the quality of data that are available across a range of communities, and to date this quality of information is not available for savannah communities. However, it is well worth mentioning the field experiments of McNaughton (1977, 1985). One of the food web topics that has provided considerable scope for debate is that of complexity and stability. The conventional wisdom was that more complex communities were more stable, and MacArthur (1955) and Elton (1958) brought together a variety of observations to support this idea. Unfor- tunately, neither ‘complexity’ (= more species, more interactions between species, greater average interaction) or ‘stability’ was rigorously defined. What is more, the original observations of MacArthur and Elton, although consistent with the idea that more complex communities are more stable, could also be seen in terms of other reasonable explanations (Begon et al. 1996). This issue was made more contentious when May (1972) introduced some mathematical descriptions (models) of food webs in which more com- plex webs were actually less stable. The Savannah Community and Its Conservation 245 In the Serengeti National Park, McNaughton (1977) chose two grassland areas that differed in the diversity of their plant communities. One was species-poor and one was species-rich. Within each area, African buffalo were allowed to graze in some plots and were excluded from others. The results are shown in Table 6.2. The more diverse plant community was subject to a greater diversity modification by the grazing buffalo. How- ever, this greater diversity impact was translated into a reduced effect on a functional property—green biomass. Although the amounts eaten were similar, growth of uneaten species in the more diverse community com- pensated for the greater consumption. Compensation by the community for grazing off-take was 83% in the species-rich community but only 9% in the species-poor community. McNaughton believes that the buffalo study provides confirmation of the idea that plant community diversity stabilises functional properties of the community to environmental perturbations. In fact the adjustments in species abundances in the more diverse commu- nity (resulting in greater diversity (H) changes) mediate functional stability. In other words, more diverse communities have a greater array of species that are able to react and compensate for what other species are doing. In 1985, McNaughton reported a more extensive study of the effects of graz- ing by zebra, buffalo, wildebeest, and Thomson’s gazelle on grassland in the Serengeti. He considered what he called biomass resistance, as a measure of community stability. This was defined as the proportion of the above- ground vegetation not eaten during either a single passage of the herd in the dry season (when wildebeest was the main grazer), or a 12 day period in the wet season (all species). The results are show in Figure 6.6. Notice that Table 6.2 The influence on biomass of grazing by buffalo (% of control) and species diversity in two areas differing in plant diversity Species diversity Not grazed Grazed Statistical significance Biomass Species-poor plot 75.9 (69.3) Species-rich plot 66.9 (11.3) Statistical significance NS (P < 0.005) Diversity Species-poor plot 1.069 1.357 NS Species-rich plot 1.783 1.302 P < 0.005 Statistical significance P < 0.005 NS NS, not significant. Species diversity is measured using the Shannon index of diversity (H) = −Σpi ln pi, where pi is the frequency of the ith species, and ln is the natural logarithm. This index takes into account the number of species, and their frequency. The latter is deemed important because a community with, say, 10 equally abundant species is considered more diverse than one with 10 species, one of which accounts for 91% and the other nine 1% each. H = 0 when only one species is present and Hmax = ln S, where S is the number of species. From McNaughton (1977). 246 THE BIOLOGY OF AFRICAN SAVANNAHS (a) (b) 80 80 60 60 40 40 20 uneaten 20 Proportion of vegetation Proportion of vegetation uneaten 0 0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 Plant diversity Plant diversity Figure 6.6 Resistance of vegetation to grazing in the Serengeti National Park, Tanzania. (a) Dry-season data (P < 0.05), (b) wet-season data (P < 0.001). Zebra (○), buffalo (□), wildebeest (●), and Thomson’s gazelle (Δ) (from McNaughton 1985). in the dry season (Figure 6.6a) wildebeest had a much greater impact on the vegetation. In both experiments grazing had a much greater impact (less vegetation uneaten) on areas of low diversity, suggesting that community stability is greater with higher diversity. The complexity and stability debate is a critical one because if more complex communities (more species, more species diversity) are more stable (persist longer, resist disturbance better) it is important to try to maintain this complexity if we wish to see savannah ecosystems survive. Another food web issue, already alluded to, involves cascading, indirect, interactions and their effects. Until we have more detailed information on savannah food webs, with interaction strengths between all species, and descriptive models to simulate the cascading interactions, we cannot be precise about these indirect interactions. Of course, at a level much lower than a complete food web, the model of Dublin et al. (1990a), referred to in Chapter 5, was an attempt to investigate these more complex multispecies interactions. However, to see something of the potential issues involved and the likely consequences of indirect interactions, we can consider Figure 6.7. This is a simplified food web for the Serengeti–Mara ecosystem taken from several sources, but mainly from Sinclair and Norton-Griffiths (1979) and Sinclair and Arcese (1995). No strengths of interaction are indicated, only their positive or negative effects, and many species are absent. However, as an example of an indirect interaction, consider what is implied if rinderpest is reduced or eliminated from the system. With the removal of rinderpest, wil- debeest numbers will increase, green grass will decrease, and consequently there will be less dry grass to burn. This will result in fewer fires (or less extensive/intense fires), and therefore greater survival of young trees. More The Savannah Community and Its Conservation 247 Negative effect on target Positive effect on target lion humans hyaena leopard wild dog rinderpest cheetah impala Grant’s and eland Thomson’s buffalo wildebeest gazelle topi elephant warthog zebra giraffe herbs green grass immature mature trees trees grasshoppers rain fire termites dry grass Figure 6.7 Simplified trophic web of the Serengeti–Mara ecosystem. Shading indicates physical factors, plants, herbivores, carnivores. young trees will mean more mature trees, which should have a positive effect on the giraffe population. The implication of the trophic networks in Figure 6.7 is that many years after the removal of rinderpest there may (other things being equal) be an increase in numbers of giraffe. In fact, such an increase was reported by Grimsdell (1979). In one area of the Serengeti estimates of giraffe numbers increased from 2,780 ± 634 (1971) to 4,970 ± 916 (1976). Pellow (1977) also had evi- dence of increasing giraffe numbers in the central area of the Serengeti. More recently, Holdo et al. (2009) have presented a Bayesian state-space model of this cascade, showing complex interactions between rinderpest, wildebeest, grass, tree cover, fire, rainfall, elephants and humans in the Serengeti eco- system. The eradication of rinderpest caused the wildebeest population to increase and the occurrence of fires to decrease. This resulted in the recovery of tree cover and shifted the amount of carbon stored in biomass and in the soil. They estimated that tree carbon in 1999 was 997 Mg C km−2 and then used this value along with other data to begin to estimate annual fluctuations in Serengeti ecosystem carbon. Their research clearly shows how a disease that affects herbivores, which in turn affects the tree:grass ratio, is linked to global warming by changing atmospheric carbon dioxide levels. These results clearly indicate that the protection of savannahs contributes to the overall health of our planet. 248 THE BIOLOGY OF AFRICAN SAVANNAHS In fact a start has been made to look at some of these interactions within the Serengeti–Mara ecosystem. In the first Serengeti book (Sinclair and Norton-Griffiths 1979), Hilborn and Sinclair (1979) constructed a model of the Serengeti ecosystem. This model focused on the interaction of rainfall, grass, wildebeest, and two predators, lion and spotted hyaena. The rainfall– grass–wildebeest component was a less sophisticated forerunner of their later model, already reported in Chapter 4 (Mduma, Hilborn and Sinclair 1998). To this basic single-species model they added predation. To simulate the interaction between predator and prey they used Holling’s disc equation (Chapter 4, Equation 3, Figure 4.16) modified into its multiprey form (Char- nov 1973). Data on total quantity of prey consumed in a year came from Schaller (1972) and prey preference came from work in Nairobi National Park (Foster and Kearney 1967, Foster and McLaughlin 1968) for wilde- beest, zebra, impala and hartebeest. For predator cub survival they used a linear relationship with food availability. Some of their results are shown in Figure 6.8. The two ‘rate of increase’ curves are for different amounts of annual rainfall and show quite clearly the effect of rain, via grass, on wilde- beest population growth (see also Chapter 4). In 1995 Hilborn et al. com- mented that the ‘ . . . model did predict quite accurately how the wildebeest would respond to the rainfall regime that has occurred in the 14 years since the model was constructed’. Where the curves cross the Nt +1/Nt = 0 line at high wildebeest density we see the effects of food limitation (Chapter 4). The curves crossing the zero line at low density is probably a result of predation. If this is a correct description of real events, it suggests that predators (lions and hyaenas) may well have a crucial effect on wildebeest population size, at very low density. At an ecological workshop held at the Serengeti Wildlife Research Centre in December 1991, Ray Hilborn and 25 other researchers spent 4 days trying to 1.1 250 mm N t + 1 150 mm 1.0 Nt 0.9 01234 567 Wildebeest population size (millions) Figure 6.8 Rate of change of wildebeest population [(Nt+1)/Nt] tplotted against wildebeest popu lation size for two dry-season rainfall regimes (from Hilborn and Sinclair 1979). The Savannah Community and Its Conservation 249 construct a second model of the Serengeti ecosystem (Hilborn et al. 1995). This model was built primarily for research coordination and policy evalua- tion. While some of the issues it addressed were clearly biological in nature (e.g. herbivore and carnivore population dynamics, species loss, vegetation change), some were more management orientated (e.g. visitor capacity, hunt- ing and poaching, economics and cash flow). The spatial scale of the model divided the ecosystem into 10 areas (4 of which include the Serengeti Nation- al Park, 1 the Masai Mara, and 5 the areas surrounding the Serengeti). The temporal scale divided the year into an 8 month wet season and a 4 month dry season, although some parts of the model did not use this intra-annual scale (e.g. human population growth). The model was subdivided into five submodels: vegetation, ungulates, predators, inside park, and outside park. Simulations started in 1960 and continued until 2020. • The vegetation submodel used dry-season and wet-season rainfall to predict dry-season old grass and new dry-season green grass. These were reduced by the percentage of each of the 10 areas that was burnt, under cultivation or woodland. • In the ungulate submodel six ‘species’ were described: wildebeest, zebra, Thomson’s gazelle, elephant, buffalo and ‘brown animals’ (topi, impala and hartebeest). For each ‘species’, the number of animals born was simply made proportional to the population size, i.e. Nt × birth rate. The popula- tion size each year (Nt +1) was made equal to that in the previous year (Nt) × dry-season survival + births − hunter kills − disease − predator kills. The key dynamic factor is dry-season survival, which was assumed to be related to the amount of dry-season grass per individual. • The predator submodel included both calculation of the kill rate of prey items and the population dynamics of the predators. Kill rate was once again modelled by a disc equation in which the variables are predator attack success and prey handling time for each predator (lion, hyaena, cheetah, leopard and wild dog). Population dynamics were simply Nt +1 = Nt × (1 + birth − mortality − number poached), with both birth and mor- tality related to predator density. • The inside park submodel had two major components, tourism quality/ growth and park revenue. Tourism quality goes up as more animals are seen, and down as more tourists are present. Animals seen was not simply a product of numbers. For example, predator numbers were weighted with regard to their ‘tourist value’, with lions considered to have half that of cheetahs and leopards, and hyaenas as zero! Tourist growth rate was relat- ed to tourism quality and the general growth rate of the tourism industry. Revenue was related to the number of tourists × the park fee ($15.00) + the fee from overnight stays in hotels ($5.00), with 50–75% of the revenue allo- cated to the park’s operating revenue and the rest going to the Tanzanian government. The anti-poaching budget (APB) of the park was a percent- age of the park’s operating revenue. 250 THE BIOLOGY OF AFRICAN SAVANNAHS • The outside park submodel had three major components: poacher effort and kill, human population growth (HPG), and changes in land use. Poaching was assumed to increase with HPG and decrease with APB. The number of kills was further influenced by species vulnerability. HPG was assumed to have a constant rate and land in wilderness was assumed to decrease in proportion to HPG. Figure 6.9 shows some of the biological output from the model. The dra- matic effect that more or less anti-poaching has on the system is clearly seen, emphasising the important role of tourism in maintaining the park budget (see later in this chapter). Of course, one problem with community models such as this is that much detail is omitted. Factors put into the model are regarded by the modellers as ‘important’. For example, grass production was assumed to be largely determined by rainfall and only dry-season mortality of ungulates is included. If these prior decisions are incorrect the model will be misleading. Nonetheless, there must be more attempts to describe these savannah communities by modelling. Not all species within a community have equal roles to play in its dynamics. This would be suspected just from an examination of the frequency distribu- tion of abundances, even without a knowledge of the species interactions. In most communities, a few species are very abundant, while many species are much less common and even rare (Magurran 2004). Savannah communi- ties are no exception (Figure 6.10). However, some species have an impact on community dynamics far in excess of their numerical abundance and Paine (1966, 1980) called these ‘keystone’ species. Sinclair and Byrom (2006) state that wildebeest have such a keystone role in the Serengeti ecosystem, 1,000 1,000 wildebeest 1,000 wildebeest 800 800 800 wildebeest 600 600 600 400 400 400 zebra zebra zebra 200 200 200 0 0 0 1960 1980 2000 2020 1960 1980 2000 2020 1960 1980 2000 2020 10,000 10,000 10,000 hyenas hyenas hyenas 8,000 8,000 8,000 6,000 6,000 6,000 4,000 4,000 cheetah 4,000 cheetah 2,000 2,000 2,000 cheetah lions lions 0 0 0 lions 1960 1980 2000 2020 1960 1980 2000 2020 1960 1980 2000 2020 Figure 6.9 Output Serengeti model. (a) 1991 values for input variables; (b) increased anti-poaching budget; (c) greatly reduced anti-poaching budget. Histograms are rainfall, both actual and predicted (from Hilborn et al. 1995). The Savannah Community and Its Conservation 251 Serengeti 18,00,000 15,00,000 12,00,000 9,00,000 Abundance 6,00,000 3,00,000 0 topi oryx zebra eland rthog roan oribi impala buffalo giraffe kongoni wa elephant steinbok waterbuck wildebeest klipspringer Grant's gazelle Thomson's gazelle Species 25,000 Tarangire 20,000 15,000 10,000 Abundance 5,000 0 oryx zebra eland impala giraffe buffalo kongoni elephant warthog bushbuck reedbuck wildebeest Grant's gazelle Thomson's gazelle Species Figure 6.10 Rank abundance of ungulates in the Serengeti ecosystem in 1991 (top) and the Tarangire ecosystem in 1990 (bottom). although it is not clear that this is in excess of their numerical abundance. However, they certainly do play a pivotal role in the dynamics of the large- mammal community, as Figure 6.7 implies. In addition, recent work has detected the impact of wildebeest on less obvious components of the ecosys- tem. By their impact on both the physical structure and species composition of the grasses and herbs of the Serengeti plains, grazed areas can support a 252 THE BIOLOGY OF AFRICAN SAVANNAHS density of butterflies two orders of magnitude greater than ungrazed areas (Sinclair and Byrom 2006). With grasshoppers the effect is negative, prob- ably because both are feeding on grasses. In ungrazed areas, 49 species of grasshopper have been recorded, while in grazed areas it can be as low as only one species. Wildebeest may even affect the bird community. Their grazing, through its effect upon grass composition and height, may affect both feed- ing and nest sites. Of the eight commonest bird species feeding in the grass layer of grazed sites, seven showed reductions of 50–80% in ungrazed sites (Sinclair and Byrom 2006). Figure 6.7 may already look complex, but all the potential unexpected, cascading, effects are certainly not present. A similar keystone role has been attributed to elephants (e.g. Laws 1970, Waithaka 1996). They have the ability to greatly modify the habitat, particu- larly in their destruction of trees, shrubs and saplings. We saw in Chapter 5 (Figure 5.8) that, in association with fire, they have the ability to convert woodland savannah into grassland savannah. In addition, studies carried out in the Aberdares, in Kenya (Waithaka 1996), have shown that elephants create and expand gaps in forests and in the process open up a more produc- tive and diversified ground layer that is subsequently exploited by a wide range of other animals. This elephant effect on savannah and forest vegeta- tion was particularly evident in many protected areas in East Africa in the 1980s (Chapter 5) when elephant numbers averaged five times those outside parks. Elephants are also a key species in maintaining and promoting the tourism industry. In a survey conducted in Kenya in the late 1980s, tourists rated the elephant the most important species. These kinds of community issues will be examined later in this chapter, when the impact of tourism and agriculture are examined. Assembly rules In Chapter 5, when we talked about interspecific competition, we comment- ed that many ecologists searched for community patterns in order to uncover the action of competition. Niche differences, discussed in Chapter 5, are one such pattern. Now we will examine some other patterns. The idea that communities have assembly rules (they are not just random assemblages), and therefore show patterns in the type of species they con- tain, has a long history in ecology. If interspecific competition is important in structuring communities, niches should be more spaced out along a resource dimension than expected by chance. Niche differences, particularly those associated with diet, are frequently manifested as morphological differences. The overdispersion of niches should therefore reveal itself as an overdisper- sion of morphological measurements. Such morphology might include bill size in birds and body size in savannah ungulates. Hutchinson (1959) collected The Savannah Community and Its Conservation 253 a number of examples, from both vertebrates and invertebrates, of sequences of potential competitors in which individuals from adjacent species had body weight ratios of approximately 2. The implication is that the larger species should be about twice as big in order to ‘escape’ interspecific competition. As we saw in Chapter 5 (Figure 5.25) there is a relationship between ener- gy turnover and body size (energy turnover = K + W 0.75). Given a series of animals, such as ruminants, with comparable digestive systems, the small- er animals require a diet with a higher proportion of protein and soluble carbohydrate at the expense of fibre. There is in fact a negative relationship between body mass and the percentage of non-stem in the diet of grazers (Owen-Smith and Cumming 1993) (Figure 6.11). Larger species bulk-feed on whole plants, while smaller species selectively graze on the more nutri- tious leaves and shoots. Because grasses tend to employ structural defence mechanisms against herbivores (Chapter 2), while dicotyledons (herbs and trees) tend to employ unpalatable secondary metabolites, this influence of diet on body size extends to browsers compared to grazers. To process grass a grazer requires not only suitably resistant teeth but also a large rumen (Chap- ter 3). The rumens of grazers therefore tend to be larger than those of brows- ers (Hofmann 1968, 1973) and this translates into the frequently observed body size difference between browsers and grazers (Figure 6.12). Given this close relationship between diet and size in savannah ungulates, it would seem that they are a likely candidate for Hutchinson’s ×2 assembly rule. Prins and Olff (1998) examined 95 African grazers with a body weight great- er than 2 kg. Most of these were mammals, but they also included two species of geese (Egyptian and spur-winged) and the leopard tortoise (Geochelone pardalis). Figure 6.13 shows their results for their total African data and for a 100 90 80 70 60 % Non-stem in diet 50 40 243 567 In weight (kg) Figure 6.11 Relationship between body weight and percentage of non-stem grass in the diet of grazing ungulates (data from Owen-Smith and Cumming 1993). 254 THE BIOLOGY OF AFRICAN SAVANNAHS 25 20 15 Number 10 5 0 6 12.5 25 50 100 200 400 Metabolic weight Figure 6.12 Frequency distribution of body sizes for African grazing (dark bars) and browsing (white bars) ungulates. Metabolic weight is kg0.75 (data from Owen-Smith and Cumming 1993). Serengeti subgroup. If sequential species obey Hutchinson’s ×2 rule and the ranked data are plotted against the natural logarithm of body weight, the line should be linear and the slope should be ln 2 = 0.693. Prins and Olff (1998) also calculated the expected line of rank versus natural log of body weight, if each body weight had the same probability of occurring (the null model if species do not interact). They argue that if every body mass, between the minimum (Wmin) and the maximum (Wmax) had an equal probability of occurring, then a non-linear line results. Weight (W) would have a uniform 10 8 6 4 In body weight (kg) 2 0 0 20 40 60 80 100 species rank Figure 6.13 Ranked ln body mass of grazers (Wi) plotted against rank number (Ri) for the whole of Africa (95 species, (circles and thick line) and for the Serengeti ecosystem (33 species, squares and thin line) (from Prins and Olff 1998). The Savannah Community and Its Conservation 255 distribution and the expected difference between consecutive species (the weight ratio, WR) would be ((Wmax − Wmin)/SR) where SR is the number of species in the assemblage. The expected value of the body mass of the ith spe- cies (Wi) with rank Ri would be W i = Wmin + (R −1)/ [(Wmax − Wmin)/(SR −1)] or, taking natural logarithms, ln(W i ) = ln{[Wmin + (R −1)]/ [(Wmax − Wmin)/(SR −1)]}. This ln(Wi) will be a non-linear function of Ri. This ‘neutral’ line, indicat- ing no competitive spacing, is also shown on Figure 6.13. Notice that the weight ratio, in combination with the range of body weights (Wmax − Wmin) will therefore determine how many species are found. That is, species rich- ness SR = (Wmax − Wmin)/WR. If the smallest and largest species are rather constant (as they are in most African savannahs) this means that the weight ratio and species richness will tend to be inversely related. The results in Figure 6.13 suggest that for the whole of Africa data, the weight ratio is only 1.066 and not the 2 suggested by Hutchinson (1959). However, the line is linear, suggesting a regular dispersion of species along this feeding axis. Of course we might expect that species taken from the whole continent of Africa might be more closely packed than expected. Competition would not push apart species that lived in separate savannah ecosystems, in differ- ent parts of Africa. In this connection the Serengeti data in Figure 6.13 is interesting. Again the ranked body weight line is linear, suggesting an even placement of species, but the weight ratio is now 1.21, suggesting that in a smaller area species might have to be more ecologically separated. Prins and Olff (1998) have pointed out that the total African data in Figure 6.13 is also interesting because there appears to be a gap between buffalo (632 kg) and the last four largest species (white rhinoceros 1875 kg, hippopotamus 1900 kg, forest elephant 2575 kg and savannah elephant 3550 kg) (Chapter 3), often referred to as megaherbivores (Owen-Smith 1988). Does this gap in the community pattern of grazing herbivores suggest that there are ‘miss- ing’ species that could fit into this grazing resource axis? From Figure 6.13 the number of these missing species can be estimated to be between 6 and 10 (Prins and Olff 1998). It is intriguing to speculate that the missing very large herbivores could be several species that are now extinct. These include the proboscid Deinotherium bozasi (in the early Pleistocene) and four ‘ele- phants’ (Elephas recki, E. iolensis, E. zulu and Loxodonta atlantica) which went extinct between 1,000,000 and 400,000 years ago. Other megagrazers that went extinct during the Pleistocene were Hippopotamus gorgops, a giant hartebeest (Megalotrogus priscus) and the giant buffalo Pelorovis( antiquus) (Klein 1988, Owen-Smith 1988). The predicted number of missing species (6–10) therefore tallies quite well with the number of recently extinct species (8) (Prins and Olff 1998). 256 THE BIOLOGY OF AFRICAN SAVANNAHS Arusha National Park Serengeti National Park Mkomazi Game Reserve Lake Manyara National Park Tarangire National Park Moyowosi-Kigosi Game Reserve Burigi-Biharamulo Game Reserve Katavi National Park Ugalla Game Reserve Selous Game Reserve Ruaha National Park 1234567 in body mass (kg) Figure 6.14 Body weight of Tanzanian bovids plotted along a ln(weight) axis, with diet indicated as grazer (white), browser (black), or mixed feeder (grey). Each point represents one species. Where two species have identical weights, one is indicated off the line. We end this examination of savannah assembly rules by presenting some unpublished data that Fiona Wragg and one of us (BS) have collated from Tanzanian protected areas (Wragg 2002). This data includes all the Tanza- nian grazing and browsing bovids, and it is displayed, in Figure 6.14, along a weight axis for those reserves with savannah habitats. Arusha National The Savannah Community and Its Conservation 257 Table 6.3 Weight ratios for 11 Tanzanian protected areas with savannah habitats Protected area Weight ratio Area (ha) Selous GR 1.30 5,233,000 Serengeti NP 1.27 3,425,100 Ruaha NP 1.29 2,595,000 Ugalla GR 1.36 850,000 Tarangire NP 1.26 680,000 Katavi NP 1.36 525,300 Burigi-Biharamulo GR 1.32 350,000 Mkomazi GR 1.39 250,000 Moyowosi-Kigosi GR 1.32 130,000 Arusha NP 1.80 13,700 Lake Manyara NP 1.45 10,833 GR, Game Reserve; NP, National Park. Park and Tarangire National Park are northern Tanzanian reserves, with ‘classic’ grass and shrub savannah. Moyowosi-Kigosi Game Reserve and Ugalla Game Reserve are western reserves with woodland (miombo) savan- nah. Selous Game Reserve and Ruaha National Park are central Tanzanian reserves, on the border between miombo and grass and shrub savannah (see Figure 1.9). Notice that as already stated (Figure 6.12), browsers are usually smaller than grazers. The weight ratios are shown in Table 6.3. All are again less than Hutchinson’s ×2 rule but show significant linear regressions of ln weight versus rank, implying that the body weights of these bovids are more regularly spaced than would be expected by chance. There do appear to be community patterns, suggesting assembly rules, for both grazing and brows- ing savannah herbivores. Island biogeography Larger islands contain more species. Although the specific relationship between number of species (S) and island area (A) can vary, it has tradition- ally been viewed as a power function: S = cAz where c is a constant giving the number of species when A has a value of 1 (often specific to a taxon), and z encapsulates the relationship between spe- cies number and area. By taking logarithms of both sides of this species–area equation, we get: log S = log c + z log A 258 THE BIOLOGY OF AFRICAN SAVANNAHS That is, on a double log plot (log S vs log A) we see a straight-line relationship between S and A, with z the slope of the line. But why is this interesting island relationship relevant to African savannahs? The species–area relationship has been found to apply to a wide variety of taxonomic groups (e.g. plants, insects, fish, birds and mammals) and to a wide variety of types of island, including ‘virtual’ islands (e.g. lakes, moun- taintops, and protected areas). Sadly, intact African savannahs are increas- ingly only found in protected areas. Some of these protected areas are now under threat. Over 200 km2 of Selous Game Reserve has been approved by the Tanzanian government for open-pit uranium mining that will pollute the air, water and soil. Selous Game Reserve was designated a UNESCO World Heritage Site in 1982, but UNESCO permitted the Tanzanian government to redraw the park boundary so that the mine would not be within the park in order to quell international outrage. No doubt more ‘islands’ will be threat- ened in the future. These national parks, game reserves and private reserves are ‘islands’ of relatively intact savannah fauna and flora in a sea of degraded habitat. What does island biogeography tell us about the numbers of species in these savannah ‘islands’? Before we consider African savannah islands specifically it will be useful to examine briefly the reasons that have been put forward for the island species– area relationship. There have been two main themes, not necessarily uncon- nected. The original explanation, put forward by MacArthur and Wilson (1967), is known as the dynamic theory of island biogeography. It proposed that species number on an island is a dynamic balance between two forces, immigration and extinction, and that these are in turn modified by island size and isolation. Figure 6.15 summarises the situation proposed by the dynamic theory. Two example immigration curves and two example extinc- tion curves are shown. When an island is empty of species any new arrival will contribute to the rate of immigration of new species. As species number on the island increases the number of new species arriving will fall, and in theory will reach zero when all the potential colonists have arrived. Islands that are near the source of the colonists (the mainland for real islands) and large islands (large targets) will have higher rates of immigration. Extinction rate will rise when there are more species on the island. This is thought to occur because with more species competitive exclusion becomes more likely and the population sizes will, on average, be lower. This latter will make spe- cies more susceptible to chance extinction. A similar reasoning would sug- gest that smaller islands would have higher rates of extinction. For example, the rate of ungulate extinction in Tanzanian reserves correlates significantly with reserve area (r = −0.93, P < 0.02)(Newmark 1996). Where the extinc- tion curve and the immigration curve cross gives the dynamic equilibrium species number for a particular island. The theory therefore explains why smaller islands have fewer species and suggests that more isolated islands will also have fewer species. The Savannah Community and Its Conservation 259 Immigration Extinction Near Small Rate Far Large Number of species Figure 6.15 MacArthur and Wilson’s (1967) equilibrium theory of island biogeography showing the rate of immigration and extinction plotted against the number of resident species on the island. The predicted equilibrium between immigration and extinction, for a large island and a near source of colonists, is shown by the ● symbol. The second major theme put forward to explain the species–area relationship is that of habitat diversity (Lack 1969, 1976). Essentially this states that lar- ger islands will have more habitats, and an area with more habitats will have more species. Of course this explanation could be incorporated within the dynamic theory because habitat diversity, linked to island size, could be one of the reasons for the different extinction curves in Figure 6.15. Protected savannah areas One of the earliest attempts to apply island biogeography ideas to savannah reserves was that of Miller and Harris (1977). They examined 13 East Afri- can savannah parks and looked at the relationship between area and species number. Their results are shown in Figure 6.16a. The fitted relationship is S = 35A0.02. The regression is not significantF ( = 0.17, P = 0.69). Soulé, Wilcox and Holtby (1979) published another set of data for 20 East African areas (Figure 6.16b) in which the species–area relationship was S = 18A0.13, and this time the regression was significant (F = 24.67, P = 0.0001). Soulé and colleagues make the point that without ‘active’ intervention, newly isolated communities such as national parks and other protected areas tend to show species loss. This is known as relaxation. In effect, the isolated area travels ‘down’ the species–area curve, from a higher point (larger area and more species), to a lower point (smaller area, fewer species). As protected areas become smaller and more isolated the rate of species loss increases because populations are smaller and therefore more prone to extinction (wild dogs 260 THE BIOLOGY OF AFRICAN SAVANNAHS (a) (b) 100 100 z = 0.13 z = 0.02 Number of species Number of species 10 10 100 1,000 10,000 1,00,000 1 10 100 1,000 10,000 1,00,000 Area (km2) Area (km2) (c) (d) 40 z = 0.03 30 z = 0.05 Number of species 10 Number of species 10 10 100 1,000 10,000 1,00000 10 100 1,000 10,000 1,00,000 Area (km2) Area (km2) (e) 40 10 z = 0.16 Number of species 1 1,000 10,000 1,00,000 10,00,000 1,00,00,000 Area (km2) Figure 6.16 Species–area graphs from five African savannah studies. Each graph shows the value of z for the data collected. (a) ‘Large mammals’ in 13 East African parks (Miller and Harris 1977); (b) ‘large mammals excluding insectivores, rodents, bats, and lagomorphs’ in 20 East African areas (Soule, Wilcox and Holtby 1979); (c) ‘ungulate species’ in 19 East African parks and reserves (Weston and Ssemakula 1981); (d) ‘large herbivores’ in 20 African reserves (East 1983); (e) ‘bovids’ in 14 Tanzanian ecological areas (Wragg 2002). The text in quotation marks indicates the species used, in the author’s own words. The Savannah Community and Its Conservation 261 in Chapter 5) and recolonisation is less likely because of intervening, unsuit- able, habitat. Of course the traditional theory of island biogeography sug- gests that the species going extinct will be lost at random, whereas in reality certain species are more prone to extinction. These might include, for exam- ple, species with larger territorial requirements, species with special habitat requirements, and perhaps species more susceptible to competition. Smaller islands will probably have fewer habitats and this will almost certainly be a major factor in the higher extinction rate found in smaller areas. Soulé and colleagues estimate that simply due to benign neglect many of the smaller African savannah reserves would lose up to 10–20% of their large-mammal species in 50 years following isolation. If this is true, it suggests that protected areas will either have to be enormous or that they will require active inter- vention to maintain an intact fauna. In response to the Soulé predictions, Weston and Ssemakula (1981) pub- lished their own dataset of the ungulate species in 19 eastern African parks and reserves (Figure 6.16c). The species–area relationship wasS = 16A0.03 and the regression was not significant (F = 0.93, P = 0.35), suggesting that smaller parks do not have fewer species. These authors are also less pessi- mistic about the survival of large-mammal species in reserve ‘islands’. They believe that the extinction rates used by Soulé et al. were too high, having been derived from studies on true oceanic islands. Savannah reserves (at least in East Africa) continue to experience immigration because, Weston and Ssemakula claim, ‘extensive lifestock ranching is not necessarily inimi- cal to wildlife’. They conclude that the ‘significance of island biogeography to the design of nature reserves is limited, at least in the savannahs’. Finally a geographically more extensive dataset was published by East (1983). He looked at ‘large herbivores’ (Figure 6.16d). The species–area relationship was S = 11A0.04 and the regression is not significant F( = 1.81, P = 0.19). Of course one of the problems here is that different studies seem to show both significant or non-significant species–area relationships. It’s not easy to see why. There are differences in the data; even the size of the same protected areas are frequently different. In some cases the species chosen are difficult to discover, because terms like ‘large mammals’ and ‘large herbivores’ are used. We end this section on species–area curves by again presenting some unpublished data that Fiona Wragg and BS have collated from Tanzanian protected areas (Wragg 2002). This data set looked specifically at bovids because a specific species–area curve probably only exists for ecologically similar species. It also looks at 14 contiguous ecological areas, rather than the 24 gazetted parks for which we have data. So for example, the ‘Serengeti island’ = Serengeti National Park + Ngorongoro Conservation Area + Maswa Game Reserve + Masai Mara National Reserve + Ikorongo-Grumeti Game Control Area and ‘Selous island’ = Selous Game Reserve + Mikumi National Park. This is quite important since two adjacent reserves clearly 262 THE BIOLOGY OF AFRICAN SAVANNAHS constitute one ‘island’. The data was compiled from a very extensive litera- ture search (over 200 sources), and confirmed by using the IUCN African Antelope Database (East 1999). The species–area relationship (shown in Figure 6.16e) is S = 2A0.16 and the regression is very significant F( = 16.15, P = 0.002). Although it is not shown in Figure 6.16 we also had data for the larger set of all ungulates. Here the species–area relationship is S = 1.5A0.19, and the regression is again very significant F( = 14.73, P = 0.002). Both these results suggest that there is a species–area relationship for both bovids and ungulates, at least in Tanzania, suggesting that the kind of effect suggested by Soulé et al. (1979) might occur for these savannah species. It should be noted that none of our protected areas had the full complement of Tanzanian bovid species (35). Extrapolating Figure 6.16e until S = 35 suggests that an area of more than 200 million km2 would be need to preserve this number of species: in other words, an area larger than Tanzania. This suggests strongly that protected areas cannot be run on the principle of ‘benign neglect’. That is, they cannot simply be established and left. Even in the largest protected areas such as the Serengeti complex and Selous complex in Tanzania, or the Kruger ecosystem in southern Africa, species will be absent or will eventu- ally be lost. Intervention and management is required, as illustrated below. One of the main problems with these virtual islands we call national parks is the fragmentation that inevitably occurs (Galvin et al. 2007). As noted above, small islands relax over time and lose species. In addition, the preven- tion of migration between these ‘island’ reserves can be crucial. Rothschild’s giraffe Giraffa( camelopardalis rothschildi) has five viable populations in pro- tected ‘islands’, in Kenya and Uganda. In Lake Nakuru National Park there was a rapid decline in numbers from 153 individuals (1995) to 62 individu- als (2002). Brenneman et al. (2009) suggest that the 1994 El Niño-induced drought, which reduced the regrowth ability of Acacia, may have contributed to this giraffe decline. They point out that under free-ranging conditions, giraffe would have had the opportunity to migrate into other areas of the Rift Valley. Many biologists believe that the re-establishment of ancient migration routes is the key to the survival of African wildlife. In 2006 an international agree- ment was signed that will create the Kavango Zambezi (KAZA) Transfron- tier Conservation Area. KAZA will link together 36 national parks located in Angola, Botswana, Namibia, Zambia and Zimbabwe. This will be the largest wildlife ‘island’ on the African continent, about 440,000 km2 in extent. How- ever, some biologists believe that fences will save lions from becoming extinct in the wild (Packer et al. 2013). The 58 authors listed for this paper suggest that fences will prevent free-ranging lions from being killed by humans when they move into nearby agricultural areas. However, another group of 56 lion experts strongly disagree with the use of fences that would further fragment natural habitats (Creel et al. 2013). Packer and co-authors based their argument solely on population density. A dense population of relatively The Savannah Community and Its Conservation 263 few lions in a small reserve near carrying capacity was considered a success, but a much larger number of lions, at a lower density, living in a large reserve was considered a failure. However, when Creel and co-authors looked at this Packer et al. data they didn’t find any relationship between fencing and den- sity when they looked at populations that didn’t exceed the carrying capacity of the environment. This lively debate highlights the complexity of the fenc- ing issue. Both groups agree on two points: fencing may help in some cases, but larger wildlife ‘islands’ are preferable and anti-poaching funding must be increased. The first phase of the establishment of Greater Limpopo Trans- frontier Park involved removal of a fence between the eastern side of Kruger National Park in South Africa and Limpopo National Park in Mozambique. Following the removal of the fence, animals from Kruger National Park were able to migrate into Mozambique, which was depleted of wildlife due to the past civil war (1977–1992). However, the negative trade-off of fence remov- al was the significant rise in rhino poaching within Kruger National Park, which not even the South African army and the use of a drone seem able to control. Money obtained from poaching is helping to fund African terrorist groups, making this a global concern. Cozzi et al. (2013) compared the effects of rivers and fences on large African carnivores. They discovered that fences were a significant obstacle for lions but were more permeable for hyaenas, African wild dogs and cheetahs. The reverse situation was observed for rivers, which were relatively permeable for lions but almost impermeable for African wild dogs and cheetahs. They found that fences helped to create short-term refuges for African wild dogs. Their research highlights the growing importance of how various types of barriers, both artificial and natural, impact various species. Davies-Mostert et al. (2013) examined how fences influence the diet and prey selection of African wild dogs. This study was done in a small (316 km2) private reserve located in Limpopo province, South Africa, that is about half the natural range size for African wild dogs. African wild dogs are cursorial predators that chase their prey until they are exhausted before killing them. Fences were observed to alter predator–prey relationships. Fence-impeded kills resulted in both quantitative and qualitative changes in predator–prey rela- tionships in that larger species, such as kudu, were twice more likely to be killed near a fence than away from a fence. Older age classes of female kudus were also killed near fences, as were a significant number of healthier adult impala males. Normally predators with a cursorial hunting strategy select for younger prey and those in poorer condition, compared to ambush preda- tors. Furthermore, the behaviour of African wild dogs was affected by fences in that they moved towards fences at the beginning of a hunt. This research is significant because in South Africa more and more predators are being introduced into new small reserves (less than 100 km2), that are often smaller than the natural home ranges of predators. Although there were short-term positive benefits for African wild dogs (i.e. an increased hunting efficiency), 264 THE BIOLOGY OF AFRICAN SAVANNAHS the authors suggest that these changes in habitat caused by fences are not sustainable in the long term because they have profound effects on predator– prey relationships. These authors conclude their study by stating that larger ‘islands’ that reduce perimeter-to-area ratios would have less negative effects on predator–prey relationships. In a rather different context, Hovestadtet al. (2005) have looked at spe- cies–area relationships in 49 West African forest–savannah ‘islands’. They examined woody plants with different modes of seed dispersal. According to island biogeography theory, for any given island, we expect steeper slopes and more pronounced spatial patterns for groups of less dispersive species. The z-value for bird-dispersed species was lower (0.11) than that for wind- dispersed species (0.27), with mammal-dispersed species taking an interme- diate value (0.16). This result suggests that, as a group, bird-dispersed species are better colonisers. The authors point out that the standard interpretation of the theory of island biogeography claims that shallow slopes in the spe- cies–area relationship imply low isolation of islands, i.e. good dispersal abili- ties of species. The results of their study appear to contradict this statement. However, the contradiction may be resolved by distinguishing between dis- persal, and the consecutive establishment of populations. Conserving savannah ecosystems Savannah wildlife has a value—ecological, scientific, financial and aesthetic. It also has a cost—loss of human life, loss of property, crops, livestock and income. These conflicting issues are briefly examined here, at the end of this book. For a more extensive account the reader should see the book edited by Prins, Grootenhuis, and Dolan (2000). This conflict between humans and wildlife is often one between the specif- ic interests of people living close to wildlife, with the less localised interests of wildlife conservation in Europe and North America. To really appreciate these issues people in the UK, for example, should imagine how they would feel about wildlife issues if bears and packs of wolves were reintroduced to Wales, North Yorkshire and Scotland? They would be an exciting tourist attraction, but how would we cope with them killing sheep or the occasional person? This is the real issue with conservation on the ground, and if wildlife is to survive, this type of human conflict issue has to be resolved. We cannot simply enclose savannahs by fences, and hope everything will be OK. Harcourt, Parks and Woodroffe (2001) in their examination of the species– area relationship for African reserves, discovered that small reserves tend to be sited in regions of high human population density (Figure 6.17a). Given the problems experienced by small reserves, mentioned in the last section, does this mean that small reserves suffer a double jeopardy? Do they lose The Savannah Community and Its Conservation 265 (a) (b) 6.0 100 80 4.0 60 40 2.0 Log reserve area 20 % Mortality by humans 0.0 0 –2.0 0.0 2.0 4.0 –0.5 0.5 1.5 2.5 Log human population density Log reserve area Figure 6.17 (a) The relationship between log area (km2) of IUCN category I and II African reserves and log local (within 50 km radius) human density (km2). The regression line is log area = 4 − 0.75 log density (F = 78.4, P < 0.0001). (b) Percentage mortality caused by humans on three carnivores, wild dog (●), lion (■), and hyaena (▲). The regression line is with the statistical outlier (Hwange, Zimbabwe), omitted (from Harcourt, Parks, and Woodroffe 2001). species because of their small size (smaller populations, fewer habitats) and because of human interference (hunting, poaching, surrounding degraded land making them more isolated)? To examine the latter effect we have data for eight protected areas, in six African countries (Hwange in Zimbabwe, Kruger in South African, Moremi in Botswana, Etosha in Namibia, Selous and Serengeti in Tanzania and Masai Mara and Nairobi in Kenya), for three carnivore species (wild dog, lion and spotted hyaena). Human density cor- relates significantly with the percentage mortality caused by humans (Fig- ure 6.17b). This suggests that humans have a serious negative effect on wildlife (see also Figure 6.7) and in fact the major conservation issues that we now consider are human induced. There will, of course, always be local con- servation issues that are important but cannot be examined in a small book of this kind. Here we consider hunting and poaching, habitat destruction and wildlife–human conflicts, and wildlife tourism. They are closely linked since ranching, for example, often creates conflicts and habitat destruction, while wildlife tourism may well be an alternative to such ranching in more arid areas. Hunting and poaching As an example of the effect that hunting and poaching can have on wildlife we will outline a series of events that took place in the Serengeti National Park in the 20th century, and that also illustrate how tourism may interact with wild- life protection. Some reference has already been made to this in Chapter 5. Of course hominids have been part of the Serengeti community since the time of Australopithecus afarensis (Peters et al. 2008, Olff and Hopcraft 2008) 266 THE BIOLOGY OF AFRICAN SAVANNAHS but only in more recent times have their descendants become a major dis- turbing influence. In April 1977 the international border between Tanzania and Kenya was closed, and it remained closed until about 1986, when it was reopened partially for tourism. However, the main tourist route between the Masai Mara Reserve, in Kenya, and the Serengeti National Park, in Tanza- nia, remained closed. This border closure immediately affected the number of tourists coming into the Serengeti National Park. In 1976 it was 70,000 and in 1977 it was only 10,000 (Figure 6.18a). Tourist numbers remained at this low level throughout the 1980s and only started to increase again in the 1990s. One consequence for the Serengeti National Park was a drop in income from visitors, leading to a drop in the operating budget of the park. This dropped continuously from 1982 to 1985 and remained low until 1987 (Figure 6.18b). As a result, the anti-poaching effort (in terms of patrol days) had dropped by the mid-1980s to 60% of that prior to the border closure (Sinclair 1995a). At about this time there had also been a considerable increase in the human population between the park and Lake Victoria to the west, often an increase approaching 15% per year. This human population increase, coupled with the lack of anti-poaching patrols, resulted in an invasion of northern and western Serengeti by poachers. The first species to be affected was the black rhinoceros, which was hunted for its horn. It lost 52% of its Serengeti popula- tion in the first year of border closure, 1977 (Figure 6.19a). By 1980 this rhino was effectively extinct in the Serengeti and this is still the situation today. The occasional individual is seen in the southern part of the park, having ‘com- muted’ from the Ngorongoro crater. About 30 black rhinos are still found in the Masai Mara. Elephant and buffalo were the next two species to show an (a) (b) 1,000 80 800 40 600 US dollars 400 20 Number of visitors 200 0 75 80 85 90 78 80 82 84 86 88 90 Year Year Figure 6.18 (a) Annual foreign visitor numbers and (b) income (solid line) and operating budget (dashed line) in US dollar equivalents for the Serengeti National Park, Tanzania. The arrowhead indicates the border closure (from Sinclair 1995a). The Savannah Community and Its Conservation 267 (a) (b) 100 14 80 12 10 60 8 40 6 4 Cumulative % killed 20 Number in thousands 2 0 0 75 1 80 85 2 90 65 70 75 80 85 90 Year Year Figure 6.19 (a) Cumulative percentage of total number killed by poachers and found by Serengeti National Park staff for black rhinoceros (squares) and elephant (triangles). Arrowhead 1 indicates border closure, arrowhead 2 indicates ivory trade ban. (b) Census of buffalo numbers in northern Serengeti. No censuses were conducted during 1977–83. (from Sinclair 1995). effect of the border closure. The elephant was subjected to trophy hunting for tusks, and buffalo were hunted for bush meat. Several hundred elephants moved to the Masai Mara where poaching, although present, was less severe. This difference in the Serengeti and Mara elephant populations eventually had a distinct effect on the vegetation of the two regions (Chapter 5). Both rhino and elephant have been placed on the Convention on International Trade in Endangered Species (CITES) list of protected species. This had a very specific positive effect on elephant poaching in the Serengeti (Fig- ure 6.19a), since there was an abrupt reduction in poaching in 1989 when the CITES ban on the ivory trade was imposed. It also appeared to have a good effect on elephant poaching generally, particularly throughout eastern and southern Africa. In Kenya, for example, elephant populations increased by 30% between 1989 and 1997. The price of ivory dropped and the internation- al trade in ivory collapsed. However, the illegal trade in ivory has once again escalated in some parts of Africa. New methods have been developed to extract DNA from seized ivory, and then determine the geographical source of the ivory (Wasser et al. 2008). This DNA forensic analysis showed that large quantities of ivory seized in Singapore and Hong Kong originated from Zambia. They indicate that between August 2005 and August 2006, 250,000 kg of ivory was smuggled into various countries and that this corresponds to approximately 38,000 poached elephants. Sadly, if poaching can’t be stopped, free-ranging elephants living in the wild will become extinct in a matter of several generations and the long-term effects of killing these ‘ecological engi- neers’ on African savannahs will be devastating. 268 THE BIOLOGY OF AFRICAN SAVANNAHS CITES forms a core piece of international legislation protecting wildlife, particularly the African elephant. It is a United Nations administered trea- ty which was set up in 1975 to control the international trade in wild flora and fauna in order to protect against overexploitation through commercial trade. It first came into force on 1 July 1975, and, as of February 2014, 180 countries had signed the treaty. However, CITES relies on specific manage- ment authorities from signatory countries to regulate and control the trade. There are three levels of regulation, depending on which of three appendices a threatened species is listed. • Appendix I includes species that are threatened with extinction and that are, or may be, affected by international trade. Savannah examples of Appendix I species (as of 13 September 2013) are cheetah, leopard, black and white rhinoceros and some populations of African elephant. • Appendix II includes species that, although not necessarily threatened with extinction, may become so unless trade is regulated, or if their parts or products cannot be easily distinguished from those of other Appendix I or II species. Trade is allowed in Appendix II species, but is strictly regu- lated and only when it has been found that it will not be detrimental to the survival of the species. Savannah examples of Appendix II species are the African elephant (Botswana, Namibia, South Africa and Zimbabwe) and lion. • Appendix III includes species that any country has identified as being exploited and/or threatened within their country, and needs the help of other countries to regulate international trade in it. Savannah examples of Appendix III species are African civets (Botswana) and aardwolf (Botswana). Although legislation, such as the CITES agreements, has helped to combat poaching, particularly for trophies, they have certainly not prevented it. Even when trophy hunting is banned (e.g. for black and white rhino) it has not pre- vented their severe reduction. There appears to be only two ways to reduce indiscriminate hunting: (1) set up protected areas, or (2) use the wildlife resources and consequently ensure its protection as an investment. This latter approach has been tried in South Africa, Namibia, Botswana, and Zimbabwe. Nelson (2007) reports that Namibia is a country where large- mammal populations are widely stable or increasing. In the late 1960s, Namibia gave private landholders rights to utilise wildlife, and extended these provisions to communal lands in the mid-1990s. These changes led to wildlife becoming an economically valued asset at the local level. Wild- life on private ranches increased by an estimated 80% from 1972 to 1992. In the communal lands of northwest Namibia’s Kunene region, aerial surveys indicated that springbok, gemsbok and mountain zebra all increased ten- fold from 1982 to 2000. The idea of using wildlife as a sustainable resource, that nevertheless maintains the savannah ecosystem intact, is one of the The Savannah Community and Its Conservation 269 conservation strategies that believes in the saying ‘use it or lose it’. Another is tourism, and we approach that topic by way of the human–wildlife conflicts imposed by agriculture and ranching. Habitat destruction and wildlife–human conflict Wildlife–human conflicts occur frequently in rural areas. African farmers sometimes lose considerable amounts of their crops to wild animals (Simons and Chirambo 1991). In Nyami Nyami district, in Zimbabwe, crop damage accounted for 96% of all complaints about animal problems (Hoare 1995). Crop loss in Laikipia district, Kenya, was estimated at between 10% and 24% of the total maize crop (Thoules 1994) and Ngure (1995) records that in 1991 Kenyans around Tsavo National Park lost crops worth an average of US$76 per farm. Farmers close to Kasungu National Park, Malawi, lost 10% of their crops to wildlife (Deodatus and Lipiya 1991). However, there is an important and intriguing misconception among farmers about the wildlife responsible for this damage. Table 6.4 shows a comparison of ranked raid frequency on farms and numbers of complaints made at Kasungu National Park, Malawi. The two rankings are significantly different: 60% of the complaints made by farmers in this 6 month period related to elephants, while in reality they ranked third after baboons and bush pig in raid frequency. In Nyami Nyami district in Zimbabwe elephants accounted for 78–80% of all complaints, with buffalo second at 15–18% (Hoare 1995). Large animals are frequently associated in the minds of farmers with ‘wildlife’, local national parks and restrictions on economic activities. They see themselves in conflict with these wildlife species. Smaller species, such as rodents, are taken for granted and simply regarded as ‘agricultural pests’ rather like fungi, nematodes and insects. Although exact figures are difficult to obtain it is known, for example, that near seven Tanzanian protected areas the main pest species were pri- mates (51.9%), bush pig (13.3%) and rodents (10.6%) (Newmark et al. 1994). In savannah areas where crop growing could be profitable the conservation Table 6.4 Comparison of raid frequency (average monthly number of raids) on farms, and number of complaints made at Kasungu National Park, Malawi, January–June 1990 Species Mean raid Raid Number of Complaint frequency ranking complaints ranking bushpig 9.63 1 5 2 baboon 5.01 2 0 5 elephant 3.01 3 15 1 vervet monkey 1.71 4 2 4 hippopotamus 0.67 5 3 3 buffalo 0.33 6 0 5 From Deodatus and Sefu (1992). 270 THE BIOLOGY OF AFRICAN SAVANNAHS solution has to be large reserves along with a well-supervised system of com- pensation for farmers. Revenue from tourism and systems like CAMPFIRE (Frost and Bond 208) will also help to change the attitude of local farmers. But the local population must be involved in this conservation process. The same type of problem exists where small carnivores take poultry (Holmern and Røskaft 2014). Using structured interviews, they found that in seven vil- lages outside Serengeti National Park, small predators were perceived to kill almost one-third of the poultry kept by households, which is equivalent to about 10% of their income. Both flock size and spatial location were impor- tant factors in perceiving predator-caused loss, raising the possibility that economic reliance on the resource in question may influence the percep- tion of predators. Kesch, Bauer and Loveridge (2014) provide an interesting insight into the problems of using game fences to guard against predators. Another type of ‘fence’ has recently been tested by King et al. (2009) to deter crop-raiding elephants. In Laikipia, Kenya, they deployed a 90 m fenceline of nine interconnected beehives (Apis mellifera scutellata) around two sides of a 2 acre (0.8 ha) farm experiencing high levels of elephant crop depreda- tion. Subsequently, compared with a nearby control farm of similar status, the protected farm experienced fewer raids. In those savannah areas where cattle are traditionally reared, the conserva- tion issues are slightly different. Here there is the possibility of using the same area for both wildlife and livestock, a situation that is not really possible with agricultural crops, or poultry. However, there are still issues of conflict. There is a widespread belief that grazing wildlife such as zebra and wildebeest com- pete with cattle for grass (e.g. Pratt and Gwynne 1977). What is the evidence for this? Prins (2000) reviewed the available literature and came to the con- clusion that despite the considerable diet overlap between the two groups (an essential prerequisite for competition) there was little evidence for food limitation of livestock populations. He suggested that competition is largely asymmetrical and diffuse, with cattle having a competitive effect on wildlife, but wildlife having little or no effect on cattle. However, controlled replicate field experiments were lacking. Since then, Young, Palmer and Gadd (2005) have published the results of a long-term enclosure experiment in Laikipia, Kenya. The experiments ran from 1995 to 2002 and consisted of a number of experimental enclosures using a series of semipermeable barriers to differentially exclude cattle (Bos indicus), megaherbivores (elephants and reticulated giraffe), and all ‘wild- life’ (herbivores >15 kg = common zebra, Grevy’s zebra, buffalo, eland, Grant’s gazelle, hartebeest, oryx and steenbok). Each plot measured 200 × 200 m (4 ha), and the following experimental treatments were applied: all large mammals excluded (O); only cattle allowed (C); only wildlife allowed (W); wildlife and cattle allowed (WC); wildlife and megaherbivores allowed (MW); all large herbivores allowed (MWC). Dung counts (as a measure of species use) in the enclosures showed that the barriers were effective in The Savannah Community and Its Conservation 271 1,000 750 500 zebra dung piles per ha 250 OWMW CWC MWC Herbivore treatment Figure 6.20 The density of zebra dung in 2000/2001 in plots from which cattle and megaherbivores had been excluded since September 1995. Error bars ±1 SE (from Young et al. 2005). excluding the target species, and zebra dung showed a number of significant patterns (Figure 6.20). There was a 44% increase in the presence of zebra dung (550 dung piles/ha in MWC and WC vs 805 in MW and W; F = 23.85, P = 0.0028), although zebra dung density was essentially the same in plots with and without megaherbivores (680 in MW and MWC vs. 670 in W and WC; F = 0.013, P = 0.91). However, there was an interesting, and significant, cattle–megaherbivore interaction. In plots without megaherbivores there was 79% more zebra dung when cattle were excluded (485 in WC vs 865 in W), while in plots with megaherbivores, but cattle excluded, there was only a 22% increase (610 in MWC vs 745 in MW; F = 11.39, P = 0.015). Young et al. (2005) also carried out vegetation surveys on their experimental plots and interestingly the results paralleled those for the zebra dung. Zebra dung density was strongly correlated with grass cover (suggesting they can track resources), which was negatively associated with cattle presence. In con- junction with published reports of strong dietary overlap (Casebeer and Koss 1970, Hoppe et al. 1977, Voeten and Prins 1999) this suggests that zebra compete with cattle for food. However, the implication is that this competition is highly asymmetrical and that wildlife (zebra) suffer more from the competition with cattle than vice versa. Compared to the total- exclusion plots, cattle alone reduced grass cover by 33%, and there was no further reduction in grass cover when wildlife were added. In contrast, wildlife alone (mostly zebra) reduced grass cover by only 14% and wildlife plus megaherbivores reduced cover by 21%. It appears that cattle can fully compensate for the absence of zebra and other wildlife. Here compensa- tion means the lack of a decrease in a shared resource (grass) when one 272 THE BIOLOGY OF AFRICAN SAVANNAHS competitor, but not the other, is excluded. In the absence of cattle, zebra and other wildlife do not compensate fully, in terms of grass cover. This con- trolled field experiment therefore confirms the conclusions of Prins (2000), previously mentioned. This conclusion might appear to indicate that livestock ranches can allow wildlife to flourish on their property without fear of it affecting their cattle. Unfortunately not all property owners believe that this is true. For exam- ple, Heath (2000) examined the estimated costs of ranching in the Laikipia district of Kenya (Table 6.5). On a typical 20,000 ha ranch he suggests that the 1,475 wildlife represent the metabolic equivalent of 1,095 cattle and that even after taking into consideration the varying degrees of overlap in diet this still amounts to some 830 cattle. Despite the claims of people like Prins (2000), Heath maintains that water shortage may often be the limiting factor in these semi-arid savannahs that are used for cattle rearing. In fact, ranch- ing in these areas is not a very secure occupation anyway. Heath (2000) gives his estimates of the anticipated profit/loss for different-sized cattle herds on a 20,000 ha ranch in Laikipia (Table 6.5). Size appears to be crucial. Small ranches are very near the loss margin, and in Laikipia wildlife are often removed from these ranches. In contrast, large ranches tolerate wildlife, and even use tourism to supplement their income. They have formed the Laikipia Wildlife Forum, established in 1992 by private and communal landowners with common interests in managing, conserving and profiting from wildlife resources. The organisation was created in response to an initiative by the Kenya Wildlife Service, designed to engage landowners and land users in the conservation and management of wildlife in non-protected areas. In addi- tion it also protects essential environmental resources such as river flow, as well as improving the livelihood and security of local people. Membership of the forum currently consists of 36 large-scale ranches, 47 community groups, 50 tour operators, 54 individuals and 8 interest groups. Since an estimated 60–70% of wildlife in Kenya is found outside protected areas, this may well Table 6.5 Estimated profit/loss from different-sized cattle herds on a 20 000 ha ranch in Laikipia, Kenya. All money in US$. Income is based on cattle having an average weight of 400 kg and an average market price of US$1.05 per kg. Direct costs include herders and veterinary inputs such as dips, salt, and drugs. Overheads include management, ancillary staff, vehicle and equipment maintenance, fuel, office, bank, insurance, and land rates No. of cattle 1,500 2,000 2,500 3,000 Income 126,000 168,000 210,000 252,000 Expenditure Direct costs 26,170 34,909 43,615 52,338 Overheads 116,541 129,457 142,403 156,643 Cash profit/loss −16,711 3,634 23,982 43,019 After Heath (2000). The Savannah Community and Its Conservation 273 Table 6.6 Tourism figures for six African countries with extensive savannah ecosystems, for 2003 Country Tourist arrivals Tourism receipts Tourism as (000s) (US$ M) % exports Botswana 975 309 NA Kenya 927 611 17.1 South Africa 6,505 5,232 11.5 Tanzania 552 441 28.1 Zambia 578 149 11.2 Zimbabwe 2,068 44 NA NA, not available. From Mitchell and Ashley (2006). be the way forward for many of these semi-arid savannahs, in both eastern and southern Africa, that are outside national parks. A good example of one such ranch within the forum is Lewa Downs, in the Meru District of Kenya. Lewa Wildlife Conservancy (LWC) is situated in Laikipia district, in the northern foothills of Mount Kenya, and includes an estimated area of 2,200 km2. The main vegetation type is Acacia savannah, grassland and some indigenous forest. Lewa was once a cattle ranch; it then became a heavily guarded black rhino sanctuary, and it is now the head- quarters for a non-profit wildlife conservancy. The Craig/Douglas family first came to Lewa Downs in 1922, and managed it as a cattle ranch for over 50 years. In the 1980s the emphasis at LWC changed and wildlife conserva- tion became the prime objective. Cattle are still farmed but in smaller num- bers. In addition to being a private conservation area for all wildlife, Lewa Downs supports a variety of programmes. These include an endangered spe- cies programme (black rhino, elephant, Grevy’s zebra, cheetah, lion, leop- ard, hyaena, and wild dogs), a research programme (including Grevy’s zebra, carnivores and general monitoring of wildlife) and a community develop- ment programme (health, education and farm development). Income comes from farming, tourism and donations. It has an annual budget of US$2.2 million and provides employment for some 450–500 people. A large part of the budget is allocated to the community projects (24% in 2006), and the endangered species programme, particularly the black rhino (24% in 2006) and Grevy’s zebra (12% in 2006) programmes. We mentioned earlier in this chapter that open-pit uranium mining will threaten Selous Game Reserve. Sadly, there are more examples of habitat destruction. Years of open-pit copper mining in Zambia has polluted the Kafue river that runs through Kafue National Park, and now a new open- pit copper mine has recently been proposed to begin operations within the Lower Zambezi National Park (LZNP). This new mining operation will likely pollute not only LZNP but also the World Heritage Mana Pools located on the 274 THE BIOLOGY OF AFRICAN SAVANNAHS Zimbabwe side of the Zambezi river. The river valley is pristine and abounds in wildlife. The Zambezi is one of the most iconic rivers in the world, flowing through Angola, Botswana, Malawi, Mozambique, Namibia, Tanzania, Zam- bia and Zimbabwe. In January 2014, the Zambian Minister of Lands, Natural Resources and Environmental Protection rejected an environmental impact study done by the Zambian Environmental Management Agency which concluded that mining in this sensitive and legally protected ecosystem was not in the best interests of Zambians. Many Zambians were shocked that the minister approved this project. Public outcry voiced by various stakehold- ers was so intense that the High Court in Lusaka ruled in late March 2014 that further research and consultation is needed before mining can begin. Zambezi Resources Limited (ZRL) is the Australian exploratory company working with the Tanzanian government on this project and their website promises that the Kangaluwi mine will create lots of jobs for local Zambians and be a model for ‘green mining’. It is difficult to imagine how any form of open-pit mining can be thought of as being environmentally friendly and once again, investment and most of the labourers for the Kangaluwi project will most likely come from China. This is troubling, because China does not have a good record for environmental protection (Tan-Mullins and Mohan 2013). Furthermore, African governments in general do not regulate how investments, Chinese or otherwise, are being used to exploit their rich, yet finite, mineral deposits. Legislation is urgently needed if national parks and reserves are to be protected, otherwise habitat destruction in Africa will be on a scale similar to that in the Amazon basin. Given that Zambia is strug- gling to clean up harmful pollution of the Kafue river watershed, the lifeline of Kafue National Park and source of water for 40% of Zambians, caused by the Chinese-run Konkola copper mines, it seems difficult to understand why the government would risk mining in yet another sensitive ecosystem. Of course the answer is obvious—the development of natural resource-based industries would improve the economy of Zambia—but at what cost? Pollutants in the Olifants river, a major river that flows through central Kru- ger National Park in South Africa, pose a serious threat to the survival of a keystone predator, the Nile crocodile (Crocodylus niloticus). The source of the Olifants river lies to the west of Kruger National Park in a highly populated agricultural region that contains several large mining operations. Unfortu- nately acid mine drainage, pesticides, polychlorinated biphenyls (PCBs), and sewage are carried by the Olifants river into Kruger National Park. Between 2008–2010 approximately 200 crocodiles died of pansteatitis, a disease that converts white fat into a hard yellow fat (Ashton 2010, Ferreira and Pienaar 2011). Since this hard yellow fat is not properly metabolised by crocodiles, they become lethargic and eventually die. This mass death of crocodiles is correlated with the building of the Massingir dam located downstream in Mozambique (Ferreira and Pienaar 2011). Research has shown that this dam functions as a sediment trap that is concentrating pollutants. All of the deep The Savannah Community and Its Conservation 275 pools that were the primary habitats of crocodiles during the dry season are filling up with toxic silt. Also, an extensive series of rapids along the Olif- ants river is disappearing due to silting and this has dramatically altered plant and animal life associated with the rapids. Sadly, the entire aquatic food web has been affected by pansteatitis and silting. As more dams will be built for hydroelectricity, more watersheds will be threatened. Pansteastitis is not restricted to crocodiles. Sharptooth catfish living in the same habitat as the affected Olifants crocodiles are dying of pansteastitis (Woodborneet al . 2012). These researchers have shown that a change in habitat is most likely responsible for this epidemic. When the river substrate was changed from rock and sand to a clay silt, catfish switched from a mostly vegetarian diet to a piscivorous diet. The Massingir dam has concentrated pollutants that have significantly altered the Olifants river food web and put a large population of crocodiles on the brink of extinction. Many other species will likely be affected. Not only is this a crisis for Kruger National Park, but the Olifants river is a major source of water for people living near the river and this water is used for agriculture and livestock. Back in 1962 Rachael Carson sounded the alarm regarding the harmful effects of pesticides on birds, yet, sadly, this continues to pose an urgent and costly threat to the long-term survival of African savannahs. Furthermore, it has recently been shown that pollutants, altered habitats, and a host of other environmental factors significantly alter the epigenetic mechanisms of plants and animals. At present, nothing is known about epigenetic modi- fications of African savannah plants and animals. New research is needed to explore this topic. Fortunately there is some good news regarding the protection of watersheds. The source of the Great Ruaha river, the Usangu wetlands, is now partially protected. After decades of deterioration caused by overgrazing of cattle, the removal of cattle resulted in the rapid recovery of the wetlands and the volume of water flowing into the Great Ruaha river doubled (Kihwele et al. 2012). Cattle were removed from the eastern region of the wetlands, so hopefully the western region of the Usangu wetlands will also be protected as this river is essential for wildlife living in Ruaha National Park and for local Africans farming near this river. The World Widelife Fund (WWF) has a project aimed to protect the Kipengere moun- tains from deforestation as these mountains supply water to the Usangu wetlands. Africa’s oldest national park, Virunga National Park in the Democratic Republic of the Congo, is now being threatened by oil exploration. Virunga National Park is a complex mosaic of savannahs, forests, wetlands, and vol- canoes and is home for the threatened okapi (Okapia johnstoni) and moun- tain gorilla (Gorilla gorilla beringei). Approximately 200 of the remaining 880 mountain gorillas live in this park. Oil concessions cover 85% of the park. Soon a UK-based oil company, Soco International, will begin seismic testing within the park. Likewise, the Congo river basin with its sensitive 276 THE BIOLOGY OF AFRICAN SAVANNAHS southern Congolian forest–savannah ecoregion contains significant deposits of copper, manganese, uranium, diamonds and oil. New roads and railways for exploration will no doubt increase habitat fragmentation, logging and poaching activities in the very heart of Africa (Reed and Miranda 2007). Another major conservation issue causing international alarm is the build- ing of a road across the northern region of Serengeti National Park. What will be the impact of this road on the annual migration? Computer model- ling suggests that the migration will be significantly impacted and that great care must be taken where this road will be built (Holdo et al. 2011). There is the real possibility that this road will end the world’s greatest migration. A road would likely increase poaching and the invasion of exotic plants. The gravel road will link coastal Tanzania with Lake Victoria and provide a corridor to the Democratic Republic of the Congo, Uganda and Rwanda. For many Tanzanians the Serengeti is a barrier holding back their economic development. However, in the next section we discuss the economic benefits of wildlife tourism. In 2010, a group of prominent ecologists held a 5 day workshop to discuss what was learned from conservation efforts in South Africa and how this information could be applied to East Africa (Beale et al. 2013). South Africa has a much longer history of conservation than East Africa, as until relatively recently many of the wilderness areas in East Africa were visited by hunters. They came up with a list of 10 lessons: • The boundaries of protected areas should coincide with intact ecosystems. • Local communities must directly benefit from protected areas. • Land use in buffer zones should be compatible with the conservation of protected areas. • Migrations and corridors between protected areas should be maintained. • Rivers and watersheds must be protected. • Law enforcement is essential. • New invasive plants should be identified before they multiply and become extremely costly and difficult to remove. • Extensive planning of where new roads will be built will reduce harmful effects to ecosystems. • The loss of heterogeneity results in unstable ecosystems. • Researchers and park managers need to work together as scientific infor- mation often does not reach practitioners, and scientists often do not con- sult with park managers. Examples of all of these lessons are discussed in this workshop. However, the authors admit that socio-political factors were not accounted for in their analysis of what needs to happen to protect African savannahs. Clearly, socio-political factors are a powerful force in determining the fate of any environment so we can only hope for a positive outcome. The Savannah Community and Its Conservation 277 Wildlife tourism The desire of overseas tourists to experience African wildlife in its natural habitat has resulted in Africa having one of the strongest tourist growth mar- kets, with most destinations showing consistently above-average increases in arrivals and receipts. Between 2000 and 2005, international tourist arriv- als to Africa increased from 28 million to nearly 40 million—an average growth of 5.6% a year, compared to a worldwide 3.1% a year. In the same period Africa’s international tourism receipts doubled from US$10.5 billion to US$21.3 billion. Total travel and tourist revenues for the whole of Africa were expected to generate $73.6 billion of gross domestic product (GDP) in 2005, equivalent to 8.8% of the regional economy. The tourist sector also provided 3,877,200 jobs directly and a total of 10,647,000 jobs indirectly in 2005, or 6.8% of all employment in sub-Saharan Africa. According to United Nations World Trade Organization figures for 2005, the number of visitors to sub-Saharan Africa from outside the region increased 13%, to 23.1 million. The most rapid growth was recorded in Kenya, where the number of visi- tors increased 26% on 2004, and Mozambique, where the number of visitors increased a massive 37%. Many of the African countries that support large areas of savannah habitat are high on the tourist list. Table 6.6 compares six countries for which figures are available for 2003. Most of this tourism is due to wildlife, and interest in African wildlife centres on savannah ecosystems with their charismatic large herbivores and carni- vores. In Botswana it is estimated that wildlife viewing accounted for about one-half of the total overseas’ tourist expenditure and generated up to $3 million income for the government in 1990. About one-half of this was from entry fees and about one-fifth from tax revenues (Modise 1996). In Kenya, in 1995, about 70% of the tourist income ($500 million) could be attributed to wildlife tourism, and government revenues from this were in excess of $20 million (Republic of Kenya 1996). Key wildlife species are especially highly valued. For example, in Kenya the net global returns to the 1.5 million wil- debeest of the Masai Mara National Reserve have been estimated at between $125 and $150 per animal per year (Norton-Griffiths 1995). In Amboseli National Park the viewing value of a lion has been estimated to be in excess of $0.5 million (Thresher 1981) and the flamingos of Lake Nakuru Nation- al Park have an annual tourist value of between $3 million and $5 million (Earnshaw and Emerton 2000). If wildlife is worth so much, via tourism, isn’t this a way to ensure that wildlife is protected? Isn’t it an investment worth protecting for its financial return, in addition to its aesthetic and scientific value? The answer is yes. But the financial returns must filter down to the local farmers, ranchers, hunters and poachers who would otherwise see a better profit in the demise of wildlife. In just over a decade there has been profound change in those employed to run Kruger National Park. Now the majority of staff, guides, and administrators 278 THE BIOLOGY OF AFRICAN SAVANNAHS who operate the extensive network of camps in the park are local people from nearby villages and towns instead of white expats. Hopefully the enlightened hiring policy of Kruger National Park will be adopted by other national parks and reserves across Africa. It is encouraging to see that at least a few camps are run by local Masai in the Serengeti–Masai Mara region; however, the vast majority of camps in Kenya, Tanzania, and Zambia continue to be owned and run by expats. A camp run by local Africans is an important value-added cultural experience for tourists. The bottom line is that local communities must significantly benefit from tourism, and perhaps from hunting schemes such as CAMPFIRE. Howev- er, the issue of trophy hunting is a complex one beyond the scope of this book. In 2013 Botswana banned all hunting and in 2013 Zambia banned trophy hunting of lions because their numbers have significantly dropped. The problem with conserving savannah ecosystems (and other ecosystems) is that we do not simply have to convince those of us that already value their scientific worth, and their exciting beauty. We also have to convince those people who see a financial gain in destroying them. Convince this lat- ter group that African savannahs should be protected, and they will be safe for posterity. The expansion of wildlife education programmes for younger generations of Africans, including taking students on field trips to national parks and reserves, will hopefully be an important factor in trying to save African savannahs from the growing threat of exponential human popula- tion growth taking place on the African continent. We hope that many of those who read this book will be young Africans. There is also global climate change and unfortunately no one can predict the outcome of complex evolu- tionary changes shaped by human activities! References Abrams P. (1983). The theory of limiting similarity. Annual Review of Ecology and Systematics, 14, 359–376. Abu Baker M. A. and Brown J. S. (2014). Foraging and habitat use of common duik- ers, Sylvicapra grimmia, in a heterogeneous environment within the Soutpansberg, South Africa. African Journal of Ecology, 52, 318–317. Andersen G. L. and Krzywinski K. (2007). Longevity and growth of Acacia tor- tilis; insights from 14C content and anatomy of wood. BMC Ecology, 7:4 doi:10.1186/1472-6785-7-4. Anderson J. M. and Coe M. J. (1974). Decomposition of elephant dung in an arid, tropical environment. Oecologia, 14, 111–125. Anderson T. M., Metzger K. L., and McNaughton S. J. (2007a). Multi-scale analysis of plant species richness in Serengeti grasslands. Journal of Biogeography, 34, 313–323. Anderson T. M., Ritchie M. E., Mayemba E., Eby S., Grace J. B., and McNaughton S. J. (2007b). Forage nutritive quality in the Serengeti ecosystem: the roles of fire and herbivory. American Naturalist, 170, 343–357. Archie E. A., Maldonado J. E., Hollister-Smith J. A., Poole J. H., Moss C. J., Fleischer R. C., and Alberts S. C.(2008). Fine-scale population genetic structure in a fission- fusion society. Molecular Ecology, 17, 2666–2679. Archie E. A., Moss C. J., and Alberts SC. (2006). The ties that bind: genetic relatedness predicts the fission and fusion of social groups in wild African elephants.Proceed - ings of the Royal Society of London Series B, 273, 513–522. Arsenault R. and Owen-Smith N. (2002). Facilitation versus competition in grazing herbivore assemblages. Oikos, 97, 313–318. Ashton P. J. (2010). The demise of the Nile crocodile Crocodylus( niloticus) as a keystone species for aquatic ecosystem conservation in South Africa: the case of the Olifants River. Aquatic Conservation: Marine and Freshwater Ecosystems, 20, 489–493. Attardo G. M., Lohs C., Heddi A., Alam U. H., Yildirim S., and Aksoy S. (2008). Analysis of milk gland structure and function in Glossina morsitans: milk protein production, symbiont populations and fecundity. Journal of Insect Physiology, 54, 1236–1242. Attwell R. I. (1966). Oxpeckers and their associations with mammals in Zambia. Puku, 4, 17–47. Aubréville A. (1963). Classification des formes biologiques des plantes vasculaires en milieu tropical. Adonsonia, 3, 221–226. Augustine D. J. (2002). Large herbivores and process dynamics in a managed savanna ecosystem. PhD dissertation, Syracuse University, Syracuse, NY, USA. Augustine D. J. and McNaughton S. J. (2004). Regulation of shrub dynamics by native browsing ungulates on East African rangeland. Journal of Applied Ecology, 41, 45–58. 280 REFERENCES Auty H., Anderson N. E., Picozzi K., Lembo T., Mubanga J., Hoare R., Fyumagwa R. D., Mable B., Hamill, L., Cleaveland S., and Welburn S. C. (2012). Trypanosome diversity in wildlife species from the Serengeti and Luangwa Valley ecosystems. PLOS Neglected Tropical Diseases, 6, e1828. Bagniewska J. M. and Kamler J. F. (2014). Do black-backed jackals affect numbers of smaller carnivores and prey? African Journal of Ecology, 52, DOI: 10.1111/ aje.12125. Bailey T. N. (1993). The African leopard. Ecology and behavior of a solitary felid. Columbia University Press, New York. Barbour M. G., Burk J. H., Pitts W. D., Gilliam F. S., and Schwartz M. W. (ed.) (1999). Terrestrial plant ecology, 3rd ed. Benjamin/Cummings Publishing Company, San Francisco, CA. Barnes M. E. (2001). Effects of large herbivores and fire on the regeneration of Acacia erioloba woodlands in Chobe National Park, Botswana. African Journal of Ecology, 39, 340–350. Beale C. M., Van Rensberg S., Bond W. J., Coughenour M., Fynn R., Gaylard A., Grant R., Harris B., Jones T., Mduma S., Owen-Smith N., and Sinclair A. R. E. (2013). Ten lessons for the conservation of African savannah ecosystems. Biological Conserva- tion, 167, 224–232. Beddington J. R., Free C. A., and Lawton J. H. (1975). Dynamic complexity in predator-prey models framed in difference equations. Nature, 255, 58–60. Begon M., Harper J. L., and Townsend C. R. (1996). Ecology. Individuals, populations and communities, 3rd ed. Blackwell Science, Oxford. Bell R. H. V. (1969). The use of the herb layer by grazing ungulates in the Serengeti National Park, Tanzania. PhD thesis, University of Manchester. Bell R. H. V. (1970). The use of the herb layer by grazing ungulates in the Serengeti. In A. Watson (ed.), Animal populations in relation to their food resources, pp. 111–123. Blackwell Scientific Publications, Oxford. Bell R. H. V. (1971). A grazing ecosystem in the Serengeti. Scientific American, 224(1), 86–93. Bell R. H. V. (1982). The effect of soil nutrient availability on community structure in African ecosystems. In B. J. Huntley and B. H. Walker (ed.), Ecology of tropical savannas, pp. 193–216. Springer-Verlag, New York. Belsky A. J. (1990). Tree/grass ratios in East African savannas: a comparison of exist- ing models. Journal of Biogeography, 17, 483–489. Belsky A. J. (1992). Effects of grazing, competition, disturbance and fire on species composition and diversity in grassland communities. Journal of Vegetation Science, 3, 187–200. Bengis R. G., Grant R., and De Vos V. (2003). Wildlife diseases and veterinary con- trols: a savannah ecosystem perspective. In J. T. Du Toit, K. H. Rogers and H. C. Biggs (ed.) The Kruger experience: ecology and management of savannah heteroge- neity, pp.349–369. Island Press, Washington DC. Benoit J. B., Attardo G. M., Michalkova V., Takac P., Bohova J., and Aksoy S. (2012). Sphingomyelinase activity in mother’s milk is essential for juvenile development: a case from lactating tsetse flies. Biology of Reproduction, 87, 1–10. Ben-Shahar R. (1990). Resource availability and habitat preferences of three African ungulates. Biological Conservation, 54, 357–365. Ben-Shahar R. and Skinner J. D. (1988). Habitat preferences of African ungulates derived by uni- and multivariate analyses. Ecology, 69, 1479–1486. REFERENCES 281 Benson-Amram S., Heinen V. K., Dryer S. L. and Holekamp K. E. (2011). Numerical assessment and individual call discrimination by wild spotted hyaenas, Crocuta crocuta. Animal Behaviour, 82, 743–752. Bercovitch F. B. and Berry P. M. (2013). Herd composition, kinship and fission– fusion social dynamics among wild giraffe. African Journal of Ecology, 51, 206–216. Berendt J. E. (1992). The third ear. Henry Holt, New York. Berriman M., Ghedin E., Hertz-Fowler C., Blandin G., Renauld H., Bartholomeu D. C., Lennard N. J., Caler E., Hamlin N. E., Haas B., Bohme U., Hannick L., Aslett M. A., Shallom J., Marcello L., Hou L., Wickstead B., Alsmark U. C. M., Arrowsmith C., Atkin R. J., Barron A. J., Bringaud F., Brooks K., Carrington M., Cherevach I., Chillingworth T., Churcher C., Clark L. N., Corton C. H., Cronin A., Davies R. M., Doggett J., Djikeng A., Feldblyum T., Field M. C., Fraser A., Goodhead I., Hance Z., Harper D., Harris B. R., Hauser H., Hostetler J., Ivens A., Jagels K., Johnson D., Johnson J., Jones K., Kerhornou A. X., Koo H., Larke N., Landfear S., Larkin C., Leech V., Line A., Lord A., MacLeod A., Mooney P. J., Moule S., Martin D. M. A., Morgan G. W., Mungall K., Norbertczak H., Ormond D., Pai G., Peacock C. S., Peterson J., Quail M. A., Rabbinowitsch E., Rajandream M., Reitter C., Salzberg S. L., Sanders M., Schobel S., Sharp S., Simmonds M., Simpson A. J., Tallon L., Turner C. M. R., Tait A., Tivey A. R., Van Aken S., Walker D., Wanless D., Wang S., White B., White O., Whitehead S., Woodward J., Wort- man J., Adams D., Embley T. M., Gull K., Ullu E., Barry J., Fairlamb A. H., Opper- does F., Barrell B. G., Donelson J. E., Hall N., Fraser C. M., Melville S. E., and El-Sayed N. M.. (2005). The genome of the African trypanosome Trypanosoma brucei. Science, 309, 416–422. Bertram B. C. R. (1979). Serengeti predators and their social systems. In A. R. E. Sinclair and M. Norton-Griffiths (ed.) Serengeti. Dynamics of an ecosystem, pp. 221–248. University of Chicago Press, Chicago. Bethune S., Griffin M., and Joubert D. (2004). National review of invasive alien species in Namibia. Ministry of Environment and Tourism, Windhoek. Bezuidenhout J. D. and Stutterheim C. J. (1980). A critical evaluation of the role played by the red-billed oxpecker Buphagus erythrorhynchus in the biological control of ticks. Onderstepoort Journal of Veterinary Research, 47, 51–75. Bioacoustics Research Program (2014). Raven Pro: interactive sound analysis software (Version 1.5). Cornell Lab of Ornithology, Ithaca, NY. Available from Boersma P. and Weenink D. (2014). Praat: doing phonetics by computer (Version 5.3.62). Available at Cambefort Y. (1984). Community structure and role in dung burial of forest and savanna Scarabaeidae dung beetles in Ivory Coast. XVII International Congress of Entomology, Hamburg, Abstract Volume, p. 339. Campbell H., Townsend I. R., Fellowes M. D. E., and Cook J. M. (2013). Thorn- dwelling ants provide antiherbivore defence for camelthorn trees, Vachellia eri- oloba, in Namibia. African Journal of Ecology, 51, 590–598. Caro T. M. (1994). Cheetahs of the Serengeti plains. University of Chicago Press, Chicago. Carson R. (1962). Silent spring. Houghton Mifflin, New York. Carter K. D., Seddon J. M., Frère C. H., Carter J. K. and Goldizen A. W. (2013a). Fission/fusion dynamics in wild giraffes may be driven by kinship, spatial overlap and individual social preferences. Animal Behaviour, 85, 385–394. Carter K. D., Brand R., Carter J. K., Shorrocks B., and Goldizen A. W. (2013b). Social networks, long-term associations and age-related sociability of wild giraffes. Ani- mal Behaviour, 86, 633–639. Casebeer R. L. and Koss G. G. (1970). Food habits of wildebeest, zebra, hartebeest and cattle in Kenya Masailand. East African Wildlife Journal, 8, 25–36. Caswell H., Reed R., Stephenson S. N., and Werner P. A. (1973). Photosynthetic path- ways and selective herbivory: a hypothesis. American Naturalist, 107, 465–480. Caughley G. (1974). Bias in aerial survey. Journal of Wildlife Management, 38, 921–933. Caughley G. (1976). The elephant problem—an alternative hypothesis. East African Wildlife Journal, 14, 265–284. Cerling A. K., Wittemyer G., Ehleringer J. R., and Douglas-Hamilton I. (2009). His- tory of animals using isotope records (HAIR): A 6-year dietary history of one fam- ily of African elephants. Proceedings of the National Academy of Science of the U S A 106, 8093–8100. Cerling T. E., Wittemyer G., Rasmussen H. B., Vollarath F., Cerling C. E., Robinson T. J., and Douglas-Hamilton I. (2006). Stable isotopes in elephant hair document migration patterns and diet changes. Proceedings of the National Academy of Sci- ences of the U S A, 103, 371–373. Chalmers M. I. (1961). Protein synthesis in the rumen. In D. Lewis (ed.), Digestive physiology and nutrition in the ruminant, pp 205–222. Butterworth, London. Charnov E. L. (1973). Optimal foraging: some theoretical explorations. PhD. Disser- tation, University of Washington. Cleaveland S., Packer C., Hampson K., Kaare M., Kock R., Craft M., Lembo T., Mlen- geya T., and Dobson A. (2008). The role of infectious diseases in the Serengeti eco- system. In:, A. R. E. Sinclair et al. (ed), Serengeti III. Human impacts on ecosystem dynamics. University of Chicago Press, Chicago. Coates Palgrave K. (1983). Trees of southern Africa. Struik Publishers, Cape Town. Coe M. J., Cumming D. H., and Phillipson J. (1976). Biomass and production of large African herbivores in relation to rainfall and primary production. Oecologia, 22, 341–354. Cole M. M. (1986). The savannas. Biogeography and geobotany. Academic Press, London. Connell J. H. (1980). Diversity and the coevolution of competitors, or the ghost of competition past. Oikos, 35, 131–138. Connell J. H. (1983). On the prevalence and relative importance of interspecific com- petition: evidence from field experiments. American Naturalist, 122, 661–696. 284 REFERENCES Cooper J. P. (ed.) (1975). Photosynthesis and productivity in different environments. Cambridge University Press, Cambridge. Cooper S. M. (1991). Optimal hunting group size: the need for lions to defend their kills against loss to spotted hyaenas. African Journal of Ecology, 29, 130–136. Cotterill F. P. D. (2003). Insights into the taxonomy of the tsessebe antelopes Damalis- cus lunatus (Bovidae: Alcelaphini) with the description of a new evolutionary spe- cies in south-central Africa. Durban Museum Novitates 28, 11–30. Cozzi G., Broekhuis M. C., Nutt J. W., and Schmid B. (2013). Comparison of the effects of artificial and natural barriers on large African carnivores: implications for interspecific relationships and connectivity. Journal of Animal Ecology, 82, 707–715. Craig G. C. (2004). Monitoring the illegal killing of elephants. Aerial Survey Standard for the MIKE programme, MIKE Technical Advisory Group. Crawley, M. J. (1983). Herbivory. The dynamics of animal-plant interactions. Blackwell Scientific Publications, Oxford. Creel S. and Creel N. M. (1996). Limitation of African wild dogs by competition with larger carnivores. Conservation Biology, 10, 526–538. Creel S. and Creel N. M. (2002). The African wild dog. Behaviour, ecology and conser- vation. Princeton University Press, Princeton, NJ. Creel S., Becker M. S., Durant S. M., Soka J. M., Matandiko W. Dickman A. J., Christianson D., Dröge E., Mweetwa T., Pettorelli N., Rosenblatt E., Schuette P., Woodroffe R., Bashir S., Beudels-Jamar R. C., Blake S., Borner M., Breitenmoser C., Broekhuis F., Cozzi G., Davenport T. R. B., Deutsch J., Dollar L., Dolrenry S., Douglas-Hamilton I., Fitzherbert E., Foley C., Hazzah L., Henschel P., Hilborn R., Hopcraft J. G. C., Ikanda D., Jacobson A., Joubert B., Joubert D., Kelly M. S., Lichtenfeld L. Mace G. M., Milanzi J., Mitchell N., Msuha M., Muir R., Nyahongo J., Pimm S., Purchase G., Schenck C., Sillero-Zubiri C., Sinclair A. R. E., Songorwa A. N., Stanley-Price M., Tehou C. A., Trout C., Wall J., Wittemyer G., and Zimmer- mann A. (2013). Conserving large populations of lions—the argument for fences has holes. Ecology Letters, 16, 1413–e3. Croft D. P., James R., and Krause J. (2008).Exploring animal social networks. Prince- ton University Press, Princeton, NJ. Croft D. P., Madden J. R., Franks D. W., and James R. (2011). Hypothesis testing in animal social networks. Trends in Ecology and Evolution, 26, 502–507. Croze H. (1974a). The Seronera bull problem. Part 1. The elephants.East African Wildlife Journal, 12, 1–28. Croze H. (1974b). The Serenera bull problem. Part 2. The trees.East African Wildlife Journal, 12, 29–48. Dabadghao P. M. and Shankarnarayan K. A. (1973). The grass cover of India. Indian Council of Agricultural Research, New Delhi. Dagg A. I. (2014). Giraffe. Biology, behaviour and conservation. Cambridge University Press, Cambridge. Dangerfield J. M. (1990). The distribution and abundance of Cubitermes sankurensis (Wassmann) (Isoptera; Termitidae) within a Miombo woodland site in Zimbabwe. African Journal of Ecology, 28, 15–20. Dankwa-Wiredu B. and Euler D. L. (2002). Bushbuck (Tragelaphus scriptus Pallas) habitat in Mole National Park, northern Ghana. African Journal of Ecology, 40, 35–41. Darwin C. (1859). On the origin of species, p. 236. John Murray, London. REFERENCES 285 Davies-Mostert H. T., Mills M. G. L., and Macdonald D. W. (2013). Hard boundaries influence African wild dogs’ diet and prey selection.Journal of Applied Ecology, 50, 1358–1366. Dean W. R. J. (2000). The birds of Angola: an annotated checklist. British Ornitholo- gists’ Union, Peterborough. Dean W. R. J., Anderson M. D., Milton S. J., and Anderson T. A. (2002). Avian assem- blages in native Acacia and alien Prosopis drainage line woodland in the Kalahari, South Africa. Journal of Arid Environments, 51, 1–19. De Boer W. F., Ntumi C. P., Correia A. U., and Mafuca, J. M. (2000). Diet and distri- bution of elephant in the Maputo Elephant Reserve, Mozambique. African Journal of Ecology, 38, 188–201. Delany M. J. and Happold D. C. D. (1979). Ecology of African land mammals. Long- man, London. Demment M. W. and Van Soest P. J. (1985). A nutritional explanation for body-size patterns of ruminant and nonruminant herbivores. American Naturalist, 125, 641–672. Deodatus E. D. and Lipiya A. K. (1991). Wildlife pest impact around Kasungu N. P. FP: MLW/87/010 Field Document No. 10, FAO, Malawi. Deodatus E.D. and Sefu L. (1992). National survey of wildlife pests. FO: MLW/87/010 Field Document No. 24. FAO, Malawi. Deshmukh I. K. (1984). A common relationship between precipitation and grass- land peak biomass for East and southern Africa. African Journal of Ecology, 22, 181–186. de Silva S. and Wittemyer G. (2012). A comparison of social organization in Asian elephants and African savannah elephants. International Journal of Primatology, 33, 1125–1141. Dharani N. (2006). Field guide to acacias of East Africa. Struik Publishers, Cape Town. Diamond J. M. (1975). Assembly of species communities. In M. L. Cody and J. M. Diamond(ed.) Ecology and evolution of communities, pp 341–444, Belknap Press, Cambridge, MA. Diamond J. M. (1986). Overview: laboratory experiments, field experiments, and natural experiments. In J. Diamond and E. J. Case (ed.) Community ecology, pp 3–22. Harper & Row, New York. Dickman C. R. (1992). Commensal and mutualistic interactions between terrestrial vertebrates. Trends in Ecology and Evolution, 7, 194–197. Dorst J. and Dandelot P. (1972). Larger mammals of Africa. Harper Collins, London. Douglas-Hamilton I, Bhalla S, Wittemyer G, and Vollrath F. (2006). Behavioural reac- tions of elephants towards a dying and deceased matriarch. Applied Animal Behav- ioural Science, 100, 87–102. Drent R. H. and van der Wal R. (1999). Cyclic grazing in vertebrates and the manipu- lation of the food resource. In H. Olff, V. K. Brown, and R. H.Drent (ed.)Herbi - vores: between plants and predators, pp. 271–299. Blackwell Science, Oxford. Dublin H. T. (1995). Vegetation dynamics in the Serengeti-Mara ecosystem: the role of elephants, fire and other factors. In A. R. E. Sinclair and P. Arcese (ed.)Serengeti II: dynamics, management and conservation of an ecosystem, pp. 71–90. University of Chicago Press, Chicago. Dublin H. T., Sinclair A. R. E., Boutin S., Anderson E., Jago M., and Arcese P. (1990a). Does competition regulate ungulate populations? Further evidence from Serengeti, Tanzania. Oecologia, 82, 283–288. 286 REFERENCES Dublin H. T., Sinclair A. R. E., and McGlade J. (1990b). Elephants and fire as causes of multiple stable states in the Serengeti-Mara woodlands. Journal of Animal Ecology, 59, 1147–1164. Ducheyne E., Mweempwa C., De Pus C., Vernieuwe H., De Deken, R., Hendricks G., and Van den Bossche P. (2009). The impact of habitat fragmentation on tsetse abundance on the plateau of eastern Zambia. Preventive Veterinary Medicine, 91, 11–18. Duff A. and Lawson A. (2004).Mammals of the world: a checklist. Yale University Press, New Haven, CT. Duncan P. (1975). Topi and their food supply. PhD thesis, University of Nairobi. Dunham K. M. (2012). Trends in populations of elephant and other large herbivores in Gonarezhou National Park, Zimbabwe, as revealed by sample aerial surveys. African Journal of Ecology, 50, 476–488. Dunham K. M. and Du Toit A. J. (2013). Using citizen-based survey data to deter- mine densities of large mammals: a case study from Mana Pools National Park, Zimbabwe. African Journal of Ecology, 51, 431–440. Durant S. M. (1998). Competition refuges and coexistence: an example from Seren- geti carnivores. Journal of Animal Ecology, 67, 370–386. Du Toit J. T. (1990). Feeding-height stratification among African browsing rumin- ants. African Journal of Ecology, 28, 55–61. Du Toit J. T. (2003). Large herbivores and savanna heterogeneity. In J. T. Du Toit, K. H. Rogers, and H. C. Biggs (ed.), The Kruger experience. Ecology and manage- ment of savanna heterogeneity, pp. 292–309. Island Press, Washington, DC. Du Toit J. T., Provenza F. D., and Nastis A. (1991). Conditioned taste aversion: how sick must a ruminant get before it learns about toxicity in food? Applied Animal Behaviour Science, 30, 35–46. DuToit J. T., Rogers K. H., and Biggs H. C. (ed.) (2003). The Kruger experience. Ecology and management of savanna heterogeneity. Island Press, Washington, DC. Earnshaw A. and Emerton L. (2000). The economics of wildlife tourism: theory and reality for landholders in Africa. In: T. T. Dolan, J. G. Grootenhuis, and H. Prins (ed.) Wildlife conservation by sustainable use, pp. 315–334. Kluwer, Dordrecht. East R. (1983). Application of species-area curves to African savannah reserves. Afri- can Journal of Ecology, 21, 123–128. East R. (1984). Rainfall, soil nutrient status and biomass of large African savannah mammals. African Journal of Ecology, 22, 245–270. East R. (1999). African antelope database 1998. IUCN/SSC Antelope Specialist Group. IUCN, Gland, Switzerland. Edkins M. T., Kruger L. M., Harris K., and Midgley J. J. (2007). Baobabs and elephants in Kruger National Park: nowhere to hide. African Journal of Ecology, 46, 119–125. Elliott J. P., McTaggart-Cowan I., and Holling C. S. (1977). Prey capture by the Afri- can lion. Canadian Journal of Zoology, 55, 1811–1828. Eloff F. C. (1984). Food ecology of the Kalahari lion Panthera leo vernayi. Koedoe, 1984 (Suppl.), 249–258. Elton C. S. (1958). The ecology of invasion by animals and plants. Methuen, London. Elton C. S. and Nicholson A. J. (1942). The ten-year cycle in numbers of the lynx in Canada. Journal of Animal Ecology, 11, 215–244. Eltringham S. K. (1974). Changes in the large mammal community of Mweya Penin- sula, Rwensori National Park, Uganda, following removal of hippopotamus. Jour- nal of Applied Ecology, 11, 855–866. REFERENCES 287 Eltringham K. S., Cooksey I. A., Dixon W. J. B., Raine N. E., Sheldrick C. J., McWil- liam N. M., and Packer M. J. (1999). Large mammals of Mkomazi. In M. J. Coe et al. (ed.) Mkomazi: the ecology, biodiversity and conservation of a Tanzanian savanna, pp. 485–504. Royal Geographical Society, London. Estes R. D. (1976). The significance of breeding synchrony in the wildebeest.East African Wildlife Journal, 14, 135–152. Estes R. D. (1991). The behaviour guide to African mammals. University of California Press, Berkeley, CA. Fant G. (1960). Acoustic theory of speech production. Mouton, The Hague. Ferdinands K., Beggs K., and Whitehead P. (2005). Biodiversity and invasive grass species: multiple use or monoculture? Wildlife Research, 32, 447–457. Ferrar A. A. and Walker B. H. (1974). An analysis of herbivore/habitat relationships in Kyle National Park, Rhodesia. Journal of the South African Management Associ- ation, 4, 137–147. Ferreira S. M. and Aarde R. J. (2009). Aerial survey intensity as a determinant of esti- mates of African elephant population sizes and trends. South African Journal of Wildlife Research, 39,181–191. Ferreira S. A. and Pienaar D. (2011). Degradation of the crocodile population in the Olifants River Gorge of Kruger National Park, South Africa. Aquatic Conservation: Marine and Freshwater Ecosystems, 21, 155–164. Field C. R. (1968). A comparative study of the food habits of some wild ungulates in the Queen Elizabeth Park, Uganda: preliminary report. In M. A. Crawford (ed.) Comparative nutrition of wild animals, pp. 135–151. Symposium No. 21, Zoologi- cal Society, London. Fischer F. and Linsenmair K. E. (2001). Spatial and temporal habitat use of kob ante- lopes (Kobus kob kob Erxleben 1777) in the Comoé National Park, Ivory Coast as revealed by radio tracking. African Journal of Ecology, 39, 249–256. FitzGibbon C. D. (1989). A cost to individuals with reduced vigilance in groups of Thomson’s gazelle hunted by cheetahs. Animal Behaviour, 37, 508–510. Foguekem D., Tchamba M. N., and Omondi P. (2010). Aerial survey of Elephants (Loxodonta africana africana), other large mammals and human activities in Waza National Park, Cameroon. African Journal of Environmental Science and Technol- ogy, 4, 401–411. Ford J. (1971). The role of trypanosomiases in Africa ecology. Clarendon Press, Oxford. Ford J. and Clifford H. R. (1968). Changes in the distribution of cattle and of bovine trypanosomiasis associated with the spread of tsetse flies (Glossina) in south-west Uganda. Journal of Applied Ecology, 5, 301–337. Foster J. B. and Kearney D. (1967). Nairobi National Park game census, 1966. East African Wildlife Journal, 5, 112–120. Foster J. B. and McLaughlin R. (1968). Nairobi National Park game census, 1967. East African Wildlife Journal, 6, 152–154. Freeland W. J. (1990). Large herbivorous mammals: exotic species in northern Australia. Journal of Biogeography, 17, 445–449. Freeland W. J. (1991). Large herbivorous mammals: exotic species in northern Aus- tralia. In P. A. Werner (ed.), Savanna ecology and management, pp. 101–105. Black- well Scientific Publications, Oxford. Frost G. H. and Bond I. (2008). The CAMPFIRE programme in Zimbabwe: payments for wildlife services. Ecological Economics, 65, 776–787. Fry C. H. and Keith S. (2004). The birds of Africa, vol. VII. Christopher Helm, London. 288 REFERENCES Fryxell J. M. (1995). Aggregation and migration by grazing ungulates in relation to resources and predators. In A. R. E. Sinclair and P. Arcese (ed.), Serengeti II: dynamics, management and conservation of an ecosystem, pp. 257–273. University of Chicago Press, Chicago. Fryxell J. M., Abrams P. A., and Holt R. D. (2008). Spatial dynamics and coexistence of the Serengeti grazer community. InA. R. E. Sinclair et al. (ed). Serengeti III. Human impacts on ecosystem dynamics. University of Chicago Press, Chicago. Gadgil M. and Meher Homji V. M. (1985). Land use and productive potential of Indian savanna. In J. C. Tothill and J. J. Mott (ed.), Ecology and management of world savannas, pp. 107–113. Australian Academy of Sciences, Canberra. Galvin K. A., Reid R. S., Behnke R. H., and Hobbs N. T. (ed.) (2007). Fragmenta- tion of semi-arid and arid landscapes: consequences for human and natural systems. Springer, Dordrecht. Garstang M., Larom D., Raspet R., and Lindeque M. (2004). Long-distance, low fre- quency elephant communication. Journal of Comparative Physiology, 190, 791–805. Gaston K. J. (1996). Species richness: measure and measurement. In K. J. Gaston (ed.), Biodiversity. A biology of numbers and difference, pp. 77–113. Blackwell Sci- ence, Oxford. Geertsema A. (1981). The servals of Gorigor. Wildlife News, 16, 4–8. Georgiadis N. J., Hack M., and Turpin, K. (2003). The influence of rainfall on zebra population dynamics: implications for management. Journal of Applied Ecology, 40, 125–136. Georgiadis N. J., Olwero N., Ojwang G., and Aike G. (2011). Reassessing aerial sam- ple surveys for wildlife monitoring, conservation, and management. In Georgi- adis N. J. (ed). Conserving wildlife in African landscapes: Kenya’s Ewaso ecosystem, pp. 1–10. Smithsonian Contributions to Zoology 632, Smithsonian Institution Scholarly Press, Washington DC. Glickman S. E., Cunha G. R., Drea C. M., Conley A. J., and Place N. J. (2006). Mam- malian sexual differentiation: lessons from the spotted hyena.Trends in Endocrin- ology and Metabolism, 17, 349–355. Global Invasive Species Database Grinnell J. and McComb K. (2001). Roaring and social communication in African lions: the limitations imposed by listeners. Animal Behaviour, 62, 93–98. Groves C. P. (2005). Order Primates. In: D. E. Wilson and D. M. Reeder (ed.), Mam- mal species of the world, pp. 111–184. The Johns Hopkins University Press, Balti- more, MD. Grubb P., Butynski T. M., Oates J. F., Bearder S. K., Disotell T. R., Groves C. P. and Struhsaker T. T. (2003). Assessment of the diversity of African primates. Interna- tional Journal of Primatology, 24, 1301–1357. Grzimek B. and Grzimek M. (1960a). Serengeti shall not die. Hamish Hamilton, London. Grzimek M. and Grzimek B. (1960b). Census of plains animals in the Serengeti National Park, Tanganyika. Journal of Wildlife Management, 24, 27–37. Gwitira I., Murwira A., Shekede M. D., Masocha M., and Chapano C. (2014). Precipi- tation of the warmest quarter and temperature of the warmest month are key to understanding the effect of climate change on plant species diversity in Southern African savannah. African Journal of Ecology, 52, 209–216. Gwynne M. D. and Bell R. H. V. (1968). Selection of grazing components by grazing ungulates in the Serengeti National Park. Nature (London), 220, 390–393. Haldane J. B. S. (1949). Disease and evolution. Ricerca Science Supplement A, 19, 68–76. Hanan N. P. and Lehmann E. R. (2011). Tree-grass interactions in savannas: para- digms, contradictions, and conceptual models. In M. J. Hill and N. P. Hanan (ed), Ecosystem function in savannas. Measurement and modelling at landscape to global scales. CRC Press, Boca Raton, FL. Happold D. and Lock J. M. (2013). The biotic zones of Africa. In: J.Kingdon et al. (ed.), Mammals of Africa. Volume I: Introductory chapters and Afrotheria. Bloomsbury, London. Harcourt A. H., Parks S. A., and Woodroffe R. (2001). Human density as an influ- ence on species-area relationships: double jeopardy for small African reserves? Biodiversity and Conservation, 10, 1001–1026. Hart B. L., Hart L., and Pinter-Wollman N. (2008). Large brains and cognition: Where do elephants fit in? Neuroscience and Biobehavioural Reviews, 32, 86–98. Hassan S. N., Rusch G. M., Hytteborn H., Skarpe C., and Kikula I. (2007). Effects of fire on sward structure and grazing in western Serengeti, Tanzania. African Journal of Ecology, 46, 174–185. Haxeltine A. and Prentice I. C. (1996). An equilibrium terrestrial biosphere model based on ecophysiological constraints, resource availability and competition among plant functional types. Global Biogeochemical Cycles, 10, 693–709. Heath B. (2000). Ranching: an economic yardstick. In H. H. T. Prins, J. G. Grootenhu- is, and T. T. Dolan (ed.) Wildlife conservation by sustainable use. Kluwer Academic Publications, Wageningen. Helm C. V. and Witkowski E. T. F. (2012). Continuing decline of a keystone species in the Kruger National Park, South Africa. African Journal of Ecology, 51, 270–279. Helsa B. I., Tieszen H. L., and Boutton T. W. (1985). Seasonal water relations of savanna shrubs and grasses in Kenya, East Africa. Journal of Arid Environments, 8, 15–31. Hensman M. C., Owen-Smith N., Parrini F., and Bonyongo M. C. (2014a). Home range occupation and habitat use of sable antelope in the Okavango Delta region of northern Botswana. African Journal of Ecology, 52, 237–245. 290 REFERENCES Hensman M. C., Owen-Smith N., Parrini F., and Bonyongo M. C. (2014b). Resource use and the nutritional status of sable antelope in the Okavango Delta region of northern Botswana. African Journal of Ecology, 52, 265–273. Herbst C. T., Stoeger A. S., Frey R., Lohshcheller J., Titze I. R., Gumpenberger M., and Fitch W. T. (2012). How low can you go? Physical production mechanism of elephant infrasonic vocalizations. Science, 337, 595–599. Herlocker D. (1975). Woody vegetation of the Serengeti National Park. Texas A&M University, College Station, TX. Hilborn R., Georgiadis N., Lazarus J., Fryxell J. M., Broten M. D., Mbano B. N. N., Murray M. G., Sinclair A. R. E., Durant S. M., Mwasaga B., Maige M. K. S., Arcese P., Albon S., Hofer H., Kapela M., Dobson A., East M., Nkya H., Dublin H. T., Packer C., Campbell K. L. I., Gascoyne S. C., Creed S. R., Hertz P., Creel N. M., and Caro T. M. (1995). A model to evaluate alternative management policies for the Serengeti-Mara ecosystem. In A. R. E. Sinclair and P. Arcese (ed.) Serengeti II: dynamics, management and conservation of an ecosystem, pp. 617–637. University of Chicago Press, Chicago. Hilborn R. and Sinclair A. R. E. (1979). A simulation of the wildebeest population, other ungulates, and their predators. In A. R. E. Sinclair and M. Norton-Griffiths (ed.) Serengeti: dynamics of an ecosystem, pp. 287–309. University of Chicago Press, Chicago. Hill M. J., Román M. O., and Schaaf C. B. (2011). Biogeography and dynamics of global tropical and subtropical savannas: A spatiotemporal view. In: M. J.Hill and N. P.Hanan (ed.), Ecosystem function in savannas. Measurement and modelling at landscape to global scales. CRC Press, Boca Raton, FL. Hirst S. M. (1969). Populations in a Transvaal lowland nature reserve. Zoologica Afri- cana, 4, 199–230. Hitchins P. M. (1968). Some preliminary findings on the population structure and status of the black rhinoceros Diceros bicornis in Hluluwe Game Reserve, Zulu- land. Lammergeyer, 9, 26–28. Hitchins P. M. (1969). Influence of vegetation types on sizes of home ranges of black rhino in Hluluwe Game Reserve. Lammergeyer, 3, 81085. Hoare R. (1995). Options for the control of elephants in conflict with people.Pachy - derm, 19, 54–63. Hofmann R. R. (1968). Comparisons of the rumen and omasum structure in East African game ruminants in relation to their feeding habits. In M.A. Crawford (ed.) Comparative nutrition of wild animals, pp. 179–194. Symposium No. 21, Zoologi- cal Society, London. Hofmann R. R. (1973). The ruminant stomach. East African Literature Bureau, Nairobi. Hoffmann A. and Klingel H. (2001). Spatial and temporal patterns in Lemniscomys striatus (Linnaeus 1758) as revealed by radio-tracking. African Journal of Ecology, 39, 351–356. Holdo R. M., Fryxell J. M., Sinclair A. R. E., Dobson A., and Holt R. D. (2011). Pre- dicted impact of barriers to migration on the Serengeti wildebeest migration. PLoS ONE, 6, e16370. Holdo R. M., Sinclair A. R. E., Dobson A. P., Metzger K. L., Bolker B. M., Richie M. E., and Holt R. D. (2009). A disease-mediated trophic cascade in the Serengeti and its implications for ecosystem C. PLoS Biology. 7, e10000210. REFERENCES 291 Holling C. S. (1959). The components of predation as revealed by a study of small- mammal predation of the European pine sawfly.Canadian Entomologist, 91, 293–320. Holmern T. and Røskaft E. (2014). The poultry thief: subsistence farmers’ perceptions of depredation outside the Serengeti National Park, Tanzania. African Journal of Ecology, 52, 334–342. Hooft P., Greyling B. J., Prins H. H. T., Getz W., Jolles A., Bastos A. D. S. (2007). Selec- tion at the Y chromosome of the African buffalo driven by rainfall. PLOS One, 2, e1086. Hooft P. van, Prins, H. H. T., Getz W. M, Jolles A. E., Wieren S E van, Greyling B. J., Helden P. D. van, and Bastos A. D. S. (2010). Rainfall-driven sex-ratio genes in African buffalo suggested by correlations between Y-chromosomal haplotype frequencies and foetal sex ratio. BMC Evolutionary Biology, 10, 106. Hopkins B. (1962). Vegetation of the Olokemeji Forest Reserve, Nigeria. I. General features of the reserve and the research sites. Journal of Ecology, 50, 559–598. Hoppe P. P., Qvortrup S. A., and Woodford M. H. (1977). Rumen fermentation and food selection in East African zebu cattle, wildebeest, Coke’s hartebeest and topi. Journal of Zoology, 181, 1–9. Houston D. C. (1979). The adaptations of scavengers. In ARE Sinclair and M Norton- Griffiths (ed.),Serengeti. Dynamics of an ecosystem, pp. 263–286. University of Chi- cago Press, Chicago. Hovestadt T., Joachim Poethke H., and Linsenmair K. E. (2005). Spatial patterns in species–area relationships and species distribution in a West African forest– savanna mosaic. Journal of Biogeography, 32, 677–684. Huntley B. J. (1978). Ecosystem conservation in southern Africa. In M. J. A. Werger (ed.), Biogeography and ecology of southern Africa, pp. 1333–1384. W. Junk, The Hague. Huntley B. J. (1982). Southern African savannas. In: B. J. Huntley and B. H. Walker (ed.), Ecology of tropical savannas, pp 101–119. Springer-Verlag, Berlin. Huntley B. J. and Walker B. H. (ed.) (1982). Ecology of tropical savannas. Springer- Verlag, Berlin. Hutchinson G. E. (1959). Homage to Santa Rosalia, or why are there so many kinds of animals? American Naturalist, 93, 145–159. Irish J. (1994). The biomes of Namibia, as determined by objective categorisation. Navorsinge van die Nasionale Museum, Bloemfontein, 10, 549–592. Jachmann H. (1991). Evaluation of four survey methods for estimating elephant den- sities. African Journal of Ecology, 29, 188–195. Jackson C. H. N. (1933). On the true density of tsetse flies. Journal of Animal Ecology, 2, 204–209. Janzen D. H. (1966). Coevolution of mutualism between ants and Acacia in Central America. Evolution, 20, 249–275. Jarman P. J. (1971). Diets of large mammals in the woodlands around Lake Kariba, Rhodesia. Oecologia (Berlin), 8, 157–178. Jarman P. J. (1972). Seasonal distribution of large mammal populations in the unflooded middle Zambezi valley. Journal of Applied Ecology, 9, 283–299. Jarman P. J. (1974). The social organisation of antelope in relation to their ecology. Behaviour, 48, 215–267. 292 REFERENCES Joy D., Feng X., Mu J., Furuya T., Chotivanich K., Krettli A. U., Ho M., Wang A, White N. J., Suh E., Beerli P., and Su X. (2003). Early origin and expansion of Plasmodium falciparum. Science, 300, 318–321. Kamler J. F., Stenkewitz U., and Macdonald D. W. (2013). Lethal and sublethal effects of black-backed jackals on cape foxes and bat-eared foxes. Journal of Mammalogy, 94, 295–306. Kassa B., Libois R., and Sinsin, B. (2008). Diet and food preference of the waterbuck (Kobus ellipsiprymnus defassa) in the Pendjari National Park, Benin. African Jour- nal of Ecology, 46, 303–310. Kat P. W., Alexander K. A., Smith J. S., and Munson L. (1995). Rabies and African wild dogs in Kenya. Proceedings of the Royal Society of London, Series B, 262, 229–233. Kebede A. T. (2009). Sustaining the Allideghi grassland of Ethiopia: influence of pastor- alism and vegetation change. PhD thesis, Utah State University, Logan. Kesch K. M., Bauer D. T., and Loveridge A. J. (2014). Undermining game fences: who is digging holes in Kalahari sands? African Journal of Ecology, 52, 144–150. Kihwele E., Mnaya B., Meng’ataki G., Birkett C., and Wolanski E. (2012). The role of vegetation in the water budget of the Usangu wetlands, Tanzania. Wetlands Ecolo- logical Management, 20, 389–398. King L. E., Lawrence A., Douglas-Hamilton I., and Vollrath F. (2009). Beehive fence deters crop-raiding elephants. African Journal of Ecology, 47, 131–137. Kingdon J. (1977). East African mammals. IIIa (carnivores). University of Chicago Press, Chicago. Kingdon J. (1989). Island Africa. The evolution of Africa’s rare animals and plants. Princeton University Press, Princeton, New jersey. Kingdon J. (1997). The Kingdon field guide to African mammals. Academic Press, San Diego, CA. Kingdon J. (2004). The Kingdon pocket guide to African mammals. A&C Black, London. Kingdon J., Happold D., Hoffmann M., Butynski T., Happold M., and Kalina J. (2013). Mammals of Africa. Volume 1 Introductory Chapters and Afrotheria. Bloomsbury, London. Klare U., Kamler J. F., Stenkewitz U., and Macdonald D. W. (2010). Diet, prey selec- tion, and predation impact of black-backed jackals in South Africa. Journal of Wildlife Management, 74, 1030–1042. Klein R. G. (1988). The archaeological significance of animal bones from Acheulean sites in southern Africa. African Archeological Review, 6, 3–25. Klemuk S. A., Riede T., Walsh E. J., and Titze I. R. (2011). Adapted to roar: functional morphology of tiger and lion vocal folds. PLOS ONE 6, e27029. Knoop W. T. and Walker B. H. (1985). Interactions of woody and herbaceous vegeta- tion in southern African savanna. Journal of Ecology, 73, 235–253. Krause B. L. (2012). The great animal orchestra. Little, Brown and Company, New York. Krause J., Croft D. P., and James R. (2007). Social network theory in the behavioural sciences: potential applications. Behavioural Ecology and Sociobiology, 62, 15–27. Krause J. and Ruxton G. D. (2002). Living in groups. Oxford University Press, Oxford. Krebs C. J. (1999). Ecological methodology. Benjamin Cummings, San Francisco, CA. Krebs C. J., Boutin S., Boonstra R., Sinclair A. R. E., Smith J. N. M., Dale M. R. T., Martin K.. and Turkington R. (1995). Impact of food and predation on the snow- shoe hare cycle. Science, 269, 1112–1115. REFERENCES 293 Krüger S. C., Allan D. G., Jenkins A. R., and Amar A. (2014). Trends in ter- ritory occupancy, distribution and density of the Bearded Vulture Gypaetus barbatus meridionalis in southern Africa. Bird Conservation International, 24, 162–177. Kruuk H. (1972). The spotted hyaena. University of Chicago Press, Chicago. Kyalangalilwa B., Boatwright J. S., Daru B. H., Maurin O., and Van der Bank M. (2013). Phylogenetic position and revised classification of Acacia s.l. (Fabaceae: Mimosoideae) in Africa, including new combinations in Vachellia and Senegalia. Botanical Journal of the Linnean Society, 172, 500–523. Lack D. (1969). The numbers of bird species on islands. Bird Study, 16, 193–209. Lack D. (1976). Island birds. Blackwell Scientific Publications, Oxford. Lamprey H. F. (1963). Ecological separation of the large mammal species in the Tarangire Game Reserve, Tanganyika. East African Wildlife Journal, 1, 63–92. Lawick H. van (1973). Solo. The story of an African wild dog. Collins, London. Lawick H. van and Lawick-Goodall J. van (1970). Innocent killers. Collins, London. Laws R. M. (1966). Age criteria for the African elephant Loxodonta africana. East African Wildlife Journal, 4, 1–37. Laws R. M. (1970). Elephants as agents of habitat and landscape change in East Africa. Oikos, 21, 1–15. Lebrum J. (1960). Sur la richesse de la flore de divers territoires africains. Bulletin des Séances de l’Academie Royale des Sciences d’Outre-Mer, NS 6, 669–690. Le Roux X., Bariac T., and Mariotti A. (1995). Spatial partitioning of the soil water resource between grass and shrub components in a West African humid savanna. Oecologia, 104, 147–155. Leuthold W. (1966). Variation in territorial behaviour of Ugandan kob, Adenota kob thomasi (Neumann 1896). Behaviour, 27, 214–257. Leuthold W. and Sale J. B. (1973). Movements and patterns of habitat utilization in Tsavo National Park, Kenya. East African Wildlife Journal, 11, 369–384. Lightfoot C. J. and Norval R. A. (1981). Tick problems in wildlife in Zimbabwe 1. The effects of tick parasitism on wild ungulates. South African Journal of. Wildlife Research, 11, 41–45. Lindeque M. and Lindeque P. M. (1991). Satellite tracking of elephants in northwest- ern Namibia. African Journal of Ecology, 29, 196–206. Linderman R. L. (1942). The trophic-dynamic aspect of ecology. Ecology, 23, 399–418. Lisein, J., Linchant, J., Lejeune, P., Bouché P., and Vermeulen, C. (2013). Aerial sur- veys using an Unmanned Aerial System (UAS): comparison of different methods for estimating the surface area of sampling strips. Tropical Conservation Science, 6, 506–520. Mabunda D., Pienaar, D. J., and Verheof J. (2003). The Kruger National Park: a century of management and research. In J. T. Du Toit, K. H. Rogers, and H. C Biggs. The Kruger Experience: ecology and management of savannah heterogeneity, pp. 3–21. Island Press, Washington, DC. Macandza V. A., Owen-Smith N., and Cain, J. W. III (2012a). Habitat and resource partitioning between abundant and relatively rare grazing ungulates. Journal of Zoology, 287, 175–185. Macandza V. A., Owen-Smith N., and Cain J. W. III (2012b). Dynamic spatial par- titioning and coexistence among tall grass grazers in an African savanna. Oikos, 121, 891–898. 294 REFERENCES Macandza V., Owen-Smith N., and Le Roux E. (2014). Faecal nutritional indicators in relation to the comparative population performance of sable antelope and other grazers. African Journal of Ecology, 52, 300–307. MacArthur R. H. (1955). Fluctuations of animal populations and a measure of com- munity stability. Ecology, 36, 533–536. MacArthur R. H. and Levins R. (1967). The limiting similarity, convergence and divergence of coexisting species. American Naturalist, 101, 377–385. MacArthur R. H. and Wilson E. O. (1967). The theory of island biogeography. Prince- ton University Press, Princeton, NJ. MacDonald D. (2001). The new encyclopedia of mammals. Oxford University Press, Oxford. Mackenzie D. I. (2012). PRESENCE software (version 6.4). Available at: McNaughton S. J. (1977). Diversity and stability of ecological communities: a comment on the role of empiricism in ecology. American Naturalist, 111, 515–525. McNaughton S. J. (1979a). Grazing as an optimization process: grass-ungulate rela- tionships in the Serengeti. American Naturalist, 113, 691–703. McNaughton S. J. (1979b). Grassland-herbivore dynamics. In A. R. E. Sinclair and M. Norton-Griffiths (ed.), Serengeti. Dynamics of an ecosystem, pp. 46–81. Univer- sity of Chicago Press, Chicago. McNaughton S. J. (1985). Ecology of a grazing ecosystem: the Serengeti. Ecological Monographs, 55, 259–294. McNaughton S. J. and Georgiadis N. J. (1986). Ecology of African grazing and brows- ing mammals. Annual Review of Ecology and Systematics, 17, 39–65. Mduma S. A. R. (1996). Serengeti wildebeest population dynamics: regulation, limita- tion and implications for harvesting. PhD thesis, University of British Columbia. Mduma S. A. R., Hilborn R., and Sinclair A. R. E. (1998). Limits to exploitation of Ser- engeti wildebeest and implications for its management. In D. M. Newbery, H. H. T. Prins, and N. Brown (ed.), Dynamics of tropical communities. 37th Symposium of the British Ecological Society, pp. 243–265. Blackwell Science, Oxford. Mduma S. A. R., Sinclair A. R. E., and Hilborn R. (1999). Food regulates the Serengeti wildebeest: a 40-year record. Journal of Animal Ecology, 68, 1101–1122. Mduma S. A. R., Sinclair A. R. E., and Turkington R. (2007). The role of rainfall and predators in determining synchrony in reproduction of savanna trees in Serengeti National Park, Tanzania. Journal of Ecology, 95, 184–196. Menaut J. C. (1971). Etude de quelques peuplements ligneux d’une savane guinéenne de Cote d’Ivoire. Thesis, University of Paris. Menaut J. C. (1983). The vegetation of African savannas. In F. Bourlière (ed.),Tropical savannas, pp. 109–150. Elsevier, Amsterdam. Miller M. F. (1995). Acacia seed survival, seed germination and seedling growth fol- lowing pod consumption by large herbivores and seed chewing by rodents. African Journal of Ecology, 33, 194–210. Miller R. I. and Harris L. D. (1977). Isolation and extirpation in wildlife reserves. Bio- logical Conservation, 12, 311–315. Mills M. G. L. (1984). Prey selection and feeding habits of the large carnivores in the southern Kalahari. Koedoe, 1984 (Suppl.), 281–294. Mills M. G. L. (1990). Kalahari hyaenas: comparative behavioural ecology of two spe- cies. Unwin Hyman, London. Mills M. G. L. and Biggs H. C. (1993). Prey apportionment and related ecological relationships between large carnivores and prey in the Kruger National Park. In N. Dunstone and M. L. Gorman (ed.), Mammals as predators, pp. 253–268. Sym- posium 65, Zoological Society, London. Mills M. G. L., Biggs H. C., and Whyte I. J. (1995). The relationship between rainfall, lion predation and population trends in African herbivores. Wildlife Research, 22, 75–88. Mills M. G. L. and Funston P. J. (2003). Large carnivores and savanna heterogene- ity. In J. T. DuToit, K. H. Rogers, and H. C. Biggs (ed.), The Kruger experience. Ecology and management of savanna heterogeneity, pp. 370–388. Island Press, Washington, DC. Mills M. G. L. and Hofer H. (compilers) (1998). Hyaenas. Status survey and conserva- tion action plan. IUCN/SSC Hyaena Specialist Group. IUCN, Gland, Switzerland. 296 REFERENCES Mills M. G. L. and Retief P. F. (1984). The response of ungulates to rainfall along the riverbeds of the southern Kalahari. Koedoe, 27, 129–141. Mills M. G. L. and Shenk TM. (1992). Predator-prey relationships: the impact of lion predation on wildebeest and zebra populations. Journal of Animal Ecology, 61, 693–702. Misra R. (1983). Indian savannahs. In F. Bourlière (ed.), Tropical savannas, vol. 13, pp. 151–166. Elsevier, Amsterdam. Mistry J. (2000). World savannas. Ecology and human use. Pearson Education, Harlow. Mitchell B. L., Skenton J. B., and Uye J. C. M. (1965). Predation on large mammals in the Kafue National Park, Zambia. Zoologica Africana, 1, 297–318. Mitchell J. and Ashley C. (2006). Can tourism help reduce poverty in Africa? Briefing Paper, Overseas Development Institute, London. Modise, S. (1996). Models of wildlife management: the Botswana experience, Afri- can Wildlife Policy Consultation , British Overseas Development Administration, London. Moe S. R., Rutina L. P., Hytteborn H., and Du Toit J. T. (2009). What controls wood- land regeneration after elephants have killed the big trees?Journal of Applied Ecol- ogy, 46, 223–230. Mooring M. and Mundy P. (1996). Interactions between impala and oxpeckers at Matobo National Park Zimbabwe. African Journal of Ecology, 34, 54–65. Mopipi K., Trollope W. S. W., and Scogings P. F. (2009). Effects of moisture, nitrogen, grass competition and simulated browsing on the survival and growth of Acacia karroo seedlings. African Journal of Ecology, 47, 680–687. Mordelet P., Menaut J. C., and Mariotti A. (1997). Tree and grass rooting patterns in an African humid savanna. Journal of Vegetation Science, 8, 65–70. Morris S., Humphreys D., and Reynolds D. (2006). Myth, marula and elephant: an assessment of voluntary ethanol intoxication of the African elephant (Loxodonta africana) following feeding on the fruit of the marula tree (Sclerocarya birrea). Physiological and Biochemical Zoology, 78, 363–369. Morrison T. A., Yoshizaki J., Nichols J. D., and Bolger D. T. (2011). Estimating sur- vival in photographic capture-recapture studies: overcoming misidentification error. Methods in Ecology and Evolution, 2,454–463. Mott J. J., Williams J., Andrew M. H.m and Gillison A. N. (1985). Australian savanna ecosystems. In J. C. Tothill and J. J. Mott (ed.), Ecology and management of the world’s savannas, pp. 56–82. Australian Academy of Sciences, Canberra. Murphy W. J., Pringle T. H., Crider T. A., Springer M. S., and Miller W. (2007). Using genomic data to unravel the root of the placental mammal phylogeny. Genome Research, 17, 414–421. Murray M. G. (1995). Specific nutrient requirements and migration of wildebeest. In A. R. E. Sinclair and P. Arcese (ed.), Serengeti II: dynamics, management and conservation of an ecosystem, pp. 231–256. University of Chicago Press, Chicago. Murray C. J. L., Rosenfeld L. C., Lim S. S., Andrews K. G., Foreman K. J., Haring D., Fullman N., Naghavi M., Lozano R., and Loez A. D. (2012). Global malaria mortal- ity between 1980 and 2010: a systematic analysis. Lancet, 379, 413–431. Muturi G. M., Poorter L., Mohren G. M. J., and Kigomo B. N. (2013). Ecological impacts of Prosopis species invasion in Turkwel riverine forest, Kenya. Journal of Arid Environments, 92, 89–97. Nelson F. (2007). Are large mammal declines in Africa inevitable? African Journal of Ecology, 46, 3–4. REFERENCES 297 Newman M. E. J. (2003). The structure and function of complex networks.SIAM Review, 45, 167–256. Newmark W. D. (1996). Insularization of Tanzanian parks and the local extinction of large mammals. Conservation Biology, 10,1549–1556. Newmark W. D., Manyanza D. N., Gamassa D. G. M., and Sariko H. I. (1994). The conflict between wildlife and local people living adjacent to protected areas in Tan- zania: human density as a predictor. Conservation Biology, 8, 249–255. Ngure N. (1995). People-elephant conflict management in Tsavo, Kenya.Pachyderm , 19, 20–25. Nikaido M., Rooney A. P., and Okada N. (1999). Phylogenetic relationships among cetartiodactyls based on insertions of short and long interspersed elements: Hip- popotamuses are the closest extant relatives of whales. Proceedings of the National Academy of Sciences of the U S A, 96, 10261–10266. Nix H. A. (1981). The environment of Terra Australis. In A. Keast (ed.), Ecological biogeography of Australia, pp. 103–133. Junk, The Hague. Noad T. and Birnie A. (1989). Trees of Kenya. Noad and Birnie, Nairobi. Norton-Griffiths M. (1973). Counting migratory wildebeest using two-stage sam- pling. East African Wildlife Journal, 11, 135–149. Norton-Griffiths M. (1978).Counting animals, 2nd ed. African Wildlife Leadership Foundation, Nairobi. Norton-Griffiths M. (1979). The influence of grazing, browsing, and fire on the vege- tation dynamics of the Serengeti. In A. R. E. Sinclair and M. Norton-Griffiths (ed.), Serengeti. Dynamics of an ecosystem, pp. 310–352. University of Chicago Press, Chicago. Norton-Griffiths M. (1995). Property rights and the marginal wildebeest—an eco- nomic analysis of wildlife conservation options in Kenya. Biodiversity and Conser- vation, 5, 1557–1577. Norval R. A., Sutherst R. W., Kurki J., Gibson J., and Kerr J. D. (1988). The effect of the brown ear tick Rhipicephalus appendiculatus on the growth of Sanga and European breed cattle. Veterinary Parasitology, 30, 149–164. Novellie P. A. and Knight M. (1994). Repatriation and translocation of ungulates into South African national parks: an assessment of past attempts. Koedoe, 37, 115–119. Nowell K. and Jackson P. (ed.) (1996). Wild cats. IUCN Status Survey and Conserva- tion Action Plan, IUCN, Gland, Switzerland. O’Connell A. F., Nichols J. D.,and Ullas K. K. (Eds.) (2011). Camera traps in animal ecology: methods and analyses. Springer, New York. O’Connell-Rodwell C. E. (2007). Keeping an ‘ear’ to the ground: seismic communica- tion in elephants. Physiology, 22, 287–294. O’Connell-Rodwell C. E., Wood J. D., Kinzley C., Rodwell T. C., Poole J. H., and Puria S. (2007). Wild African elephants (Loxodonta africana) discriminate between familiar and unfamiliar conspecific seismic alarm calls.Journal of the Acoustic Soci- ety of America, 122, 823–830. O’Connor T. G., Haines, L. M., and Snyman H. A. (2001). Influence of precipitation and species composition on phytomass of a semi-arid African grassland. Journal of Ecology, 89, 850–860. Ogutu J. O. and Owen-Smith N. (2005). Oscillations in large mammal popula- tions: are they related to predation or rainfall? African Journal of Ecology, 43, 332–339. 298 REFERENCES Ogutu J. O., Owen-Smith N., Piepo H. P., and Said M. Y. (2011). Continuing wildlife population declines and range contraction in the Mara region of Kenya during 1977–2009. Journal of Zoology, 285, 99–109. Ogutu J. O., Piepho H. P., Dublin H. T., Bhola N., and Reid R. S. (2007). El Nino-South- ern Oscillation, rainfall, temperature and Normalized Difference Vegetation Index fluctuates in the Mara-Serengeti ecosystem.African Journal of Ecology, 46, 132–143. O’Kane C. A. J., Duffy K. J., Page B. R., and Macdonald D. W. (2014). Model highlights likely long-term influences of mesobrowsers versus those of elephants on wood- land dynamics. African Journal of Ecology, 52, 192–208. Okitsu S. (2005). Factors controlling geographical distribution in savanna vegetation in Namibia. African Study Monographs, Suppl.30, 135–151. Oldeland J., Schmiedel U., and Jürgens N. (2010). Life forms along the BIOTA transects—new perspectives on biomes in southern Africa. In U. Schmiedel and N. Jürgens (ed.), Biodiversity in southern Africa. Volume 2: Patterns and processes at regional scale, pp. 136–144. Klaus Hess Publishers, Göttingen. Olff H. and Hopcraft J. G. C. (2008). The resource basis of human-wildlife interaction. In:, A. R. E. Sinclair et al. (ed.), Serengeti III. Human impacts on ecosystem dynam- ics. University of Chicago Press, Chicago. Olivier P. I., Ferreira S. M., and van Aarde R. J. (2009). Dung survey bias and ele- phant population estimates in southern Mozambique. African Journal of Ecology, 47, 202–213. Ottichilo W. K., De Leeuw J., Skidmore A. K., Prins H. H. T., and Said M. Y. (2000). Population trends of large non-migratory wild herbivores and livestock in the Masai Mara ecosystem, Kenya, between 1977 and 1997. African Journal of Ecology, 38, 202–216. Owen-Smith N. (1982). Factors influencing the consumption of plant products by large herbivores. In: B. J. Huntley and B. H. Walker (ed.), Ecology of tropical savan- nahs, Springer-Verlag, Berlin. Owen-Smith N. (1988). Megaherbivores. The influence of very large body size on ecol- ogy. Cambridge University Press, Cambridge. Owen-Smith N. (1990). Demography of a large herbivore, the greater kudu, in rela- tion to rainfall. Journal of Animal Ecology, 59, 893–913. Owen-Smith N. and Cumming D. H. M. (1993). Comparative foraging strategies of grazing ungulates in African savannah grasslands. Proceedings of the XVII Interna- tional Grassland Congress. Owen-Smith N. and Mason D. R. (2005). Comparative changes in adult vs. juvenile survival affecting population trends of African ungulates. Journal of Animal Ecol- ogy, 74, 762–773. Owen-Smith N., Mason D. R., and Ogutu J. O. (2005). Correlates of survival rates for 10 African ungulate populations: density, rainfall and predation. Journal of Animal Ecology, 74, 774–788. Owen-Smith N. and Ogutu J. O. (2003). Rainfall influences on ungulate population dynamics in the Kruger National Park. In J. T. DuToit, K. H. Rogers, and H. C. Biggs (ed.), The Kruger experience. Ecology and management of savanna heteroge- neity, pp. 310–331. Island Press, Washington, DC. Packer C. (1990). Serengeti lion survey. Report to TANAPA, SWRI, MWEKA, and the Game Department. Tanzania National Parks, Arusha. Packer C., Loveridge A., Canney S., Caro T., Garnett S. T., Pfeifer M., Zander K. K., Swanson A., MacNulty D., Balme G., Bauer H., Begg C. M., Begg K. S., Bhalla S., REFERENCES 299 Bissett C., Bodasing T., Brink H., Burger A., Burton A. C., Clegg B., Dell S., Delsink A., Dickerson T., Dloniak S. M., Druce D., Frank L., Funston P., Gichohi N., Groom R., Hanekom C., Heath B., Hunter L., DeIongh H. H., Joubert C. J., Kasiki S. M., Kissui B., Knocker W., Leathem B., Lindsey P. A., Maclennan S. D., McNutt J. M., Miller S. M., Naylor S., Nel P., Ng’weno C., Nicholls K., Ogutu J. O., Okot-Omoya E., Patterson B. D., Plumptre A., Salerno J., Skinner K., Slotow R., Sogbohossou E. A., Stratford K. J., Winterbach C., Winterbach H., and Polasky S. (2013). Conserv- ing large carnivores: dollars and fence. Ecology Letters, 16, 635–641. Paine R. T. (1966). Food web complexity and species diversity. American Naturalist, 100, 65–75. Paine R. T. (1980). Food webs: linkage, interaction strength and community infra- structure. Journal of Animal Ecology, 49, 667–685. Palmer T. M., Doak D. F., Stanton M. L., Bronstein J. L., Kiers E. T., Young T. P., Goheen J. R., and Pringle R. M. (2013). Synergy of multiple partners, including freeloaders, increase host fitness in a multispecies mutualism. Proceedings of the National Academy of Science of the U S A, 107, 17234–17239. Parker G., Sundaresan S., Chege G., and O’Brien T. (2011). Using sample aerial sur- veys to estimate the abundance of the endangered Grevy’s zebra in northern Kenya. African Journal of Ecology, 49, 56–61. Patterson B. D., Kays R. W., Kasiki S. M., and Sebestyen V. M. (2006). Developmental effects of climate on the lion’s mane (Panthera leo). Journal of Mammalogy, 87, 193–200. Pearsall W. H. (1957). Report on an ecological survey of the Serengeti National Park, Tanganyika. Oryx, 4, 71–136. Pellow R. (1977). Serengeti Research Institute, Annual Report, 1975–1976, pp. 49–67. Tanzania National Parks, Arusha. Pellow R. (1981). The giraffe (Giraffa camelopardalis tippelskirchi Matschie) and its Acacia food resource in the Serengeti National Park. PhD thesis, University of London. Pennycuick L. (1975). Movements of migratory wildebeest population in the Seren- geti area between 1960 and 1973. East African Wildlife Journal, 13, 65–87. Penzhorn B. L. (1971). A summary of the re-introduction of ungulates into South African National Parks (to 31 December 1970). Koedoe, 14, 145–159. Peters C. R., Blumenschine R. J., Hay R., Livingstone D. A., Marean C. W., Harri- son T., Armour-Chelu M., Andrews P., Bernor R., Bonnefille R., and Werdelin L. (2008). Paleoecology of the Serengeti-Mara Ecosystem. In A. R. E. Sinclair et al. (ed.), Serengeti III. Human impacts on ecosystem dynamics. University of Chicago Press, Chicago. Pfefferle D., West P. M., Grinnell J., Packer C., and Fischer J. (2007). Do acoustic features of lion, Panthera leo, roars reflect sex and male condition?Journal of the Acoustic Society of America, 121, 3947–3953. Pienaar U. de V. (1961). A seocnd outbreak of anthrax amongst game animals in the Kruger National Park. Koedoe, 4, 4–17. Pienaar U. de V. (1963). The large mammals of the Kruger National Park–their distri- bution and present-day status. Koedoe, 6, 1–263. Pienaar U. de V. (1967). Epidemiology of anthrax of wild animals and the control of anthrax epizootics in the Kruger National Park, South Africa. Federation Proceed- ings, 26, 1496–1502. Pienaar U. de V. (1969). Predator-prey relationships amongst the larger mammals of the Kruger National Park. Koedoe, 12, 108–176. 300 REFERENCES Pijanowski B. C.,Villanueva-Rivera L. J., Dumyahn S. L., Farina A., Krause B. L., Napoletano B. M., Gage S. H., and Pieretti N. (2011). Soundscape ecology: the sci- ence of sound in the landscape. BioScience, 61, 203. Pimm S. L. (1991). The balance of nature? University of Chicago Press, Chicago. Plantan T. B. (2009). Feeding behaviour of wild and captive oxpeckers (Buphagus spp.): a case of conditional mutualism. PhD thesis, University of Miami. Plantan T. B., Howitt M., Kotze A., and Gaines M. (2013). Feeding preferences of the red-billed oxpecker, Buphagus erythrorhynchus: a parasitic mutualist? African Journal of Ecology, 51, 325–336. Poole J. H., Payne K., Langbauer W.R., and Moss C. J. (1988). The social contexts of some very low frequency calls of African elephants. Behavioural Ecology and Sociobiology, 22, 385–392. Poole J. H., Tyack P. L., Stoeger-Horwath A. S., and Watwood S. (2005). Elephants are capable of vocal learning. Nature, 434, 455–456. Pratt D. J. and Gwynne M. D. (1977). Rangeland management and ecology in East Africa. Hodder & Stoughton, London. Prins H. H. T. (1996). Ecology and behaviour of the African buffalo. Social inequality and decision making. Chapman & Hall, London. Prins H. H. T. (2000). Competition between wildlife and livestock in Africa. In H. H. T. Prins, J. G. Grootenhuis, and T. T. Dolan (ed.), Wildlife conservation by sustain- able use. Kluwer Academic Publications, Wageningen. Prins H. H. T., de Boer W. F., van Oeveren H., Correia A., Mafuca J., and Olff H. (2006). Co-existence and niche segregation of three small bovid species in south- ern Mozambique. African Journal of Ecology, 44, 186–198. Prins H. H. T., Grootenhuis J. G., and Dolan T. T. (ed.) (2000). Wildlife conservation by sustainable use. Kluwer Academic Publications, Wageningen. Prins H. H. T. and Olff H. (1998). Species richness of African grazer assemblages: towards a functional explanation. In D. M. Newberry, H. H. T.Prins, and N. Brown (ed.), Dynamics of tropical communities, pp. 449–490. Blackwell Scientific, Oxford. Prins H. T. and Van Der Jeugd H. P. (1993). Herbivore population crashes and wood- land structure in East Africa. Journal of Ecology, 81, 305–314. Pritchard J. K., Stephens M., and Donnelly P. (2000). Inference of population struc- ture using multilocus genotype data. Genetics, 155, 945–959. Prugh, L. R., Stoner, C. J., Epps, C. W., Bean, W. T., Ripple, W. J., Laliberte, A. S., and Brashares, J. S. (2009). The rise of the mesopredator. BioScience, 59, 779–791. Pyšek P., Jarošik V., Hulme P. E., Pergl J., Hejda M., Schaffner U., and Vila M. (2012). A global assessment of invasive plant impacts on resident species, communities and ecosystems: the interaction of impact measures, invading species’ traits and environment. Global Change Biology 18, 1725–1421 1737. Rasmussen H. B., Wittemyer G., and Douglas-Hamilton I. (2006). Predicting time- specific changes in demographic processes using remote-sensing data. Journal of Applied Ecology, 43, 366–376. Raunkier C. (1934). The life forms of plants and statistical plant geography. Clarendon Press, Oxford. Redfern J. V., Viljoen P. C. Kruger J. M., and Getz W. M.. (2002). Biases in estimating population size from an aerial census: a case study in the Kruger National Park, South Africa. South African Journal of Science, 98, 455–461. Reed E. and Miranda M. (2007). Assessment of the mining sector and infrastructure development in the Congo basin region. WWF, Washington, DC. REFERENCES 301 Reid R. S. and Ellis J. E. (1995). Impacts of pastoralists on woodlands in south Turkana, Kenya: livestock-mediated tree recruitment. Ecological Applications, 5, 978–992. Republic of Kenya. (1996). Economic Survey. Government Printers, Nairobi. Rich S. M., Leendertz F. H., Xu G., LeBreton M., Djoko C. F., Aminake M. N., Takang E. E., Diffo J. L. D., Pike B. L., Rosenthal B. M., Formenty P., Boesch C., Ayala F. J., and Wolfe N. D. (2009). The origin of malignant malaria. Proceedings of the Nation- al Academy of Science of the U S A, 106, 14902–14907. Rich S. M. and Xu G. (2011). Resolving the phylogeny of malaria parasites. Proceed- ings of the National Academy of Science of the U S A, 108, 12973–12974. Riggio J., Jacobson A., Dollar L., Bauer H., Becker M., Dickman A., Funston P., Groom R., Henschel P., de Iongh H, Lichtenfeld L., and Primm S. (2013). The size of savan- nah Africa: a lion’s (Pantera leo) view. Biodiversity and Conservation, 22, 17–35. Ritchie E. G. and Johnson C. N. (2009). Predator interactions, mesopredator release and biodiversity conservation. Ecology Letters, 12, 982–998. Roca A. L., Georgiadis N., Pecon-Slattery J., and O’Brien S. J. (2001). Genetic evi- dence for two species of elephant in Africa. Science, 293, 1473–1477. Roemer G. W., Gompper M. E., and Van Valkenburgh B. (2009). The ecological role of the mammalian mesocarnivore. BioScience, 59, 165–173. Royama T. (1996). A fundamental problem in key factor analysis. Ecology, 77, 87–93. Rutberg A. T. (1987). Adaptive hypotheses of birth synchrony in ruminants—an interspecific test. American Naturalist, 130, 692–710. Rutherford M. C. (1997). Categorisation of biomes. In M. C. Rutherford and R. H. Westfall (1994). Biomes of southern Africa: an objective categorization, pp. 1–94. Memoirs of the Botanical Survey of South Africa 63, National Botanical Institute, Cape Town. Rutherford M. C. and Westfall R. H. (1994). Biomes of southern Africa: an object- ive categorization. Memoirs of the Botanical Survey of South Africa 63, National Botanical Institute, Cape Town. Ryan S. J., Knechtel C. U., and Getz W. M. (2007). Ecological cues, gestation length, and birth timing in African buffalo (Syncerus caffer). Behavioural Ecology, 18, 635–644. Sankaran M., Hanan N. P., Scholes R. J., Ratnam J., Augustine D. J., Cade B. S., Gignoux J., Higgins S. I., Le Roux X., Ludwig F., Ardo J., Banyikwa F., Bronn A., Bucini G., Caylor K. K., Coughenour M. B., Diouf A., Ekaya W., Feral C. J., Feb- ruary E. C., Frost P. G. H., Hiernaux P., Hraber H., Metzger K. L., Prins H. H. T., Ringrose S., Sea W., Tews J., Worden J., and Zambatis N. (2005). Determinants of woody cover in African savannas. Nature, 438, 846–849. Sarmiento G. (1996). Biodiversity and water relations in tropical savannas. In O. T. Solbrig, E. Medina, and J. F. Silva (ed.), Biodiversity and savanna ecosystem pro- cesses. A global perspective, pp. 61–75. Springer-Verlag, Berlin. Sarmiento G. and Monasterio M. (1983). Life forms and phenology. In F. Bourlière (ed.), Tropical savannas, pp. 109–150. Elsevier, Amsterdam. Savage A. F., Cerqueira G. C., Regmi S., Wu Y., El Sayed N. M., and Aksoy S. (2012). Transcript expression analysis of putative Trypanosoma brucei GPI-anchored surface proteins during development in the tsetse and mammalian hosts. PLoS Neglected Tropical Diseases, 6, e1708. Schaller G. B. (1972). The Serengeti lion. University of Chicago Press, Chicago. Schoener T. W. (1983). Field experiments on interspecific competition. American Naturalist, 122, 240–285. 302 REFERENCES Scholes M. C., Scholes R. J., Otter L. B., and Woghiren, A. (2003). Biogeochemistry: the cycling of nutrients in the Kruger National Park. In J. Du Toit, H. Biggs, and K. H. Rogers (ed.), The Kruger experience: ecology and management of savanna heterogeneity, pp 130–148. Island Press, Washington, DC. Scholes R. J. and Archer S. R. (1997). Tree-grass interactions in savannas. Annual Review of Ecology and Systematics, 28, 517–544. Scholes R. J. and Walker B. H. (1993). An African savanna: synthesis of the Nylsvley study. Cambridge University Press, Cambridge. Scrimshaw N. S., Taylor C. E., and Gordon J. E. (1968). Interactions of nutrition and infection. Monograph series 57, World Health Organization, Geneva. Seber G. A. F. (1982). The estimation of animal abundance and related parameters. Macmillan, New York. Seghieri J. (1995). The rooting patterns of woody and herbaceous plants in a savanna: are they complementary or in competition? African Journal of Ecology, 33, 358–365. Sharam G. J., Sinclair A. R. E., and Turkington R. (2006). Establishment of broad- leaved thickets in Serengeti, Tanzania: The influence of fire, browsers, grass -com petition and elephants. Biotropica, 38, 599–605. Shmida A. and Wilson M. V. (1985). Biological determinants of species diversity. Journal of Biogeography, 12, 1–20. Shorrocks B. (1978). The genesis of diversity. Hodder & Stoughton, London. Shorrocks B. and Cokayne A. (2005). Vigilance and group size in impala (Aepyceros melampus Lichtenstein): a study in Nairobi National Park, Kenya. African Journal of Ecology, 43, 91–96. Shorrocks B., Cristescu B., and Magane S. (2008). Estimating density of Kirk’s dik-dik (Madoqua kirkii Günther), impala (Aepyceros melampus Lichtenstein) and com- mon zebra (Equus burchelli Gray) at Mpala, Laikipia District, Kenya. African Jour- nal of Ecology, 46, 612–619. Shorrocks B. and Croft D. P. (2009). Necks and networks: a preliminary study of population structure in the reticulated giraffe (Giraffa camelopardalis reticulata de Winston). African Journal of Ecology, 47, 374–381. Sih A., Crowley P., McPeek M., Petranka J., and Strohmeier K. (1985). Predation, competition, and prey communities: a review of field experiments. Annual Review of Ecology and Systematics, 16, 269–311. Sillans R. (1958). Les savanes de l’Afrique centrale. Lechevalier, Paris. Sillero-Zubiri C., King A. A., and Macdonald D. W. (1996). Rabies and mortality in Ethiopian wolves (Canis simensis). Journal of Wildlife Diseases, 32, 80–86. Simon N. (1962). Between the sunlight and the thunder: the wildlife of Kenya. Collins, London. Simons H. W. and Chirambo P. C. (1991). Wildlife pest impact around Liwonde National Park, March–April (1990). FO: MLW/87/010 Field Document No. 11, FAO, Malawi. Sinclair A. R. E. (1974). The natural regulation of buffalo populations in East Africa. III. Population trends and mortality. East African Wildlife Journal, 12, 185–200. Sinclair A. R. E. (1975). The resource limitation of trophic levels in tropical grassland ecosystems. Journal of Animal Ecology, 44, 497–520. Sinclair, A. R. E. (1977). The African buffalo. A study of resource limitation of popula- tions. University of Chicago Press, Chicago. REFERENCES 303 Sinclair, A. R. E. (1979). Dynamics of the Serengeti ecosystem: process and pattern. In A. R. E. Sinclair and M. Norton-Griffiths (ed.),Serengeti: dynamics of an ecosystem, pp 1–30. University of Chicago Press, Chicago. Sinclair A. R. E. (1985). Does interspecific competition or predation shape the Afri- can ungulate community? Journal of Animal Ecology, 54, 899–918. Sinclair A. R. E. (1995a). Serengeti past and present. In A. R. E. Sinclair and P. Arc- ese (ed.), Serengeti II: dynamics, management and conservation of an ecosystem, pp. 3–30. University of Chicago Press, Chicago. Sinclair A. R. E. (1995b). Population limitation of resident herbivores. In A. R. E. Sinclair and P. Arcese (ed.), Serengeti II: dynamics, management and conservation of an ecosystem, pp 194–219, University of Chicago Press, Chicago. Sinclair A. R. E. (2012). Serengeti story: life and science in the world’s greatest wildlife region. Oxford University Press, Oxford. Sinclair A. R. E. and Arcese P. (ed.) (1995). Serengeti II: dynamics, management and conservation of an ecosystem. University of Chicago Press, Chicago. Sinclair A. R. E. and Byrom A. E. (2006). Understanding ecosystem dynamics for con- servation of biota. Journal of Animal Ecology, 75, 64–79. Sinclair A. R. E., Grant J., Olff H., Mduma S. A. R., Galvin K., and Sharon G. J. (2008). Historical and future changes to the Serengeti ecosystem. In:, A. R. E. Sinclair et al. (ed.), Serengeti III. Human impacts on ecosystem dynamics. University of Chicago Press, Chicago. Sinclair A. R. E., Mduma S., and Brashares J. S. (2003). Patterns of predation in a diverse predator-prey system. Nature, 425, 288–290. Sinclair A. R. E. and Norton-Griffiths M. (ed.) (1979).Serengeti: dynamics of an eco- system. University of Chicago Press, Chicago. Sinclair A. R. E. and Norton-Griffiths M. (1982). Does competition or facilitation regulate migrant ungulate populations in the Serengeti? A test of hypotheses. Oecologis, 53, 364–369. Skinner J. D. and Smithers R. H. N. (1990). Family Felidae—cats. In The mammals of the southern African sub-region, 2nd ed., pp. 391–423. University of Pretoria. Smit P. (2004). Prosopis: A review of existing knowledge relevant to Namibia. Namib- ia Scientific Society, 52, 13–40. Smith P. P. (1998). A reconnaissance survey of the vegetation of the North Luangwa National Park, Zambia. Bothalia, 28, 197–211. Smith S. M. and Cain J. W. (2008). Foraging efficiency and vigilance behaviour of impala: the influence of herd size and neighbour density. African Journal of Ecol- ogy, 47, 109–118. Smith J. E., Kolowski J. M., Graham K. E., Dawes S. E., and Holekamp K. E. (2008). Social and ecological determinants of fission-fusion dynamics in the spotted hyaena. Animal Behaviour, 76, 619–636. Smith J. E., Powning K. S., Dawes S. E., Estrada J. R., Hopper A. L., Piotrowski S. L., and Holekamp K. E. (2011). Greetings promote cooperation and reinforce social bonds among spotted hyaenas. Animal Behaviour, 81, 401–415. Smuts G. L. (1978). Interrelations between predators, prey and their environment. BioScience, 28, 316–320. Solbrig O. T. (1996). The diversity of savanna ecosystems. In O. T. Solbrig, E. Medina, and J. F. Silva (ed.), Biodiversity and savanna ecosystem processes. A global perspec- tive, pp. 1–27. Springer-Verlag, Berlin. 304 REFERENCES Solomon T., Snyman H. A., and Smit G. M. (2007). Rangeland dynamics of Southern Ethiopia: (2). Assessment of woody vegetation structure in relation to land use and distance from water in semi-arid Borana rangelands. Journal of Environmental Management, 85, 443–452. Soltis J. (2009). Vocal communication in African elephants (Loxodonta africana). Zoo Biology, 28, 1–18. Soulé M. E., Wilcox B. A., and Holtby C. (1979). Benign neglect: a model of faunal collapse in the game reserves of East Africa. Biological Conservation, 15, 259–272. Spiegel O., Getz W. M., and Nathan R. (2013). Factors influencing foraging search efficiency: Why do scarce Lappet-faced vultures outperform ubiquitous white- backed vultures? American Naturalist, 181, 102–115 Spinage C. A. (1973). A review of ivory exploitation and elephant population trends in Africa. East African Wildlife Journal, 11, 281–289. Stanton M. L., Palmer T. M., and Young T. P. (2002). Competition-colonization trade- offs in a guild of African acacia-ants. Ecological Monographs, 72, 347–363. Stapley L. (1998). The interaction of thorns and symbiotic ants as an effective defence mechanism of swollen-thorn acacias. Oecologia, 115, 401–405. Steenkamp H. E. and Chown S. L. (1996). Influence of dense stands of an exotic tree, Prosopis glandulosa Benson, on a savanna dung beetle (Coleoptera: Scarabaeinae) assemblage in southern Africa. Biological Conservation, 78, 305–311. Steffan W. I., Walker B. H., Ingram J. S. I., and Koch G. W. (1992). Global change and terrestrial ecosystems: the operational plan. IGBP, Stockholm. Stevenson Hamilton J. (1911). The relation between game and tsetse-flies. Bulletin of Entomological Research, 2, 113–118. Stevenson Hamilton J. (1947). Wildlife in South Africa. Cassell, London. Stone G. N., Willmer P., and Rowe A. J. (1998). Partitioning of pollinators during flowering in an African Acacia community. Ecology, 79, 2808–2827. Stuart C. and Stuart T. (1997). Field guide to the larger mammals of Africa. Struik, Cape Town. Sueur J., Pavoine S., Hamerlynck O., and Duvail S. (2008). Rapid acoustic survey for biodiversity appraisal. PLOS One, 3, e4065. Sundaresan S. R., Fischhoff I. R., Dushoff J., and Rubenstein D. I. (2007). Network metrics reveal differences in social organization between two fission–fusion spe- cies, Grevy’s zebra and onager. Oecologia, 151, 140–149. Swann D. E., Kawanishi K., Palmer J. (2010). Evaluating types and features of camera traps in ecological studies: a guide for researchers. In: A. F. O’Connell, J. D. Nich- ols, and U. K. Karanth (ed.) Camera traps in animal ecology: methods and analyses. Springer, Tokyo. Swynnerton G. H. (1958). Fauna of the Serengeti National Park. Mammalia, 22, 435–450. Talbot L. M. and Talbot M. H. (1963). The wildebeest in western Masai-land, East Africa. Wildlife Monograph No. 12, The Wildlife Society, Bethesda, MD. Tan-Mullins M. and Mohan G. (2013). The potential of corporate environmental responsibility if Chinese state-owned enterprises in Africa. Environment, Develop- ment and Sustainability, 15, 265–284. Taylor R. J. (1984). Predation. Chapman & Hall, New York. Thompson J. J. (1989). A comparison of some avian census techniques in a popula- tion of lovebirds at Lake Naivasha, Kenya. African Journal of Ecology, 27, 157–166. REFERENCES 305 Thoules C. R. (1994). Conflict between humans and elephants on private land in northern Kenya. Oryx, 28, 119–127. Thresher P. (1981). The economics of a lion.Unasylva , 33, 34–36. Torchio P. and Manconi M. (2004). Rugged pursuit. Swara, 27, 70–72. Trinkel M. and Kastberger G. (2005). Competitive interactions between spotted hyaenas and lions in the Etosha National Park, Namibia. African Journal of Ecol- ogy, 43, 220–224. Tucker C. J. (1979). Red and photographic infrared linear combinations for monitor- ing vegetation. Remote sensing of Environment, 8, 127–150. Tucker C. J. J., Townsend J. R. G., and Goff T. E. (1985). African landcover classifica- tion using satellite data. Science, 227, 369–375. Turner A. and Anton M. (1997). The big cats and their fossil relatives. Columbia Uni- versity Press, New York. Tutchings A., Fennessy S., Marais A., and Fennessy J. (2013). Africa’s giraffe. A conser- vation guide. Giraffe Conservation Foundation, Windhoek. Tyson P. D. (1986). Climatic change and variability in southern Africa. Oxford Univer- sity Press, Oxford, UK. Vandermeer J. (1984). The evolution of mutualism. In B. Shorrocks (ed.),Evolution - ary Ecology, pp. 221–232. Blackwell Scientific Publications, Oxford. Vandvik V. and Birks H. J. B. (2002). Pattern and process in Norwegian upland grass- lands: a functional analysis. Journal of Vegetation Science, 13, 123–134. Van Langevelde F., Tomlinson K., Barbosa E. R. M., de Brie S., Prins H. H. T., and Hig- gins S. I. (2011). Understanding tree-grass coexistence and impacts of disturbance and resource variability in savannas. In:, M. J. Hill and N. P. Hanan (ed.), Ecosystem function in savannas. Measurement and modelling at landscape to global scales. CRC Press, Boca Raton, FL. Varley G. C. and Gradwell G. R. (1970). Recent advances in insect population dynam- ics. Annual Review of Entomology, 15, 1–24. Vermeulen C., Lejeune P., Lisein J., Sawadogo P., and Bouché P. (2013). Unmanned aerial survey of elephants. PLoS ONE, 8, e54700. Vesey-Fitzgerald D. F. (1960). Grazing succession among East African game mam- mals. Journal of Mammalogy, 41, 161–172. Vesey-Fitzgerald D. F. (1973). Browse production and utilization in Tarangire Nation- al Park. East African Wildlife Journal, 11, 291–306. Vesey-Fitzgerald D. F. (1974). Utilization of the grazing resources by buffaloes in the Arusha National Park, Tanzania. East African Wildlife Journal, 12, 107–134. Vila M., Espinar J. L., Hejda M., Hulme P. E., Jarosik V., MaronJ. L., Pergl J., Schaf- fner U., Sun Y., and Pyšek P. (2011). Ecological impacts of invasive alien plants: a meta-analysis of their effects on species, communities and ecosystems.Ecological Letters, 14, 702–708. Voeten M. M. and Prins H. H. T. (1999). Resource partitioning between sympatric wild and domestic herbivores in the Tarangire region of Tanzania. Oecologia, 120, 287–294. Wafula M. M., Atimnedi P., and Tumwesigye C. (2007). Managing the 2004/05 anthrax outbreak in Queen Elizabeth and Lake Mburo National Parks, Uganda. African Journal of Ecology, 46, 24–31. Wagner A. P. (2006). Behavioral ecology of the striped hyaena (Hyaena hyaena). PhD thesis, Montana State University. 306 REFERENCES Waithaka J. M. (1996). Elephants: a keystone species. In T. R. McClanahan and T. P. Young (ed.), East African ecosystems and their conservation, Box 11.2, p. 284. Oxford University Press, Oxford. Walter H. (1971). Ecology of tropical and subtropical vegetation. Oliver & Boyd, Edinburgh. Walter, H. (1973). Vegetation of the Earth in relation to climate and the ecophysiological conditions. Springer-Verlag, Berlin. Wang J. and Aksoy S. (2012). PGRP-LB is a maternally transmitted immune milk pro- tein that influences symbiosis and parasitism in tsetse’s offspring.Proceedings of the National Academy of Science of the U S A, 109, 10552–10557. Wang S., Ghosh A. K., Bongio N., Stebbings K. A., Lampe D. J., and Jacobs-Lorena M. (2012). Fighting malaria with engineered symbiotic bacteria from vector mosqui- toes. Proceedings of the National Academy of Science of the U S A, 109, 12734–12739. Warner R. R. and Chesson P. L. (1985). Coexistence mediated by recruitment fluctuations: a field guide to the storage effect. American Naturalist, 125, 769–787. Wasser S. K., Clark W. J., Drori O., Kisamo E. S., Mailand C., Mutayoba B., and Stephens M. (2008). Combating the illegal trade in African elephant ivory with DNA forensics. Conservation Biology, 22, 1065–1071. Waters A. P., Higgins D. G., and McCutchan T. F. (1991). Plasmodium falcipa- rum appears to have arisen as a result of lateral transfer between avian and human hosts. Proceedings of the National Academy of Science of the U S A, 88, 3140–3144. Watts A. C., Perry J. H., Smith S. E., Burgess M. A., Wilkinson B. E., Szantoi Z., Ifju P. G., and Percival H. F. (2010). Small unmanned aircraft systems for low-altitude aerial surveys. Journal of Wildlife Management, 74, 1614–1619. Webb W. L., Lauenroth W. K., Szarek S. R., and Kinerson R. S. (1983). Primary pro- duction and abiotic controls in forests, grasslands and desert ecosystems in the United States. Ecology, 64, 134–151. Weeks P. (1999). Interactions between red billed oxpeckers, Buphagus erythro- rhynchus, and domestic cattle, Bos taurus, in Zimbabwe. Animal Behaviour, 58, 1253–1259. Weeks P. (2000). Red-billed oxpeckers: vampires or tick birds? Behavioural Ecology, 11, 154–160. Weissengruber G. E., Forstenpointner G., Peters G., Kubber-Heiss A., and Fitch W. T. (2002). Hyoid apparatus and pharynx in the lion (Panthera leo), jaguar (Panthera onca), tiger (Panthera tigris), cheetah (Acinonyx jubatus) and domestic cat (Felis silvestris, F. catus). Journal of Anatomy, 201, 195–209. West P. M. and Packer C. (2002). Sexual selection, temperature, and the lion’s mane. Science, 297, 13339–11343. Weston D. and Ssemakula J. (1981). The future of the savannah ecosystem: ecological islands or faunal enclaves? African Journal of Ecology, 19, 7–19. Wey T., Blumstein D. T., Shen W., and Jordan F. (2008). Social network analysis of animal behaviour: a promising tool for the study of sociality. Animal Behaviour, 75, 333–344. White F. (1983). The vegetation of Africa. UNESCO, Paris. Whitehead H. (2009). SOCPROG programs: analysing animal social structures. Behavioural Ecology Sociobiology, 63, 765–778. REFERENCES 307 Whitehouse A. M., Hall-Martin A. J., and Knight M. H. (2001). A comparison of methods used to count the elephant population of the Addo Elephant National Park, South Africa. African Journal of Ecology, 39, 140–145. Whittaker R. H. (1975). Communities and ecosystems, 2nd ed. Macmillan, New York. Whyte I. J. (1985). The present ecological status of the blue wildebeest (Connochaetes taurinus taurinus, Burchell, 1823) in the central district of the Kruger National Park. MSc thesis, University of Natal, Pietermaritzburg, South Africa. Williams B. K., Nichols J. D., and Conroy M. J. (2002). Analysis and management of animal populations: modeling, estimation, and decision making. Academic Press, San Diego. Williams D., Pettorelli N., Henschel J., Cowlishaw G., and Douglas C. M. S. (2014). Impact of alien trees on mammal distributions along an ephemeral river in the Namib Desert. African Journal of Ecology, 52, DOI: 10.1111/aje.12134. Willmer P., Stone G. N., and Mafunde D. (1998). Ants, pollinators and acacias in Mkomazi. In M. Coe et al. (ed.), Mkomazi. The ecology, biodiversity and conser- vation of a Tanzanian savanna, pp. 337–360. Royal Geographic Society (with the Institute of British Geographers), London. Wilson D. E., Cole F. R., Nichols J. D., Rudran R., and Foster M. S. (1996). Measuring and monitoring biological diversity. Standard methods for mammals. Smithsonian Institution Press, Washington. Wirtz P. and Lörscher J. (1983). Group sizes of antelopes in an East African park. Behaviour, 84, 135–155. Wittemyer G. (2001). The elephant population of Samburu and Buffalo Springs National Reserve, Kenya. African Journal of Ecology, 39, 357–365. Wittemyer G., Ganswindt A., and Hodges K. (2007a). The impact of ecological vari- ability on the reproductive endocrinology of wild female African elephants. Hor- mones and Behaviour, 51, 346–354. Wittemeyer G., Rasmussen H. B., and Douglas-Hamilton I. (2007b). Breeding phen- ology in relation to NDVI variability in free-ranging African elephant. Ecography, 30, 42–50. Woodborne S., Huchzermeyer K. D. A., Govender D., Pienaar D. J., Hail G., Myburgh J. G., Deacon A. R., Venter J., and Lubcker N. (2012). Ecosystem change and the Olifants river crocodile mass mortality events. Ecosphere, 3, 87. Woodford M. H. and Sachs R. (1973). The incidence of cysticercosis, hydratidosis, and sparganosis in wild herbivores of the Queen Elizabeth National Park, Uganda. Bulletin of Epizootic Diseases of Africa, 21, 265–272. Woodroffe R., Ginsberg J., and Macdonald D. (1997).The African wild dog. IUCN, Gland. Woodward A. and Reed J. D. (1995). Intake and digestibility for sheep and goats con- suming supplementary Acacia brevispica and Sesbania sesban. Animal Feed Science and Technology, 56, 207–216. World of Wildlife (1971). Orbis Publishing, London. Wragg F. C. (2002). Biodiversity and conservation of African mammals. PhD thesis, University of Leeds. WWF (2011). Camera traps—more on camera traps. World Wildlife Fund— Wildlife Conservation, Endangered Species Conservation. World Wildlife Fund. Yadava P. S. (1991). Savannas of north-east India. In P. A. Werner (ed.), Savanna ecol- ogy and management, pp. 3–10. Blackwell Scientific Publications, Oxford. Young T. P., Palmer T. M., and Gadd M. E. (2005). Competition and compensation among cattle, zebras, and elephants in a semi-arid savanna in Laikipia, Kenya. Bio- logical Conservation, 122, 351–359. Young T. P., Stubblefield C. H., and Isbell L. A. (1997). Ants on swollen-thorn acacias: species coexistence in a simple system. Oecologia, 109, 98–107. Zimmerman D. A. et al. (1996). Birds of Kenya and northern Tanzania. Princeton University Press, Princeton, NJ. Index Note: The suffix ‘f’ following a page reference indicates a figure and ‘t’ indicates a table. A Acacia nilotica (Egyptian thorn) 38t, 44, 48 aardvark (Orycteropus afer) 22, 112 Acacia nilotica subulata 20 aardwolf (Proteles cristatus) 17, 127 Acacia reficiens 38t absolute population estimates 138 Acacia robusta (splendid) 46, 48, 50t Acacia 7, 18, 34, 43, 45, 62, 190, 193, 194 Acacia robynsiana 18 ants 66, 231–4 Acacia senegal (three thorned aka gum arabic) 18, giraffes 89 37f, 38t, 44, 49, 50t insects 66 Acacia seyal var. fistula 233 Lewa Downs (Kenya) 273 Acacia seyal (white thorn) 19, 49, 50, 215 Mimosoideae subfamily 43 Acacia sieberiana (paperbark) 37, 37f, 50 mutualistic interactions 228 Acacia thomasii 38t phenology 37f Acacia tortilis (umbrella thorn) 16, 18, 44, 48, 50t, pods 44, 228 64, 193 seasonal flowering patterns 38t mutualistic interactions 228 seeds 228 seasonal flowering patterns 38t Serengeti National Park 50t Acacia xanthophloea (fever tree) 50t, 67 Acacia brevispica (Senegalia)(prickly aka wait a bit Acacia zanzibarica 38t, 232 thorn) 38t, 45 Acrida 67 Acacia bussei 38t Acrotylus elgonensis 67 Acacia clavigera 64 Acrotylus patrrualis 67 Acacia–Commiphora 16, 98 Aculeiferum senegal 44 Acacia drepanolobium (whistling thorn) 29, 34, 45, Addo Elephant National Park (South Africa) 148 46, 50, 64, 139, 140, 151, 188, 231f, 232, 233, 233f Adenolobus 18 herbivore damage 189f Advanced Very High Resolution Radiometer seasonal flowering patterns 38t (AVHRR) 93 tagging 140, 188 aerial surveys 146–50 Acacia erioloba 232 Africa 1, 7 Acacia erubescens (blue thorn) 20, 45, 46 average rainfall 12f Acacia etbaica 38t, 194 ecological features of savannahs 8t Acacia gerrardii (grey-haired aka Gerrard’s thorn) 46, location of savannahs 2, 5 50t, 64 African blackwood (Dalbergia melanoxylon) 19 Acacia hebeclada tristis (candle pod) 20, 46 African buffalo Syncerus( caffer) 17, 19, 20, 22, 84f, Acacia hockii 50t, 215 154, 155–61 Acacia karroo 214 see also buffalo (Bovidae family, Tribe Bovini) Acacia kirkii (Kirk’s aka flood plain) 20, 46 African civet (Civettictis civetta) 130 Acacia laeta (black hooked thorn) 16, 47 African elephant (Elephantidae family) 112–16 Acacia mellifera (hook thorn aka wait a bit African genets 129–30 thorn) 37f, 44, 47, 193 African horse sickness 210 Acacia montis-ustii (Brandberg acacia) 18 African mahogany (Afzelia africana) 51 Acacia nebrownii 18 African palm civet (Nandinia binotata) 130 Acacia nigrescens (nob thorn) 19, 44, 47 African striped weasel (Poecilictis libyca) 133 310 INDEX African white-backed griffin (Gyps africanus) 72 antipredation 204, 206f African wild ass (Equus africanus) 17 ants 6, 22, 66, 231–4 African wild cat (Felis silvestris lybica) 126, 217 Aonyx capensis (otter) 133 African wild dog (Lycaon pictus) 22, 131, 225, 263 Aonyx congica (otter) 133 Afrohippus taylori 66–7 Apis 66 Afrotheria 74 Apis mellifera 66 Afzelia africana 23, 24 apparent competition 182, 183f age specificity 154, 154f, 156, 157f, 161 Arecaceae (Palm family) 42 Akagara National Park (Rwanda) 106 arid/eutrophic savannahs 168 Albizia antunesiana 21 ‘arid’ zone 11 Alcelaphini tribe (hartebeests, topi, wildebeests and Aristida 19 impala) 104–8 Aristida adscensionis (common needle grass) 16, 39t Allerton-type herpes virus 160 Aristida stipoides (needle grass) 16, 39t Allideghi Wildlife Reserve (Ethiopia) 216 artificial markings 140–2 Aloe 16 Artiodactyla 75 Aloe vera 42 artiodactyls 75 Amaryllidaceae (daffodils) 36 Arundinoideae 40 Amboseli National Park 18, 134, 277 Arusha National Park (Tanzania) 185, 256–7, 256f, amensalism 182 257t Anacardiaceae (mango family) 54 Asian elephant (Elephas maximus indicus) 7 anaplasmosis 160 assembly rules 252–7 Andropogon 16, 24, 40 assimilation efficiency 243 Andropogoneae 40 attack, risk of 203, 203f, 204 Andropogon gayanus (bluestem grass) 39t Australia 1, 5–7, 7t, 8, 8t Andropogon greenwayi (grass) 39t, 61, 186t Australopithecus afarensis 265 Angola 19, 20, 21, 262 avian fauna see birds Angolan giraffe (smoky giraffe)(G. c. angolensis) 89, 90 B angular-stemmed commiphora (Commiphora B. a. africanus (oxpecker) 235 karibensis) 53 babesiasis 160 animals 65–136 baboons 42, 44, 54, 135–6, 269, 269t artificially marked 140–1 badgers (Mustelidae family) 133 birds 69–74 Baikiaea plurijuga 20 identifying individuals 139 Baikiaea woodlands 20 insects 65–9 Balanitaceae (desert date family) 52–3 mammals 74–136 Balanites aegyptiaca 16, 19, 33, 50 Aningeria altissima 24 Balanites angloensis 20 Annona senegalensis 24 bamboos 40 annuals 36 Bambusoideae 40 Anogeissus leiocarpus 19, 23 banded mongoose (Mungos mungo) 130t Anopheles 67 banded weasel (Poecilictis libyca) 133 antelope 17, 197, 204, 206f baobab (Adansonia digitata) 19, 34, 54 dwarf 110–12 baobab family (Bombacaceae) 54–5 feet 75 Baphia massaiensis obovata 21 horse 99–102 bat-eared fox (Otocyon megalotis) 17, 18, 20, 132, 198 prey 198 bats 7t, 18 rumination 75 Bauhinia petersiana serpae 21 spiral-horned 95–9 bearded vulture (Guptaetus barbatus see also roan antelope (Hippotragus equinus); sable meridionalis) 149 antelope (Hippotragus niger) beechwood (Faurea saligna) 42 anthrax (Bacillus anthracis) 170, 207, 209, 210 bees (Apis mellifera scutellata) 44, 232, 270 antibodies 207 beira (Dorcatragus megalotis) 17 Antilopini tribe (gazelles and dwarf antelopes) beisa oryx (Oryx gazella beisa) 17, 99, 100f, 101 108–12 Bengal tigers (Panthera tigris tigris) 7 INDEX 311 Benin 103 black wildebeest (Conochaetes gnou) 107 Berlinia grandiflora 24 blade (grass leaf) 34 best-fit population model 165f, 166 blesbok 218 Bignoniaceae (Jacaranda family) 56–7 blesmols 135 bioacoustics 117–20 blowflies 209 biomass 238, 245t blue monkey (Cercopithecus mitis) 23 of carnivores 242 blue thorn (Acacia erubescens)(Senegalia of herbivores 242 erubescens) 45 –rainfall relationship 168, 169f, 181 blue wildebeest (Connochaetes taurinus taurinus) 106 resistance 245 body size 198, 204, 222, 223, 228 BIOME 3 computer model 2 –abundance relationships 138 biomes 2f, 31 energy turnover and 223, 253 birds 5, 7t, 17, 42, 65, 69–74 ungulates 252, 254f dawn chorus 120 vultures 72 distribution of 70t body weight 142, 161, 223, 235, 253, 253f, 254, 255 effects of pesticides on 275 bovids 256f, 257 metal rings 140 protein turnover and 223 miombo woodland savannah 23 ungulates 224f South America 5 vultures 72 vultures 71–3 bohor reedbuck (Redunca redunca) 104, 169f weavers 73–4 bongo (Tragelaphus eurycerus) 24 wildebeest (Connochaetes taurinus) and 252 Bontebok National Park (southern Africa) 217 birth 135, 137, 153, 154, 156 boomslang (Dispholidus typhus) 65 buffalo 89 borassus palm (Borassus aethiopum) 23, 33, 34, 42 elephant 114, 115 Boscia 43 lesser kudu 99 Boscia albitrunca 18 lion 123 Boscia foetida 18 roan antelope 102 Boscia senegalensis 16, 43 sable antelope 102 Bos indicus 270 steenbok 111 Boswellia carteri 16, 53 zebras 83 Boswellia papyrifera 19 biting flies 160 Botswana 20, 262, 273t, 277, 278 black and white colobus (Colobus angolensis) 23 Bovidae family 89–112 black-backed jackal (Canis mesomeles) 17, 20, 113f, buffalo 89–92 123, 131, 198, 209, 217 dwarf antelopes 110–12 skull 122f gazelles 108–10 weights of mammal prey consumed by 197f hartebeests, topi, wildebeests and impala 104–8 black-faced impala (Aepyceros melampus petersi) 18, horse antelopes 99–102 20, 107 reedbucks, kob and waterbuck 102–4 black-footed cat (Felis nigripes) 126 spiral-horned antelopes 95–9 blackheaded (village weaver)(Plocus cucullatus) 74 bovids 184, 194, 256, 256f, 257, 260f, 261, 262 black-hooked thorn (Acacia laeta)(Senegalia Brachiaria brizantha (common signal grass) 39t laeta) 47 Brachiaria deflexa (signal grass) 19, 39t black mamba (Dendroaspis polylepis) 65 Brachystegia 20, 33, 51 black rhinoceros (hook-lipped)(Diceros bicornis) 18, Brachystegia allenii 21 20, 22, 34, 83–5, 84f, 169f Brachystegia bakerana 21 ear notch markings 141, 141f Brachystegia boehmii 21 endangered species programme 273 Brachystegia floribunda 21 head 86f Brachystegia glaberrima 21 Serengeti 191, 266, 267f Brachystegia gossweilerii 21 skull 80f Brachystegia longifolia 21 teeth 79 Brachystegia spiciformis 21, 51 tree damage 189 Brachystegia taxifolia 21 black sable (Hippotragus niger niger) 102 Brachystegia utilis 21 312 INDEX Brachystegia wangermeeana 21 predator–prey interactions 197, 198, 207 Brandberg acacia (Acacia montis-ustii) 18 rainfall and abundance of 170 Brazil 3 rumination 75 brindled gnu see wildebeest (Connochaetes taurinus) sarcoptic mange 210 broad-leaved beechwood (Faurea specios) 42 ticks 160 brown capuchin monkey (Cebus apella) 3 see also African buffalo Syncerus( caffer) brown hyaena (Hyaena brunnea) 20, 123, 127–8, buffalo-weavers 69 209 Buffon’s kob Kobus( kob kob) 103 browsers 78–9, 170, 188, 189, 193, 222 Buphagus erythrorhynchus 235 Acacia 50, 194, 228 Burchell’s zebra (Equus quagga burchelli) 235 body size 253, 254f Burigi-Biharamulo Game Reserve (Tanzania) 256f, 257t bushbuck 99 burke (Burkeae africana) 21, 23, 24, 52 eland 97 burning 7, 32, 36 gazelles 108 Burseraceae (myrrh family) 53 giraffes 78, 89 bushbuck (Tragelaphus scriptus) 17, 22, 78, 80f, 99, greater kudu 98 100f, 139, 169f horse antelopes 99–102 bush fires 33, 36, 63 impala 78, 107 bush pigs 269, 269t Kirk’s dik dik 111 bush squirrels 135 lesser kudu 98 bush-strikes 69 niche separation 224, 224f bushytailed mongoose (Bdeogale crassicauda) 22 –plant interactions 188–9 bustards 69 rhinoceros 83, 86f butterflies 44, 232, 252 seasonal intake of protein and energy 244f Butyrospermum paradoxum 24 spines 34 spiral-horned antelopes 95–9 C springbok 110 C3 plants 1, 114 zebras 82 C4 plants 1, 114 browsing 216 Caesalpinioideae 51 height stratification 224 Caessalpiniodes 43 bruchid beetles 44, 234 camels 75 buffalo (Bovidae family, Tribe Bovini) 38, 78, 89–92, camel’s foot trees (Bauhinia genus) 51 168, 187f, 217, 228, 266, 269 CAMPFIRE 270, 278 biomass–rainfall relationship 169f Canada 120, 194 body weights 224f candelabra tree (Euphorbia candelabrum) 54 cyclic grazing 186 candle-pod acacia (Acacia hebeclada)(Vachellia diseases 160 hebeclada) 46 dung 77 cane-rats 135 estimating numbers of 144 Canidae family (dogs and jackals) 130–2 feet 75 canids 197, 208 food webs 245 canine distemper 210 group (herd) size in 205f canine distemper virus 207, 208 –human conflict 269, 269t Cape fox (Vulpes chama) 18, 132, 198 hunting 267 cape grey mongoose (H. pulverulenta) 130t influence on biomass of grazing by 245t Cape griffin vulture (Gyps coprotheres) 72, 73 lion predation 160 cape hare (Lepus capensis) 217 Masai Mara ecosystem 152, 153f Cape (red) hartebeest (Alcelaphus buselaphus mortality 159–60 caama) 105 niche separation 224 Caper family (Capparaceae) 43 nutrition 160–1 Capparis spinosa 43 poaching 173 capybara (Hydrochaeris hydrochaeris) 3, 5, 8 population numbers in the Kruger National Park in caracal (Caracal caracal) 22, 126–7, 197f relation to rainfall 176f carbon dioxide 247 predation rates 171t carcass biomass 178 INDEX 313 carnivores 22, 116–33, 194, 263 Chrysophyllum perpulchrum 24 badgers, weasels, polecats and otters (Mustelidae Chrysopogon aucheri (Aucher’s grass) 39t family) 133 Cissus cornifolia 19 cats (Felidae family) 123–7 Cissus mildbraedii 36 dogs and jackals (Canidae family) 130–2 Cisticola grass warblers 69 genets, civets and mongooses (Viverridae family CITES 267, 268 and Herpestidae) 129–30 civets 130 group living 121 climatically determined savannahs 29 hunting behaviour 121 climatic effects 152, 210 hyaenas (Hyaenidae family) 127–9 climatic patterns 9–14 teeth 116–21 coastal topi (Damaliscus lunatus topi) 105 carrion 235 coconut palm (Cocos nucifera) 42 carrying capacity 167, 263 coexistence 212, 215, 224 caryopsis 35 tree–grass 29, 213, 215, 215f cascading indirect interactions (in communities) 246 Coke’s hartebeest (Alcelaphus buselaphus cokii) 105, cassava plant (Manishot esculenta) 53 152, 221, 221t Cassia subfamily (Caessalpiniodes) 43 Cola gigantea 24 catenas/catenary sequences 62 Cola laurifolia 24 catfish 275 coliiformes 69 Catophractes alexandri 18 Colombia 4 cats (Felidae family) 123–7, 197, 198, 207 Colophospermum mopane 20 cattle 6, 208, 270, 271, 272t, 273, 275 Combretaceae (Terminalia family) 55–6 cellulase 1 Combretum 18, 55, 89 cellulose 1, 5, 75, 76, 135 Combretum apiculatum 19, 20 Cenchrus biflorus (African foxtail) 16, 39t Combretum collinum 24 Cenchrus ciliaris (common African foxtail) 19, 39t Combretum engleri 20 cerrado 3, 8 Combretum imberbe 19 cestodes 160 Combretum molle 21, 64 Cetacea 75 commensalism 182 cetaceans 75 Commiphora 16, 19, 20, 53 Cetartiodactyla 75 Commiphora abyssinica 16, 53 cetartiodactyles 75 Commiphora africana (poison-grub) 16, 19, 53 chacma baboon (Papio ursinus) 135 Commiphora karibensis (angular-stemmed chaco 4 commiphora) 53 chamaephytes 31, 32, 32t Commiphora mollis (soft-leaved commiphora) 53 cheetah (Acinonyx jubatus) 17, 18, 19, 20, 22, 113f, Commiphora trothae 64 125–6 common duiker (Sylvicapra grimmia) 217 effects of rivers and fences on 263 common eland see eland (Taurotragus oryx) exposure to anthrax and feline coronavirus 207 common hartebeest (Alcelaphus buselaphus) 104 identifying individuals 139 common hippopotamus (Hippopotamus kleptoparasitism and 227 amphibius) 87–8 population estimates 179t common (plains) zebra (Equus quagga) 17, 22, 82, predation data for 199t 84f predator–prey interactions 201 common (small-spotted) genet (Genetta genetta) weights of mammal prey consumed by 197f 129 chevrotains 75 common (southern) reedbuck (Redunca chimpanzee (Pan troglodytes) 23 arundinum) 104, 217 China 274 common warthog (Phacochoerus africanus) 86–7 chital/spotted deer (Axis axis) 7 common waterbuck (Kobus ellipsiprymus Chloridoideae 40 ellipsiprymus) 22, 102 Chloris gayana (Rhodes grass) 35, 39t common wildebeest (Connochaetes taurinus) 100f, Chloris pycnothrix 62 106 Chobe National Park (Botswana) 21, 227 community patterns 213, 252, 255, 257 Chrysobalanaceae (Mobola family) 43 compensatory growth 184 314 INDEX competition Cymbopogon afronardus (blue citronella) 39t apparent 182, 183f Cymbopogon plurinodis (citronella) 39t balanced 215 Cynodon dactylon (Bermuda grass) 39t, 185, 186, exploitation 210 187f field manipulations 212 Cynometra vogelii 24 interference 210 Cyphostemma 18 past 213, 221 between trees and grass 213–16 D two-species 211 Dactyloctenium aegyptium (common crowfoot) 16, 39t competitive interactions 182, 210–13 dama gazelle (Nanger dama) 17 interspecific 210, 212, 214, 217, 220, 230 Damaliscus lunatus 105 intraspecific 210 Damaliscus superstes 105 kleptoparasitism 225–7, 234 Daniellia oliveri 24 resource competition 217–25 date palm (Phoenix dactylifera) 42 trees and grass 213–16 death 67, 81, 93, 116, 153, 154, 156, 181 complex communities 244 of matriarch elephant 114 complexity and stability 243, 244 numbers estimation and 142 connectance 243 nutrition and 161 conservation 237–8 and pollution 274 assembly rules 252–7 predator 162, 171 energy flow and food webs 238–52 probability of 155 habitat destruction and wildlife–human rates 68, 137, 172 conflict 269–76 risk of 203, 204f hunting and poaching 265–9 road 131 island biogeography 257–9 declining species 152 protected savannah areas 259–64 decomposers 241 of savannah ecosystems 264–5 decomposition rate status 81–2 of dung 150 wildlife tourism 277–8 of litter 6, 8t consumers 241 deer 5, 7 primary 241 defassa waterbuck (Kobus ellipsiprymus defassa) 17, secondary 241, 243 100f, 102–3 consumption efficiency 243 defecation rate 150 Convention on International Trade in Endangered Demidoff’s galago (Galagoides demidoff ) 133 Species (CITES) 267, 268 Democratic Republic of Congo 24 Copaifera baumiana 21 density dependence 167, 177 copper mining 273, 274 density independence 167 coprophagy (reingestion of faeces) 5 density (trees) 19 correspondence analysis 225 desert date (Balanites aegyptiaca) 52–3 Coryphosima stenoptera 66 desert date family (Balanitaceae) 52–3 counting errors 152 desert dwarf mongoose (Helogale hirtula) 130t Crematogaster mimosae (ant) 231f, 233 desert warthog (Phacochoerus aethiopicus) 17, 87, 235 Crematogaster nigriceps (ant) 231f, 233 detection, risk of 201, 203, 204 Crematogaster sjostedti (ant) 231f, 233, 234 detritivores 66, 66t, 238 critically endangered species 82, 85, 102, 105 Dialium engleranum 20 crocodiles 274–5 Diamphidia beetle 53 crop loss 269 diastema 79, 135 Crotalaria 16 dibatag (Ammodorcas clarkei) 17 Croton 64 Dichanthium papillosum 20 Ctenium concinnum 40 Dichrostachys cinerea 20 Ctenium newtonii (sickle grass) 39t, 40 dicotyledons 41, 78, 253 Ctenium somalense 40 Digitaria eriantha (Panola grass) 19, 39t culm 34 Digitaria macroblephora (woolly finger grass) 39t, 61 Cymbopogon 19, 40 dik dik 20, 193, 205 INDEX 315 dilution effect 201, 203f elephant (Elephantidae family) 17, 18, 19, 20, 22, Diospyros elliotii 24 113f, 193, 266 Diplorhynchus condylocarpon 21 Acacia and 44, 194 direct effects 188, 237 biomass–rainfall relationship 169f disease 67, 159, 171, 207–10, 234, 247 crop raiding 270 birds and 69 dietary history 115 from domestic dogs 131 dung 77 endemic 181 evolution 112 outbreaks 171 feeding facilitation and grazing succession 228 spread of 92 forest (Loxodonta africana cyclotis) 24, 116 see also rinderpest habitat destruction and wildlife–human disturbance driven savannahs 29 conflict 269, 269t, 270 dogs (Canidae family) 130–2 as a keystone species 252 see also jackals (Canidae family); wild dog (Lycaon and marula fruit 54 pictus) Masai Mara ecosystem 153f doka savannah 21, 23 mixed feeding 78 donkeys 6 poaching 191, 267, 267f dorcas gazelle (Gazella dorcas) 17 population estimates 150 dormice 135 savannah (Loxodonta africana) 112–16 droughts 22, 34, 216 social network analysis 90 dry season 3, 5, 9, 11, 13 teeth 81 dung tree damage 189 beetles 66, 216 trees, fire and 190 counts 150 tusks 81 decomposition rate 150 vocalisations 116, 118–19 dwarf antelopes (Bovidae family, Antilopini and elephant grass (Pennisetum purpureum) 18 Neotragini tribes) 110–12 Elephas iolensis 255 dwarf mongoose (Helogale parvula) 121, 130t Elephas recki 255 dynamic theory of island biogeography 258 Elephas zulu 255 endangered species 82, 83, 89, 98, 105, 131 E endemism 21 ear notch markings 141 endotherm vertebrates, production efficiency 243 ebony family (Ebenaceae) 56 energy flow, and food webs 238–52 ebony tree (Diospyros ebenum) 56 energy turnover 223, 253 Echinochloa 19 Enhanced Vegetation Index (EVI) 94 Echinochloa colona 19 Enneapogon cenchroides (Grey head grass) 19, 39t Echinochloa colonum (Barnyard grass) 39t Enteropogon macrostachyus (Bush rye) 39t Echinochloa pyramidalis (Antelope grass) 39t Equidae family see zebras (Equidae family) ecological features of savannahs 8t Eragrostis 16, 19 ecological niches 211, 211f Eragrostis lehmanniana (Lehmann’s lovegrass) 39t ecotone 200 Eragrostis patens (Desert lovegrass) 39t ectoparasites 235 Eragrostis superba (Masai lovegrass) 39t ectotherm vertebrates, production efficiency 243 Eriocephalus 18 Egyptian cobra (Naja haje) 65 Ethiopia 16 Egyptian goose 253 Ethiopian wolf (Canis simensis) 17, 131, 208 Egyptian mongoose (Herpestes ichneumon) 130, 130t Etosha National Park (Namibia) 21, 119, 209, 227 Egyptian thorn (Acacia nilotica)(Vachellia Euarchontoglires 74 nilotica) 44, 48 Euarchordata and Glires 74 Egyptian vulture (Neophron percnopterus) 72, 73 Eucalyptus 6, 7 eland (Taurotragus oryx) 17, 20, 22, 84f, 95–7, 193, 207 Euclea 64 biomass-rainfall relationship 169f Euclea pseudebenus 56 Masai Mara ecosystem 152, 153f Euphorbia 16, 34 population numbers in the Kruger National Park Euphorbia family (Euphorbiaceae) 53–4 in relation to rainfall 176f Euphorbia guerichiana 18 316 INDEX Eurasian otter (Lutra lutra) 133 freshwater habitats 70 Europe 9, 42, 264 fringe-eared oryx (Oryx gazella callotis) 99, 101 excretion 242 frogs 7t extinct in the wild species 81 fundamental niche 211 extinction 258, 261 extinction curves 258, 259 G extinct species 81, 255 gallery forests 24 Gazella thomsonii albonotata (Mongalla gazelle) F 109 fatty acids 77, 78 gazelles (Gazella) 75, 77f, 108–10, 197 Faurea saligna 21 see also Grant’s gazelle (Nanger granti); Thomson’s Faurea speciosa 21 gazelle (Eudorcas thomsonii) feeding facilitation, and grazing succession 228–30 geese 253 feeding height stratification 224 gemsbok (Oryx gazella gazella) 18, 20, 99, 101, 171t, Felidae family (cats) 123–7, 197, 198, 207 268 felids 208 genets (Viverridae family and Herpestidae) 129–30 feline coronavirus 207 miombo genet (Genetta angolensis) 22 feline genet (Genetta felina) 129 geophytes 31, 32, 32t, 33, 36 feline immunodeficiency type viruses 207 Geosphere—Biosphere Programme (IGBP) 238 fencing 262–3 geoxyles 36 fennec fox (Vulpes zerda) 131 Gerbillus bottai 16 fever tree (Acacia xanthophloea) 33 Gerbillus muriculus 16 fire 22, 33, 36, 63, 186, 191, 214, 247 Gerbillus nancillus 16 elephants, trees and 190 Gerbillus stigmonyx 16 in miombo woodland 22 gerbils (Gerbillus) 16 reduction of risk 186 gerenuk (giraffe necked antelope)(Litocranius see also burning walleri) 17, 78, 108, 211f Flacourtia indica 21 giant anteater (Myrmecophaga tridactyla) 3, 4 flamingos 277 giant buffalo (Pelorovis antiquus) 255 fleeing 204, 207 giant diospyros (Diospyros abyssinica) 56 floral richness 25, 25f giant eland (Taurotragus derbianus) 19, 24 florets 35 giant hartebeest (Megalotrogus priscus) 255 flowers 35, 36 giant river otters (Pteronura brasiliensis) 4 food chains 238 giant sable (Hippotragus niger variannii) 102 food supply 155, 162, 164, 168, 185 giant squirrels 135 food webs 243–52 giraffe (Giraffa camelopardalis) 88–9, 173, 211f, 224, cascading interactions 246 232, 247 complexity and stability 243, 244 Acacia and 34, 44 energy flow and 238–52 biomass–rainfall relationship 169f inside park submodel (Serengeti) 249 competitive interactions 213 outside park submodel (Serengeti) 250 feet 75 predator submodel (Serengeti) 249 Masai Mara ecosystem 152, 153f ungulate submodel (Serengeti) 249 population numbers in the Kruger National Park vegetation submodel (Serengeti) 249 in relation to rainfall 176f foot and mouth disease 210 predation rates 171t forest buffalo (red buffalo)(Syncerus caffer nanus) 89 predator–prey interactions 197, 198 forest cats (Felis) 4 predator removal 174f forest elephant (Loxodonta cyclotis) 116 rumination 75 forest genet (Genetta maculata) 129 skull 80f forests 14 social network analysis 90–2 forest-savannah mosaic 23–4 survival rates 178t foxes 132 teeth 79 frankincense 16, 53 giraffe-necked antelope (gerenuk)(Litocranius freezing 205 walleri) 17, 78, 108, 211f INDEX 317 Giraffidae family see giraffe Giraffa( camelopardalis) reedbuck 102 Global Change and Terrestrial Ecosystems roughage 79 (GCTE) 238 rumens 253 global warming 247 seasonal intake of protein and energy 244f Glossina 68 spiral-horned antelopes 97 Glossina palpalis 68 teeth 79–81 GmmMGP 68 Thomson’s gazelle 109 goats 232 topi 79, 105 golden jackal (Canis aureus) 131, 197f waterbuck 78, 102 Gombe Stream National Park (Tanzania) 23 wildebeest 78 gorilla (Gorilla gorilla) 133 see also zebras (Equidae family) Gran Sabana (Venezuela) 5 grazing 33, 184, 228–30, 245t Grant’s gazelle (Nanger granti) 17, 100f, 108–9, 152, cyclic 186 153f, 169f, 170, 221t, 224f dry season 185 grass 16, 19, 38–41, 39–40t, 41f, 184 lawns 185, 229 branching 35 McNaughton’s grazing model 184f competition between trees and 213–16 resistance of vegetation to 246f cover 16 succession and feeding facilitation 228–30 Eragrostis lehmanniana (tufted perennial grass) 28 wet season 185 florets 35 greater honey guide (Indicator indicator) 133, 227 fruit 35 greater kudu (Tragelaphus strepsiceros) 17, 18, 20, 22, grain 35 97–8, 169f, 170, 217 greenness 221 Greater Limpopo Transfrontier Park 263 height 186, 211, 221 Great Ruaha river 275 inflorescence 41 green-hair tree (Parkinsonia africana) 51 long 66 green monkey (Chlorocebus sabaeus) 134 medium-height 228 Grevillea family (Proteaceae) 42 production 66, 67, 164, 243, 250 Grevy’s zebra (Equus grevyi) 17, 83, 84f, 90, 273 rainfall, plant biomass and grass–tree mixture 27– Grewia tenax 193, 194 30 Grewia villosa 20 Serengeti 186t grey-haired acacia (Gerrard’s thorn) (Acacia short 66, 228 gerrardii)(Vachellia gerrardii) 46, 64 and shrub savannah 15–18 griffin vultures (Gyps species) 72, 235 stem 34 grivet monkey (Chlorocebus aethipos) 134 stoloniferous perennial grass (Tragus ground pangolin (Smutsia temminckii) 22 koeleriodes) 28 ground squirrels 135 tall 102, 104, 126 group living 121, 123 Themeda triandra (tufted perennial grass) 28 group size 201, 203, 204, 206f tufted 18, 28 Guayanas 5 tussock 34, 36 Guenther’s dik-dik (Madoqua guentheri) 111–12 grasshoppers (Orthoptera) 6, 7t, 66–7, 242, 252 Guibourtea coleosperma 20, 21 grazers 22, 78, 222, 223f guiding behaviour 227 Australia 6 Guinea 23 body size 253, 254f Guinea baboon (Papio papio) 135 buffalo 78, 89 guinea fowls 69 bulk 78 Guinean forest–savannah mosaic 23 community pattern 255 Guinean savannahs, Raunkier life forms 32t cyclic 186 gundis 134 hartebeests 79, 105 hippopotamus 87 H horse antelopes 101 habitat impala 107 destruction and wildlife–human conflict 269–76 kob 103 diversity 259 pigs see pigs (Suidae family) hammerkop (Scopus umbretta) 69 318 INDEX hares 198 horse antelopes (Bovidae family, Hippotragini hartebeests (Bovidae family, Alcelaphini tribe) 79, tribe) 99–102 104–8, 153f, 169f horses 6, 75, 77 Coke’s 105, 152, 221, 221t hosts, parasites and 207–10 harvester termite (Hodotermes mossambicus) howler monkey (Alouatta caraya) 3 132 human African trypanosomiasis (HAT) 68, 69 Hausa (Villier’s) genet (Genetta thierryi) 129 human population growth (HPG) 250 heart-fruit (Hymenocardia acida) 53 Human sleeping sickness 68 height (trees) 19, 30, 33 hunting 265–9 helmet-shrikes 69 cursorial 121 hemicryptophytes 31, 32, 32t, 33 schemes 278 herbaceous plant species 3, 30 stealthy style 121 herbivores 1, 38 hyaenas (Hyaenidae family) 127–9, 171t, 198, 208, Acacia and 228 209, 263 diet 115 see also brown hyaena (Hyaena brunnea); spotted insect 7 hyaena (Crocuta crocuta); striped hyaena interactions between plants and 182, 184–94 (Hyaena hyaena) invertebrate 1 Hymenocardia acida 24 niche differentiation among 217–25 Hyparrhenia 19, 24, 40 South America 5 Hyparrhenia cymbaria (coloured hood grass) 39t tree recruitment 192 Hyparrhenia diplandra (sword hood grass) 39t vertebrate 1 Hyparrhenia dissoluta (yellow hood grass) 39t herbivory 186, 214 Hyparrhenia filipendula (fine hood grass) 39t herb layer biomass 27, 28 Hyparrhenia rufa (brown hood grass) 39t herbs 229 hyraxes 112 Herpestidae family (gents, civets and mongooses) 129–30 I Heteropogon contortus (Spear grass) 39t immigration 258 hiding 205 curves 258 highveld protea (Protea caffra) 42 impala (Aepyceros melampus) 17, 20, 100f, 107–8, hindgut fermentation 75, 77 173, 193, 224 Hippopotamidae family (hippopotamuses) 87–8 biomass–rainfall relationship 169f Hippopotamus gorgops 255 body weights 224f hippopotamus (Hippopotamus amphibius) 20, 22, 75, counted during the Mara transects 221t 84f, 169f, 217 defences against predation 201 anthrax 210 density in predator removal area 174f feet 75 feeding preference of 188f habitat destruction and wildlife–human Masai Mara ecosystem 153f conflict 269t mixed feeding 78 skull 80f mutualistic interactions 235, 236 teeth 79 niche separation 224 Hippotragini tribe (horse antelopes) 99–102 population numbers in the Kruger National Park Hippotragus niger kirkii 102 in relation to rainfall 176f Hippotragus niger niger 102 predation rates 171t Hippotragus niger roosevelti (horse antelope) 102 survival rates 178t Holling’s disc equation 165, 248 vigilance in 202f honey badger see ratel (honey badger)(Mellivora Imperata cylindrica (cotton grass) 39t capensis) India 1, 7 honeybees (Apis) 66 Indian rhinoceros (Rhinoceros unicornis) 7 honeyguides 69 Indigofera basiflora (herb) 61 hooded vulture (Necrosyrtes monachus) 72, 73 Indigofera schimperi 20 hook-lipped rhinoceros see black rhinoceros (hook- ‘inhibiting’ 191 lipped)(Diceros bicornis) insectivores 22 Horn of Africa 17 insects 65–9 INDEX 319 inselbergs 66 Kavango Zambezi (KAZA) Transfontier inside park submodel (Serengeti) 249 Conservation Area 262 interactions, between species 137 Kenya 166, 262, 277, 278 intermediate grasslands (zone III) 61 tourism figures for 273t International Biological Programme (IBP) 238 Wildlife Service 272 internodes 34 see also Laikipia (Kenya) interspecific competition 211, 212, 214, 217, 220, ‘key factor’ 159 230 keystone species 251 mutualistic interactions 217–18 k factor analysis 137, 158f, 159, 161, 163 resource competition 217–25 Khaya grandifoliola 24 tree and grass competition 213–16 Kigelia aethiopica 19 intertropical convergence zone (ITCZ) 10–12, 10f, ‘killing’ 171, 172, 191 12f, 57 killing power (k value) 157 intraspecific competition 210, 215, 219 kill rate 249 invasive plants 216–17 Kirkia acuminata 19 invertebrates 6 Kirk’s acacia (Acacia kirkii)(Vachellia kirkii) (flood- production efficiency 243 plain acacia) 46–7 Iridaceae (irises) 36 Kirk’s dik dik (Madoqua kirki) 18, 111–12, 113f island biogeography 257–9 Klaserie Private Nature Reserve (South Africa) 92 dynamic theory of 258 kleptoparasitism 225–7, 234 equilibrium theory of 259f klipspringer (Oreotragus oreotragus) 78, 217 protected savannah areas 259–64 kob (Kobus kob) 103–4, 169f Isoberlinia angolensis 21 Konkola copper mines (Zambia) 274 Isoberlinia doka 23 kopjes 66 isotope-ratio mass spectrometry (IRMS) 115 Kordofan giraffe Giraffa( camelopardalis IUCN 86, 87, 95, 99, 101, 102, 124, antiquorum) 89 126, 135 korrigum (Damaliscus lunatus korrigum) 105 African Antelope Database 262 Kruger National Park (South Africa) 21, 38, 92, 101, conservation categories 81–2 173, 179 ivory 267 baobab trees 55 browsers 188, 224 J greater kudu 170 Jacaranda family (Bignoniaceae) 56–7 herbivore populations 195 jacaranda tree (Jacaranda mimosifolia) 56–7 marula (Sclerocarya birrea) 54 jackals (Canidae family) 131–2, 198, 208 poaching 263 Jackson’s hartebeest (Alcelaphus buselaphus pollution in the Olifants river 274 jacksoni) 17, 105 predator–prey interactions 172, 196 jaguar (Panthera onca) 3, 4 prey 128, 131 Jatropha campestris 20 ungulates 176f jaw muscles 79 white rhinoceros reintroduction to 217 jerboas 135 wildlife tourism 277–8 Julbernardia 20 kudu 20, 97–9, 170, 209 Julbernardia globiflora 21 Acacia 228 Julbernardia paniculata 21 competitive interactions 213 feeding preference of 188f K niche separation 224 Kafue National Park (Zambia) 198, 273, 274 population numbers in the Kruger National Park Kalahari 128, 131 in relation to rainfall 176f Kalahari sand terminalia (Terminalia predation rates 171t brachystemma) 56 predator–prey interactions 263 kangaroo 6, 8 survival rates 178t karoo bushes/shrubs 18 see also greater kudu (Tragelaphus strepsiceros); Kasungu National Park (Malawi) 269 lesser kudu (Tragelaphus imberbis) Katavi National Park (Tanzania) 256f, 257t Kyllinga 61 320 INDEX L Lewa Downs (Kenya) 273 Laikipia (Kenya) 232, 269, 270, 272, 272t Lewa Wildlife Conservancy (LWC) 273 Laikipia Wildlife Forum 272 Lichtenstein’s hartebeest (Alcelaphus Lake Edward plain 36 lichtensteinii) 20, 22, 105 Lake Manyara National Park (Tanzania) 155, 170, life forms 31–6 171, 186, 256f, 257t annuals 36 Lake Nakuru National Park 262, 277 perennials with above ground seasonal Lake Victoria 58, 63 vegetation 34–6 lamina (grass) 34 perennials with above ground woody structures lammergeier (bearded vulture)(Gypaetus barbatus (trees and shrubs) 33–4 meridionalis) 73 ligule (grass) 34 Lamto savannahs (West Africa) 33 Lily family (Liliaceae) 36, 42 land use 151, 250, 276 Limpopo National Park (Mozambique) 54–5 Lannea humilis 215 Lincoln–Peterson estimate 143 Lannea schimperi 19 lion (Panthera leo) 123–4, 171, 179, 209, 235, 262, lappet-faced vulture (Torgus tracheliotus) 73 263, 277 large false mopane (Guibourtia coleosperma) 52 canine distemper virus 208 large-leaved euclea (Euclea natalensis) 56 feline immunodeficiency type virus 207 large-leaved munondo (Julberlandia paniculata) 52 forest-savannah mosaic 24 large-leaved terminalia (Terminalia mollis) 56 grass and shrub savannah 17, 18 large sour plum (Ximenia caffra) 42 identifying individuals 139 large-spotted genet (Genetta tigrina) 122f, 129 kleptoparasitism 226 larks 69 Lake Manyara National Park 171 Laurasiatheria 74 manes 124 leadwood (Combretum imberbe) 55 metal tags 140 least concern species 82, 126, 130 population estimates 179t aardwolf 127 predation rates on large herbivores 171t buffalo 95 predator–prey interactions 196, 197, 198, bushbuck 99 199t Grant’s gazelle 109 roar 119–20, 124 Kirk’s dik dik 112 skull 122f puku 104 tree and shrub savannah 19, 20 reedbucks 104 trophy hunting 278 roan antelope 102 weight 121 sable antelope 102 weights of mammal prey consumed by savannah baboon 135 197f springbok 110 woodland savannah 22 warthogs 87 livestock ranches 272 wildebeest 107 lizards 7, 7t leaves 34, 37 llanos del Orinoco 4, 5 lechwe (Kobus leche) 22, 198 llanos de Moxos 4 Leguminosae (pod-bearing family) 43–52 locusts 42 lelwel (Alcelaphus buselaphus lelwel) 104–5 long grass plains (zone IV) 61, 66 leopard (Panthera pardus) 17, 18, 19, 20, 22, 24, 125 Lotka–Volterra model 194, 211, 214 estimating population of 143 Loudetia 24 and feline immunodeficiency type viruses 207 Loudetia kagerensis (sour russet grass) 39t identifying individuals 139 Loudetia simplex (common russet grass) 39t population estimates 179t lovebirds (Agapornis sp.) 143 predation 171t, 199t lower risk species 86, 104, 105 weights of mammal prey consumed by 197f Lower Zambezi National Park (LZNP) 273 leopard tortoise (Geochelone pardalis) 65, 253 Loxodonta atlantica 255 Lepus capensis 198 Luangwa Valley (Zambia) 208 Lepus saxatilis 198 Lutra maculicollis (otter) 133 lesser kudu (Tragelaphus imberbis) 17, 78, 98–9 lynx (Felis lynx) 194 INDEX 321 M miombo (msasa)(Brachystegia spiciformis) 51–2 macropods 6 miombo (Muuyombo) tree (Prince-of-Wales’ mahobohobo (Uapaca kirkiana) 54 feathers)(Brachystegia boehimii) 21, 51 malaria 67–8 miombo woodland savannah 20, 21–2, 25, 33, 42 Malawi 19, 21, 208, 274 missing species 255 malbrouk monkey (Chlorocebus cynosuros) 134 mixed feeders 78, 222 malnutrition 171, 177, 210 Mkomazi Game Reserve (Tanzania) 18, 37, 66, 146t, see also under nutrition 232, 256f, 257t mammals 74–136 mobola family (Chrysobalanaceae) 43 Afrotheria 74 mobola plum (Parinari curatellifolia) 43 carnivores 116–33 MODIS IGBP Land Cover Classification 14 Euarchontoglires 74 Mongalla gazelle (Gazella thomsonii albonotata) forest-savannah mosaics 24 109 Indian savannahs 7 mongooses (Viverridae family and Herpestidae) 121, Kakadu National Park (Australia) 7t 130, 130t Laurasiatheria 74 Cynictis penicillata 198 non-ruminant ungulates 82–8 monkeys 3 primates 133–5 monocotyledons 41, 78 rodents 135–6 Monocymbium ceresiiforme (wild oatgrass) 39t ruminant ungulates 88–116 Monotes glaber 21 tree and shrub savannah 19 montane forest 16 ungulates 74 mopane tree (Colophospermum mopane) 19, 20, 52 Xenarthra 74 Moringa ovalifolia 18 maned wolf (Chryocyon brachyurus) 3, 4 morphology and life history 30–8 mango family (Anacardiaceae) 54 life forms 31–6 Manipur brow antler deer (Cervus eldii eldii) 7 phenology 37–8 Manyara National Park (Tanzania) 198 mortality 154, 156 ‘many eyes’ theory (increased vigilance) 201 density dependent 154 marabou stork (Leptoptilos crumeniferus) 234–5 density independent 154 Maranthes polyandra 24 rates 156 Marquesia macroura 21 young versus adult 173 marsh deer (Blastocerus dichotomus) 5 Morus mesozygia 24 marula (Sclerocarya birrea) 54 mosquito 67 Masai giraffe Giraffa( camelopardalis tippleskirchi) 89 mountain acacia (Brachystegia glaucascens) 52 Masai Mara ecosystem 128, 151, 218, 220, 229 mountain gorilla (Gorilla gorilla beringei) 275 Acacia woodland savannah in 191f mountain reedbuck (Redunca fulvorufula) 104 population trends for 12 wildlife species in 153f mountain zebra (Equus zebra) 18, 20, 268 Masai Mara Game (National) Reserve (Kenya) 18, mousebirds 69 151, 173, 207, 277 Moyowosi-Kogosi Game Reserve 256f, 257, 257t Massingir dam, Mozambique 274 Mozambique 277 mathematical description/model 137, 163, 190, 244 msasa 51–2 mean annual precipitation 2f, 29, 30f, 213–14 muhutu (Terminalia brownii) 56 Mediterranean Coastal Biotic Zone 14 multiple mark–release–recapture methods 143 meerkat (Suricata suricata) 130t munondo (Julberlandia globiflora) 52 megaherbivores 270 Murchison Falls National Park (Uganda) 24 Melanthera marlothiana 20 Muridae family 217 Mesopsis abbreviatus 66 Mustelidae family 133 mesquites (Prosopis spp.) 51 mutanga (Elaeodendron buchanii) 56 metal ring tagging 140 mutualistic interactions 182, 227–8 microsatellite analysis 96 Acacia ants 231–4 midgrass savannahs 6 feeding facilitation and grazing succession 228–30 milk proteins 68 scavenging vultures 234–5 Mimosoideae 43 myrrh 16, 53 miombo genet (Genetta angolensis) 22 myrrh family (Burseraceae) 53 322 INDEX N not threatened species 82 nagana 68 Nubian giraffe Giraffa( camelopardalis Nairobi National Park (Kenya) 18, 105 camelopardalis) 89 Nakuru National Park (Kenya) 236 numbers estimation 138–53 Namibia 19, 21, 30, 207, 216, 262, 268 aerial surveys 146–50 see also Etosha National Park (Namibia) African buffalo 155–61 narrow-leaved mahobohobo (Uapaca nitida) 54 artificial markings 140–2 natural experiments 172, 213, 217, 218 multiple mark–release–recapture methods 143 natural markings 139–40 natural markings 139–40 Nazinga Game Ranch (Burkina Faso) 150 removal method 143–4 near threatened species 86, 99, 101, 106, 125, 128 road transects 144, 147, 151t nematodes 160 wildebeest 161–3 Neotragini tribe (gazelles and dwarf antelopes) nyala (Tragelaphus angasii) 20 108–12 Nyami Nyami district, (Zimbabwe) 269 neutralism 182 Nylsvley ecosystem (southern Africa) 241 niche differentiation 217–25, 252 buffalo 217, 224 O Coke’s hartebeest 221, 221t Ochna pulchra 21 giraffe 224 oil exploration 275 Grant’s gazelle 221t, 222 oil palm (Elaeis guineensis) 42 herbivores 217–25 okapi (Okapia johnstoni) 275 hippopotamus 217 Olaceae (sour plum family) 42–3 impala 221t, 222, 224 Olifants river, pollutants 274–5 kudu 224 olive baboon (Papio anubis) 135 oribi 217 open tree savannah 17 steenbok 217, 224 open woodland savannah 64, 108 Thomson’s gazelle 218, 219–20, 221t, 222 Orchidaceae (orchids) 36 topi 221, 221t, 224 oribi (Ourebia ourebi) 19, 22, 169f, 173, 174f, 197, warthog 221t 217 waterbuck 221t Orthoptera (grasshoppers) 6, 7t, 66–7, 242, 252 white rhinoceros 217 oryx (Oryx gazella) 99, 169f, 170 wildebeest 218, 219, 221, 221t Oryza barthii (wild rice) 39t zebra 218, 219, 221, 221t ostrich (Struthio camelus) 69, 70–1, 153f niches otters (Mustelidae family) 133 acoustic niche hypothesis 120 outside park submodel (Serengeti) 250 ecological 211, 211f oxpeckers 69, 235–6 fundamental 211 realised 211 P root 214, 215 palatability 37 separation 215, 221, 224, 224f pale fox (Vulpes pallida) 132 see also niche differentiation palm (Borassus aethiopum) 34 Nile crocodile (Crocodylus niloticus) 24, 65, 274 palm civet (Nandinia binotata) 130 nitrogen 6, 38, 43, 161, 186, 188 palm family (Arecaceae) 42 nob-thorn (Acacia nigrescens)(Senegalia pampas deer (Ozotoceros bezoarticus) 5 nigrescens) 47 Paniceae 40 nodes 34 panicle 35, 41, 51 nodulated legumes 227 Panicoideae 40 nonruminants see ungulates Panicum coloratum (blue guinea grass) 16, 39t Normalized Difference Vegetation Index (NDVI) 92, Panicum maximum (slender guinea grass) 39t, 63 93–4 Panicum turgidum (Taman guinea grass) 16, 39t North African banded weasel (Poecilictis libyca) 133 pansteatitis 274, 275 North America 9, 240, 264 Pantanal, Brazil 3 northern Congolian forest-savannah mosaic 23 paperbark acacia (Acacia sieberiana)(Vachellia not endangered species 89, 108, 134 sieberiana) 50 INDEX 323 Papilionoideae (pea subfamily) 43 poaching 191, 250, 263, 265–9 parasites 160 pod-bearing family (Leguminosae) 43–52 and hosts 207–10 Pogonarthria squarrosa (herringbone grass) 19, 40t parasitic weavers 69 poison-grub commiphora (Commiphora Parinari congensis 24 africana) 16, 19, 53 Parinari curatellifolia 21, 24 polecats (Mustelidae family) 133 Parkia biglobosa 24 Mustela putorius 133 Parkinsonia africana 18 pollination 37, 227 park revenue 249 pollution 274 patas monkey (Erythrocebus patas) 134 Pooideae 40 pathogens 207 population pea subfamily (Papilionoideae) 43 control 181, 181f Peltophorum africanum 20 densities 138 Pennesetum mezianum (grass) 63 sizes 138 Pennisetum 19 population estimates 138–53 Pennisetum clandestinum (kikuyu grass) 39t aerial surveys 146–50 Pennisetum mezianum (bamboo grass) 40t, 61, 186t camera traps 150 Pennisetum purpureum (elephant grass) 35, 40t counting dung 150 Pennisetum schimperi (wire grass) 40t, 83 multiple recapture methods 143 Pennisetum ultramineum (grass) 186t removal method 143–4 pentastomids 160 sample surveys 149 Pentzia 18 total surveys 149 perennials unmanned aerial systems (UAS) 149 with above ground seasonal vegetation 34–6 population models 151, 163–7 with above ground woody structures (trees and buffalo 159 shrubs) 33–4 wildebeest 163–6 with non-lignified storage organs 34 zebra 166–7 with permanent above ground woody porcupines 135 structures 33 potential evaporation 240 with woody underground storage organs 34 PRAAT 117 perissodactyls 75 precipitation 2 pesticides 275 predation 230, 237 PGRP-LB 68 competitive interactions 181 phanerophytes 31, 32t, 33 rates 171t phenological niche separation 215 regulating populations 172 phenology 37–8 predator kills 171 photosynthesis 238–9 predator populations 179 photosynthetically active radiation (PAR) 239 predator–prey cycles 194 photosynthetic carbon reduction pathway (Calvin- predator–prey interactions 182, 183–210, Benson cycle) 1 211, 263 physical attributes 194 herbivores and plants 184–94 physical dimensions 211 parasites and hosts 207–10 phytochorian 21 predators and prey 194–207, 206f pigs (Suidae family) 75, 79, 86–7 predator(s) Piliostigma thonningii 24 attack success 249 plains zebra (Equus burchelli) 60, 166, 171t confusion 201 plant(s) forest-savannah mosaic 24 biomass 27–30 grass and shrub savannah 17 food species separation 221 mutualistic interactions 234 interactions between herbivores and 194 removal 173 parts eaten separation 221 submodel 249 primary productivity 93 tree and shrub savannah 20 Plasmodium 67, 68 pregnancy rate 164, 165 ploceine weavers 69 prey handling time 249 324 INDEX prickly acacia (Acacia brevispica; Senegalia) rainforests 8, 22, 24, 25, 56, 70, 89 (wait-a-bit-thorn) 45 ranching, costs of 272 primary producers 238 raptors 21 primary production 8t, 168, 170, 185, 238, 241, 243 ratel (honey badger)(Mellivora capensis) 133, 227 above ground 185f, 241 rat-like murids 135 efficiency 243 realised niche 211 gross 238 reciprocal negative intraspecific effects 182 net 238, 239t, 240, 240f, 241f red-billed quelea (Quelea quelea) 74 primates 22–3, 133–5, 269 red bushwillow (Combretum apiculatum) 55 proboscid (Deinotherium bozasi) 255 red colobus (Procolobus species) 23 Prosopis 216–17 red fox (Vulpes vulpes) 131 Prosopis africana 19, 51 red-fronted gazelle (Eudorcas rufifrons) 17 Protea caffra (highveld protea) 42 red hartebeest (Alcelaphus buselaphus caama) 105 Proteaceae (Grevillea family) 42 red kangaroo (Macropus rufus) 6 protected areas 259–64 red-leaved rock fig Ficus( ingens) 37 protection 227 red-tailed monkey (Cercopithecus ascanius) 23 protein 49, 68, 69, 78, 93, 94, 103, 223, 253 reduction (population) 156, 159 crude 78t, 160, 161 Reduncini tribe see kob (Kobus kob); reedbucks milk 68 (Bovidae family, Reduncini tribe); waterbuck shortage 243 (Kobus ellipsiprymus) turnover 223 reedbucks (Bovidae family, Reduncini tribe) 104 Proteus 42 Reeve’s muntjac (barking deer)(Muntiacus reevesi) 7 Pseudocedrela kotschyi 24 reintroductions 217 pseudogalls 231, 233 relative population estimates 138 Pteleopsis anisoptera 20 relaxation 259 Pterocarpus antunesii 20 removal method 143–4 Pterocarpus erinaceus 24 resource competition see niche differentiation Pterocarpus santalinoides 24 resource dimensions 211, 252 puff adder (Bitis arietans) 65 resource utilisation curves 212 puku (Kobus vardonii) 22, 104, 198 respiration 238, 241, 242 purple-leaved albizia (Albizia antunesiana) 51 reticulated giraffe (Giraffa camelopardalis purple-pod terminalia (Terminalia prunoides) 56 reticulata) 17, 84f, 89, 140f, 270 pygmy hippopotamus (Hexaprotodon liberiensis) 87 identifying individuals 139–40, 140f social network analysis 90 Q tree damage 189 quadrats 150 ‘reversing’ 191 Queen Elizabeth National Park (Uganda) 24, 210, 217 Rhaphotitha 67 quiver tree (Aloe dichotoma) 18, 42 Rhigozum 57 Rhigozum brevispinosum 20 R Rhigozum trichotomum (karoo shrubs) 18 rabies 207–8 Rhigozum virgatum 20 radio transmitter tracking 141–2 rhinoceros (Rhinocerotidae family) 75, 77, 81, 83–6 raid frequency on farms 269, 269t see also black rhinoceros (hook-lipped)(Diceros rainfall 2, 3, 9, 11, 25, 214 bicornis); white rhinoceros (square-lipped) Australia 5–6 (Cerathotherium simum) and buffalo 170, 176f Rhodesian teak (Baikiaea plurijuga) 51 competitive interactions 214 Rift Valley fever 210 dependence 167 rinderpest 170, 190, 208, 225, 246, 247 -mediated density dependence 167 buffalo and 160, 170 plant biomass and grass–tree mixture 27–30 epidemics of 209t and species richness patterns 25 in the Serengeti 159, 160, 161, 170 and survival 175 wildebeest and 161, 170, 179, 218, 247 and ungulate biomass 168 see also disease see also mean annual precipitation Rio Branco-Rupunini, Brazil 5 INDEX 325 rivers, effects on large carnivores 263 scavengers 227 road transects 144, 145, 147, 151t scavenging vultures 234–5 roan antelope (Hippotragus equinus) 19, 20, 101–2, Schizachyrium semiberbe (crimson bluestem) 40t 169f, 170, 176f, 178t Schmidtia bulbosa (zand kweek grass) 40t rockfowls 69 Schmidtia pappophoroides 19 rock wallaby 6 Schoenefeldia gracilis (spiky grass) 16, 40t rodents 18, 135–6, 269 scimitar oryx (Orya dammah) 16 root niches 214 seasonal habitat changes 198 roots 35 secondary consumption 242f rope squirrels 135 secretary bird (Sagittarius serpentarius) 69, 71 Rothschild’s giraffe Giraffa( camelopardalis Securinega virosa 20 rothschildi) 89, 262 seed dispersal 227, 264 roughage eaters see grazers selective eaters see browsers Ruaha National Park (Tanzania) 18, 256f, 257, 257t, Selous Game Reserve (Tanzania) 23, 202, 203, 256f, 275 257, 258, 273 Rukwa valley (Tanzania) 228 group (herd) size in 205f ruminants weight ratios 257t digestion 227 Selous’s mongoose (Paracynictis selousi) 22 teeth and jaw muscles 79–81 semi-arid bush savannah 232, 272 see also ungulates Senegalia 43, 44 rumination 75 separation by habitat 221 Ruppell’s fox (Vulpes ruppelli) 131 Serengeti 16, 131, 170, 220, 223, 229 Ruppell’s griffon vulture Gyps( ruepellii) 72, 73 annual grass production in 66 migrating wildebeest 218–21 S model 249 sable antelope (Hippotragus niger) 20, 22, 100f, 102, protein in grass species in 78t 169f, 170, 178t rank abundance of ungulates in 251f safari parks 18 weights of mammal prey consumed by carnivores Sahel 15, 16 in 197f Sahelian savannahs, Raunkier life forms 32t wild dog (Lycaon pictus) 226f sambar (Cervus unicolor) 7 Serengeti-Mara ecosystem 185f, 190 Samburu National Reserve (Kenya) 18, 114 model 248 sarcoptic mange 210 predation and population numbers in 172 sausage tree (Kigelia africana) 57 rainfall and green biomass 57–60 savannah baboons (Papio species) 23, 122f, soils, fire and vegetation 61–4 134–5 trophic web 247f savannah buffalo Syncerus( caffer caffer) 89 ungulate migrations in 219f savannah elephant (Loxodonta africana) 112–14 vegetation patterns 57–64, 63f savannah foxes 132 Serengeti National Park (Tanzania) 17, 18, 155, 179, savannahs 1–26 207, 227, 245, 256f Africa 5, 9–26 effects of hunting and poaching 265 Australia 5–7 group (herd) size in 205f climatic patterns 9–14 predator composition in 180f comparison of major savannahs 7–8 road building across 276 distribution of 2–3, 3f tourism 266 forest-savannah mosaic 23–4 weight ratios 257t grass and shrub 15–18 serval (Leptailurus serval) 22, 113f, 126, 197f India 7 Setaria incrassata (setaria) 16, 40t profiles 15f Setaria sphacelata (common setaria) 40t South America 3–5 sharptooth catfish 275 species richness patterns 24–6 Shepherd’s tree (Boscia albitrunca) 43 tree and shrub 18–21 Shimba Hills National Park (East Africa) 102 vegetation patterns 14–15 shoebill (Balaeniceps rex) 69 woodland 21–3 short grasslands (zone II) 61 326 INDEX shrikes 69 South African ground squirrel (Xerus inauris) 198 shrub and grass 21 South America 1, 3–5, 8, 8t shrub and woodland savannah 62 southern giraffe (Cape giraffe)(Giraffa camelopardalis shrubs 33–4 giraffa) 89 shrub savannah (bushveld) 17, 20 southern reedbuck (Redunca arundinum) 22, 169f sickle bush (Dichrostachys cinerea) 51 South Luangwa National Park (Zambia) 23 sickle-leaved albizia (Albizia harveii) 50 species–area relationships 138, 258, 260f, 261, 262, side-striped jackal (Canis adustus) 20, 22, 131 264 silver terminalia (Terminalia sericea) 56 species diversity 138, 245t single-species populations 137–81 species interactions 182–236 arid/eutrophic savannahs 168 amensalism 182 biomass–rainfall relationship 168, 169f, 181 apparent competition 182, 183f changes in 153–63 commensalism 182 disease 159 competition interactions 182, 210–13 estimating animal and plant numbers 138–53 herbivore–plant 182 and rainfall 168 mutualism 182, 227–8 soil nutrients 168 neutralism 182 see also numbers estimation; population models parasites and hosts 207–10 sitatunga (Tragelaphus spekii) 22 predator–prey 182, 183–210 skulls species richness 137 of carnivores 122f patterns 24–6 of ungulates 80f Speke’s gazelle (Gazella spekei) 17 slender mongoose (Herpestes sanguineus) 130t sphingomyelinase 68 small false mopane (Guibourtia conjugata) 52 spikelets 35 small sour plum (Ximenia americana) 42 spines 43 small-spotted genets (Genetta genetta) 198 spiral-horned antelopes (Bovidae family, smelly boscia (Boscia foetida) 43 Strepsicerotini tribe) 95–9 snakes 7t, 65 Spirostachys africana 20 snowberry tree (Securinega virosa) 53 splendid acacia (Acacia robusta)(Vachellia snowshoe hares (Lepus americanus) 194 robusta) 48 social network analysis 90–2 Sporobolus consimilis 61 soft-leaved commiphora (Commiphora mollis) 53 Sporobolus marginatus (grass) 61, 186t soil 14, 17, 22 Sporobolus nitens 41f alkaline 19 Sporobolus pellucidus (dropseed) 40t alluvial 19 Sporobolus pyramidalis (Pyramid dropseed) 40t Australia 6 Sporobolus robustus (reed dropseed) 40t moisture 14, 28 Sporobolus spicatus (spike dropseed) 40t, 200 nutrients 168 spotted hyaena (Crocuta crocuta) 17, 20, 22, 113f, rainfall and 14 122, 128–9, 179, 181, 209 Serengeti 61–4 antibodies 207 volcanic 16 canine distemper virus 208 Solanum incanum (herb) 61 culling in Kruger National Park 172 Solanum panduraeforme 188 ear notch markings 141 solar radiation 238–9, 239f kleptoparasitism 225 Somalia 16 population estimates 143, 179t Somali-Masai dry savannah 16 predator–prey interactions 196, 199t Sorghum 19 skull 122f Sound Analysis Pro (SAP) 117 weights of mammal prey consumed by 197f soundscape ecology 120 springbok (Antidorcas marsupialis) 18, 110, 169f, source–filter theory 118 171t, 217, 268 sour plum family (Olaceae) 42–3 spring hare 135 sourveld, southern Africa 38 spur-winged goose 253 South Africa 20, 216, 273t square-lipped rhinoceros see white rhinoceros see also Kruger National Park (South Africa) (square-lipped)(Cerathotherium simum) INDEX 327 squirrels 135 mounds 43, 61 stable communities 244 Tetraponera penzigi (ant) 231f, 233 stalking and/or ambush 121 theileriases 160 starvation 81, 171, 234 theileriosis 210 steenbok (steinbok)(Raphicerus campestris) 20, 111, Themeda triandra (red oat grass) 16, 40t, 61, 188 188f, 205, 217, 224 therophytes 31, 32t, 33 Stereospermum kunthianum 19, 24 Thomson’s gazelle (Eudorcas thomsonii) 17, 60, Stipagrostis 18 109–10, 219–20, 221, 221t Stipagrostis uniplumis (silky bushman grass) 19, 40t biomass–rainfall relationship 169f Strepsicerotini tribe (spiral-horned antelopes) 95–9 body weights 224f striped hyaena (Hyaena hyaena) 17, 22, 128 feeding facilitation and grazing succession 228, striped polecat (zorilla)(Ictonyx striatus) 133 229, 229f, 230 striped weasel (Poecilictis libyca) 133 Masai Mara ecosystem 152, 153f Struthioniformes 69 niche differentiation 218, 219–20 Strychnos 21, 24 population estimates of 219f Strychnos cocculoides 21 predation rates 171t Strychnos spinosa 21 predator–prey interactions 201 subshrubs 36 predator removal and 173, 174f subungulates 112–16 Thornicroft’s giraffe (Giraffa camelopardalis Sudan–Zambezian savannahs, Raunkier life thornicrofti) 22, 23, 89, 90 forms 32t thorns 43 sugarbirds 69 thorn trees 43 Suidae family (pigs) 75, 79, 86–7 thorn-tree zone 64 sun squirrels 135 three-thorned (gum arabic acacia)(Acacia senegal) survival rates 175, 178t (Senegalia senegal) 49 survivorship curves 166 tiang (Damaliscus lunatus tiang) 105 Swayne’s hartebeest (Alcelaphus buselaphus tick-borne diseases 210 swaynei) 105 ticks 160 sweetveld (southern Africa) 38 tillering 35 time foraging 201 T time vigilant 201 Tamarindus indica 19 topi (Damaliscus lunatus) 79, 100f, 152, 170, 174f, tannins 188, 188f 221, 221t, 224 tantalus monkey (Chlorocebus tantalus) 134 biomass–rainfall relationship 169f Tanzania 16, 101, 273t, 278 body weights 224f see also Lake Manyara National Park (Tanzania); feeding facilitation and grazing succession 228 Mkomazi Game Reserve (Tanzania); Selous population trend in the Masai Mara Game Reserve (Tanzania) ecosystem 153f Tarangire National Park (Tanzania) 17, 18, 256f, 257, topi (Damaliscus lunatus jumela) 105, 173 257t tora (Alcelaphus buselaphus tora) 105 Techlea 64 tourism 191, 265, 266, 269, 270, 273t teeth 79–81, 116–21 quality 249 Temminck’s ground pangolin (Smutsia temminckii) 22 wildlife tourism 277–8 temperature 2, 7, 9, 12f, 25, 32, 57, 97, 101, 155 Tragus koelerioides (creeping carrot grass) 40t Australia 5 Transvaal reserve (South Africa) 110 vegetation patterns 14 tree and shrub savannah 18–21 Terminalia 19, 23, 24 ‘tree population’ model 191 Terminalia brachystemma 21 trees 33–4, 41–57 Terminalia family (Combretaceae) 55–6 Baobab family (Bombacaceae) 54–5 Terminalia mollis 64 bark thickness 33 Terminalia prunoides 20 Caper family (Capparaceae) 43 Terminalia sericea 20 competition between grass and 213–16 termites (Cubitermes sankurensis) 1, 6, 7t, 61, 66, 144 cover 30, 214 miombo woodland savannah 22 desert date family (Balanitaceae) 52–3 328 INDEX trees (continued) diet 222f ebony family (Ebenaceae) 56 feet 76f euphorbia family (Euphorbiaceae) 53–4 feral 6 Grevillea family (Proteaceae) 42 gazelles and dwarf antelopes 108–12 herbivore damage on 189, 189f giraffes 88–9 Jacaranda family (Bignoniaceae) 56–7 grazers 78 leaf fall 34 group (herd) size in 205f leaves 34 hartebeests, topi, wildebeests and impala 104–8 Lily family (Liliaceae) 36 hindgut fermentation 76, 77f, 79 mango family (Anacardiaceae) 54 hippopotamuses 87–8 Mimosoideae 43 horse antelopes 99–102 mobola family (Chrysobalanaceae) 43 interspecific competition 217 myrrh family (Burseraceae) 53 kob (Kobus kob) 103–4, 169f palm family (Arecaceae) 42 mixed feeders 78 pod-bearing family (Leguminosae) 43–52 non-ruminants 82–8 rainfall and growth of 28 perissodactyls 75, 81 recruitment 192, 193, 216 pigs 75, 79, 86–7 roots 34 reedbucks 104 sour plum family (Olaceae) 42–3 rhinoceroses 83–6 spines 33–4 roughage grazers 79 Terminalia family (Combretaceae) 55–6 ruminants 75, 88–116 xeromorphic leaves 34 Serengeti-Mara ecosystem 60, 251f see also tree and shrub savannah skulls 80f trematode worms 160 South America 5 trophic networks 247 species–area relationships 262 Trypanosoma 68, 160 spiral-horned antelopes 94–9 Trypanosoma brucei gambiense 68 submodel 249 Trypanosoma brucei rhodesiense 68 teeth 79–81 trypanosomes 68–9 teeth and jaw muscles 79–81 trypanosomiasis 67 waterbuck (Kobus ellipsiprymus) 102–3 Tsavo (East and West) National Parks (Kenya) 18, 269 zebras 82–3 tsessebe (Damaliscus lunatus lunatus) 22, 105, 169f, uranium, open-pit mining 258, 273 170, 176f, 178t urea recycling 77 tsetse flies (Glossina spp.) 67, 68, 69, 143, 160, 190 Urochloa species 19 turacos 69 Usangu wetlands 275 Turkwel riverine forest (Kenya) 216 utilisation function 211 two-species competition 211 V U Vachellia 43, 44 Uapaca kirkiana (mahobohobo) 21, 54 Vangueria infausta 21 Ugalla Game Reserve 256f, 257, 257t variable combretum (Combretum collinum) 56 Uganda 24, 262 vegetation 27–64 Ugandan kob (Kobus kob thomasi) 24, 103 grasses 38–41, 39–40t, 41f umbrella thorn (Acacia tortilis)(Vachellia tortilis) lily family (Liliaceae) 42 48–9 morphology and life history 30–8 under nutrition 159, 160 palm family (Arecaceae) 42 UNESCO 258 patterns 14–15, 57–64 unexpected effects 237, 252 rainfall and green biomass 57–60 ungulates 1, 5, 22, 74–82, 204, 208 rainfall, plant biomass and grass-tree mixture artiodactyls 75, 81 27–30 body weights 224f Serengeti-Mara ecosystem 57–64 browsers 78–9 soils, fire and 61–4 buffalo 89–92 submodel 249 bulk grazers 78 trees 41–57 INDEX 329 velvet-leaved combretum (Combretum molle) 56 wet season 3, 11, 13, 60 Venezuela 4, 5 ‘wet’ zone 11 verbal description/model 163 whales 75 vertebrates, production efficiency 243 whistling thorn (Acacia drepanolobium; vervet monkey (Chlorocebus pygerythrus) 23, 44, Vachellia) 45, 64, 139, 231f, 232 134, 269t white-backed griffin vulture 73 Vigna friesiorum 36 white bauhinia (Bauhinia petersiana) 51 Virunga National Park (Democratic Republic of the white-bearded wildebeest (Connochaetes taurinus Congo) 275 mearnsi) 106 vitelline masked weaver (Plocus velatus uluensis) 74 white-ear kob (Kobus kob leucotis) 103, 145 Vitex species 24 white-headed vulture (Trigonoceps occipitalis) Viverridae family (genets, civets and mongooses) 72 121, 129–30, 130t white rhinoceros (square-lipped)(Cerathotherium volcanic grassland 16 simum) 20, 79, 85–6, 86f, 217 vulnerable species 82, 88, 98, 99, 101, 106, 108, 116, white-tailed deer (Odocoileus virginianus) 5 124, 126 white-tailed mongoose (Ichneumia albicauda) vultures 21, 65, 71–3, 209, 234–5 130, 130t white thorn (Acacia seyal)(Vachellia seyal) 49–50 W wild cat (Felis sylvestris) 22, 197f wait-a-bit-thorn (hook thorn) (A. mellifera) wild dog (Lycaon pictus) 113f, 123, 131, 179, 181, (S. mellifera) 47 197f, 198, 203, 207 wallaby 6 grass and shrub savannah 17 warthog (Phacocoerus africanus) 84f, 152, 173, 174f, population estimates 179t 176f, 221t predation data for 199t biomass–rainfall relationship 169f skull 122f Masai Mara ecosystem 153f tree and shrub savannah 19, 20 predation rates 171t woodland savannah 23 skull 80f wildebeest (Connochaetes taurinus) 8, 78, 106, 229f, survival rates 178t 248f, 250–1, 270 teeth 79 adult mortality 162 wasp (Dinarmus magnus) 234 biomass–rainfall relationship 169f waterbuck (Kobus ellipsiprymus) 87, 102–3, 152, 170, and birds 252 221t, 235 body weights 224f biomass–rainfall relationship 169f counting of 148 population numbers in the Kruger National Park cyclic grazing 186 in relation to rainfall 176f decline in population of 172 population trend in the Masai Mara dung 77 ecosystem 153f effect on other ungulates 219 predation rates 171t feeding facilitation and grazing succession 228, survival rates 178t 229f, 230, 230f water buffalo 6 habitat preferred 222 water shortages 272 lion predation on 172 water storage facilities (trees) 34 in Masai Mara 221t weasels 133 migration of 17, 57, 60, 218–21 Mustela nivalis 133 numbers estimation, changes in 154 weavers 65, 73–4 population estimates of 219f weight ratios 253, 255, 257, 257t population model 163–6 west African giraffe (Giraffa camelopardalis population numbers 161–3 peralta) 19, 89 population numbers in the Kruger National Park western hartebeest (Alcelaphus buselaphus major) in relation to rainfall 176f 19, 104, 105 predation rates 171t western rhigozum (Rhigozum brevispinosum) 57 predator deaths 162 western woody euphorbia (Euphorbia predator–prey interactions 197 guerichiana) 54 rinderpest and 170, 179, 190, 209, 246, 247 330 INDEX wildebeest (Connochaetes taurinus) (continued) yellow mongoose (Cyntis penicillata) 130t, 198 risk of attack 203 yellow tree bauhinia (Bauhinia tomentosa) 51 risk of death 202, 204f sarcoptic mange 210 Z survival rates 178t Zambezian forest-savannah mosaic 23 tree and shrub savannah 20 Zambezian region 19 woodland savannah 22 Zambezi Resources Limited (ZRL) 274 wild green-hair tree (Parkinsonia africana) 51 Zambezi river 274 wildlife–human conflict 269–76 Zambia 88, 262, 267, 274, 278 wildlife tourism 277–8 ‘cathedral mopane’ 19 wild water buffalo (Bubalus arnee) 7 copper mining 273 wolves 208 tourism figures for 273t wooded grassland 24 zebras (Equidae family) 81, 82–3, 219, 221, 270, 271 wood-hoopoes 69 biomass–rainfall relationship 169f woodland savannah 14, 21–3 body weights 224f see also miombo woodland savannah common/plains zebra (Equus quagga) 82 woody plants, seed dispersal 264 dung 77, 271f woody species 16, 151, 193, 215 feeding facilitation and grazing succession 228, World Heritage Mana Pools 273 229, 229f, 230f world-wide distribution of savannahs 2–9 feet 75 Africa 5 hindgut fermentation 75, 77f Australia 5–7 identifying individuals 139 comparision of major savannahs 7–8 lion predation on 172 India 7 numbers counted during the Mara transects 221t South America 3–5 population estimates 219f World Wide Fund for Nature (WWF) 275 population model 166–7 population numbers in the Kruger National Park X in relation to rainfall 176f Xenarthra 74 predator–prey interactions 172 Ximenia americana 20 Serengeti-Mara 219f Ximenia caffra 20 skull 80f survival rates 178t Y teeth 79 yearling/adult ratio 164 Zimbabwe 20, 25, 207, 262, 273t yellow armadillo (Euphractus sexcinctus) 3 Ziziphus mucronata 19 yellow baboon (Papio cyanocephalus) 134 zorilla (Ictonyx striatus) 133