The Flexible Phenotype ‘This Text Is a Must for Anybody Who Has Remained Curious About the Ways Animals, Including Humans, Deal with Their Environment

The Flexible Phenotype ‘This text is a must for anybody who has remained curious about the ways animals, including humans, deal with their environment. It’s scientifi cally sound and at the same time it’s a gripping story.’ Dr Hans Hoppeler, Institute of Anatomy, University of Bern, Switzerland and Editor-in-Chief of Journal of Experimental Biology

‘Written with zest, a sense of fun, and a deep love of nature, The Flexible Phenotype offers biologists a real synthesis of ecology, physiology, and behaviour, based on in-depth empirical research. The adventures of the subject of many of these studies, the red knot, a migrant shorebird, captivate the imagination, show us how physiology and morphology express ecology, and make this book not only an important and truly integrated study in biology, but also a pleasure to read.’ Professor Eva Jablonka , Cohn Institute, Tel-Aviv University, Israel

‘Even amongst mammals and birds, animals are diverse. Biologists have long sought the evolutionary pressures that have led to particular body designs and lifestyles. I believe that to do so requires an approach that considers the interaction of body design and behaviour in an ecological context. Too often these components are considered in relative isolation. In contrast, this book is wonderfully broad and holistic, integrating across levels. Even though the book is written in an accessible style that will entertain anyone who is interested in nature, it is serious science.’ Professor John McNamara , University of Bristol, UK

‘Change rules every individual’s world, whether imposed by temporal variation in the environment or by movement through variable surroundings, and survival and reproductive success depend on adaptive responses to this change. Drawing on experience in fi eld and laboratory research, and integrating modern ideas about acclimatization and the optimizing of individual behavior, physiology, and morphology, the authors have produced an accessible, highly readable, and stimulating synthesis of the fl exible phenotype. Using examples drawn from their own work on migrating shorebirds, as well as myriad other organisms, the authors show how individuals respond to change by altering their structure and function through a variety of behavioral and physiological mechanisms. Long-standing traditions of research in physiological, behavioral, and evolutionary ecology are brought completely up-to-date in this timely treatment of organisms in their changeable worlds. The ways in which mechanisms of individual response promote survival and productivity are keys to understanding how organisms will adjust, or fail to do so, in the face of habitat alteration and global climate change.’ Professor Robert E. Ricklefs , University of Missouri - St. Louis, USA

‘In their book Piersma and van Gils provide a timely summary of the reawakening in our knowledge of phenotypic fl exibility in the context of comparative biology. They convincingly remind us of how it is a key component of the whole process by which an organism interacts with its environment. Written in an engaging style which draws the reader into the salient issues of the day with everyday examples this book is not only a landmark in the fi eld, but an entertaining read as well. It will therefore appeal to readers across the spectrum, from interested amateur naturalists, via students of physiological and behavioural ecology to established professional researchers.’ Professor John Speakman , University of Aberdeen, UK The Flexible Phenotype

A body-centred integration of ecology, physiology, and behaviour

Theunis Piersma & Jan A. van Gils

1 Great Clarendon Street, Oxford ox2 6dp 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 in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With ofﬁ ces in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © Theunis Piersma and Jan A. van Gils 2011 The moral rights of the authors have been asserted Database right Oxford University Press (maker) First published 2011 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, 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 book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by CPI Antony Rowe, Chippenham, Wiltshire

ISBN 978–0–19–923372–4 (Hbk.) 978–0–19–959724–6 (Pbk.)

10 9 8 7 6 5 4 3 2 1 d e d i c a t e d t o Rudi Drent (1937–2008)

ethologist-extraordinaire who never lost sight of ecological context This page intentionally left blank Contents

1. Introduction 1 The migrant shorebird story ...... 1 Bodies express ecology ...... 2 What is an organism anyway? ...... 4 Organization of the book ...... 5 Scope and readership ...... 7 Acknowledgements ...... 8

Part I Basics of organismal design

2. Maintaining the balance of heat, water, nutrients, and energy 15 Dutch dreamcows do not exist ...... 15 Hot bodies in the cold ...... 16 Thermometers do not measure feelings ...... 20 Balancing water ...... 23 Elements of a body ...... 26 Birds are not airplanes ...... 28 Shorebird insurance strategies ...... 30 Dying strategically ...... 31 Synopsis ...... 32 3. Symmorphosis: principle and limitations of economic design 33 A well-trained man, a frog, and a hummingbird ...... 33 Economy of design ...... 34 Symmorphosis: the principle and the test ...... 36 Safety factors ...... 40 Multiple design criteria ...... 43 One more problem: the climbing of adaptive peaks ...... 44 In addition to oxygen, ﬁ res need fuel too ...... 46 Testing symmorphosis in shorebird food-processing systems ...... 47 Synopsis ...... 49

Part II Adding environment

4. Metabolic ceilings: the ecology of physiological restraint 55 Captain Robert Falcon Scott, sled dogs, and limits to hard work ...... 55 The need for a yardstick ...... 57 Peaks and plateaus: what is true endurance? ...... 60 viii CONTENTS

Regulation of maximal performance: central, peripheral, or external? ...... 61 Temperature and the allometric scaling constant ...... 65 Protecting long-term fi tness assets: factorial scopes and optimal working capacity revisited ...... 66 The evolution of laziness ...... 68 Evolutionary wisdom of physiological constraints ...... 69 Beyond Rubner’s legacy: why birds can burn their candle at both ends ...... 70 Hard-working shorebirds ...... 72 Synopsis ...... 75 5. Phenotypic plasticity: matching phenotypes to environmental demands 79 Adaptive arm-waving, and more ...... 79 Use it or lose it ...... 81 The dynamic gut ...... 84 ‘Classical’ phenotypic plasticity: developmental reaction norms ...... 85 Seasonal phenotype changes in ptarmigan, deer, and butterfl ies ...... 90 Environmental variability and predictability, and the kinds of phenotypic adjustments that make sense ...... 93 Degrees of fl exibility ...... 95 Direct costs and benefi ts, their trade-offs, and other layers of constraint ...... 96 Phenotypes of fear ...... 98 Plasticity: the tinkerer’s accomplishment? ...... 101 Phenotypic fl exibility in birds ...... 102 Synopsis ...... 106

Part III Adding behaviour

6. Optimal behaviour: currencies and constraints 109 When the going gets tough, the tough get going ...... 109 Loading leatherjackets ...... 110 Better lazy than tired ...... 114 More haste less speed ...... 115 Oystercatchers pressed for time ...... 117 Informational constraints: getting to know your environment ...... 119 Informational constraints: getting to know your patch ...... 121 Do updating animals really exist? ...... 122 The psychology of decision-making ...... 125 Ideal birds sleep together ...... 128 Synopsis ...... 130 7. Optimal foraging: the dynamic choice between diets, feeding patches, and gut sizes 131 Eating more by ignoring food ...... 131 A hard nut to crack ...... 133 It takes guts to eat shellﬁ sh ...... 135 Optimal gizzards ...... 140 Synopsis ...... 144 CONTENTS ix

Part IV Towards a fully integrated view

8. Beyond the physical balance: disease and predation 147 Running with the Red Queen ...... 147 The responsive nature of ‘constitutive’ innate immunity ...... 149 Body-building to defy death ...... 151 Coping with danger ...... 155 Predicting carrying capacity in the light of fear ...... 158 Synopsis ...... 159 9. Population consequences: conservation and management of fl exible phenotypes 161 The Holy Grail of population biology ...... 161 Dredging out bivalve-rich intertidal fl ats: a case study on red knots ...... 161 Population consequences for the molluscivores ...... 162 Which individual red knots made it through? ...... 165 Migrant fl exibility and speed of migration ...... 166 Global change and phenotypic change: plasticity prevails ...... 167 How fl exible phenotypes cope with advancing springs ...... 169 When organisms can cope no more: limits to phenotypic change ...... 172 Synopsis ...... 173 10. Evolution in fi ve dimensions: phenotypes first! 174 Flexible phenotypes and the study of adaptation ...... 174 Separating the environment from the organism ...... 175 . . . and putting them back together: phenotypes fi rst! ...... 176 Genotypes accommodating environmental information? ...... 176 Enter the tarbutniks, and niche construction ...... 179 Evolution in four or fi ve dimensions? ...... 180 Context, please! An orchestra in need of a theatre ...... 182 Synopsis ...... 184

References 185 Name Index 219 Subject Index 222 This page intentionally left blank CHAPTER 1 Introduction

Since an organism is inseparable from its environment, any person who attempts to understand an organism’s distribution must keep constantly in mind that the item being studied is neither a stuffed skin, a pickled specimen, nor a dot on a map. It is not even the live organism held in the hand, caged in the laboratory, or seen in the fi eld. It is a complex interaction between a self-sustaining physico-chemical system and the environment. G.A. Bartholomew (1958 : p. 83)

George Bartholomew was not the only one to waterbirds, and especially the shorebirds that struggle with phenotypes. We all do—whether bred in the meadows around our villages to on a personal level, because our bodies and then disappear for the winter, or migrated performances change with age for better or for through in great numbers in spring and in worse, or professionally, because we are farm- autumn. In the late 1980s, as budding profes- ers, veterinarians, nurses, medical doctors, ath- sional scientists, we quite naturally chose a letes, physiotherapists, ecologists, evolutionary coastal shorebird known to use contrasting biologists or students of these matters. Our migration routes, the red knot Calidris canutus , bodies, and those of the organisms that com- as the focal species in which to evaluate mand our attention, never stay the same, and, the energetic consequences of this fascinating as we will discover in this book, neither are intercontinental migration behaviour. In The they quite what they seem to be at fi rst sight. Netherlands, red knots come in two types. One population, breeding on the tundra of north- The migrant shorebird story central Siberia, passes by in autumn to spend the winter in tropical West Africa and makes a Although encompassing the whole living world, brief touchdown during the return migration in this book was inspired by, and fi nds a focus in, late spring. Another population, from northern the spectacle of seasonal migration, the way that Greenland and the far north-eastern Arctic of birds and other animals travel the world. Canada, comes to us in early autumn, only to Migration would appear a natural fascination leave in spring, thus having to endure the wet for anyone with a burning desire to explore and and windy, and sometimes frosty, conditions discover. How do animals fi nd their way? How typical of winters in north-western Europe. do birds cope with the radically different condi- Field studies in both West Africa and the tions that they encounter on their northern Dutch Wadden Sea are revealing. We can look breeding grounds compared with their tropical at the birds’ behaviour, assess their diet, which wintering areas? Why do they carry out their consists mostly of molluscs, study their move- extensive seasonal movements in the fi rst place? ments and foraging activity over the intertidal Are they as free as they appear to be? fl ats in relation to tide and time of day, and Born and raised as we were in the water-rich look at changes in body mass—and the size of Netherlands, our youthful fascination with the body stores—in the course of the winter. biology of migration would soon fi nd a focus in However, for a full evaluation of the distinct

1 2 THE FLEXIBLE PHENOTYPE

energetic balances of the tropical- and the tem- and distribution from various directions, perate-winterers, we had to bring birds into including, simultaneously, the ecological, the captivity, so that we could measure digestive behavioural, and the physiological angles. To effi ciencies as a function of prey type, and met- be sure, a jack of all trades runs the risk of abolic rates as a function of thermal conditions. being master of none, which may explain why, Rather than feeding them the small bivalves at a time when the term ‘integrative’ has that red knots typically eat and ingest whole in become part of the name of highly cited jour- the fi eld, we fed them the only thing we could nals and university departments, truly inte- afford to give them: protein-rich fi sh food-pel- grative studies of organisms, with due lets, prey that did not need crushing in their attention to physiology, behaviour, and ecol- muscular gizzards, as was their wont. When, ogy, and their inter-relationships, are still thin after some time, we compared the bodies of on the ground. This understanding, together free-living and long-term captive red knots, with our conviction that ecological studies of we were in for the shock that eventually trig- organisms that ignore the key elements of gered this book: wild and captive birds could their physiology, physiological studies that not have been more different! At similar body disregard the organisms’ ecology and behav- masses, the captive red knots were fatter, with ioural studies carried out in ignorance of the much smaller digestive organs than their wild environmental context, might all be a waste of counterparts. Rather than the expected 8–10-g time, were all factors that inspired the writing gizzard, the captives eating soft food pellets of this book. had a 1–3-g gizzard. All of a sudden, the long- term captives no longer seemed a particularly Bodies express ecology good model for birds in the fi eld. But of course, and this insight came obligingly By developing empirical arguments to describe fast, the drastic changes in body composition in the intimate connections between (animal) general, and in gizzard size in particular, of birds bodies and their environments, we have tried brought from one (outdoor) environment into a to formulate a true synthesis of physiology, very different (indoor) environment, provided behaviour, and ecology. We shall review the us with the key to the natural ways in which red principles guiding current research in eco- knots cope with the environmental variations physiology, behaviour, and ecology, and illus- encountered in the course of their seasonal trate these using as wide a range of examples migrations. Our little revelation, that organis- as possible. The results of our long-term mal structure tends to change in concert with research programme on migrant shorebirds— the ambient environment in fully adaptive ways, and the foods they eat—will provide the uni- was not new. George A. Bartholomew, for one, fying narrative. Chapters in the fi rst half of the had the insight almost half a century earlier. book all end with sections showing relevant However, it did make it clear that, in order to examples from the shorebird world. As shore- understand migrant birds, we needed a much birds travel from environment to environment, tighter integration of ecological, physiological, the changing natures of their bodies refl ect the and behavioural approaches than was usual at changing selection pressures in these different the time. And there seems no reason why this environments. The book and its examples will should not be true for any other organism. be incremental. Each successive chapter will Since then, we have tried to elaborate on build on the previous ones and add another this insight, studying problems of behaviour level of explanation to the fascinating INTRODUCTION 3

complexity of integrative organismal biology, reproductively mature organisms throughout illustrating how changes in an individual’s their life, phenotypic fl exibility —reversible with- shape, size, and capacity are a direct function in-individual variation—has remained little of the ecological demands placed upon it. In explored and exploited in biology (Piersma and essence, the book is about the ways in which Drent 2003 ). This is surprising, because, as we bodies express ecology . shall discover, intra-individual variation most Bodies can ‘express ecology’ by being suffi - readily provides insights into the links between ciently plastic, by taking on different structure, phenotypic design, ecological demand func- form or composition in different environments. tions (performance) and fi tness ( Feder and Watt Part of the phenotypic variation between organ- 1992 ). isms, especially differences between isolated The subtitle of our book contains the words populations and unrelated individuals, may be ‘ecology’, ‘physiology’, and ‘behaviour’, but fi xed and refl ect differences in genetic make-up not the word ‘evolution’. This is not because ( Mayr 1963 ). Some of the variation develops in we have neglected evolution in any way. interaction with the particularities of the envi- Indeed, we fully subscribe to Theodosius ronment in which an organism fi nds itself. This Dobzhansky’s classic dictum that ‘nothing in part, indicated by the term phenotypic plasticity , biology makes sense except in the light of evo- can be further subcategorized on the basis of lution’ ( Dobzhansky 1973 ). Evolution is not in whether phenotypic changes are reversible and our title because it is somewhat on the periph- occur within a single individual, and whether ery of what we want to reveal concerning the the changes occur, or do not occur, in seasonally intimate connections between environment predictable, cyclical ways ( Table 1 ). The non- and phenotype. Rather than discussing evolu- reversible phenotypic variation between geneti- tion explicitly, such as the exciting but poorly cally similar organisms that originates during understood relationships between phenotypic development, developmental plasticity , has attrac- plasticity and evolutionary change or innova- ted much empirical and theoretical attention, tion (as examined in the impressive treatise by including the publication of several monographs West-Eberhard 2003 ), we shall assign evolu- ( Rollo 1995 , Schlichting and Pigliucci 1998 , tionary thinking to a back-seat, as we discuss, Pigliucci 2001a ). In contrast, the subcategory of on the basis of reversible phenotypic traits, the phenotypic plasticity that is expressed by single many ecological aspects of organismal design.

Table 1. Mutually exclusive defi nitions of the four historically most commonly used categories of phenotypic plasticity (after Piersma and Drent 2003 ; note that previous workers used less restrictive defi nitions). Phenotypic plasticity itself indicates the general capacity for change or transformation within genotypes in response to different environmental conditions.

Plasticity category Phenotypic change is Variability occurs within Phenotypic change is reversible a single individual seasonally cyclic

Developmental plasticity No No No Polyphenism 1 No No Yes Phenotypic fl exibility Yes Yes No Life-cycle staging2 Yes Yes Yes

1 Can be regarded as a subcategory of developmental plasticity. 2 Life-cycle staging is a subcategory of phenotypic fl exibility. 4 THE FLEXIBLE PHENOTYPE

Even current authors who do explicitly discuss the role of phenotypic plasticity in micro- evolutionary adaptation to new environments DEMOTYPE (e.g. Price et al . 2003 , Ghalambor et al . 2007 ), ecological while they may incorporate developmental plas- interactions sexual selection ticity into their arguments, tend not to mention HOTYP the existence of the reversible plasticity categories ET E ( Table 1 ). Does this indicate ignorance or over- HENOTYPE P le adult t exib rait sight, or is it the case that understanding the role (fl s) E-CY of phenotypic fl exibility in evolutionary change IF CL L E still lies beyond grasp? In the fi nal chapter of

( c y ) c s l S this book we shall provide our views on this i S a it c a T A E tr lly G ult subject and conclude that ‘nothing in evolution v a r yi n g a d P H E E N O T Y P ( makes sense except in the light of ecology’. fixed adulttraits) TIONAL EFFECTS (NATURAL SELECTION) TIONAL EFFECTS (NATURAL

envionment + time of yearr development

What is an organism anyway? PHENOTYPE

ACGT At this point we should make clear how we INTERGENERA look at an organism, both with respect to its development ‘visible’ incarnations in the course of develop- GENOTYPE ment from egg to maturity, and also the ‘invisible’ continuity between generations ( Piersma 2002 ). The whole ‘life-history’ cycle is written Figure 1. An organism (in this case a red knot, see Box 1 ) in the genes, is expressed in the phenotype and can be regarded as consisting of various intra-generational stages (indicated here by the terms genotype, phenotype, and is evolutionarily evaluated in the demotype others), in the course of which the infl uence of environment (Fig. 1 ). The hierarchy of ‘life-history struc- and time of the year increases. The ‘life-history strategy’ that is tures’ ( Ricklefs 1991 ) and their transformations most successful, under the environmental conditions that an are inseparable from each other and exist only animal encounters, yields the highest numbers of offspring in with reference to their environmental context. the next generation (i.e. the most common demotype, the variant with the highest fi tness). The sketches on the right In brief, the phenotype is composed of fi xed illustrate (from bottom to top) the phenomena referred to by traits (e.g. sex, beak size, feather colour, skeletal the genotype (chromosomes, part of a DNA sequence and the dimensions), as well as many fl exible morpho- double helix), the early phenotype (egg and freshly-hatched logical and behavioural traits. Single genotypes chick), aspects of the adult phenotype that are fi xed and those may show variation in fi xed adult traits (devel- that are fl exible (migration from Canada to Europe versus opmental plasticity), variation that comes Siberia to Africa; short- versus long-billed birds, and large- versus small-gutted birds, respectively), including aspects of about during development as a consequence of the phenotype that vary predominantly on a seasonal basis variation in the environment by the action of (life-cycle stages such as seasonal changes in plumage). Traits reaction norms (Schlichting and Pigliucci 1998 ). that can only be described in relation to environmental In addition, there are traits that remain fl exible conditions are indicated by the term ethotype (for example, a even at maturity (phenotypic fl exibility), and a trait like daily energy expenditure). The resulting reproductive success of such genetically instructed types, the inter- subcategory of fl exible traits is formed by those generational effect, is called the demotype (in this case that show predictable seasonally cyclic changes. illustrated by different numbers of short- and long-billed birds, This latter category of phenotypic variation or of large- and small-gutted birds). was named life-cycle stages by Jacobs and Based on Piersma (2002 ). INTRODUCTION 5

Wingfi eld (2000 ). Life-cycle staging specifi cally ecology, physiology, and behaviour. Let us call refers to the occurrence of seasonally-struc- them ‘rules of organismal design’ (including tured sequences of ‘unique’ phenotypes with constraints on design). In Chapter 2 we start respect to state (e.g. reproductive or not; moult- off from two relevant physical laws: (1) organ- ing or not) and appearance (e.g. nuptial plum- isms cannot escape the fi rst law of thermody- age or not). Similarly, a subcategory of namics (the law of conservation of energy) and developmental plasticity called polyphenism (2) they are bound by the second law of ther- refers to season-specifi c occurrences of particu- modynamics (the law of spontaneous ‘spread- lar phenotypes. The best-known examples are ing-out’ of energy and matter). In order to those butterfl ies in which the phenotypes of reduce the amount of entropy in their bodies, successive generations, during any given year, organisms have actively to maintain a balance will depend on the season of hatching (Shapiro in terms of their gains and losses of water, 1976 , Brakefi eld and Reitsma 1991 , Nijhout nutrients and energy, and solutes (micronutri- 1999 ). Both life-cycle staging and polyphenism ents). Depending on what needs to be kept in may be under the infl uence of endogenous balance, and the size and nature of the body programmes, especially such as that of the cir- involved, accounting takes place over varying cannual clock system (Gwinner 1986 ). time-scales. In Chapter 3 we introduce the Phenotype , as classically understood, is background of the cost–benefi t analyses of something that one can measure in an organ- functional capacities that recur throughout the ism independently of its environment, whereas book. Organs are always part of chains of the ethotype has no meaning except when set in processing units within bodies. For this reason an environmental context ( Ricklefs 1991 ). The there is no point in spending the energy and ethotype (derived from the now outdated label nutrients that would be required to build for the study of animal behaviour: ethology) organs with excess capacity relative to that of encompasses the behavioural dimension of the other organs in the chain. This is the prin- phenotypes, and includes factors such as the ciple of symmorphosis that predicts close energy requirements of an individual as a quantitative coupling of different performance measure of its performance in its environment. measures within single organisms. As com- In principle, the fi tness values of all these mon sense as this may seem, there are never- pheno- and ethotypic variants could be quan- theless several issues that need to be discussed. tifi ed (Nager et al . 2000 ), and in Ricklefs’ (1991) These have to do with multiple functions, terminology one would call this entity the reserve capacities of specifi c organs under demotype . The demotype, or fi tness of an organ- extreme and dangerous conditions, and the ism, is a function of both ecological interac- problematic climbing of ‘adaptive peaks’ tions and sexual selection processes. Fitness ( Dawkins 1996 ). Some of the empirical argu- determines which of the competing ‘units of ments build around the technique of ‘allomet- sequenced structures and transformations’ ric scaling’, a concept that is briefl y introduced (i.e. organisms) will survive in nature’s never- in that chapter. ending struggle ( Fig. 1 ). In Chapter 4 we make our fi rst excursion from physiology to ecology, moving away Organization of the book from physical boundary factors and design principles to the real world. In this chapter This book is built around a logically arranged we examine the extremes of adaptiveness constellation of fi rst principles in energetics, as we discuss peak-performance rates, also 6 THE FLEXIBLE PHENOTYPE

known as ‘metabolic ceilings’. The idea that The cost–benefi t approach introduced in the working capacity of organisms would Chapter 3 is incomplete because it is limited to have some sort of maximum level dictated by physical responses. After all, the behavioural size and other design constraints (known as dimension of phenotypes (the ethotypes) can maximum sustained metabolic rates), has respond even more rapidly to changing envi- inspired a large body of ecophysiological ronments than bodies can. The behaviour of research, including the search for allometric real animals at particular times and places is constants. We shall review the fi eld and illus- not only infl uenced by locally achievable ener- trate the (conditional) presence and workings getic costs and benefi ts, but also by the tnessfi of such ceilings by examining shorebirds currency that the animal has been naturally working hard, both under cold conditions selected to optimize. In Chapter 6 we discuss (during breeding in the High Arctic and under how currencies and design constraints infl u- harsh climatic conditions in the non-breeding ence optimal behavioural strategies, especially season) and also during trans-oceanic non- with respect to time and place allocations. In stop migration fl ights. We conclude that met- the process we review the use of the word abolic ceilings are functional solutions to (1) ‘optimal’ in evolutionary ecology. Having the environmental conditions in which an introduced this additional layer of complexity, organism lives and (2) the precise ways in in Chapter 7 we focus on a particular class of which hard work precipitates death, a con- behaviour: the choice of diets and feeding clusion that reverberates through much of patches. Foraging must guarantee a steady what we will repeatedly encounter in the rest income at the lowest possible price. We intro- of the book. duce the elegant theories of optimal diet choice Expanding on our introductory Table 1 , we and the use of feeding patches. Until now such illustrate in Chapter 5 how phenotypes, at theories have assumed rather constant envi- several different levels and over several dif- ronments and bodies. Here, based on our own ferent time- and developmental scales, suc- work on shorebirds, we extend the story to ceed in matching their environments. This include the dynamics of organ size, energy chapter is not simply a lengthy review of odd expenditure, and prey quality. These observa- organismal phenomena. By describing the tions accord with new versions of foraging great variation of ways in which organisms models that allow for fl exible adjustments of adjust to variable environmental demands, food-processing capacities. we emphasize the immense ‘creativity’ of the From here onwards, building on the basics evolutionary process in response to ecologi- laid out in the previous chapters, the book cal challenges. After a discussion on the pos- becomes increasingly ‘integrative’. There is sible costs of a capacity to change, we raise more to nature than nutrients and energy questions about the extent to which pheno- equivalents: only organisms that successfully typic adjustments are fully genetically escape predation and resist disease long instructed or self-organized on the basis of enough will leave offspring. In Chapter 8 we organism–environment feedback loops. The bring these factors into context by introducing reversibility of some such changes is illus- the ways in which animals cope with the trated by the way in which the organs of birds threats of disease and predation. The immune are continuously adjusted to changes in system is not only amazingly complex, but it is energy expenditure, food quality, and other shown to be highly fl exible as well. In addition environmental factors. to simple behavioural measures to avoid being INTRODUCTION 7 killed by predators, birds also appear to have ments ( West-Eberhard 2003 )? The writing of physiological adjustments in store. We intro- this fi nal chapter on evolutionary implications duce ways in which the fi elds of predation, was diffi cult. It most clearly represents this disease, and foraging can be merged under the book’s attempt to ‘grope publicly for solutions single equation of ‘common currencies’. to recalcitrant conceptual issues of the day’ In Chapter 9 we come close to fi nding the ( Gottlieb 1992 ). Holy Grail of population biology by reaching an understanding of how environmental con- Scope and readership ditions and organismal strategies translate into individual fi tness (the demotype), and We trust that this book will encourage a fur- hence into changes in the size (and the distri- ther integration of ecology, physiology, and bution) of populations. We illustrate how the behaviour, and will in turn foster collaborative relationships between variable features of research agendas between various kinds of the environment (i.e. prey quality) and of the organismal scientists, beyond early proposals organism living in that environment (i.e. giz- such as those of Bill Karasov (1986 ). Although zard size or processing capacity) combine to migrant shorebirds are the principal characters affect subsequent survival and population in some parts of our story, the arguments sizes of red knots. We also examine how organ- developed here are of much broader relevance. isms cope with various kinds of global change. We hope that this book will be read by a wide Do the adjustments shown refl ect changes at audience of professionals (obviously includ- the level of the genome, or do they refl ect phe- ing advanced students) who work with organ- notypic plasticity? isms, whether they have medical, veterinarian In Chapter 10 we try to bring everything or biological backgrounds. It should attract together under an evolutionary spotlight. In readership among active researchers and his masterful book-length essay on the basics scholars from a wide range of taxonomic spe- of evolutionary organismal biology, Denis cializations. Given our own background, our Noble ( 2006 ) stopped short of using the term approaches will be most easily recognized and ‘ecology’ with a discussion of organs and sys- appreciated by those working at the interface tems, which he described as the ‘orchestra’. of physiology, behavioural ecology, and evolu- Here we provide a venue in which that ‘orches- tionary biology, but we also hope to heighten tra’ can be located, a proper ‘theatre’ composed the interest of both hard-core evolutionary and of the ever-stringent ecological context. Indeed, molecular biologists and practising pheno- many of the riddles in understanding the dis- type-specialists, including those in the medi- tribution of red knots, other shorebirds, and cal and veterinary professions. many other organisms, could not have been Although aimed at a diverse audience, this explained without insights into their energet- is still a ‘scientifi c’ book that offers in places ics (and those of their prey), organ capacities, complicated, but fundamentally instructive, and phenotypic fl exibility. What we provide drawings and diagrams, uses technical terms here is a discussion of the degree in which fl ex- (that we have tried to explain as we encoun- ibility of phenotype is relevant to the evolu- ter them in the course of the story) and refer- tionary process. For example, do genes take ences to the primary literature. At the same the lead during evolutionary adaptive change, time, we have tried hard to engage our readers or is there merit in the suggestion that genetic by developing narratives of excitement and changes actually follow phenotypic adjust- discovery. 8 THE FLEXIBLE PHENOTYPE

In the end, this book should appeal to all unlikely still to exist in the event of an earth- those who expound a truly integrative ecosystem collapse), we fi nd it very strange to approach to understanding the basis of our believe that humans could sustain themselves existence on Earth, i.e. of biological systems in in space in the fi rst place. As this book makes a global, ecological setting. We hope this will abundantly clear, we are connected to earth as include theoretical physicist Stephen Hawking with an umbilical cord; our functioning organ- ( Fig. 2 ), the author of the best-selling A Brief ism simply needs earth. Whether we think History of Time (1988). At a pre-fl ight news con- about the bacteria in our guts, the peace of ference before his fi rst experience of weight- mind that comes from strolling through earthly lessness, Hawking said that he wanted to habitats, the way that our developing bodies encourage public interest in space exploration. need normal gravity and even pathogens ( Zuk With an ever-increasing risk of being wiped 2007 ), or the multifold ways in which bodies out on Earth, he argued that humans would express ecology, human beings, alongside need to colonize space. Apart from the need for most other organisms, are a physical extension large human populations and strong societies of it. Bodies are earth. to succeed in space travel (which is perhaps Acknowledgements Our decades-long work with shorebirds was the impetus for what we have presented here, especially because we saw it as part of a much broader and as yet unfi nished canvas. Over time our impatience for a unifying thesis grew and we realized that the quickest way to this end was to attempt to do it ourselves, to write a book. In an interview with Christine Beck ( 2004 ), Bart Kempenaers, then the newly appointed research director of the Max Planck Institute for Ornithology in Seewiesen, Germany, nicely expressed how we felt when he said that ‘books are exciting because they go beyond pure fact, because they give shape to developing ideas, disclose backgrounds, reveal associations; books change perspectives’. However, we were somewhat naïve in Figure 2. After a life immobilized by a debilitating thinking that our ‘task would involve little neuromuscular disorder (but still able to study the bone- more than writing down all we knew before crushing gravity of black holes), Stephen Hawking, a theoretical we forgot it’ (paraphrasing Dolph Schluter physicist, enjoyed a total of four minutes of zero gravity in a 2000 , p. vii, replacing ‘I’ with ‘we’). We can modifi ed Boeing 727 fl ying a series of parabolic arcs between now certainly concur with him that ‘as the 11 and 8 km height in April 2007. Weightlessness is achieved book got underway the limits of our knowl- during the 45-degree nose dives. In Chapter 5 we discuss the effects of long-term zero gravity conditions on body function edge became distressingly apparent, and we and organ sizes. now feel we learned most of the contents along Photo freely available at www.ZeroG.com . the way’. INTRODUCTION 9

Box 1 Brief description of the book’s empirical backbone: a molluscivore migrating bird interacting with its buried prey

Red knots ( Fig. 3 ) are close to what an ‘average bird’ would probably look like with respect to mass, size and dimensions. They are slightly bigger than a blackbird Turdus merula (Old World, southeast Australia, New Zealand), or an American robin Turdus migratorius (New World), but they show much greater seasonal variation in mass and feather colour than either of these thrushes. This has to do with the very long migrations of red knots that breed on barren tundra as far north as one can get. Outside the breeding season, one fi nds red knots throughout the world except Figure 3. A red knot on Banc d’Arguin, Mauritania, a Antarctica, but only in marine coastal habitats, wintering area in West Africa, that is ingesting a hard- shelled food item: a bivalve. The back-and-white margins usually large wetlands with extensive intertidal of the wing covering feathers give it away as a juvenile foreshores (Piersma 2007 ). The discontinuous bird, only a few months old when photographed shortly circumpolar breeding range of red knots after capture in December 2004. The bird carries rings that incorporates the breeding areas of at least six combine the colours Red, White, and Blue, and is known populations ( Fig. 4 ), populations that are to us as R6WRBB. Molecular sexing on the basis of DNA suffi ciently distinct morphologically to count in a small blood sample ( Baker et al. 1999 ) showed it to as subspecies. These subspecies are certainly be a male. R6WRBB has been sighted many times in distinct when it comes to their migratory Mauritania for a further two years after capture. trajectories and the seasonal timing of their Photo by Jan van de Kam. movements, and show little or no overlap at the fi nal wintering destinations, with a limited overlap during south- and northward migration. Arguably the most fascinating molluscs, sometimes supplemented with aspect of the whole substructuring of the easily accessed softer prey, such as shrimp- world’s red knots is the suggestion, based on and crab-like organisms and even marine genetics, that all of the world’s six or seven worms ( Piersma 2007). This is not because subspecies have diversifi ed recently from a red knots necessarily prefer hard-shelled single founder population that survived the molluscs (in fact they do not, when given the last glacial maximum of ca. 20 000 years ago choice), but because they are specialized in ( Buehler et al. 2006 , Buehler and Piersma fi nding and processing such prey, to the 2008 ). detriment of being able to fi nd the actively Red knots are specialized ‘molluscivores’ crawling soft-bodied worms and small (i.e. predators eating only shellfi sh), but crustaceans on which other sandpipers briefl y shift their diet to surface arthropods specialize. In fact, one of their sensory during the summer breeding season on capacities, the ability to use self-induced High-Arctic tundra. Wherever people have pressure gradients around hard objects in studied their diet, red knots eat hard-shelled soft, wet sediments, has not been described

continues 10 THE FLEXIBLE PHENOTYPE

Box 1 Continued

C. c. piersmai C. c. rogersi C. c. roselaari C. c. rufa C. c. islandica C. c. canutus

islandica tus rogersi nu roselaari ca

i a

m s r

e i rufa

Figure 4. Global distribution of the six recognized subspecies of red knots during the breeding and the non-breeding season. Polygons in the arctic depict breeding ranges; circles depict principal wintering areas and the diameter of the circles indicate the relative number of birds using each area. Arrows depict north- and southward migration routes and use of en route sites. Projected above the contemporary distribution of red knots is a phylogram summarizing the genetic population structure as presently understood ( Buehler and Baker 2005 ). This fi gure is modifi ed after Buehler and Piersma ( 2008 ), an update of the original synthesis by Piersma and Davidson1992 ( ).

for any other animal. The ‘remote detection’ same way as molluscivore-diving ducks, of buried hard-shelled prey is probably ingest their prey whole, then crush the shell enabled by their bill-tip organ, the dense in the muscular part of the stomach, the conglomerations of pressure sensors (Herbst gizzard (the glandular stomach is corpuscles) clustered in forward-pointing rudimentary; Piersma et al . 1993b , van Gils ‘sensory pits’ in the outside surfaces of the et al . 2003a, 2006a). Crushed shell material tips of both upper and lower mandible is usually voided as faeces and not as pellets, ( Piersma et al . 1998 ). which is the habit of related shorebirds, such Unlike other molluscivore shorebirds, such as dunlin Calidris alpina and redshank Tringa as oystercatchers, that remove the fl esh from totanus . As a consequence of the work the shell with their stout bill, red knots, in the carried out on the shell material by the INTRODUCTION 11

gizzard and intestine, and in order, perhaps, to move away fast ( Vermeij 1987 , 1993 ). Red to prevent the wear and tear infl icted by shell knots do eat gastropod molluscs (snails, fragments on the sensitive intestinal wall, usually protected by a heavy or profusely both gizzard and intestine are relatively heavy ornamented shell), but most of their diet in shorebirds that eat hard-shelled-prey. In an consists of bivalves, and these bivalves allometric comparison among 41 shorebird employ all the different strategies just listed species, the red knot came out as the one ( Stanley 1970, Piersma et al . 1993a ). All with the largest gizzard for its body mass bivalves are protected by hard valves, but ( Battley and Piersma 2005). shallowly buried species, such as edible Although red knots are able to crush and cockles Cerastoderma edule , are much more process reasonably heavily built molluscs, heavily armoured than deep-buried bivalves, they rather prefer, as we shall see, thin- such as the Baltic tellin Macoma balthica and shelled prey. It also helps their food-fi nding if the other relatively thin-shelled tellinids. the shellfi sh live close to the sediment There are no examples of intertidal bivalves surface, or on top of it, and if they are not exhibiting ridges or spines, although the attached to any hard surface, such as a piece heavily ridged venerid Placamen gravescens of rock, a mussel- or oyster-bank, or from north Australia comes close. Mussels of mangrove roots. The state of traits that red the family Mytilidae come in heavily knots, indeed most predators, prefer, armoured forms, or may secure themselves to obviously precisely mirrors the state of traits rocks or conglomerates of conspecifi cs with that molluscs use to avoid being eaten. As so-called ‘byssal threads’. A few bivalves of prey they can develop heavy or extravagant intertidal fl ats, such as the very thin-shelled armature, they can attach themselves to a Siliqua pulchella that lives in the soft muds of hard substrate, they can sit deep in the Australasia, are fast burrowers and swimmers sediment beyond the length of a probing bill, and still have a chance to get away when or, upon discovery by a predator, they can try detected by a probing shorebird.

We dedicate this book to our mentor and ing ambitious, if not daring, programmes of teacher, Rudi Drent. We are sad that he did not fi eld research, whilst still attending to the live to be part of the entire writing process. physiological nuts and bolts that are better Sharing thoughts with him and entertaining studied in laboratory settings. We appreciate, ideas about our science were among the most and cherish, his many fi ghts for the institu- formative and fantastic experiences in our pro- tional embedding of our endeavours, the very fessional lives. Rudi infl uenced this book in force that gave our generation a chance in many more ways than by introducing us early science. on to the concepts of optimal working capaci- Yes, fi nishing this book has been a long jour- ties, also known as ‘metabolic ceilings’ ney. We would like to acknowledge the trust, ( Chapter 4 ). As an ethologist, Rudi certainly patience, encouragement, help and tolerance never lost sight of ecological context. As proof the people that have to live with us in their fessor in animal ecology at the University of daily lives—both our colleagues at work and Groningen he created a great ambiance for our family members. TP is grateful to Petra de budding ecologists. He was forever stimulat- Goeij for creating great ambiances for book 12 THE FLEXIBLE PHENOTYPE writing. JAvG admires the patience of his wife, Hans Hoppeler, Ewald Weibel, Maurine Dietz Ilse Veltman, and their children, Lieke and and Rob Bijlsma provided helpful comments Jort, which they expressed so many times dur- on Chapter 3 . We enjoyed interacting with Tim ing the writing process. We are also grateful to Noakes with respect to Chapter 4 , and appre- the artists who helped make this book what it ciated the constructive comments by John is. Dick Visser prepared all the new artwork Speakman, Bob Gill, Bob Ricklefs, Kristin and, as usual, it was both exciting and a Schubert (who additionally provided material great pleasure to work with him. We thank for an entire paragraph), Bernd Heinrich, Irene Barbara Jonkers for her smart cover design, Tieleman, Joost Tinbergen, John McNamara, Jan van de Kam for providing stunning pho- Maurine Dietz, Klaas Westerterp, and Dan tos, and Ysbrand Galama for the dreamcow of Mulcahy. For comments on Chapter 5 we thank Chapter 2 . TP thanks David Winkler and the Matthias Starck, Paul Brakefi eld, Chris Cornell Laboratory of Ornithology for gener- Neufeld, Rick Relyea, and Massimo Pigliucci. ous hospitality during a minisabbatical in Ola Olsson commented on Chapter 6 , John Ithaca—laying the ground for Chapter 10 . We McNamara on Chapter 7 , Piet van den Hout are very grateful to our contacts at Oxford on Chapter 8 and Christiaan Both helped with University Press, especially editors Ian Chapter 9 . Ritsert Jansen tinkered with the fi rst Sherman and Helen Eaton, for their guidance, and the last chapters, but what we may really encouragement, patience, and general good need is a new book! Eva Jablonka, Massimo faith in the enterprise. Marie-Anne Martin Pigliucci, and Bob Gill helped with the fi nal provided the index, was a great text editor and chapter, as did the 2009/2010 cohort of the a constant source of encouragement during University of Groningen TopProgramme in the fi nal phases. Evolution and Ecology: Adriana Alzate-Vallejo, Two long-time friends from Canada, Hugh Lotte van Boheemen, Rienk Fokkema, Jordi Boyd and Jerry Hogan, read all chapters. Apart van Gestel, Oleksandr Ivanov, Hernan Morales, from giving us confi dence, they kept us honest Froukje Postma, Andrés Quinones, and Michiel in the use of English, the terminology of proxi- Veldhuis. Finally, we thank those who either mate and ultimate causes, and helped us with supplied or helped us fi nd original photos or context. We are very grateful for their endur- artwork: John Speakman, Colleen Handel and ing support. Bob Gill provided a refreshing Bob Gill, Chris Neufeld, Mike Stroud, Duncan review of the introductory chapter. Irene Irschick and Maria Ramos, Ola Olsson, and Tieleman shined her bright light on Chapter 2 . Alexander Badyaev. Part I

Basics of organismal design This page intentionally left blank CHAPTER 2 Maintaining the balance of heat, water, nutrients, and energy

Dutch dreamcows do not exist It was recognized long ago that all living organisms have to face this basic physical law. We grew up in the Dutch countryside. Land of Max Rubner (1854–1932), a German physiolo- cattle. One of us vividly remembers the cartoon gist, measured the heat produced by a dog in the milking shed at one of the local farms: placed in a metabolic chamber for 45 days! He the Dutch dreamcow, with udders at both ends found that this amount (17 349 kilocalories = to produce milk, a backside at each end to pro- 72 588 kJ) precisely matched the energy that duce manure and no mouth to feed ( Fig. 5 ). was released when combusting the dog’s food Such a cow would defy the fi rst law of thermo- (17 406 kilocalories), correcting for the amount dynamics, the law of ‘conservation of energy’: of energy retained in the dog’s faeces. He con- a perpetuum mobile that could not exist. Applied cluded that the only source of heat production to cows, the fi rst law simply states: what comes by his dog was the chemical energy captured out (nutrients and energy in the form of milk in the dog’s food ( Chambers 1952 ). No one and manure), needs fi rst to go in (nutrients and since Rubner has refuted these simple physical energy in the form of food). The exact form of facts, and no one ever will. energy is not important as it may be trans- Beside the fi rst law that can never be broken, formed from one (food) into another (milk, organisms also need to cope with the ever- manure). This necessitates work and the energy present second law of thermodynamics. Any (in the form of heat) lost along the way. system, be it dead or alive, tends to show an increase in disorder, called entropy, over time. Eggs break rather than unbreak, coffee grows cold rather than hot, and people grow older rather than younger. In order to live, organisms need to fi ght against this spontaneous ‘spreading-out’ of energy and matter. To reduce the amount of entropy in their bodies, organisms actively maintain some sort of balance in terms of the gains and losses of heat, water, nutrients, and energy, and solutes (micronutrients). Stated simply: they have to eat (Fig. 6 )! If they do not, their bodies will soon reach a state of much Figure 5. Cartoon of the Dutch dreamcow, a cow that produces twice the amount of milk and manure but does not greater entropy, a state that we call death. have to eat. A living perpetuum mobile! With this inevitable trend toward increased Drawing by Ysbrand Galama. entropy, useful chemical energy eventually

15 16 BASICS OF ORGANISMAL DESIGN

H2O problem by depositing ‘blubber’, a sparsely vascularized layer of fat directly under the skin. The ability to maintain constant body nitrogenous waste temperatures was not seized upon by the land- dwelling reptiles, but once feathers and hairs

H2O had evolved, it became possible to maintain relatively constant core temperatures.

CO2 Great benefi ts were thus to be gained. Places H2O where insuffi cient solar radiation was availa- carbohydrates proteins ble to heat up a cold body could now be inhab- fats ited. In fact, even the coldest corners of the Earth became suitable habitats—think of pen- Figure 6. For all animals, this rhinoceros included, the following is true: eating is nothing more than increasing guins living in Antarctica or polar bears Ursus the entropy of your food, in order to prevent an increase in the maritimus doing the same around the North entropy of yourself. By breaking down food molecules into Pole. At a biochemical level, effi cient enzymes smaller molecules, free energy is released that can then be used with narrow temperature-tolerance ranges to drive energy-requiring reactions within the animal. could now be utilized. Furthermore, as neuro- Based on Eckert et al . ( 1988 ). logical reactions run faster at higher body temperatures, information could now be processed faster as well. Very rapid responses can be vital degrades into a ‘waste’ form of energy: heat. when trying to catch a prey or escape a preda- Heat is considered a useless by-product, as it tor. Finally, warm blood, especially blood that cannot be converted back into other forms of is warmed up during a fever, keeps pathogenic energy. In hot climates, heat can even be a fungi out ( Robert and Casadevall 2009 ). problematic by-product, as it may raise body Obviously, the advantageous warm-blooded temperature to excessively high levels. But in lifestyle also carries with it a large cost: much cooler climates, heat is mostly a useful by- more food is needed to cover the greater heat product, as it helps an animal to control its production. We will come back to that in a body temperature and thus the rate of chemi- moment, but fi rst we want to explain why the cal reactions taking place inside its body. How terms warm- and cold-blooded are inappropri- animals regulate their heat balance will be the ate, even though they have been in use since fi rst issue dealt with in this chapter. Aristotle ( Karasov and Martínez del Rio 2007 ). Under hot circumstances, the blood of a ‘cold- Hot bodies in the cold blooded’ organism (e.g. a lizard on a rock) can in fact be warmer than that of the ‘warm-blooded’ Sometime during our evolutionary history we creature sitting right next to it (e.g. a gerbil). To crawled out of the water as fi shlike vertebrates get rid of this inconsistent terminology, modern ( Dawkins 2005 , Shubin 2008 ). Compared with classifi cations are based on the constancy in air, water has a much higher heat capacity and body temperature, giving us the terms homeo- heat conductivity, making it diffi cult for our therm (homos = same) and poikilotherm ( poikilo aquatically living ancestors to maintain a tem- = varied). Another way of classifying the ther- perature differential between their bodies and mal biology of animals is by recognizing the their surroundings. Modern aquatically living heat source that keeps the body warm: endo- mammals, such as seals, have solved this therms (endo = within) mostly use their own MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 17 metabolism for this, while ectotherms ( ecto = kinds: snow buntings Plectrophenax nivalis , jays outside) use external heat. Most endotherms are Perisoreus canadensis , arctic gulls Larus hyper- homeotherms (e.g. the gerbil) and most ecto- boreus, weasels Mustela rixosa , lemmings therms are poikilotherms (e.g. the lizard), but Dicrostonyx groenlandicus rubricatus , ground there are exceptions (e.g. parasites living inside squirrels Citellus parryii, arctic foxes Alopex a bird or a mammal are typical homeothermic lagopus , Eskimo dogs Canis familiaris , and polar ectotherms). In this chapter we shall mostly deal bear cubs Thalarctos maritimus. With this diver- with homeothermic endotherms. sity of animals locked up in out- and indoor Let us now return to the costs of being a homeo- cages, there were times when the Point Barrow therm. Much of what we now know about tem- laboratory must have looked like a small zoo. perature regulation in homeotherms is due to Scholander and his colleagues predicted the pioneering work of Per Scholander that simple Newtonian cooling laws governed ( Scholander et al. 1950a ). This man, who moved the heat loss of an animal, i.e. that the heat loss to the United States after growing up and being (and thus heat production under a constant trained as a botanist in Norway, had a lifelong core temperature) would be proportional to fascination with the physiological responses to the temperature gradient between the animal extreme conditions of both plants and animals. and its surroundings ( Newton 1701 , Mitchell His adventurous mind, together with the need 1901 ). They were right. Cooling down the met- to take measurements along an ‘extreme condi- abolic chamber, such that the gradient between tion axis’, brought him around the world at a body temperature and environmental temper- time when not so many could do so ( Scholander ature doubled, did indeed double an animal’s 1990 ). For his work on thermoregulation, his metabolic rate. However, this simple and ele- original plan was to build a mobile laboratory gant result was found only when the environ- in a huge military troop glider (an engineless mental temperature was below a certain value. aircraft), which could then be brought to remote Above this threshold, the so-called lower criti- places (e.g. the inaccessible and practically cal temperature, metabolic rate remained unknown Prince Patrick Island in the Canadian remarkably stable; in this region, heat as a Arctic Archipelago was on his mind). His boss ‘waste product’ of resting metabolism could thought this idea was insane, but at the same be used to maintain constant body tempera- time suggested building a more permanent lab- ture. As shown later (e.g. King 1964 ), an upper oratory at an alternative location, Point Barrow critical temperature also exists. Above this on Alaska’s North Slope. And this he did. threshold additional costs come into play in In Alaska, Scholander measured the costs of order to prevent overheating. homeothermy by placing individual animals The lower critical temperature differed enor- in small metabolic chambers. During trials mously between their study animals: lem- lasting eight to twelve hours, an individual’s mings and weasels reached this threshold

O2 -consumption and CO 2 -production were between 10 and 20 °C, while arctic foxes could quantifi ed across a wide range of temperatures go down to –30 °C or more before they elevated (from the lowest Alaskan winter temperatures their metabolic rate. 1 In general, Scholander of –40 °C up to over +30 °C generated by an et al . found that the lower critical temperature electric heater). The animals were of various was lower in larger animals than in smaller

1 Note that a recent study on arctic foxes found much higher values ( Fuglesteg et al . 2006 ). It was argued that Scholander’s foxes were not at rest during the measurements. 18 BASICS OF ORGANISMAL DESIGN

400 Afric 0.07 Raccoon Vidua Arctic Ground Squirrel LemmingW an Dall Sheep easel Arctic Fox Coati 300 Manakin 0.06 Arctic Jay ARCTIC Wolf Caribou 0.05 200 Grizzly Polar Bear Bear (dry) Arctic Gull, Arctic Fox 0.04 Rabbit and larger mammals 100 Beaver (dry)

metabolic rate (basal = 100) metabolic rate Marten 0.03 Three-toed Sloth –70–60 –50 –40 –30–20 –10020 10 30 insulative value Night Monkey Two-toed Sloth ° White-faced ambient temperature ( C) 0.02 TROPICAL Monkey Lemming Marmoset Figure 7. The metabolic response to cold in arctic (solid lines) Weasel Paca Coati 0.01 Rabbit and tropical (dashed lines) birds and mammals as measured Shrew Polar Bear (wet) by Scholander et al . ( 1950a ). Each species’ basal metabolic rate Squirrel Beaver (wet) 0.00 has been standardized to a value of 100 to facilitate comparison 0 10 20 30 40 50 60 70 between species. hair length (mm) From Scholander et al . ( 1950a ). Figure 8. Hair length largely determines the insulative capacity animals (Fig. 7 ). There are two reasons for this. of a mammal’s pelt. Mainly for this reason, arctic mammals (open dots) are better insulated than tropical mammals (closed dots), as One reason, already recognized by Scholander their hairs are usually longer. In addition, for a given hair length, et al. (1950b ), is that larger animals have longer tropical pelts have a slightly lower insulative value than arctic fur or longer feathers, which trap the heat pelts. Note that immersion in water (closed triangles) reduces an more effi ciently ( Fig. 8 ). The other reason is animal’s capacity to maintain body heat enormously. that, with an increase in body mass, body sur- Based on Scholander et al . ( 1950a ). face area, through which most heat is lost, increases more slowly (mass2/3 ) than does metabolic rate (mass 3/4 ; Schmidt-Nielsen 1984 ; see the poles, tells us that this rule does not give Box 2). Thus, at least within a given species, robust predictions inter specifi cally. the larger an individual, the better it would be On the subject of latitudinal gradients, able to live in a cool climate, and the more dif- Scholander was interested to fi nd out whether fi cult it would be to live in a warm climate; tropical animals, for other reasons than body heat dissipation would be more diffi cult for size, had different lower critical temperatures larger bodies. This principle underpins a much from arctic animals. Thus, after spending two older ‘rule’, named after Bergmann ( 1847 ). winters at Point Barrow (at 71 ° N), he and his Bergmann’s rule on the basis of surface-to- team went to Barro Colorado Island, Panama volume ratios states that body size increases (at 9 ° N), in order to take similar measurements, when going from equator to pole. There is but this time on sloths Choloepus hoffmanni , good evidence that this rule does indeed hold monkeys Aotus trivirgatus and Leontocebus geof- intra specifi cally, both among mammals froyi, raccoons Procyon cancrivorus, coatis Nasua ( Ashton et al . 2000 , Freckleton et al . 2003 , Meiri narica , jungle rats Proechimys semispinosus , and and Dayan 2003 ) and birds ( Ashton 2002 , Meiri two tropical bird species Pipra mentalis and and Dayan 2003 ), though it is sometimes ques- Nyctidromus albicollis . The results were aston- tioned whether heat conservation is the driv- ishing: tropical animals began to increase their ing force ( Ashton 2002 , Ho et al . 2010 ). The metabolic rate when the air was cooled down existence of elephants, rhinoceros, and giraffes, just a few degrees below the ambient air tem- all distributed in the tropics but not around perature, i.e. in the range of 20 to 30 °C! As MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 19

raccoons and jungle rats are not smaller than arctic zones. This difference is now widely weasels and lemmings, this difference could accepted ( Wiersma et al. 2007 ). not be explained by body size. Scholander et al. Although tropical nights can sometimes be skinned several of their study animals and quite chilling, there is of course a good reason found that a piece of tropical pelt had a lower for tropically dwelling animals not to have insulative capacity than an equally sized piece warm fur or a high BMR: during the heat of of arctic pelt; this result could be partly explained the day, they run the risk of being overheated. by the fact that tropical fur is generally sparser Counter-intuitively, in order to prevent over- and shorter ( Fig. 8 ). As human beings, we are a heating, metabolic rate increases above the so- good example of how the low insulative value called ‘upper critical temperature’, due to of our ‘pelt’ over-rules the effect of a relatively heat-dissipating mechanisms, such as sweat- large body size. Our lower critical temperature ing or panting. At a certain point (i.e. at an lies between 27 and 29 °C (Du Bois 1936 , ambient temperature near core temperature), Winslow and Herrington 1949 ). Needless to say, these measures are insuffi cient and core tem- this is with our clothes off. Scholander et al . perature will increase ( Fig. 9 )—a situation ulti- (1950a ) concluded: ‘Man is indeed a tropical mately leading to death (i.e. a lost fi ght against animal, carrying his tropical environment with him.’ Note that better insulation in northern arctic/temperate species is not a general rule. 6 For example, Tieleman et al . ( 2002 ) found ‘desert A larks’ (exposed to cold nights) were better insu- h)) 5

lated than equally sized larks living in the tem- / (g × 2 4 O

perate zone. 3 Apart from body size and the insulative 3 value of fur, there may be another reason why arctic homeotherms can better withstand the 2 cold than their tropical equivalents: they have 1 a bigger ‘engine’, a higher basal metabolic metabolic rate (cm rate (BMR). This would have the effect of low- 0

ering an animal’s lower critical temperature. Is ° C) 45 44.5 there evidence for higher BMRs in arctic mam- 44 mals and birds? Scholander et al. ( 1950c ) con- 43 42.4 cluded that there were no general BMR 42 41.3 41.9 42.1 42.0 differences between ‘their’ tropical and arctic 41 41.1 B 40 animals, notwithstanding a few exceptions ( rectal temperature (the ‘lazy-acting’ sloths had relatively low 0152105 205 3035 40 ° metabolic rates, unsurprisingly). This was air temperature ( C) consistent with a study on fi ve tropical bird Figure 9. (A) A typical Scholander curve across its full range, species by Vleck and Vleck ( 1979 ). However, extending beyond the upper critical temperature (UCT, the later studies, both on birds (Weathers 1979 , temperature at which metabolic rate goes up again after a plateau value), applied to data on metabolic rates in white- Hails 1983 , Kersten et al. 1998 , Wiersma crowned sparrows Zonotrichia leucophrys gambelii . Body et al. 2007 ) and mammals ( Lovegrove 2000 ), temperature is given in (B), and note that it increases when the consistently found BMRs that were 20–40% ambient temperature is still below the UCT. lower in the tropics than in the temperate and Modifi ed from King ( 1964 ). 20 BASICS OF ORGANISMAL DESIGN

Nature’s second law, occurring at core temper- So far we have seen very signifi cant differ- atures of 46–47 °C; The Guinness Book of ences between arctic and tropical animals in Records reports a lethal core temperature terms of their thermoregulation. Red knots, of 46.5 °C in humans; McWhirter 1980 ). key players in this book, breed in the High Scholander et al. did not ‘heat up’ their study Arctic and some winter in the tropics. They animals enough to reach upper critical tem- thus need to tolerate both cold and heat stress peratures, but with experimental temperatures within their same small bodies. No wonder approaching +35 °C they must have brought that fl exibility became a way of life for these them very close. In fact, the existence of the impressive globetrotters. And the differences upper critical temperature was established between arctic and tropical environments can later in studies on birds, such as the one by get even greater when we consider more than King ( 1964 ). just air temperature alone. With lower critical temperatures being lower in arctic animals as an adaptation to the cold, Thermometers do not measure we may expect upper critical temperatures to be feelings higher in tropical animals as an adaptation to the heat. This has never been thoroughly sur- ‘It feels cold’ is an expression we often hear. It veyed along the arctic–tropics gradient, but captures the notion that air temperature alone there is some sparse evidence from a compari- does not always describe our sense of heat loss. son between birds living in hot deserts and Strong winds on a sunny, seemingly warm day, related species from the temperate zone. make us reluctant to take off our sweaters and Tieleman et al. ( 2002 ) contrasted two temperate- we would certainly leave them on if a cloud living larks (skylarks Alauda arvensis and wood- moved in front of the sun, taking away the pleas- larks Lullula arborea ) with two desert-dwelling ant warmth of the sunlight. The way to encap- larks (hoopoe larks Alaemon alaudipes and sulate all the factors that affect an animal’s heat Dunn’s lark Eremalauda dunni), and did indeed balance is to talk in terms of ‘environmental fi nd that the desert birds had higher upper criti- temperature’ or ‘operative temperature’. In cal temperatures (38–44 °C vs. 35–40 °C). Dutch, we literally use the term gevoelstempera- However, this effect was not found in a general tuur , ‘feelings temperature’, when we speak analysis containing more desert and non-desert about environmental temperature. species ( Tieleman and Williams 1999 ). Three Thus, in addition to air temperature, envi- mechanisms may underlie this possible adapta- ronmental temperature is affected by convection of tropical/desert homeotherms: not only tion through wind, different forms of radiation do they produce less heat (because of a lower and conductance (Fig. 10 ). Furthermore, ani- BMR) and are better able to get rid of this heat mal properties such as size, shape, and colour (because of thinner fur or sparser plumage— infl uence the thermoregulatory signifi cance of except for the better insulated desert larks just given levels of wind and radiation ( Porter and discussed), but they may also be able to tolerate Gates 1969 ). Therefore, under specifi ed weather higher core temperatures. For example, variable conditions, environmental temperature is dif- seedeaters Sporophila aurita , small passerines ferent for different animals! The early labora- that live in humid lowland tropics and thus tory measurements on thermoregulation by have diffi culty in cooling off via evaporation, Scholander and others had ignored these com- are able to tolerate core temperatures approach- plexities, and it was only after George Bakken ing 47 °C ( Weathers 1997 )! had published an infl uential theoretical paper MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 21

radiation (atmosphere) SUN

reflected sunlight direct sunlight evaporation radiation dust and particles (animal) radiation (vegetation) scattered sunlight

WIND

reflected sunlight radiation (ground) conductance

Figure 10. Overview of the ways in which an animal can lose or gain thermal and radiant energy due to the effects of weather. Adapted from Porter and Gates ( 1969 ).

on how to ‘sum up’ these different factors heat required to keep the mount’s temperature ( Bakken 1976 ), that biologists started to quan- constant in a natural thermal environment to a tify environmental temperatures. reference laboratory environment with known There are basically two ways in which oper- wind speed (usually 1 m/s). For example, if ative temperatures can be quantifi ed. First, all the mount in a natural setting requires the components making up operative tempera- same amount of heat per time unit as the same tures can be measured independently, after mount in a standard laboratory setting at which the Bakken equations ( Bakken and –20 °C, then the standard operative tempera- Gates 1974 , Bakken 1976 ) can be used to inte- ture equals –20 °C (even though the air tem- grate these different measures (e.g. Gloutney perature in the natural setting may have been and Clark 1997 ). Secondly, taxidermic models tens of degrees higher). can be used to measure operative temperature What new insights have been gained, (reviewed by Dzialowski 2005 ). Such copper compared to those already obtained by casts mimic the morphology and absorptivity Scholander and his team, since biologists of an animal, but with these models the effect escaped from their confi ned laboratory set- of wind on heat loss is ignored. A heated taxi- tings and measured the actual thermal condi- dermic mount measures so called ‘standard’ tions experienced by animals in the wild? First, operative temperature, which combines all the use of heated mounts made it possible to factors affecting heat fl ow, including wind. It convert the ‘old’ standard laboratory measure- does so in a direct way, by relating the rate of ments on metabolic rate into an ecologically 22 BASICS OF ORGANISMAL DESIGN

and behaviourally important parameter—the 3 maintenance metabolism of free-living birds in their natural environment (Wiersma and Piersma 1994 ). Maintenance costs represent 2 the sum of BMR and the extra costs of thermoregulation at environmental temperatures 1 BMR metabolism (W)

below the thermoneutral zone. Second, it predicted maintenance ARCTIC CANADA WADDEN SEA WADDEN SEA WADDEN yielded insights into the thermal properties of 0 ICELAND ICELAND available microhabitats, so that we now under- JFMAMJJASO ND stand how animals can reduce thermoregula- Figure 11. Variation in maintenance costs of islandica knots tory costs via microhabitat selection (e.g. throughout their annual cycle. At their non-breeding grounds Wiersma and Piersma 1994 , Cooper 1999 ) or (the Dutch Wadden Sea), during the coldest period of winter, islandica knots frequently exceed the inferred metabolic ceiling prevent overheating (e.g. Walsberg 1993 ). It of 4.5 BMR, assuming an average activity cost of 1.5 W. also made us realize that animals in open land- Adapted from Wiersma and Piersma (1994 ). scapes, where strong winds can prevail and where microhabitat selection is simply impossible, experience much greater heat losses than wintering period, when fuelling demands expected on the simple basis of air tempera- larger organs (see Chapter 5 ), leading to ele- ture alone. vated BMRs ( Piersma et al. 1996a , 1999b ), and This brings us back to red knots, which dwell thus, presumably, to additional heat produc- on wide, open mudfl ats during most of their tion. Furthermore, the growing layer of fat lives. The islandica subspecies overwinters in hampers the capacity to lose all the extra heat, northwest Europe, notably the Wadden Sea and and, as if that were not enough, the change into the British estuaries, where, during the coldest an often darker breeding plumage also increases period of winter, maintenance plus activity heat absorption ( Battley et al. 2003 ). costs frequently exceed the inferred metabolic How do tropically wintering shorebirds solve ceiling of 4.5 BMR (Fig. 11 ; Wiersma and Piersma all these heat problems? Most signifi cantly, they 1994 ; see Chapter 4 ). In contrast, the tropically seem to reduce their BMR upon arrival in their wintering subspecies (notably canutus and piers- warm winter quarters (Kersten et al. 1998), prob- mai ) experience the other side of the same coin: ably through the shrinkage of several visceral on a wide, open mudfl at there is no shade to organs ( Piersma et al . 1996a , Guglielmo and hide from a burning sun! At midday in such Williams 2003 , Vézina et al. 2006 ). Their BMRs conditions, operative temperature easily are 30% lower than those of comparable shore- exceeds body temperature (Rogers et al. 2006 ); birds wintering in the temperate zone (Kersten yet shorebirds cannot permit themselves to et al. 1998 ). Besides this, they have several other remain inactive during the hottest part of day, behavioural options available. For example, by since feeding activity is tidally, rather than diur- increasing the blood fl ow through their legs nally, regulated. We have seen that resident (giving them ‘hot legs’; Fig. 12 ), knots standing tropical animals tend to have lower BMRs, on dry sand can lose 16% of their metabolic heat which prevent heat-load problems; but it is production at 34°C air temperature (Wiersma known that shorebirds have relatively high et al . 1993 ). As the thermal conductivity of water BMRs ( Kersten and Piersma 1987 ), which make is 25 times that of air, knots can potentially lose them little heat machines. It is most likely that a lot of heat when standing in water ( Battley the worst problems occur during the end of the et al. 2003 ). Furthermore, by raising their back MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 23

tend to ‘seek’ randomness by evaporation. It is the faster-moving molecules that are able to do this, i.e. those which have the most kinetic energy. Those molecules which are unable to escape have a lower average kinetic energy, and thus the overall temperature of the remaining liquid water decreases. This is the simple physical principle of excessive heat loss via evaporative cooling. There are two main avenues along which water can evaporate from an animal’s body: via the skin (cutaneous evaporative water loss, CEWL) and via the lungs (respiratory evaporative water loss, REWL). Mammals have evolved a special way to speed up their cutaneous water loss: their skin is full of sweat glands. Birds lack sweat glands ( Calder and King 1974 ), but some of them living in extremely hot environments have found a remarkable alternative. Turkey vultures Cathartes aura, black vultures Figure 12. Infrared photos of red knots at two ambient Coragyps atratus, and wood storks Mycteria temperatures. The temperature scale given on the right is unique americana all excrete onto their legs when for each photo. By increasing the blood fl ow through their legs, exposed to increasing environmental tempera- and by raising the feathers on their back, knots can increase their tures, a phenomenon known as urohidrosis heat loss at high ambient temperatures such as 34 °C. ( Kahl 1963 , Hatch 1970 , Arad et al. 1989 ). This See Wiersma et al . ( 1993 ). excretion takes place at a high rate (approximately every 5 minutes), with each ‘dropping’ containing about 5 ml of liquid—a stinky, but feathers (ptiloerection), they reduce heat gain effective, way to keep cool! and increase convective and evaporative cool- In addition to evaporation, water leaves a ing ( Fig. 12 ; Battley et al. 2003 ). One other avenue body by excretion of (normal) faeces and urine for evaporative cooling (losing heat by bringing as a means of eliminating indigestible matter body water from a liquid into a gaseous state) is and toxic products of metabolism (and through to speed up respiration. However, the resultant the occasional production of eggs or milk). panting may solve a heat-load problem, only to With 99% of all molecules in a body being lead to another diffi culty. How to maintain a water molecules (Robbins 2001 ), there is a clear water balance when using so much (osmoti- need to compensate for all this water loss. Part cally free) water for evaporative cooling, whilst of this compensation comes naturally, as water being surrounded by salty seawater? is a by-product of metabolism (so-called oxidative water). Furthermore, food contains a Balancing water variable amount of water that can also be utilized. Finally, all remaining water defi cits can Evaporative water loss is an inevitable conse- be made up by drinking. quence of thermodynamics’ second law: water Until recently, and still echoed in some schol- molecules, nicely ordered in an animal’s body, arly textbooks, it was thought that birds living 24 BASICS OF ORGANISMAL DESIGN in hot and dry places showed no specifi c phys- there is less need for evaporative cooling. iological adaptations to reduce water loss (e.g. Second, as ventilation frequency is reduced, Maclean 1996 , Hill et al. 2004 ). Mammals were less water is able to escape via warm and known to have specifi c water-loss adaptations, water-saturated breath. So Bartholomew’s such as the unusual ability to concentrate conclusions were incorrect, and ecophysiolog- urine found in desert-dwelling kangaroo rats ical textbooks will need to update their section Dipodomys merriami ( Walsberg 2000 ), but birds on desert birds. were thought to prevent dehydration prob- Pigeon racers know it too: the fi rst thing lems simply by their behaviour (e.g. hiding in avian migrants do upon arrival is to drink the shade, or fl ying to a water pool when heavily to quench their thirst (Biebach 1990 , thirsty) and not by physiological means. This Pennycuick et al. 1996 , Schwilch et al . 2002 ). idea probably stems from the work of George Being aloft for a few days apparently places Bartholomew and colleagues, who concluded, heavy demands on a bird’s water economy, on the basis of almost a decade of work, that which is not surprising, as fl ying and drinking desert birds lack any physiological specializa- are simply incompatible. In addition, meta- tion ( Bartholomew and Cade 1963 , Dawson bolic water production is relatively low during and Bartholomew 1968 ). As Bartholomew did fl ight, despite an obviously high metabolic his work in an area that is relatively young on rate. This is because the main fuel, fat, has a an evolutionary time-scale (the deserts of the low water yield per unit energy ( Jenni and south-western United States), the generality of Jenni-Eiermann 1998 ). On top of this comes this conclusion was questioned by Joe Williams water loss through respiratory evaporation, and Irene Tieleman (2005). In a collaboration and this amount can be considerable, as respi- with Haugen and Wertz, they carried out their ration rates are ‘sky high’ during fl ight ( Carmi work on Old World larks and showed that the et al. 1992 , Klaassen 1995 ). upper skin layer in larks living in the hyper- There are, therefore, both behavioural and arid deserts of Arabia has a different lipid physiological solutions to all these potential structure from that of related larks living in dehydration problems in long-distance migrants. The Netherlands, and that this diminishes the The simplest solution is not to fl y during the rate of cutaneous water loss in the Arabian heat of the day in order to prevent extraordi- larks (Haugen et al. 2003a ). They further nary water losses. There are many records of showed that one of these Arabian desert larks, birds travelling by night and resting by day, the hoopoe lark, could fl exibly increase skin notably among trans-Sahara migrants (Biebach resistance to water vapour by changing the 1990 , Biebach et al. 2000 , Schmaljohann et al . lipid structure of its skin in response to increas- 2007 ). Another option may be to make regular ing temperatures (Haugen et al. 2003b ). The ‘drink stops’. Indeed, large waterfowl do make other avenue of evaporative water loss, via frequent stops of up to a couple of hours when respiration, was also found to be lower in travelling long distances (e.g. Green et al. 2002 ). desert larks. Compared to Dutch skylarks and These breaks are too short to replenish energy woodlarks, the Arabian desert larks had lower stores, which require a period of weeks in BMRs ( Tieleman et al. 2002 ), a result consistent these species ( Lindström 2003 ), but quite long with the earlier mentioned general trend of enough to ‘refi ll’ with enough water. Flying at lower BMR in the warm tropics. A reduced high altitudes is another way to reduce respi- BMR reduces evaporative water loss in two ratory water loss. Although the air gets thin- ways. First, less heat is produced and thus ner and thus more air needs to pass through MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 25 the lungs, the reduced air temperature also switch to energy-poor but water-rich (62%) reduces the temperature of the exhaled air and insects. This shows that food may sometimes thereby the amount of saturated water that it be too dry—but, as we shall see next, food may can contain (Carmi et al. 1992 , Klaassen 1995 , also be too wet! Schmaljohann et al . 2008 , 2009 ). On a physio- When red knots feed on shellfi sh under tem- logical level, birds may use protein rather than perate winter conditions, they ingest 27 times fat as their source of energy. Per unit distance the amount of water per day than is predicted fl own, the amount of water produced when on the basis of their body mass alone ( Visser burning protein is six times that of burning fat et al. 2000 ). They achieve this notable water ( Jenni and Jenni-Eiermann 1998 )! Although uptake for two reasons. First, shellfi sh contain this mechanism has been found in dehydrat- little energy relative to the amount of water ing humans (Bilz et al. 1999 ), so far it has not (0.3 kJ per g wet mass, while, for example, the been explicitly investigated in migratory birds. knots’ insectivorous summer diet contains 4.9 However, it is well known that migrant birds kJ per g wet mass). Second, the knots’ energy catalyse protein during their fl ights (e.g. demands during winter are relatively high, for Piersma and Jukema 1990 ) and the striking reasons explained in the previous section. observation of migrant birds found in the Thus, it may seem that, whenever enough har- Sahara with good fat stores but with low breast vestable molluscs are available, red knots muscle scores (L. Jenni unpublished data as should never run out of water. There is one cited in Klaassen 2004 ) points in the right problem, however. All this ingested water is direction. too salty to become part of the bird’s body fl u- In spite of all these water problems that ids or to be used for evaporation. Red knots, migration may entail, the idea that dehydration together with other marine molluscivores, risk may govern migratory behaviour is not need a highly effi cient salt-excretion system widely accepted. Klaassen ( 2004 ) suspects that ( Peaker and Linzell 1975 , Hughes et al. 1987 ). this is because lowered body water contents And yes indeed, they do have one: corrected have only rarely been found in long-distance for body mass differences, their nasal salt trans-desert migrants ( Biebach 1990 , Landys glands were the largest in a comparison with et al. 2000 ). However, as noted by Klaassen, 21 other Charadriiformes species ( Staaland body water content remains remarkably stable 1967 ; Fig. 13 ). In spite of this, it may be that the in animals dying from water restriction (Chew constraint on food intake and the ability to 1951 , 1961 , Dawson et al. 1979 ), and it is thus a evaporate osmotically free water lies in the poor indicator of dehydration stress. bird’s capacity to excrete salt. Indeed, knots Indirectly, the (un)availability of water may reduced their food intake when the water pro- affect migratory performance during the fuel- vided was experimentally changed from fresh ling phase. Blackcaps Sylvia atricapilla stop- to salt ( Klaassen and Ens 1990 ). Nevertheless, ping-over in dry fruit plantations in southern over a time period of just a few days, the knots Israel during southward migration gain body adjusted to the salty water and increased food mass faster when provided with extra fresh consumption again, suggesting that they were water ( Sapir et al. 2004 , Tsurim et al. 2008 ). The able to increase the capacity of their salt glands. extra water available enables them to take up This fl exible adjustment in the capacity and higher quantities of fat-rich but water-poor size of the nasal glands has also been found in (35%) fruits. During water-restricted condi- several other avian species (Peaker and Linzell tions, the blackcaps stop feeding on fruits and 1975 ). 26 BASICS OF ORGANISMAL DESIGN

which he explained by the continuous degrada- 1000 tion of phytoplankton keeping this ratio in the water column constant ( Redfi eld 1958 ). In more recent times, larger datasets and more precise Red Knot measurements have yielded some small modi- 100 fi cations here and there, but overall they still support the generality of the magic ratio 106:16:1 nasal gland mass (mg) in the offshore ocean ( Hoppema and Goeyens 10 1999 ). 10 100 1000 body mass (g) Unlike the elemental equality between the sea and its phytoplankton, most ‘higher’ Figure 13. Scaled to body mass, nasal salt glands are largest organisms face a dramatic difference between in red knots in a comparison with 21 other Charadriiformes the elemental composition of their food and species. Based on data presented by Staaland ( 1967 ). their own bodies. Take termites, for example, whose bodies have a C:N ratio of approxi- Elements of a body mately 5, but consume dead wood with ratios between 300 and 1000 ( Higashi et al. 1992 ). It is obvious that shellfi sh-eating knots ingest That represents a lot of wood that needs to be more salt than they need, but, on the other hand, chewed before a termite satisfi es its nitrogen they cannot live without a certain amount of it. needs! Since photoautotrophs are almost The sodium in salt is needed for the regulation always made up of much more carbon than of body fl uid volume and osmolarity, acid–base their herbivorous consumers, a surplus of car- balance and tissue pH, muscle contraction and bon is a general phenomenon in herbivores. nerve impulse transmission ( Robbins 2001 ). As With the initial focus in ecological stoichiome- well as sodium, there are many more elements try on small ectothermic organisms, notably that need to be kept in balance in an animal’s the freshwater cladoceran Daphnia ( Sterner body, since defi ciencies and imbalances of and Hessen 1994 , Urabe et al. 1997 ), stoichiom- minerals are considered to be an important etrists mostly considered the surplus of C as a determinant of animal condition, fertility, pro- problem rather than a benefi t, as it reduces the ductivity, and mortality ( Underwood 1977 ). effi ciency of growth in such organisms The fi eld that studies the balance between the ( Anderson et al. 2005 , Hessen and Anderson elemental make-up of animals and their food is 2008 ). However, this point of view changed called ‘ecological stoichiometry’, an area of drastically when they turned their attention to research that has fl ourished over the last dec- the endotherms. ade ( Sterner and Elser 2002 , Elser and Hamilton Marcel Klaassen and Bart Nolet (2008) 2007 ). One of the best-known applications of hypothesized that endothermy is the perfect stoichiometric principles to ecology is the solution to excess carbon: just by ‘burning it Redfi eld ratio, named after Alfred C. Redfi eld, off’ endothermic animals get rid of their sur- an oceanographer from Harvard and the Woods plus carbon, while at the same time taking Hole Oceanographic Institute. He discovered a advantage of the heat-release to keep their remarkably constant ratio between carbon (C), bodies warm. They tested this idea by allomet- nitrogen (N), and phosphorus (P) (106:16:1), rically comparing nitrogen requirements rela- both in the world’s oceans and in the phyto- tive to energy requirements in birds, eutherian plankton living in them (Redfi eld 1934 ), and mammals, marsupials, and reptiles. Birds and MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 27 mammals did indeed have relatively low nitro- pods must have been partly carnivorous in gen requirements compared to reptiles, which order to grow, a pattern still seen in juvenile explains why they are able to tolerate lower marine iguanas Amblyrhynchus cristatus ( White quality food (i.e. food with higher C:N ratios). 1993 ). However, if growing larger has been the In fact, even among birds and mammals, the evolutionary solution that overcame one die- phylum with the highest energy requirements tary problem in sauropods, it may have led to (birds) had the lowest (relative) nitrogen yet another dietary problem in its place. Larger requirements. This way of thinking sheds new animals need relatively bigger bones ( Schmidt- light on the evolution of endothermy. As men- Nielsen 1984 ; and already recognized by tioned earlier in this chapter, one of the advan- Galileo in 1637). Of all the organs in a verte- tages of being homeothermic is the ability to brate, bones contain the highest levels of phos- capture cold-blooded, and thus slower, prey. It phorus ( Elser et al. 1996 ; Fig. 14 ), as they are is thus generally thought that homeothermy made up of hydroxyapatite, a phosphorus originated in small predatory carnivores mineral. Adult sauropods were thus likely to ( Bennett and Ruben 1979 , Hayes and Garland have high P-demands ( Lane 2009b ). Being 1995 ). In contrast, the stoichiometric hypothe- giants, they may have fulfi lled these needs by sis of Klaassen and Nolet opens up the possi- consuming the leaves of large and fast-grow- bility that endothermy began in herbivores ing trees, which contain relatively high frac- eating too much carbon to meet their nitrogen tions of phosphorus. Although these can only requirements. As is often the case with ideas be mere speculations, it is clear that living in a about evolutionary origins, the hard-to-prove true Jurassic Park was not at all easy! facts lie hidden in the fossil records, and we The stoichiometry of bones brings us neatly have not yet heard the last word on this subject back to red knots. Apart from phosphorus, ( Lane 2009b )! bones also contain a lot of calcium. When ‘Burning it off’ may be one solution to the reproducing, most arthropod-eating birds need problem of excess carbon, but growing very large may be yet another way to deal with it. 50:1 15:1 7:1 Long ago, in the Jurassic era, typical plants 16 skin were conifers, cycads, and ferns, which have heart inherently low nitrogen content. The higher 12 kidney liver carbon dioxide concentrations in the air dur- brain muscle ing that period made things worse, suppress- % N 8 ing the nitrogen levels in those plants even 1:1 further. Jeremy Midgley and colleagues ( 2002 ) 4 blood skeleton postulated the idea that sauropod dinosaurs could only live on such low quality food by 0 growing very large. With an increase in body 024 6128 10 % P size, relative metabolic rate and relative growth rate (both expressed per unit body mass) Figure 14. Stoichiometry of organs: both blood and, especially, decline. The lower the metabolic rate and the skin are very protein-rich and thus have high N:P ratios. By contrast, bones are rich in phosphorus, due to the deposition of growth rate, the less protein, DNA and RNA apatite (a group of phosphate minerals), and thus have a low N:P needs to be produced, which tempers the ratio. Percentages refer to dry masses, while the white lines refer nitrogen needs and thus allows a low-quality to atomic ratios. diet. Midgley et al. suggest that juvenile sauro- From Elser et al. ( 1996 ). 28 BASICS OF ORGANISMAL DESIGN extra calcium in order to produce strong egg hold the record in skeletal calcium dynamics shells, and many obtain this by supplementing ( Piersma et al. 1996b ). But such an increase is their insectivorous diet with molluscs, e.g. still small compared to their real asset: red landsnails in the case of great tits (Graveland knots, together with bar-tailed godwits and et al. 1994 ). On their tundra breeding grounds, other long-distance migrant shorebirds, can red knots are specialist insectivores too, and store fuel in the form of fat in such amounts female knots may thus also face calcium defi - that upon take-off on long fl ights half of their ciencies prior to egg-laying. As knots daily body can be pure lipid (e.g. Piersma and Gill ingest huge amounts of calcareous shell mate- 1998 ). At other times, e.g. upon arrival or when rial when not on the tundra (which is during close to starvation, their bodies contain very lit- most of their life cycle), one would expect them tle fat. The amazing dynamics of fat in shore- to store calcium in their bones before departure bird fat stores are a consequence of variable to the breeding grounds—and indeed they do. energy balance states. However, the highest rates of calcium accumulation in their bones occur after arrival at the Birds are not airplanes breeding grounds (Fig. 15 ), presumably from ingested teeth and bones extracted from lem- Red knots ready to take off for their 2800-km ming carcasses (MacLean 1974 ). The most long non-stop journey from Selvogur, south- likely functional explanation for not storing all west Iceland, to their Canadian/Greenlandic the required calcium at the stopover is that breeding grounds carry about 80 g of fat. At lighter bodies make for cheaper fl ying. If the the end of the fl ight, there is 15–40 g of fat left remaining calcium can be collected upon ( Piersma et al. 1999b , Morrison et al . 2005 ). arrival, then that is what should be done. With These knots build up and break down this an almost 50% increase in skeleton mass, knots amount of fat four times a year (fuelling four non-stop fl ights), a seasonal routine that is maintained throughout an average life of 5–10 ICELAND ELLESMERE ICELAND 1.6 ISLAND years. Fat is by far the most cost-effective way to store chemical energy. Its energy density 1.5 females males amounts to 38 J/g, compared with approxi- 1.4 mately 5 J/g for wet protein or 4 J/g for wet 1.3 carbohydrates. The main reason for this differ- 1.2 ence is that fat can be stored ‘dry’, whereas 1.1 relative ash mass relative stored protein and glycogen typically contain 1.0 >70% water. Even without water, fat (40 J/g) is 0.9 more than twice as energy-rich as protein (18 10 20 30 9 J/g) or carbohydrates (18 J/g; Jenni and Jenni- May June July Eiermann 1998 ). Furthermore, the energetic Figure 15. Changes in ash mass of the skeleton of red knots maintenance cost for adipose tissue (the tissue during their stopover in Iceland and while at their Canadian holding body fat) is at least an order of magni- breeding grounds at Ellesmere Island. The ash of skeleton tude lower than that for other tissues ( Scott contains 25–30% calcium (Grimshaw et al . 1958 ), and thus and Evans 1992 ). It is, therefore, not surprising these changes also refl ect changes in a knot’s calcium stores. Ash masses are scaled relative to the average ash mass of that fat was long thought to be the only source males in Iceland. of energy during migratory fl ights ( Odum From Piersma et al . ( 1996b ). et al. 1964 ). Birds were thought of as airplanes, MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 29 fi lling up and emptying their fuel tanks, while metabolism and chemical body composition holding structural mass constant. changes during long-term fasting in these pen- Gradually this ‘aircraft-refuelling paradigm’ guins (Robin et al. 1988 ) and in other birds (Le was abandoned. Studies accumulated evidence Maho et al. 1981 , Cherel et al. 1988 ), and con- that fuelling birds also increased the amount of cluded that three distinctive fasting phases can protein tissue, e.g. in the form of larger fl ight generally be recognized ( Fig. 16 ). During the muscles or a bigger heart (in order to gain short phase I, characterized by a rather rapid power; e.g. Fry et al. 1972 , McLandress and body mass decline, the animal adapts to long- Raveling 1981 , Piersma et al. 1999b ), while term fasting by decreasing its protein catabolism simultaneously reducing the mass of their and mobilizing its lipid stores. The loss of body nutritional organs before take-off (so as to fl y mass is tempered during phase II, when the ani- with a total mass as low as possible; e.g. Piersma mal has fully switched to relying on its fat stores. 1990 , 1998 , Piersma and Gill 1998 ). In fact, pro- During the last phase, phase III, fat stores have tein may be used as actual fuel: the central nerv- run out and the animal starts to burn structural ous system is unable to oxidize fatty acids and tissue, thus catabolising mostly protein again thus needs glucose as an energy source, which and losing body mass at a high rate. As the ani- is derived from amino acids (Jenni and Jenni- mal is now burning its ‘true self’, it will face Eiermann 1998 ). In addition, the burning of fat death unless it soon fi nds food. itself requires some protein because intermedi- ates in the citric acid cycle need to be replen- phase: ished on a continuous basis in the form of amino

START I II III DEATH acids ( Jenni and Jenni-Eiermann 1998 ). Further- A BODY MASS LOSS more, as burning protein yields so much more metabolic water than burning fat, a fuelling migrant storing extra protein may prevent (mass/time) dehydration during the fl ight. Finally, muscle body mass loss tissue undergoes continuous damage during strenuous exercise and needs to be repaired, B FAT CATABOLISM probably immediately after ending a long-distance fl ight ( Guglielmoet al. 2001 ). A bird in fl ight burns both fat and protein fat-metabolites ( Jenni-Eiermann et al . 2002 ), but the ratios of fat (mass/volume) ( β -OHB) in blood catabolism and protein catabolism may change with depletion of the stores. For example, when C PROTEIN CATABOLISM fat stores become depleted the body needs to rely on protein as the prime source of energy ( van der Meer and Piersma 1994 , Jenni-Eiermann (mass/time) et al . 2002 ). Such prolonged fasting periods nitrogen excretion occur, for example, in male emperor penguins time course of fast Aptenodytes forsteri, which do not eat for as long Figure 16. Conceptualization of the physiological changes as 115 days while incubating the eggs at colonies that occur in the course of prolonged fasting: (A) changes in far from the sea in Antarctic winter conditions, body mass, and shifts in the metabolic dependence on (B) fat losing 40% of their initial body mass ( Groscolas and (C) protein. 1986 ). Yvon Le Maho and his team studied the From van der Meer and Piersma ( 1994 ). 30 BASICS OF ORGANISMAL DESIGN

Shorebird insurance strategies 220 islandica canutus Animals are able to anticipate a forthcoming 200 fasting period or a period of high endurance expenditure by storing extra fat mass, thereby 180 extending the ‘economic’ phase II of fasting. This is most obvious in migrant birds fuelling 160 for their long-distance journeys. In a non-mi- body mass (g) 140 gratory context, birds and mammals can also

lay down extra fat stores, which enable them 120 to overcome a period of probable food scarcity. Take, for example, red knots during winter in 100 the Dutch Wadden Sea. Not only are the envi- ASO ND JFMAMJ JASO ronmental conditions harsh (low ambient tem- Figure 17. Summary of the annual cycle in body mass of peratures), they are also unpredictable. a typical islandica knot wintering in northwest Europe and Mudfl ats can be ice-covered for periods of a typical canutus knot wintering in West Africa. Both sub- more than a week, and even low tides can be species fuel up four times a year, each time covering a long too high for enough mudfl ats to become non-stop fl ight of 1850–5100 km. Additionally, islandica deposits extra mass during midwinter. exposed (due to storms). Red knots wintering From Piersma and Davidson ( 1992 ). in the tropics face much better and much more predictable conditions: there is never frost and strong onshore winds are uncommon. Moreover, because most tropical wintering 135 2 occasional scarcity sites are located so close to the open ocean, no scarcity 2 tidal height is hardly ever affected by wind 0 130 2 ( Wolff and Smit 1990 ). Indeed, these differ- 1 2 ences in environmental predictability seem to 1 0 be refl ected in the midwinter body mass pat- 2 125 terns of Wadden Sea wintering and tropically wintering red knots, respectively. The body 1 2 mass of knots wintering in the Wadden Sea body mass (g) average 120 0 0 increases from 130 to 160 g during winter, 0 –2 while it remains low at 120–125 g in knots win- –4 012 tering in West Africa ( Fig. 17 ). 115 In order to put these suggestive descrip- 0 10 20 30 40 50 60 70 80 tional data to a true test, we performed an number of days since capture indoor experiment under controlled condi- Figure 18. Body mass of red knots in relation to their tions. Captive knots (of the islandica subspecies experience of food scarcity. Four fl ocks of seven birds wintering in The Netherlands) were either occasionally had no food for 1 or 2 days (the number above the offered food on a continuous basis, or were closed dots indicating the number of days without food in the preceding week). Three fl ocks of seven birds (open dots) always given food at random moments in time. This had food. The inset shows the difference in weekly change in regime was maintained for about ten weeks body mass between starved and fed birds as a function of the with clear-cut results: knots that always had number of days without food in the preceding week. food available were always about fi ve grams Based on T. Piersma et al . unpubl. data. MAINTAINING THE BALANCE OF HEAT, WATER, NUTRIENTS, AND ENERGY 31 lighter than the birds whose food was taken minimal thermostatic cost = lowest BMR level away at stochastic moments (Fig. 18 ). It is 120 A likely that this difference in body mass mainly refl ects a difference in the amount of fat stored, 110 but it is evident that birds can indeed insure 100 r2 = 0.68 themselves against elevated levels of starva- 90 tion risk (see also Piersma and Jukema 2002 , MacLeod . 2005 ). 80 et al body mass (g) 70

Dying strategically 60 With the end of this chapter comes a discussion 2.0 B on the end of a life. You have just been reading 1.5 STRUCTURAL FAT that shorebirds wintering under harsh and cold 1.0 conditions have more fat stores to deplete before fat mass (g) fat they die than conspecifi cs wintering in the more 0.5 predictable and mild tropics. But there is a snag 0.0 in this argument: analysis of many hundreds of 35 C carcasses of starved oystercatchers and red knots has revealed that dying in milder climates 30 r2 = 0.61 occurs with lower lean body masses than dying 25 under harsher wintering conditions ( Piersma et al. 1994a ). Red knots starving in a typical west 20 Dutch Delta Wadden Sea European winter had 25 g more fat-free mass Mauritania left than conspecifi cs that died in West Africa dryfat-free mass (g) 15 Netherlands capt. Guinea-Bissau capt. (Fig. 19 ). The same principle applied to oyster- 10 S. Africa capt. catchers: individuals wintering in Scotland died 0 123 at a 30–50 g greater body mass (with almost no maintenance metabolism (W) fat left) than individuals wintering in the milder Figure 19. Starved bodies of red knots differ in mass and Wadden Sea. How can we explain these remark- composition depending on where they died. (A) body mass, (B) able observations? extracted fat mass, and (C) fat-free dry mass are plotted as a function Think of a bird’s body as consisting of an of the maintenance costs in the 5-day period preceding death. essential part, without which it cannot survive, From Piersma et al . ( 1994a ). and a remaining part consisting of reserves and stores. The essential part cannot be cat- Because the demands for a given body mass abolized, since if this were to happen, the ani- are higher under a high working level than mal would die immediately because of under a low working level (i.e. because of dif- insuffi cient energy output. Instead, the energy ferences in maintenance costs), the point of output comes from the reserves, i.e. the non- death occurs at a higher body mass when the essential part, and needs to supply the entire animal is working harder ( Fig. 20 ). body, essential and non-essential components This higher body mass at death, when work- included. A bird dies of starvation when the ing hard, offsets the life-prolonging effect of energy supply is less than the energy demand. midwinter fattening in temperate wintering 32 BASICS OF ORGANISMAL DESIGN

ESSENTIAL PART NON-ESSENTIAL PART it is ‘burned out’. This point shows that a body echoes an individual’s ecology. As we will point out again and again, bodies are by no means static machines carrying a single-sized fuel tank, running on a single type of fuel. On MR-GENERATED the contrary, their fuel tank can shrink and expand, it can use multiple types of fuel, and metabolic rate MAXIMUM the size and capacities of the engines can be adjusted. It is indeed fair to say that ‘birds are not airplanes’. MR demanded at low work level MR demanded at high work level

0 100 total body mass (%) Synopsis Figure 20. Why work level may determine body mass at Animals face two relevant physical laws: (1) death: both the energy production (thick line) and the energy demand (thin lines; two work levels) vary as a function of total they cannot escape the fi rst law of thermody- body mass. However, because only the fat-free component of a namics (the law of conservation of energy), body can produce energy, while it is the entire body that and (2) they are bound by the second law of demands energy, hard-working birds die (death occurs when thermodynamics (the law of spontaneous production equals demand) with greater body mass. ‘spreading-out’ of energy and matter). In order From Piersma et al . ( 1994a ). to reduce the amount of entropy in their bodies, organisms have actively to maintain a bal- areas. Based on the wintering body masses of ance in terms of the gains and losses of heat, non-fasting individuals, it has been calculated water, nutrients, and energy. We have seen that red knots can go without food for 4–5 various traits that organisms have evolved to days, irrespective of where they live. The aver- keep these budgets in balance across a wide age knot wintering in the Dutch Wadden Sea range of environmental challenges. In the next has about 20 g of fat and 25 g of protein stored, chapter we will consider how these traits are while a knot wintering in West Africa has only themselves balanced against each other in 3 g of fat but 55 g of protein to catabolize before order to prevent excess capacity. CHAPTER 3 Symmorphosis: principle and limitations of economic design

A well-trained man, a frog, most other features of the mammalian respira- and a hummingbird tory system indicated that the capacity of one part of the system was closely matched to the ‘If a well-trained man runs, the mitochondria of capacity of other parts (Weibel et al . 1991 , 1998 ). his muscles can consume over fi ve liters of oxy- Animals are not built in wasteful ways, animals gen every minute. To maintain oxidative metab- are ‘economically designed’. olism at such a level, a continuous fl ow of One organismal design feature must surely oxygen through a series of connected compart- have been key to this insight: the realization ments has to be maintained: from the pool in that most organ systems are arranged in series, environmental air, oxygen is carried into the rather than in parallel. The respiratory chain lung by inspiration, is transferred to the red that leads up to muscular work is clearly serial. blood cells, moved into the tissues by circula- First, oxygen from the air is extracted by the tion, and fi nally reaches the cells and their mito- lungs and captured by the blood. The oxygen chondria by diffusion. Conversely, the carbon is then transported by the blood that is pumped dioxide produced during the metabolism of by the heart and delivered to the cells where it organic compounds needs to be moved along is ‘burned’ in the mitochondria to generate much the same pathway and by similar mecha- ATP to make the muscle work. In a summary nisms, only in the reverse direction.’ So began address in 1985, Ewald Weibel clarifi ed this the paper in which Richard Taylor and Ewald point with a cartoon of a frog ( Fig. 21 ). The car- Weibel (1981) articulated the principle of ‘sym- toon illustrates the sequence from fuel to force, morphosis’ for the fi rst time. Symmorphosis also incorporating aspects of the nervous con- represented the core idea germinating from a trol of muscles, for the sake of physiological meeting of two minds who both fi rmly believed completeness. In a summary of the amazing that animals ‘are built reasonably’ (Weibel and feats of aerodynamic performance by hum- Bolis 1998 ). Taylor worked at Harvard University mingbirds, Raul Suarez ( 1998 ) presented a on the effi ciency of animal locomotion, Weibel somewhat evolved ‘frog-cartoon’, a hovering at the University of Berne in Switzerland on the hummingbird ( Fig. 21 ). His account also design of well-functioning lungs. Jointly they focused on the respiratory chain, in this case to set up a research programme to fi nd out whether show how hummingbirds are able to generate lungs were adjusted to the maximal rate of oxy- the highest known mass-specifi c metabolic gen consumption in running animals of differ- rates among vertebrates ( Suarez 1992 ). ent size ( Weibel 2000 ). Although the answer Hummingbirds can do this because they show: would eventually be negative with respect to (1) high oxygen-diffusing capacity in the lungs, the uptake of oxygen in the lungs of mammals, (2) high pumping performance by the heart,

33 34 BASICS OF ORGANISMAL DESIGN

FUEL FORCE CONTROL

1985 1998

Figure 21. Schematic diagrams of a sitting frog and a fl ying hummingbird to show how organ systems, such as those for respiratory performance, are arranged in sequence. The frog is directly from Weibel (1985 ), who used it to show how the heart is coupled to the lungs, the blood stream to the heart, the mitochondria to the bloodstream, and the muscular sarcomeres to the mitochondria. He also indicated how muscle function is controlled by the nervous wiring from spine to bone and muscle. The hummingbird on the right is from Suarez (1998 ). It provides a visual template to make the point that hovering is made possible by large pectoral muscles, powered by abundant mitochondria to which oxygen from the lungs and metabolic fuels from the gut must be transported at high rates.

(3) high rates of oxygen delivery to muscle all, ‘evolution is a design process that, as in the fi bres due to dense capillary networks, (4) high fi rst chapter of the Book of Genesis, makes intracellular densities of mitochondria that something and then looks at it to see if it’s any also show (5) internal surface enlargements to good’ ( Vogel 2003 , p. 497). What we discuss in boost ATP production, and (6) high concentra- this book is what Richard Dawkins (1995 ), with tions of enzymes involved in the energy his usual fl air for understatement, has called metabolism of the muscles. Thus, humming- ‘God’s utility functions’. What is maximized in birds are specialized in burning fuel fast; of evolution (i.e. utility in economics) is long-term course, they can only do so if there are enough reproductive success (fi tness). In a world of lim- substrates to burn. Note that in Suarez’ ver- ited resources and rampant competition, a good sion of the cartoon, that small piece of intestine way to achieve this is the wise use of energy is a little more visible than in Weibel’s version and supplies. This includes, of course, the evo- (Fig. 21 ). In this way Suarez emphasized the lutionary engineering of bodies. So, indeed, importance of fuel sequestration in the gut, the ‘engineering’ is another word that we will use. parallel organ system with probably the tight- In fact, according to Daniel Dennett ( 1995 ), the est coupling to the respiratory system. marriage of biology and engineering is the central feature of the Darwinian revolution. Economy of design Enter industrialist Henry Ford (1863–1947), prolifi c inventor, father of mass production In an age when problematic arguments against and assembly lines, and, of course, maker of evolution by proponents of intelligent design the model T-Ford, the low-cost, mass-produced have not quite been put to bed (e.g. Coyne vehicle that made cars a commodity. Henry 2009a ), it may seem hazardous to use the word Ford wrote up his ‘life and works’ together ‘design’. Yet, because design arguments are so with Samuel Crowther (1922), and this book helpful in appreciating the workings of evolu- makes abundantly clear how ‘economic’ Ford tion, this is precisely what we plan to do. After was with respect to almost everything. In SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 35 regard to the economy of production and the extravagance ( Briffa and Sneddon 2007 , but see product itself, he had this to say: ‘the less com- Oufi ero and Garland 2007 ). In today’s world of plex an article, the easier it is to make, the evolved economic competition, the sex appeal cheaper it may be sold, and therefore the of cars is certainly a much greater selling factor greater number may be sold’. In many ways, than it was in the days of Henry Ford. the concept of greatest numbers sold parallels Nevertheless, a vindicated simple example of the Darwinian idea that only lineages with energy minimization in nature (indeed, it was more successful offspring than competing lin- published in the journal Nature ), was provided eages will prevail. And small cost differences by Richard Taylor, one of the fathers of the sym- do seem to make the difference. An appealing morphosis principle ( Hoyt and Taylor 1981 ). example was provided by Daniel Dykhuizen When a horse selected her own speeds, walking, (1978 ) who compared the growth of two trotting or galloping, at each of the three gaits strains of the common intestinal bacterium she chose a narrow range of non-overlapping Escherichia coli, a wild type and a mutant una- speeds (Fig. 22 ). The same horse was then run ble to synthesize the amino acid tryptophan. on a treadmill where she could be coaxed into In chemostats (vessels with a constant supply using other than her preferred speeds. With a of nutrients), the two strains were brought mask on, her oxygen consumption rates were together. Without a tryptophan supply in the measured. The measurements ( Fig. 22 ) clearly chemostat’s solution, the mutants would die demonstrated that whether she walked or trot- out rapidly. However, in a solution with tryp- ted or galloped freely, the horse always chose tophan, the mutant outcompeted the wild the speeds where her oxygen consumption rates type. The real surprise was that, by not having (i.e. energy expenditure) were lowest for that to incur the biosynthetic costs of tryptophan, particular gait. the winning mutant only saved a tiny fraction, Let us now go back to Henry Ford one more an estimated 0.01%, on the wild-type’s energy time. According to an anecdote expounded by budget. In another—mammalian—example, Nicholas Humphrey (1983 ), Ford commissioned by not developing and maintaining a useless a survey of car scrapyards to see whether there visual system in their pitch dark world, mole were Model T parts that never failed. It turned rats appeared to free up 2% of their energy out that almost anything could break: axles, budgets (Cooper et al . 1993 ). With a 200 times brakes, pistons, steering wheels, whatever you greater saving than the mutant gut-bacteria, will. There was one exception: pins called ‘king- one can imagine how quickly mole rats went pins’ invariably had years of functional life left in blind after they became permanent under- them. ‘With ruthless logic’ (Humphrey 1983 ), ground citizens. Ford concluded that the kingpins of Model T However, completely simple, one- dimensional were too good for their job and should be made examples of the economic imperative in biologi- more cheaply. If he had wanted to build a cal design are not so easy to fi nd (although sev- Mercedes, the pins would have been fi ne. He eral, more complicated, examples will be worked would, however, have been obliged to improve up in the rest of this chapter). In fact, that other all other parts of the car and make it much more powerful Darwinian mechanism, sexual selec- expensive. The point is that a ‘design had to bal- tion, or the acquisition of good sexual partners ance. Men die because a part gives out. Machines on the mating market ( Cronin 1991 , Ridley 1994 , wreck themselves because some parts are weaker Miller 2001 ), will generate the opposite of thrifty than others. Therefore, a part of the problem in designs in nature; it leads to a fair amount of designing a universal car was to have as nearly 36 BASICS OF ORGANISMAL DESIGN

symmetry). In 1981, Taylor and Weibel provided the following defi nition: ‘state of structural design commensurate to functional needs result- walk trot gallop ing from regulated morphogenesis, whereby the

10 formation of structural elements is regulated to satisfy but not exceed the requirements of the 8 functional system’. Symmorphosis thus predicts 6 that all structural elements of a body, or at the 4 least its subsystems, are fi ne-tuned to each other and to overall functional demand (Weibel 2000 ). 2 Because a serial system is as strong as the weak- number of choices for different gaits different of choices for number 0 est link, any element in the chain that would be 28 stronger than the weakest would be wasteful. 26 Most research efforts to develop empirical 24 tests of the symmorphosis principle have been

per m distance) 22

2 devoted to the respiratory chain as the model 20 system ( Fig. 23 ). Such tests require a comparison 18 of performance levels of different functional ele- 16 ments in the respiratory chain with overall respi- 14 ratory performance. However, comparisons at cost of transport (ml O 12 just any level of oxygen consumption, carried 0 1234567out within individuals or groups, will be unin- running speed (m/s) formative. This is because, at submaximal per- Figure 22. Horses show three distinct gaits (walk, trot, or formance levels, all functioning elements of the gallop), each of which is carried out over a limited range of respiratory chain have to be in balance: there are preferred speeds (top panel). When horses were forced to move at a range of speeds on a treadmill, the preferred running no escape pathways for oxygen once it is taken speeds at each of the three gaits corresponded to where they up by the bloodstream. The level of performance achieved minimal costs of transport (lower panel). of the different respiratory elements will there- Compiled from Hoyt and Taylor ( 1981 ). fore be determined by the demand for oxygen at the receiving end, the mitochondria (di Pamprero 1985 , Lindstedt and Conley 2001 , Burness 2002 , as possible all parts of equal strength consider- Darveau et al. 2002 ). Only at truly maximal per- ing their purpose’ (Ford and Crowther 1922 ). formance levels by the entire organism, i.e. when The message is loud and clear: economic designs animals work so hard that they reach maximal avoid unnecessary excess, a kind of symmorphic oxygen consumption rates (Fig. 24 ), and when logic avant Taylor and Weibel (1981 ). 90% of both the blood and the oxygen in the bloodstream are delivered to the muscles ( Figs. Symmorphosis: the principle 23 and 24 ), is it informative to determine if work- and the test ing capacities of elements in the chain leading up to muscular work (the supply side) match oxy- The term symmorphosis comes from Greek, as gen consumption rates (the demand side). technical biological terms tend to, with ‘morpho- Keeping in mind the fi rst law of thermodynam- sis’ meaning ‘formation’ and ‘symmorphosis’ ics (Chapter 2 ) , working capacities of elements of literally meaning ‘balanced formation’ (think of the respiratory chain cannot be lower than those SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 37

STRUCTURE PROCESS Mouth O 2 pulmonary Lung AIR ventilation BLOOD pulmonary diffusion

Heart cardiac output / circulatory convection Blood vessels

Capillaries skeletal muscle CELL diffusion

Mitochondria mitochondrial oxidative phosphorylation ATP ATP ATP Ribosomes

Muscle tissue Membrane pumps Sacromeres synthesis ion transport muscle work LOW HIGH metabolic rates Figure 23. Schematic representation of the cascade of structural elements and functional steps in a mammalian respiratory chain, from the fl ow of oxygen into the lungs to the supply of ATP to fuel action by the contractile elements of the muscles, the sarcomeres. On the left of the diagram, some of the structural elements are mentioned; and on the right, some of the processes involved in getting oxygen to the mitochondria for ATP generation. The underlined structures and processes were examined in the greatest dedicated empirical test of the symmorphosis principle so far, and are discussed in detail in the text. At the bottom end of the scheme it is shown how, at rest, during low active metabolic rates, energy demands are mainly due to the maintenance processes in the tissues; for example, protein synthesis by ribosomes (left) or ion transport by membrane pumps (centre). With increasing levels of exercise, muscle work (right) takes over the respiratory cascade from lung to capillaries and mitochondria, and the ease with which oxygen can be transported through the cascade limits the maximal oxygen consumption rate. Compiled from Lindstedt and Jones ( 1987 ), Weibel et al . ( 1991 ), and Weibel ( 2000 , 2002 ). achieved by the whole organism. What is critical mammals in which peak aerobic performance in tests of the symmorphosis principle is to see levels had been measured. Animals that varied whether the working capacities of any of the ele- from mice, mongooses, wildebeest, lions, to ments may be greater than necessary. eland (the largest antelope), steers, and horses The research programme that Taylor and (spanning four orders of magnitude in body Weibel and their colleagues set up to test sym- mass) were exercised in what was no doubt an morphosis contained two aspects. First of all equally large variety of treadmills.1 They then they went to East Africa to enlarge the range of sacrifi ced the experimental animals. Using

1 As recounted by friend and colleague Steven Vogel (2001 ) ‘in this business, data do not come easily. Determining condition is tricky, and providing motivation is worse; animals don’t have the investigator’s sense of the importance of working as hard as possible. On one occasion, a cheetah expressed its antipathy toward task and taskmas- ter by jumping off the treadmill and doing Dick [Taylor] quite a lot of damage.’ 38 BASICS OF ORGANISMAL DESIGN

SYSTEMIC PULMONARY MUSCULAR RENAL SPLANCHIC OTHER

liver

muscles brain LR lungs etc. REST AT kidneys gut heart

LR DURING MAXIMUM EXERCISE DURING MAXIMUM

Figure 24. Distribution of blood fl ow among the different circulatory subsystems in an adult human at rest (top) and during maximum exercise (bottom). The width of the blood vessels indicate, in linear ways, the relative blood fl ows (i.e. 6 l/min through the pulmonary system at rest, and 25 l/min during maximum exercise). Although the lung receives 100% of the blood fl ow in both states, muscle blood-fl ow is 20-fold higher during intense exercise (22 l/min) than during rest (1.2 l/min), at the expense of blood fl ow to the kidneys and the liver. Blood fl ow through the rest of the body (especially the brain) is defended during exercise. Inspired by, and based on, Weibel ( 2000 ).

specialist histological and morphological (3) the capacity of the bloodstream to transport expertise from a wide range of laboratories, oxygen bound to haemoglobin (indexed by the they determined (from the end to the begin- red blood cells in circulation), and (4) the capac- ning of the chain, see Fig. 23 ): (1) the functional ity of the lungs to capture oxygen from inhaled capacity of muscle mitochondria to generate air (indexed by the oxygen diffusing capacity) ATP (indexed by mitochondrial volume; there ( Weibel et al. 1991 ). Covering the great range of was no evidence for species differences in mito- body masses, the mammal data showed that chondrial characteristics), (2) the capacity of the slopes of the allometric regressions of mito- the blood circulation to deliver oxygen to the chondrial and capillary volumes were similar muscle cells (indexed by capillary volume), to the slope of maximal oxygen consumption SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 39 on body mass ( Fig. 25 ). These two elements of 1000 /s) 2 the respiratory chain apparently were in fi ne whole animal balance with maximal performance on the 100 slope = 0.86 treadmills. However, the slope of 1.08 of pul- 10 monary oxygen diffusing capacity on body mass was signifi cantly higher than the slope of 1

0.86 for maximal oxygen consumption on body 0.1 mass, suggesting overcapacity of the lungs, especially for larger mammals. 10 /mmHg/s) maximal oxygen consumption (ml O maximal oxygen These fi ndings were confi rmed in a second 2 type of what Weibel and colleagues have called 1 slope = 1.08 an adaptive rather than an allometric compari- 0.1 lungs son. Keeping body mass approximately con- 10,000 stant, they went on to compare the same 0.01 1000 variables of maximal aerobic performance and 0.001 respiratory chain elements between athletic slope = 0.89 100 oxygen diffusing capacity (ml O oxygen bodies and less athletic forms. On the premise muscle capillaries that athletic types would show higher maximal 10 performance levels than non-athletic types capilary (ml) volume 1 (which they did, see the data points for the East 10,000 African mammalian athletes in Fig. 25 ), one 1000 would again expect the capacity of elements of slope = 0.87 the respiratory chain to match maximal oxygen 100 mitochondria performance. Comparing dog (athletic) with goat (non-athletic), human athlete with couch 10 potato, pony with calf, and horse with steer (ml) mitochondrial volume 1

( Table 2 ), the Weibel–Taylor team was able to 0.010.1 1 10 100 1000 confi rm the fi ndings from the allometric com- body mass (kg) parison. Mitochondrial, capillary, and circulat- Figure 25. Double-logarithmic plots of the critical peaking red blood cell volumes were close matches performance measure (maximal oxygen consumption rate) and with maximal oxygen-consumption rates, but various elements, both structural and functional, in the respiratory the dog, the athlete, the pony, and the horse chain (see Fig. 23 ) against body mass of mammals that vary in had smaller oxygen-diffusing capacities rela- size by four orders of magnitude. Particularly ‘athletic’ species are indicated by the fi lled symbols. The smallest species was the tive to their maximal oxygen-consumption mouse Mus musculus, the second smallest the dwarf mongoose rates than their non-athletic counterparts, so Melogale pervula, and the two largest, the horse Equus ferus that the ratios were signifi cantly smaller than 1 caballus and the steer Bos primigenius . (Table 2 ). Assuming that human athletes start Compiled from Weibel ( 2000 ). off at the same level as their sedentary counterparts, the athletes must have upregulated all measured parts of the respiratory chain except prepared for lower than normal partial oxygen the diffusing capacity of the lungs. pressures, such as those found at high altitudes Why should lungs, in humans and other and in underground burrows. In that case, mammals, have excess capacity? Weibel ( 2000 ) excess lung capacity would give mammals a presents two possible explanations. First, he greater freedom of choice with respect to habi- considers it possible that lungs in general are tat use. This would be important as, and here 40 BASICS OF ORGANISMAL DESIGN

Table 2. A test of symmorphosis in the respiratory chain of mammals based on comparisons between the ratios of several down-chain respiratory variables (either morphological or physiological, see Fig. 23 ) and maximal oxygen consumption of four pairs of athletic versus non-athletic forms of similar body mass. A ratio of 1 is expected if the down-chain respiratory variables match maximal oxygen consumption in similar ways in athletic and non-athletic forms. Underlined ratios differ signifi cantly from 1. Simplifi ed from Weibel et al . ( 1991 ) and Weibel ( 2000 ).

Athletic/Non-athletic Mitochondria Blood circulation Heart Lungs form Mitochondrial Capillary volume/ Red blood cells Oxygen-diffusing volume/maximal maximal oxygen in circulation/ capacity/maximal oxygen consumption maximal oxygen oxygen consumption consumption in circulation

25–30 kg dog/goat 1.20 1.26 1.08 0.61 70–80 kg athlete/sedentary 1.03 0.77 0.96 0.67 person 150 kg pony/calf 0.90 0.89 0.79 0.65 450 kg horse/steer 1.00 0.82 1.00 0.76

is the second explanation, lungs have ‘very good news for human patients who lose part limited malleability of structure’. Unlike the of a lung, or even a whole lung. From this per- heart and the mitochondria that can increase spective, the argument of supposed wasteful- their volume with training, to generate greater ness is turned on its head. Perhaps, in some oxygen-diffusing capacity, alveolar surface cases, excess capacities are not just wasteful, areas can no longer increase in mature mam- but safe ( Mauroy et al . 2004 ). mals. Although in humans a modest increase A celebrated example of safety factors is of lung performance is possible with training their relevance to the construction of elevator ( Hochachka and Beatty 2003 ), attempts to train cables, especially in fast passenger lifts. Jared guinea pigs Cavia porcellus from the time of Diamond ( 1993 ) begins a wonderfully written weaning were not successful in increasing review of safety factors with a very frightening lung-diffusing capacity (Hoppeler et al . 1995 ). account of what it would be like to fi nd your- One could say that lungs are constructed in a self in a crowded skyskraper elevator when somewhat wasteful (and permanent) way so the cables break. ‘Fortunately, it’s rare in prac- that their capacity is on the safe side. tice that elevator cables snap. The reason is that they’re designed by engineers to support a strain up to 11 times the maximum strain Safety factors that they would face when lifting their advertised maximum load at their designed speed.’ That the lungs of large mammals have approx- However, Diamond continues, ‘you may not imately double the capacity than required, wish to learn . . . that safety factors of steel even at high oxygen-consumption rates, is bridges and buildings are only about 2’.2 Safety

2 At the very time this part of the text was formulated, a bridge across the Mississippi, part of interstate 35W connecting Minneapolis with the sister city of St. Paul, collapsed during the evening rush hour, leaving many people dead and injured. Such accidents are not unique and led Gordon (1978 ) to make the comment that ‘engineering amounts to applied theology’. SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 41

factors, then, are estimates of the magnitude of broken, teeth could not be replaced in a pre- the load on a device that just causes failure, dentist world. divided by the maximum load the same device A capacity for compensation when things go is expected to experience under normal condi- wrong may explain why paired organs, such as tions (Vogel 2003 ). Safety factors of fast pas- lungs or kidneys, usually have smaller safety senger lifts are among the highest reported for factors than unpaired organs, such as the pan- man-engineered structures ( Fig. 26 ). Human creas ( Fig. 26 ). It provides the logic of why a teeth, with a safety factor of 15, are the cham- healthy person can afford to donate a kidney to pions among biological structures. This can a grateful transplant patient, and it can also only be regarded as a good thing, as, once explain why pancreatic cancer is usually only

safety factor 15 human teeth 14

12 cable of fast passenger elevator 1 11

10 human pancreas 5

9 pigeon feather shaft 8 8 cable of slow passenger elevator 14 7 cat intestinal brush-border frequency of fatigue fracture cable of slow freight elevator arginine transporter wooden building 6 wing bones of a flying goose 5 cable of powered dumb waiter leg bones of a galloping horse 4 tree trunks human kidneys wing bones of a fruit bat 3 leg bones of a hopping kangaroo mouse intestinal brush-border sucrase leg bones of a running ostrich steel building or bridge 2 lungs of a cow dragline of a spider 1 shell of a squid ENGINEERED STRUCTURES0 BIOLOGICAL STRUCTURES Figure 26. A review of ‘safety factors’ in structures built by humans and in a great variety of biological structures. Despite an average safety factor of 4.8 for bending loads, leg bones of galloping horses do sometimes break from material fatigue, and the incidence of such fractures increases towards the periphery. Safety factors are from Diamond ( 1998 ) and Vogel ( 2003 ); the diagram with the leg-bone fractures is from Alexander (1998 ) compiled from data in Currey (1984 ). Note that safety factors, such as those presented here, can also refl ect a close match of system design to unrecognized and unscrutinized performance requirements (Salvador and Savageau 2003 ). For those of you who are curious about ‘dumb waiters’, these are the miniature lifts by which hotels transport room-service meals from the kitchen in the basement to the rooms at higher fl oors ( Diamond 2002 ). 42 BASICS OF ORGANISMAL DESIGN

Table 3. Nine considerations that would either increase (+) used by birds to build their feathers ( Bonser or decrease (−) the setting of safety factors in design 1996 , Weber et al . 2005 ). While wear and tear on processes. Modifi ed from Diamond ( 2002 ). feathers may well explain their relatively high Consideration Increase or decrease safety factor, it certainly explains their regular of setpoint value replacement during moult. Coeffi cient of variation of load + At the low end of safety factors higher than Coeffi cient of variation of + one (remember that 1 indicates symmorphosis capacity as originally understood) are structures such as Deterioration of capacity with + the spider’s dragline. Draglines just have to time carry the spider’s own weight, and some of its Cost of failure + prey. In cases when local failure actually gives a Cost of initial construction – fi tness advantage, safety factors can thus better Cost of maintenance – be lower than 1 and do not quite deserve the term Cost of operation – any more ( Vogel 2003 ). In the summer of 1999, Opportunity cost of occupied – during an icebreaker-based expedition organ- space ized by the Swedish Polar Secretariat, we vis- Cost of rebuilding – ited a series of tundra sites across the Canadian Arctic. We were interested in the spatial varia- detected when pancreatic enzyme outputs tion of predation pressure by shorebirds on have dropped to only 10% of normal peak val- insects. Research partner Jens Rydell then pro- ues (DiMagno et al . 1973, Diamond 2002 ). That posed to count, as a measure of predation pres- the wing bones of birds have a safety factor of sure, the number of legs remaining on mature approximately 6, whereas their feather shafts cranefl ies (Tipulidae; they start off with six). have a safety factor of 9 (Fig. 26 ), probably has With too few cranefl ies captured per site, the to do with the higher variability of the over- plan did not quite work, but the story makes load on structures, such as fl ight feathers, when the point that giving up a leg grabbed by a bird compared with wing bones ( Corning and can be a good thing for a cranefl y, if it prevents Biewener 1998 ). It is not diffi cult to imagine it from being eaten. There are other examples that the wing bones of fast-fl ying animals, such where places of weak construction in a struc- as birds and bats, face greater overloads (e.g. ture pay off: self-amputations occur in spiders when hitting or avoiding obstacles in mid-air) (eight legs to lose) and lizards (one tail to spare, than the leg bones of hopping kangaroos, but the tail will grow back). Nicely obeying the which in turn may face greater stresses than evolutionary engineering principles advocated the leg bones of ostriches running on fl at in this book, lacertid lizards Podarcis spp. that ground. These examples clearly suggest some live in places with high predation pressure, of the general engineering considerations sum- release their tails more readily when receiving marized in Table 3 : with an increase in the coef- the proper prompting than similar lizards that fi cients of variation of load or capacity, and have less to fear (Cooper et al . 2004 ). Birds adap- with an increase in the costs of failure (high for tively drop their tail feathers when pursued or fast elevators, lower for slow elevators and grabbed by raptors for exactly the same reason negligible for dumb waiters), safety factors ( Lindström and Nilsson 1988 , de Nie 2002 , should go up. This is also true if the capacity of Møller et al . 2006 ). Trees may survive a storm if structures to withstand overloads deteriorates some weaker branches snap and streamline the with time; an example would be the keratin remaining structure ( Heinrich 1997 ). At this SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 43 point you may want to go to the garden shed to metabolic rate of lungs is smaller than the met- verify whether those long-legged spiders, the abolic rate of the skeletal musculature, the harvestmen (order Opiliones), do indeed move oxygen-diffusing surface of the lungs is the around well with only 5, 6 or 7 legs! relatively cheap link in the respiratory chain. For spiders, the combined costs of losing a Therefore, he predicted, lungs would be leg and rebuilding it, or losing a silky dragline designed for higher capacities than the rela- and renewing it, are relatively cheap. As listed tively expensive mitochondria! in Table 3 , with increasing costs of construction, maintenance, and operation, safety factors should go down. Cost considerations may Multiple design criteria explain why, despite safety factors close to 5, race horses nevertheless break leg bones, and, Since their grand summary in 1991 of tests of particularly, why the frequencies of fatigue the symmorphosis principle, by detailed stud- fractures during or prior to falling increase in ies on serial capacities in the respiratory chains the bones at the extremities ( Fig. 26 ); these of mammals, and in the face of serious con- bones are made of the same material as those structive criticism (e.g. Garland and Huey closer to the core of the horse. The costs of 1987 , Lindstedt and Jones 1987 , Dudley and transport during running (which, as we have Gans 1991 , Garland 1998 , Gordon 1998 , seen, Hoyt and Taylor’s horse wanted to mini- Bacigalupe and Bozinovic 2002 ), Weibel and mize, Fig. 22 ), will go up a great deal if the associates have regularly updated their analy- lower leg bones are given extra strength and ses (Weibel et al . 1992 , 2004 , Weibel and weight (Alexander 1984 , 1998 ). After all, these Hoppeler 2005 ). They clearly stand by their parts have to be powered with the greatest conclusion that respiratory chain elements momentum. Incidentally, the lack of weight in such as the vascular supply network and the the extremities explains why long- and thin- energy needs of the cells active during maxi- legged birds, like shorebirds, have relatively mal work are quantitatively coupled to low costs when they run (Bruinzeel et al . 1999 ). maximal metabolic rates; these elements thus This brings us back to previous considera- have safety factors close to 1. Although sym- tions about weaker than average links in serial morphosis is no longer a hotly debated issue constructions such as the respiratory chain, or ( Suarez and Darveau 2005 ), the previous dis- indeed, the legs of a horse. As explained by cussion illustrates that the principle of elegant Alexander ( 1997 , 1998 ), if elements in a chain economic design needs qualifi cation. differ in the coeffi cients of variation with Let us get back to lungs. Although we have respect to load, capacity, and/or cost (Table 3 ), not acknowledged this yet, they do more than the performance of the system is optimized by catch oxygen for the circulation of blood and devoting resources to extra capacity for the eventual delivery to the mitochondrial power- more variable or the cheaper element. Quite houses. Lungs also service the body by remov- literally, in a chain made partly of expensive ing carbon dioxide, one of the products of gold links and partly of cheap iron links, mak- metabolism, from the bloodstream. The remo- ing the gold links very slightly weaker and the val of carbon dioxide and the regulation of iron links much stronger would result in a blood pH (a delicate function of carbon diox- stronger chain for the same price. This is some- ide level in the body circulation) may put dif- what counter-intuitive, but means, as poin- ferent design constraints on the lungs ted out by Alexander ( 1998 ), that because the ( Lindstedt and Jones 1987 , Dudley and Gans 44 BASICS OF ORGANISMAL DESIGN

Figure 27. The behavioural and morphological trick of the peacock, a butterfl y that goes under the Latin nameInachis io, to combine multiple design criteria. For camoufl age they keep the wings folded and only show their drab undersides; for advertisement they show the full glory of their colourful upper wings with eyespots. Photos by Harm Alberts (from www.waarneming.nl ).

1991 ). Water loss via respiration and the asso- leons can show rapid adaptive colour change, ciated maintenance of heat balance may also e.g. Stuart-Fox et al. 2008 ). Accordingly, some impose demands on the respiratory system. of the critics argue that the best we can Dudley and Gans (1991 ) point out that this expect from the form and function of most may be especially true for the ungulates of the natural structures is that they represent ‘con- tropical open savannahs that were so promi- tinua of imperfection’ (Dudley and Gans nent in the empirical verifi cation of the sym- 1991 ). morphosis principle ( Weibel et al . 1991 ). In some cases it is possible to do justice to multiple natural selection criteria in a single One more problem: the climbing design ( Fig. 27 ). The shape, surface area, and of adaptive peaks strength of butterfl y wings are likely to be adaptations primarily to the aerodynamic A fi nal criticism of the symmorphosis princi- demands placed on them. The colours of wings ple, related to the previous ones, is that natural are subject to a different selection regime. selection will not necessarily lead to optimal Somehow, butterfl ies have to fi nd a balance design ( Gould and Lewontin 1979 , reiterated between the opposing forces of crypsis and by Lindstedt and Jones 1987 , Dudley and Gans sexual advertisement, and also take into 1991 ) because selection pressures will change account thermoregulation (Cott 1957 , Kingsolver with time, i.e. that we are studying ‘moving 1988 , Dudley and Gans 1991 , Houston et al . targets’ ( Gordon 1998 ). For these reasons, the 2007 ). The peacock, and many other butter- structure and function of organisms, and the fl ies, have resolved this engineering confl ict by parts thereof, would have diffi culty in climb- putting their colourful sexual advertisement ing any adaptive peak ( Dawkins 1996 ). This on the upper wings, and the cryptic coloration may all be true, but such criticisms give us no on the under wings ( Fig. 27 ). Simply by clos- reason to abolish the helpful notion of sym- ing its wings, a peacock turns from a strikingly morphism. Quite the opposite: the various visible object into a cryptic form that is hard to caveats mentioned suggest additional layers fi nd. Behavioural solutions like this may be of adaptation and make our lives as biologists rare (but note that cuttlefi sh, fi sh, and chame- so much more interesting! SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 45

For an example, let us get back to race horses, to the appropriately named ‘thoroughbreds’. 200 St. Leger (2937 m)

For thousands of years, and especially over the 190 last few centuries, these horses have been bred for racing performance. Selective breeding 180 confers a consistent and intense directional 170 selection pressure over many generations

winning time (sec) Oaks (2423 m ) Derby (2423 m ) ( Gaffney and Cunningham 1988 , Bailey 1998 ). 160 Despite the best efforts of horse breeders in A over 40 countries that have more than half a 150 million thoroughbreds to work with, racing 1850 1870 1890 1910 1930 1950 1970 1990 performance, the type of hard muscle work 250 that is highly correlated with peak aerobic power ( Jones 1998a ), has levelled off over the 240 last century ( Fig. 28 ). Yet Jones ( 1998a, 1998b ) 230 reports that selective breeding has indeed time (sec) pushed up the performance of various ele- 220 ments of the equine respiratory system. Most B interestingly, even the oxygen-diffusing power 210 of the lungs of these thoroughbreds, an ele- 1900 1920 1940 1960 1980 2000 ment that earlier was reported to have excess 42 EPO capacity, has reached symmorphosis with the rest of the respiratory chain. Artifi cial selection 38 has done the job it set out to do. 34 "epidemic" At this point sports enthusiasts will think of of cycling deaths similar patterns of levelling-off in athletic per- 30 formances by humans. Rates of decrease in the men’s world record running times began to average speed (km/h) 26 slow down after about 1960 (Nevill and Whyte 22 C 2005 , Noakes 2006 , Denny 2008 ). The women’s 1900 1920 1940 1960 1980 2000 world record progression showed dramatic improvements between 1965 and 1980, but has Figure 28. Limits to running performance improvements also levelled off over the past 25 years. These of race horses (A) and human athletes (B), and cycling speeds of human athletes during the Tour de France (C). (A) data suggest that runners today are at the lim- Changes in the winning times over the 2400-m long Derby its of achievable biological capacity (Denny and the 3-km St. Leger races in the UK (from Jones 1998a ); 2008 ). This, of course, is not because human (B) long-term changes in the record times for the Olympic runners as a group have been selectively bred 1500 m for men (from Vogel 2001 ); and (C) average racing in the way that thoroughbreds have been. speeds in the Tour de France. Note that the drug erythropoietin (EPO) was first introduced in the late-1980s Training schemes have improved. Also, cham- and may have increased the mortality level since then (from pion athletes originate from an ever wider Noakes 2007 ). range of human populations. The élite marathon runners of today usually come from high- altitude populations, especially in East Africa, 46 BASICS OF ORGANISMAL DESIGN

people that are adapted to live under relatively are interesting for their own reasons, and it is low partial oxygen pressures and may show worth reminding ourselves that when it comes 50% greater oxygen-diffusion capacity than to whole body performance, the oxygen and lowlanders ( Hochachka et al . 2002 ; inciden- fuel transport chains have to complement each tally, athletes have also become relatively taller other. and more slender, Charles and Bejan 2009 ). To the best of our knowledge, nobody has This hints at something we have just learned: carried out formal analyses of symmorphy that directional selection on the performance across all the successive elements in the food of the respiratory chain increases aerobic per- processing chain (Fig. 29 ). Is the level at which formance. In any case, now that the limits to performance have almost been reached in the world of athletes, future improvements in world records will require the use of more effective performance-enhancing drugs or the introduction of gene doping. This may already have happened in the Tour de France, the classic cycling tournament ( Noakes 2007 ). A plateau level that had been maintained for three decades was exceeded when the new drug erythropoietin (EPO, a red blood cell booster) was fi rst introduced in the late 1980s ( Fig. 28C ). In biology, ‘metabolic ceilings’ really may exist. We will develop the subject in the next chapter.

In addition to oxygen, fi res need fuel too As Max Kleiber (1961 ) implied with the title of his animal energetics classic The Fire of Life , oxygen needs a substrate to burn. So far, the discussion has focused on the respiratory chain: how animals make sure that there is enough oxygen for the mitochondria to oxidize fuels to generate ATP and thus power body work. Unlike oxygen, fuels can be stored in the body, and this complicates the one-to-one relationship between supply and demand, a simplifi cation that made Figure 29. The ‘barrel model’ of an organism’s energy balance. the respiratory chain such an attractive system Input constraints (maximum rates of foraging, digestion, and to study ( Taylor et al . 1996 , Weibel et al . 1996 ). absorption) are engaged in series; different outputs (heat production, mechanical work, tissue growth) are parallel and independently Just like the respiratory chain, the food-to-fuel- controlled. If the sum of output rates does not match the input, the pathway (which includes ingestion, digestion, balance is buffered by the storage capacity of the system. Note that absorption, and assimilation), exhibits serial the fi rst spigot always ‘leaks’ (basal heat loss of the organism). organization ( Fig. 29 ). Food-processing chains From Weiner (1992 ). SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 47 fuel becomes available to the body tissues lim- nutritional components (sugars, carbohydrates, ited by rates of prey capture, maximum food- lipids, etc.) are passively or actively absorbed processing, by rates of digestion in the stomach in a variety of ways. They had to develop dif- and intestine, by rates of absorption in the ferent assays for different nutrients. intestine or by rates of assimilation and bio- The evidence accumulated by the Diamond chemical adjustment in the liver? Clearly, an group is consistent with ‘reasonably designed’ animal built with small safety factors for maxi- nutrient-processing systems. With manipu- mum food-intake rates would have to live in lated variations in food-intake rates, the exper- very predictable worlds with very predictable imental mice showed parallel variations in the energy expenditures. In practice, instantane- size of the main digestive organ, the small ous food-intake rates are usually at least one intestine ( Fig. 30A ). Diamond ( 2002 ) suggests order of magnitude greater than long-term that, with increasing food-intake rates, the food-intake rates (e.g. Piersma et al . 1995b , van excess capacity of the intestine gradually Gils and Piersma 2004 , Jeschke 2007 ), so 10 may diminishes because ‘a mouse cannot increase serve as a preliminary minimum estimate of its intestinal mass indefi nitely’. This becomes the safety factor for food-intake rates. Skipping clear when the measured index of absorption the second funnel in Weiner’s diagram ( Fig. 29 ) rates of one of the nutrients, glucose, is plotted for lack of data on digestive rates, we come to against glucose intake rates (Fig. 30B ). With an the third, the one symbolizing absorption rates. increase in nutritional demands because of In this case, we can count ourselves lucky, as in lactation in the cold, the safety factor, that the 1990s a whole body of work was carried under relaxed conditions is close to 3 (see out by Jared Diamond and his research group Fig. 26 ), levels off at 1. on rates of intestinal assimilation in house mice As we shall see in the next chapter, the ques- ( Toloza et al . 1991 , Diamond and Hammond tion whether the respiratory chain and the 1992 , Hammond and Diamond 1992 , Hammond food-processing chain show any degree of et al. 1994 , 1996a , Weiss et al . 1998 , O’Connor symmorphy cannot be separated from ques- and Diamond 1999 , O’Connor et al . 1999 , Lam tions about the biological engineering con- et al. 2002 ; summary in Diamond 2002 ). Their straints to maximal and long-term aerobic approach differed from that to respiratory performance, on the design of ‘metabolic ceil- chains, as they did not look at maximal per- ings’, so to speak. Before we turn our attention formance over a range of species. Instead, they to this question, let us briefl y examine sym- examined capacities within a single species morphosis in the food-processing systems of manipulated to function at different levels of our prime witnesses in this book: the shore- energy expenditure. In small (female) mam- birds and, especially, red knots. mals one can do this with ambient temperature and the presence or absence of lactation for different numbers of pups. This is a valid approach Testing symmorphosis in shorebird for the food-processing chain because, unlike food-processing systems oxygen that cannot escape once it has entered the bloodstream, nutrients do have ‘escape Diamond ( 2002 ) made a plot of small-intestine pathways’ after being absorbed by the intes- mass against food-intake rate in mice experienc- tine. They do not have to be burned at once, but ing different levels of energy expenditure ( Fig. 30 ), can be stored. However, Diamond and his asso- in order to test whether indices of capacities of ciates faced the complication that different the foraging- and the absorption-funnels ( Fig. 29 ) 48 BASICS OF ORGANISMAL DESIGN

5 A predicted on the basis of the symmorphosis principle, in an allometric comparison between 4 the organ sizes, intestine and liver mass correlate closely with gizzard mass (Fig. 31A , B). 3 Because of the need to crush shells and process shell fragments, the gizzards and intestines of 2 the fi ve hard-shelled prey specialists in the graph (red knot, great knot, surfbird, purple small intestine mass (g) 1 sandpiper, and rock sandpiper), have high

0 workloads relative to other organs in the diges- 0 5210 15 20 530tive system ( Battley and Piersma 2005 ). This –1 food intake (g d ) may explain why several of these mollusc specialists have small livers relative to gizzard 3 B and intestine ( Fig. 31B ). The intraspecifi c comparison for red knots (Fig. 31C , D) is even more interesting. In the 2 plots of intestine and liver mass against gizzard mass, the data seem to fall into two safety factor safety 1 groups. As expected, intestine mass ( Battley and Piersma 2005 ), shows a positive correlation with gizzard mass. The slope of this rela-

0 tionship is much steeper during the spring 0 10 20 30 40 50 stopover in Iceland than in winter ( Fig. 31C ). glucose intake (mmole d–1) Liver mass does not correlate with gizzard Figure 30. (A) Mass of the small intestine of female mice mass at all in winter, but does so during refuel- increases with increasing food intake during lactation and at ling at the spring stopover site ( Fig. 31D ). This low ambient temperatures. Virgin mice at temperatures of 23 °C and 5 °C are indicated by fi lled circles and triangles, is an exciting fi nding, as red knots have been respectively; lactating mice at 23 °C and 5 °C are indicated by shown to select spring stopover sites on the open squares and triangles, respectively. (B) The safety factor basis of high food-quality, i.e. places with prey for the small intestinal brush border glucose transporter SGLT1 that have a lot of meat and little shell material of female mice goes down with glucose intake rates. Data are ( van Gils et al . 2005c , 2006a ). In winter, gizzard from the same experiments as in (A) plus experiments involving size increases are explained to a large extent by increased pup mass. Virgin mice at 23 °C and 5 °C are indicated by fi lled circles and squares, respectively; lactating mice are the decrease in the ratios of meat to shell mass indicated with open symbols, for 23 °C (downpointing triangle, ( van Gils et al . 2003a , 2006b ). diamond) and 5 °C (squares) environments, and at 23 °C with We suggest that two mechanisms infl uence increased pup mass (upward pointing triangle). the degree of symmorphy within the food- Data are from Hammond and Diamond ( 1992 , 1994 ), and from processing chain of red knots and other shore- Hammond et al. ( 1994 ); the compilations are from Diamond ( 2002 ). birds: (1) the need for shell processing (scaled to body mass and size-related energy requirements, see Box 2 ), determines the size of both do correlate. We have plotted intestine mass gizzard and intestine, but not that of the liver, (third funnel) and liver mass (an assimilation (2) the quantity of food processed determines funnel not illustrated in Fig. 29 ) against gizzard the size of the intestinal tract (to facilitate mass in a range of shorebird species ( Fig. 31 ). As absorption and/or to adjust to proper levels of SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 49

100 AC24

16 10 12

fresh intestine mass (g) 4 1 0 B 16 D

12 10 8

fresh liver mass (g) fresh liver 4 1

0 1 10 6810 12 14 fresh gizzard mass (g) fresh gizzard mass (g) Figure 31. Examination of the symmorphosis design principle predictions for the digestive system of shorebirds, based on comparisons between the size of the gizzard as the fi rst element in the digestive chain, and intestine and liver as subsequent elements, both in an interspecifi c comparison between shorebird species (A, B) eating either hard-shelled (open circles) or other prey (fi lled circles) and in an intraspecifi c comparison for red knots (C, D) that either fi nd themselves in a refuelling situation during northward migration in Iceland (open circles) or in Western Europe in winter (fi lled circles). Linear regression lines are shown where slopes are signifi cantly different from 0. In A and B, the fi ve hard-shelled prey specialists, from the lightest to the heaviest, are rock sandpiper Calidris ptilocnemus , purple sandpiper Calidris maritima , red knot, surfbird Aphriza virgata , and great knot Calidris tenuirostris. Although technically we have worked with stomach masses, in shorebirds the proventriculus (glandular stomach) is so small ( Piersma et al . 1993b) that we are justifi ed in talking about gizzards (muscular stomachs) here. Based on data presented in Piersma et al . ( 1999b ), Battley and Piersma ( 2005 ), and T. Piersma (unpubl. data).

immune defence; Piersma et al . 1999b , Baker the respiratory chain, is now widely accepted et al . 2004 ). Greater assimilation rates of nutri- as a useful, heuristic design principle. As we ents by the liver (e.g. during stopovers) induce will see later in the book, like other criteria increases of liver capacity. Do we really see used in optimality-driven evaluations of symmorphosis here? We believe we do, but as organismal performance, symmorphosis is usual it is symmorphosis with a qualifi cation. better seen as a useful null hypothesis, than as a hypothesis with very precise and rigid crite- Synopsis ria for rejection. The discussion of safety factors has demonstrated how an evaluation of Symmorphosis, the principle that evolved cases where simple, economy-based expecta- body designs avoid excess capacity, e.g. in cas- tions are not upheld, develops our biological cades of serial physiological processes, such as insight. 50 BASICS OF ORGANISMAL DESIGN

Box 2 Introduction to allometry: assessing relative changes in body parts and whole organisms and their performance levels

Arithmetic offers neat tricks, and one of the A b = 1.0 neatest and most widely applied set of tricks 10000 in comparative biology is the use of bmax = 0.92 1000 logarithms and exponents (‘power b = 0.67 relationships’) in so-called ‘allometric scaling’ 100 b = 0.82 exercises ( Dial et al . 2008 ). With the use of min logarithms, it is generally possible to unbend 10 curvilinear relationships, to transform them into linear ones of the kind suitable for maximal metabolic rate 1 standard statistical examinations (regression, covariance analyses, etc.). When an organ or B b = 1.0 a performance measure, such as oxygen- 1000 consumption rate, does not change in bmax = 0.79 100 constant proportion to its body mass, we say b = 0.67 that it scales allometrically (allo = different, 10 b = 0.76 metric = measure). Isometric scaling (iso = min same) represents the special case of 1

geometrically similar objects that show basal metabolic rate constant proportional changes with size: at 0.1 constant shape, surface area will simply scale to length with an exponent of 2, and volume 1 10 100 1000 body mass will scale to length with an exponent of 3 (the fi tted regression line will go through the Figure 32. Estimates of b, the scaling exponent, for the origin). As the ground rules of proper allometric relationships of maximum metabolic rate (A) and arithmetic conduct for the use of logarithms basal metabolic rates (B) on body mass in mammals using the and exponents are presented in several recent ‘allometric cascade model’ (an algebraic model based on Wagner ecophysiological textbooks (see, for example, ( 1993 ) that integrates estimates of all possible rate-limiting steps determining metabolic rate) implemented by Darveau et al . Willmer et al . 2000 , p. 43, and Karasov and (2002 ) and Hochachka et al. (2003 ). The range of achieved Martínez del Rio 2007 , p. 14), there is no scaling exponents for maximum metabolic rate and basal need to repeat them here. These books, and metabolic rate shown here represent the outcomes of sensitivity all-time or recently published ‘classics’ like analyses for reasonable range of estimates in up to 10 potential Peters ( 1983 ), Calder ( 1984 ), Schmidt-Nielsen metabolic rate-limiting steps. For reference, slopes of 1.0 and ( 1984) and Bonner ( 2006 ), can be consulted 0.67 are shown. Note that for simulation purposes, both for many biological applications. metabolic rate and body mass values are given in arbitrary units. Ever since Kleiber’s analysis of the Modifi ed from Darveauet al . ( 2002 ). relationship between metabolic rate, MR, and body mass, M (published fi rst in 1932, but Much paper has been printed with single- Kleiber 1961 is more widely available), cause explanations for b , which, depending on ecophysiologists and, at a later stage, one’s sample of data points, is often close to ecologists, have been intrigued by the power 0.75, but not always. Chapter 3 has made function (or allometric relationship): MR = a M b abundantly clear that metabolic or oxygen- (where a , the intercept, is the scaling constant consumption rates may refl ect many different and b , the slope, is the scaling exponent). potential rate-limiting steps. It is this realization SYMMORPHOSIS: PRINCIPLE AND LIMITATIONS OF ECONOMIC DESIGN 51 that caused Charles Darveau, his supervisor lower scaling exponent b for basal metabolic Peter Hochachka and their associates, to try rate compared with maximum metabolic rate and operationalize b not to refl ect a single ( Fig. 32 ), can be explained by different rate-limiting process, but to refl ect the whole rate-limiting steps being most important in cascade of key steps in the complex pathways the two exercise conditions. For basal of energy demand and energy supply ( Darveau metabolic rate, oxygen-delivery steps of the et al. 2002 , 2003 , Hochachka et al. 2003 , kind discussed in the symmorphosis literature Suarez et al. 2004 ; the immediate criticisms by and in this chapter (relatively low exponents), Banavar et al. 2003 , West et al. 2003 are also contribute almost nothing to the overall instructive to read and to get a feel for this scaling exponent b . In contrast, for maximum enormous literature). metabolic rates, b is controlled entirely by the The team of Hochachka and Darveau oxygen delivery steps (relatively high comes to the conclusion that the generally exponents). This page intentionally left blank Part II

Adding environment This page intentionally left blank CHAPTER 4 Metabolic ceilings: the ecology of physiological restraint

Captain Robert Falcon Scott, sled ories (or 4.2 million kJ) each (Fig. 33 ). The dogs, and limits to hard work South African sports physiologist Timothy Noakes, author of the runner’s bible Lore of 1 If only Robert Falcon Scott had followed up Running ( Noakes 1992 ), comes to the conclu- Ernest Shackleton’s suggestion to use sled sion that this was ‘one of the greatest human dogs to reach the South Pole in the cold performances of sustained physical endurance Antarctic summer of 1911–12, he might have of all time’ ( Noakes 2006 , 2007 ). beaten his Norwegian adversary Roald Scott’s team, and other teams travelling Amundsen to the pole. Amundsen did use across the Antarctic continent at the time dogs to pull sledges, sacrifi cing some to feed ( Noakes 2006 ), certainly spent far more energy the others as they went. More importantly, if in total than endurance runners and cyclists Scott had listened to Shackleton, and not during famous lower latitude sporting events insisted either on bringing back 60 kg of what ( Fig. 33 ). However, it was the long duration that turned out to be effectively useless ‘geological made the polar exertions so physiologically and specimens’ from the pole on the then almost psychologically demanding. Enduring several empty sledges, he and his companions might months of immense physical challenge is well have made it back to civilization rather beyond what most of us can imagine. Still, from than dying 300 km short of their base, in late a physiological point of view, it is perhaps fairer March 1912 ( Huntford 1985 , Solomon 2001 , to make comparisons of performance levels Fiennes 2003 , Noakes 2006 ). Of the fi ve-man over intervals of one day, such as the detailed party, Evans died in February and Oates in reconstructions of the daily energy balances of mid-March. By the time Scott, Wilson, and Scott, other explorers and endurance athletes, Bowers lost their lives, they had been man- reconstructed by Mike Stroud ( 1998 ). An hauling sledges for some 159 consecutive days, English medical doctor turned exercise physi- of which the last 60 were in exceptionally bit- ologist and adventurer ( Stroud 2004 ), Stroud ter cold, covering a total distance of 2500 km used himself and fellow expeditioner, Sir and expending an estimated 1 million kilocal- Ranulph Fiennes, as guinea pigs during sev-

1 It only takes a few small steps to get from Antarctic explorer Robert Falcon Scott to our own world. This is because his son, Sir Peter Scott (1909–89), an avid ornithologist, sportsman, painter and founder of the Wildfowl and Wetlands Trust, as well as the Worldwide Fund for Nature (and designer of its panda logo), was the mentor of Hugh Boyd, now an eminence grise in waterbird ecology. Hugh, with a career within the Canadian Wildlife Service, has been a great source of inspiration and friendship to us.

55 56 ADDING ENVIRONMENT

believes that starvation actually killed the men in Scott Polar Expedition Captain Scott’s party. In fact, Stroud and Fiennes 1911–1912 almost killed themselves during their attempt to Bunion Derbies 1928/1929 ski across the whole of Antarctica in 1992–93 (Stroud 2004 ). Before their eventual rescue, they RAAM cycling race 2005 had lost over 20 kg each. Although humans can Tour de France endure several days of heavy physical exercise 1989 with limited food rations (Guezennec et al. 1994 , Six Day Running Race Hoyt et al. 2006 ), by day 95 of the expedition, 1989 Stroud and Fiennes showed considerable reduc-

021 3 4 million tions in muscle power, reductions that were total energy expenditure per person (kJ) linked to marked decreases in cytoplasmic and mitochondrial enzyme activities ( Stroud et al . Figure 33. Reconstructed total energy expenditure levels averaged per individual participants for fi ve events that Noakes 1997 ). Still, Stroud (1998 ) claims that ‘the restric- ( 2006 ) considers to represent the limits of human endurance. tion on food intake was imposed by the self-con- From bottom to top: (1) the six day pedestrian running races in tained nature of the expedition rather than by the Madison Square Garden, New York; (2) the 21-stage Tour de ability to eat or absorb more food’. Two other France cycling race; (3) the non-stop cycling Race Across America studies on endurance exercise levels that app- (RAAM); (4) the ‘Bunion Derbies’ foot race over 4960 km from roach or exceed those of the arctic explorers, Los Angeles to New York in 1928 and in the reverse direction in 1929 (Berry 1990 ); and (5) the heroic performance of Captain dealing respectively with Tour de France cyclists Robert Falcon Scott’s team who man-hauled their sleds across (Westerterp et al. 1986 ) and Swedish cross-coun- the Antarctic for 159 days in 1911–12. The inset shows Scott’s try skiers ( Sjodin et al. 1994 ), show rather smaller team at the South Pole on 18 January 1912. Standing from the defi cits, or even complete energy balance. The left are Henry (‘Birdie’) Bowers, Captain Scott, and biologist Dr data point for the Swedish skiers (Table 4 ) sug- Edward Wilson. Sitting from the left are Lawrence Oates and Edgar Evans. None survived the return journey. gests that an adult human male who expends 30 Compiled from Noakes ( 2006 ). MJ per day must be close to his ‘metabolic ceiling’, i.e. the maximum energy output that can possibly be sustained by an organism that eats eral Arctic and Antarctic sledge-hauling trips, enough to stay in energy balance. and went on to publish the results of an impres- The energy expenditure of the sled dogs, sively wide range of physiological measure- which in 1911–12 successfully pulled Roald ments, often painfully obtained ( Stroud et al . Amundsen from the Antarctic rim to the South 1993 , 1997 ). Pole and back over a period of 97 days, was Stroud’s summary (Table 4 ) makes clear that reconstructed to be approximately 21 MJ per polar explorers who expended 25–30 MJ per day day ( Orr 1966 , Pugh 1972 , Campbell and always incurred substantial energy defi cits. They Donaldson 1981 ). Although this is a very high ate only 60–70% of their daily energy require- level, Alaskan sled dogs racing for 70 hours in ments and, not surprisingly, lost about a kilo bitter cold conditions expended twice that every four days. For this reason, Stroud ( 1987b ) much (47 MJ/day, Hinchcliff et al . 1997 ).2 A s

2 This expenditure comes with considerable costs, including the death of dogs during the races. The level of performance asked from these dogs, even when trained, leads to gastric ulcers and increased intestinal permeability ( Davis et al . 2005 , 2006 ), as well as a degree of muscle damage ( Hinchcliff et al . 2004 ). METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 57

Table 4. A summary of human studies (male subjects only) on energy intake and expenditure per individual during prolonged intense exercise, and the extent to which energy intake during these periods could cover expenditure. Expanded from Stroud ( 1998).

Event Duration Energy Energy Defi cit Reference (days) intake (MJ/ expenditure1 (MJ/day) day) (MJ/day)

Captain Scott’s polar party >150 ca.18 ca. 25 7 Stroud (1987b) South Pole expedition 70 21 25 4 Stroud (1987a) North Pole expedition 48 19 30 11 Stroud et al . (1993) Trans-Antarctic expedition fi rst 50 20 32 12 Stroudet al . (1997) last 45 22 21 -1 Tour de France 20 >25 34 <9 Westerterp et al . (1986) Military mountain training 11 13 21 8 Hoyt et al . (1991) Military arctic training 10 >11 18 <7 Jones et al . (1993) Military jungle training 7 17 20 3 Forbes-Ewan et al . (1989) Cross-country skiing 7 31 30 -1 Sjodin et al . (1994) Sahara-multi-marathon 7 15 27 12 Stroud (1998)

1 Most expenditure values are based on isotope-dilution studies; see Speakman ( 1997a ) for the complete introduction. However, the value for the 70-day South Pole expedition is based on an energy balance evaluation. All values for Scott’s party were estimated. In view of the uncertainties around each of the estimates, values presented are rounded to the nearest MJ. Human subjects tend to under-report their own food intake; where such effects were most obvious, reported energy intake estimates are minima.

pointed out by Noakes ( 2007 ), with a total writers would, or four to fi ve times their rate expenditure of energy that approached the of energy expenditure during sleep. This level level of Scott and his comrades (Fig. 33 ), happens to be close to the multiple expended Amundsen’s sled dogs might well have by animals highly motivated to work; birds achieved ‘the greatest animal “sporting” per- taking care of nestlings, for example, or small formance’ of all time. But sled dogs weigh only mammals nursing pups (Drent and Daan 1980, about half as much as humans, and are very Hammond and Diamond 1997 ). Ever since different animals anyway. How should we Krogh ( 1916 ) started measuring metabolic compare the athletic achievements of British rates of animals—that is, the metabolic rate of polar explorers and Norwegian sled dogs? homeotherms under standardized conditions: inactive, in the thermoneutral temperature zone, in the absence of digestion, growth, The need for a yardstick moult, ovulation or gestation—animal physi- While hard at work on the text of this book, we ologists have used this strictly defi ned level of have been deskbound all day. However tired energy expenditure, named ‘basal metabolic we feel at night, our body machinery has been rate’ or BMR (see Box 3 ), as the yardstick for running at low levels. Moving around in the comparisons between different activity levels house with some thinking and writing does within and between species ( King 1974 ). not take much more effort than sleeping; in Realizing that maximal efforts shown by energetic terms, it takes about 70% more effort hard-working animals may well refl ect a com- ( Black et al . 1996 ). The Arctic explorers, cyclists, promise between physiological capacity (max- and skiers that we have just talked about, imal physical output) and the ‘willingness’ to expended about three times as much as book- incur elevated mortality risks due to short and 58 ADDING ENVIRONMENT

long-term consequences of increased perform- The expenditure levels of Captain Scott, Mike ance, Drent and Daan (1980) introduced the Stroud, and Sir Ranulph Fiennes clearly do not concept of ‘maximal sustained working lev- count during the times they were crossing els’, the aerobic capacity for work of animals parts of the Antarctic land-ice on foot (they as it is limited by evolutionarily shaped phys- relied quite heavily on their endogenous nutri- iological constraints. They expressed it in ent reserves), but those of the Swedish cross- multiples of BMR (Fig. 34 ). Animals that would country skiers studied by Sjodin et al . ( 1994 ) choose3 to ignore this ‘physiological warning do, and supposedly that of the lumberjack in level’ would lose body condition and face pre- Fig. 34 does too. We now need data on human cipitous increases in mortality risk. A similar BMR to bring the values in Table 4 into per- evolutionary concept was earlier formulated spective. Here we hit a snag, as BMR varies by Royama ( 1966 ). He called it the ‘optimal widely as a function of size, sex, and age (e.g. working capacity’, defi ned as the energetic Henry 2005 , Froehle and Schoeninger 2006 , performance level of parents beyond which Wouters-Adriaans and Westerterp 2006 ), and they would suffer from increased risks due to it may even vary within individuals (at least physical fatigue, infection, and predation. somewhat, Johnstone et al . 2006 )—to wit the Rudi Drent’s and Serge Daan’s 1980 paper central theme of this book! Celebrating the ‘The prudent parent: energetic adjustments in 75th anniversary of Scott’s epic walk, during avian breeding’ not only set the stage for evo- their fi rst Antarctic mission in 1986–87, Stroud lutionary explanations of animal energetic and Fiennes had their BMR measured on the performance levels (Peterson et al . 1990 , Weiner way back, two weeks after they had reached 1992 , Hammond and Diamond 1997 ), but also the South Pole. Despite considerable body formulated a widely used organismal invest- mass losses, BMR was 60% higher than before ment model that makes the distinction between they started the ‘walk’ ( Stroud 1987b ). ‘capital’ spenders that rely on stored resources, In the absence of BMR measurements on the and ‘income’ spenders that rely on concurrent same subjects that yielded the empirical data intake of nutrients or energy (Tammaru and on rates of energy expenditure during pro- Haukioja 1996 , Jonsson 1997 , Bonnet et al . 1998 , longed intense exercise (Table 4 ), we must rely Meijer and Drent 1999 , Boyd 2000 ). Despite its on appropriate data published independently. modest outlet—the Dutch ornithological jour- On the basis of Henry’s (2005) compilation of nal Ardea—‘The prudent parent’ has become 10,552 human BMR values, for an athletic one of the most highly cited papers in Caucasian man, a BMR of 6–7 MJ/day would ecology. represent a fair estimate. On this basis, polar Paraphrasing Drent and Daan twice, one explorers would have lived at a level of ca. 4 could argue that ‘maximum sustained work- times BMR (Table 4 ), while the well-fed ath- ing levels’ are only achieved by animals that letes taking part in the Tour de France and are ‘income’ spenders, as they maintain energy cross-country skiing would have exceeded this balance over the period of peak performance. lumberjack-level ( Fig. 34 ) and probably came

3 The use of the words ‘choose’ and ‘willingness’ refl ects the typical ‘slang’ of behavioural ecology. What we mean to express here and later, when expressions such as ‘animals that would choose to do this rather than that’ are used, is to say that ‘animals that are built to do this rather than that, would have the highest fi tness’. The latter is a more evolutionarily respectful use of words, but it is much more elaborate. METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 59

Figure 34. Original version of the cartoon in which Rudi Drent and Serge Daan (1980) indicated that what they called ‘maximum sustained working level’ of parent birds tending nestlings, usually approached values close to four times basal metabolic rate, a level also achieved by humans doing heavy labour, such as lumberjacks (Brody 1945 ). Their bird data referred to long-eared owls Asio otis ( Wijnandts 1984 ), house martins Delichon urbica ( Bryant and Westerterp 1980 ), starlings Sturnus vulgaris , glaucous gulls Larus glaucescens , and turtle doves Streptopelia risoria ( Brisbin 1969 ). closer to 5 times BMR. A small proportion of mass. Kirkwood ( 1983 ) was the fi rst, but seems estimates of the ratio between sustained meta- to have done this rather hastily; his is an odd bolic rates and BMR of birds and mammals are and incomplete collection. He assembled 21 higher than 5 ( Bryant and Tatner 1991 , miscellaneous estimates for metabolizable Hammond and Diamond 1997 , Speakman energy intake in domestic and non-domestic 2000 ). A ‘sustained metabolic scope’ (hence- mammals and birds kept under ‘intensely forth called factorial scope) of 5 may, therefore, demanding conditions’. He found the overall represent a reasonable fi rst guess at a maximal exponent of 0.72 to be close to published expo- physiological working level of seriously chal- nents for BMR in mammals and birds (i.e. the lenged animals that still maintain energy lines for BMR and maximum metabolizable balance. energy intake run parallel), the level to be on It was just a matter of time before someone average 3.6 times BMR, and concluded that fi rst recognized the usefulness of allometry 1700 kJ/kg0.72 per day ‘may prove to be a valu- (see Box 2 ) for scaling this level of energy able yardstick with which to compare intakes expenditure by making double-logarithmic of various species under natural and experi- plots of sustained metabolic rates on body mental conditions’. Applied to our human 60 ADDING ENVIRONMENT

athletes, his equation would predict a 70-kg 14 human to be able to expend 36 MJ/day, and to

eat and digest and assimilate enough food to 12 remain in energy balance. Only cyclists in the Tour de France come close, but incur an intake 10 defi cit (Table 4 ). Nevertheless, Kirkwood’s birdflight in windtunnel preliminary allometric yardstick has been transoceanic flight by shorebirds 8 maximal metabolic rate (exercise- induced) widely used. In the meantime, his yardstick has been refi ned for specifi c groups of animals summit metabolic rate (cold-induced) 6 ( Lindström and Kvist 1995 ). sustained metabolic rate

maximum sustained working level 4 Peaks and plateaus: what is true energy expenditure (multiples of BMR) (multiples energy expenditure endurance? 2 basal metabolic rate In 1990, Charles Peterson, Ken Nagy, and Jared 0 Diamond drew the fi rst representation of the minutes hours days weeks empirical fi nding that the longer an activity is time intervals over which levels are measured maintained, the lower the possible power out- or relevant put can be, with a decline towards some Figure 35. Unlike basal metabolic rate, the power output of asymptotic value that is only a few times BMR exercising animals is a negative function of the duration for ( Fig. 35 ; Peterson et al . 1990 ). Relying on anaer- which the particular activity level is maintained. Inspired by obic metabolism, humans and other verte- Peterson et al . ( 1990 ), the energy expenditure levels in this fi gure were compiled from Drent and Daan (1980), Dawson brates can achieve energy expenditure levels and Marsh ( 1989 ), Hammond and Diamond ( 1997 ), Chappell up to a 100 times BMR for a few seconds et al . ( 1999 , 2003 , 2007 ), Videler ( 2005 ), Vézina et al . ( 2006 ), (Bennett and Ruben 1979 ). This should just be and Gill et al. (2005 , 2008). Note that especially cold-induced enough to outrun or outfl y a predator, which summit metabolic rate and exercise-induced maximal metabolic after all faces the same problem, namely the rate can vary a great deal among individuals, species and higher build-up of toxic levels of lactic acid during taxonomic levels. anaerobic ATP production. Still, exercise- trained animals may maintain energy expend- wheels to the point of exhaustion while meas- iture levels of 20–30 times BMR for many uring oxygen consumption (a measure called

minutes ( Schmidt-Nielsen 1984 ), though activ- maximal metabolic rate or VO 2max; see previous ity levels maintained for an hour or more chapter, especially Fig. 25 and Chappell et al . rarely, if ever, exceed 15 times BMR ( Fig. 35 ). 1996 , 1999 , 2007 , Bishop 1999 , Bundle et al . Exercise physiologists have developed dif- 1999 , Hammond et al . 2000 ). In many such ferent experimental procedures to assay meta- studies, BMR is measured in the same individ- bolic peak performance. For example, they uals, even though factorial scopes (i.e. multiples may expose animals to severe thermoregula- of BMR) vary a lot. Averaged for different age- tory demands until they show a decrease of or sex-categories or species, summit and maxi- core temperature (yielding a measure called mal metabolic rates usually fall between 5 and summit metabolic rate ; e.g. Swanson 1993 , 10 times BMR ( Fig. 35 ). Actually, these values Chappell et al . 2003 , Swanson and Liknes 2006 , seem rather low given the short intervals (min- Vézina et al. 2007 ) or they may chase them for utes rather than hours) over which such peak several minutes on treadmills or in running performance measurements are made. METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 61

Another interesting point is that, although length of the Pacifi c from Alaska to New both exercise- and cold-induced metabolic Zealand—a distance of 11,000 km—in single rates strongly involve skeletal muscle work (in non-stop fl ights that may last nine days or running and shivering, respectively), animals even a few more ( Gill et al . 2009 ). On the basis tend to have higher factorial scopes for exer- of reconstructed fat and protein losses during cise-induced power output than for cold- such fl ights, the godwits would be flying at induced power output (Hinds et al . 1993 ). In energy expenditure levels of about 9 times birds, exercise-induced scopes are higher than BMR (Gill et al . 2005 ). This level is consistent cold-induced scopes ( Dawson and Marsh 1989 , with modern wind-tunnel data, but still curi- Hinds et al . 1993 ), but in some small mammals, ously high compared to the trendline ( Fig. 35 ). cold-induced factorial scope is higher than It is almost double the long-term levels for sus- exercise-induced scope (Chappell et al . 2003 ). tained metabolic rate in hard-working mam- These differences may be due to the fact that mals and non-fl ying birds, yet takes place over birds carry a relatively larger musculature for similar lengths of time. locomotion than mammals, i.e. the fl ight mus- A big difference, of course, is that long-dis- cles, with the discrepancy between the sizes of tance migrants are not in energy balance. They great and small fl ight muscles negatively do not eat during the fl ight, but deplete the affecting shivering capacity ( Hinds et al . 1993 , sizeable stores of fat and proteins that are and see Piersma and Dietz 2007 ). In addition, accrued before they take off (Piersma and Gill mammals, but not birds, can use ‘brown fat’ to 1998 , Gill et al . 2005 ). Indeed, despite showing generate heat for thermogenesis (i.e. ‘non- evidence of remarkable protein saving (Jenni shivering thermogenesis’; see Cannon and and Jenni-Eiermann 1998 ), such migrant shore- Nedergaard 2004 , Morrison et al . 2008 ). With birds not only empty their abdominal and sub- their two channels of heat generation, mam- cutaneous fat stores, but, with the exception of mals achieve high cold-induced factorial lungs and brain, also burn up belly organs and scopes. other proteinaceous tissues in fl ight ( Battley Experimental measurements of the cost of et al . 2000 ). This raises the possibility that sus- bird fl ight, carried out in wind-tunnels over tained metabolic rates are constrained by the many hours (Pennycuick et al . 1997 , Kvist et al . capacity of animals to digest and assimilate 2001 , Ward et al . 2004 , Engel et al. 2006 , Schmidt- food, something that even active animals do Wellenburg et al . 2006 , 2008 ) also vary widely, not have to do as long as they rely on stores. Is but gravitate towards values around 9 times there evidence that food processing and assim- BMR ( Videler 2005 ). During such long and ilation, i.e. the lower funnels that feed into intense exercise, it was not the thrush nightin- Weiner’s (1992) ‘animal barrel’ (Fig. 29), could gale Luscinia luscinia called ‘Blue’ that fl ew indeed determine plateau energy expenditure continuously for 16 h that became exhausted, levels? but the human experimentalists that needed to stand guard at the wind-tunnel from 7 a.m. the Regulation of maximal performance: fi rst day to 1 a.m. the next (Lindström et al . central, peripheral, or external? 1999 , Klaassen et al . 2000 , Å. Lindström pers. comm.). Here let us return to the topic of the design of Using the latest satellite-transmitter technol- metabolic ceilings left in the previous chapter, ogy, it has now been demonstrated that bar- and the question of whether the food-process- tailed godwits Limosa lapponica cross the entire ing chain and respiratory chains show any 62 ADDING ENVIRONMENT degree of symmorphy. To set the scene, let us In principle the greatest constraint could be fi rst look at an advanced reincarnation of the availability of energy in the environment, Weiner’s barrel model ( Fig. 36 ). This diagram the energy intake rate (bow-tie A). However, summarizes the combined insights of University we have already made the point that it is of Aberdeen biologists John Speakman and unlikely that intake rate generally constrains Elz˙bieta Król (2005) after many years of experi- maximal performance—it certainly did not in mental studies on metabolic ceilings. Figure 36 the well-fed Swedish skiers or Tour de France makes clear that the energy absorbed into the cyclists—since maximum instantaneous food body (bow-tie B) is not the only process that can intake rates tend to be so much larger than constrain sustained energy intake and expendi- overall, long-term intake rates ( Jeschke et al . ture; four other processes (bow-ties A and C to 2002 , van Gils and Piersma 2004 , Jeschke 2007 ). E) could do this as well. Furthermore, in birds, food-supplementation

Environment Brain

Fat storage Food A Protein energy (growth) Glycogen B

Gut Metabolism storage

Absorbed Exported energy by gut energy

D C

Faecal Excreted energy energy (by-products of waste metabolism e.g. urea)

Work E Total heat production

Figure 36. Diagram summarizing energy fl ows in an animal. Arrows indicate energy fl ows, and the bow-ties A to E indicate possible rate-constraining processes (see text). Energy is ingested at the mouth and enters the alimentary tract where there is a storage capacity to accommodate excess intake. Some of the ingested energy cannot be absorbed by the gut and is eliminated as faecal waste. The energy that is absorbed into the body may be diverted into different types of storage. Boxes represent the steady- state sizes of stores and arrows the movement of energy into and out of these stores. There is long-term storage in the form of fat deposited into adipose tissue and short-term storage in the form of glycogen in the liver. Energy may also be diverted into growth, which we could call protein storage, and such energy may also be withdrawn. During conversions, some energy is lost as heat. Some energy is lost in materials that are eliminated as by-products of metabolism (such as the nitrogenous end-products of protein metabolism: urea, uric acid, and ammonia). The remaining energy can either be exported from the animal as specifi c compounds (milk during lactation, for example), or used for metabolism. This includes the energy used to fuel cellular processes (i.e. BMR), thermoregulation, and muscle contraction. Some energy is exported as mechanical work (locomotion, for example). Adapted from Speakman and Król ( 2005 ). METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 63 experiments have rarely managed to increase metabolic ceilings (Table 5 ). Attention has measures of peak performance (Meijer and focused on sustained energy intake rather than Drent 1999 , Speakman and Król 2005 ). As was expenditure because during reproduction ‘con- pointed out by John Speakman (pers. comm.), siderable amounts of energy are exported as the very existence of storage organs at the milk’ ( Speakman and Król 2005 ). In small entrance to the gut (crop, stomach; see Kersten mammals, pregnancy increases energy demand and Visser 1996 ) suggests that the rate at which by about 60%, while during peak lactation, energy can be harvested often exceeds the rate energy intake quadruples ( Johnson et al . at which it can be processed. In principle, it is 2001a ). To test whether this represents the also possible that metabolic ceilings are lim- metabolic ceiling, rodent mothers have been ited by the capacity to excrete metabolic end- given more young (Hammond and Diamond products, e.g. the capacity of the kidneys to 1992 , Künkele 2000 , Johnson et al. 2001a ), excrete urea (bow-tie C). In lactating mice fed young for a longer time ( Hammond and a high-protein diet, the latter may actually Diamond 1994 ), or more hungry young occur (J.R. Speakman pers. comm.). (Laurien-Kehnen and Trillmich 2003 ). In none For a long time there were two main hypoth- of these studies did food intake increase. Pup eses to explain the regulation of maximal per- masses typically went down, which makes formance. One, related to the capacity of the sense given that the same amount of food had alimentary tract to absorb energy (the food- to be shared by more pups. To test whether processing chain), is called the central limitation this apparent limit was centrally determined, hypothesis (bow-tie B). The other, related to the diet quality was reduced (Speakman et al . capacity to expend energy at the different sites 2001 , Hacklander et al . 2002 , Valencak and of utilization (the respiratory chain, with the Ruf 2009 ). Mice, and also European hares, muscles, testicles, ovaries, and, in mammals, responded by eating and assimilating more brown adipose tissue and mammary glands as food, suggesting the ability of the digestive the sites of heat production) is called the periph- tract to upregulate capacity. eral limitation hypothesis (bow-tie D) (Ricklefs In order to increase selectively the demands 1969 , Hammond and Diamond 1997 , Hammond of the peripheral systems, Perrigo ( 1987 ) forced et al . 2000 , Bacigalupe and Bozinovic 2002 , lactating mice to run for their food, but found Speakman and Król 2005 ). Although the null that, despite free access to food, neither of the hypothesis of symmorphosis predicts that nei- two mouse species tested showed elevations ther of them would be limiting on its own of food intake. Furthermore, when Kristin (Chapter 3 ), these two alternatives have Schubert and colleagues (2009) applied a simi- attracted a lot of elegant experimental testing, lar paradigm in laboratory mice, they found especially in small rodents. The rigour of the that minimum metabolic rate and daily energy tests in this research programme, and the fact expenditure (DEE) were both actually reduced that they turned up an exciting new hypothe- (see below, ‘The evolution of laziness’). When sis, justifi es some elaboration. peripheral demands were elevated, by making Rodents breed fast, have fur, but can easily lactating rats and mice simultaneously preg- be exposed to temperatures that will elevate nant, food intake did not increase either their metabolism, and have mammary glands ( Biggerstaff and Mann 1992 , Koiter et al . 1999 , that produce, and teats that deliver, milk. Johnson et al. 2001c ). However, in all studies These characteristics make small mammals where rodents were exposed to cold ( Table 5 ), excellent models for experimental studies of food intake did increase, especially during 64 ADDING ENVIRONMENT

Table 5. What can lift the metabolic ceiling? Nine experimental ways to examine the various physiological mechanisms that could limit sustained energy intake in rodents. Expanded from reviews in Speakman and Król ( 2005), Król et al . ( 2007 ) and Speakman ( 2008 ). No. Manipulation ‘Taxon’ examined Reference

1 Increase litter size Swiss Webster mouse Hammond and Diamond (1992) Guinea pig Cavia Künkele (2000) porcellus MF1 laboratory mouse Johnson et al . (2001a) 2 Prolong lactation Swiss Webster mouse Hammond and Diamond (1994) 3 Increase demand of precocial pups by Guinea pig Laurien-Kehnen and Trillmich witholding food early in lactation (2003) 4 Decrease quality of food so that they have MF1 laboratory mouse Speakman et al . (2001) to eat and digest a greater volume European hare Lepus Hacklander et al . (2002) europaeus European hare Valencak and Ruf (2009) 5 Force lactating animals to run for food Deer mouse Peromyscus Perrigo (1987) maniculatus Wild-type house mouse Perrigo (1987) Mus domesticus Outbred house mouse ICR Schubert et al . (2009) (CD-1) 6 Make them simultaneously pregnant and Rockland-Swiss mouse Biggerstaff and Mann (1992) lactating Brown rat Rattus Koiter et al . (1999) norvegicus MF1 laboratory mouse Johnson et al . (2001c) 7 Increase temperature gradient between Swiss Webster mouse Hammond et al . (1994) animal and environment Mouse Apodemus Koteja (1995) fl avicollis Deer mouse Koteja (1996) Cotton rat Sigmodon Rogowitz (1998) hispidus Deer mouse Hammond and Kristan (2000) MF1 laboratory mouse Johnson and Speakman (2001) House mouse Koteja et al . (2001) 8 Surgically remove some mammary glands Swiss Webster mouse Hammond et al . (1996b) 9 Shave fur of mothers to increase heat loss MF1 laboratory mouse Król et al. (2007)

lactation! All this suggested that performance Kimberly Hammond and colleagues ( 1996b ) limits were not set centrally. That mice could tested this by surgically removing half of the increase food-intake rates when muscles and mammary glands. If the capacity of these tis- brown adipose tissue were made to work sues were fl exible, it would expand, but it did harder in the cold, suggested that it was actu- not. This seemed consistent with the peripheral ally the capacity of the mammary glands that limitation hypothesis until Speakman’s group, set the upper limit. using the doubly-labelled water technique in METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 65 novel ways (Speakman 1997a ), began to meas- Speakman ( 2003a , 2003b ) concluded that per- ure milk production ( Johnson and Speakman formance in lactating mice must be limited 2001 , Król and Speakman 2003a , b , Król et al . by the capacity of the female mouse to dissi- 2003 ). As noted earlier, the mice ate more, the pate heat. They put this to a test by shaving colder it got. However, with increased heat some mice during lactation (Król et al . 2007 ). loss, milk production also went up and so did As shown in Fig. 37 , this experiment con- the mass of pups at weaning! Król and fi rmed that shaved mothers ate more and had heavier litters with heavier individual pups. Król and Speakman were quick to recognize 30 * that the problem of heat dissipation is proba- * * bly rather general, and is likely to constrain 25 metabolic performance in many ecological

20 contexts. Apart from the well-known link between high temperature, solar radiation, 15 and humidity, and the reduction in milk pro- food intake (g/day) intake food shaved mothers control mothers A duction in dairy cattle ( Thompson 1973 , King 10 et al . 2006 , Nassuna-Musoke et al . 2007 ), sheep 120 Ovis aries ( Abdalla et al. 1993 ), and pigs Sus * scrofa ( Black et al . 1993 , Renaudeau and Noblet 100 * * * 2001 ), heat-dissipation limitations might help 80 explain latitudinal and altitudinal trends in 60 clutch sizes of birds (larger where it is colder; Bohning-Gaese et al . 2000 , Cooper et al . 2005 ,

litter mass (g) 40 but see McNamara et al. 2008 ). It could also 20 B provide a reason why humans, one of the few 0 mammals that can sweat profusely, are able to 10 * outrun pretty much any running game in the * * midday heat ( Heinrich 2001 ). And it could 8 * help explain why long-distance migrant birds 6 tend to fl y at night ( Klaassen 1996 ), and why fuelling rates of migrant shorebirds are so low 4 in the tropics ( Battley et al . 2003 , Piersma et al .

pup body mass (g) 2 C 2005 ). 0 0 42 6810 12 14 16 18 20 day of lactation Temperature and the allometric scaling constant Figure 37. Food intake of 20 shaved and 20 unshaved control MF1 laboratory mouse mothers during (A) lactation, (B) the As most of us know from experience, bodies heat mass of their litters, and (C) the mass of their individual pups. up during exercise. In running horses, (near- From days 6 to 12 the mothers had dorsal fur shaved off to lethal) body temperatures as high as 45 °C have increase heat loss capacity. Values are means ± 1 SD. Stars indicate days with signifi cant differences (P < 0.05) between been measured ( Hodgson et al. 1993 ). Just as the the shaved and control groups. capacity to lose heat can limit sustained perform- Compiled from Król et al . ( 2007 ). ance in lactating mice, rising body temperatures 66 ADDING ENVIRONMENT

can put a limit to the duration or the level Protecting long-term fi tness assets: of intense exercise ( Weishaupt et al . 1996 , factorial scopes and optimal working Fuller et al . 1998 , González-Alonso et al . 1999 , capacity revisited Arngrimsson et al . 2004 ). Noting that the allometric scaling constant for maximal metabolic rate When one builds bigger bodies to ingest, in mammals (ca. 0.80) is consistently higher than absorb, excrete, and do work, the expand- that for BMR (0.70–0.75; Taylor et al . 1981 , Bishop ing machinery requires ever more energy for 1999 , Savage et al . 2004 , Weibel et al . 2004 ), maintenance. Such increases in maintenance Gillooly and Allen ( 2007 ) proposed that the requirements will be refl ected by increasing steeper size-dependence of maximal metabolic BMR levels. A summary of the early studies on rate than of BMR may be explained by larger mice whose energy expenditure had been ele- mammals achieving the highest muscle temper- vated in a variety of ways (see Table 5 ), shows atures at maximal activity. The larger the animal, BMR to go up with sustained metabolic rate the more unfavourable its surface–volume ratio (Fig. 38A ). This effect comes with an increase when losing, rather than retaining, heat. In gen- in body mass, due to hypertrophy of the eral, chemical reaction rates increase exponen- liver, digestive tract, and mammary glands tially with temperature, as does metabolic rate ( Hammond et al . 1994 , Speakman and Johnson ( Gillooly et al . 2001 ). 2000 , Johnson et al . 2001b ). As a consequence Gillooly and Allen (2007 ) suggest that the of having put in place increasingly powerful short-term maximal metabolic rate (technically machinery, factorial scope (the ratio between sustained energy expenditure and BMR) called VO2max ) is a function of body temperature. This led them to the somewhat counter- climbs to an asymptote of about 7 times BMR intuitive prediction that mammals should be ( Fig. 38B ; Hammond and Diamond 1997 ). able to achieve higher maximal metabolic rates We have come full circle. During our quest at higher ambient temperatures, because heat to establish whether maximal aerobic per- loss is more restricted, leading body tempera- formance is limited centrally, by digestion and tures to rise more. Craig White and colleagues assimilation systems, or peripherally, by the ( 2008 ) tested this hypothesis on the basis of power that can be generated by the receiving published maximum aerobic performance organs, experimental evidence has not been measurements in rodents, and rejected it. They uniformly consistent with either alternative. In a sense, this celebrates the notion of symmor- found no relationship between VO2max and testing temperature, which varied from –16 ° to phosis, the principle of logical, economic ani- +30 °C. Interestingly, they did fi nd that animals mal design (Chapter 3 ) : symmorphic design that were acclimated at lower temperatures would have negated evidence in favour of achieved higher maximal metabolic rates. Cold either hypothesis for a single site limiting aer- necessitates shivering, and thus a more intense obic performance. That rates of heat loss best use of muscle tissue. This, as we shall see in the explain the limits to aerobic performance in next chapter, usually leads to hypertrophy. In lactating laboratory mice, which only push up fact, within species, such chronic muscle-medi- their metabolic ceilings when it is cold enough ated thermogenesis can lead to greater increases or when insulation is poor enough, does not in aerobic capacity, whether measured as maxi- detract from the fact that mouse mothers can mal or summit metabolic rate, than endurance only do this by building a bigger body, a big- training (Schaeffer et al . 2001 , 2003 , 2005 , Vézina ger metabolic machine. Thus, they still face a et al . 2006 ). maximum factorial scope of 7. METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 67

evolutionary sense if working hard comes at a 50 survival cost ( Valencak et al. 2009 ). If very hard 40 work precipitously increases the likelihood of death (e.g. because of free-radical derived oxi- 30 dative DNA and tissue damage; Harman 1956 , 20 Alonso-Alvarez et al . 2004 , Wiersma et al . 2004 , Cohen et al. 2008 ), without leading to com- 10 A pensatory increases in reproductive output basal metabolic rate (kJ/day) basal metabolic rate 0 (Williams 1966a , 1966b , Lessells 1991 , Daan 8 and Tinbergen 1997 ), evolutionary trade-offs would select for animals that are not prepared 6 to work harder than what we can call the ‘optimal working capacity’ (Fig. 39 ; Royama 1966 ). 4 On the basis of fi eld data on provisioning birds,

factorial metabolic scope factorial B V 0 c 0 100 200 300 sustained metabolic rate (kJ/day) Vp clean Figure 38. How the laboratory mouse equivalent of basal metabolic rate (BMR) goes up with experimentally increased fitness Vp dirty levels of energy expenditure due to hypertrophy of various organ groups (A), and how simultaneously the factorial scope of sustained energy expenditure on BMR reaches plateau levels A of ca. 7 times BMR (B). Adapted from Hammond and Diamond ( 1997 ).

V clean So, if animals are pushed very hard, under fitness some conditions (e.g. when enabling for greater V dirty heat loss), they can raise the ceiling from working at 5 times BMR (Fig. 37 ) to working at 7 B times BMR (Fig. 38 ). As we have seen, endur- 0 2468 ance athletes in energy balance can push their working level (relative to BMR) performance levels to 5 times BMR, but not further. We have also seen that free-living birds, Figure 39. The marginal value of work: Royama’s (1966) concept of an optimal working level during reproduction except in the case of marathon migrants, gener- couched in terms of a simple optimality plot of (A) the two ally do not work harder than 4 times BMR (Fig. main fi tness components (the residual reproductive value of 34 ). The considerable gap between the maxi- the parent Vp and the reproductive value of the clutch Vc) or mum sustained working level of 4 times BMR (B) their sum as a function of the parent’s relative metabolic that hard-working parent birds are prepared to rate. Two sorts of environment are distinguished: a ‘clean’ one in which the chance of catching a disease is small, and a ‘dirty’ give (Drent and Daan 1980), and the physiolog- one with a higher chance of becoming infected. ical maxima of 5–7 times BMR that can be Inspired by fi g. 13 in Drent and Daan (1980) and achieved under exceptional conditions, makes fi g. 1 in Daanet al . ( 1990b ). 68 ADDING ENVIRONMENT

optimal working capacity is seldom higher Selman et al . 2008 , Schubert 2009, Vaanholt than 4 times BMR (Drent and Daan 1980) and et al . 2009 ). Basically, elevating DEE did not most often lower (Bryant and Tatner 1991 , not shorten life span. Tinbergen and Dietz 1994 ). Speakman (1997b ) Whether or not enhanced energy expendi- found the same in small mammals, stating that ture has life-shortening effects, there is still these animals ‘routinely live their lives at well evidence that mammals, as well as birds, pre- below their physiological capacities, in the fer to work at optimal levels, where metabolic same way that we drive our cars at well below investment in reproduction is concerned. their mechanical capacities’. Lactating mouse mothers challenged to work for food support make the ‘decision’ not to work above a certain level—or indeed, to The evolution of laziness increase body mass as required for maintaining metabolic machinery—even though this As we have seen, there seems abundant evidence results in a linear decrease in milk energy out- that animals, even when confronted with chal- put and thereby the biomass of weaned litters lenges such as experimentally increased brood (Schubert et al. 2009). Perrigo ( 1987 ) earlier sizes that might gain them more offspring, are observed that two wild-type mouse species simply not prepared to work harder than 4 times differed in their willingness to work in sup- BMR (Drent and Daan 1980) or, indeed, harder port of offspring. He attributed this to their than necessary ( Masman et al. 1988 ). Nevertheless, different life-history strategies. While Mus only a few experiments have demonstrated that musculus are opportunistic breeders capable of ‘increased daily work precipitates natural death’. maintaining fertility whenever food is availa- An important example is the study by Serge ble, Peromyscus maniculatus are strictly sea- Daan and co-workers (1996) with that very title, sonal, and presumably face a higher in which they analysed time till death of 63 free- opportunity cost of litter production. This spe- living kestrels Falco tinnunculus that had raised cies appeared more determined to save a broods with manipulated sizes. They were able smaller number of pups when Mus mothers to demonstrate that kestrels with larger broods had already abandoned their litters. incurred increased mortality the following win- Earlier we encountered the élite cyclists who ter compared with those with normal or reduced perform at very high levels during the three brood sizes. weeks of the Tour de France (Westerterp et al . Experiments with insects confi rmed these 1986 , Table 4 ). Interestingly, in the Vuelta a trends. When honey bee workers Apis mellifera España, a similarly gruelling cycling race but a were forced to fl y with extra loads and work week shorter, total energy expenditure was really hard, they showed decreased life spans similar ( Lucia et al . 2003 ). This suggests that ( Wolf and Schmid-Hempel 1989 ). When house when expending energy, humans, and proba- fl ies Musca domestica were prevented from fl y- bly other animals, account for time. Animals ing by keeping them in small fl asks, so that are only prepared to deliver a certain level of their energy expenditure decreased, they had effort over certain time periods, a phenome- increased life spans ( Yan and Sohal 2000 ). non called ‘pacing’ in the world of sport Interestingly, extra work load experiments on (Foster et al . 2005). In the words of Tim Noakes laboratory rats and various kinds of mice have (pers. comm.): ‘Train hard when it is impor- yielded equivocal results ( Holloszy and Smith tant, and rest when it is important. And don’t 1986 , Navarro et al . 2004 , Speakman et al . 2004b , confuse the two.’ We suggest that the time METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 69 spans over which the accounting of metabolic We believe that such insights need to come performance takes place in different animals from adequate incorporation of ecology into could be a rich area of experimental research. what at fi rst sight may seem a purely physio- Free-living animals are resistant to experi- logical problem. The impressive portfolio of mental manipulations to increase their instan- studies on highly inbred strains, or genetically taneous work levels, for what we can now see modifi ed lab mice, will thus need expansion to are sound reasons. Unless some mechanistic a variety of free-living animals, with due atten- shortcuts are found to the actual expenditure tion to ecological context for an appreciation of decision processes (e.g. Speakman and Król the evolutionary causes of metabolic ceilings. 2005 ), it will be impossible to verify experimentally the optimality model of Fig. 39 along Evolutionary wisdom of physiological the whole range of work levels. Nevertheless, constraints it seems likely that within taxa the ‘death on working level curves’ would be determined Unlike bar-headed geese Anser indicus that by ecological conditions, e.g. the general envi- have evolved particular muscle phenotypes to ronmental or social stress levels that have migrate over the Himalayas at heights of 9000 m become known as ‘allostatic loads’ ( McEwen ( Scott et al . 2009 ), humans clearly are a low- and Wingfi eld 2003 , Wingfi eld 2005 , Rubenstein altitude species. When Reinhold Messner and 2007 ), or the levels of pathogens in the envi- his team were getting close to the summit at ronment ( Piersma 1997 ). 8848 m of Mount Everest on 8 May 1978, expe- Such variations suggest that there is scope riencing an equivalent of 7% oxygen rather for studies in which relationships between than the 21% at sea level, they found it increas- energy expenditure levels and survival are ingly hard to stay on their feet ( Messner 1979 ): studied, either in a strict comparative context ‘every 10–15 steps we collapsed into the snow (comparing relationships in groups with dif- to rest, then crawled on’. Their leg-muscle ferent environments) or in experimental con- fatigue could not have been due to lactic acid texts (by additionally manipulating aspects of poisoning because, paradoxically, observa- the environments in which the expenditure– tions on human exercise at low partial oxygen survival relationships are studied). During pressures show that, whereas lactic acid levels such endeavours it remains important to pay should go up with increasing hypoxia at high attention to constraints other than work level, altitudes, they actually go down (Bigland- especially time limitations, that prevent ani- Ritchie and Vollestadt 1988 ). Also, paradoxi- mals from keeping up with the hard work and cally, rather than going up to compensate for from eating for very long (e.g. Kvist and the decreasing partial oxygen pressures at Lindström 2000 , Tinbergen and Verhulst 2000 ). great heights, heart rates actually go down When interpretable aspects of the ecological with altitude (Sutton et al . 1988 ). This change is context can be linked to maximal performance, related to blood being shunted away from the or rather to the precise levels and relationships muscles (hence the fatigue experienced by between experimentally manipulated work Messner and his comrades; Amann et al . 2006 ) levels and both reproduction and survival and towards the brain—the ‘selfi sh’ organ that ( Fig. 39 , and see Bouwhuis et al . 2009 for a pos- controls oxygen delivery to the system ( Noakes sible fi eld setting), we may achieve deeper lev- et al . 2001 , 2004 ). Noting that, during impor- els of understanding of the selection factors tant games, athletes can deliver an end spurt at driving the height of metabolic ceilings. the point where their muscles must show 70 ADDING ENVIRONMENT

greatest fatigue, Noakes ( 2007 ) concludes that evolution to deliver sensible, evolutionarily it is also the brain that regulates exercise per- informed, behaviours. The evolutionary algo- formance by arranging an anticipatory rithms underpinning such sensible behaviour response pattern: only the brain has the knowl- prevent animals from ‘overdoing it’ and losing edge that the fi nish is close. In more general life and limb whilst doing so. This was the terms, he writes, ‘the limits of our endurance ‘wisdom’ that slowed Messner down before he lie deeply in the human brain, determined by reached Mount Everest’s summit. And it is this our heredity and other personal factors yet to wisdom that makes kestrels ‘lazy’ and unwill- be uncovered’ ( Noakes 2006 ). ing to spend much more than 500 kJ/day, or 4 To this we would add that the limits to times BMR ( Fig. 40 ; Deerenberg et al . 1995 ). endurance lie in any animal’s brain. Brains have been shaped in the process of organic Beyond Rubner’s legacy: why birds can burn their candle at both ends In 1908, the German animal physiologist Max 3 survivors Rubner published a comparison between intake

2 rates of six domestic animals (guinea pig, cat Felis domesticus , dog, cow, horse, and human) 1 and their life span (Rubner 1908 ). Despite a

number of birds number 50,000-fold difference in body mass between the 0 smallest and the largest species, he was struck 1.0 by the fi nding that there was only a five-fold 0.8 variation in what we will call here the mass-spe-

0.6 cifi c lifetime energy expenditure (the multiplica- tion of mass-specifi c rate of energy expenditure 0.4 and maximum life span). Humans had by far local survival rate 0.2 the largest mass-specifi c lifetime energy expend- 0.0 iture within this group of species; taking them 3 non-survivors out of the comparison, the range of variation was reduced to a factor of 1.5. 2 About a decade later, the American biologist Raymond Pearl ( 1922 , 1928 ) used Rubner’s 1

number of birds number conjecture as the basis of his ‘rate-of-living’

0 theory: the idea that animals have a certain 200 300 400 500 600 allotment of energy to spend during a lifetime, DEE (kJ/day) or a fi xed number of heartbeats (e.g. Calder Figure 40. Local survival rates of kestrels Falco tinnunculus in 1984 ). In Eric Charnov’s words we could call the northern Netherlands as a function of daily energy this a ‘life-history invariant’ ( Charnov 1993 ). expenditure (DEE) during the raising of chicks. The upper and According to the rate-of-living theory, the ele- lower panels show the frequency distributions of experimental vated energy expenditure levels of Captain birds that were either (top) or not (bottom) recaptured in the study area during the next breeding season. The middle panel Scott and his comrades during their Antarctic shows the logistic regression curve of survival on DEE ( X ² = exploits would have shortened their lives, 5.66, df =1, P = 0.02). even if they had survived the attempt to reach From Deerenberg et al . ( 1995 ) . the South Pole fi rst. From our point of view, METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 71 the rate-of-living theory is attractive because it mammal, we would like to discuss what may could provide a link between metabolic per- actually be pretty good evidence for a certain formance and typical life-history variables, invariance in lifetime energy expenditure. The such as annual survival and life span. But how case is reminiscent of the levelling-off of athletic invariant is lifetime energy expenditure, or performance measures in humans and other rather, at what taxonomic level is it invariant, domestic animals (Fig. 28). In the endeavour to and what does this tell us about the evolution- increase economic productivity of farm animals ary levelling of metabolic ceilings? by ‘improved’ keeping conditions, better food, The most recent, and the most comprehen- and intense genetic selection, it is now widely sive, assessment of Rubner’s idea that animals acknowledged that clever breeding and hus- have a certain amount of energy to spend was bandry schemes cannot prevent the most highly made by John Speakman and associates productive animals from being most at risk ( Speakman et al . 2002 , Speakman 2005, Furness from behavioural, physiological, and immuno- and Speakman 2008 ). As summarized in Fig. logical problems ( Rauw et al . 1998 , Lucy 2001 ). 41 , among species there is about a 20-fold (in For example, in the USA, the average milk mammals, note natural log scales) to 40-fold yields of dairy cows quadrupled between 1950 (birds) variation in lifetime energy expendi- and 2007, but despite earlier maturation, the ture per gram of body mass. It does not seem overall reproductive life spans of these cows to matter whether lifetime energy expenditure nearly halved ( Knaus 2009 ). Livestock breeders is (under-) estimated on the basis of BMR, or seem to be faced with Charnov’s life-history properly estimated using measurements of invariance, although they are not quite pre- daily energy expenditure. There are clear pared to accept this reality (e.g. Liu et al . 2008 ). declines of mass-corrected lifetime energy That birds tend to have so much more time, expenditures with body size ( Fig. 41 ) and, and so much more energy, to spend in a life- therefore, lifetime expenditures can hardly be time than mammals, is correlated with the called invariant. Lifetime expenditures cer- fi nding that rates of bodily decline (i.e. rates of tainly appear taxon-specifi c, a conclusion that ageing) are lower in birds than in mammals shows most clearly in a comparison between (Holmes and Austad 1995a , 1995b , Ricklefs lifetime energy expenditures of mammals and 1998 , 2000 , Ricklefs and Scheuerlein 2001 , birds (Fig. 41 right panels). Indeed, as related Costantini 2008 , Holmes and Martin 2009 ). to us by our text editor Marie-Anne Martin, Getting a bit ahead of ourselves, this may be ‘my friends in North Wales are always sad because birds, at least the fl ighted ones (and when they keep losing their pet rats and fer- also bats, the only group of fl ying mammals), rets, while my friend in Gloucestershire has have experienced selection pressures that have just incorporated her 54-year old African grey allowed the evolution of biochemical mecha- parrot into her will’. This is the more surpris- nisms that reduce the wear and tear on tissues ing since birds generally have higher rates of ( Barja et al. 1994 , Perez-Campo et al . 1998 ). The energy expenditure than mammals (e.g. Koteja mechanisms, which relate to mitochondrial 1991 , Ricklefs et al . 1996 ). The average bird has functioning and the ways that electron leakage 3–4 times more kJ to spend on BMR (Fig. 41 ), and free-radical production are balanced under and also on DEE (Speakman 2005 ), than the conditions of high and low energy demands, average mammal. are just beginning to be revealed ( Buttemer Before returning to why an average bird et al. 2008 , Costantini et al . 2008 , Dowling and spends so much more energy than an average Simmons 2009 , Monaghan et al . 2009 ). We leave 72 ADDING ENVIRONMENT

BMR-based DEE-based BMR-based

11 70 10 60 9 50 8 40 30

7 MAMMALS 20 6 10 5 0

11 30 10 absolute frquency (lifetime expenditure, kJ/g) expenditure, (lifetime e 9 20 log 8 BIRDS 7 10 6 5 0 0 2 46810 12 02468 5 6 78910 loge (body mass, g) loge (lifetime expenditure, kJ/g) Figure 41. Contrasts between mammals (top panels) and birds (bottom panels) in mass specifi c lifetime energy expenditure (kJ/g) as a function of body mass (left two sets of panels) or as a frequency distribution (right set of panels). Values were calculated by multiplying life span of a species either by measured BMR or measured daily energy expenditure (DEE). Compiled from Speakman ( 2005 ), where the statistical details can also be found. Note that mass-specifi c lifetime energy expenditure declines signifi cantly with body mass (see also Furness and Speakman 2008 ).

readers to consult Nick Lane (2005 ), in his Monaghan et al . 2008 ). Our reading of the splendid Power, sex, suicide: Mitochondria and the literature suggests that any kind of hard work, meaning of life, for a summary of the fi eld and perhaps above taxon- (or rather, ecology-) the development of this complicated, but dependent thresholds, does come with wear hugely exciting, story. What is most important and tear. A precipitous increase in the likeli- here is that the capacity to fl y seems to go with hood of organ or performance failure, and lower extrinsic (ecological) mortality, such as mortality associated with increases in energy the likelihood to be depredated (Austad 1997 ), expenditure (Ricklefs 2008 ), would explain and that lower extrinsic mortality, in combina- why animals are reluctant habitually to spend tion with the high metabolic demands of fl ight, as much as they are physiologically capable of may well be driving the appearance of the ( Valencak et al. 2009 ). Indeed, it would explain mitochondrial mechanisms that reduce intrin- why special kinds of motivation are needed to sic wear and tear (Lane 2005 ). lift our own modest human metabolic ceilings. At this point, we have returned to ecology, the driver of what may at fi rst have seemed Hard-working shorebirds purely mechanistic, physiological processes that one can study in isolation from the messy In the 1970s and 1980s, biologists interested in outside world ( Partridge and Gems 2007 , animal energetics were still pre-occupied with METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 73 the measurement, comparison, and interpreta- investment (the clutch) at a small place (the tion of BMR in as wide a variety of animals as nest), they may be tricked into entering traps possible (e.g. Aschoff and Pohl 1970 , McNab twice. A series of expeditions to the most 1986 , 1988 , 2002 ). As a University of Groningen northerly lands on earth (Alerstam and Jönsson undergraduate, one of us (TP) joined the fray, 1999 ), enabled us to collect enough data on and, together with Marcel Kersten, started energy expenditure of shorebirds incubating a measuring BMR in several species of shorebird clutch to test the idea that, during that phase held in captivity. It soon became clear that the of their annual cycle, they show relatively high relatively high values of BMR in shorebirds levels of energy expenditure (Piersma et al . that we measured, and that had been found by 2003 ) in comparison with bird species at lower others, were no artefact of method, but a phe- latitudes ( Fig. 42 ). The small shorebirds, for nomenon typical of the group (Kersten and which the costs of thermoregulation will be Piersma 1987 ), at least of those living in north relatively highest, actually showed energy temperate cold climates ( Kersten et al . 1998 , expenditures that would have been predicted Kvist and Lindström 2001 ). Not only shore- by Kirkwood’s (1983) equation to represent birds, but also songbirds, were found to have metabolic ceiling levels. higher BMRs than other birds ( Aschoff and These incubating shorebirds were in energy Pohl 1970 ). Cage-existence metabolism (a balance (Tulp et al . 2009 ). However, typical for ‘standardized’ measure based on energy intake long-distance migrants that make very long, within the confi nements of aviaries; Kendeigh non-stop fl ights (Newton 2008 ), shorebirds go 1970 ) was consistently high as well. At the through many phases when they are not in time, high BMR values were interpreted as refl ecting the high maintenance requirements of the considerable body machinery needed to support high levels of performance in real life 3.2 kmax ( Kersten and Piersma 1987 ; and see Box 3 ). As we have seen, doubly-labelled water 2.8 2 18 ( H2 O) dilution rates in the body fl uids of (re-) sampled animals, including athletes, pro- 2.4 vide robust integral measurements of energy expenditure of animals that go about their own log DEE (kJ/day) 2.0 business in the intervening interval (Speakman 1997a ). The critical requirement is recapture. 1.6

During the non-breeding season, when most 0.8 1.2 1.6 2.0 2.4 2.8 shorebirds fi nd safety in numbers as they roam log body mass (g) around in huge fl ocks ( Piersmaet al . 1993c , Figure 42. Daily energy expenditure (DEE) as a function of van de Kam et al . 2004 ), recapture within inter- body mass in tundra-breeding shorebird species (closed vals of a few days is impossible (but see Castro triangles) and one long-tailed skua Stercorarius longicauda, et al . 1992 for a single example). Thus, assess- (the largest species) during the incubation phase. These are ments of energy expenditure levels during the compared with DEE during incubation of lower latitude bird species (open symbols; from Tinbergen and Williams 2002 ). non-breeding season have to be based on indi- Shorebirds are indicated with triangles, other species with rect methods ( Wiersma and Piersma 1994 , circles. kmax is the suggested upper limit to daily metabolizable Cartar and Morrison 2005 ). During incubation, energy intake for homeotherm animals ( Kirkwood 1983 ). however, when shorebirds have made an From Piersma et al . ( 2003 ). 74 ADDING ENVIRONMENT energy balance on a daily basis, particularly amongst the highest values measured in refu- during the fuelling phase and when burning elling birds (Lindström 1991 , 2003 ). off all the stored energy during the fl ight phase Migrating shorebirds may have the capacity ( Piersma 1987 ). Anders Kvist and Åke Lindström to process food and store fuel at such fast rates ( 2003 ) capitalized on the availability of hungry because they are usually in a hurry and because individuals of a large variety of shorebirds mak- they make such long-distance fl ights. They are ing a short stop during southward migration in in a hurry because they need to get to the arctic the Baltic. These shorebirds have a great urge breeding grounds in good time (Gudmundsson to fuel up quickly and move onward, and they et al . 1991 , Drent et al . 2003 ), or because getting just love mealworms Tenebrio sp. In this setting, back early to their southward staging areas or Kvist and Lindström were able to show that wintering grounds is benefi cial (Zwarts et al . shorebirds as a group have unrivalled capaci- 1992 , Alerstam 2003 ). As we have earlier ties to process food and refuel fast (Fig. 43 ). alluded to ( Fig. 35 ), during long-distance BMR was measured concurrently with the fl ights of many thousands of kilometres, shore- intake measurements, and most species had birds spend the equivalent of 8–10 times BMR energy assimilation rates higher than 6 times ( Pennycuick and Battley 2003 , Gill et al . 2005 ). BMR. This enabled them to increase mass by as In the course of such non-stop fl ights, shore- much as 10% of fat-free body mass per day, birds not only deplete much of their fat store,

n = Philomachus pugnax m (189, 136 g) 2 Pluvialis apricaria (187, 161 g) 1 Pluvialis squatarola (177, 165 g) 1 Tringa nebularia (157, 132 g) 1 Calidris canutus (128, 103 g) 3 Tringa totanus (116, 98 g) 3 Arenaria interpres (111, 88 g) 3 Philomachus pugnax f (106, 86 g) 5 Gallinago gallinago (89, 89 g) 3 Tringa ochropus (72, 63 g) 1 Tringa glareola (61, 51 g) 13 Calidris ferruginea (56, 48 g) 14 Calidris hiaticula (55, 48 g) 5 Actitis hypoleucos (49, 38 g) 6 Calidris alpina (46, 41 g) 30

01246810 20 5 10 15 20 energy assimilation rate (x BMR) Δm (%) Figure 43. Maximum energy assimilation and fuel deposition rates of captive shorebirds intercepted during southward autumn migration in southern Sweden. Parenthetical values following the species name represent average body mass of the birds on the day (24-h period, see text) when maximum energy assimilation rate was measured and average body mass when BMR was measured. Energy assimilation rates (grey bars to the left) are presented as multiples of BMR to normalize for differences in size and metabolic capacity between animals. The black bars to the right represent the mass increase per 24 h (∆m), the 24-h period with the maximum energy assimilation rate, expressed in % of fat-free body mass. Error bars represent standard deviations. Modifi ed from Kvist and Lindström ( 2003 ). METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 75

120°E150°E180°E 150°W120°W the time and energy allocated to migration, but also minimize the risk of mortality from preda- 60°N tors, parasites, and pathogens. Bar-tailed godwits and other shorebirds appear to have been

30°N selected to work routinely as hard, and almost as long, as the arctic explorers with whom we started this chapter. They also deplete their 0° bodies to the same extent. However, by making a living along the edges of our planet, so to speak, in places that some parasites, pathogens, 30°S and predators have not been able to reach (e.g. 2000 km Piersma 1997 , 2003 , Mendes et al . 2005 , Gill et al. 2009 ), godwits may exploit ecological con- Figure 44. When the sky is the limit: southward fl ight tracks of ditions that do not make them pay the price (of nine bar-tailed godwits that maintained estimated rates of energy expenditures of 9 times BMR for up to 9.4 days. Individual godwits enhanced subsequent mortality) that usually were fi tted with satellite transmitters and followed during their comes with very hard work. fl ights during August–October 2006 and 2007. Small fi lled circles denote satellite-informed locations collected during 6–8-h intervals, and solid lines show interpolated 24–36-h tracks between Synopsis reporting periods. The intensity of shading of the land-masses correlates with increasing lack of vegetation cover. It is clear that animals are not usually willing to Modifi ed from Gillet al . (2009). perform at levels, or for lengths of time, of which they should be maximally capable. We think that this strategy has been selected by the but also lose protein, due to atrophy of organs costs of high performance levels, in the form of in the body cavity and even the working fl ight subsequently reduced survival or future repro- muscles ( Piersma and Jukema 1990 , Battley duction. The somewhat lopsided scientifi c et al . 2000 ). Developments in satellite-transmit- efforts to determine the proximate limits to ani- ter technology have recently made it possible mal performance, by looking at the metabolic to establish that such performance levels are machinery rather than at complete life-histories maintained for at least 9 days in bar-tailed in naturalistic settings, has made an evolution- godwits ( Fig. 44 ). Maintaining an estimated ary understanding of metabolic ceilings still metabolic rate of 8–10 times BMR for more beyond reach. In relative terms, the long-term than 9 days represents a combination of meta- metabolic ceiling of ca. 7 times Basal Metabolic bolic intensity and duration that is unprece- Rate in challenged individuals may be real and dented in the current literature on animal general, because greater performance requires energetics ( Fig. 35 ). larger machinery that is ever more expensive On the basis of the arguments developed to maintain. However, our reading of the litera- earlier in this chapter, there should be good ture suggests that absolute limits to perform- ecological reasons why these bar-tailed god- ance are mostly relative to ecological context, wits are prepared to work so hard for so long which sets the severity of the survival punish- (note that their annual survival is also very ment for over-exertion. Over generations, the high, >96%, P.F. Battley pers. comm.). Gill et al . many dimensions of ecology thus determine (2009 ) propose that, by making a non-stop how far individual animals, including human trans- Pacifi c fl ight, godwits not only minimize athletes, are prepared to go.

76 ADDING ENVIRONMENT

Box 3 BMR: interesting baseline, epiphenomenon, or both?

‘The most remarkable thing about the apteryx Ricklefs et al. ( 1996 ) discussed whether (the kiwi) was its small respiratory system, metabolic ceilings should be seen as a direct suggesting that in the wild this must be a shy, function of BMR, or whether these quantities patient, creepy little bird, with little inclination should be treated as energetic consequences to exert itself much and therefore little need to of different aspects of an organism’s breathe heavily’. So wrote Charles Darwin in performance. They concluded that the his early notebook ‘D’, according to David relationship between BMR and energy Quammen ( 2006 ). It is not unusual for Charles expenditure in the fi eld ‘may be fortuitous Darwin’s writings to encapsulate many later rather than direct’. We propose that the developments in biology. In the present case, relationship is indeed indirect, but certainly reference to ‘little need to breathe heavily’ not fortuitous. In the most general correctly anticipated the fi nding that the interpretation BMR refl ects the size and fl ightless kiwis of New Zealand have the maintenance cost of the metabolic machinery lowest BMR-values reported in birds (McNab (much like the fuel usage of an idling motor), 1994 , 1996 ); mention of ‘its small respiratory an idea fi rst explicitly formulated by our system’ refl ected the small size of fl ight research associate Marcel Kersten in 1985 muscles and other organs associated with ( Kersten and Piersma 1987 ). Depending on respiration ( McNab 1996 ); and the observation the ecological context, the machinery may be that ‘in the wild this must be a shy, patient, necessary for animals to achieve high rates of creepy little bird’ anticipated the general energy expenditure during strenuous exercise tendency for high BMRs to be associated with (requiring large pectoral muscles and hearts high levels of energy expenditure under fi eld and viscous blood) or to achieve the storage conditions ( Scholander et al. 1950c , Kersten of large amounts of energy rich fuel each day and Piersma 1987 , Daan et al. 1990b , (requiring large guts and livers). Thus, the Speakman et al. 2004a ). graphical model presented by Piersma et al. The rather strictly defi ned, standardized ( 1996a ; Fig. 45A ) to show how BMR and nature of BMR (the metabolism of inactive fi eld metabolic rate would be functionally homeotherms that are not digesting, coupled is incomplete. In refl ecting the size growing, moulting, ovulating or gestating and capacity of the metabolic machinery, under thermoneutral conditions) and the BMR does not determine a general ceiling, potential to correlate BMR with life-history but rather, by virtue of the size and capacity and biogeographic variables in ever more of particular organ groups, sets particular sophisticated ways, have made BMR kinds of ceiling (Fig. 45B ). Thus, similar levels immensely popular with physiologists and of BMR can refl ect the maintenance ecologists alike; according to the ISI Web of requirements of widely varying machinery, Knowledge, well over 1000 published papers not only between and within species, but have BMR as their topic. See Lasiewski and even within individuals (Piersma 2002 , Ksia˛ z˙ ek Dawson ( 1967 ), Weathers ( 1979 ), McNab et al . 2009 , Fig. 45B ). This led another ( 1986 , 1988 ), Elgar and Harvey ( 1987 ), research associate of ours, François Vézina, to Lovegrove ( 2000 ), Tieleman and Williams refl ect with a sigh that ‘BMR is a package ( 2000 ), White and Seymour ( 2004 , 2005 ), deal of so many different things’. Anderson and Jetz (2005 ) and White et al . Although incomplete because of this ( 2007 ) for examples. compartmentalization, the rationale of using METABOLIC CEILINGS: THE ECOLOGY OF PHYSIOLOGICAL RESTRAINT 77

HIGH CEILING A

LOW CEILING TR physiological scope + flight TR + flight forage forage

1.0 BMR BMR 0.7

'low gear' 'high gear'

DIFFERENT KINDS OF HIGH CEILING B

TR + A KIND OF LOW CEILING flight

TR TR + + flight flight

exercise exercise exercise nutritional organs nutritional organs nutritional rest rest rest BMR FMR BMR FMR BMR FMR

selection selection pressure pressure C Environmental variation Env. 1 Env. 2 Env. 3 exercise exercise organs exercise nutritional organs nutritional nutritional

on organ rest rest rest size and BMR BMR BMR capacity? on BMR?

Figure 45. An old (A) and a new (B) version of a diagram to demonstrate the possible functional link between the heights of metabolic ceilings and BMR, and a diagram of the question whether natural selection pressures work on BMR or rather the component parts of the metabolic machinery that require a certain BMR for maintenance (C). In (B) attention is given to the idea that BMR refl ects the summed contribution of size-related maintenance requirements of different organ systems (assuming invariable metabolic intensity). The vertical double-arrows in (A) and (B) indicate the ‘absolute scope’ for enhanced sustained performance given the size of the metabolic machinery. Panels (A) and (B) are based on Piersma ( 2002 ). continues 78 ADDING ENVIRONMENT

Box 3 Continued

BMR as a yardstick to evaluate or Diamond 1995 , Ksia˛ z˙ ek et al . 2004 , 2009 ) interspecifi cally scale maximal metabolic and (3) be selected for in the wild ( Broggi performances, seems nevertheless quite valid et al . 2005 ), BMR may be too much of an as long as the appropriate, context-specifi c, epiphenomenon to be a good trait for values for BMR are used ( McKechnie et al . natural selection to work on ( Fig. 45C ). We 2006 ). However, even though BMR may (1) suggest that natural selection fi rst and be repeatable within individuals ( Hayes et al . foremost targets the size and capacity of 1998 , Bech et al . 1999 , Horak et al . 2002 , different organs and organ groups, and that Labocha et al . 2004 , Rønning et al . 2005 , BMR will follow as a result (cf. Jackson and Vézina and Williams 2005 ), (2) have a Diamond 1996 , Williams and Tieleman 2000 , heritable component ( Konarzewski and Sadowska et al. 2008 ). CHAPTER 5 Phenotypic plasticity: matching phenotypes to environmental demands

Adaptive arm-waving, and more . . . enabled Kerry Marchinko ( 2003 ) to show that, when displaced to other fl ow regimes within Barnacles are peculiar crustaceans. Protected in Barkley Sound on the Pacifi c coast of Vancouver cone-shaped shells, fi rmly attached to rock and Island, they will quickly adjust the length and other hard substrates, they weather the pound- strength of their arms ( Fig. 46D ), even when ing of waves, the attacks by predators, the occa- mature. Thus, cirral nets appear fully reversi- sional desiccation that comes with living on the bly plastic throughout life. high shore and the high temperatures on sunny Their crustacean affi liation also shows when days. They feed on small planktonic items, it comes to sex and reproduction. Unlike many which they fi sh from the surrounding water by marine invertebrates that engage in so-called a ‘net’ made up of what their arthropod ances- ‘free spawning’—jettisoning sperm and eggs try has provided them with: six pairs of strongly in the water, where they fi nd each other by modifi ed arms. These arms are called cirri, and chance (Luttikhuizen et al. 2004 )—barnacles the fi shing net a cirral net ( Fig. 46A , B ). get intimate. Indeed, as documented by It is easy to imagine that, when the water is Charles Darwin ( 1854 ), at lengths up to eight calm, even wide and fragile nets will work, times body length, barnacles possess the long- whereas under strong water fl ows such nets est penises in the animal world (Neufeld and would be ripped to the side and become inef- Palmer 2008 ). This goes together with the fact fi cient at best, and damaged at worst. Several that most are hermaphrodites—combining species of intertidal barnacles indeed carry maleness with femaleness—but apparently long arms on sheltered shores, and short arms they do not do it on the cheap and self-fertilize. on exposed shores ( Fig. 46C ; Arsenault et al . Glued to the substrate, barnacles cannot cud- 2001 , Marchinko and Palmer 2003 ). At similar dle up; they have to use a long penis to fi nd, densities of edible plankton, high water fl ows reach, and then impregnate fastened but fertile will compensate for the smallness of the cirral neighbours ( Klepal 1990 , Murata et al . 2001 ). net. Under very exposed conditions, however, Now, what was true for the feeding arms form- the adaptive miniaturization comes to a halt ing the cirral net, would also be true for a penis ( Li and Denny 2004 ), as barnacles increasingly extending out of the father’s cone: at high begin to limit cirral netting to times of rela- water-fl ow rates, a long penis would be hard tively slow water fl ow ( Marchinko 2007 , Miller to manoeuvre properly. Building on the previ- 2007 ). Barnacles are quite easily dislodged ously mentioned barnacle work at Barkley from where they live and can be artifi cially Sound, Chris Neufeld and Richard Palmer glued on to other pieces of rock or shell. This (2008) indeed found that penises of Balanus

79 80 ADDING ENVIRONMENT

A B

C D 1 mm 5.2

4.8 0 days

4.4

4.0 7 days 3.6

3.2 length of the sixth arm (mm) 18 days 2.8

1 2345 maximum water velocity of 35 days breaking waves (m/s)

Figure 46. Barnacles Balanus glandula extend their arms to form cirral nets with which to withdraw small planktonic food items from the overlying water. Their cirral nets are cast wide in the calm waters of protected shores (A), but the arms are short in the waves of exposed shores (B), these examples representing opposing ends of a gradient (C). On the photos, in addition to folding out their cirral nets, the barnacles have also extended their considerable penises (see Fig. 47 ). Experimental displacement to protected shores from another protected (left) or a wave-beaten environment (right), shows the arms (in this case the sixth) to adjust adaptively within a couple of weeks. Compiled from Neufeld and Palmer ( 2008 ) (A, B), Arsenault et al . ( 2001 ) (C), and Marchinko ( 2003 ) (D). glandula became shorter and thicker at the base interfere with foraging (Klepal et al. 1972 , Hoch on the more exposed shores ( Fig. 47A , B ). For 2008 ). While barnacles do not shed sperm and reasons just outlined, the variations in penis eggs, rather they shed free-living larvae. At dimension made sense ( Fig. 47C , D). Perhaps temperate latitudes, where larval food supply not surprisingly, penis length and thickness may vary during the year (Yoder et al . 1993 , were also perfectly reversibly plastic (Fig. 47E , Dasgupta et al . 2009 ), the reproduction of bar- F). Not only do penis size and shape refl ect the nacles, it turns out, is also strongly seasonal. fl ow regime of the waters surrounding its Long before the reversibility of penis size was owner, in the Eastern Atlantic, lonely barna- examined, Margaret Barnes ( 1992 ) had con- cles have larger penises relative to crowded cluded that some barnacles get rid of their ones ( Hoch 2008 ). penises altogether outside the mating season. A big penis will come at a cost though, as it What is left during the ‘off season’ is a rather needs to be grown and maintained, and it may cost-free ‘stump’. PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 81

A 0.7 B 10.5 10.0 0.6 9.5

9.0 0.5 8.5 penis length (mm)

body size (soma mass) body size 8.0 penis basal width at standard

penis length (mm) at standard 0.4 0 12345 0 12345 maximum water velocity of breaking waves (m/s)

CDdirection of flow

protected-shore penis

reachable area exposed-shore penis

7.0 E F 0.36 6.5 0.34 6.0 0.32 5.5

5.0 0.30

0.28

4.5 penis length (mm) body size (soma mass) body size

penis basal width at standard 0.26 penis length (mm) at standard 4.0 protected exposed protected exposed shore shore shore shore

Figure 47. With increasing wave action, the penis of the barnacle Balanus glandula becomes shorter (A) and thicker at the base (B). This makes sense because the area within reach of the penis, and therefore the number of potential mates, is a function of its length and strength; in calm water, more area would be reached by a long and slender penis (C) and in strongly fl owing water, by a short and strong penis (D). Experimental displacements from either a protected shore (fi lled dots) or an exposed shore (open circles) shows penis length (E) and basal width (F) have mutually adjusted after 20 weeks. Means with SE. Compiled from Neufeld and Palmer ( 2008 ).

Use it or lose it during their time in space ( Lackner and DiZio Just as reproductively inactive barnacles do 2000 , Payne et al . 2007 ). Remember the sight of away with their penises, astronauts (who weakened astronauts returning from Apollo speak English) and cosmonauts (who speak missions on the TV screen? Without gravity Russian) quite unwillingly lose body parts a human environment is less demanding in 82 ADDING ENVIRONMENT

normal of the limbs, the effects of weightlessness are not only easier to predict on the basis of fi rst principles (Fig. 49A ), but the predictions are also easier to verify experimentally ( Fig. 49B , C ). Importantly, some muscle groups are wasted more than others (Vico et al . 1998 , Payne et al . Bed rest, some Excercise, 2007 ). Because they bear the body’s weight, the diseases, and pregnancy weightlessness ‘anti-gravity’ muscles of thigh and legs would be expected to show atrophy when they are pushed out of work during space fl ight or during experimental weightlessness on Earth ( Fig. 49 ); and they do ( Payne et al . 2007 ). Various types of immobilization experiments successfully mimic the gross effects of weightlessness in atrophy space (Fitts et al . 2001 , Adams et al . 2003 ). Muscles of the lower leg, assayed either by computed hypertrophy tomography or MRI scanning, showed losses, Figure 48. Visualization of the degree of atrophy of human which levelled off at 20% after 40 days (Fig. 50 ), hearts after several weeks of space travel or bed rest, as well with little evidence of differences between earth- as the degree of hypertrophy that comes with exercise and or space-bound microgravity effects. Despite pregnancy. the best attempts at giving them replacement Modifi ed after Hill and Olson ( 2008 ). exercise, after six months aboard the International Space Station, crew members still had lost 13% of their calf muscle volume and 32% of the peak several different ways and this shows upon power of their legs (Trappe et al . 2009 ). Not only the return to Earth. is muscle volume reduced during disuse, vari- One of the fi rst things to be affected is the ous metabolic changes occur that include a heart (Fig. 48 ), which may shrink by as much decreased capacity for fat oxidation. This can as a quarter after one week in orbit ( Hill and lead to the build-up of fat in atrophied muscle Olson 2008 ). Heart atrophy is correlated with ( Stein and Wade 2005 ). decreases in blood and stroke volume, blood In addition to muscle loss, and apart from a pressure, and reduced exercise capacity deterioration of proper immune functioning ( Convertino 1997 , Payne et al . 2007 ). Especially during space travel ( Sonnenfeld 2005 , Crucian after several months in the International Space et al. 2008), arguably the most fearsome effect on Station, and particularly in women, upon bodies is bone loss. Although their hardness and return to Earth, astro- and cosmonauts experi- strength, and the relative ease with which they ence dizziness and blacking-out because blood fossilize, give bones a reputation of permanence, does not reach the brain in suffi cient quantities bone is actually a living and remarkably fl exible ( Meck et al . 2001 , Waters et al . 2002 ). tissue. In the late-nineteenth century, the German Six weeks in bed leads to similar atrophy of anatomist Julius Wolff ( 1892 ) formulated that the heart as one week in space, suggesting that bones of healthy persons will adjust to the loads either weightlessness itself, or the concomitant that they are placed under ( Fig. 51A ). A decrease reduction in exercise, causes heart atrophy in load will lead to the loss of bone material at ( Perhonen et al . 2001 ). For the muscles and bones appropriate places, and an increase to thicker PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 83

A B C

normal gravity zero gravity

bed rest

Figure 49. How a change of gravitational forces (size of arrows indicates relative forces) affects the muscles and bones of the leg (A), and some of the contraptions used in human experiments on the effects of use and disuse on muscle and bone size, strength, and functioning (B, C). Diagram (B) shows how loads are taken off the lower limb without constraining the knee or ankle, whereas (C) shows how bed-rested women received compensatory treadmill running in a low-pressure chamber to mimic exercise for astronauts. Compiled from Lackner and DiZio ( 2000 ), Tesch et al . ( 2004 ), and Dorfman et al . ( 2007 ). bone ( Fig. 51B ). It is no surprise then that, in the 51C ). With considerable individual variation, gravity-free environments in space, bones will during half a year in space, cosmonauts lost up de-mineralize, especially the lower limbs that to a quarter of the tibial bone material ( Vico et al . normally counteract the effects of gravity (Fig. 2000 ). Although bone-formation continues under zero gravity conditions—this was established in chicken embryos sent to space (Holick space flight 1998 )—relative to bone-resorption, bone-forma- 25 immobilization on earth tion rates decline (Holick 2000 ). What has been 20 of greatest concern is that, unlike muscle loss that levels off with time ( Fig. 50 ), bone loss seems 15 to continue steadily with 1–2% a month of 10 weightlessness. During a 2–3 year mission to Mars, space travellers could lose up to 50% of 5 their bone material, which would make it impos-

decrease in leg muscle size (%) size decrease in leg muscle sible to return to Earth’s gravitational forces. 0 0 10 20 30 40 100 Thanks to Wolff’s Law, space agencies will have duration of unloading experiment (days) to become very creative in addressing the issue Figure 50. Relative loss of the lower leg muscles (triceps of bone loss during fl ights to Mars ( Lackner and surae, gastrocnemius, soleus, or some combination thereof) as DiZio 2000 , Whedon and Rambaut 2006 ). So far, a function of the duration of unloading either during space boneless creatures, such as jellyfi sh, are much fl ight or in immobilization experiments on Earth (this includes limb immobilization by casting, unilateral lower limb suspension more likely to be able to return safely to Earth as illustrated in Fig. 49 B, and bed rest). after multi-year space trips than people. Gravity Modifi ed from Adamset al . ( 2003 ). is a Mars bar. 84 ADDING ENVIRONMENT

Bone deposition –3000 compressive surface –2000 increased decreased –1000 0 strain strain neutral axis

microstrain 1000 2000 tensile surface Customary strain level

bone geometry decreased increased at baseline strain strain loading-induced adaptation Bone loss A B

2 0 –2 –4 –6 –8 –10 –20

differences vs initial values (%) vs initial values differences –22 –24 C before after 6-month 6-month flight recovery Figure 51. The use–disuse principle formulated as a functional adaptation model for bones (A), an illustration of the principle for an ulna of a rat (B), and a test of disuse–use effects in human cosmonauts (C). In the latter case, the loss of bone mineral density during a 6-month space fl ight, and the degree of recovery during the subsequent half year on Earth, were established. Compiled from Ruff et al . ( 2006 ) (A), Robling et al . ( 2002 ) (B), and Vico et al . ( 2000 ) (C).

Bone loss during space travel certainly The dynamic gut brings home the point that ‘if you don’t use it, you lose it’, but what about ‘if you use it It is anybody’s guess what happens to the again, do you regain it’? Within six months of innards of space travellers, but on the basis of their return to Earth, cosmonauts did indeed what we see in other vertebrates (Starck 2003 , show partial recovery of their tibial bone Naya et al . 2007 ), in the course of journeys mass (Fig. 51C ). Obviously most of them beyond the atmosphere, their digestive tracts needed more time, but even after one year of may well show change. To get a feel for what is recovery, bone losses in men who had been possible, let’s look at snakes, animals that go experimentally exposed to three months of through veritable ‘gastrointestinal rebirths’ permanent bed rest were not fully compen- ( Pennisi 2005 ). Pythons, boas, rattlesnakes, sated, though their calf muscles had fully and vipers are snakes that employ sit-and-wait recovered much earlier ( Rittweger and foraging strategies ( Greene 1997 ). This makes Felsenberg 2009 ). them go without meals for so long that their PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 85 stomachs, intestines, and accessory organs snakes and crocodiles belong ( Starck 2003 ), shrink and become ‘dormant’ (Secor et al . 1994 , but in all these groups, organ size and capacity Secor and Diamond 1995 , Starck and Beese are remarkably responsive to changes in 2001 , Ott and Secor 2007 ). When these snakes demand. We will elaborate examples from do catch the occasional prey, it can be almost birds later in this chapter. as big as themselves. Prey capture, killing by constriction, and swallowing through the ‘Classical’ phenotypic plasticity: extended gape (Fig. 52A ), elicits a burst of developmental reaction norms physiological activity, with drastic upregula- tion of many metabolic processes (reviewed in Let’s just briefl y think about space travel again. Secor 2008 ). Immediately, the heart starts to Although chicken embryos still show bone for- grow ( Fig. 52C ). With a doubling of the rate of mation when taken to space (Holick 1998 ), on heartbeat, and as blood is shunted away from the basis of what we now know it would be the muscles to the gut (Starck and Winner impossible for a human child to grow up under 2005, Starck 2009 ), blood fl ow to the gut zero gravity conditions and still survive the increases by an order of magnitude. Within functional demands of a return to Earth. two days, the wet mass of the intestine more ‘Developmental plasticity’ is the name for this than doubles (Fig. 52D ): this change is addi- category of phenotypic plasticity, the one in tional to growth, involves the rehydration of which environmental conditions during ontog- previously inactive cells, and the incorpora- eny determine the size, shape, construction, and tion of lipid droplets (J.M. Starck pers. comm.). behaviour of the eventual mature phenotype. This is followed by 100% increases of liver and There are innumerable examples of develop- pancreas mass, and a 70% increase of kidney mental phenotypic plasticity, and this exhilarat- mass (Fig. 52C ). At the same time, stomach ing variety of ways in which different aspects of acidity increases (a drop in pH from values of environments shape different organismal traits 7–8 to 1–2), ensuring that the skeleton of the is illustrated in Fig. 53 . Picking up the barnacle ingested prey disappears within about six theme from earlier in this chapter, some of the days (Fig. 52B ). Many amazing changes, at intertidal acorn barnacles Chthamalus anisopoma several different levels of physiological action, develop a bent shell when grown in the pres- take place in the intestine, including the instan- ence of predatory snails; this form is indeed taneous lengthening of microvilli ( Fig. 52E ). more resistant to predation than the normal All in all, the stomach operates for less than morph ( Lively 1986 , 1999 , Lively et al . 2000 , a week, but the intestines carry on for another Jarrett 2008 ). week to fi nish the job. Interestingly, part of the A classic case of developmental phenotypic winding down of the gut includes the genera- plasticity is the variation in the degree to which tion and banking of new cells lining the intes- water fl eas Daphnia show hoods, helmets, tine, thus preparing the pythons for the next spines, and longer tails in response to preda- meal ( Starck and Beese 2001 ). And what diges- tors, or to the chemical identifi ers (the ‘kai- tively goes for pythons also goes for another romones’) that they leave in the water (e.g. reptile, the caiman Caiman latirostris from Tollrian and Dodson 1999 , Caramujo and South America (Starck et al . 2007 ). The precise Boavida 2000 , Petrusek et al . 2009 , Riessen and mechanism of the restructuring of the gut and Trevett-Smith 2009 ). This variation was fi rst associated organs may differ between mam- described by the enlightened German zoolo- mals, birds, and the sauropsid reptiles to which gist Richard Woltereck ( 1909 ) in clones of water 86 ADDING ENVIRONMENT

1 day postfeeding 5 heart 1.4 pancreas

1.2 4 1.0 3 3 days postfeeding 0.8

2 0.6

20 liver 8 kidneys wet mass (g) wet 6 days postfeeding 16 6 12 B 8 4

fasted 0 5 10 15 015 015 days postfeeding C

2 days postfeeding

10 days postfeeding

D E 01 3 6 15 days postfeeding Figure 52. After a Burmese python Python molurus has captured a prey item, killed it by asphyxation, and started swallowing it (A), much of the snake’s internal physiology gears into action. (B) Shows how the bones of a rat are digested in the acidic stomach within 6 days post-feeding. In the fi rst few days after ingestion, the internal organs show growth spurts and then a slow return to the old levels. Within 2–3 days, the intestine has grown to full capacity (D), including a drastic lengthening of the microvilli (E, the bars in this image represent 1 μm). Compiled from Lignot et al . ( 2005 ) and Secor ( 2008 ). PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 87 fl eas (see Harwood 1996 , Sarkar 2004 ). Finding ENVIRONMENTAL GRADIENT that different pure strains of Daphnia show similar variation in hood size, variation that converged under particular environmental conditions, Woltereck began to doubt the dis- without predatory snailwith predatory snail tinction between genotype and phenotype, a intertidal barnacle contrast that was just becoming part of the Chthamalus anisopoma mental armoury of biologists at the time. Woltereck was straightened out by the Danish plant physiologist and budding geneticist waterflea without predatory with predatory Daphnia pulex Wilhelm Johannsen ( 1911 ). Johannsen sug- Diptera larvae Diptera larvae gested, fi rst in two books in Danish and then in an address to the American Society of Naturalists in 1910 (with which he reached the world), that what genotypes do is not to

instruct for a single phenotype, but to instruct of spider web for the ways that phenotypes respond to envi- bistriata Parawixia ronmental variation ( Roll-Hansen 1979 ). This normal, with small prey during termite swarms, large prey relationship between trait values and specifi c environmental conditions has since been called what Woltereck named it: ‘Reaktionsnorm’ or ‘reaction norm’ or ‘norm of reaction’ ( Pigliucci

2001a , 2001b ). The norm of reaction of barna- sea urchin cles to predatory snails, and of water fl eas to Diadema antillarum at low density at high density predatory insects or fi sh, is now also known as an example of the category of ‘induced defences’ ( Tollrian and Harvell 1999 ). Figure 53 shows fi ve more reaction norms refl ecting several other types of developmen- scleractinian coral scleractinian

tal plasticity. Spiders weaving a coarser damicornis Pocillopora grained and bigger web when they can catch at high water flowat low water flow large prey (in this case fl ighted termites), could be said to show ‘induced offence’ (cf. Padilla 2001 , Kopp and Tollrian 2003 , Miner et al . 2005 , Kishida et al . 2009 ). Caribbean sea buttercup urchins Diadema antillarum have a hard calcite heterophylly density dependence induced offence induced defence

Ranunculus flabellarisRanunculus on land in water

Figure 53. Examples of phenotypic plasticity in invertebrate animals and in plants, along with some of the common denominators of the particular class of plasticity. seasonal butterfly

Compiled from (from top to bottom): Lively ( 1999 ), Lampert et al . polyphenism Precis almana ( 1994 ), Sandoval ( 1994 ), Levitan ( 1989 ), Kaandorp ( 1999 ), Schlichting and Pigliucci ( 1998 , p. 56), and Nijhout ( 1999 ). in dry season in wet season 88 ADDING ENVIRONMENT

Table 6. A summary of content and scope of some recent books dedicated to the phenomena of phenotypic plasticity.

Reference Book title No. of pages Content and scope

Rollo (1995) Phenotypes, their epigenetics, 463 An early, but very wide-ranging, synthesis ecology and evolution by a molecular biologist of the genetic background and possible hierarchical organization of phenotypic construction, or organismal design, with due recognition of the role of ecological factors. Schlichting and Pigliucci (1998) Phenotypic variation: a 387 The textbook on phenotypic evolution from reaction norm perspective combined genetic and developmental perspectives. Well-illustrated. Norms of reaction are the core idea to explore the basics and evolutionary consequences of phenotypic responses to environmental variation. Tollrian and Harvell (eds) (1999) The ecology and evolution of 383 Explores and reviews a wide range of inducible defenses cases of developmental plasticity that fall under the banner of ‘induced defence mechanisms’: from the use of armour, via unpalatability, to immunity and behaviour as environmentally induced mechanisms for bodily defence. Pigliucci (2001a) Phenotypic plasticity: beyond 328 A solid attempt to develop mainly genetic nature and nurture interpretations of phenotypic plasticity as the core concepts in evolutionary biology. West-Eberhard (2003) Developmental plasticity and 794 Almost encyclopaedic review of the evolution importance of developmental processes for evolutionary innovation. Develops the notion that in evolution phenotypes change fi rst, and that genes follow. DeWitt and Scheiner (eds) (2004) Phenotypic plasticity: 247 With a loose focus on developmental functional and conceptual processes and their genetic approaches. interpretation (reaction norms), the chapters explore a wide range of issues to do with phenotypic plasticity. Turner (2007) The tinkerer’s accomplice: how 282 Formulates the role of self-organization, in design emerges from life interaction with replicators on one side itself and the environment on the other, in phenotypic expressions. It develops the idea that organisms show design, not because their genes instruct them to, but because agents of homeostasis build them that way. Blumberg (2009) Freaks of nature: what 326 This is mostly a book about development, anomalies tell us about but one in which the importance of development and evolution interactions between developing bodies and their environments is strongly emphasized. PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 89

Gilbert and Epel (2009) Ecological developmental 480 Fantastically comprehensive and biology: integrating well-illustrated textbook on the ways epigenetics, medicine, and that environments affect development evolution of animals, and why this knowledge is important in medicine and evolutionary biology.

skeleton but are able to shift body size in phenotypic plasticity is the size-, age-, condi- response to variations in local food condition-, and/or context-dependent sex change tions. Levitan ( 1989 ) experimentally altered that occurs in some plants, and in animals as competitor densities, and showed that body diverse as annelids, echinoderms, crustaceans, size increased and decreased to levels pre- molluscs, and, most famously, in fi sh dicted from fi eld-based food abundance– ( Polikansky 1982 , Munday et al. 2005 , 2006 ). body size relationships. By adjusting All the ‘textbook’ examples amalgamated in structural body size, the urchins reduced Fig. 53 represent relatively well-studied cases maintenance costs and thereby optimized of phenotypic plasticity. Some, such as the reproduction and survival according to local water fl eas of Woltereck and subsequent food availability ( Levitan 1989 ). Although authors, have been important in the develop- body size is altered, the mouth structures ment of quantitative genetic-related theories (Aristotle’s lantern) of the urchins, and thus of phenotypic plasticity (Pigliucci 2001a ). What their capacity to eat, remained unchanged are missing from this list of examples are the ( Levitan 1991 ). extensive, but usually reversible, internal and Depending on the type of water fl ow, scler- external changes that occur in vertebrate bod- actinian corals build either denser or more ies, changes that have played major roles in spaced out structures ( Kaandorp 1999 , Todd the previous and in the present chapter, but 2008 , who still owe us a term for this category are almost absent from the burgeoning litera- of phenotypic plasticity). Buttercups ture on phenotypic plasticity (see Table 6 ). Ranunculus have ‘proper’ wide leaves when With a notable exception ( Turner 2007 ), the growing on land and almost fi lamentous phenotypic plasticity literature has mostly leaves when growing in water, representing a passed by the staggeringly large body of phe- category of plant plasticity called heterophylly nomena related to internal (physiological), (Cook and Johnson 1968 ). The last example in external, and seasonally cyclic (moult) and Fig. 53 is quite typical of the insect world. behavioural plasticity, especially in the higher Depending on season, a tropical butterfl y vertebrates. To make sure that readers will Precis almana may either show angular wing have encountered the full range of plasticity shapes and a dull brown colour, so that it phenomena by the end of this chapter, let’s resembles a dead leaf (in the dry season), or now look at the reaction norms of male and show rounded and colourful wings with eye- female rock ptarmigan Lagopus mutus , show- spots (in the wet season) (Nijhout 1999 ). This ing phenotypes that vary with time of year, category of phenotypic plasticity is known as snow cover, and, in the case of males, the avail- (seasonal) polyphenism. A fi nal category of ability of mates. 90 ADDING ENVIRONMENT

Seasonal phenotype changes in hunting gyrfalcons Falco rusticolus . When oppor- ptarmigan, deer, and butterfl ies tunities for mating disappear, the male ptarmigan camoufl age themselves by soiling their Rock ptarmigan are a species of grouse that lives white plumage and so are dirtiest at the start of on the arctic tundra. For most of the year the tun- incubation (Montgomerie et al. 2001 ). Only when dra is covered by snow, and both sexes are cam- a clutch is lost to predators and their mates oufl aged by white plumage (Montgomerie et al . become fertile again, do the males clean up. They 2001 ). In the brief arctic spring, when the snow eventually moult into a cryptic summer plum- melts and is replaced by summer tundra, the age once the season with mating opportunities females moult their white plumage in favour of really has ended. Thus, the visual aspect of the green and brown feathers (Fig. 54 ). However, the plumage in females is determined by its role for males do not immediately moult their white camoufl age only. In males, plumage not only feathers and are exceptionally conspicuous for serves the dual role of camoufl age in winter some time, being attractive not only to prospect- (when white) and in late summer (when coming females, but also easily discovered targets for prising a cryptic patterning of browns), but also

males

females

D J annual c N ete yc F pl le m o O c M a

S A

A M JJ

Figure 54. Annual cycle of the external appearance of male and female rock ptarmigan in the Canadian Arctic in relation to changes in the camoufl age afforded by the tundra habitat (indicated by the texture of the middle circle). This illustrates the power with which such intra-individual changes can be interpreted in a functional context. The changes in plumage represent a change in feathers, except for the change from a pure white to a dirty white colour of males in late June–July, which is the result of soiling. From Piersma and Drent ( 2003 ) based partly on data from Montgomerie et al . ( 2001 ). PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 91

Figure 55. Variation in antler size and shape of extant and extinct ‘true’ deer Cervidae. Compiled from Emlen ( 2008 ). the sexually selected, ornamental role (white) in Emlen 2008 ). What makes deer antlers so differ- the mating season. The soiling of the white plum- ent from functionally similar horns and tusks in age when mating opportunities are no longer other ungulates is that they are as seasonal as the present (but when the moult into a cryptic plum- plumages of ptarmigan. Antlers fi rst grow at age has not yet begun) provides a behavioural puberty, and successive sets of antlers developed ‘quick fi x’ to a critical phenotypic allocation by individuals tend to increase in size and com- problem. plexity (Goss 1983 , Bubenik and Bubenik 1990 , Showing off is especially important in males Lincoln 1992 ). Whereas the small forest deer of that are in a position to monopolize more than the tropics may regenerate antlers at any time of one female during their fertile periods, and this the year, at temperate latitudes deer show syn- is what deer do, at least as long as they live in chronized seasonal cycles of growth, calcifi ca- semi-open habitats. The true deer (Cervidae) are tion, cleaning, casting, and regeneration (Lincoln unique among animals in having males that 1992 , Goss 1995 ). Regeneration of antlers occurs show temporary bony outgrowths on their outside the reproductive season, in a period heads, the antlers (Fig. 55 ). 1 Antlers are ornamen- of low testosterone, through a process that is tal weapons that should impress both sexual different from wound healing in that it is a stem competitors and prospective mates (Geist 1998 , cell based process ( Kierdorf et al. 2007 ). Growing

1 In reindeer and caribou Rangifer tarandus, females also have antlers, as they have to compete with the males and amongst each other for scarce fodder under arctic winter snow ( Lincoln 1992 ). 92 ADDING ENVIRONMENT antlers are covered by a ‘velvet skin’, which is shed when the underlying bone calcifi es and dies, yielding the insensitive antlers that can be used as weapons during the rut. Hard antlers DRY WET remain attached to the skull as long as the stags 25 have raised testosterone levels (3–9 months, 250 depending on species), after which the connec- 20 200 C) tion is severed and the antlers are cast (Lincoln ° 1992 ). Overall, the phenotype of individual male 15 150 deer is as seasonally variable as it is in male 10 100 ptarmigan. Butterfl ies only live a couple of months. 5 50 (mm) mean rainfall However, just like ptarmigan and most other ( mean temperature 0 0 organisms, butterfl ies experience seasonal envi- J JASONDJFMA M ronments and, in response, many butterfl y spe- Developmental plasticity cies show seasonal polyphenism: within single genotypes, but in successive generations and Hormones in Small eyespots larva/pupa Large eyespots, depending on the time of year, they may show Cryptic wing More conspicuous pattern wing pattern strikingly different phenotypes ( Kingsolver Large body size Small body size 1995a , 1995b , Nijhout 1999 ). Well-studied exam- Dry season Wet season ples are the Bicyclus butterfl ies from East Africa. phenotype phenotype Like other tropical butterfl ies, they are exposed Long life Short life to an alternation of dry and wet seasons. From Low metabolism High metabolism Thrifty/quiescent Active behaviour June to October it is dry and mostly relatively behaviour Hormones in Fast reproduction cold, but from November through April it is wet Delayed reprod. adult and initially quite hot (Fig. 56 ). In the dry season, the butterfl ies are large and cryptically col- Adult acclimatization oured. During the wet season, they show Figure 56. Dry and wet seasonality in Malawi, East Africa, and abundant eye-spots and a somewhat smaller the phenotypic responses by Bicyclus anynama butterfl ies that body size (Windig et al . 1994 , Brakefi eld et al . have one generation of inconspicuous large morphs in the dry 2007 ). The alternative morphologies can be gen- season, and two generations of more conspicuous and smaller erated from split broods reared at either high or morphs in the wet season. Larvae of the last wet-season cohort develop with declining temperatures and produce the cohort of low temperatures ( Kooi and Brakefi eld 1999 ). dry-season morphs without eye-spots. The diagram illustrates The butterfl ies of the dry season emerge around that this seasonal polyphenism is steered by both developmental the end of the rains and have to survive almost plasticity (leading to the conditional metamorphosis induced in half a year before they can lay eggs on fresh the larvae and expressed at pupation) and adult acclimatization green grass at the start of the next wet season. To (maternal, e.g. hormonal, contributions to the eggs laid in whatever season). survive that long, they rely on their cryptic col- Compiled from Brakefi eld and Reitsma ( 1991 ), and Brakefi eldet al . ouration and a quiescent lifestyle, including low ( 2007 ). metabolic rates (and a long life span under laboratory conditions). The ornamented butterfl ies of the wet season come in two generations. The producing a slightly longer lived cohort, that fi rst, which hatch from eggs laid by their cryptic then produces the cryptic cohort that endures parents, immediately mate and lay again, thus the dry conditions, and so on. The wet season PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 93 butterfl ies show the distinct eye-spots; these active phenotype ( Murton and Westwood may thwart attacking birds ( Kodandaramaiah 1977 ). To perform optimally under a wide et al. 2009 ) and are important during mate choice range of environmental conditions (variations ( Constanzo and Monteiro 2007 , Oliver et al . that are often cyclic), a long-lived individual 2009 ). The colourful wet season butterfl ies are must track or anticipate the external changes also much more active, have a higher metabo- by regulating gene expression to adjust its lism and a shorter life-span (even when not morphology, physiology, and behaviour exposed to predation in the laboratory). A won- ( Willmer et al . 2000 ). The cyclically varying derful combination of fi eld and laboratory phenotypic expressions of these adjustments research enabled Paul Brakefi eld and co-work- within an individual rock ptarmigan, for ers (2007) to establish that the seasonal plasticity example, are thus called ‘life-cycle stages’ shown by Bicyclus butterfl ies represents aspects (Jacobs and Wingfi eld 2000 , Piersma 2002 , that are typical of both ‘classical’ developmental Ricklefs and Wikelski 2002 ). The seasonal plasticity (hormonal effects in larva and pupa) template for such sequences might be pro- and physiological acclimatization (induced in vided by the natural photoperiodic rhythm the reproductive adult stage and maternally and/or by an endogenous circannual pace- contributed to the eggs) ( Fig. 56 ). maker ( Gwinner 1986 ). In addition, tempera- The extents of phenotypic change shown by ture, rainfall, food, or densities of conspecifi cs ptarmigan and butterfl ies are rather similar, might give supplementary information, which but their life spans (and life-cycle characteris- individuals could use to ‘fi ne-tune’ the timing tics) are not. Individual ptarmigan may expe- of their phenotypic transformations (Wingfi eld rience many seasons, individual butterfl ies a and Kenagy 1991 ). The polyphenisms of short- single season at best. What is called seasonal lived organisms, as in many insects, refl ect a polyphenism in insects is called life-cycle stag- similar strategy ( Shapiro 1976 , Danks 1999 ). ing in birds and mammals. And whereas the The accuracy with which future environ- genetic-developmental architectures behind mental conditions can be predicted would the phenotypic expressions are likely to be determine the kind of phenotypic plasticity very different in birds and butterfl ies, in both that one might expect to evolve ( Levins 1968 , groups seasonally varying ecological shaping Moran 1992 , Padilla and Adolph 1996 ). In factors to do with feeding, the avoidance of unpredictable environments, where there is predation, and mating, can functionally insuffi cient environmental information, or if explain their strikingly variable phenotypes. the wrong environmental information is used, an organism might end up with a phenotype Environmental variability and that does not quite match its current environ- predictability, and the kinds of ment (Hoffman 1978 ). Indeed, such mismatches phenotypic adjustments that make are considered to be an important cost of devel- sense opmental plasticity (see Table 7 ). However, this potential cost would disappear for organisms Different activities related to reproduction and that are capable of fast and reversible pheno- survival (breeding, moult, migration, hiberna- typic change. We will get back to this issue after tion, etc.) are usually separated in time within we have addressed ‘trade-offs’. individuals, tend to occur at predictable times Carrying on from our introductory classifi - of the year in seasonal environments, and are cation in Chapter 1 ( Table 1 ), Fig. 57 represents accompanied by changes in the reproductively the best attempt at present to place the different 94 ADDING ENVIRONMENT

Table 7. Inventory of costs of, and limits to, phenotypic classes of phenotypic plasticity within a two- plasticity. Developed from DeWitt et al . ( 1998 ), Pigliucci ( 2001a ), dimensional space, bounded by axes that rep- Lessells ( 2007 ), Valladares et al . ( 2007 ), and Vézina et al . ( 2010 ). resent (1) the degree of environmental Costs of plasticity Limits to plasticity variability (relative to the organism concerned) and (2) the degree to which this variability is Maintenance: energetic Phylogenetic ‘constraint’: however costs of sensory and adaptive a kind of plasticity predictable (Drent 2004). Highly predictable regulatory mechanisms, would be, in any organism changing environments would select for and of any additional some forms of plasticity are polyphenism in short-lived organisms and life- structures. impossible because of historic cycle staging in long-lived organisms. The lower genetic, developmental or the predictability of environmental variation, sensory constraints. Production: excess costs of Damage-related constraints : in the better it is for organisms to respond oppor- producing structures plants, phenotypic plasticity tunistically, rather than seasonally scheduled. plastically rather than as may be limited by herbivory Developmental plasticity would then describe fi xed genetic responses. because grazing damage the kind of variable responses: organisms prevents achievement of the encountering unpredictably variable environ- optimal phenotype. Information acquisition : Phenotypic compromise : resource ments in the course of their life would benefi t investments (risk, time, constraints may bias functional from plasticity being reversible, i.e. showing energy) for sampling the adjustment to one trait at the phenotypic fl exibility . If environmental variation environment, including cost of another. cannot be predicted, organisms might go into lost opportunities (e.g. bet hedging (generating differently adaptive mating, foraging). phenotypes at random) in shorter-lived organ- Developmental instability : Information reliability : the phenotypic imprecision environmental cues may be isms. In theory at least, especially longer-lived may be inherent for unreliable or changing too animals could also cope with extravagant environmentally contingent rapidly. changeability of the environment by not adjust- development, which can ing the phenotype at all, i.e. show robustness . result in reduced fi tness under stabilizing selection. Genetic : due to deleterious Lag time : the response may start between within individuals/generations individuals/generations effects of plasticity genes too late compared with the through linkage, time schedule of the

HIGH polyphenism life-cycle pleiotropy, or epistasis environmental change, leading staging with other genes. to maladaptive change. Developmental range: plastic genotypes may not be able to developmental phenotypic express as wide a range of plasticity flexibility/ acclimatization adaptive phenotypes as a predictability polytypic population of specialists would, e.g. because of limitations in the workings of bet-hedging robustness

neuroendocrine control system. LOW Epiphenotype problem : the LOW variability HIGH plastic responses could have Figure 57. Different classes of phenotypic plasticity (or the evolved recently and still lack of it in the case of robustness) arranged according to an function like an ‘add-on’ to the axis of environmental variability over time and an axis of basic developmental environmental predictability over time. The categories on the machinery rather than as an left show plasticity between individuals (and thus, between integrated unit; this may generations), and on the right within individuals. compromise their performance. Inspired by Meyers and Bull ( 2002 ) and adapted from Drent (2004). PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 95

Degrees of fl exibility Cerebral cortex as old as you are In the previous chapter we encountered sled Visual cortex Cerebellum as old as you are slightly younger dogs that, along with human polar explorers, than you are were the true champions of endurance exer- Teeth as old as the time cise. Arctic sled dogs hard at work would be since they grew expected to show many physiological and internal morphological changes compared to Intercostal muscle 15.1 years Heart muscle cells sled dogs on holiday. In summer, the absence 6 years yonger of snow and ice puts sled dogs out of work. than you are Nadine Gerth and co-workers (2009) used the contrast between summer and winter for a Fat cells comparison of muscle size and ultrastructure 9.8 years in Greenlandic sled dogs. During the summer Gut (not the lining) holiday, dogs had thinner muscle, the slim- 15.9 years Gut epithelium ming representing changes in mitochondrial 5 days numbers, lipid droplet sizes, and numbers of contractile myofi brillar elements. Gerth et al . (2009 ) were surprised that capillary networks Skin (outer layer) 2 weeks in the muscles did not change with season, Red blood cells and did not explain it. Could there be inherent 120 days differences between tissue types in the degree to which they can adjust to changing func- Bone tional demands (Lindstedt and Jones 1987 )? 10 years And what measure would best refl ect a capacity for change? Cell turnover perhaps? A fruitful approach to look at cell turnover in different tissues, especially the longer-lived ones, was the serendipitous use of peaks in 14 C in the atmosphere generated by nuclear bomb Figure 58. Most of your cells are younger than you are (at 14 least the DNA in the cells is), but there is a lot of variation testing in the 1960s. This C is incorporated between cell types. into body tissues. Spalding et al . ( 2005a ) used Modifi ed from Vince ( 2006 ), partly based on Spaldinget al . ( 2005a , DNA from stored tissues to estimate their age 2005b 2008 ) and Bergmann et al . ( 2009 ). at the moment of sampling (often after death). They verifi ed the method with the enamel of that we discussed earlier. However, we have teeth that, once formed, are not rebuilt (Spalding also shown that hearts show fast changes in et al . 2005b ). Whereas the lining of the gut and size in response to changes in workload, so the outer skin show fast DNA turnover, indi- how does that tally with the cardiac muscle cating fast cell-replacement (Fig. 58 ), the DNA cells living so long (Bergmann et al . 2009 )? The in fat cells, muscle cells, and bone cells is not discrepancy is due to cell components other replaced in many years. The oldest tissues in than the nucleus being built and broken-down, any person are the heart (cardiac muscle cells) rather than whole cells being renewed and and various components of the brain. removed. Myofi brillar volume may change, The high turnover of gut epithelium tallies and the densities and total volume of mito- well with the high fl exibility of the intestine chondria may quickly change too ( Hoppeler 96 ADDING ENVIRONMENT

and Flück 2002 ). Note that the brain is mainly these tissues, show greater and faster responses composed of very old cells ( Fig. 58 ). Brains to changes in demand than others. thus beautifully combine permanence (constancy of cell numbers) with change (new neu- Direct costs and benefi ts, their trade- ral connections made all the time) ( Edelman offs, and other layers of constraint 2004 ). A study on Andean toads Bufo spinulosus sug- Do you still remember that a big penis might gests that there may be generalizable, tissue-spe- hinder the waving of cirral nets to catch plank- cifi c, capacities to respond to changing conditions. tonic food items by barnacles ( Hoch 2008 )? Daniel Naya and co-workers (2009) compared Then here is a somewhat similar example, the sizes of various organs as a function of sea- where the trade-off between benefi ts and costs son (comparing summer, when the animals are of a sizeable reproductive organ has been most active, with winter) and their feeding state mapped in much more detail. Enter arachnids, (summer animals were exposed to a two-week or spiders. In most species, males are smaller fast or fed). The degree of hypertrophy of the than females, but in some species, males are active over the inactive stage in the two situa- dwarfs (Vollrath 1998 ). This might be because tions was clearly correlated (Fig. 59 ): heart, lungs, such small males can use ballooning on silky and the large intestine never showed signifi cant threads to fi nd distant females, or because they change, but stomach, small intestine, and liver would be no threat or competitor for the showed great change in both contexts. The highly females, which would ease the access to what, fl exible small intestine fi ts our expectations, as in spiders, is the discerning and ruling sex. In does the infl exible lung (see Chapter 3 ). cobweb spiders of the genus Tidarren, the males Nevertheless, we seem far from clearly under- are so tiny, about 1% of the female’s mass, that standing why some tissues, or components of it raises ‘a delicate problem’ ( Vollrath 1998 ): if the mating organs, a paired structure with the name pedipalps (which they fi ll with the sperm

no significant change re. food supply significant change re. food supply to be pumped into the female), have shrunk in

stomach proportion to the body, they would no longer fi t snugly—key to lock—into the female’s fat bodies small intestine

re. season re. receptive organ, the epigyne. Tidarren spiders

liver significant change have solved this problem by the tiny males carrying relatively massive copulatory organs, to kidneys the extent that this makes the spiders pretty heart clumsy and reduces their chances in the scram- (summer over winter) (summer over

re. season re. ble to get to a fertile female before others degree of change re. season of change re. degree large intestine lungs

no significant change (Ramos et al. 2004 ). The tiny Tidarren males have a radical solution, as they actively break degree of change related to feeding state (feeding or fasting) off the carrying arm of one pedipalp before fi lling it with sperm and before they approach a Figure 59. Degree of change in the dry mass of various female. The loss of one pedipalp speeds them organs of mature Andean toads Bufo spinulosus as a function up considerably, and also makes them more of either feeding and fasting (X-axis) or season (summer compared with winter, Y-axis). The degrees of change refl ect the enduring athletes ( Fig. 60 ). This partial emas- fi rst two principal components in a factor analysis. culation has clear dividends and may not even Based on Naya et al . ( 2009 ). be as costly as it would seem at fi rst sight. PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 97

Female Tidarren tend to consume males they developing such horns, Leigh Simmons and have mated with and mostly give them the Douglas Emlen (2006) could show that the chance to insert only a single pedipalp. Still, growth of these horns actually occurred at the these spiders may trade insemination capacity expense of an organ that also helps to secure for speed and endurance, providing an exam- progeny—large testes. All the forms of pheno- ple of a trade-off that leads to the severing of a typic plasticity embody such trade-offs. Theory body part that gets in the way at some stage. has it that for phenotypic plasticity to evolve, Trade-offs are about alternative investments, no single phenotype can be optimal in all envi- the allocation problem of where, and in which ronments experienced by the organism (Via parts of the phenotype, to invest scarce and Lande 1985 , Moran 1992 ). Thus, there must resources (e.g. Stearns 1989 , 1992 , Gervasi and be trade-offs (and reliable cues to inform the Foufopoulos 2008 ). Many dung beetle males organism about the state of its variable envi- develop large horns on the head or on the ronment). Investments in one direction neces- thorax, with which they compete for females. sarily, due to time, energy, and nutritional In a manipulative study where larvae of limitations, prevent investments in another Onthophagus nigriventris were prevented from direction (for a model and a test with Daphnia , see Hammill et al . 2008 ). Only fantasy creatures called ‘Darwinian demons’ would be able to respond continuously and at many levels to changes in trade-offs by adjusting phenotype. Dung beetles that are able to predict that they A B can monopolize the female once they have secured it, would invest in horns; those that 6 predict they would not, would play the sperm- 5 4 competition game and invest in testes. The 3 Darwinian demonic dung beetles would be 2 able continuously to adjust the relative sizes of 1 their horns and testes in response to fl uctuat- maximum speed (cm/s) maximum 0 ing ecological and social environments. 2000 Although dung beetles show considerable

1500 developmental versatility ( Emlen et al . 2007 , Rowland and Emlen 2009 ), Darwinian demonic 1000 dung beetles do not exist. There are always 500 limits to plasticity.

time until exhaustion (s) time until exhaustion 0 This is all we might have said on trade-offs two palps one palp and phenotypic plasticity, except that in the Figure 60. Minute males of the cobweb spider Tidarren pertinent literature there exists another layer sisyphoides before (A) and after (B) removal of one half of the of concern: the trade-off between being pheno- paired ejaculatory organs, the pedipalps. The pedipalps are typically plastic, or not being plastic at all indicated by the two (A) or single (B) arrows; the scale bar ( Newman 1992 , Via et al . 1995 , DeWitt et al. represents 1 mm. Box plots of maximum speed before and after 1998 , Pigliucci 2001a ); this is related to the idea pedipalp removal (n = 16) and time till exhaustion in two groups of spiders still having two pedipalps (n = 15), or a single that plasticity itself is a trait with a genetic (n = 17) one, are shown below the photos. basis ( Relyea 2005 ). Perhaps there are true Compiled from Ramos et al . ( 2004 ). costs to plasticity itself (Table 7 ). Such costs 98 ADDING ENVIRONMENT could include the energetic costs of the main- However, the last two limits-categories are tenance of sensory and regulatory structures somewhat controversial. As Massimo Pigliucci enabling plasticity, or the excess costs of pro- ( 2001a ) points out, for an intra-individual ducing a plastic rather than a fi xed (robust) phenomenon it is odd to juxtapose individual structure. Adaptive plasticity makes it neces- constraint with population variation, as sary to acquire information about the environ- DeWitt et al. (1998 ) did in their formulation of ment, and such acquisition will cost time. a ‘developmental-range’ limit. We left it in Plastic structures may (or may not) be more Table 7 because it is quite possible that control unstable than developmentally fi xed struc- mechanisms (e.g. of a neuroendocrine nature) tures, and if there were to be such things as limit the way in which environmental infor- plasticity genes, in some combinations with mation is fed into the functionally adjusting other genes they could be deleterious. In a phenotype ( Lessells 2007 ). review, Pigliucci ( 2001a ) could not come up The ‘epiphenotype problem’, the issue that with much evidence in favour of any such plastic responses may be recent add-ons that costs. But perhaps all these theoretically pos- are not yet genetically and developmentally sible costs of plasticity are not as far-fetched as hard-wired and may therefore compromise they might at fi rst sight appear. In the rstfi spe- performance, is interesting on two accounts. cifi c study on the costs of plasticity (Relyea First, it is quite possible that many forms of 2002a ), evidence for such costs appears ‘perva- plasticity represent self-organized rather than sive’. Comparing sibships of larval frogs that developmentally tightly orchestrated struc- possessed different degrees of phenotypic tures (Turner 2007 ). We shall get back to that in plasticity, depending on the trait examined a minute. Second, ‘a plastic response may start and the environment experienced, increased as an epiphenotype and subsequently be inte- plasticity could either have positive, negative, grated by natural selection attempting to or no effects on fi tness components, such as reduce the costs and limits of the new response’ degrees of trait development, mass gains, and ( Pigliucci 2001a , p. 176). In our fi nal chapter survivorship. Thus, now that the appropriate we will return to plasticity as a creative force studies are being done (Lind and Johansson in evolution ( West-Eberhard 2003 ). In a cele- 2009 ), there is increasing evidence that plastic- brated class of developmental plasticity that ity per se may come at a cost. so far has escaped notice, all these elements of The mere fact that Darwinian demons do costs, benefi ts, and limits nicely combine. Let’s not exist, actually demonstrates that there are look now at frogs and newts, and especially limits to plasticity, and we provide a list of their tadpoles! possibilities in Table 7 . Let us quickly run through them. Phylogenetic constraints are Phenotypes of fear obvious, damage through herbivory may prevent plants from achieving locally optimum That amphibians live up to their name, that phenotypes (Valladares et al. 2007 ), organisms they never fully made the transition from living may have to compromise between different in water to living on land, actually gives them a adjustments in the light of resource limita- lot of ecological opportunity. By their enormous tions ( Vézina et al . 2010 ), environmental infor- life-history versatility, they can capitalize on the mation is bound to be unreliable at times, and resources offered by ponds, pools, and puddles, lag-times between environmental changes and even the underlying soil ( Newman 1992 ). and phenotypic responses must be an issue. Toads can survive years in dry desert soil, frogs PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 99 can survive deep-frozen ground, and when the work by Rick Relyea and his team has turned rains come, or the spring arrives, the sleeping the American wood frog Rana sylvatica ( Fig. 61 ) beauties wake up, display, and quickly mate into the best-understood example of what it and lay eggs, so that the larvae can use the tem- means to have a very plastic larval stage. porary, ephemeral, pool- and puddle-habitats Relyea ( 2003a ) showed that when tadpoles to mature and metamorphose in a life-history were exposed to the smells and actions of dif- stage that can endure the subsequent lean times ( Wilbur and Collins 1973 , Wilbur 1990 , Relyea 2007 ). Amphibians in Paradise, salamanders that have found permanent and predator-poor, deep-water bodies, forego that robust land-living stage altogether. Most famous among these reproductively active larvae is the Mexican A B axolotl Ambystoma mexicana ( Gould 1977 , Ryan and Semlitsch 1998 ). 500 C What is of most concern here though, is not 0.020 400 the metamorphosis into the adult stage, but life 0.010 300 as a larva. Look at a tadpole: it is a cute little 0.000 200 package of developing protein. Tadpoles make tail shape –0.010

final mass (mg) 100 high-quality, easily digested food for anybody –0.020 that is interested. Apart from hiding (and thereby 0 refraining from foraging and thus not growing), 30 0.010 their only defence is an ability to fl ee fast, to con- 0.005 29 fuse, or, in the case of many toads, possess a dis- 0.000 tasteful skin. Under threat it helps to have a 28 –0.005 relatively big and muscular tail ( Fig. 61 ), as it shape muscle may enable the tadpole to swim fast (Dayton et stage developmental 27 –0.010 al . 2005 ). A large tail fi n could also act as a lure to 30 0.010 length attract the attention of predators away from vul- width 0.005 depth nerable parts of the body ( Doherty et al . 1998 , 25 Johnson et al. 2008 ). In many species of frog, tad- 0.000 20 body shape poles from ponds with predators are smaller, activity (%) –0.005 with relatively shorter bodies and deeper tail 15 –0.010 fi ns than tadpoles from predator-free environ- no predator predator no predator predator ments ( Relyea 2001a ). Indeed, one of the most Figure 61. An adult wood frog Rana sylvatica (A), typical common tadpole predators, dragonfl y larvae, responses of wood frog tadpoles to either the presence (B, preferentially kill tadpoles with relatively shal- bottom) or the absence (B, top) of predatory dragonfl y larvae, low and short tail fi ns, and narrow tail muscles with a drawing by Rick Relyea, where the difference in tail ( Van Buskirk and Relyea 1998 ). Tadpoles lend depth between the two morphs is shaded (B, inset). The themselves well to experimentation, as artifi cial compilation in (C) shows the multiple phenotypic effects of caged dragonfl y predators on growth, development, activity, ponds, either stocked or not stocked with and relative morphology of wood frog tadpoles. The enclosed or free-swimming tadpole-predators morphological traits were size-adjusted by taking mean residual of various kinds, and with different levels of values (± SE) from regressions on mass. food or food-competitors, are easily set up. The Compiled from Relyea ( 2001b ) (B) and Relyea ( 2002a ) (C). 100 ADDING ENVIRONMENT ferent predators (there was a choice of four) or highly visible, large, and colourful tail fi ns, combinations thereof, they discriminated and whereas the presence of fi sh induced achro- produced predator-specifi c larval phenotypes. matic tails ( Touchon and Warkentin 2008 ). When predators were combined, tadpoles gen- Tadpoles of the tree frog Hyla versicolor altered erally developed phenotypes that were typical their defences when 10 different prey (them- of the more risky type, suggesting that they selves being one of the types) were either perceived the risk of combined predators as crushed by hand or consumed by dragonfl y the risk of the most dangerous one. Tadpoles larvae. Across all prey types, crushing induced not only have to survive, they also have to only a subset of the defences that were induced grow. Wood frogs that face high levels of com- by consumption, suggesting once more that petition from other tadpoles increase their for- tadpoles can assess quite precisely the degree aging activity and develop larger bodies. These and type of threat and respond accordingly larger bodies contain relatively larger scraping ( Schoeppner and Relyea 2005 ). mouthparts and longer guts, which improve Nevertheless, even the plasticity of tadpoles the animal’s ability to fi nd, consume, and fi nds a limit. When Schoeppner and Relyea digest food, but this ability comes at the cost of ( 2008 ) exposed wood frog tadpoles, and when a large tail, which makes tadpoles more sus- Teplitsky et al. (2005b ) exposed tadpoles of a ceptible to predators ( Relyea and Auld 2004 , European frog Rana dalmatina, to a gradient of 2005 ). densities of a particular predator, all measured The extraordinary ability of tadpoles to traits exhibited graded responses that levelled sense changes in their environment and off with increased predation risk. This sug- respond in precise and often appropriate ways, gests that there is either an organizational limit is illustrated, for example, by wood frog tad- to plasticity or a ‘functional limit’ that refl ects poles in experimental ponds that appeared to the high costs of the defensive phenotype. Part respond differentially to variations in the per of these costs might be incurred later in life. capita food levels and to changes in conspecifi c Wood frog tadpoles reared with caged dragon- density alone (Relyea 2002b ). When European fl y larvae, rather than without, showed rela- common frog tadpoles Rana temporaria were tively large limbs and narrower bodies as fully exposed to non-lethal caged predators (drag- grown frogs ( Relyea 2001b ); the survival value onfl y larvae) at low tadpole densities they of these later adjustments was not established. showed morphological as well as behavioural However, four months after metamorphosis, defences. At high densities and with severe common frogs that developed from predator- competition for food they remained active and exposed, rather than predator-free, tadpoles only adjusted their morphology ( Teplitsky and swam more slowly with less endurance—the Laurila 2007 ). Similarly, when tadpoles were opposite of what happened in the larval stage exposed to fi sh (that pursue their prey), they ( Stamper et al . 2009 ). This suggests that the lar- built deeper and longer tails and bigger tail val adjustments may come with a long-term muscles. In the presence of dragonfl y larvae fi tness cost. (sit-and-wait predators), they built only deeper We round up with the amazing story of two tails (Teplitsky et al . 2005a , Wilson et al . 2005 ), competing amphibian larvae studied in Japan- suggesting that tails as lures were the selective a frog Rana pirica and a salamander, or newt, force behind oversized tail fi ns. Indeed, in a called Hynobius retardatus . Tadpoles of many a Neotropical tree frog Dendrosophus ebraccatus , salamander or toad have tendencies to become the presence of dragonfl y larvae induced cannibalistic, but they only do so at high con- PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 101 specifi c densities ( Pfennig 1992 , Michimae ing, including the famous ‘spandrels’-paper 2006 ), and, preferably, if these conspecifi cs are by Stephen Jay Gould and Richard Lewontin not closely related ( Pfennig and Frankino (1979), and assembles many convincing argu- 1997 ). They then develop a broad head and ments that good adaptive design is usually gape large enough to swallow tasty nieces and the result of the self-organization that logi- nephews. They also turn predatory when there cally follows from growing structures inter- are many frog tadpoles around, as the mechan- acting with their immediate physical ical vibrations of fl apping tadpole-tails are suf- environments. In his own words: ‘organisms fi cient for this transformation to take place are designed not so much because natural ( Michimae et al . 2005 ). In a wonderful example selection of particular genes has made them of a phenotypic arms race, in response to sala- that way, but because agents of homeostasis manders becoming broad-headed and danger- build them that way’. Natural selection, the ous, frog tadpoles developed thicker bodies tinkerer, has an accomplice. and a comb on their backs to prevent ingestion For an interesting exposition of this notion by the salamander larvae; to which salaman- we go back to antlers. Unlike skeletal bones ders responded by further enlarging gape size that are adaptively moulded because of the ( Kishida et al . 2009 )! Interestingly, frog tad- particular strains that they experience (Fig. poles do not become bulky and high-bodied 51 ), antlers grow in the ‘assiduous avoidance solely in the presence of salamander tadpoles, of strain’ by virtue of a specialized, blood-rich, they only adjust when the salamander tad- and highly innervated type of skin called vel- poles become dangerous, i.e. express the pre- vet. Protected and nurtured by the velvet, car- dacious phenotype (Kishida et al . 2006 ). A fi nal tilage grows until specialized bone cells called subtlety is that when the threat of either preda- osteoblasts, moving in from the blood, eventu- tory larval salamanders or dragonfl ies is taken ally mineralize it. As long as the cartilage stays away, the specifi c morphological defences are abreast of the osteoblasts, the antlers grow, lost ( Kishida and Nishimura 2006 ). This but, as soon as the mineralizing osteoblasts reversal indicates once more that induced have taken the upper hand, growth stops. The defences can come with serious costs velvet then dies and, with its naked ossifi ed ( Hoverman and Relyea 2007 ), and thus that antlers, the stag is ready for the rut. The ratio fl exibility, or reversibility, pays if it is still an between ossifi cation and cartilage growth option developmentally ( Relyea 2003b ). determines the size and shape of the antlers. In small deer the ratio is high, so antlers stay Plasticity: the tinkerer’s small and pointed (Fig. 62A ). In large deer the accomplishment? ratio between ossifi cation and cartilage growth is low, so antlers can grow as ‘expansive broad With a few exceptions (e.g. Newman and plates’ ( Turner 2007 ). After the rut, the antlers Müller 2000 , Blumberg 2009 ), the literature on are shed. The new set usually resembles the phenotypic plasticity is permeated with the old set except that, if a deer is well fed, it is a idea that adaptive phenotypic responses, be bit larger. That antlers grow larger with age is they reversible or irreversible, cyclic or not, probably self-organized. Reproductive and are instructed from the genome. Of course, stress hormones combine with direct diet cues they will be, at some level at least. In his eye- to affect the relative rates of ossifi cation and opening book The tinkerer’s accomplice, J. Scott cartilage growth, and thereby antler size and Turner (2007 ) builds on early lines of reason- shape. 102 ADDING ENVIRONMENT

Now comes a surprise. When, during some ible again the next year, even if the velvet is stage in a deer’s life, antlers incur damage at fi ne then! The injured antlers are cast, but the one side because a bit of velvet got damaged memory of their shape apparently is not. during antler growth, this damage will be vis- Turner ( 2007 ) argues that the rich infi ltration of the velvet with both sensory fi bres (routed

Slow Similar rates Rapid through the trigeminal nerve and the thala- cartilage growth of cartilage cartilage growth mus) and so-called autonomic fi bres, which Rapid growth & Slow 240 mineralization mineralization mineralization control a whole range of physiological functions, suggests that ‘antlers are in the middle 200 of a feedback loop in which the trigeminal sensory nerves convey information to the brain 160 about the growing antler’s shape, while pattern of growth is mediated by information 120 streaming out on autonomic nerve fi bres, per- antler length (cm) 80 haps through controlling patterns and rates of blood supply to the various parts of the grow- 40 A ing antler’ (Fig. 62B ). This interaction would not only explain why growth deformities are 0 0 40 80 120160 200 240 copied in successive years, but also why ant- shoulder height (cm) lers show such exquisite bilateral asymmetry. There are no genes that directly instruct for Brain map B antler size and shape. Antler form naturally results from variations in the relative rates of cartilage growth and ossifi cation (which may partly year 2 be under genetic instruction), and the build-up year 1 of shape memories in the brain. Now think of the Trigeminal Autonomic amphibian phenotypes of fear. Could the tail sensory output input shape of tadpoles ‘simply’ result from differ- Reticular Hypothalamus formation ences in the kinds of exercise induced by the Growing antler presence of particular tadpole predators? pattern information Thalamus Limbic system Pattern matching information Phenotypic fl exibility in birds

Figure 62. (A) Antler length is a function of the shoulder In the most celebrated case of evolution in the height of the deer stag that carries it. The outcome is determined world of birds, the radiation of Darwin’s fi nches by a ‘race’ between how rapidly cartilage grows and how on the Galápagos Islands ( Lack 1947 , Grant rapidly it mineralizes, a ratio that is body-size dependent. In small deer, mineralization relative to cartilage growth rates is 1986 , Weiner 1994 , Grant and Grant 2008 ), the high, which leads to small, single-point antlers. In the largest trait of most interest is beak size. Depending on deer, the ratio is small and this produces the broad plate-like rainfall, plant growth, and the production of antlers carried by moose Alces alces and the extinct Irish elk seeds of different sizes, in different series of Megaloceros giganteus. (B) A diagram outlining the ways in years either big-beaked or small-beaked fi nches which antler shape may be memorized in the brain and how have survival advantages. As beak size has a this memory feeds back into the shape of the antler during the subsequent growth cycle. genetic underpinning, selection events show Put together from fi gures in Turner ( 2007 ). up in the beak-size trajectories of the relatively PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 103 isolated island populations. Immersed, as we Female red knots even show seasonal varia- are here, in examples of phenotypic plasticity, tion in the mass of their skeleton: they store one cannot help but wonder why small-beaked calcium before going to the trouble of laying a Darwin’s fi nches, living during times when clutch of four eggs ( Piersma et al . 1996b ). A bigger beaks would be advantageous, do not phenotypic aspect not studied in shorebirds, themselves build bigger bills? but carefully documented in songbirds, is the Although oystercatchers Haematopus ostrale- seasonal waxing and waning of song nuclei in gus can and do modify bill size and shape the brain ( Nottebohm 1981 , Smith et al . 1997 , ( Hulscher 1985 , van de Pol et al . 2009 ), Darwin’s Tramontin et al . 2000 ; note that even modern fi nches apparently cannot. In their very humans leading domesticated lives show sea- advanced synthesis of 34 years of work on nat- sonal change in parts of the brain, Hofman and ural selection in action in Darwin’s fi nches, Swaab 1992 ). Within a week after shifting Peter and Rosemary Grant (2008) do not even Gambel’s white-crowned Sparrows Zonotrichia use the word plasticity! Perhaps the material leucophrys gambelii from short to long days, the with which, and the way that, bills are built number of neurons in a brain area called the only allows for limited plasticity (Gosler 1987 ). High Vocal Centre increased by 50,000, or 70%, This contrasts with most other parts of bird as did the volume of this nucleus (Tramontin bodies that do show plasticity, mostly in the et al . 2000 ). Perhaps surprisingly, the morpho- categories of life-cycle staging ( Murton and logical changes were faster than the develop- Westwood 1977 , Wingfi eld 2008 ) and pheno- ment of the song control circuitry and the typic fl exibility (this book). In fact, one of the degree of stereotypy of the songs themselves. few examples of developmental plasticity in These took some extra weeks to mature fully. birds is the one for oystercatcher bills where, In view of the focus of much of the rest of to an extent at least, youngsters learn from this book on foraging, distributional decisions, their parents to feed on a particular diet and time-allocation problems, we round up by ( Sutherland 1987 , Sutherland et al . 1996 ). examining in greater detail the machinery for Depending on the prey types that they select, digestion and assimilation in birds, and their their bill is worn and strengthened in particu- plasticity. The digestive system of birds can be lar ways (yes, self-organization!), even though divided into several main components (Table there remains fl exibility later in life. 9 ). In any one taxon, each of these digestive Rather than structural body size, what pre- organs will have morphological characteristics dominantly varies in birds are body mass (an that functionally refl ect the evolutionary his- indicator of nutrient stores) and the size of tory and contemporary ecology of the taxon most of the organs ( Table 8 ), with the excep- ( Stevens and Hume 1995 ). tion of the lungs (Piersma et al . 1999b ). This is A large gut may provide the ability to use especially the case in those birds exposed to foods of low quality that are diffi cult to handle seasonally widely different conditions, either mechanically or chemically. The size of the because they migrate from arctic or temperate digestive tract should, therefore, refl ect the ben- to tropical climates, or precisely because they efi ts of having organs of that size, in balance stay put at highly seasonal northern or south- with the costs of maintaining that system. ern latitudes. There is also considerable intra- Before assessing degrees of variation in organ individual variation in correlated metabolic size, it is worth considering some general fac- attributes (Vézina et al . 2006 , 2007 ) and aspects tors that may set upper bounds to the size of of the immune system (Buehler et al . 2008a ). digestive systems. For example, it is obvious 104 ADDING ENVIRONMENT

Table 8. Many phenotypic traits of shorebirds (red knot, great knot, bar-tailed godwit, and others) show considerable intra- individual phenotypic change (C; relative to minimum) in the course of the seasons (S) and/or in relation to ambient environmental conditions (E).

Trait C S/E References

Body mass 2.5 × S, E Piersma et al . (2005) Skeleton 1.4 × S Piersma et al . (1996b) Fat stores 150 × S, E Piersma et al . (1999b) Flight muscles 3 × S, E Dietz et al . (2007) Blood haematocrit 1.5 × S, E Landys-Ciannelli et al . (2002) Heart 1.4 × S, E Piersma et al . (1999b) Liver 2.5 × S?, E Piersma et al . (1999b) Gizzard 7 × E Piersma et al. (1993b) Intestine 4 × E Battley et al . (2005) Skin 1.7 × S Battley et al . (2000) Testicle 30 × S T. Piersma unpubl. data Feather mass 1.3 × S Piersma et al . (1995a) Basal metabolic rate 2.5 × S, E Piersma et al . (1995a) Peak metabolic rate 1.6 × S?, E Vézina et al . (2006) Basal corticosterone level 20 × S Piersma et al . (2000) Daily peak melatonin titer 10 × S Buehler et al . (2009c) Total leukocyte count 4 × S, E Buehler et al . (2008a) Microbial killing effi ciency 10 × S, E Buehler et al. (2008a) Preen wax composition Complete qualitative shift S Reneerkens et al . (2002, 2007)

Table 9. Subdivision of the major components of the digestive system in birds. Based on Klasing ( 1998 ), Battley and Piersma ( 2005 ), and J.M. Starck (pers. comm.).

Organ Alternative name/components Primary functions

Oesophagus + crop Food storage, movement of food toward proventriculus Proventriculus Glandular stomach Gastric secretion Gizzard Muscular stomach Crushing or grinding food, mixing food with gastric secretions Small intestine Intestinum tenue, composed of duodenum and Enzymatic digestion, absorption of digestive end the combination of jejunum and ileum (the products ‘ jejunoileum ’) Ceca Microbial fermentation, water and nitrogen absorption, immunosurveillance Rectum Colon, large intestine Electrolyte, water and nutrient absorption Cloaca Storage and excretion of urine and faeces Liver Metabolism of absorbed nutrients, production of bile acids and bile salts Pancreas Secretion of digestive enzymes PHENOTYPIC PLASTICITY: MATCHING PHENOTYPES TO ENVIRONMENTAL DEMANDS 105 that there is a physical limit to the volume avail- reduce basal energy requirements ( Battley et al . able for organs within the abdominal cavity. Fat 2001 , Konarzewski and Diamond 1995 ). deposits will further reduce the abdominal Under the simplifying assumption that space for the digestive tract. For example, size (or mass) indicates functional capacity, migratory shorebirds deposit a thick layer of fat Fig. 63 summarizes the magnitude of changes around the gizzard, covering the intestines and in three major gastrointestinal components, rectum, and pressing up against the tail end of the gizzard, gut or small intestine, and liver. the skeleton. In great knots Calidris tenuirostris Data are separated into four taxonomic about to depart on northward migration from groups: galliformes (primarily grouse and Australia, fat in the abdominal cavity made up quail), songbirds, shorebirds, and waterfowl. 38% of the total abdominal tissue mass ( Battley Even though only few of the studies in the and Piersma 2005 ). At such times the profi le of review explicitly focused on documenting the abdomen changes, with fat birds showing a such changes, organ mass variations of ‘bulge’ behind the legs ( Owen 1981 , Wiersma 20–80% were common across a wide range and Piersma 1995 ). Confl icts for space between of species; length changes of the intestine fuel stores and organs may limit the maximum typically were lower, generally 10–20%. size of the digestive organs. Earlier we discussed whether tissue-related Internal organ mass could directly affect differences in fl exibility could be due to dif- fl ight costs and performance of birds. Because ferences in cell turnover, based on carbon- the costs of fl ight increase with body mass isotope signatures of historically spiked ( Kvist et al. 2001 ), and as manoeuvrability is DNA (Spalding et al. 2005a ). A method that impaired at higher masses ( Dietz et al . 2007 ), informs about the whole cell content, rather minimization of digestive organ mass will be than DNA only, may be more appropriate, an important consideration. In a comparative and measurements of retention times of car- study of raptors, Hilton et al . ( 1999 ) showed bon-isotopes supplied in food fulfi l this con- that ‘pursuing’ species (that require fast and dition. In a study on zebra fi nches Taeniopygia manoeuvrable fl ight) had shorter guts than guttata, Bauchinger and McWilliams (2009 ) ‘searching’ species, presumably refl ecting the found retention times varied between 8 days pursuers’ need to minimize mass. The shorter for small intestine, 10–13 days for stomach, retention time of these species would reduce kidney, liver,and pancreas, 17–21 days for the period of time they had food in their guts, heart, brain, blood, and fl ight muscle, and also keeping body mass low, though at a cost of 26–28 days for leg muscle and skin. Thus, lowered digestive effi ciency. Sedinger ( 1997 ) they found that tissues with a tendency to showed that grouse have caeca 4–5 times longer show large changes (notably the small intes- than waterfowl feeding on the same diet, and tine and liver) had fast turnover rates, suggested that the small caeca in waterfowl whereas less dynamic tissues, such as the refl ect a balance between the costs of fl ight for fl ight and leg muscles, had slower turnover waterfowl and the benefi ts of the caeca for rates. Tissue-specifi c turnover rate may thus nutrient balance. And even in animals at rest, partially determine the magnitude of organ mass-specifi c maintenance costs of the intes- fl exibility. tine and liver are greater than those of muscle Clearly, if we keep an open mind and our eyes and adipose tissue (Krebs 1950 , Blaxter 1989 , open, pretty much all parts of bodies appear to Scott and Evans 1992 ). Reductions in metaboli- vary with time, age, demand, and environment. cally active tissues, therefore, dramatically To ascertain the functionality of these changes 106 ADDING ENVIRONMENT

n = 88739 n = 64420 we need to do much more than ‘just’ record the 150 gizzard mass liver mass morphological changes. In the next few chapters 120 we considerably expand our view beyond bodily structures to examine active free-living organ- 90 isms, especially birds, as an integrated whole. 60

30 Synopsis

0 ‘A single genotype can produce many differ- n = 35414 n = 76 15 100 ent phenotypes, depending on the contin- gut mass gut length gencies encountered during development. 80 That is, the phenotype is the outcome of a complex series of developmental processes 60 absolute change (% of initial value) that are infl uenced by environmental factors 40 as well as by genes. Different environments can have an effect on the outcome of devel- 20 opment that is as profound as that produced 0 by different genes’. To this succinct description of phenotypic plasticity by Nijhout

waterfowl waterfowl ( 1999 ), we would like to add the possibility galliformessongbirdsshorebirds galliformessongbirdsshorebirds that much of the adaptedness shown by phe- Figure 63. Absolute change (either increase or decrease) in notypes is actually driven by self-organiza- digestive organs of birds. Data represent maximum percentage tional processes (Turner 2007 ). Reproductively changes recorded in studies for age- or sex-classes of birds, due to factors including diet type or quality, migration, breeding, successful phenotypes will satisfy ecological and food intake. Measurements from wild and captive birds are demands by optimizing the balance between included. The gut category represents data for both the entire various cost and benefi t functions of specifi c gut (including caeca and rectum) and the small intestine only. kinds of phenotypic (trait) variation. In brief, Sample sizes for each taxonomic group are given above the box most traits of all organisms are responsive to or points. Boxes represent the 25–75 percentiles (divided by particular environment features over shorter the median), whiskers represent the 10 and 90 percentiles, and outliers are shown as dots. For groups with only a few values, or longer time scales. Consequently taking individual data points are plotted. ecological context into account will enrich From Battley and Piersma ( 2005 ). biology on all fronts. Part III

Adding behaviour This page intentionally left blank CHAPTER 6 Optimal behaviour: currencies and constraints

When the going gets tough, the tough effl uent’, these snails adaptively increase the get going thickness of their shells ( Trussell and Smith 2000 ). However, as Bibby et al. (2007 ) discov- Rapidly rising concentrations of atmospheric ered, this only occurs when the water is not carbon dioxide help to heat up our globe. Not acidifi ed. Periwinkles tasting crab-juice only that, but the increase in carbon dioxide under normal pH conditions increased their also means that the seas and oceans around us shell thickness by 0.05 mm after 15 days, are becoming more acid. Compared to pre- whereas those exposed in acidifi ed crab industrial times, the pH of seawater has effl uent were unable to thicken their shells dropped by 0.1 units and is predicted to be 0.4 (Fig. 64A ). This would bring unresponsive units lower by the end of this century ( Orr periwinkles into real danger if they were et al . 2005 ). The corresponding increases in dis- exposed to real crabs, rather than merely to solved hydrogen ions (25% and 150%) will crab effl uent, but in fact they have a simple have a heavy impact on sea life, especially on way out of this potential problem. Snails those creatures that build a calcareous exoskel- may have a reputation of being slow and liv- eton or protective shell around themselves ing virtually sedentary lives, but they do (Vermeij 1987 ). Think of the python that we have behaviour as an option. The acidifi ed discussed earlier, which lowers the pH of its snails compensate for their lack of morpho- stomach fl uids in order to dissolve the bones logical defence by increasing their avoidance of the ingested prey. Calcareous material sim- behaviour, i.e. by crawling out of dangerous ply dissolves under acid conditions, and the water (Fig. 64B ). current acidifi cation of our world seas is appar- This is exactly the point that we shall make ently enough to create problems for corals, repeatedly in this chapter: animals behave, molluscs, and some plankton (e.g. Riebesell and their behaviour is the most fl exible and et al . 2000 ). Not perhaps as serious as for a rat fastest way to respond to environmental in a python’s stomach, but suffi cient to turn change. Think about the ptarmigan encoun- their daily lives upside down. tered in the previous chapter (Montgomerie Enter periwinkles Littorina littorea, small et al . 2001 ), that soiled their snow-white feath- gastropods living on rocky coasts. Just as in ers, thereby camoufl aging themselves: it takes most other molluscs, their calcareous shell weeks to moult into an inconspicuous tundra- acts as a protective shield against predators. like plumage, but only minutes to take a dirty Although hard and seemingly unchangea- mud bath. There are numerous examples ble, their shell is actually quite plastic. Even where behaviour outruns morphological when exposed to only the taste of shell- change as a way to cope with environmental crushing crabs when they are reared in ‘crab change.

109 110 ADDING BEHAVIOUR

much this decision1 contributes to her lifetime 0.06 * A reproductive success. Would she have done 0.04 better if she had produced four eggs (because she would not have had to work so hard, thus 0.02 yielding an extra breeding season), or would 0.00 half a dozen eggs have maximized her contribution to the population’s gene pool? This way

thickness increase (mm) thickness –0.02 of thinking about behaviour gave birth to a new fi eld of science, behavioural ecology, which 20 * B fl ourished from the late 1960s onwards, and is

15 still very much alive and kicking (see textbooks by Krebs and Davies 1978 , 1984, 1991, 1997 and 10 more recently by Danchin et al . 2008 ). Attempting to determine the contribution of 5 * a fi ve-egg clutch, or any alternative clutch size,

% avoidance response % avoidance 0 to a kestrel’s fi tness is a daunting task. That normal pH low pH normal pH low pH this is possible is shown by the elegant fi eld + cue + cue experiments of Serge Daan and co-workers Figure 64. (A) When exposed to the taste of crabs, periwinkles ( Daan et al . 1990a ). They manipulated brood increase the thickness of their shell, but are only able to do so size by moving eggs around between different under normal pH conditions. In acidifi ed seawater, the snails cannot build thicker shells, but (B) use a behavioural way to avoid nest boxes. Some pairs had to raise fewer predation by crawling out of the water. Asterisks indicate young than their original clutch size, others treatments that are signifi cantly different from other treatments. had to raise enlarged broods, while the brood Modifi ed from Bibby et al . ( 2007 ). size in a third group was left untouched. The fate of the adults and their young was followed over the subsequent years, enabling the esti- Loading leatherjackets mation of survival and reproductive rates. After many years of data collection it became In this chapter we will discuss behavioural fl ex- evident that the fi tness-value of the unmanip- ibility, and we will approach it from a functional ulated brood sizes was highest, because of a angle. It was ethology-godfather Niko Tinbergen reduced parental fi tness in enlarged broods who taught us that one (out of four) ways of and a reduced fi tness of the clutch in reduced looking at behaviour is to consider function . We broods. can ask how much an observed behavioural However elegant and stimulating such fi eld- choice contributes to an animal’s fi tness, and based estimates of the fi tness-consequences of also how much the alternative, unchosen life-history decisions may be, they do not allow options would have contributed (Tinbergen estimates of the fi tness-consequences of the 1963 ). For example, if a female kestrel lays a more short-term decisions that animals may clutch of fi ve eggs, we would like to know how make many times per day. Think of a parent

1 Behavioural ecologists use the word ‘decision’ to indicate that an animal has behaved in one of several possible ways (McFarland 1977 ). They do not suggest that the animal has consciously calculated which behaviour would be optimal in any particular case. Possible mechanisms that animals may use to reach such ‘decisions’ are considered later in this chapter. OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 111 starling Sturnus vulgaris that has just delivered behaviour (Stephens and Krebs 1986 , Sih and a ‘beakful’ of fi ve, rather than four or six, leath- Christensen 2001 ). Going back to the starling erjackets (Tipula larvae) to her offspring wait- delivering food to its hungry young, a candi- ing in the nest. Parent starlings line up multiple date currency to maximize is the daily amount prey items in their beak before returning from of food delivered to its offspring . While doing so, the feeding site to the nest (so-called ‘central- the starling faces a trade-off between, on the place foraging’; Orians and Pearson 1979 ), and one hand, minimizing the daily time fl ying they thus need to decide on exactly how many back and forth between the foraging site and prey items to upload into their beaks ( Fig. 65A ). the brood (by collecting as many leatherjack- While collecting the food at the foraging site ets as it is possible to stuff into the beak) and, (mostly leatherjackets; Tinbergen 1981 ), they on other hand, minimizing the daily time face diminishing returns, because each extra spent at the foraging site (by returning each prey item in the beak means a reduced effi - time to the brood with just a single leather- ciency for fi nding and handling the next one jacket). Because the variable to maximize is ( Fig. 65B ; this has been coined ‘the loading known and quantifi able (the currency), the effect’ by Kramer and Nowell 1980). Would the optimal number of prey to deliver can now be starling that fl ew off with five leatherjackets calculated (Fig. 65B ). Interestingly, the opti- have done better (in fi tness terms) if she had mal number depends on the average time it continued foraging and collected another prey takes to travel between the nest and the feed- item before taking off, or should she already ing patch: the shorter the time it takes to travel have stopped after she had found the fourth (on average), the fewer prey should be brought leatherjacket? It may not matter at all, and cer- back upon return ( Fig. 65C ). tainly not for a single round-trip, but what if Alex Kacelnik ( 1984 ) took advantage of this she always loaded fi ve prey items rather than quantitative prediction in order to fi nd out four or six, every breeding season, her whole whether the suggested candidate currency life long? With about 6000–8000 prey deliveries best described the number of prey uploaded per brood ( Tinbergen 1981 , Kacelnik 1984 ), an as a function of the distance fl own, or whether occasional second brood within the same sea- alternative currencies performed better. On son, and 2–3 breeding seasons per lifetime, this the Dutch isle of Schiermonnikoog, he trained adds up to roughly 20,000 such prey-delivery free-ranging parent starlings to come to a decisions per lifetime. The fi tness-consequences wooden tray on which they could collect meal- of each of these short-term decisions are small worms for their young. By dropping single and this makes it diffi cult to quantify their pay- mealworms through a long plastic pipe, at offs. For this reason, behavioural ecologists increasing intervals, Kacelnik generated the work with measurable surrogates for fi tness, birds’ own loading curve and recorded the so-called ‘currencies’. number of mealworms that the trained star- Just as we all use currencies in our daily lings would take at each visit. Each day, he lives to express the value of our goods in would randomly move the tray to a different Euros, pounds, dollars, or yen, behavioural site, such that, by the end of the experiment, he ecologists use currencies to express the fi tness- had manipulated the starlings’ travel distance, value of behavioural options. As there is likely ranging from 8 to 600 m. The results were to be a premium on collecting as much energy astounding (Fig. 65D ). The number of meal- per unit time as possible, energy intake rate is worms went up as a function of travel time. often used as a currency to explain foraging What is more, the results also fi tted closely 112 ADDING BEHAVIOUR

7 prey 8 prey load 8 optimum 6

2 1 prey B A

optimum travel time searching time

7 6 5 optimum (long) 4 optimum (short) 3 load size 2 C 1 D 0 long short 0 2040 60 80 100 travel time searching time round trip travel time (s) Figure 65. (A) The complete foraging cycle of a central-place foraging parent starling, which alternates between collecting prey items at a feeding site and delivering them to its young in the nest. (B) The loading or gain curve at the feeding site describes the number of prey items loaded in the starling’s beak as a function of search time. As the beak-load increases, the effi ciency offi nding and loading the next item decreases, resulting in a decelerating loading curve. The optimal number of items to load at the average site (seven in this example) depends on the average travel time between patches and can be found by the tangential solution shown. (C) The less time it takes to travel (on average), the fewer prey should be brought back upon return. (D) Kacelnik’s (1984) data (dots) fi tted best with the prediction based on net family gain rate maximization (line). Compiled from Shettleworth ( 1998 ) (A) and Krebs and Davies (1991) (B–D).

with the prediction based on the maximization sume itself rather than to deliver to the young, of the net amount of energy gained by the entire etc.; see Tinbergen 1981 ), but the decision stud- family as a whole, thus subtracting the metabolic ied here is: how many prey items to load before rates of the parent and the chicks from the taking off? Then, after defi ning thecurrency , gross amount of energy collected. However, the unit in which the pay-offs of the taken deci- we should add that the predictions of the other sion are expressed, it becomes clear what con- three currencies were not too far off this mark straints the parent starling faces: travel time, either! search time for each consecutive prey item, This example shows the use of decisions, ‘loading time’ required to pick up each newly currencies, constraints, and optimality in encountered prey item, and the metabolic rate behavioural ecology. We must take into account in both the parent (in the feeding patch, in the that the starling needs to make many more air, and in the nest) and its offspring (while decisions, which are not treated here (e.g. begging and while resting). If any of these con- where to feed, how many prey items to constraints could be alleviated (shortened in the OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 113

A case of time factors, lowered in the case of metabolic demands), the starling would then optimum obtain higher net benefi ts. Once the constraints (long and short) are know, then the trade-offs can be defi ned. In the starling’s case, the major trade-off is fi nding the right balance between the daily time spent in the feeding patches and the daily time devoted to travelling. Having identifi ed the long short travel time searching time trade-offs, the optimal solution to the decision of interest can be calculated. This way of stud- 8 B ying behaviour gives the behavioural ecologist 7 predictions, and as we shall see next, empirical 6 falsifi cation of clear-cut predictions is an effec- 5 tive way to make progress in science (Popper 4 1959 ). load size 3 A couple of years after Kacelnik carried out 2 his elegant fi eld experiment in The Netherlands, 1 0 a remarkable fi nding would shake up his ear- 0 20 40 60 80 lier conclusions. At the Oxford University round trip travel time (s) Farm in Wytham, Kacelnik’s PhD student Innes Cuthill repeated the Schiermonnikoog experiment ( Cuthill and Kacelnik 1990 ), but this time offered free-ranging trained starlings a linear, rather than a decelerating, loading curve ( Fig. 66A ). Thus, once it had arrived at the tray, a starling was again offered one meal- C worm at a time, but now with a constant time interval in between two consecutive meal-

100 D worms (so-called ‘sudden-death patches’; 386 371 365 346 403 371 371 375 358 329 364 365 405 356 358 327 354 351 356 321 Brunner et al . 1996 ). No more diminishing 332 356 394 339 80 returns, but a constant supply of food up until

60 a total of eight mealworms had arrived on the starling’s dish (which represents the maximum

% visits 40

20 perch with the light off to the one with the light on; the light turns off at the perch at which it arrives, turning on the light at 0 1 2534 6 the other end. This goes on until the travel requirements have bird been met, after which food is offered at the feeder in the centre of the cage. (D) Percentage of patch visits at which individual Figure 66. (A) When the loading curve is linear, load size starlings consumed all eight mealworms that were offered. should always be maximal, irrespective of travel time. (B) Shadings indicate different treatments (travel distance and Surprisingly, parent starlings brought home larger loads when steepness of the gain curve were varied), numbers indicate travelling for a longer time. Different symbols denote different sample sizes. individuals. (C) Experimental set-up mimicking patch use. In Compiled from Krebs and Davies (1991) (A-B), Shettleworth ( 1998 ) order to travel between ‘patches’, the starling hops from the (C), and Cuthill and Kacelnik ( 1990 ) (D). 114 ADDING BEHAVIOUR number of mealworms that a starling can pos- constraints that had to be redefi ned, the next sibly carry at any one time). The optimality scientifi c discovery story tells us how failing model now predicts that the starling should optimality models can inform us about alter- always collect all eight prey items before tak- native currencies . ing off, irrespective of travel distance (Fig. 66A ). So what did the starlings actually do? Better lazy than tired They ignored the theoretical prediction by bringing home more prey when travelling fur- For the most part, foraging theoreticians have ther (Fig. 66B ). Imagine the amazement on the included ‘time’ as the denominator in their faces of Cuthill and Kacelnik! Yet, unabashed, favourite currency, meaning that something they continued with another experiment that ought to be maximized per unit of time . For would eventually bring back light into the new example, the number of leatherjackets col- darkness. The linear loading curves were now lected per minute , or the net amount of Joules offered to self-feeding starlings that had no gained during the total day. This makes much brood to care for, but still needed to travel to sense, since foraging takes time, and animals move between ‘patches’ ( Fig. 66C ). Guess also need to devote some of their daily time to what? Those self-feeding starlings did exactly activities such as fi nding a mate, building a what they were ‘supposed’ to do: they gave up nest, incubating eggs, and so on. It also makes virtually every food patch after collecting all sense when animals are in a hurry—when eight mealworms, irrespective of the distance migrants need to fuel up to reach the breeding travelled ( Fig. 66D ). The only difference grounds in time, for example. In such cases, between these latter two experiments was the they devote all the time possible to foraging fact that chick-feeding starlings had to ‘upload’ ( Piersma et al. 1994b ) and maximize the (net) and carry their prey items. Cuthill and energy income per day. Kacelnik, therefore, concluded that, with For this reason, it did not make sense when increasing load size, some cost function would foragers were often observed feeding at a increase, whether it was the energy expendi- slower pace than was possible, even foragers ture while in the patch, or the energy expendi- that were trying to fuel up as quickly as they ture while in fl ight, or the time it takes to return could, as you will see shortly. Dungeness crabs to the nest. Such costs had been overlooked Cancer magister were eating smaller bivalves when Kacelnik ( 1984 ) suggested that parent than expected from rate-maximizing princi- starlings should maximize net family gain . Even ples ( Juanes and Hartwick 1990 ), Lapland though this currency did account for parental longspurs Calcarius lapponicus were not maxi- energy expenditure, it did so in a load-size mizing the delivery rate to their offspring independent manner. because they fl ew slower than they ‘should’ Such is the power of optimality models. As ( McLaughlin and Montgomerie 1990 ), as did we have just demonstrated, a clash between black terns Chlidonias nigra ( Welham and theory and facts can lead to new ideas about Ydenberg 1993 ). Honey bee Apis mellifera the true constraints and currencies that under- workers only partially fi lled up their crops, lie the observed behaviour. Optimality models whereas they should have fi lled them up com- are by defi nition over-simplistic, but, when pletely if they were to maximize the rate of falsifi ed, they form a guide for further research. delivering nectar to the hive ( Schmid-Hempel Just as in everyday life, making mistakes is the et al. 1985 ). This list could be extended further, best way to learn! Having seen an example of but in all these examples it turned out that OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 115 another energetic currency was in fact being diminishing returns when emptying their maximized, namely, the gross amount of patches (because searching through the patches energy gained per unit of energy expended, occurred in a non-systematic way, leading to called ‘effi ciency’, and formally written as: longer time intervals between consecutive prey encounters). This gave rise to the critical b / c , question of this study: at what (diminishing) where b denotes gross energy intake rate and c intake rate would knots give up a patch and denotes metabolic rate. switch to a new patch ( Fig. 67A )? Initially, the reasons for this effi ciency max- Given the time of year, we expected knots imization were not understood. In fact, effi - to give up their patches so that they maxi- ciency maximization was dismissed as a mized their short-term net energy intake nonsensical currency by the fi rst foraging the- rate, i.e. their net intake calculated over total oreticians ( Stephens and Krebs 1986 ). They foraging time, taking into account the time reasoned that, if 2 J can be gained for 1 J of and energy cost required to make brief fl ights expenditure, and the alternative would yield between patches. This meant that, on the one 5 J while expending 3 J, then the second hand, patches should not be exploited for option, the least effi cient option (5/3 < 2/1), too long, as the birds would lose too much should be preferred, simply because it yields time to increasingly long probing intervals, a higher net gain (5–3 > 2–1). Only if expend- whereas, on the other hand, giving up the ing energy bears additional costs should it patches too soon would cost too much travel pay off to expend energy parsimoniously. We time, and especially travel energy, when refer back to Chapter 4 for an in-depth treat- moving between different patches (Fig. 67B , ment of non-energetic costs due to energy C). To our initial surprise, the red knots expenditure (such as shortened life span, probed much too long in their patches and increased risk of immediate death, hampered were more reluctant to move on to a new immune response). Although such non-ener- patch than we had anticipated. Initially, we getic costs may explain effi cient and prudent interpreted this as ‘lazy behaviour’, since behaviour, more recent foraging models have the time devoted to the most costly activity, shown that there may nevertheless be purely the short fl ights between patches, was less energetic, rate-maximizing reasons for being than half of what we expected. an effi cient forager. Let us now examine this Were the knots maximizing their effi ciency idea more closely. by obtaining every Joule of energy income as cheaply as possible, irrespective of the time it More haste less speed took? If so, would this increase their life- expectancy at a time of year when they actu- Even when in captivity, red knots fuel up rap- ally need to fuel up as quickly as possible? idly in spring and deposit the same amount of This could be the case, but it seemed more fat as do their free-ranging conspecifi cs (Cadée likely that it would be preferable to speed up et al. 1996 ). It was at this time of year that we the fuelling process, the more so because the performed an experiment with four red knots, birds weighed less than was normal for the which were allowed to exploit depletable last days of April. patches in a large outdoor intertidal aviary The answer was provided by new the- (van Gils et al. 2003b ). In the same way as the oretical work from McNamara and Houston starlings we discussed earlier, the knots faced ( 1997 ), but also by Ydenberg and co-workers 116 ADDING BEHAVIOUR

arrival departure prey encounter digestive & handling pause search

time budget

long-term metabolic rate (W) 2.0 3.0 4.0 5.0 30 Onet D 25 Omin 20 15

long-term net 5 digestive constraint

energy intake rate (W) rate energy intake 0 y = –x –5 Orest

0.8 B C

0.6

0.4 at departure (#/s) 0.5

potential encounter 0.2 rate

0.0 0 51510 20 0 5110 520 initial prey density (#/patch) % of foraging time spent travelling Figure 67. (A) Typical patch-feeding situation for a red knot: between arrival and departure from a patch, the bird alternates between searching for prey (open bars), handling prey (grey dots), and digesting prey (grey bar). (B) The potential prey encounter rate at which patches were given up was independent of their initial prey density. (C) The higher the potential prey encounter rate at which patches are given up, the more time knots will devote to travelling. (D) Long-term net energy intake rate is determined by the giving-up policy, peaking at relatively high quitting intake-rates (Onet ). However, because red knots face a digestive bottleneck on their gross energy intake rate, their long-term net intake rate cannot exceed the diagonal line indicating this constraint (which has a slope of –1 because net intake (Y-axis) = gross intake (intercept) – cost (X-axis)). Therefore, they should not opt for Onet , but rather feed at a slower pace by alternating between resting (Orest ) and giving up their patches at lower potential intake rates (Omin ), in which case their long-term net intake rate is maximized at the triangle. This is exactly what the birds did, as indicated by the grey arrow going from (B), via (C) to (D). Modifi ed from van Gilset al . ( 2003b ). OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 117

( Ydenberg et al . 1994 , Ydenberg and Hurd amount of fl esh that can be assimilated daily 1998 ), and Hedenström and Alerstam ( 1995 ). (about which we shall learn much more in the Energy-limited foragers, i.e. animals that face next chapter). a constraint on the daily amount of energy that The intake rate at which our birds left their can be spent or assimilated, cannot feed dur- patches did indeed maximize the foraging ing the entire day, as they need to ‘respect’ gain ratio while they were feeding. To inves- their energetic bottleneck (see Chapter 4 ). Such tigate how this strategy would maximize foragers, and this may sound counter-intui- overall net energy intake, we used the graphi- tive, should actually maximize (a modifi ed cal approach advocated by McNamara and form of) effi ciency while foraging, in order to Houston ( 1997 ) as shown in Fig. 67D . The tri- maximize the net daily energy uptake across angle in that graph (the long-term net energy the entire day . The form of effi ciency that should intake rate) cannot exceed the diagonal line be maximized has been coined the ‘foraging indicating the digestive constraint. The more gain ratio’ ( Hedenström and Alerstam 1995 ) or this point is located to the left on the diago- simply the ‘modifi ed form of effi ciency’ nal, the higher the long-term net energy ( Houston 1995 ), and takes account of the meta- intake rate (since the diagonal is going ‘down- bolic rate when not foraging, i.e. the resting hill’). This left-most location on the diagonal metabolic rate r : is achieved by alternating between resting (option O in the graph) and maximizing the b / ( c – r ) . rest foraging gain ratio while feeding (option O min ,

This apparent paradox has a relatively simple which is found by drawing a line from O rest — explanation. If the gross amount of energy that dashed in this case—that is tangental to the can be assimilated cannot exceed a certain thick curve defi ning the continuous range of threshold, then it pays to gather this amount of feeding options). energy as cheaply as possible, taking into Of course, the argument for effi ciency maxi- account r , the metabolic rate while not feeding mization only holds as long as there is enough (the faster the threshold is reached, the more time to reach the digestive bottleneck. When energy is lost to non-feeding behaviour). The there is only a limited amount of foraging time, less energy that is expended in total, the more animals are said to be time constrained rather energy remains. With respect to our knots, than energy constrained. When ‘heavily’ time they were indeed energy-limited: the amount constrained, the best thing such animals can of energy they could assimilate during any do, when aiming to maximize net daily energy one day was constrained by the rate at which intake, is to maximize their net intake rate they were able to process their shellfi sh prey. while feeding. It is as simple as that and means As knots ingest their prey whole, they must that animals may switch ‘tactics’, depending internally crush and process the bulky shell- on whether they face a time constraint or an material, which then needs to pass rapidly into energy constraint, as exemplifi ed below. the digestive system in order to make room for new food. The crushing of the shell takes place Oystercatchers pressed for time in the muscular stomach, called the gizzard ( Piersma et al. 1993b ), and it is especially the The aviaries in which we offered depletable size of this part of the ‘gut’ that determines the patches to knots were used to study foraging daily amount of shell mass that the birds can oystercatchers Haematopus ostralegus 17 years process (van Gils et al . 2003a ), and thus the earlier by Cees Swennen and others (1989). 118 ADDING BEHAVIOUR

summer autumn winter They wondered how animals deal with 100 shrinking foraging time, such as often hap- 80 pens to birds exploiting the intertidal zone 60 during low tide. Strong onshore winds and 40 periods of frost may dramatically shorten a time (%) shorebird’s foraging time ( Zwarts et al . 1996 ). 20 A If such time restrictions are predictable and 0 12 846 10 68512 animals can foresee them, can they compen- daily foraging time allowed (h) sate by working harder? During the 1980s, rate-maximization was still the dominant B b/c 60 maximum rate currency in foraging theory, and animals 6 3 4 5 50 maximum were expected to obey this ‘rule’ under all efficiency 2 40 circumstances. Seven individual oystercatchers were 30 b=c 1 allowed to exploit cockles Cerastoderma edule , 20 but had to retreat into their roosting cage when 10 energy intake rate b (W) rate energy intake the tide came in. As it was possible to manipu- 0 0 r 5 10 15 20 25 30 late the tidal length in these cages, the birds energy expenditure rate c (W) were offered normal tides with 6 hours of low tide, but also shorter tide intervals of 5, 4, 3, cannot meet 432 1 2.5, and even 2 hours. These tidal regimes were requirement BEST TACTIC 400 C made predictable by offering the same tidal length for a period of about a week (Swennen 300 et al . ( 1989 ) also performed a series with unpre- 200 dictable feeding periods). Seemingly violating the rate-maximizing paradigm, by working 100 total net daily gain (kJ) harder when their available foraging time was

0 shortened, the oystercatchers showed that 0 2846 10 12 they did not always work at maximum rates. daily foraging time allowed (h) The birds spent proportionally more time for- Figure 68. (A) Oystercatchers devoted more of their time to aging during shorter low tides (Fig. 68A ), foraging (lower bars) with a predictable shortening of the found their prey items faster, and it also took low-tide period (middle bars denote resting and upper bars them less time to handle and ingest them. At denote preening). (B) Harder work (horizontal axis, with r the time of his experiment, Swennen may have indicating resting metabolic rate) leads to diminishing wondered why his oystercatchers did not foraging returns (vertical axis). Numbers indicate different rates maximizing different currencies (net energy intake rate always work at their maximum rate; nowa- is maximized in 4, effi ciency is maximized in 2, and the days the surprise may be the other way around foraging gain ratio is maximized in 1). (C) Under a digestive constraint, total net daily intake is maximized by maximizing the foraging gain ratio (tactic 1 in (B)). However, this strategy requires much foraging time. With a shrinkage of the available foraging time, an oystercatcher does better to switch Modifi ed from Swennenet al . ( 1989 ) (A) and Ydenberg and Hurd ( 1998 ) strategies, with maximization of the net energy intake rate (B–C), making a slight correction in (B) by setting resting metabolic rate r (tactic 4 in (B)) being the best tactic for the shortest available at 4 rather than 5 (otherwise the foraging gain ratio would be maximized feeding times. while resting, a point that was overlooked by Ydenberg and Hurd). OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 119 for us. Why do oystercatchers work harder Informational constraints: getting when foraging times are shortened? As we to know your environment have just seen, energetically bottlenecked foragers should maximize their (modifi ed form The case of the oystercatchers exemplifi es two of) effi ciency in order to maximize their net major constraints in a single story—energetic daily gain. Oystercatchers are a classical exam- ceilings and time constraints—but there is, in ple of digestively bottlenecked foragers fact, a third, more hidden constraint that also ( Kersten and Visser 1996 , Zwarts et al . 1996 ), underlies the entire study. This is the degree to and should thus feed at a relaxed pace under which the oystercatcher’s brain is able to all circumstances, whether the tide intervals absorb, process, and remember information are short or long (a further consideration being about the current tidal regime. Swennen and that feeding at a high rate may incur additional his colleagues trained the birds for a couple of non-energetic costs, such as bill damage; days before they got acquainted with their lat- Rutten et al . 2006 ). Again theoretical work est tidal regime. As a subsidiary test, they also solved this mystery, this time by Ydenberg and performed some trials in which each day a Hurd ( 1998 ). new, completely randomly chosen tidal regime The simple explanation is that, with a pro- was offered. Not surprisingly, the oystercatch- gressive shortening of their foraging time, ers did not respond to tide length and were oystercatchers did get rid of their digestive unable to meet their daily energy demands constraint, but acquired a new time constraint when exposed to shortened low tides. in return. Modelling gross food intake as a Knowledge about your environment can be a function of workload ( Fig. 68B ), Ydenberg matter of life and death! and Hurd (1998 ) made clear that, in order to The starlings that were the earlier stars of continue maximizing net daily energy gain, the show should also be aware of an impor- oystercatchers should switch currencies when tant environmental parameter: their long-term faced with a progressive shortening of forag- average intake rate, determined by feeding ing time. When there are more than 9.5 hours policy and also by the average distance of feeding per day (i.e. almost 5 hours per between their patches and their nests. The tidal cycle), oystercatchers should maximize rate-maximizing starlings should leave each the modifi ed form of effi ciency, the foraging patch when the instantaneous, or marginal , gain ratio (tactic 1 in Fig. 68C ). Under a tighter intake rate drops below the long-term average regime, this policy becomes time limited and intake rate. For this reason, this model has the birds do better by switching to effi ciency been coined the ‘marginal value theorem’ maximization (tactic 2 in Fig. 68C ), allowing (MVT; Charnov 1976a ). Thus, the graphs in themselves once more to obtain the maximum Fig. 65B and C, represent the average situation amount of gross energy that they can assimi- in the forager’s environment. Only the average late during the day. However, when there are travel time to patches and the average gain fewer than 7 hours of daily feeding (each low curve predict the optimal average number of tide lasting 3.5 hours at most), birds do better leatherjackets loaded. However elegant the to switch to tactic 3 (Fig. 68C ), and they should graphical approach of the MVT, as exempli- switch to instantaneous rate maximization fi ed in Fig. 65B and C, the implication is that when feeding times fall below 6 hours per this approach does not work when predicting, day (low tides of 3 hours or less; tactic 4 in for a single environment ( Fig. 69A ), the optimal Fig. 68C ). leaving policy for different patches at different 120 ADDING BEHAVIOUR

distances from the central place (and this has recent within-environment study, the nega- often been overlooked, as recently stressed by tive load—distance effects are confi rmed for Olsson et al . 2008 ). In fact, within a single envi- overwintering Bewick’s swans Cygnus colum- ronment with no variation in patch quality, bianus bewickii ( van Gils and Tijsen 2007 ). because all patches should be left at the same Although these birds do not bring food to a marginal intake rate, similar load sizes should central place, they do get together at a central be brought back home from all possible dis- place, each night, when roosting at a lake or tances. When the cost of carrying a load a fl ooded fi eld. Exploiting a mosaic land- increases with load-size (as should be the case scape of recently harvested sugarbeet fi elds, for the parent starlings; Cuthill and Kacelnik the birds feed for less time in distant fi elds, 1990 ), but also when there are additional, thus quitting those fi elds at higher intake load-size independent costs of travelling, neg- rates ( Fig. 70 ). More examples of negative ative load-distance correlations are expected load—distance effects were detected in situa- (Olsson et al . 2008 ). tions where predation risk increased further Most studies on central-place foraging away from a central refuge ( Brown and Kotler have been carried out between different ‘envi- 2004 ). We will come back to such anti-preda- ronments’ (Fig. 69B ) and generally fi nd position issues in Chapter 8 . tive load-distance correlations (reviewed in To keep up with changes in your environ- Stephens and Krebs 1986 ). However, in a ment, it is critical to be able to perceive these changes. In fact, by each day placing the artifi cial

6 OBSERVED FUELLING RATE

B (#/h) quitting dropping rate 0 0 2846 10 12 14 distance to roost (km)

Figure 70. A within-environment effect of central-place foraging in Bewick’s swans: the birds gave up their foraging sites (fi elds of sugar beet remains) at much higher feeding rates (expressed as easier to observe dropping rates) when these were located further from the roost (dots are means, Figure 69. Cartoons of central-place foraging, with the central bars are standard errors). Marginal value theorem predictions place in the middle (bird on the circle), surrounded by patches were met since the marginal net intake rates at quitting (squares) of various quality (shading). The effects of travel distance (calculated from the quitting dropping rates) precisely on foraging decisions can be quite different when studying them matched the long-term net intake rates (calculated from the (A) within environments or (B) between environments. Most observed rates of increase in their abdominal profi les), as studies comply with the between-environment comparison. indicated by the thick grey bar. Modifi ed from Olsson et al . ( 2008 ). Modifi ed from van Gils and Tijsen ( 2007 ). OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 121 feeding patch at a different distance from the of no intake between prey captures? And, sec- nest, Kacelnik ( 1984 ) changed the starlings’ ondly, if they are able to get some measure of environment from day to day, making this instantaneous intake rate, should they really study a between-environment study. As sug- base their foraging decisions on this measure, gested by the positive load—distance effects or are there better alternatives? (Fig. 65D ), the starlings did perceive these rapid Before the marginal value theorem was pub- changes as environmental changes and lished ( Charnov 1976a , Charnov and Orians responded adaptively. How quickly starlings 1973 ), there were ideas about how discrete-prey can learn about their environment was exam- foragers could estimate their own intake rate ined in a follow-up study by Cuthill et al . ( 1994 ). and how they could decide when to quit a patch. In a laboratory set-up, starlings had to get used A simple and attractive idea was that such for- to new travel distances every two days. After agers should give up a patch after a fi xed time- the sudden increase or decrease of travel dis- interval of not fi nding a prey item, called the tance, the birds gradually adjusted their load Giving-Up Time (GUT). Niko Tinbergen meas- size in the correct direction (larger loads for ured it in carrion crows Corvus corone searching longer distances). After six full cycles of travel for camoufl aged eggs ( Tinbergen et al . 1967 ), and patch use, the birds reached the asymp- and Croze (1970) came up with a strong and totic behaviour that they maintained until the intuitive argument why animals may use GUTs: next distance-change. Quite a steep learning ‘the GUT is taken to be a quality of the crow’s curve! But note that in this indoor laboratory persistence—an expression of the amount of study, one thing was kept simple for the birds: effort the predator is willing to allot in pursuing the type of patch offered was always the same. one more of a particular prey’. In a classical Out in the fi eld, patches differ in their quality experiment, Krebs et al . ( 1974 ) concluded that and the amount of prey they contain, and in black-capped chickadees Parus atricapillus the next section we will discuss how birds get adopted a constant GUT when giving up their to know their patch. patches (but others noted that too few prey were taken to be conclusive about this, e.g. Iwasa et al . Informational constraints: getting 1981 ). to know your patch However elegant and natural the GUT concept may seem, it did not fi t with the equally The function describing the cumulative energy elegant marginal value theorem (MVT). First, gain when exploiting a patch is usually drawn the MVT requires a continuous, uninterrupted as a smooth curve (e.g. take a look at Fig. 65B intake in which GUTs cannot exist. Second, and C). Such a continuous, uninterrupted even if the MVT is slightly modifi ed, such that intake may be applicable for foragers sucking it can make predictions about GUTs, then a up liquid food, such as nectar-eating hum- constant GUT turns out to be suboptimal. In mingbirds ( Roberts 1996 ). However, for many that case, foragers should be more patient foragers, food comes in discrete lumps at (apply longer GUTs) in richer patches, pro- rather stochastic moments in time (e.g. Fig. vided that they can recognize patch quality 67A ), yielding stepwise gain curves ( Fig. 71A ). instantly ( McNair 1982 ). But even this latter We have assumed that many foraging deci- rule does not actually work in practice, as it sions are based on instantaneous intake rate, requires patch quality to be recognized but how can and how should such foragers instantly. Most foragers, especially those that measure their own intake rate if there are bouts feed on cryptic prey, need sampling time to 122 ADDING BEHAVIOUR

identify patch quality and can, therefore, not 6 A adopt McNair’s ‘fl exible GUT rule’. Such naïve 5 foragers should fi nd some way to integrate the 4 information that their foraging success contains, so that they can estimate patch quality 3 (i.e. prey density) while feeding and determine 2 a good moment to give up the patch. They

number of prey found of prey number 1 would do really well if they could combine such ‘patch sample information’ with an a pri- 0 ori expectation about prey density, based on B 8 the environment-wide frequency distribution of prey densities, i.e. if they updated their for- 7 aging information ( Oaten 1977 , Iwasa et al . 6 1981 , McNamara 1982 , Olsson and Holmgren 1998 , Green 2006 , Olsson and Brown 2006 , van 5 Gils 2010 ). The patch-sample information need 4 not be complicated; it only needs the total

instantaneous intake rate instantaneous intake number of prey found in a patch and the total 3 search time elapsed ( Iwasa et al . 1981 ; Fig.

C 71A ). In most natural environments, where 7 prey densities follow a clumped frequency distribution, the perceived prey density at which such updating foragers give up their 6 patches would not be constant, but rather an increasing function of patch exploitation time (Fig. 71B ). This is because, in such environ-

potential intake rate potential intake 5 ments, every encounter with a new prey makes it more likely that there is more prey, and thus 0.0 0.10.2 0.3 0.4 0.5 0.6 0.7 perceived prey density increases with each search time consecutive prey encounter. Therefore, updating foragers should be more patient during the Figure 71. (A) Feeding on discrete prey yields a stepwise gain curve (line), with moments of prey-encounter alternating with initial phase of depleting a patch and ‘wait for bouts of no intake of rather unpredictable length. Optimal good news’ about patch quality in the form of moments to leave such a patch (grey dots) depend on the prey encounters. If they do so, they would combination of the number of prey items found and the total time leave patches at a constant potential intake rate it took to fi nd them. (B) Fluctuation of the estimated instantaneous ( Fig. 71C ), where the potential intake rate is intake rate for a Bayesian forager in a clumped environment the intake expected over the remainder of the experiencing the gain curve plotted in (A). This estimate goes down with search time, but jumps up with each prey encounter patch occupation ( Olsson and Holmgren 1998 ). (arrows). In order to maximize its long-term intake rate, the forager Everything comes to those who wait. should not leave its patch at a certain critical instantaneous intake rate (dashed line), but should instead be a little more patient and leave when the estimates have reached the grey dots. (C) By doing Do updating animals really exist? so, it leaves at constant estimated potential intake rates (i.e. the intake rate that potentially can be achieved during the rest of the If animals can combine prior information patch visit). about the environment with new information Modifi ed from Dall et al . ( 2005 ) (B–C). about a current option (be it a patch, a prey, a OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 123 mate, or whatever), they can markedly your chances would only increase from 1 in 3 improve their assessment of the value of the to 1 in 2, irrespective of whether you switch current option, and thus improve fi tness or not! (Olsson and Brown 2010 ). This is the essence Since the seminal paper on Bayesian forag- of ‘Bayesian’ updating, named after Thomas ing by Oaten (1977 ), relatively few tests of this Bayes (1702–61), an Anglican priest inter- rapidly expanding family of models exist. ested in probability theory. Just to give you a Nevertheless, a recent review paper strongly simple example of Bayesian updating in suggests that animals are capable of Bayesian practice: imagine a patch in which a forager updating: in ten out of the eleven studies that has found two prey items during the fi rst were compared, behaviour was consistent minute of search. Should it continue search- with Bayesian-updating predictions (Valone ing or should it leave this patch? If the envi- 2006 ). Most tests were performed on foraging ronment is structured such that a patch can animals, but there is also the case of female only contain two prey items at most, then the peacock wrasses Symphodus tinaca searching forager should defi nitely move on. By con- for males ( Luttbeg and Warner 1999 ). Even trast, if patches can contain many more than estuarine amphipods Gammarus lawrencianus just two items, then the forager should stay, use Bayesian updating when selecting their especially if it took little time to fi nd these mates ( Hunte et al . 1985 ). As echoed by Valone two items. Thus, knowing your environment (2006 ), the strongest support comes from stud- (in a statistical, probabilistic sense) greatly ies in which prior knowledge about the envi- improves your assessment abilities; without ronment is experimentally manipulated, such knowledge about your environment, it is as in the recently published work on Bayesian much harder to make the right choices. A bumblebees Bombus impatiens ( Biernaskie et al . popular example of Bayesian updating in 2009 ). In this study, bees were each offered humans is the so-called ‘Monty Hall prob- patches of 12 fl owers, of which five fl owers lem’ ( Selvin 1975 ), named after an American contained nectar (a 30% sucrose solution). The quizmaster. Imagine you are participating in bees were fooled by the other seven fl owers, as a TV game in which you have to select one they contained only water, and no sugar. out of three closed doors ( Fig. 72 ). Behind Optimal foraging in such a uniform environ- one of these doors stands a car, behind each ment would indicate a greater tendency to of the other two doors stands a goat. You will leave the 12-fl ower patch with each successive take home either a car or a goat, depending encounter of a nectar fl ower ( Fig. 73A ), because on which door you choose. Once you have each encounter means that the patch will har- selected your door, the friendly quizmaster, bour one nectar fl ower less in the future, as the who knows what stands behind each door, bee empties each encountered nectar fl ower. helps you by opening one of the other two And that is indeed what they do ( Fig. 73B ). doors for you, with a goat behind it. Knowing Now comes the evidence that the bees were this, or, in Bayesian terms, updating your prior indeed Bayesian bees: when offered a differ- expectation with new sampling information , ent environment, with highly variable patches should you switch to the other closed door? (in which the 12 fl owers contained either one The answer is yes; by switching, your chance or nine nectar fl owers), the bees switched pol- of winning the car increases from 1 in 3 to 2 icies according to the theory. The tendency to in 3. The intuitive, but incorrect, answer is stay in the patch should (Fig. 73C ) and did that there is no need to switch doors, since (Fig. 73D ) peak when encountering a second 124 ADDING BEHAVIOUR

NO SWITCHING SWITCHING

1 2

Host reveals 12either goat

Player picks car (probability 1/3) Switching results in the other goat

Host must 12reveal goat 2 12

Player picks goat 1 Switching wins (probability 1/3)

Host must 12reveal goat 1 1 2

Player picks goat 2 Switching wins (probability 1/3)

Figure 72. How to win a car that is behind one of three closed doors? Pick a door, and you will have a 1 in 3 chance of fi nding a car behind your door, provided you do not switch (three cartoons on the left). If you decide to switch after the quizmaster reveals one of the two goats (three cartoons on the right), your chance of winning the car increases to 2 in 3. The crucial point is this: when the quizmaster opens a door to reveal a goat, this does not give you any new information about what is behind the door you have chosen, so the chance of there being a car behind a different door remains 2 in 3; therefore, the chance of a car being behind the remaining door must be 2 in 3. Modifi ed from http://en.wikipedia.org/wiki/Monty_Hall_problem . nectar fl ower in that patch, since from then on a phenotype can respond to a changing envi- the bee can be sure that it is exploiting a rich ronment through behaviour! patch containing nine nectar fl owers (note Red knots can also behave like Bayesians, as that because of depletion, every subsequent revealed by the four captive birds exploiting encounter with a nectar fl ower in the same depletable patches in our large outdoor inter- patch reduces the potential value of the patch). tidal aviary ( van Gils et al . 2003b ). Though we The bees learned about the new environment did not manipulate their prior environmental in a matter of hours; a good example of the expectations, we were able to distinguish effectiveness, speed, and fl exibility with which between the best rule for Bayesians, in which OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 125

6 A C to forage like Bayesians. Ola Olsson and co- 3 workers (1999) cut off 354 branches on which 0 individual woodpeckers had been foraging for known times. By subsequently X-raying these –3 branches, they were able to count the number –6 of prey items remaining in them (the so-called

3 B D ‘Giving-Up Density’ GUD; Fig. 74A ). They found more prey items in those branches that 2 were occupied longest (Fig. 74B ), a diagnostic 1 tendency to stay for Bayesian foragers adopting the potential 0 assessment rule when exploiting clumped –1 prey distributions (Fig. 71B ). In fact, such posi- –2 tive correlations were not only found within –3 individual birds, but also between birds.

0 153 42 0126374859Realizing this, Olsson et al. then made an number of rewards important step. They reasoned: if animals respond more or less optimally to the quality Figure 73. Tendencies to stay in a patch of fl owers related to of their environment (e.g. travel distances, the number of nectar fl owers encountered within that patch: food availability), we can fl ip the optimality (A) as expected for the uniform environment; (B) the observed argument around and use the observed behav- tendency decreases with each successive encounter; (C) as expected for the highly variable environment; (D) the observed ioural variation between woodpeckers to make tendency peaks at the second encounter. inferences about the quality of their individual Modifi ed from Biernaskieet al . ( 2009 ). environments. High marginal intake rates at patch departure (equivalent to high GUDs) should, according to MVT, refl ect high long- they ‘wait for good news’ and allow low esti- term average intake rates. The idea that opti- mates of current prey density during the initial mal foraging theory could be used in a reversed phase of patch exploitation (dots in each of manner had been suggested much earlier by three graphs in Fig. 71 ; the so-called ‘potential Wilson (1976 ), but was only picked up after value assessment rule’; Olsson and Holmgren the seminal paper by Brown ( 1988 ). And it 1998 ), and the second best rule for Bayesians, yielded sensible patterns too: woodpeckers in which patches are left at a constant estimate leaving their branches at high GUDs did of current prey density (dashed horizontal line indeed raise more fl edglings than woodpeck- in Fig. 71B ; the so-called ‘current value assess- ers leaving them at low GUDs ( Fig. 74C ). ment rule’; Iwasa et al . 1981 ). The knots turned out to use the best rule, and thus to have the The psychology of decision-making patience to stay long enough in seemingly poor patches that initially rendered nothing Animals are not statisticians, and it is diffi cult because of bad luck. By doing so, they left their to conceive that animals can calculate probabil- patches at a constant potential intake rate ( Fig. ities. Even comparing the least amount of 67B as conceptualized in Fig. 71C ), rather than patch-sample information, the number of prey at a constant instantaneous intake rate. captured and the cumulative search time, with In a further remarkable study, lesser spotted the optimal stopping points may be a bridge woodpeckers Dendrocopos minor were shown too far (Fig. 71A ). Nevertheless, we have just 126 ADDING BEHAVIOUR

When prey densities follow a clumped dis- A tribution among patches (i.e. many poor patches and rather few good ones), a Bayesian forager’s estimate of the local prey density 4 would decrease with elapsed search time, but

) would make an upwards jump each time a 2 3 prey is encountered ( Fig. 71B and C and Fig. 73C ; or a fall, as in Fig. 73A in a uniform environment). This idea is based on purely mathe- 2 matical models, but, interestingly, the same idea occurred in an early empirical paper in average GUD (#/dm average 1 which the author speculated about the psy- B chology of patch departure in a parasitoid wasp Nemeritis ( Waage 1979 ; Fig. 75 ). Waage 0 0 10 20 30 40 50 did not use the terminology of a forager’s esti- average PRT (s/dm2) mate of prey density, but rather of a forager’s ‘tendency to stay in the patch’, a motivational 6 variable in the terminology of psychologists ( Geen 1995 ). In the words of Dick Green ( 1984 ), 5 think of this tendency ‘as being determined by a simple wind-up toy that originally is wound 4 up enough to run for one unit of time; but each 3 time a prey is found the toy is wound up enough to run for two additional units of time’. 2

number of fledglings number Evidence for such prey-induced increments/

1 decrements in motivation has accumulated C during the last few years, especially in the lit- 0 erature on parasitoid foraging ( Wajnberg 2006 , 0 12 average GUD (#/dm2) Biernaskie et al . 2009 ). In the form of survival analysis (notably Cox’s proportional hazards Figure 74. (A) X-ray image of a dead branch containing larvae model; Cox 1972 ), the incremental (or decre- (small, dark wormlike shapes) that remained after woodpeckers mental) effect of each prey encounter can be had searched for them. (B) Giving-up density (GUD) of larvae detected and estimated in terms of decreases increased as a function of patch residence time PRT, with each (or increases) in the chance of leaving the patch dot denoting the averages per individual bird. (C) Breeding at a given time ( Pierre and Green 2007 ). The success, expressed as the number of fl edglings per nest, increased as a function of an individual’s average GUD. realistic aspect of the proposed incremental/ Modifi ed from Olsson et al . ( 1999 ). decremental mechanism is that it does not require counting abilities on the part of the reviewed some evidence for Bayesian decision- animal: the forager can forget about previously making in animals, so the obvious next ques- encountered prey, it just needs to ‘wind up its tion is how they do this. Have animals evolved toy’ upon catching a new prey item (for which simpler ‘rules of thumb’ that may approximate it does not need much brain capacity; Holmgren to the optimal policies ‘invented’ by theoretical and Olsson 2000 ). What it does require though, ecologists ( McNamara et al . 2006 )? is a good sense of time, since the leaving OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 127

tendency goes down over time. Animals do measured the pecking intensity after the last have an idea of time (Aschoff 1954 ), but how prey of the patch had been offered, and showed accurately can they assess the duration of short that the starlings had a good sense of time. time intervals such as occur during patch Pecking intensity peaked exactly when the exploitation? food-item-that-never-came should have come Again it is starlings, those lab-rats of behav- ( Fig. 76 ). This result was upheld for different ioural ecologists, who can provide some of the interval durations ( Fig. 76), which provided answers. Brunner, Kacelnik, and Gibbon ( 1992 ) insight into the underlying mechanism. wanted to know if animals were able to assess The most popular mechanistic view of tim- the duration of time intervals. So they trained ing in animals is presented by the so-called captive starlings to reproduce time intervals ‘scalar expectancy theory’ (SET; Gibbon 1977 , behaviourally. In the forage-and-travel cages Shettleworth 1998 , Malapani and Fairhurst discussed earlier (Fig. 66C ), the birds could 2002 ). This assumes that a time interval is per- earn food by pecking at a so-called ‘pecking ceived, encoded, and then compared to a key’. However, food always came at fi xed time stored representation of similar intervals. intervals (Fig. 66A ), and it made no sense to However, the memory for these similar inter- peck before this interval had passed. Each vals is imperfect, as the animal draws a ran- patch ended with a sudden depletion, which dom sample from a distribution of values was made unpredictable by offering a variable centred on the currently perceived interval. number of food items per patch. Brunner et al . SET predicts that the variance of the memorized distribution increases with the length of the perceived time interval, but in such a way

4 oviposition effect oviposition time since last oviposition 3

0.8 2 responsiveness 1.6 pecks/s

come 3.2 time 1 6.4

timing of ovipositions should have when food 12.8 0 25.6 fixed interval (s) fixed twice three times Figure 75. The proximate mechanism of patch-departure after interval the fixed the fixed oviposition in the parasitoid wasp Nemeritis as envisioned by interval interval Waage ( 1979 ). The responsiveness or tendency to stay declines over time, but increases with each encounter with a host. How Figure 76. Starlings have a good sense of time. In an indoor much it increases, depends in a simple manner on the time since experiment where they had to peck for food, average pecking the last oviposition (inset). When the responsiveness drops to intensity peaked when the food item that never came should have zero, the wasp leaves the current patch. come, and this result was repeated for different fi xed intervals. Based on Waage ( 1979 ). Based on Brunner et al . ( 1992 ). 128 ADDING BEHAVIOUR that the coeffi cients of variation of all stored large home-range, there is no way a solitary distributions remain constant. Indeed, in the forager can keep up with all these dynamic starling experiment, the coeffi cient of variation resource changes just on its own. What may be did not change with the duration of the time a good patch today, can be a poor patch tomor- interval ( Fig. 76 ). Note, however, that compet- row, and where to fi nd good alternatives in ing alternative ideas about timing in animals that case? Joining knowledgeable individuals do exist. For example, support for a mecha- on their daily foraging trips will greatly nism called ‘learning to time’ (LeT) has been enhance survival probabilities in temporarily found in the psychologist’s lab rats—pigeons dynamic environments, and a communal roost ( Machado and Keen 1999 ). Irrespective of the or a breeding colony may be a good starting precise underlying mechanism, the message point for naïve foragers opting for a ‘guided here is clear: animals can keep track of time food tour’ ( Ward and Zahavi 1973 ). Think quite accurately. about scavenging birds, such as ravens or vultures, feeding on highly ephemeral food ‘patches’ that may only be around for a day or Ideal birds sleep together two. These birds, particularly ravens Corvus In this chapter we have so far considered for- corax, have been known actively to recruit con- aging animals as solitary individuals living in specifi cs from a central roost upon encounter- uncertain environments in which other indi- ing a sizeable carcass (Wright et al . 2003 ). But viduals play no role. However, the opposite is roosts may also be places where plenty of ‘pas- often the case: individuals of many species sive information’ is at hand: by-products of feed together, travel together, roost together, the performance or activities of others may and breed together. Although there are obvi- mirror the quality of the resources having been ous costs involved in living gregariously— fed upon. For example, fatness, plumage qual- think of competition for space, food, and ity, or roost arrival time may be indicative cues mates—there are also many benefi ts, and that are passively being advertised (so-called improved information gain is probably one of ‘inadvertent social information’; Danchin et al . them (Danchin et al . 2004 , Bonnie and Earley 2004 ), but may also provide the naïve individ- 2007 , Seppänen et al . 2007 , Valone 2007 ). For ual with useful information on the past forag- example, think of the Bayesian bird using its ing success of others (Bijleveld et al . 2010 ). Red own foraging success (‘personal information’) knots roost together too: thousands of individ- to assess patch quality. If this very same ani- uals get together on the shore whenever the mal shares this patch with others and is able to incoming tide fl oods their muddy foraging keep track of their success too (‘public infor- grounds. The following story suggests that mation’), it may reach much better estimates they too may greatly benefi t from each other in in a much shorter time (Valone 1993 , Valone fi nding their hidden food. and Giraldeau 1993 , Templeton and Giraldeau The null model of animal distribution, the 1996 , Smith et al . 1999 ). Scaling things up, so-called ‘ideal free-distribution’ model using public information may be essential to ( Fretwell and Lucas 1970 ), assumes ideal forag- keeping track of continuously changing ers, meaning that they are omniscient with environments. respect to the spatial distribution of their food. Let us stick to foraging: over time (days, They thus know where to go, and pay no price weeks), food stocks change due to prey growth, in getting there (the latter referring to the free reproduction, emigration, and mortality. In a aspect of the model). Both assumptions are of OPTIMAL BEHAVIOUR: CURRENCIES AND CONSTRAINTS 129 course over-simplifi cations, but the model has mudfl ats in between the mainland and the been very useful as a stepping stone in studies barrier islands (Kraan et al . 2009 ). At high tide, of interference competition (van der Meer and these birds get together at Richel, a wide, open Ens 1997 ). We wanted to see how far this model sandfl at of about 1 km 2 in the mid-north of the would bring us when trying to explain where area ( Fig. 77A ). During fi ve consecutive late red knots would feed in our ‘backyard’ study summers (1996–2000), we caught red knots at area, the Wadden Sea ( van Gils et al . 2006c ). this roost, radio-tagged them, and followed The western part of the Wadden Sea (Fig. their whereabouts throughout our study area. 77A ) is home to about 40,000 islandica knots, At the same time, we mapped their food which feed during low tide on the intertidal resources in great detail by sampling each

A RICHEL N VLIELAND Richel- Grienderwaard area GRIEND Ballast- plaat West

FRIESLAND 5 km

100 B

40 omniscient? travel costs? NO NO NO YES relative use of space (%) relative 20 YES NO YES YES 0

Westwad Richelwad Ballastplaat Grienderwaard Figure 77. (A) Map of the western Dutch Wadden Sea, with the roost at Richel in the mid-north and the four main feeding grounds distinguished (West, Richel-area, Grienderwaard, and Ballastplaat). (B) The use of these feeding grounds by red knots, predicted by the four different scenarios (symbols connected by lines), and as observed by radio-telemetry (grey bars, including standard errors). Modifi ed from van Gilset al . ( 2006c ). 130 ADDING BEHAVIOUR year the macrozoobenthos at nearly 2000 sta- close. With a highly dynamic food supply that tions laid out in a grid at 250-m intervals. is consumed by many, grows fast, and hides in Combining measured tidal heights with fi ne- the mud, it is out of the question that only per- scaled elevation data, we could predict the sonal sampling information can keep knots daily exposure time for each station. Food up-to-date. They must somehow exchange densities and exposure time were the main public information, be it actively or passively, input variables in a so-called ‘stochastic and their central roost is likely to play an dynamic programming model’ in which we important role as an information centre simulated red knot movements. We consid- ( Bijleveld et al . 2010 ). ered four possible scenarios. If knots had no With this report from the Wadden Sea, we knowledge about their environment but were have set the scene for the next chapter(s), in to travel for free, they would distribute them- which ‘Wadden Sea knots’ will feature again. selves randomly, only respecting the differen- There will be one important difference, how- tial availability of mudfl ats due to differences ever. So far we have assumed that all knots in elevation (fi lled dots in Fig. 77B ). If, instead, have similar bodies. In the next chapter, we these naïve birds travelled at a cost, they will focus on these individual red knots, and would not get too far from the roost and all we will see that different individuals do differ- hang around near Richel during low tide ent things. Much of this behavioural variabil- (open dots in Fig. 77B ). If they ‘would know ity can be explained by their fl exible bodies, better’, have complete knowledge of their bodies that we ‘looked into’ before releasing ‘foodscape’, and travelling was for free, they them once again into their wide open windy would go to the best feeding grounds landscape. (Ballastplaat), as much as the tide allowed them (this particular mudfl at lies relatively Synopsis low and is therefore exposed relatively briefl y; closed triangles in Fig. 77B ). On the other Behaviour is a rapid way for animals to hand, if such knowledgeable birds did not respond fl exibly to a changing environment. travel for free, they would get to Ballastplaat The study of behaviour has often been much less often, as this site is located furthest approached from a functional angle, which from the roost, and would instead spend more has allowed researchers to make clear-cut, fal- time feeding nearby (but skipping the food- sifi able predictions about behaviour. Knowing poor area called ‘Westwad’; open triangles in the constraints faced by an animal (including Fig. 77B ). informational constraints), and assuming a The last prediction matched best with the currency that approximates fi tness, research- movement data of our radio-tagged birds (bars ers can qualify decisions as optimal or subopti- in Fig. 77B ): knots seem to be ideal, but not mal. This chapter provides a stepping stone to free. It was no surprise to us that they were not the next chapter, where we adopt this success- free, since they have to pay a price for travel- ful optimality approach to interpret and pre- ling, but we did not expect that the fi t with the dict fl exible organ size changes within ideal/omniscience prediction would be so shorebirds. CHAPTER 7 Optimal foraging: the dynamic choice between diets, feeding patches, and gut sizes

Eating more by ignoring food refrain from consuming every prey item that it encountered. A decade later, the model was In the previous chapter we demonstrated the refi ned by others (e.g. Pulliam 1974 , Werner and intellectual elegance of behavioural optimality Hall 1974 , Charnov 1976b ), but the general reasoning, a concept that enables ecologists to structure and predictions remained unchanged. make stepwise scientifi c progression by means The basics are these: a forager can either search of Popperian falsifi cation of quantitative predic- for prey, or handle an encountered prey item, tions. In this chapter we will show how our for- but it cannot do both things at the same time. aging work on red knots was built on this Therefore, handling a prey item carries a time approach, step by step. Initially we thought that cost, as searching has stopped for a moment. the standard optimal-diet model of behavioural Sometimes it may be more fruitful to neglect an ecology would be suffi cient to explain the com- encountered prey item and to continue search- position of the menu of red knots functionally. ing for a better one. This is the case when the However, this proved not to be the case, and ‘intake rate while handling the item’ is less than only after including the process of digestion into the maximum possible intake rate averaged the diet model were we able to explain diet selec- over a long period of foraging. The ‘intake rate tion. This in turn led to a deeper understanding while handling’, usually referred to as ‘profi ta- of the knots’ related selection criteria, notably of bility’, is the key parameter in the model and the habitats they made use of, and the strong dif- equals the ratio between the energy content of a ferences between different phenotypes in these prey item and the required handling time. The selection criteria. Not only could we then under- main prediction of this simple diet model is thus stand why different phenotypes responded dif- clear cut and intuitively appealing: in rich envi- ferently to the different food patches on offer, but ronments, where high long-term average intake also how the idea worked in reverse—the food rates can be achieved, foragers should be more supply moulding the phenotypes. Elaborating selective; in poor environments, diets should be these observations will be the aim of this chapter. broader. Its simple elegance also appealed to the But let us begin fi rst with a brief review of the godfather of ‘behavioural ecology’, Lord John R. standard diet model that we abandoned so soon Krebs, who tested this model in captive great after embracing it. tits Parus major . Back in 1966, Robert MacArthur and Eric Krebs et al . ( 1977 ) offered their tits segments Pianka developed a simple but effective model of mealworms as prey. These came in two that showed how a forager, trying to maximize forms: long pieces (eight segments long) and its long-term average food-intake rate, should short pieces (four segments). By adding a small

131 132 ADDING BEHAVIOUR

piece of plastic tape to the short ‘worms’, the ate many more large ones than small ones ( Fig. handling times of both types were kept more 78 ). However, even at this high intake rate, or less similar (the birds had to keep this prey they did not ignore the smaller worms com- under one foot to peel off the tape). In this way, pletely. Krebs and co-workers suggested that the profi tability of each type varied by a factor the birds were actually sampling the smaller of two—either four or eight segments of energy prey every so often as it passed by because per prey. The prey items were offered sequen- they needed to test its profi tability. Such sam- tially by putting them on to a conveyer belt pling was an activity not included in the origi- one-by-one. The eight-segment long—most nal model. profi table—prey passed by at three different This study by Krebs et al. is one of the very rates. The crux of the experiment is that only at many examples that show that the classical the highest encounter rate would the tits’ long- optimal-diet model is a ‘strong’ model, a model term intake rate be above the profi tability of that has successfully been ‘confi rmed’ many the shorter worms. Thus, only at this highest times since its conception in 1966 (reviewed by delivery rate should the birds ignore the Schluter 1981 , Stephens and Krebs 1986 , Perry shorter worms. And that is indeed more or less and Pianka 1997 , Sih and Christensen 2001 ). what they did. At the two lower prey encoun- Perry and Pianka (1997 ) conclude that the nar- ter rates the birds were not selective and took rowing down of a forager’s breadth of diet small and large worms in equal proportions, with an increase in food abundance ‘may be while at the highest encounter rate the birds optimal foraging theory’s most robust theorem to date’. On the other hand, the classical optimal-diet model has also been refuted many NON-SELECTIVE SELECTIVE times. Sih and Christensen ( 2001 ) conclude 100 that empirical studies on immobile prey yield 90 results that are usually in line with the predic- 80 tions of the model, whereas studies on mobile

70 prey often refute them. Perhaps this should not have surprised us. An important aspect of 60 prey behaviour, the possibility of escaping profitability (mg/s) of unprofitable prey profitability (mg/s) of unprofitable 50 when encountered, is not dealt with in the

percentage profitable prey in diet prey percentage profitable 40 model, and thus the standard model should

123456 not have been tested in cases involving mobile intake rate (mg/s) if feeding on profitable prey only prey. We should have realized that the standard model captures the trade-off between Figure 78. In an experimental situation, profi table and wasting time and gaining energy for a specifi c unprofi table prey were offered to great tits. The tits largely situation, and cannot be applied directly to all obeyed the optimal diet hypothesis by being non-selective under relatively low intake-rates, whilst selecting (mostly) for possible foraging scenarios. As it was, with red the profi table type under relatively high intake-rates. The thick knots feeding on highly sessile mollusc prey, stepwise line and the shading give the theoretical prediction, we initially expected a close match between where the shift from being non-selective (non-shaded) to being the standard theory and our observations. In selective (grey shaded) occurs when the intake rate on the fact we found the opposite: a clash between profi table type only would be equivalent to the profi tability of the unprofi table type, which is just over 4 mg/s. Dots are means the theory and the facts. Could the prey’s hard- and error bars give 95%-confi dence intervals. shelled armour be the possible explanation for Modifi ed from Krebset al . ( 1977 ). these discrepancies? THE DYNAMIC CHOICE BETWEEN DIETS, FEEDING PATCHES, AND GUT SIZES 133

A hard nut to crack the intake rate in a fi eld situation. Here knots need to search for their food rather than to pick Red knots are mollusc-specialists (see out items from a tray. Digestion can take place Chapter 1 ), but, remarkably, molluscs are not during foraging, and as long as food is found their favourite meal. If captive knots are at a slower pace than the digestive system can offered the choice between 15-mm long hard- process it, there is no digestive constraint shelled cockles Cerastoderma edule and 50-mm ( Jeschke et al . 2002 ). However, when prey items long shrimps Crangon crangon, they will go are found and handled faster than they can be for the shrimps ( van Gils et al . 2005a ). Only processed (as in ‘luxurious’ experimental set- after depleting the tray of shrimps will the tings), knots face a digestive bottleneck. In this birds switch to the cockles. With only case there is a premium on picking out those MacArthur and Pianka’s (1966) diet model in food items that contain the highest amount of mind, this diet choice cannot be explained, as metabolizable energy per unit shell material. a shrimp is more energy rich (1–2 kJ of metab- The cockles in the choice-experiment had olizable energy) than a cockle (less than half about 400 mg of shell mass each; the large a kJ) but takes proportionally more time to shrimps had only 80 mg of indigestible cara- handle (more than half a minute rather than pace mass. This makes the ratio of metaboliz- the fi ve seconds or less required to handle a able energy per mg ballast material, the cockle). This makes shrimps half as profi table so-called ‘digestive quality’, 20 times higher (40–50 Watts) as cockles (80–90 Watts). Knots for shrimps (1500 J per 80 mg) than for cockles should like cockles better than shrimps! Why (350 J per 400 mg). This potentially enables does this not seem to be the case? The answer lies in the knots’ habit of ingesting their prey whole, even shelled prey. 1.000 Because of this, they need to crush their food in their guts, and they do this in their muscu- S lar stomach, the gizzard, before digestion can 0.100 S commence (Piersma et al . 1993b ). The crushing and the subsequent processing of the shell material take time. In fact, in an experimental M M 0.010 L setting where knots were offered no choice, (prey/s) rate intake L but only one prey type at a time in large quantities, we found that this time-cost scaled 0.001 directly with the amount of indigestible shell 1 10 100 1000 material in any prey item: the more shell mate- shell mass (mg/prey) rial a prey contained, the longer the retention Figure 79. No matter whether a prey type contained a little or time in the digestive system (van Gils et al . a lot of shell mass per item (horizontal axis), the knots’ intake 2003a ). This constraint turned out to be very rate (vertical axis) would be constrained by a constant maximum robust: whichever prey type that we offered, rate of processing shell mass of about 2.6 mg/s (fi tted regression our hungry knots, ‘with eyes bigger than their line with a slope of –1 for log-transformed axes). Six different prey-types were offered: Baltic tellins Macoma balthica (squares) stomach’, could not ingest their prey faster and cockles Cerastoderma edule (dots), both in three different than a rate of 2.6 mg of dry shell material per size classes (S = small; M = medium; L = large). Symbols give second ( Fig. 79 ). Nevertheless, this ultimate means, bars give standard errors. processing rate need not necessarily constrain Modifi ed from van Gilset al . ( 2003a ). 134 ADDING BEHAVIOUR knots to metabolize energy from shrimps (2.6 material digestively (black dot in Fig. 81 ). mg per second × 1500 J per 80 mg = 48.8 Watts) Because knots have such a large gizzard (rela- at a rate that is 20 times faster than from cock- tive to body mass their gizzards are among the les (2.6 mg per second × 350 J per 400 mg = 2.3 largest of any shorebirds; Battley and Piersma Watts). Are you still following us? We have 2005 ), this capacity is relatively high (2.6 mg of just given you the explanation for the knots’ preference for shrimps over cockles when either come in excess quantities. Our observa- A tions fi nally knocked the classical diet model patch B on the head, at least in situations where knots were eating invertebrates ( van Gils et al . patch A 2005a ).

GRIEND If knots like shrimps so much better than 2 km cockles and other bivalves, why have they 100 become mollusc-specialists and eat cockles B 80 and other hard-shelled bivalves most of the soft time? Why are they not shrimp-specialists or, 60 more generally, crustacean-specialists? Red 40 Available knots are sensorily specialized to detect hard- 20 hard prey density (%) prey shelled prey in soft wet sediment (Piersma et 0 al . 1998 ), but the answer lies in a crucial differ- 10 C ) ence between our laboratory studies and the 2 fi eld situation. In the wild, food does not occur in excess quantities. It is not laid out conspicu- 1 ously in plastic trays, but scattered and hiding (g AFDM/m in the sediment. Knots need to search for it. 0.1 What is more, shrimps are relatively rare, biomass prey Available 8 D while cockles and other low-quality molluscs 6 are relatively abundant. In addition, shrimp 4 patches contain rather few cockles and typical cockle patches harbour low shrimp densities (J/mg)

(Fig. 80 ). For this reason, patch choice comes quality Digestive 2 before prey choice: if knots want to feed on A B shrimps, they need to go somewhere other Patch than if they plan to feed on cockles. Due to the Figure 80. (A) Hard and soft food is often found in different relatively low abundance of shrimps and other localities, exemplifi ed here for the late-summer situation in 1998, high-quality crustaceans ( Fig. 80C ), knots’ near the isle of Griend, western Dutch Wadden Sea. (B) Virtually intake rates when feeding in a typical shrimp- all prey in patch A were hard-shelled, while in patch B about half patch are constrained by the low density of the were soft bodied. (C) Patch A contained an order of magnitude shrimps (grey dot in Fig. 81 ). In contrast, due more ash-free dry mass (AFDM) of food per square meter than patch B, (D) but of much lower quality, largely due to the absence to the relatively high abundance of cockles, of soft-bodied crustaceans. Box-and-whisker plots in (C) and and due to their heavy shells, knots’ intake (D) give mean (large dot), median (horizontal line within box), rates when feeding in a typical mollusc-patch inter-quartile range (box), range (bars) and outliers (small dots). are constrained by the capacity to process shell From van Gils et al . ( 2005b ). THE DYNAMIC CHOICE BETWEEN DIETS, FEEDING PATCHES, AND GUT SIZES 135

4 It takes guts to eat shellfi sh Every year, during the late-summer arrival 3 period, we catch red knots at Richel, their main shrimp high-tide roost in the Dutch Wadden Sea. The 2 usual aim is to colour-ring them in order to estimate year-specifi c survival rates (see cockles 1 Chapter 9 ). On some of these occasions, we energy intake rate (W) rate energy intake also glued tiny VHF-radios (1.3–1.8 g) to the back of individual knots. This enabled us to 0 0 1234 ascertain which individuals visited the shrimp- prey biomass (g AFDM/m2) and crab-patches and which did not. Figure 81. Functional responses of knots to densities of Initially, we found no patterns in the behav- shrimp (grey line) and cockles (black line). Cockles usually occur iour. In terms of body size and body mass or in much higher densities than shrimps. Therefore, long-term sex, the ‘odd’ crustacean-feeders were indis- intake rate in a typical cockle patch (black dot) is higher than in tinguishable from their shellfi sh-feeding coun- a typical shrimp patch (grey dot), even though the intake rate in terparts. However, things became clearer when the cockle patch is constrained by the digestive processing rate (black horizontal line as opposed to the black dotted line, which we began to use ultrasonography. By measur- would be the intake rate of cockles without a digestive constraint; note that in the latter case the asymptote of the ‘cockle functional response’ would exceed the asymptote of B A the ‘shrimp functional response’, as cockles are more profi table than shrimps). Scaling of functional responses is based on both patch experimentally determined parameters. A

1.0 B dry shell material per second). This means that 0.8 the digestively constrained intake rate when 0.6 they are feeding on cockles and other molluscs is usually higher than that when they are feed- 0.4 ing on high-quality crustaceans, as the latter is proportion of constrained by the density of the crustaceans 0.2 high-quality prey in diet high-quality prey (Fig. 81 ). This would explain why knots under 0.0 fi eld conditions usually ignore the soft food 0 2 46810 12 that they like so much in the laboratory. gizzard mass (g) However, there is yet another element in Figure 82. (A) The hard-shelled-prey patch A ( Fig. 80A ) was this fascinating story! Each year during late visited by knots having a large gizzard, while the soft-bodied- summer, the time when knots return from their prey patch B was visited by knots having a small gizzard. Birds tundra breeding grounds to the Wadden Sea, that alternated between both patches had intermediate gizzard we fi nd fl ocks of knots that do feed in the rela- sizes (for explanation of box-and-whisker-plots see Fig. 80 ). (B) tively poor shrimp- and crab-patches. These As predicted by the Digestive Rate Model (grey bars, see p. 138), and as empirically confi rmed by dropping analyses birds seem to ignore the much more abundant (dots and error bars, representing means and standard errors), cockles and other molluscs. In order to explain the proportion of soft-bodied prey in the diet declined as a this, we need to call organ fl exibility into the function of the knots’ gizzard mass. equation at this point. From van Gils et al . ( 2005b ). 136 ADDING BEHAVIOUR

ing gizzard width and height ultrasonographi- food? This may be the case, but it did not seem cally, we were able accurately to assess gizzard to make sense from an optimality point of mass in live knots ( Dietz et al . 1999 ). This view, since intake rates are much higher when turned out to be a methodological break- feeding on abundant hard-shelled food than through. These measurements revealed that on scarce crustaceans ( Fig. 81 ). However, these the birds with the smallest gizzards fl ew to the calculations were made for birds with large crustacean-patches, birds with the largest giz- gizzards. If knots with a small gizzard differ in zards visited the cockle banks, whereas birds their digestive processing rates from knots with intermediate gizzard masses alternated with a large gizzard, then consumption of soft between both ( Fig. 82 ; van Gils et al . 2005b )! food may make sense after all. We had discovered phenotype-related habitat Their capacity for phenotypic change could selection. However, these observations did not now be used in an experiment. By varying yet explain why crustacean-patches were vis- the hardness of the food on offer we experi- ited (in fact, we simply reasoned that feeding mentally manipulated gizzard size in order on hard-shelled food is always more benefi - to study the functional consequences in terms cial). Moreover, it left us with a new issue, a of digestive capacity. A wonderful tool that true ‘chicken-and-egg problem’. Did the soft- had been so successful in behavioural ecol- food feeders eat soft food because their giz- ogy, the experimental manipulation of a zards were so small, or were their gizzards so behavioural choice (see Chapter 6 ), could small because they ate soft food? To answer this question, we needed to go back to the laboratory. We already suspected wild Knots (n = 72) 20 that gizzards become smaller when birds eat well fed starved food that is easy to process. For example, our 10 captive knots are kept on a diet of artifi cial, protein-rich trout-food pellets, a form of soft 0 food that does not require crushing. Dissecting captive Knots (n = 10) some of those captive birds revealed gizzards 40 of around 3 g, while the average knot in the wild has a gizzard that is three times as large 30 ( Fig. 83 ; Piersma et al . 1993b ). Nevertheless, the proof of the pudding is in the eating! Thus, relative frequency (%) relative 20 we performed a controlled experiment to test our ‘gut feeling’ about food affecting gizzard size. By alternately offering captive knots soft 10 and hard food, and monitoring their gizzards 0 1 2 3 cm ultrasonographically, we found that the giz- 0 234567 8910 11 12 13 14 15 zards varied in mass synchronously with the fresh mass of gizzard (g) dietary regime: over a period of about a week’s time they became smaller when birds fed on Figure 83. Wild knots (top graph) have much larger gizzards soft food and became larger when birds ate than captive knots (bottom graph), even if the former have starved to death in severe winter weather (darker bars in top hard-shelled food ( Fig. 84 ; Dekinga et al . 2001 ). graph). Inset in bottom panel shows cross-sectioned gizzards of Did some knots develop a small gizzard sim- a 130-g wild (top) and a 125-g captive (bottom) knot. ply because they happened to feed on soft From Piersma et al . ( 1993b ). THE DYNAMIC CHOICE BETWEEN DIETS, FEEDING PATCHES, AND GUT SIZES 137

pellets pellets that, across different prey types varying in the amount of shell mass per item, the amount of 12 shell mass processed remained constant within each ‘gizzard-group’ (at 0.24 mg/s in the small 10 gizzards and 2.58 mg/s in the large gizzards; 8 Fig. 85A ). The result was upheld when allow- 6 ing small-gizzard birds to develop a large

gizzard mass (g) gizzard and vice versa . Between individuals, 4

2 1.000 A 0 0 10 20 30 40 50 60 70 time (days)

0.100 2.58 mg/s Figure 84. Flexible changes in the gizzard of captive red knots that were offered hard and soft food alternately: coming 0.24 mg/s straight from the fi eld where the birds ate hard-shelled bivalves, they were offered soft trout pellets for the fi rst 2½ weeks of 0.010 captivity, which halved their average gizzard size. Then they intake rate (prey/s) were offered hard-shelled mussels Mytilus edulis for more than a month, during which gizzard mass increased back to fi eld 0.001 dimensions (matching our expectations, gizzards were largest 1 10 100 1000 when the birds were offered large mussels, which have lower shell mass (mg/prey) digestive qualities than the small mussels offered later). This 10.0 pattern was repeated throughout the rest of the experiment: gizzard atrophy when given pellets and gizzard hypertrophy 2 when back on mussels again. These repeated measurements . G P = c within the same individuals could only be done with the use of ultrasonography. 1.0 From Piersma and Drent ( 2003 ).

now be applied to manipulate a ‘physiologi- B

P = processing rate (mg shell/s) 0.1 cal choice’, an individual’s gizzard mass. Thus, we manipulated gizzard mass in a few 24683 579 G = estimated gizzard mass (g) captive knots to obtain two groups: birds with small gizzards and birds with large giz- Figure 85. (A) Intake rate is constrained by the rate of zards. We then offered ad libitum hard-shelled processing shell material, with the latter differing tremendously food to these groups, one type of prey at a between large and small gizzards. Whereas birds in the large- gizzard group were able to process 2.58 mg of dried shell time, but in large, excess quantities. The birds material per second (fi lled symbols), birds in the small-gizzard were kept hungry, which motivated them to group could only process 0.24 mg per second (open symbols). feed at maximum rates. Prey species and size classes as defi ned for Fig. 79 . (B) Lumping The results were very clear cut. Maximum the different prey species and size classes together, but plotting feeding rates were one order of magnitude shell-mass processing rate against an individual’s gizzard mass, we found a quadratic relationship between the two. Box-and- lower for knots with small gizzards than for whisker-plots, as explained in Fig. 80 , with grey-fi lled boxes for knots with large gizzards (Fig. 85A ; van Gils birds in the large-gizzard group and non-fi lled boxes for birds et al. 2003a ). Processing shell material was in the small-gizzard group. clearly the bottleneck, as indicated by the fact From van Gils et al . ( 2003a ). 138 ADDING BEHAVIOUR

the shell mass processing rate increased quad- digestive capacity (Fig. 86A ). At a given prey ratically with an increase in gizzard mass (Fig. density (solid line in Fig. 86A ), intake rate is 85B ). With this simple quadratic function to constrained by rates of processing for small hand, we could now go back into the fi eld and gizzards, and thus an increase in gizzard mass recalculate the intake rates of the different leads to an increased intake rate. However, phenotypes in the different food patches. above a certain gizzard mass, processing For each possible gizzard mass, at each food capacity is so high that it exceeds the rate at patch, we calculated the maximum intake rate which food is found and ingested, and thus it that a feeding knot could attain. When there becomes the food density that constrains the are multiple prey types within a single patch, intake rate at large gizzard masses (thus obey- this maximum is predicted by the so-called ing Holling’s (1959) classical disc equation: Digestive Rate Model (DRM), which not only Fig. 86A ). takes account of a prey type’s abundance, but Now, when there are two types of patches, also of its digestive quality (see Hirakawa 1995 those containing low densities of high-quality for details and van Gils et al . 2005a , 2005b food (crustacean patches) and those contain- and Quaintenne et al . 2010 for applications). ing high densities of low-quality food (cockle Generally speaking, the model predicts that patches), then knots with small gizzards will intake rate depends on both food density and achieve their maximum intake rate in low-

A eat high-quality prey eat low-quality prey 15 H 30 o ll in g 's d is 10 c curve 20

0 energy intake rate (W) 1 B

energy intake rate (W) rate intake energy 0

00 10 H pre 800 15 y density600 (# mix 10 400 L 5 C 200 food quality / m 0 gizzard mass (g) 2 0 ) 0 246810 12 gizzard mass (g)

Figure 86. (A) The functional response of red knots potentially facing a digestive constraint. With an extremely large gizzard of 15 g, intake rate increases at a decelerating rate, as given by Holling’s disc model. However, for smaller gizzards, the intake rate is only given by the disc model for low food densities. At higher food densities, intake is constrained by the rate of processing shell material, which increases quadratically as a function of gizzard mass. Solid line gives the dependence of intake rate on gizzard size, e.g. food density of 100 items per m 2 . (B) Energy intake rate as a function of gizzard mass in two different patches: a high-density patch, but offering low-quality food (solid line, the same as the solid line in (A)) and a low-density-but-high-quality patch (dashed line). This shows that rate-maximizing knots should select the high-quality food when their gizzard is small (< 6 g), but should switch to low- quality food when their gizzard is large (³ 6 g). (C) As we have seen before (Fig. 82A ), this is exactly what our radio-marked knots did. Modifi ed from van Gilset al . ( 2005b ). THE DYNAMIC CHOICE BETWEEN DIETS, FEEDING PATCHES, AND GUT SIZES 139 density/high-quality patches, while knots timing the daily departure and arrival from/at with large gizzards will achieve their maxi- the central roost of our radio-tagged knots mum intake rate in high-density/low-quality ( Fig. 87A ; see Fig. 87B on how they do this). patches ( Fig. 86B ). Put simply, this is because Besides the risk of not fi nding enough high- knots with large gizzards hardly ever face a quality food, such long working days are an digestive constraint. For them food quality is obvious cost of having a small gizzard (see unimportant, it is the food quantity that mat- Chapter 4 for a discussion of the long-term ters. For small-gizzard knots facing a stringent survival costs of hard work). Therefore, one digestive bottleneck, it is the other way around: obvious new question cries out for an answer: quality matters, quantity does not (above some if having a small gizzard is so costly, and if giz- threshold minimum density). Thus, although zard size is a fl exible trait that individuals can food choice will determine gizzard size, it is fi ne-tune themselves, why do some individu- now clear, from a functional perspective, that als ‘decide’ to have a small gizzard? gizzard size also determines food choice (Fig. The answer has to do with the migratory 86C ). lifestyle of knots. Because the costs of fl ight Even though knots with small gizzards increase with body mass (Kvist et al . 2001 ), attain their maximum intake rate at soft-food long intercontinental non-stop fl ights are likely patches, their intake rate is still lower than in to be cheaper when unnecessary body weight knots with large gizzards feeding in hard- is ‘thrown overboard’. Indeed we found that, shelled mollusc patches ( Fig. 86B ). To compen- in the days before take-off for a long fl ight, and sate for this lower intake rate, knots with small in the days just after their return, digestive gizzards are predicted to feed longer on a daily organs in knots are relatively small ( Piersma basis. And indeed they do, as we found out by et al. 1999a , 1999b ). This also occurs in other

18 constraint set by high tide A B 0.5 1.0 Richel 1.5 12 14 2.0 Grienderwaard 15 16 15 0.0 Ballastplaat 17

Richel-flats tide extension 12 13

0.0

9 12 daily feeding time (h)

6 10 km

0 2 4 6810 12 gizzard mass (g)

Figure 87. (A) The smaller their gizzard, the longer the radio-marked knots were away from their roost at Richel, feeding on the mudfl at instead (means ± standard errors). (B) In the western Dutch Wadden Sea, knots can extend their daily feeding period by 4–5 h beyond the usual 12 h by moving eastwards along with the outgoing tide (numbered dots; also tide-isoclines are indicated, expressed in hours delay per tidal cycle relative to the tide at Richel). Inset shows a knot with a small (left) and a knot with a large (right) gizzard. Modifi ed from van Gilset al . ( 2005b , 2007a ). 140 ADDING BEHAVIOUR

migratory shorebirds ( Piersma 1998 , Piersma A and Gill 1998 ), in waterfowl ( van Gils et al . 600 Kirkwood-Kvist/Lindström 2008 ), and in songbirds ( Bauchinger et al . 2009 ). 400 Consequently we think that the inter-indi- intake vidual variation in gizzard masses, which we see at any given moment during the late-sum- 200 expenditure mer arrival period, is due to different individ-

uals arriving in the Wadden Sea on different satisficing 0 net rate-maximising dates ( Dietz et al . 2010 ). Thus, birds with small increased digestive prey quality B gizzards are likely to have just arrived and are 600 in the process of growing a larger gizzard. It Kirkwood-Kvist/Lindström takes at least a week (under relaxed laboratory conditions; Dekinga et al . 2001 ) and up to three 400 re intake itu end weeks (under harsher, natural, fi eld condi- exp tions; Piersma et al . 1999a ) to increase gizzard 200 size from small to large. Large-gizzard knots are, therefore, likely to have been around for a couple of weeks already. For several reasons, 0 departure dates from the arctic differ between increased thermoregulatory costs C individuals: males care for the brood, females 600 Kirkwood-Kvist/Lindström leave right after hatching, juveniles need some daily energy expenditure or metabolizable intake (kJ) extra weeks to complete their growth, and 400 failed breeders leave before successful breed- intake ers ( Tomkovich and Soloviev 1996 ). Thus, we expenditure do not think that there are typical small-giz- 200 zard knots and typical large-gizzard knots, but rather that timing differences between indi- 0 viduals are causing gizzard mass differences 0 5 10 15 20 between individuals at any given time. gizzard mass (g) Figure 88. (A) A knot can optimize its gizzard size in two ways. First, by balancing energy income with energy expenditure Optimal gizzards (which both increase as a function of gizzard mass), leading to We have now come full circle with the giz- a ‘satisfi cing’ gizzard. Second, by maximizing the difference between energy income and energy expenditure, leading to a zard—food interplay (the ‘chicken-and-egg ‘net rate-maximizing’ gizzard. (B) An increase in the food’s problem’): gizzard size determines food choice, digestive quality will reduce the optimal size of both gizzard but food also determines gizzard size. The lat- types. (C) By contrast, an increase in a knot’s energy expenditure ter can now be understood functionally: if (e.g. through an increase in thermoregulatory costs) will enlarge knots on a daily basis need to assimilate as the optimal satisfi cing gizzard, while leaving the rate-maximizing gizzard unaltered. much energy as they expend, and if there is a From van Gils et al . ( 2005c ). cost to carrying and maintaining a large gizzard, then it pays to maintain a gizzard that is just large enough to obtain the daily necessary coined the term ‘satisfi cing gizzard’; solid amount of kiloJoules (a gizzard for which we arrow in Fig. 88A ). Soft, high-quality food, THE DYNAMIC CHOICE BETWEEN DIETS, FEEDING PATCHES, AND GUT SIZES 141

with little shell mass relative to the amount of 100 4 meat, can be processed rapidly (solid arrow in A 80 Fig. 88B ) and thus would require a smaller giz- 3 zard than hard, low-quality food (solid arrow 60 in Fig. 88A ). Thus, offering soft trout-food pel- 2 lets yields small gizzards, while offering hard- 40 Crustacea shelled mussels yields large gizzards (Fig. 84 ; Mya

diet composition (%) 1

20 Hydrobia prey quality (kJ/g shell) Dekinga et al . 2001 ). A proximate mechanism Cerastoderma Macoma explained functionally. 0 0 JMSDFMA J J A ON This reasoning implies that we can think of time of year gizzards as having an optimal size for each 300 B 12 particular ecological context. They can be too 1 11 small, thus not yielding enough daily energy, 2 kJ/g shell 10 200 D or they can be too large, thus costing too much N J 9 energy to carry and maintain (van Gils et al . 3 S O F 8 2006a ). Getting back to the optimality models M 4 Kvist 7 100 A 5

that had worked so well in behavioural ecol- ood- A 6

required gizzard mass (g) kw ogy (see Chapter 6 ), it occurred to us that we daily shell mass intake (g) 7 6 J M Kir may be able to predict optimal gizzard sizes for 8 9 300 400 500 600 different ecological scenarios, in which both daily energy requirement (kJ) food quality and metabolic requirements vary 15 independently. This would imply that the C smallest gizzards are best when food quality is net rate-maximizer high under low metabolic demands, while 10 large gizzards are required in the opposite case ( Fig. 88 ). satisficer We added one extra dimension to these gizzard mass (g) 5 predictions. As knots are long-distance migrants, they do not always just need to bal- 0 JMSDFMA J JA ON ance energy income with energy expenditure. time of year There are times of the year, when (net) income Figure 89. (A) Monthly composition (bars scaled on left axis) needs to be maximized, which is when the and quality (dots scaled on right axis) of the diet of knots birds are fuelling up for migration. At these wintering in the Dutch Wadden Sea. (B) How diet quality times, they may need a gizzard that is larger (diagonal lines, having a slope of 1) and daily energy expenditure than a satisfi cing gizzard, a gizzard that is so (horizontal axis) together determine the amount of shell material large that it is possible for a knot to reach the that needs to be processed daily (left vertical axis), and thus determine the require gizzard mass (right vertical axis) for upper limit to its daily metabolizable energy satisfi cing knots (monthly dots connected by lines in lower left intake, a gizzard for which we coined the corner) and for rate-maximizing knots (monthly dots in vertical term a ‘rate-maximizing gizzard’ (open Kirkwood-Kvist bar). (C) Gizzard masses of islandica knots in the arrows throughout Fig. 88 ; this upper limit Dutch Wadden Sea fi t the satisfi cer prediction throughout most was determined by Kvist and Lindström of the year, except during periods of fuelling, when gizzards fi t ( 2003 ) when they offered easily digestible the rate-maximizing prediction (especially in spring, but even during early winter when fuelling for a small midwinter body mealworms in excess quantities to captive mass peak). fuelling red knots; see Fig. 43). From van Gils et al . ( 2003a ). 142 ADDING BEHAVIOUR

Formulating the proper cost and benefi t func- consecutive years (1998–2002), with 1998 offer- tions (plotted conceptually in Fig. 88 ), we were ing the best-quality food and 2002 the worst- now able to generate month-specifi c, year- quality food. This variation was refl ected in round predictions for the optimal gizzard mass the birds’ gizzards, with the smallest gizzards of knots wintering or fuelling in the Dutch found in 1998 and the largest ones in 2002 ( Fig. Wadden Sea (van Gils et al . 2003a ). The benefi t 90 ). In Chapter 9 we will come back to this function was determined by the seasonal varia- gradual decline in food quality over the years, tion in diet quality, as observed through our and discuss its causes, and consequences, for analysis of knot-droppings (Fig. 89A and diago- red knot survival. nal lines in Fig. 89B ); the cost function was given With the gizzard model working so well in by the seasonal variations in metabolic require- our own backyard, the Dutch Wadden Sea, we ments, which were assessed earlier using heated started thinking globally, and wondered whether taxidermic mounts (horizontal axis in Fig. 89B ; we could explain gizzard mass variation between Wiersma and Piersma 1994 and Chapter 2 ). We knot-staging sites worldwide. Digging through predicted that the largest gizzards would occur the literature and scrolling through old data-fi les in midwinter when demands are high and food on food quality, we were able to predict ‘over- quality is low, whereas the smallest gizzards were predicted for late summer, when demands are lowest and food quality is still rather high 0.6 A ( Fig. 89B ). In the Wadden Sea, knots fuel-up 0.5 during spring, and thus, in spite of the high- 0.4 quality food during that time of year, we pre- 0.3 dicted relatively large, rate-maximizing ‘spring 0.2 gizzards’ (Fig. 89B ). When we collated all the gizzard data that we had built up over almost 0.1 20 years of research (920 individual knots; measured initially by dissection of carcasses, diet quality (g flesh/g shell) and, in later years, by ultrasonography), we found a fi t with the predictions that was even 12 B better than we could ever have dreamed. 10 Throughout the year, gizzard size varied in response to demands and food quality, with 8 knots maintaining a rate-maximizing gizzard 6 in spring and a satisfi cing gizzard for the rest of 4 the year; it even seems that during early winter gizzard mass (g) their gizzards are accommodated to fuel their 2 moderate midwinter mass peak ( Fig. 89C ). 0 1998 1999 2000 2001 2002 The gizzard optimality model also worked year well when predicting variation in gizzard masses between years during the late-summer Figure 90. (A) In the Dutch Wadden Sea, diet quality declined arrival period ( van Gils et al . 2006a , 2006b , between 1998 and 2002, and, as expected, (B) red knots responded by increasing the size of their gizzard. Box-and- 2007a ). Again based on analysis of droppings, whisker-plots as in Fig. 80 ; lines representing regressions (straight we found some considerable variation in late- lines) and their 95%-confi dence intervals (curved lines). summer food quality across a period of fi ve Modifi ed from van Gilset al. ( 2006b ). THE DYNAMIC CHOICE BETWEEN DIETS, FEEDING PATCHES, AND GUT SIZES 143 wintering gizzard masses’ for fi ve out of the six net rate- existing knot subspecies, and for three of those 16 maximiser we could additionally fi nd food data from their spring stopover sites ( van Gils et al. 2005c ). The 12 latter yielded the surprise. Where we had satisficer expected to predict large fuelling gizzards dur- 8 ing spring, enabling a rapid build-up of energy stores, we found that the predicted rate-maxi- gizzard mass (g) 4 mizing gizzards at the spring stopover sites were of similar size (or even smaller in the case of the STOPOVER WINTERING nominate subspecies) than the satisfi cing giz- 0 zards predicted for the wintering sites (Fig. 91 ). rufa rogersi We were surprised by our own model, although canutus islandica piersmai the reason is obvious. Food qualities were twice Figure 92. Observed mean gizzard masses (± standard errors) as high at the spring stopover sites as at the win- at wintering grounds (fi lled symbols) and stopover sites (open tering sites (Fig. 91 ). Once again our predictions symbols) match well with predicted gizzard masses (grey bars ± were refl ected in the data: during winter, giz- standard errors, with bottom of each bar giving the local zards varied between sites as expected (from satisfi cer prediction and top of each bar giving the local rate- maximizing prediction). about 6 g in Roebuck Bay, northwest Australia to From van Gils et al . ( 2005c ). about 10 g in Mauritania, West Africa), while the ‘spring gizzards’ hardly showed any variation and averaged out relatively small at about 8 g This led us to infer that knots select their (Fig. 92 ). spring stopover sites on the basis of food quality. Selection of such ‘hotspots’ would speed up

migration. This is for two reasons. First, better )

canutus 2003

700 islandica ( quality food enables faster fuelling. This is 600 rufa m 16 500 rogersi ö because the rate-maximizing gizzard can then 14 400 piersmai be smaller and is thus energetically cheaper, 300 12 1 kJ/g shell Kvist/Lindstr leading to higher net energy intake rates (com- 10 200 pare the lengths of the double-headed open 2 8 arrows in Figs 88A and B). Second, with hardly 100 3 any size difference between the satisfi cing win- 4 6 predicted gizzard mass (g) 5 tering gizzard and the rate-maximizing spring

predicted daily shell mass intake (g) predicted daily shell mass intake 6 7 8 gizzard, knots lose none of the precious time 200 300 400 500 600 that would otherwise be given over to increas- daily energy requirement (kJ) ing gizzard mass (Piersma et al . 1999a , 1999b ). Figure 91. Using the same framework as explained in Fig. In their site selection, knots may even follow a 89B , we calculated satisfi cing gizzard masses for fi ve knot northward ‘wave’ in food quality, just as arctic- subspecies at their wintering grounds (fi lled symbols). Additionally, breeding geese are thought to follow a north- for three of them we could also make predictions on rate- ward ‘green wave’ in the quality of their food maximizing gizzard masses at their stopovers (open symbols). during spring migration ( Drent et al . 1978 ). This This shows that, due to an increase in food quality (crossing several food-quality isoclines), little change in gizzard size is is because mollusc prey is usually of best required when moving from wintering to stopover site (arrows). quality during spring, due to the production of Modifi ed from van Gils et al . ( 2005c ). nutritious gametes (Honkoop and van der 144 ADDING BEHAVIOUR

Meer 1997 ). Since spring occurs later in the year constant environments and bodies. Based on with increasing latitude, there may be a ‘fl esh our own work on shorebirds, we here take wave’ carrying knots to their breeding grounds these models a step further by including the (also see Piersma et al . 1994b ). Two observations dynamics of organ size, energy expenditure, on spring fuelling rates (published by Piersma and prey quality. Our observations on diet et al . 2005 ) are in accordance with this idea: (1) choice, patch use, and stopover-site selection in non-seasonal environments, the tropics, fuel- by red knots accord with new versions of ling rates are lower than in seasonal environ- foraging models that allow for fl exible ments; (2) in seasonal environments actually adjustments of food-processing capacities. experiencing the austral autumn, Southern Furthermore, not only behaviour is opti- Hemisphere stopovers, ‘spring’ fuelling rates mized with respect to processing capacity— are lower than at Northern Hemisphere processing capacity itself is optimized with stopovers. respect to physiological demand. Red knots fi ne-tune the size of their digestive organs, Synopsis notably their gizzard, such that energy income balances with energy expenditure, Optimal foraging theory has produced ele- while, at times of (migratory) fuelling, giz- gant models of diet choice and habitat use. zards are enlarged to such an extent that net Until now such models have assumed rather energy income is maximized. Part IV

Towards a fully integrated view This page intentionally left blank CHAPTER 8 Beyond the physical balance: disease and predation

Running with the Red Queen Arctic and alpine’ or ‘decreasingly marine/ saline’, the covariation shows up clearly ( Fig. Our prime witnesses in this book, red knots, 93 ); ruffs Philomachus pugnax and red knots breed in the High Arctic, about as far north as occupy the outlying positions. When the inter- one can get. Specialized molluscivores outside specifi c associations were subjected to formal the breeding season, red knots shift to eating statistical tests, the empirical linkages between surface arthropods during the summer months the degrees of northerly breeding and the use of on the tundra (see Chapter 1 ). Because they are marine or coastal non-breeding habitats were able to live off intertidal bivalves as well as confi rmed ( Piersma 1997 ); and these associa- surface arthropods, there is no obvious reason tions do not stop at shorebirds. Comparisons why red knots should necessarily restrict among gulls and terns, ducks and passerines all themselves either to tundra during the breed- suggest that only the far-northern breeding taxa ing season (they ought to be able to fi nd such uniquely use coastal or saline habitats (Piersma arthropod prey in temperate meadows as well; 2007 ). This biogeographic pattern is not linked Schekkerman et al . 2003 ) or to coastal intertidal to any sort of trophic specialization. habitats during the non-breeding season (they In The Red Queen: sex and the evolution of human should also be able to use freshwater wetlands nature , Matt Ridley ( 1994 ) explores sex, or rather with shellfi sh resources, such as the Niger the reasons why such a seemingly wasteful life- fl oodplains in Mali; Zwarts et al . 2009 ). history phenomenon is the rule rather than the Expanding this comparison, other High exception, and concludes that sex evolved to Arctic-breeding shorebird species also winter in fend off pathogens and parasites. By continu- marine coastal wetlands, including grey plover ous intergenerational re-arrangement of the Pluvialis squatarola , sanderling Calidris alba , pur- genome—which is what sex is all about (Lane ple sandpiper Calidris maritima , and bar-tailed 2009a )—especially the genes coding for pro- godwit Limosa lapponica . These species provide a teins that are part of the line of defence against striking contrast with their more southerly invaders of the body, organisms try to remain breeding congeners (respectively Eurasian one step ahead of their adversaries. This mech- golden plover Pluvialis apricaria , Temmick’s stint anism of an ‘evolutionary arms race’ ( Van Valen Calidris temminckii , dunlin Calidris alpina , and 1973 ), in this case between pathogens and their black-tailed godwit Limosa limosa) that addition- hosts, was named the ‘Red Queen hypothesis’, ally, or uniquely, rely on freshwater wetlands after Lewis Carroll’s character from Through the during the non-breeding season. Indeed, when looking-glass ( Carroll 1960 ). the breeding and wintering habitats of the 24 In the relevant part of the wonderland story, different species of the sandpiper subfamily Alice is running as fast as she possibly can Calidrinae are ranked as ‘decreasingly High

147 148 TOWARDS A FULLY INTEGRATED VIEW

can do, to keep in the same place. If you want 25 24 23 to get somewhere else, you must run at least 22 9 twice as fast as that!’ If temperate freshwater 20 11 15 habitats offer greater disease pressure than 10 14 High Arctic and marine habitats, shorebirds in 13 15 7 freshwater habitats have to run faster than shore- 21 8 19 birds in High Arctic and marine habitats 6 10 5 (Piersma 1997 , 2003 ). Indeed, ‘there are ticks 16 12 and lice, bugs and fl ies, moths, beetles, grass- 20 DECREASINGLY MARINE DECREASINGLY 4 hoppers, millipedes and more, in all of which rank order winteringrank habitat 5 2 1 males disappear as one moves from the tropics 18 17 1 3 to the poles’ ( Ridley 1994 ). This increase in 0 50 10 15 20 25 parthenogenetic (male-less) reproduction is rank order breeding habitat consistent with the idea that sex exists to beat DECREASINGLY ARCTIC pathogens, but also, and this is the point in the Figure 93. Signifi cant correlation between the extent to present discourse, with the idea that in the which shorebird species belonging to the subfamily Calidrinae High Arctic, pathogen pressure is low. (Scolopacidae) breed in decreasingly arctic/alpine, open and The patterns also carry another implication. climatically extreme habitats (rank order breeding habitat) and Although shorebirds breeding furthest north the extent to which they winter increasingly in freshwater rather than marine habitats (rank order wintering habitat). The and wintering in shoreline habitats cover the linear regression line is shown to lead the eye. 1: Aphriza greatest migration distances between suitable virgata ; 2: Calidris tenuirostris ; 3: Calidris canutus ; 4: Calidris patches of habitat, they, and not the freshwater alba; 5: Calidris pusilla; 6: Calidris mauri; 7: Calidris rufi collis ; 8: species, might be able to pull off such energeti- Calidris minuta; 9: Calidris temminckii; 10: Calidris subminuta ; cally demanding migrations because they do 11: Calidris minutilla; 12: Calidris fuscicollis; 13: Calidris bairdii ; 14: Calidris melanotos ; 15: Calidris acuminata ; 16: Calidris not need to invest as heavily in immune ferruginea; 17: Calidris maritima; 18: Calidris ptilocnemis ; 19: defence ( Piersma 1997 ). Alternatively, because Calidris alpina; 20: Eurynorhynchus pygmeus; 21: Limicola of the low pathogen pressures, ‘clean’ polar falcinellus ; 22: Micropalama himantopus ; 23: Tryngites and marine habitat would allow high levels of subrufi collis ; 24: Philomachus pugnax . energy expenditure without negative fi tness From Piersma ( 2003 ). consequences (see Chapter 4 ). It goes without saying that scientists have gone out to test for themselves whether the abundance of disease organisms is actually alongside the Red Queen, yet notices that their lower in some environments than others. There position does not change. Almost breathless, is increasing evidence indicating that blood par- Alice mentions to the Queen that in her country asites and/or their vectors are indeed relatively ‘you’d generally get to somewhere else—if you sparse in both High Arctic and marine/saline ran very fast for a long time, as we’ve been habitats compared with temperate and tropical doing’. ‘A slow sort of country!’ said the Queen. freshwater habitats (e.g. Bennett et al . 1992 , ‘Now, here, you see, it takes all the running you Figuerola 1999 , Jovani et al . 2001 , Mendes et al .

1 This citation was incorrectly attributed to Daly and Wilson ( 1983 ). In addition, and ironically, it was cited back to front (confi rmed by M. Ridley pers. comm.). BEYOND THE PHYSICAL BALANCE: DISEASE AND PREDATION 149

2005 ). We would also expect contrasts in overall immune challenges, these have the additional immune investments, or aspects of immunity, benefi t of being measurable from samples between High Arctic/marine species and lower taken on single capture occasions. latitude/freshwater species—contrasts that Three techniques are now commonly applied: would follow logically from the differential (1) the killing of standardized microbes by exposure to disease pressures during either the fresh or stored deep-frozen and then thawed egg- or chick- phases or later in life—but these blood or plasma ( Tieleman et al . 2005 , Millet have proven to be diffi cult to investigate and to et al. 2007 ), (2) concentrations of different leu- confi rm ( Mendeset al . 2006a , 2006b ). kocyte types (Campbell 1995 ), and (3) ‘complement’ and ‘natural antibody’ assays (Matson et al . 2005 ). Microbial killing of particular strains The responsive nature of of the Gram-negative Escherichia coli bacterium, ‘constitutive’ innate immunity the Gram-positive Staphylococcus aureus bacterium, and the yeast Candida albicans should Do freshwater specialists exposed to a greater indicate the overall capacity to limit microbial density and variety of disease organisms, espe- infections ( Millet et al . 2007 ). Leukocyte con- cially in the tropics, indeed ‘run faster’ than birds centrations provide information on circulating that limit their presence to marine, saline, or immune cells but have the problem that they High Arctic environments? We must be honest are also indicators of health and ongoing infec- now and let you know that we will not be giving tions, and are thus potentially induced you an answer here. Science has simply not yet ( Campbell 1995 ). However, differential leuko- conquered much of this territory; the main hur- cyte counts can be used in multivariate analy- dle being the measurement of ‘running speeds’ sis in terms of their relationship to functional in immunological terms. Assessments of habitat- measures of immunity, such as microbial kill- related differences in immunity between species ing (see Fig. 94 ). Heterophils and eosinophils are hampered by the immense complexity of the mediate innate immunity against novel patho- immune system ( Janeway et al. 2004 ). gens and are important phagocytes, whereas In a straightforward and perhaps somewhat monocytes link innate and acquired defence. simplistic categorization, the immune system Lymphocytes mediate pathogen-specifi c anti- can be divided into innate (non-specifi c) and body and cell-mediated responses of the acquired (specifi c) arms, and further divided acquired immune system (Campbell 1995 ). into constitutive (non-induced) and induced Complement cascade and natural antibodies branches ( Schmid-Hempel and Ebert 2003 ). link innate and acquired immunity, thus pro- Obviously, the acquired and induced branches viding the fi rst line of defence against spread- of the immune system should closely refl ect ing infections including viruses ( Ochsenbein the diseases in the environments in which the and Zinkernagel 2000 ). Are you still with us? organisms have found themselves. To be able A body of work on immunity in red knots to separate organism and environment heuris- carried out by Debbie Buehler and co-workers tically, it therefore makes sense that, for a fi rst has shown that these innate measures of immu- characterization of the immune system of free- nity are repeatable within individuals ( Buehler living animals, one should turn to the non-in- et al. 2008a ). However, even under constant duced (constitutive) and non-specifi c (innate) environmental conditions, they nevertheless parts. Representing circulating levels of varied with season (Buehler et al . 2008b ). The immune proteins, rather than responses to immune measures did not respond to 150 TOWARDS A FULLY INTEGRATED VIEW

immunity measures that are the most costly to 1.0 lymphocytes monocytes maintain. Overall, the story seems to be that 0.8 even these ‘least responsive measures of immu- 0.6 heterophils nity’ are actually quite responsive ( Buehler S. aureus killing 0.4 et al . 2009a ), especially to ambient disease pres- eosinophils sures (Buehler et al . 2008b ). Comparing three PC2 0.2 groups of red knots ( Fig. 94 ), two of which were hemolysis 0.0 C. albicans killing sampled immediately after capture in the hemagglutination Wadden Sea and one of which was maintained –0.2 E. coli killing in a hyper-clean aviary (running seawater to –0.4 A keep the fl oor and the little mud patch clean, –0.2 0.20.0 0.4 0.6 0.8 1.0 and weekly cage disinfection), showed captive

0.8 birds to have reduced capacities to kill the bac- free terium S. aureus and the yeast C. albicans , as Sep. 2005 free well as lower levels of heterophil and eosi- 0.4 July 2006 captive nophil leukocytes, when compared to their

PC2 wild counterparts. In a principal component 0.0 analysis (Fig. 94 ), the affected variables fell on a single axis that seemed to refl ect reduced –0.4 B phagocytosis and infl ammation-based immu- –1.0 –0.5 0.0 0.5 1.0 nity in the aviary birds. Not surprisingly, this PC1 correlated with densities of microbial colonies Figure 94. Structure of the measures of constitutive innate measured per gram of mud, being lower in the immunity in captive and free-living red knots. In panel (A) the axes regularly disinfected aviaries than in the fi eld. represent the fi rst two components of a principal component There is no doubt that the immune system analysis. Vectors are the loadings of each immune measure and the represents one of the most fl exible parts of the length of the vector indicates how much of the variation in an phenotype. There is also no doubt, and even a immune measure is explained by the two axes. Dashed vectors are little evidence, that the interactions between the measures best explained by PC3 (not shown). The angle between two vectors gives the degree of correlation between them. Adjacent pathogens and parasites in an environment and vectors are highly correlated with each other, orthogonal (90º) the various branches of immunity of potential vectors are uncorrelated, and vectors pointing in opposite directions hosts occurring there greatly affect the use of (180º) are negatively correlated. (B) The groupings of principal these environments. The literature on mammals component scores (captive red knots, free-living birds caught in offers several examples of the pre-emptive pres- September 2005 and free-living birds caught in July 2006). For each group the mean ± SE of PC1 is plotted against the mean ± ence of deadly pathogens, such as the pandemic SE of PC2. Captive and free-living birds are distinguished on PC1 of viral rinderpest that swept through East but not on PC2. Africa in the late-1800s and left vast areas unus- Modifi ed from Buehler et al . ( 2008b ). able by a variety of ungulates for decades (Price 1980 , Dobson 1995 ). In the same category falls sleeping sickness, caused by a blood parasite experimental differences in temperature (and and transmitted by tsetse fl ies, which makes thus to differences in the costs of living), except parts of the African savannah inhospitable for when the birds were brought under stress by humans and their livestock ( McNeill 1976 ). In limiting feeding hours ( Buehler et al . 2009b ). addition to the avoidance of habitats in order to Under energy stress, red knots cut down on the prevent a disease, habitat use is also affected BEYOND THE PHYSICAL BALANCE: DISEASE AND PREDATION 151

95B ). That in turn reduced their rate of accumu- 400 A lating body fat (Fig. 95C ). In spite of this, and a 300 few other exceptions ( Cornell 1974 ), we conclude that it has as yet been diffi cult to build in 200 the presence or absence of more subtly acting 100 disease organisms to explanatory models of n bites per dropping habitat use. However, this has been much easier 0 with the more obvious threats to life and limb B 1200 represented by the predators.

800 Body-building to defy death

bite rate (#/h) bite rate 400 For red knots and other shorebirds, peregrine falcons Falco peregrinus are lethal enemies ( Fig. 0 96 ). Even though peregrines can be held 0.03 C responsible for only a small fraction of the deaths of red knots (van den Hout et al . 2008 ),

0.02 the knots’ use of sites and their fl ocking behaviour betrays a deep-seated fear of the falcons ( Piersma et al . 1993c , Dekker 1998 , Dekker and 0.01 Ydenberg 2004 , van den Hout et al . 2010 ). Staying away from falcons is one way of cop- fuelling rate (API-units/day) fuelling rate 0.00 NO YES ing with the threat they pose (Ydenberg et al . infected 2004 , 2007 , Leyrer et al . 2009 ), but this is not Figure 95. (A) Bewick’s swans infected with low-pathogenic always possible, especially since they live in avian infl uenza took fewer bites per faecal dropping than their such open habitats in which cover is obstructive healthy counterparts, a sign that they experienced poor digestion. rather than protective (as opposed to most pas- (B) Because of this digestive bottleneck, ill swans took fewer bites serines; Lima 1993 , van den Hout et al . 2010 ). per hour, (C) leading to slower fuelling rates (expressed as the daily increase in a bird’s Abdominal Profi le Index, API). An interesting situation occurs when such Modifi ed from van Gilset al . ( 2007b ). migrating shorebirds have fuelled up to embark on long-distance fl ights. They are heavy and, at least in red knots, the growth of once it is infected, even if this is by a non-lethal the pectoral muscle does not keep pace with pathogen. GPS-tagged Bewick’s swans Cygnus increases in fat loads (Fig. 97A ). The predicted columbianus bewickii that were found to be loss of aerodynamic performance (Dietz et al . infected by a mild forms of avian infl uenza 2007 ) was demonstrated in a simple experi- (H6N2 and H6N8), stayed longer at their win- ment. Every week the aviaries at the Royal tering grounds and travelled back to their Netherlands Institute for Sea Research are breeding grounds at a slower pace than unin- cleaned, and we use this time to weigh the fected individuals (van Gils et al . 2007b ). They birds and to examine them for moult and foot had hampered digestion ( Fig. 95A ; consistent infections. Animal caretaker Maarten Brugge with the idea that there is a trade-off between trained the red knots to fl y from his hand back immune function and digestion; Adamo et al . to their cage. This involved several metres of 2010 ), which constrained their intake rate (Fig. fl ight, followed by a 90º turn and a descent 152 TOWARDS A FULLY INTEGRATED VIEW

1.7 A

1.6

1.5

1.4

1.3 log pectoral muscle mass (g)

1.2 160 g

Figure 96. Photographic illustration of fast pure-rotational 2.0 2.1 2.2 2.3 2.4 log body mass (g) banks by dunlin Calidris alpina chased by a peregrine falcon Falco peregrinus . 1.0 B The photo, from the World Wide Web, was taken by 0.8 an anonymous photographer. 0.6 into the aviary. Maarten kept a score of whether 0.4 or not the birds were able to make this sharp turn. At body masses higher than 160 g, an 0.2 increasing proportion of birds failed to make it fraction that failed to turn 0.0 (Fig. 97B ), landing somewhere in the corridor 105 125 145 165 185 205 225 and having to walk back and jump over the body mass (g) doorstep into the aviary. Red knots build their Figure 97. Pectoral muscle mass and the ability to make 90 pectoral muscles up to a certain point only, degree turns in red knots. (A) The allometric relationship between and need to accept, or otherwise behaviour- pectoral muscle mass and body mass where the grey line is the ally compensate for, the associated risks of theoretical prediction based on aerodynamic principles and the decreased fl ight performance during the nalfi black line the continuous piecewise allometric regression through days of fuelling ( Piersma et al . 2005 ). the data. The dotted vertical line marks 160 g, the body mass level at which red knots start to encounter manoeuvrability problems. Of course, birds could also assess whether or Panel (B) shows how the fraction of birds failing to make the 90 not predators are around, and then ‘decide’ degree turn increases steeply above weights over 160 g. whether or not to make bodily adjustments. In Modifi ed from Dietzet al . ( 2007 ). an indoor experiment, ruddy turnstones Arenaria interpres were exposed to daily but unpredictable disturbances by either a gliding sparrow- As fl ight performance is expressed as the hawk Accipiter nisus model or, alternatively, by a ratio between pectoral muscle mass and body gliding black-headed gull Larus ridibundus mass to the power 1.25 ( Dietz et al . 2007 ), an model ( van den Hout et al . 2006 ). During the alternative way for a bird to boost fl ight abil- days of raptor-disturbance, the birds grew larger ity is to reduce body mass rather than to boost pectoral muscles (open dots in Fig. 98 ), while pectoral muscle size. In fact, ‘fl ight perform- during the days of gull-disturbance their pecto- ance’ may be too rough a measure to grasp all ral muscles would shrink. ‘Playing’ with pecto- essential aspects of a bird’s escape abilities. ral muscles in this way changed the birds’ fl ight Turnstones feed in small fl ocks on sandy performance in response to shifts in danger. beaches or rocky shores, often in the close BEYOND THE PHYSICAL BALANCE: DISEASE AND PREDATION 153

1.51 co-workers predicted that knots would reduce Red Knot Ruddy Turnstone body mass rather than increase pectoral mus- 1.50 0.088 cle mass, since being lighter reduces inertia, thus allowing tighter turning angles ( van den 1.49 0.082 Hout et al . 2010 ). They repeated the experiment with the stuffed sparrowhawk model, 1.48 0.076 but now substituting red knots for turnstones. 1.47 0.071 Instead of increasing pectoral muscle mass, the log pectoral muscle mass knots indeed reduced their overall body mass, 0.066 1.46 maintaining a more or less constant pectoral 2.04 2.06 2.08 2.10 2.12 muscle mass (solid dots in Fig. 98 ). This has log body mass made us wonder if perhaps new species’ Figure 98. Phase space with lines for equal fl ight capacity names are in order: how about ‘stones’ and (PMM/BM1.25 ) (Dietz et al. 2007 ), for ruddy turnstones (van den ‘turnknots’ rather than turnstones and knots? Hout et al. 2006 ) and red knots (van den Hout et al . 2010 ), Although the ability to escape from an showing that both species increase fl ight capacity in response to approaching predator is an important skill, it raptor-model intrusions: red knots by decreasing body mass and ruddy turnstones by increasing pectoral muscle size (all data may trade off with the ability to process food were log-transformed). during times of food scarcity. A starving bird From van den Hout et al . 2010 . catabolizes its own protein stores, and thus needs to ‘decide’ between burning down different organs. Will it sacrifi ce its flight capaci- vicinity of obstructive cover, and raptor ties or its food-processing capacities, or both at attacks are thus usually by surprise. Little the same time? Analysing the composition of time remains for turnstones to escape, and 103 knot bodies provided the answer (Dietz they would thus benefi t most if they could and Piersma 2007 ). These birds had all been rapidly accelerate in order to reach a safe des- collected in The Netherlands in winter, and tination, such as water or salt marsh. had either died through starvation during Acceleration demands the build-up of fast- severe winter weather or by accident during twitch muscle fi bres, and thus turnstones normal winter conditions (by light-house colli- should enlarge their pectoral muscles, rather sions or during capture). Although all organs than reduce overall body mass, in order to considered (gizzard, liver, intestines, pectoral boost ‘fl ight performance’ and thus increase muscles, leg muscles) were smaller in the escape chances. Red knots, however, are typi- starved birds than in the non-starved birds, cal ‘farshore foragers’, feeding in huge fl ocks the reductions were much more serious in the on wide-open mudfl ats, often as far from the pectoral muscles (a decline by c . 60%) than in coastline as possible ( Piersma et al . 1993c ). For the gizzard (c . 21%), especially with respect to them, there is much more time left between the reduction in total body mass ( c. 33%; Fig. predator detection and escape: acceleration 99 ). In fact, the starved birds had gizzards of capacity may therefore not be their fi rst prior- about 8–8.6 g fresh mass, just enough to bal- ity. Red knots will team up, performing united, ance energy income with energy expenditure erratic display fl ights that confuse the on a daily basis (the satisfi cing policy; see approaching raptor; this requires high turning Chapter 7 ). So, the capacity to process food at manoeuvrability of all fl ock members. Using this stage appears more important than pre- aerodynamic theory, Piet van den Hout and venting capture and being eaten. After all, 154 TOWARDS A FULLY INTEGRATED VIEW impaired escape abilities need not lead to on resting shorebirds. The alternative is Richel: a death, as long as there are no successful preda- completely fl at sandbank, enabling a wide-open tory attacks, but turning off digestive capacity unobstructed view of the potentially dangerous will certainly lead to death. sky, but usually rather far from good feeding Such a food-safety trade-off and how it affects grounds. Where to rest during high tide: at the the phenotype is also evident in roost selection risky but energetically cheap Griend, or at the by knots. Safety at the roost is important, but safer but energetically more expensive Richel? safe roosts may not always be located near good The optimal solution may differ between differ- feeding areas. Therefore, birds may face a trade- ent phenotypes, as revealed by a between-years off between maximizing roost safety and mini- comparison of roost-use by 26 radio-tagged mizing daily travel distance from and to the knots (van Gils et al . unpubl.data). In years when roost. This trade-off is evident in red knots win- almost all roosting took place at Richel, the knots tering in the western Dutch Wadden Sea, where had small pectoral muscles relative to their body there are basically two major roosts (Piersma mass; while in years of relatively greater roost- et al . 1993c ). There is Griend, a small island in ing at Griend, knots carried relatively large pec- the midst of good feeding grounds, but rather toral muscles ( Fig. 100 ). These inter-annual dangerous because of a recently elevated dike differences in roost-use, and hence in pheno- that enables raptors to perform surprise attacks typic appearance, may refl ect variation between years in food distribution or raptor abundance. For example, in a typical ‘Richel-roosting year’, such as 1997, there may have been good food gizzard 4 patches near Richel or there may have been relatively many raptors around (the latter increasing predation danger more steeply at the 3 endiked Griend than at the wide-open Richel).

2 0.065 ) fat-free dry organ mass (g) 1 1.25

0.060 1999 10 pectoral muscles

2000 1998 8 0.055

6 1997 flight capacity (PMM/BM 4 0.050 30 40 50 60 70 80 90 100 2 relative use of Richel as a roost (% of tides) fat-free dry organ mass (g)

0 Figure 100. Phenotypic adjustments of fl ight capacity to 60 80 100 120 140 160 roost safety. The more often knots roosted at the relatively safe body mass minus fat-free dry organ mass (g) Richel (relative to the number of times they roosted at Griend), Figure 99. Starved knots (closed dots) have smaller organs the smaller their fl ight capacity, i.e. pectoral muscle mass/body than non-starved knots (open dots), but the reductions are mass 1.25 . much stronger in the pectoral muscles than in the gizzard. Based on J.A. van Gils, A. Dekinga, and T. Piersma et al . Modifi ed from Dietz and Piersma ( 2007 ). (unpubl. data.) BEYOND THE PHYSICAL BALANCE: DISEASE AND PREDATION 155

Coping with danger food patches at the same instantaneous intake rate in order to maximize its long-term aver- Besides the bodily adjustments just described, age intake rate. However, by including fear in animals can respond behaviourally to an this model, an animal trading-off risk of star- increased predation danger. We have already vation with risk of predation should give up mentioned ‘teaming-up’ or fl ocking as an dangerous patches at higher instantaneous effective way to dilute the risk of being caught intake rates than safe patches. This prediction among fl ock mates ( Bednekoff and Lima 1998 ). has been tested and verifi ed many times since Selecting a safer habitat and actively looking Brown’s (1988) seminal paper (reviewed by out for predators (vigilance) are two other Brown and Kotler 2004 ). This fundamental behavioural ways to reduce predation risk principle provides a strong link between the (note that predation risk can be managed by behaviour of fearful prey and the organisation an organism, whereas predation danger is a of the species communities of which they are feature of the environment which is uncontrol- part. lable and refl ects predationrisk if no anti- For example, reintroducing wolves into predation measures were taken; Lank and Yellowstone National Park brought back the Ydenberg 2003 ). extensive aspen and willow forests, a typical In most of these cases, the animal pays a North-American landscape that had largely price for its safer life. Joel Brown termed this disappeared over the last 150 years (Ripple ‘the cost of predation’ ( Brown 1988 ), and he and Beschta 2005 ). The strong linkage between argued that this price can be expressed in wolves and plants came about not so much terms of a loss of energy intake. Feeding in because of wolves killing the major consumers dense fl ocks usually leads to greater interfer- of aspen and willow, elks, but merely due to ence and reduced food-intake rates ( Vahl et al . wolves scaring elks out of potentially danger- 2005 ), looking out for potential threats detracts ous habitats. The slippery slopes of streams, from fi nding food ( Radfordet al . 2009 ), and where elks browse aspen branches, are espe- safer habitats may be nutritionally poorer hab- cially risky. The mere avoidance of this ripar- itats (Cresswell 1994 ). In the long run, these ian zone by elks after the wolves showed up lower intake rates have knock-on effects on an triggered a cascade of ecological events. It animal’s life history. For example, female elk ‘released’ aspen and willow growth along Cervus elaphus give birth to fewer young when streams, which brought back beavers; their wolves Canis lupus are abundant (Creel et al . dams then slowed down water fl ow and 2007 ). When experimentally exposed to streams started to meander more, thereby gen- enhanced predation danger, snowshoe hares erating a mosaic of productive biodiverse Lepus americanus not only gave birth to fewer microhabitats for birds and butterfl ies ( Ripple young, those young were also smaller and and Beschta 2004 ). This is just one example out lighter ( Sheriff et al . 2009 ). of many showing the non-lethal fear effects of The discovery of a common—energetic— predators on their prey ( Werner and Peacor currency to express ‘predation risk’ gave rise 2003 , Schmitz et al . 2004 ). Fear is a driving eco- to a new and fl ourishing fi eld within animal logical force! ecology, the ‘ecology of fear’ ( Brown et al . Note that for the same risk of predation, not 1999 ). Patch-use theory played a central role, every individual experiences the same cost of with the classical marginal value theorem pre- predation. Thus, some individuals showing dicting that a forager should give up all its lower fear-levels will exploit dangerous 156 TOWARDS A FULLY INTEGRATED VIEW

patches to a greater extent than others. This 0 may seem odd at fi rst, as dying in the claws –5 or jaws of a predator is the end of everyone’s –10 M. balthica 15mm fi tness. However, we should not forget that –15 the cost of predation actually refers to the loss –20 in energetic foraging opportunities due to preventing being killed. Foregone energetic –25 income is most costly to those individuals –30 that are most in need of energy, such as lean –35 S. plana 35mm

individuals with small energy stores. For deviation from mean body mass (%) –40 them, the fi tness value of one unit of energy is 0 12345678 relatively large. Hence, one unit of predation depth layer (cm) risk is relatively small on their energetic scale. Figure 101. Relative body mass of two marine bivalve species A good example of lean individuals accepting in the upper layer of the sediment (the upper 1 cm, the upper greater risks of predation than heavier indi- 2 cm, etc.). viduals can be found in marine benthic inver- Modifi ed from Zwarts and Wanink ( 1991 ). tebrates, notably bivalves. Being burrowed in the sediment, they protect themselves against predation by shorebirds and fi sh. The deeper being easier to catch (think of the heavy red the safer; but this trades off with food intake, knots failing to turn into their cage; Fig. 97B ). as the food needs to be collected from the sea- However, because for them the fi tness value of fl oor surface or the water column using an food is lower than for light individuals, they extendable ‘trunk’, the inhalent siphon. In show less risky behaviour by devoting more this dilemma, shallowly burrowed individu- time to vigilance and feeding in poorer but als indeed have signifi cantly lower body safer patches. Because of this hiding behaviour masses than deeply burrowed bivalves (Fig. of the most well-nourished prey individuals, 101 ; Zwarts and Wanink 1991 ). Lean individ- hungry predators usually face the most mar- uals appear more willing to give up some of ginal prey in their quest for food. their safety in exchange for a higher energy income. The same patterns have been found in migratory songbirds stopping over at a small 50 offshore island in the German part of the North 40 Sea. Dierschke ( 2003 ) analysed the body weights of birds caught by local cats or raptors 30 and compared them to the body weights of 20 birds trapped for ringing purposes. He found that the lightest individuals had the highest % birds depredated 10 chance of falling victim to predation (Fig. 102 ). 0 Lank and Ydenberg ( 2003) correctly pointed 0–20 20–40 40–60 60–80 80–100 out that this result does not imply that lighter body mass percentiles individuals are easier to catch (as Dierschke Figure 102. Distribution of body masses of depredated birds. erronously suggested). Presumably it is the About half of the killed birds belonged to the 20% lightest birds. other way around, with the heavier individuals Modifi ed from Dierschke ( 2003 ). BEYOND THE PHYSICAL BALANCE: DISEASE AND PREDATION 157

Although not necessarily lean, migratory 3.5 A birds fuelling for migration are also in great need of energy. This is especially so when there 3.0 is little time left to fuel, such as in long-distance 2.5 migrants on their fi nal stopover on the way to the High Arctic; for example, in bar-tailed god- 2.0 wits Limosa lapponica . Wintering in West Africa 1.5

abdominal profile score lapponica and stopping over in May in the Dutch Wadden taymyrensis Sea, the taymyrensis -subspecies faces a tight 1.0 travel schedule to arrive on time at its Siberian 1 815 22 29 613 20 27 April May breeding grounds in June (Drent and Piersma 1990 , Piersma and Jukema 1990 ). This greatly 80 B ) contrasts with the lapponica -subspecies, which 2 has the whole winter to fuel up for migration 60 as it spends the winter resident in the Wadden Sea. Moreover, this subspecies breeds rela- 40 tively close by in northern Scandinavia and thus requires less fuel for migration. As a 20 result, in the Dutch Wadden Sea, taymyrensis prey availability (prey/m 0 fuels up to 10 times faster than lapponica ( Fig. meadow 0 500 1000 1500 103A ; Duijns et al . 2009 ) and the fi tness-value distance from cover (m) of one unit of energy will be much greater in Figure 103. (A) Based on changes in the abdominal profi le, taymyrensis . This should reduce the cost of pre- the taymyrensis subspecies of the bar-tailed godwit fuels up dation, which in turn would make them accept faster and to a greater extent than the lapponica subspecies. greater predation risks. These subspecifi c dif- (B) Furthermore, they feed in higher prey densities in meadows ferences are indeed refl ected in their feeding and closer to the coastline than the lapponica subspecies. The light grey area on the left gives the 95% confi dence interval distributions on the mudfl at, where they face a around the average distance to the coast for taymyrensis ; the steep food-safety trade-off: taymyrensis is con- dark grey area on the right does so for lapponica; and the sistently found feeding near shore, where both darkest grey area in the middle gives the overlap. food abundance but also predation danger are Modifi ed from Duijns et al. ( 2009 ). relatively high; lapponica prefers to feed in the safer outer-coast mudfl ats at the expense of fi nding much lower food densities ( Fig. 103B ; prospects in need of protection; Clark 1994 ). Duijns et al . 2009 ). An individual’s fi tness prospects (equivalent Independent of the short-term value of to its reproductive value) usually increases energy, more long-term fi tness prospects, with age, as older individuals have higher such as the annual survival probability, may survival chances and have greater reproduc- also differ between individuals. According to tive capacities. Indeed, in red knots (van den the modern, fear-based foraging theory, bet- Hout et al . in prep), and also in redshanks ter fi tness prospects should increase the cost Tringa totanus ( Cresswell 1994 ), juveniles are of predation, thus increasing the fear response usually found in the riskier but more produc- for a given risk of predation (this principle tive habitats whereas the adults feed in the has been coined the ‘asset-protection princi- safer but effectively poorer habitats. In par- ple’, with the asset being an animal’s fi tness rotfi sh Scarus spp. and Sparisoma spp., 158 TOWARDS A FULLY INTEGRATED VIEW

reproductive value also increases with body Predicting carrying capacity size and, indeed, the largest individuals in the light of fear showed the greatest fear response towards an approaching scuba diver (Gotanda et al . 2009 ). We have seen that animals give up dangerous Getting back to burrowing invertebrate crea- patches at higher food densities than safe tures, the degree of risk acceptance in the patches. This notion should be included in form of burrowing depth refl ected the per- models that try to predict how many days ani- ceived fi tness prospects by knots’ favourite mals can live on a given supply of food ( van prey, the tellinid bivalves Macoma balthica . Gils et al. 2004 ). This number of days is usually Even though they are tiny little animals with called an area’s ‘carrying capacity’ ( Sutherland nerve nodes rather than a brain, the accuracy and Anderson 1993 ). We can think of an area of their perceptions was amazing. Years in as having a carrying capacity of, let’s say, 1000 which Macoma burrowed shallowly (accept- bird-days, which means that there is enough ing greater risks) were followed by popula- food for a single bird to live for 1000 days, two tion declines, whereas the years of deep birds for 500 days, or 1000 birds for a single burrowing were followed by population day. These calculations always take into increases ( Fig. 104 ; van Gils et al . 2009 ). account the time that animals need to harvest Obviously, Macoma , but this holds for all ani- their food, and thus there is a critical food- mals, cannot directly assess its future survival density below which it will take more time to chances, but must rely on some sort of cue, collect the daily required amount of food than which correlates with these prospects. Food there are hours in a day. abundance, food quality, or body condition Leaving predation aside, one would expect may often be reliable cues. that, eventually, all patches would be harvested down to a level where gross food-intake rate would just counterbalance an animal’s proportion accessible (%) minimal energy requirements. In those rare 50 75 100 cases where all patches are equally costly to 0.2 exploit in terms of energy expenditure, this would lead to similar ‘giving-up densities’ of 1997 0.1 2000 food across all patch types (Fig. 105A ). This 2006 1998 was confi rmed recently in an experimental 0.0 study on bank voles Myodes glareolus that were 2004 2005 offered identical patches and which experi- –0.1 2002 1999 2003 enced virtually no predation threat ( Eccard population growth rate population growth –0.2 2001 and Liesenjohann 2008 ). However, the metabolic exploitation costs can frequently differ 0.8 1.0 1.2 1.4 1.6 between patches. Under such circumstances, proportion accessible (arcsine-transformed) costly patches should be given up at higher Figure 104. The asset-protection principle at work: the population threshold food-densities than cheaper patches of Macoma balthica increased after individuals burrowed themselves ( Fig. 105B ). This was verifi ed in Bewick’s relatively deeply, protecting their asset against predation. After swans feeding on pondweed tubers, which left individuals burrowed themselves relatively shallowly, the population would decline. This suggests that Macoma is able to assess its own behind more food in shallow water, where clay fi tness perspectives reasonably well. bottoms made the tuber-harvest more costly From van Gils et al . ( 2009 ). than in sandy patches ( Nolet et al . 2001 ). BEYOND THE PHYSICAL BALANCE: DISEASE AND PREDATION 159

A BC gressively more time in the dangerous patches. t=0 1 This stops once the intake rate in the danger- 2 3 ous patches has also dropped below the mini- 4 5 6 mally required level. From then on animals 7 8 either starve or leave the area: carrying capac- 9 10 ity has been reached. Foragers thus effectively 11 12 leave an imprint of their so-called ‘landscape 13 14

15 minimal energy requirement of fear’ (Laundré et al . 2001 ), a heterogeneous gross intake rate intake gross 16 y=x map of variable giving-up food densities that refl ects the spatial heterogeneity in predation costs. For example, small mammals often αβ αβ αβ retreat to their burrows upon predator detection, and thus unharvested food densities are metabolic rate 0 predation risk usually higher further out from the burrows, Figure 105. An area’s carrying capacity predicted by the original yielding physical maps with burrows sur- (A) and modifi ed versions (B–C) of Sutherland and Anderson’s rounded by contours of low predation costs (1993) depletion model. In these examples, the area consists of ( van der Merwe and Brown 2008 ). a β two patches, and , which initially contain similar prey densities. In human-dominated areas, roads often The parameters of the functional response are kept similar across shape the landscape of fear. This is true, for the fi gures, which enable direct comparisons between predictions of carrying capacity. Initially, at the onset of exploitation, intake example, of geese and swans that prefer agri- rates are high. Due to depletion, intake rates gradually decline cultural fi elds furthest removed from human (downward arrows), with the intake rates at time t given by grey access. Only when the food in such areas is lines. Minimal energy requirements are given by dashed horizontal depleted will they accept greater risks and line intersecting the three subfi gures. feed closer to roads ( Gill et al . 1996 ; van Gils Modifi ed from van Gilset al . ( 2004 ). and Tijsen 2007 ). Even predators themselves, such as brown bears Ursus arctos , live in such Consequently, the clay habitat had a lower car- road-based landscapes of fear. Moose Alces rying capacity than the sandy habitat ( Nolet alces sometimes take advantage of this by giv- and Klaassen 2009 ). ing birth on or near roads in the ‘safe’ shelter As predation risk can also be expressed of cars scaring off the bears hungry for baby ‘energetically’ in the form of a predation cost, moose ( Berger 2007 ). Having ended here with predation risk can be included in these ener- human-shaped landscapes of fear, we will con- gy-budget-based carrying-capacity calcula- tinue in the next chapter with human-wrought tions. Initially, when there is still plenty of fearsome landscapes: intertidal mudfl ats emp- food throughout the environment, dangerous tied and destroyed by large-scale mechanical patches will be avoided by foragers, and only fi shery activities. safe patches will be exploited ( Fig. 105C ). This will change once the difference in intake rate Synopsis between both types of patches equals the difference in predation cost: from that moment The responses of animals to diseases and pre- onwards, animals will alternate between both dation are diverse and fl exible. The immune patch types. Once the intake rate in the safer system is one of the most fl exible parts of the patches has dropped below the minimally phenotype, which can be turned up or down required level, the foragers will spend pro- depending on potential pathogenic threat 160 TOWARDS A FULLY INTEGRATED VIEW

levels. Insightful knowledge on the fl exible threat of predation. By fl exibly modifying immune system is still thin on the ground, as body composition, prey can reduce their risk immunity is so amazingly complex. What we of predation. Birds ‘play’ with pectoral muscle do know, however, is that shorebird species mass and total body mass, depending on the that breed in the relatively clean High Arctic, need to escape through acceleration or through winter in relatively clean saline habitats; manoeuvrability. Obviously, the most immedi- whereas species that breed at lower latitudes ate response to predation threat is through with higher pathogen pressures, winter in behaviour: visually scanning the environment, pathogen-rich freshwater environments. It is coming together in fl ocks, or switching habi- possible that High-Arctic breeders are the only tats. In such ways parasites and predators ones able to make super long-distance migra- strongly affect an area’s carrying capacity for tions because they do not need to invest so particular animals, something that is gradu- much in immune defence. We understand ally being realized and incorporated in applied much more about phenotypic responses to the contexts and conservation thinking. CHAPTER 9 Population consequences: conservation and management of fl exible phenotypes

We are deep into the book now, and you have Dredging out bivalve-rich intertidal read much about phenotypes in interaction fl ats: a case study on red knots with their environments. We have come to understand diet choice and the distributions Tidal fl ats occur along shallow seas with soft of animals because we took phenotypic change sediment bottoms, in areas where the tidal into account. That is all very interesting in its range is at least a metre or so ( Eisma 1998 , own right, but is it enough? Lenihan and Micheli 2001 , van de Kam et al. 2004 ). The lowest parts of intertidal areas are usually barren except for the occurrence of The Holy Grail of population biology seagrass meadows and reefs formed by invertebrate ‘ecosystem engineers’, such as oysters The population-oriented ecologists amongst Ostrea , mussels Mytilus , or tubeworms us are usually after even ‘holier grails’: the Sabellaria . Intertidal areas are inundated at understanding of changing numbers of organ- least once a day. On the landside, fl ats may be isms, and of variable population sizes bordered by salt marshes, or, in the warmer ( Kingsland 1995 , Hilborn and Mangel 1997 , regions of the world, by mangrove forests. Vandermeer and Goldberg 2003 , Ranta et al. The international Wadden Sea, a shared ter- 2006 ). Building on what we have learned about ritory of The Netherlands, Germany, and specialized molluscivore shorebirds in the pre- Denmark, is reckoned to be one of the biggest vious two chapters, we now want to take the intertidal areas in the world (van de Kam et al . next step by disentangling the reasons behind 2004 , Lotze et al . 2005 ), a distinction shared their changing numbers. Can we understand with the Yellow Sea (see below). Mostly 7 to 10 population dynamics better if we include the km wide, it stretches along 450 km of coast- notion of phenotypic fl exibility? In addition, line. During spring low tides a total of 490,000 given that the world’s ecology is in constant ha of intertidal mud is exposed. Compared fl ux, we would like to establish what differ- with other intertidal areas worldwide, the ences fl exible rather than infl exible phenotypes Wadden Sea harbours particularly large shell- could make to what animals can do or can tol- fi sh populations (Piersma et al . 1993a ). In 2009 erate on this rapidly changing planet. This it received recognition as a UNESCO World may allow us to predict population conse- Heritage Site. The Netherlands’ 120,000 ha of quences somewhat outside the current range the Wadden Sea are designated a State Nature of environmental conditions. To set the scene, Monument, and is protected under various we will start with a case study on a bird that other national and international conventions. will not surprise you at all—the red knot. Despite this high number of formal conservation

161 162 TOWARDS A FULLY INTEGRATED VIEW accolades, until 2004, three-quarters of the and sandgapers Mya arenaria , polychaetes, and intertidal fl ats of the Dutch Wadden Sea were crustaceans such as shorecrabs Carcinus mae- open to mechanical dredging for edible cock- nas. More indirectly, and over longer time les Cerastoderma edule ( Swart and van Andel scales, sediments lost fi ne silts during dredg- 2008 , Piersma 2009 ). Most parts are still open ing events, and changing sediment character- to other forms of bottom disturbance with fi sh- istics may have led to long-term reductions in ing gear of various severity (Watling and Norse settlement rates by both cockles and Baltic tell- 1998 ). ins ( Piersma et al . 2001 , Hiddink 2003 ). Suction-dredging of cockles took place from big barges with a draft of only 80 cm (Fig. Population consequences for the 106E ). The dredges, suspended from either molluscivores side of the barge, are about a metre wide and have water jets to loosen up the top layer of Several decades of serious fi shing pressure not the sediment. This is then scraped off by a only led to a stark loss of the shellfi sh resources, blade at a depth of approximately 5 cm. Within but also resulted in a serious decline of many the dredge, a strong water fl ow ensures that molluscivore birds in the Wadden Sea. The small objects are pushed through a screen. oystercatchers, most of which are resident in The remaining items, such as large cockles, The Netherlands, showed a steep decline, are sucked on board for further sorting and incurring the greatest losses when high food cleaning. After dredging, the intertidal fl ats demands coincided with restricted access to look ‘scarred’ and can remain so for a whole partially ice-covered mudfl ats during severe year ( Fig. 106F ). In the case of mechanical winters (Camphuysen et al . 1996 , van Roomen cockle fi shing in the Wadden Sea, GPS loggers et al. 2005 ). Eider ducks Somateria mollissima , were installed to record the position and consisting of local breeders and a much larger fi shing-activity of each barge (Fig. 106A –D). contingent migrating in from the Baltic, also On the basis of these records, an average 4.3% showed food-scarcity related mass starvation of the intertidal fl ats was affected by dredging and movements out of the Wadden Sea each year (Kamermans and Smaal 2002 ). ( Camphuysen et al . 2002 ). Oystercatchers and Cumulatively, between 1992 and 2001, 19% of eiders are large birds. Their actual death is eas- the intertidal area was affected by mechanical ily noticed. This is in marked contrast to the cockle-dredging at least once (Kraan et smallest dedicated molluscivore, the red knot, al. 2007). If this still seems a small percentage which shows rather cryptic mortality in the to you, note that cockle-dredgers selected form of declining head counts. Following ear- areas that red knots would select too, and that lier losses (Piersma et al . 1993c ), red knots by far the greatest loss of feeding area ( Fig. declined by 42% between winters in the late- 107C ) was accounted for by cockle-dredging 1990s and the early 2000s (Fig. 107A ). A decrease (van Gils et al. 2006b , Kraan et al . 2009 ). in annual local survival from 89% to 82% per A direct, immediate effect of dredging was year could account for almost half of this the complete removal of all organisms larger decline ( Fig. 107B ). The remaining red knots than 19 mm in the top layer. As the dredged moved elsewhere in Western Europe to winter sites were usually the most biodiverse ( Kraan (Kraan et al. 2009 ). Red knots en route to winter- et al . 2007 ), dredging also affected smaller ing areas in West Africa must store fuel for cockles, other bivalves such as blue mussels fl ight and thus have greater energy demands. Mytilus edulis, Baltic tellins Macoma balthica , Their numbers in late summer showed an even Netherlands

Germany

Belgium

1998

N D N LA E LI V E

FRIESLAN

55km km A

2000

B F

2001 1998

N D

LIELAN VLIELANV

FRIESLAN

C 55km km G

2002 2005

D H

Figure 106. Summary of the benthic sampling efforts in the western Dutch Wadden Sea and the extent of mechanical cockle- dredging in late-1998 (A), 2000 (B), 2001 (C), and 2002 (D). Dredged sampling stations are indicated by a fi lled dot, sampling stations not affected by mechanical dredging by open circles (1999 not shown as no dredging took place). (E) Cockle-dredging barges in action and (F) an aerial view of the surface scars of cockle dredging six months after the event (for scale, note the two human fi gures). Distribution of sampling locations where red knots could have achieved the required minimal intake rate of 0.3mg AFDM/s in 1998 (G) and 2005 (H), based on the measured harvestable food densities and a diet-selection model. Compiled from Kraan et al . ( 2007 , 2009 ); photographs by M. de Jonge (E) and J. de Vlas (F). 164 TOWARDS A FULLY INTEGRATED VIEW

80 the empirical relationships between prey densities, sizes and quality, and potential intake 60 rates (van Gils et al . 2006a , 2007a ), we can pre-

40 dict intake rates for every place on the mud- (x1000) fl ats at which data on food abundance and 20 quality were collected. On the basis of our number of Red Knots number A 0 knowledge of energy requirements as a func- 100 tion of weather conditions and storage strategies, we can additionally determine the 90 minimal intake rate necessary for the particular category of birds that we are interested in B 80 ( Piersma et al . 1995b ). On this basis, the disap- annual survival (%) annual

50 pearance of red knots ( Fig. 107A ) could be 8000 explained by the decline in the extent of suita- 40 6000 ble intertidal fl at ( Fig. 106G and H). Because 30 the rates of decline in red knot numbers ( Fig. 4000 20 107A ) and suitable foraging area ( Fig. 107C )

2000 10 were similar, average densities per unit of suit- C able mudfl at remained unchanged ( Fig. 107D ). suitable foraging area (%) foraging suitable suitable foraging area (ha) foraging suitable 0 0 As we have seen in Chapters 6 and 7 , in mol- 25 luscivore birds that ingest their prey whole, 20 energy intake rate is limited by the processing 15 capacity of shell material in the gut ( van Gils 10 et al . 2003a , Quaintenne et al . 2010 ). With all their potential for phenotypic change, in prin- 5 suitable foraging area foraging suitable number of knots per ha number D ciple red knots could build larger gizzards, 0 '96/'97–'01/'02 '02/'03–'05/'06 extend the range of suitable prey (by including items with less shell per unit shell mass) and Figure 107. (A) Decline of red knot numbers (box plots) and (B) their annual local survival (means ± SE) between thereby almost double the area of suitable winters in the intervals 1996⁄1997–2001⁄2002 and 2002⁄2003– mudfl at ( Fig. 108A ). In fact, they did grow 2005⁄2006. (C) Shows the decrease of suitable foraging area larger gizzards ( Fig. 108B ), but not by much. for red knots in the western Dutch Wadden Sea between the The measured increase of 0.4 g would have led two periods (left axis: suitable foraging area expressed in ha; to an expansion of suitable foraging grounds right axis: suitable foraging area in % of total area). (D) Because by only 225 ha, or less than 10% ( Fig. 108A ). the ‘slopes’ of the decline in numbers (A) and suitable area (C) are similar, the average number of red knots per ha of suitable Why did red knots not increase gizzard size foraging area remained constant. Box plots give the mean further? Enlargements of gizzards involve (large dot), median (horizontal line inside the box), interquartile increases in a number of cost factors that we range (box), range (bars) and outliers (asterisk). did not account for. These cost factors include Adapted from Kraan et al . ( 2009 ). the energy costs of growth and maintenance of a larger gizzard (van Gils et al . 2003a ) and steeper decline over the same period than that effects of added mass on manoeuvrability dur- of the over-wintering birds ( Kraan et al . 2010 ). ing escape from aerial predators (Chapter 8 , These patterns could all be explained by Dietz et al. 2007 , van den Hout et al . 2010 ). changes in the food-base of red knots. Using Minimizing the overall rate of energy expendi- CONSERVATION AND MANAGEMENT OF FLEXIBLE PHENOTYPES 165

'96/'97–'01/'02 50 being built up to that size threshold, the birds 8000 '02/'03–'05/'06 have to rely on stores that they already carry 40 ( van Gils et al . 2006b ). In the absence of an 6000 energy store (‘money in the bank’), gizzards 30 cannot be rebuilt as ‘the income is too small to 4000 20 pay the carpenter’. Birds will starve locally or, more likely, move away swiftly to try and fi nd 2000 suitable foraging area (%) suitable foraging area (ha) 10 places with higher prey quality.

0 0 Which individual red knots made '02–'05 n = 35 it through? In the Dutch Wadden Sea, from the late-1990s '96–'02 n = 90 onward, the area of mudfl at suitable for foraging red knots declined, not only because the 0 5101520 densities of suitable prey declined, but also gizzard mass (g) because the quality, the fl esh-to-shell ratios, of Figure 108. (A) Suitable foraging area in the western Dutch these prey declined (van Gils et al . 2006b ). Wadden Sea as a function of gizzard mass during the period During these years, in a series of late summers 1996⁄1997–2001⁄2002 (upper line) and during 2002⁄2003– and autumns, we studied red knots returning 2005⁄2006 (lower line), predicted for a minimally required from the High Arctic breeding grounds in intake rate of 0.3 mg AFDM/s (left axis: suitable foraging area expressed in ha; right axis: suitable foraging area in % of total northern Greenland and north-eastern Canada, area). (B) Distribution of gizzard masses (g) in the two periods a time of year when their arthopod foods do (in box plots; see previous fi gure). Even though the increase in not require particularly large gizzards ( Piersma gizzard size between the two periods was small (0.4 g), it was et al . 1999b , Battley and Piersma 2005 ). signifi cant (GLM using 125 noninvasively measured gizzards Returning to the Wadden Sea after breeding, between September and April, and year nested within period; red knots would encounter different prey F 3,121 = 5.76; P = 0.001). Indicating the range, the grey bar equals the mean ± SD gizzard mass from 1996 to 2005. There qualities in different years, with a general are vertical projections of mean gizzard mass per period on to downward trend. Variations between years in suitable mudfl at area. gizzard mass of birds captured during these Adapted from Kraan et al . ( 2009 ). times refl ected prey quality not just qualita- tively, but also quantitatively ( Fig. 109 ). ture by carrying the smallest possible gizzard These non-invasively measured gizzard for the energy budget to be in balance, at the sizes (Dietz et al . 1999 , Starck et al . 2001 ) of expense of the area where they could profi tably individually colour-marked birds were the forage, the birds may have found some kind of basis of an important discovery. Red knots that optimum. However, red knots may also have were subsequently resighted in the study area failed to increase gizzard size further simply had signifi cantly larger gizzards at first cap- because they were unable to do so. At any par- ture than birds that were not seen again ( Fig. ticular prey quality on offer, gizzard sizes 109 ). Birds that disappeared from view may below the required threshold do not allow the either have died or moved away permanently. digestive processing of enough shell material Of course, red knots with small gizzards could for enough energy to be extracted to cover the have increased gizzard size after capture; in daily costs. This means that whilst gizzards are fact, local survival was best quantitatively 166 TOWARDS A FULLY INTEGRATED VIEW

2005 , Schekkerman et al . 2003 ), organ fl exibil- 8 ity is clearly important. Our latest example 7 also demonstrated how events during one time of year (in this case the building up and 6 use of stores for the migration from arctic 5 breeding grounds to temperate wintering seen again areas) affect the survival chances upon arrival 4 not seen again

(required) gizzard mass (g) from that migration (in this case affected either 0.10 0.12 0.14 0.16 0.18 by having the stores to build up a gizzard in prey quality (g flesh/g shell) years with low prey quality, or not). What we have implied, but have not quite shown Figure 109. The gizzard mass required to maintain energy balance is predicted to decline with function of prey quality directly, is that the particular capacity for organ (solid line). Gizzards of individually colour-marked birds seen change enhances performance beyond what again within a year after measurement in late summer and would be possible with fi xed organ sizes. For autumn fi tted the predicted quantitative relationship (grey an illustration, we stay with red knots but dots; mean ± SE), whereas gizzards of birds not seen again move half a world away to the East-Asian– within a year were signifi cantly smaller (open dots; the two Australasian fl yway. groups almost entirely overlap in the poorest-quality year). Adapted from van Gils et al . ( 2006b ). Migrant fl exibility and speed explained under the assumption that, after of migration capture, birds increased their gizzards by an average 1 g ( van Gils et al . 2006b ). However, The red knots ‘wintering’ in the tropics of north- since birds with smaller gizzards also had west Australia breed on tundra on the remote lower body masses ( van Gils et al . 2006b ), the New Siberian Islands in the Arctic Ocean and, birds with especially small gizzards would not coincidentally, have been named after one of us have been able to make the change. With no by the Russian ornithologist Pavel S. Tomkovich ‘money in the bank’, they would be the ones to ( 2001 )— Calidris canutus piersmai . We were sur- move and/or to die. prised to discover that these ‘piersmai -knots’ All this material on red knots in the Wadden leave northwest Australia only in early to mid- Sea goes to show that, by using their capacity May ( Battley et al . 2005 ), more than a month to change the size and shell-processing capac- after the exodus of most other locally wintering ity of the gut, red knots have an ability to cope migrant shorebirds (Tulp et al . 1994 ). At this with variations in prey quality between years. time they have less than a month to negotiate It has also become clear that limits to this 10,400 km of airspace before having to arrive on capacity to change, e.g. by not having the the tundra in early June. On the basis of resight- stores to ‘pay’ for the change, will affect the ings of colour-ringed birds, we know that they birds’ decision to stay or to move on, and will stage somewhere in the Yellow Sea. With one probably make the difference between life and million hectares of mud, the Yellow Sea mud- death. For red knots to capitalize on what the fl ats are more extensive than those of the whole world offers them by way of particular, Wadden Sea, but, unlike the Wadden Sea, they season-specifi c resources (high quality shell- are still being reclaimed and transformed with fi sh at temperate and tropical latitudes when great speed and effi ciency ( Fig. 110 ; Barter 2002 , not breeding, safety and abundant arthropods van de Kam et al . 2008 ). We now know that dur- at high latitudes when breeding; Piersma et al . ing northward migration red knots choose the CONSERVATION AND MANAGEMENT OF FLEXIBLE PHENOTYPES 167

arrival (H.-Y. Yang et al. in prep.), this example emphasizes that fl exible migrants will be faster migrants, and certainly will have more options during lean times. It also indicates that changes in prey quality at staging sites easily interfere with the annual cycles of such demanding migrants.

Global change and phenotypic change: plasticity prevails Figure 110. That the world is changing fast is illustrated here The reclamation of mudfl ats currently taking by newly built motorways crossing mudfl ats near Incheon, place in the Yellow Sea ( Fig. 110 ) is just the tip Greater-Seoul, South-Korea. The extent of areas where shorebirds and fi shermen can both fi nd their marine resources of an iceberg of human-wrought change to the is steadily becoming smaller. planet (e.g. Vitousek et al . 1997 ). The Millenium Photographed on 7 September 2007 by Jan van de Kam. Ecosystem Assessment ( 2005 ), carried out at the initiative of the United Nations, reports that 60% of the world’s ecosystem services extensive western mudfl ats bordering Bohai have been degraded. Many Mediterranean Bay, in China, a few hundred km east of its capi- and temperate habitats had already been con- tal Beijing ( Battley et al . 2005 , Yang et al . 2008 ). verted by 1950. In the tropics this process of Depending on the quality of the prey that conversion and degradation of habitats is still they fi nd in the Yellow Sea (expressed in terms in full swing (Fig. 112 ). With a quarter of the of units of energy per unit dry shell mass), and earth’s land surface now cultivated, and with depending on the body mass they have to much of the seafl oor affected by fishing gear reach before the onward fl ight to the New ( Pauly 2007 ), the biosphere is heavily affected Siberian Islands, red knots might well have by humankind. Even the current warming of enough time to refuel in less than the single the atmosphere is attributed in part to the month available ( Fig. 111 ). However, for birds emission of greenhouse gases by an ever with infl exible gizzards, prey quality in China expanding human population (e.g. Allen et al . has to be fi ve times the measured prey quality 2009 ). Despite the considerable power of in northwest Australia (Fig. 111B ); whereas for organisms to cope with environmental change, birds with fl exible gizzards, prey quality is not for many, the complete disappearance of such a constraint ( Fig. 111D ). Still, at prey (micro-) habitats exceeds their ability to cope. qualities somewhat lower than those in Species extinction rates are now 100–1000 Australia, red knots would have to double the times the background rate ( Lawton and May size of a 7-g gizzard upon arrival in China ( Fig. 1995 , Millenium Ecosystem Assessment 2005 ). 111C ), for which they would need to bring As biologists there is thus much for us to stores (‘money in the bank’) and which would deplore, but the good news is that by now our take time (about a week under the best of con- trade has quantifi ed biological phenomena ditions; Dekinga et al . 2001 ). Although the for long enough to look for patterns in what recently measured quality of knot-food in phenotypic changes can teach us. Andrew Bohai Bay suggests that red knots in recent Hendry and Michael Kinnison (1999) endeav- years had rather little adjusting to do upon oured to assemble and maintain a compilation 168 TOWARDS A FULLY INTEGRATED VIEW

inflexible gizzard flexible gizzard

4 A C

no change required 0

–2 required change in gizzard mass (g)

–4 PREY QUALITY IN AUSTRALIA

50 B D

estimated depart 185 g 20 fuelling window depart 160 g

10 required fuelling time (days) 0 0.11 10 0.11 10 prey quality (kJ/g shell)

Figure 111. Fuelling schedules of red knots preparing to migrate onward to High Arctic breeding grounds from an intertidal staging area in the Yellow Sea. Left-hand plots (A and B) are for birds that do not adjust their gizzard mass after arrival from northwest Australia in the Yellow Sea staging area (infl exible gizzards), and right-hand plots (C and D) are for birds that do (fl exible gizzards). Prey quality in the Yellow Sea is varied along the X-axes. The upper plots show the predicted change in gizzard mass for birds that would fuel as fast as possible (note that in A an invariant gizzard is modelled so no change occurs). The lower plots show the time required to fuel up to departure masses of either 160 g or 185 g as a function of prey quality and, in (D), gizzard mass. The estimated ‘fuelling window’ based on observed arrival and departure dates is shaded, as is, for reference, the prey quality in northwest Australia. From Battley et al . ( 2005 ). of quantitative changes of animal phenotypes fact, phenotypic change due to human-driven through time (up to 200 generations). They disturbances of various kinds can be very included phenotypic changes at single sites, abrupt. Of obvious interest in the context of but also changes taking place when animals this book is the question whether these phe- were moved, or moved themselves, from one notypic changes refl ect changes in genetic site to another. Phenotypic change can be make-up, i.e. demonstrating natural selection expressed in two alternative, but correlated, in action (Endler 1986 ), or whether they metrics called Darwins and Haldanes (Hendry mainly represent phenotypic plasticity. and Kinnison 1999 ). The simplest is the On the basis of comparisons between assess- Haldane, the absolute change in standard ments of phenotypic change in free-living deviations of a trait per generation. In their organisms, Hendry et al . ( 2008 ) concluded that latest summary of the data, Hendry et al . ‘our analysis suggests a particularly important (2008 ) show that phenotypic change is perva- contribution from phenotypic plasticity’. In an sive in all studied contexts ( Fig. 113 ). independent meta-analysis for genetic responses Nevertheless, rates of phenotypic change were to climate change, Gienapp et al . ( 2008 ) simi- measurably greater in habitats changed by larly concluded that ‘clear-cut evidence indicat- humans compared with ‘natural contexts’. In ing a signifi cant role for evolutionary adaptation CONSERVATION AND MANAGEMENT OF FLEXIBLE PHENOTYPES 169

mediterranean forests, woodlands, and shrub range expansion after introduction temperate forest steppe and woodland self expansion temperate broadleaf and mixed forests tropical and sub-tropical new host dry broadleaf forests flooded grasslands and savannas introduction tropical and sub-tropical grasslands,savannas, in situ anthopogenic and shrublands disturbance tropical and sub-tropical coniferous forests in situ natural variation deserts

montane grasslands 0 123456 and shrublands phenotypic change (Haldane numerator) tropical and sub-tropical moist broadleaf forests Figure 113. Phenotypic changes associated with each temperate category of environmental change (see text for details). The data coniferous forests boreal summarized here are based on the mean value per system of the forests loss by 1950 Haldane metric for phenotypic change (the absolute change in loss between 1950 and 1990 tundra standard deviations per generation calculated as the difference projected loss by 2050 between two sample means divided by the within-population 0–10 20 40 60 80 100 standard deviation and the number of elapsed generations; fraction of potential area converted Haldane 1949 , Gingerich 1993 ). Boxes include 50% of the data, central horizontal lines represent medians, whiskers contain the Figure 112. There are numerous ways to visualize what remainder of the data excluding outliers (open circles). humans do to planet earth. By way of example, we have chosen Based on a compilation of 2414 Haldane rates. a summary of historical, ongoing, and projected conversions of Modifi ed from Hendry et al . ( 2008 ). terrestrial biomes worldwide from the United Nations’ Millenium Ecosystem Assessment. A biome is the largest unit of ecological classifi cation that can be conveniently recognized spring from the effects of the environment, to across the globe. The potential area of biomes was based on conclude that there was no evidence for any soil and climatic conditions and forms the basis for establishing the extent to which these had been converted by 1950, genetic change. between 1950 and 1990, and how much will have been Perhaps such conclusions were made rather converted by 2050. Near-terrestrial biomes, such as mangroves reluctantly; after all, quantitative geneticists and intertidal fl ats, were not included because their areas were strive to measure just that: quantitative genetic too small to be accurately assessed. Most of the conversion of changes. This leads us to think that the evidence biomes is to cultivated systems. from phenotype monitoring in a changing From Millenium Ecosystem Assessment ( 2005 ). world so far provides abundant evidence that to ongoing climate warming is conspicuously phenotypic responses to environmental change rare’. A case in point is a 47-year study on ring- refl ect phenotypic plasticity rather than genetic billed gulls Larus novaehollandiae in New change. Let us now examine some more exam- Zealand (Teplitsky et al . 2008 ). Over time these ples of this in relation to our changing climate. birds became smaller and lighter, a change that could be correlated with increasing local tem- How fl exible phenotypes cope peratures. Using advanced quantitative genetic with advancing springs modelling, Céline Teplitsky and co-workers were able to separate the expected effects of As we have admitted in Chapter 1 , the two genes that individuals would pass on to off- authors are both products of the Dutch 170 TOWARDS A FULLY INTEGRATED VIEW

8 A 800 C

6 700 4

2 600 0 temperature sum ( ° C) temperature 2 average temperature ( ° C) temperature average r = 0.33 −2 500

30 BD40 35 25 30 20 25 15 20

10 15

2 first egg date (since 1 March) 5 date (1 = April 1) mean laying 10 r = 0.36

1900 1920 1940 1960 1980 2000 1960 1970 1980 1990 2000 2010 year year Figure 114. Changes in spring temperatures in The Netherlands (A, average temperature between 16 February and 15 March) and in Oxfordshire, UK (C, sum of daily maximum temperatures between 1 March and 25 April) in relation to the fi rst egg date of northern lapwing in the province of Fryslân (B) and mean laying of great tits in Whytham Wood, Oxford (D). Compiled from Both et al . ( 2005 ) and Charmantier et al . ( 2008 ).

countryside, a totally man-made temperate Then as now, fi nding the fi rst egg of the sea- latitude human habitat. For us northerners, son brings honour and fame in its wake, includ- the arrival of spring is a happy event, and any ing a reception by royalty or their provincial changes in the timing of seasonal temperature representative, and ensuing newspaper cover- rises ( Fig. 114A ) will become a topic of conver- age ( Jensma 2009 ). Using these archival services sation. In the bad old days, one of the most we were able to go back more than a century to welcome features of spring was the sudden plot fi rst egg dates in Fryslân ( Fig. 114B ). These availability of fresh protein in the form of eggs varied by almost a month. Over this period of the earliest laying bird species. Mothers sent there have been huge changes in the manage- their children out to gather them, and, even ment of the grasslands where lapwings breed though our own mothers did not encourage (e.g. Beintema et al. 1985 , Schekkerman et al . us, at this point we have to admit that part of 2008 , Schroeder et al . 2009 ), but, to our surprise, our ecological education was learning to fi nd most of the variation in fi rst egg dates was (and then take and eat) the eggs of lapwings explained by weather, with little variation left Vanellus vanellus , a harvesting tradition still to account for statistically by meadowland- clung to in the province of Fryslân (Jensma management ( Both et al . 2005 ). A decrease in 2009 ). spring temperatures from 1900 to 1940 coin- CONSERVATION AND MANAGEMENT OF FLEXIBLE PHENOTYPES 171 cided with later fi rst egg dates. During the leaves. Caterpillar-phenology is tightly cou- period of temperature increase that followed, pled with temperature (Both et al . 2009 ), such lapwing eggs were found increasingly early. that pied fl ycatchers that advance egg-laying Similar patterns of change in the timing of in warmer springs produce more offspring reproduction relative to temperature changes than fl ycatchers that do not ( Bothet al . 2006 ). have now been established in several birds, with The adaptiveness of earlier egg-laying in pied nest-box breeding species yielding the most fl ycatchers is thus well-established. Whether detailed information because of ease of study. this phenotypic change also involves any An increase in spring temperature in the south- genetic change remains to be seen ( Both and ern UK ( Fig. 114C ), was correlated with an Visser 2005 , Both 2010 ). advance in laying date of the famous great tits of Wytham Woods, Oxford (Fig. 114D ). In this analysis ( Charmantier et al . 2008 ), and in a similar study of great tits in The Netherlands ( Nussey 20 et al . 2005 ), the change in timing is explained by 18 individual adjustment of behaviour, rather than by genetics. Hanging on to their predominantly 19 17 23 quantitative genetic paradigm, both studies 6 16 25 emphasize that selection would favour highly 15 22 phenotypically plastic individuals. 4 5 21 3 12 13 Observational studies at single sites where 14 trait changes are correlated with singular cli- 8,9,10,11 matic change variables, run the risk of confus- 7 24 ing correlation with causation. Capitalizing on 2 Europe-wide and largely independent studies 1 on another nest-box breeder, the long-distance 0.4 migrating pied fl ycatcher Ficedula hypoleuca 16 20 0.2 18 22 that winters in sub-Sahelian West Africa, 19 1 17 Christiaan Both and 22 co-authors established 0.0 23 near-experimental evidence that changing 2 3 –0.2 15 21 4 breeding times are actually causally related to 24 8 7 25 5 11 9 14 climate change ( Fig. 115 ). As we have seen –0.4 12 6 10 13 (Fig. 114A and C), over the last half-century, per year) date (days Slope laying –0.6 Western Europe has experienced steadily –0.10 –0.05 0.00 0.050.10 0.15 0.20 warming springs. Further south and north in Slope temperature (°C per year) Europe this is not the case (Fig. 115 ). It is Figure 115. The timing of pied fl ycatcher breeding has been impressive to fi nd that pied ycatchersfl bred followed over long periods in many nestbox studies distributed increasingly earlier in areas where tempera- across Europe. At some study sites there was a trend towards tures had gone up, but delayed egg-laying in colder springs (open circles), at some there was mild warming areas where temperatures had gone down! (open triangles, between 0 and +0.08°C per year), and at the rest there was a clear spring temperature increase (closed dots, Despite the suggestion made by their name, in more than +0.08°C per year). The inset shows that, depending many woods, pied fl ycatchers feed their young on the temporal change in spring temperature, pied fl ycatchers by relying on a peaked resource of small cater- either delayed or advanced egg laying. pillars that feast on crops of freshly budded Modifi ed from Both et al . ( 2004 ). 172 TOWARDS A FULLY INTEGRATED VIEW

For now we conclude that the phenotypic changes observed in so many taxa and systems in seeming response to climate change ( Parmesan and Yohe 2003 , Walther et al . 2005 ), do indeed demonstrate the adaptive power of fitness phenotypic plasticity. Arguably, for conservation managers, it becomes most interesting when the limits to such plasticity are reached plasticity mitigation pathology (for more tit examples, see Visser et al . 2003 , environmental change 2004 ), or even when the environmentally induced changes to the phenotype become Figure 116. Degrees and means by which organisms can maladaptive ( Monaghan 2008 ). When this cope with different degrees of environmental change: a three- phased reaction norm. Within a certain range of environmental happens, conservation-conscious governments change, development of an alternative phenotype has no that halt the dredging or reclamation of mud- negative fi tness consequences. Beyond this range, fi tness fl ats ( Fig. 110 ), reversing destruction of habi- declines because the development of an optimal phenotype is tats (Fig. 112 ), or even stabilizing or reversing constrained. As long as further phenotypic adjustments are atmospheric agents of climate change, could able to mitigate negative effects on fi tness, the decline is slow, make the greatest difference. but outside this ‘mitigation zone’, pathologies develop and fi tness drops steeply. From Monaghan ( 2008 ). When organisms can cope no more: limits to phenotypic change tions, initially adaptive adjustments may even Red knots returning light and with a small giz- turn into pathologies ( Sapolsky 2004 ). This zard from the tundra were less likely to stay at will reduce fi tness even further ( Fig. 116 ). traditional mudfl at sites in The Netherlands Moving to another habitat patch is a coping and survive (Fig. 109 ). They illustrate the prin- strategy that is relevant as long as suitable alter- ciple that, when times get tough, organisms natives remain, and as long as organisms have may be able to cope locally, may have the both the capacity and the resources to make a option to move, or may even be forced to move. Many organisms cannot make a move— move. Coping locally or moving would entail think of barnacles on a rock—and simply have small energy costs, so that, over a certain range to adjust their phenotype when the environ- of environmental changes, fi tness cost could ment changes. A category of coping, particu- be negligible (Fig. 116 ). At some level of envi- larly relevant to humans, is the downstream ronmental change, however, organisms are no effect of adjustments made early in life on health longer able to express the locally optimum and vigour later in life. In a study of the joyful phenotype, and in this range, the necessary dancing corvids called red-billed choughs phenotypic adjustments do entail fi tness costs. Pyrrhocorax pyrrhocorax on the Scottish island of For an example, recall from Chapter 4 that at Islay, Jane Reid and co-workers (2003) showed some levels of work (energy expenditure) that birds raised in poor environments carried required to survive demanding conditions this stamp from their childhood throughout (including the provisioning of offspring), there life: they survived a little worse and showed are negative downstream consequences, nota- lower breeding success, even if they had mean- bly on survival. And when the going really while moved to the best chough habitats later gets tough, under severely stressful condi- in life. In choughs, chicks tend to cope with CONSERVATION AND MANAGEMENT OF FLEXIBLE PHENOTYPES 173 days of food shortage (and no growth), by faster genes infl uences phenotypes. Whether there compensatory growth as soon as the food sup- is more to phenotypic evolution than gene ply improves ( Schew and Ricklefs 1998 , Metcalfe selection, and the very role of phenotypic and Monaghan 2001 ). That such growth accel- plasticity in evolution, will be the subject of erations may actually come with costs later in the next and fi nal chapter. life was shown in an elegant experiment with zebra fi nches. Here the degree of growth com- Synopsis pensation was negatively correlated with the speed of learning a simple discriminatory task Humans modify habitats at a fast pace. With later in life ( Fisher et al . 2006 ). the disappearance of species-specifi c ‘good’ Similar negative delayed effects of com- habitat there is no way to avoid individual dis- pensatory growth on cognition have been appearance and species extinctions. Species reported for small-born human infants (e.g. that persist tend to show considerable direc- Estourgie-van Burk et al. 2009 ). It is the tional phenotypic change across generations. absence of a ‘silver spoon’ that tragically Despite the best attempts to demonstrate that condemns some human groups to disease such changes refl ect genetic change, rmfi evi- and poverty, partly because of the delayed dence for such micro-evolutionary processes detrimental effects of poor nutrition in early in multicellular organisms is still quite limited. life (Gluckman et al . 2009 ), or even during the Instead, plants and animals have great pheno- pregnancy of the mother ( Roseboom et al . typic plasticity, which, in several well-studied 2006 ). What is worse, longitudinal studies of examples from the global ‘experiment’ of cli- children born in Amsterdam immediately mate change, helps these organisms to adjust after the Dutch winter famine at the end of and survive. Unless environmental conditions the Second World War, showed that pre-natal become so bad that coping strategies generate famine may lead to poorer health, even in the delayed negative effects, or even phenotypic offspring of these post-war children (Painter breakdown, the various kinds of phenotypic et al . 2008 ). That health transcends genera- plasticity certainly help organisms to survive tions suggests that information other than in a world of change. CHAPTER 10 Evolution in fi ve dimensions: phenotypes fi rst! . . . actually, if you say that genetics is relating variation in the genome to variation in phenotype, there is more accessible variation in expression than there is in sequence, and there is more variation in phenotype between cells and tissues and organs than between people. P. Brown in an interview by J. Gitschier ( 2009 )

And the older identical twins become, the eas- individual’s shape, size, and capacity will often ier it is to tell them apart! Not surprisingly, the be direct functions of the ecological demands shaping roles of development and environ- placed upon them. Bodies express ecology. ment thus leave their marks on bodies that What we have been talking about are adapta- carry identical genomes. Bodies are fl exible, tions, the functionally appropriate adjustment they change with age and with environment, of phenotypes. In assembling our case, we have but how does this relate to evolution? In this tried to foster the view that, for evolutionary chapter we will elaborate, perhaps in a some- insight, phenotype–environment interactions what partisan manner, the view that there is need to be studied functionally (Feder and Watt much more to evolution than selection of 1992 , Watt 2000 ). Rather than emphasizing sim- genes, which is the single axis of inheritance ply that the capacity for phenotypic change is an acknowledged in the ‘Modern Synthesis’ (e.g. adaptation (which it might well be; see Chapter 5 Mayr 1982 , 2002 , Coyne 2009a ). As scientists and Pigliucci 2001a , 2001b ), inter- and especially we seem to be on the verge of a ‘post-Modern intra-individual trait variation can be used to Synthesis’ in which the fl exible phenotype, in evaluate the ‘goodness of design’ criterion for interaction with its environment, will take cen- phenotypic adaptation. As argued by Piersma tre stage. and Drent ( 2003 ), intra-individual phenotypic variations can demonstrate how alternative Flexible phenotypes and the study phenotypic representations of a single genotype of adaptation represent adaptations. This is particularly appropriate when interspecifi c comparisons This book has presented many examples of cannot be made. For example, to overcome the organisms responding to environmental chal- absence of matched food-specialists within the lenges, including humans during space-travel same taxon when studying digestive adapta- and competing in sports, and migrant birds tions, Levey and Karasov ( 1989 ) compared the conquering much of the globe in the course of fl exible phenotypes of birds that were primarily their demanding annual itineraries. We hope to frugivorous in autumn and insectivorous in have succeeded in building an ever more pow- spring. erful case that organisms and their particular Perhaps most importantly, reversible intra- environments are inseparable: changes in an individual phenotypic variation gives us a

174 EVOLUTION IN FIVE DIMENSIONS: PHENOTYPES FIRST! 175 powerful tool to study phenotypic design 1992 ). The question of whether this evolution- experimentally ( Sinervo and Basolo 1996 , ary moulding is strictly a process of novel Sinervo 1999 ), almost in the same way that genetic instruction, or mostly one of the self- behavioural ecologists have so successfully organized nature of pre-existing variations unravelled the functions of behavioural traits ( Newman and Müller 2000 , Camazine et al . (Krebs and Davies 1987 ). In one tradition, such 2001 , Turner 2007 ), cries out for much more ‘phenotypic engineering’ has made use of hor- investigation than we have the opportunity to mone-implants, including the male sex hor- indulge in here. Consequently we shall be mone testosterone ( Sapolsky 1997 , Ketterson obliged to leave it to one side and focus on the and Nolan 1999 ). In the dark-eyed juncos Junco links between phenotypic plasticity and hyemalis studied by the extended wife-and- evolution. husband team of Ellen D. Ketterson and Val Nolan (Ketterson et al . 1991 , 1996 ), testosterone Separating the environment affected an array of phenotypic traits and life- from the organism . . . history characteristics. This wide range of (interacting) effects actually limited the power In an insightful essay, Richard Lewontin and of the hormonal implant approach. One of the Richard Levins (2007) suggest that the most examples of how natural and experimental powerful contribution made by Charles variations of phenotype can enlighten issues Darwin to the development of modern biology of evolutionary adaptation, is spelled out in was not his satisfactory evolutionary mecha- Chapters 6 and 7 . It is concerned with the vari- nism, but his ‘rigorous separation of internal ably large muscular gizzards of birds eating and external forces that had, in previous theo- hard-shelled prey ( Piersma et al . 1993b ). Our ries, been inseparable’. Before Charles Darwin, work with red knots showed that birds not during the times of Lamarck and Charles’ only increased gizzard size when they shifted grandfather Erasmus, many scientists enter- feeding from soft to hard food, but also tained the thought of evolution involving increased their capacity to process that hard modifi cation by descent. However, by allow- food, and vice versa ( van Gils et al . 2003a ). As ing animals to propagate acquired characters we have seen in Chapter 9 , in an ecological in their offspring, they ‘totally confounded context where large gizzards are necessary, inner and outer forces in an unanalyzable individual red knots that managed to build up whole’ ( Lewontin and Levins 2007 ). Charles larger gizzards after returning from the Arctic Darwin separated a force (mutation) that was tundra, were the ones who survived longest entirely internal to the organism and deter- ( van Gils et al . 2006b ). mined the variation between individuals, from At this point in the argument, the study of another force (natural selection) that was appropriate phenotypic adjustment within entirely external and determined which of generations meets the study of cross-genera- the variations would survive and reproduce tional aspects of phenotypic adjustment— (survival of the fi ttest). Half a century evolution proper. In evolutionary biology, a after Darwin’s (1859) ‘ The origin of species ’, phenotypic trait can be considered an adapta- Johannsen ( 1911 ) proposed another distinction tion only if there is evidence that it has been between the ‘internal and external forces’ moulded in specifi c ways during its evolution- when he considered the genotype to be the ary history to make it more effective for its embellishment of a genome with its genetic particular role ( Williams 1966b , West-Eberhard instruction, and the phenotype to be the physi- 176 TOWARDS A FULLY INTEGRATED VIEW

cal entity that interacts with the outside world. modular organization of the phenotype. Step (4) Evolution happens because variable pheno- argues that organisms that respond in a functional types are differentially successful in propagat- manner to new conditions are selected . This is fol- ing the genetic instructions that they carry to lowed by step (5), where genetic changes that subsequent generations. make the adaptive phenotypic change more precise This separation of internal and external and less costly are selected . forces undoubtedly led to the incredibly suc- Step (2) represents the process of phenotypic cessful modern reductionist biology that we accommodation, the ‘self-righting property’ of witness today. However, the many examples developmental systems that ensures that any of externally-driven but internally-processed intrinsic change will trigger a sequence of reg- and reversible intra-individual phenotypic ulatory changes to generate an integrated phe- responses that have been reviewed in this book notype automatically ( Alberch 1989 , Blumberg illustrate the limitations of a world view that 2009 , Coyne 2009b ). With respect to step (3), ignores the complex interplay between geno- arguments too complicated to be brief about type, body, and environment. According to here (but expounded in West-Eberhard 2003 , Lewontin and Levins ( 2007 ), the ‘separation Kirschner and Gerhart 2005 ), have been devel- [between internal and external forces] is bad oped to suggest that a modular design best biology and presents a barrier to further enables a developing phenotype to sort itself progress’. out adaptively under new environmental conditions (see also Parter et al . 2008 ). Step (5) represents the concepts of genetic assimilation . . . and putting them back together: and genetic accommodation ( Crispo 2007 ). phenotypes fi rst!

In her book of epic proportions Developmental Genotypes accommodating plasticity and evolution , Mary Jane West- environmental information? Eberhard ( 2003 ) sought to give developmental biology back its rightful place in evolutionary These confusing and related concepts, genetic biology.1 Most relevant for our current stream assimilation and accommodation, originated of thought is her fi ve-step scenario for the role over 100 year ago ( Baldwin 1896 , Morgan 1896 , of phenotypic plasticity in the generation of Osborn 1897 , Waddington 1942 ). They have novelty, as well as adaptedness, in evolution generated an interesting theoretical literature (this summary is adapted from Jablonka (e.g. Pigliucci 2001a , West-Eberhard 2003 , 2006 ). Lande 2009 ), which we shall have to leave to Step (1), and we have seen abundant exam- one side in order to concentrate on the main ples of this, is the realization that organisms are purposes of this book. Following Pigliucci et al . inherently responsive to environmental changes (2006 ), we have chosen to use to use the term (italics for emphasis). Step (2) states that such genetic assimilation for the important, but diffi - phenotypic responses, which often are adaptive, cult, concept that, after a shift in environmen- represent the ‘normal’ reaction to new challenges, tal conditions, developmental plasticity can where ‘normal’ means that this capacity is built-in . produce a partially adapted phenotype, which Step (3) has not been introduced so far. It states subsequently becomes fi ne-tuned to the new that responses to changed conditions depend on the conditions by natural selection ( Lande 2009 ).

1 The burgeoning literature on the role of plasticity in evolution is testimony to West-Eberhard’s success. EVOLUTION IN FIVE DIMENSIONS: PHENOTYPES FIRST! 177

In a hierarchical order, natural selection would changing environments requires generation of represent the central force selecting the fi ttest novel developmental variation, but heritabil- phenotypes under all conditions. In changing ity of such variation should, by defi nition, environments these would be represented by limit the range of future variability’. Figure properly adjusted phenotypes from a plastic- 117A illustrates, in the most general terms, ity range (i.e. because the organism is able to how genetic accommodation could take place. perform the appropriate phenotypic accom- Summarizing many years of fi eldwork on the modation). If environmental conditions per- rapidly colonizing house fi nch Carpodacus sist, any costs of plasticity (see Table 7 ) would mexicanus (a species that until the 1850s was ensure that novel phenotypes become geneti- limited to the warmer parts of west-central cally assimilated. USA and northern Mexico), Badyaev ( 2009 ) You will have noticed, of course, that the sketches a scenario of hormone-related induc- evolutionary scenario of West-Eberhard ( 2003 ) tion of novel phenotypes and their ‘subsequent represents a radically different view on how directional cross-seasonal transfer of a subset evolution proceeds from that which we have of developmental outcomes favoured by natu- all learned at school. In the ‘Modern Synthesis’, ral selection’ ( Fig. 117B ). This scenario would random mutations in the genome have to gen- account for the recent fast and locality-specifi c erate new phenotypic variants, of which the changes in phenotype and life-history traits of appropriate versions would then be selected this very successful colonizer. in the ambient environment. Appropriate phe- Threespine sticklebacks Gasterosteus aculea- notypic adjustments in new environments tus live in marine, as well as limnic, environ- have to await appropriate mutations and sub- ments, and are a celebrated case of adaptive sequent selection. This takes time, and many radiation (e.g. Schluter 1993 , 2000 ). After the animals have to die for a phenotypic change to retreat of the glaciers, over the last 15,000 take place, generating the ‘cost of natural selec- years, throughout the northern hemisphere tion’ or ‘mutational load’ identifi ed by Haldane sticklebacks have become isolated in lakes and ( 1957 , see Nunney 2003 ). According to this often show a kind of sympatric speciation established view, the natural order of things is (with genetic differentiation) towards a bot- that genes change fi rst and that phenotypes fol- tom-living blunt-headed form that specializes low . The more comprehensive view, which in feeding on bottom-living prey, such as insect incorporates considerations from developmen- larvae, and a mid-water form with a long beak tal biology and acknowledges the important that specializes in feeding on zooplankton. role of phenotypic plasticity ( West-Eberhard According to the genetic assimilation scenario, 2003 , 2005 ), states exactly the opposite: pheno- one would expect the original marine form to types change fi rst and genes follow . express full developmental plasticity for the What West-Eberhard ( 2003, 2005 ) has quite appropriate shapes of bottom- or mid-water- successfully engaged (see Gottlieb 1992 for an living fi sh and the specialized forms to have earlier attempt), is the research agenda ‘for the lost this capacity. Although there was evidence mechanisms that connect induction and reten- for some subtle differences in plasticity in the tion of within-generation accommodations in bottom-living form (Wund et al . 2008 ), all pop- evolutionary lineages’ ( Badyaev 2009 ). The ulations—despite their genetic divergence— struggle we face reconciling variability and displayed the whole range of functionally heredity has succinctly been stated by appropriate plasticity, but with little evidence Alexander Badyaev ( 2009 ) as ‘adaptation to for genetic assimilation ( Fig. 118 ). 178 TOWARDS A FULLY INTEGRATED VIEW

environment selection shifts only in new environment phenotype phenotype phenotype

environment environment environment

maternal behaviour

phenotype demotype prolactin environmental corticosterone shift androgens hormones developmental fitness variation

oogenesis and sex determination

environmental induction reproductive homeostasis: production of novel natural selection on novel of behavioural and channelling and accommodation developmental variation developmental variants and physiological variation of induced variation of the phenotype reproductive homeostasis

Figure 117. An explanation of how genetic assimilation should work (A), and one of the fi rst empirical examples of how it does work (B). In (A) the population is initially occupying a single environment. When the environment changes, a pre-existing reaction norm allows the population to persist because it produces novel phenotypes with no initial genetic change. If natural selection keeps operating in the new environment, the novel phenotype may become genetically fi xed (assimilated), and the original reaction norm may lose plasticity if there are costs associated with maintaining plasticity when it is not favoured. Panel (B) summarizes the origin of novel adaptations during range expansion by house fi nchesCarpodacus mexicanus . New environmental conditions induce novel behavioural and physiological variations. The involvement of a single set of hormonal mechanisms in regulating environmental assessment, incubation behaviour, and oogenesis, ensures the channelling and accommodation of the induced variation, resulting in production of novel functional phenotypic variants. Subsequently, natural selection acts on the resulting developmental variants and, if the fi tness benefi ts persist, developmental outcomes of the environmental induction would do too. Compiled from, respectively, Pigliucci et al . ( 2006 ) and Badyaev ( 2009 ).

Nevertheless, support for the notion of variations enabled Edgell et al . ( 2009 ) to show genetic assimilation is now accumulating. In that a critical behavioural anti-predation trait littorinid snails of the rocky shores of Atlantic (soft tissue withdrawal in the presence of North America, different populations, which crabs) actually lost its plasticity within a few are genetically indistinguishable, have histori- hundred years. A similar case was made by cally been exposed to an invasive predatory Aubret and Shine ( 2009 ) on the basis of newly crab for variable lengths of time. These time established island populations of tiger snakes EVOLUTION IN FIVE DIMENSIONS: PHENOTYPES FIRST! 179

reared on a diet of: ‘tarbut’ ( Avital and Jablonka 2000 , Jablonka bottom-living prey open water-living prey and Lamb 2005 ). Tarbutniks are unique in having perfect DNA-maintenance systems; all

marine these imaginary animals are genetically identical to each other. One would almost say that tarbutniks are unable to evolve. However, because of the downstream effects of chance events during development (we are reminded lake of the degree to which identical human twins bottom can differ, and see Vogt et al . 2008 for a wonderful example on variability in tarbutnik-like, source population clonal crayfi sh), tarbutniks do show variations in size and shape and the colour of their fur. lake open Tarbutniks live in small family groups and water join their parents on their foraging trips. In this way they learn about their environments, which foods are good and which foods are bet- Figure 118. Effect of either a diet of bottom-living or open- ter avoided (Fig. 119 ). Their capacity to learn water living prey on head shape in threespine sticklebacks Gasterosteus aculeatus reared from parents from the as young animals, but also as adults, means (evolutionary) source population in the sea (top), from a that new adaptive behaviours accomplished descendant lake population that lives on bottom-dwelling prey either by accident, or by copying others, will (middle) and a lake population with a mid-water distribution be transmitted to younger generations. During (bottom). The ‘deformation’ grids are exaggerated four times to courtship, tarbutniks have the habit of offering highlight shape differences. Compiled from Wund et al . ( 2008 ). nuptial gifts (males) and either accepting these gifts or not (females). Now imagine a tarbutnik female that has been raised in a family that Notechis scutatus around the Australian conti- is fond of berries. She fl atly refused the atten- nent. These snakes benefi t from large heads in tions of a nut-offering male, but accepts the places where large prey are available. Such male that is offering berries (Fig. 119 ). large heads are achieved by phenotypic plas- Everything else being equal, the population ticity in 30-year young populations but seem would steadily separate itself into a berry- and fi xed by genetic assimilation in 9000-year old a nut-loving clan. populations. Whatever the evidence for genetic The clans would diverge even more rapidly assimilation, these examples illustrate once if the berry-loving tarbutniks were to hit upon more that genotypes and phenotypes show a the invention of protecting and nurturing fair degree of decoupling. berry-bearing plants; or if nut-loving tarbutniks were to start cultivating nut-bearing Enter the tarbutniks, and niche bushes. At this point tarbutniks have ‘con- construction structed their niche’ (Laland et al . 2008 ), and this niche cannot but reciprocally infl uence Tarbutniks are hypothetical small rat-like ani- what the tarbutniks will look like. Imagine mals with long-whiskered snouts and long that berry-bearing plants are small, but that tails with two streamers ( Fig. 119 ). They were nut-bearing bushes need climbing into. The named after the Hebrew word for culture, i.e. tarbutniks that love nuts would train their 180 TOWARDS A FULLY INTEGRATED VIEW

Figure 119. Socially mediated learning in tarbutniks, long-tailed and long-snouted vertebrates that are genetically identical and have perfect DNA-maintenance systems so that genes never change. Chance events during development ensure that, just like identical twins, they do show small differences in size, shape, fur colour and various aspects of their learned behaviour. In (A) a young tarbutnik is introduced to carrots by its mother, which it consequently eats throughout life (B). In (C) a berry-loving female responds to the advances of a berry-offering male, but (D) shows that she rejects the advances of a male offering nuts. Illustrations by Anna Zeligowski, from Jablonka and Lamb ( 2005, p. 157 and p. 160).

young to be agile climbers from a tender age isms construct their own niches ( Odling-Smee onwards. On the basis of use–disuse and other et al . 1996 , 2003 , Laland et al . 1999 , Olff et al . self-organizing principles that we have 2009 ), so will the niches ‘construct’ the organ- encountered in this book, even our untrained ismal phenotype. Here is another reason why, eyes would be able to tell a nut-loving tarbut- even in the evolutionary sense, organisms and nik from a berry-loving one. environments are inseparable. Let us now, for the sake of argument, bring back in genes and genetic variation. Let us Evolution in four or fi ve dimensions? turn tarbutniks into something that vaguely resembles them, the short-lived Australian In their imaginative book Evolution in four marsupial mice that go under the genus name dimensions , Eva Jablonka and Marion Lamb Antechinus , for example. What we would (2005) expound the view that, in addition to the undoubtedly see in antechinal tarbutniks is co- single evolutionary dimension that we all are evolution between habits and genes; there used to, and work with—genetic inheritance— would be ‘reciprocal causation’ ( Fig. 120 ; there are three more ways that informational Laland et al . 2008 ). In the same way that organ- carry-over between generations can generate EVOLUTION IN FIVE DIMENSIONS: PHENOTYPES FIRST! 181

Natural selection evolution ( Fig. 121 ). With the tarbutniks, we have just been introduced to what they call ‘the behavioural inheritance system’ ( Avital and Jablonka 2000 ). This represents the building of Reciprocal the niche; routines can be taught to offspring ENVIRONMENT Causation ORGANISM and thus transferred down the generations. In addition, they distinguish an inheritance route of extra-genetic information through the germline called ‘cellular epigenetic inheritance’ ( Jablonka and Raz 2009 ), from an inheritance Niche construction route that includes maternal effects called Figure 120. A schematic summary of the two-way interactions ‘developmental inheritance’ ( Mousseau and between organisms and environment called ‘reciprocal Fox 1998 , Bateson et al . 2004 , Groothuis et al . causation’ combines the forces of natural selection (organisms 2005 ). The fi nal inheritance pathway is one that shaped by the environment) and niche construction (environ- humans have specialized in, a mechanism using ments shaped by the organism). Another way of describing reciprocal causation is to say that it represents the evolu- symbols and including the writing of books, we tionary feedbacks between phenotype-induced changes to their call culture. ambient environments (niche) and the changed selection There are various ways to cut the cake, of pressures by these environments on the phenotype (body). course. Jablonka and Lamb ( 2005 , 2007 ) some- Diagram designed by F.M. Postma and J. van Gestel. how decided that their fi ve inheritance path-

culture

symbolic

niche behavioural

body

developmental EPIGENETIC INHERITANCE cellular epigenetic

genetic germ gamete INHERITANCE PATHWAYS GENERATION X GENERATION X+1 GENERATION X+2 Figure 121. As well as the genetic variation transmitted from generation to generation through the germline, there are four additional ‘epigenetic’ inheritance pathways. The fi rst of those, ‘cellular epigenetic inheritance’, transmits acquired information via the gametes. The remaining three are all ‘somatic’ and include ‘developmental inheritance’ (e.g. through maternal effects), ‘behavioural inheritance’ (through niche construction), and ‘symbolic inheritance’. One of the reasons that it took biologists so long to distinguish these layers is that they are thoroughly intertwined. There is also ‘reciprocal causation’, especially between niche and body. Note that symbolic inheritance seems specifi c to organisms that read books, listen to CDs and watch movies and TV. Developed from Jablonka and Lamb ( 2005 , 2007 ), and correspondence with E. Jablonka. 182 TOWARDS A FULLY INTEGRATED VIEW

ways ( Fig. 121 ) represent four evolutionary Context, please! An orchestra in need dimensions, hence the name of their book; of a theatre they merged the cellular-epigenetic and developmental pathways into one and called it ‘the In The music of life: Biology beyond the genome , epigenetic inheritance system’. Here we stick the renowned physiologist Denis Noble ( 2006 ) to the fi ve inheritance pathways identifi ed: takes issue with what he regards as the unwar- genetic, cellular-epigenetic, (somatic) develop- ranted ‘reductionist causal chain’, which sees mental, behavioural, and symbolic. The latter organisms as simple expressions of their four may also be combined under the liberal genetic instruction (Fig. 122A). He compares banner ‘epigenetic inheritance’ ( Fig. 121 ). the organs, the systems of the body, to an These fi ve inheritance pathways, combined orchestra. What he calls ‘downward causa- with the quiet power of reciprocal causation, tion’, the higher level controls on gene expres- make phenotypes (and perhaps even geno- sion and the internal developmental feedbacks, types!) respond much faster than what was is compared to the conductor of that orchestra. ‘allowed’ under the one-dimensional, geneti- As a biomedic, perhaps he should be excused cally straightjacketed, ‘Modern Synthesis’ out- (though perhaps he should not be), but he and lined by Ernst Mayr and others. Phenotypes many others (e.g. Nüsslein-Volhard 2006 ) fail respond and interact at many different levels. to provide their orchestras with somewhere to The concept of a ‘fl exible phenotype’ repre- play, theatres that will co-determine whether sents the core of post-modern evolutionary an orchestra sounds good, fi nds a receptive thinking. audience and is worth the governmental sub-

Organism Behaviour Organs Higher-level triggers of Organismal cell Tissues performance signalling Organ systems Cells Higher-level controls of Organs gene Sub-cellular expression Tissues mechanisms Cells Pathways Organelles

Proteins Proteins, etc. Protein Strength of machinery DNA reads genes natural or sexual Genes selection

Figure 122. Two visualizations of the types of interaction embodied in phenotypes. In (A), Denis Noble (2006 ) takes issue with the simplistic reductionist causal chain (small upward arrows), pointing out that at all levels of biological organization there are loops of interacting ‘downward and upward’ causation, here exemplifi ed by the controls of gene expression by higher organizational levels than the gene. In (B), Garland and Kelly ( 2006 ) use the same upward causal chain of increasing complexity to make the point that natural and sexual selection will act most strongly on the highest levels of biological organization. Natural selection typically does not act on genetic variation. EVOLUTION IN FIVE DIMENSIONS: PHENOTYPES FIRST! 183 sidies, public donations or philanthropic environmental context, might well all be a efforts. The theatre of an organism, of course, waste of time. Drastically paraphrasing is the environment. It is the environment that Dobzhansky ( 1973 ) we believe that nothing in will determine whether the orchestra will suc- evolution makes sense except in the light of ceed. But is there an audience that determines ecology! whether in a particular theatre a particular Richard Lewontin ( 2000 ), on the last pages orchestra sounds good? Not visibly, certainly, of his long essay on the interactions between but natural and sexual selection can be consid- genes, organisms, and environments The tri- ered to enact the role of invisible spectators. ple helix: gene, organism, and environment , As Noble wrestled with downward causa- laments that all too often research efforts tion and his musical metaphors, the evolution- focus on the measurable, rather than on the ary physiologists Theodore Garland and Scott frontiers of insight. The invention of protein- Kelly designed a diagram (Fig. 122B) that gel electrophoresis to assay genetic diversity embodies some aspects of Noble’s representa- led a whole generation of evolutionary genet- tion and some aspects of the diagram with icists to do just that, assay genetic diversity. which we started this book ( Fig. 1 ): the posi- The advent of ever more powerful gene- tioning of a developing phenotype in the con- sequencers may now limit biology to the text of its environment. Garland and Kelly questions that can be answered by DNA- (2006 ) make the point that natural and sexual sequences. What has driven the writing of selection act most strongly on aspects of the this book is the realization that there is a con- phenotype at the highest organizational levels, tinuum between organisms and environ- ‘because they are the most strongly correlated ments that needs much more attention in the with Darwinian fi tness’, the demotype. complementary realms of physiological, Excitingly, it is at the highest organizational behavioural, ecological, and evolutionary layers of an organism, in the phenotype, the study—the notable complication that even ethotype, and the demotype (Fig. 1 ), where the ‘small changes in the environment lead to infl uence of the environment is most pervasive small changes in the organism that, in turn, and where reciprocal causation feedbacks will lead to small changes in the environment come to the fore. ( Lewontin 2000 ). What we have come to real- Perhaps then, behaviour has primacy in ize as well is that life scientists must not only evolution ( Bateson 2004 , Aubret et al . 2007 )? include the ecological theatre in their think- The tarbutniks of Avital, Jablonka, and Lamb ing and their measurements, but also that in show how behavioural change and learning the biology of the future there should be equal can be the motor of evolution in a genetically attention to the two, complementary, roles of invariant organism. Behaviour, however, can the genetic coding of some processes and the also act as a buffer against change. A move to a self-organizational nature of how these proc- less stressful environment and a change of diet esses are played out in the development and can shield organisms from strong directional functioning of any organism. The fact that selection ( Duckworth 2009 ). Nevertheless, we ‘organisms are designed not so much because have made our point: ecological studies of natural selection of particular genes has made organisms that ignore the key elements of their them that way, but because agents of homeos- physiology, physiological studies with disre- tasis build them that way’ (Turner 2007 ), may gard for the organisms’ ecology, and behav- well be the most important reason that organ- ioural studies carried out in ignorance of the isms appear so wonderfully adapted to their 184 TOWARDS A FULLY INTEGRATED VIEW environments and their environments to change needs systems of inheritance, but we them. also fi nd that the genetic system that we all work with is just one of fi ve such possible Synopsis inheritance systems. Since organisms not only provide the beginnings and nurture of their In this book we have offered a progression of offspring, but also build the environments in ideas that support the view that, although which they and their offspring live, there are bodies and environments are recognizable very tight feedbacks of reciprocal causation, entities, they really are inseparable. In this both in the development of organisms and in fi nal chapter we extend this view to evolu- the relationships between the developing tion—the inheritance of, and selection for, ran- organism and their environments. Bodies are domly variable phenotypic traits across earth, and we would do well to acknowledge generations. What we fi nd is that evolutionary that in the ways that we study them. References

Abdalla, E.B., Kotby, E.A., and Johnson, H.D. (1993). cise performance via effects on peripheral locomotor Physiological responses to heat-induced hyperthermia muscle fatigue in humans. Journal of Physiology 575 , of pregnant and lactating ewes. Small Ruminant Research 937–52. 11 , 125–34. Anderson, K.J. and Jetz, W. (2005). The broad-scale ecol- Adamo, S.A., Barlett, A., Le, J., Spencer, N., and Sullivan, ogy of energy expenditure of endotherms. Ecology K. (2010). Illness-induced anorexia may reduce trade- Letters 8 , 310–18. offs between digestion and immune function. Animal Anderson, T.R., Hessen, D.O., Elser, J.J., and Urabe, J. Behaviour 79 , 3–10. (2005). Metabolic stoichiometry and the fate of excess Adams, G.R., Caiozzo, V.J., and Baldwin, K.M. (2003). carbon and nutrients in consumers. American Naturalist Skeletal unweighting: spacefl ight and ground-based 165 , 1–15. models. Journal of Applied Physiology 95 , 2185–2201. Arad, Z., Midtgard, U., and Bernstein, M.H. (1989). Alberch, P. (1989). The logic of monsters: Evidence for Thermoregulation in Turkey vultures: Vascular anat- internal constraint in development and evolution. omy, arteriovenous heat exchange, and behaviour. Geobios 19 , 21–57. Condor 91 , 505–14. Alerstam, T. (2003). Bird migration speed. In Avian migra- Arngrimsson, S.A., Petitt, D.S., Borrani, F., Skinner, K.A., tion, eds P. Berthold, E. Gwinner, and E. Sonnenschein, and Cureton, K.J. (2004). Hyperthermia and maximal pp. 253–67. Springer-Verlag, Berlin. oxygen uptake in men and women. European Journal of Alerstam, T. and Jönsson, P.E. (1999). Ecology of tundra Applied Physiology 92 , 524–32. birds: Patterns of distribution, breeding and migration Arsenault, D.J., Marchinko, K.B., and Palmer, A.R. (2001). along the Northeast Passage. Ambio 28 , 212–24. Precise tuning of barnacle leg length to coastal wave Alexander, R.McN. (1984). Optimum strengths for bones action. Proceedings of the Royal Society, London, B 268 , liable to fatigue and accidental fracture. Journal of 2149–54. Theoretical Biology 109 , 247–52. Aschoff, J. (1954). Zeitgeber der tierischen Tagesperiodik. Alexander, R.McN. (1997). A theory of mixed chains Naturwissenschaften 41 49–56. applied to safety factors in biological systems. Journal of Aschoff, J. and Pohl, H. (1970). Der Ruheumsatz von Theoretical Biology 184 , 621–36. Vögeln als Funktion der Tageszeit und des Körpergrösse. Alexander, R.McN. (1998). Symmorphosis and safety Journal für Ornithologie 111 , 38–47. factors. In Principles of animal design: The optimization Ashton, K.G. (2002). Patterns of within-species body size and symmorphosis debate, eds E.R. Weibel, C.R. Taylor, variation of birds: strong evidence for Bergmann’s rule. and L. Bolis, pp. 28–35. Cambridge University Press, Global Ecology and Biogeography 11 , 505–23. Cambridge. Ashton, K.G., Tracy, M.C., and de Queiroz, A. (2000). Is Allen, M.R., Frame, D.J., Huntingford, C., Jones, C.D., Bergmann’s rule valid for mammals? American Naturalist Lowe, J.A., Meinshausen, M., and Meinshausen, 156 , 390–415. N. (2009). Warming caused by cumulative carbon emis- Aubret, F. and Shine, R. (2009). Genetic assimilation and sions towards the trillionth tonne. Nature 458 , 363–68. the postcolonization erosion of phenotypic plasticity in Alonso-Alvarez, C., Bertrand, S., Devevey, G.L., Prost, J., island tiger snakes. Current Biology 19 , 1932–36. Faivre, B., Sorci, G., and Loeschcke, V. (2004). Increased Aubret, F., Bonnet, X., and Shine, R. (2007). The role of susceptibility to oxidative stress as a proximate cost of adaptive plasticity in a major evolutionary transition: reproduction. Ecology Letters 7 , 363–68. Early aquatic experience affects locomotor performance Amann, M., Eldridge, M.W., Lovering, A.T., Stickland, of terrestrial snakes. Functional Ecology 21 , 1154–61. M.K., Pegelow, D.F., and Dempsey, J.A. (2006). Arterial Austad, S.N. (1997). Why we age: What science is discovering oxygenation infl uences central motor output and exer- about the body’s journey through life . Wiley, New York.

185 186 REFERENCES

Avital, E. and Jablonka, E. (2000). Animal traditions: behav- Bateson, P. (2004). The active role of behaviour in evolu- ioural inheritance in evolution . Cambridge University tion. Biology and Philosophy 19 , 283–98. Press, Cambridge. Bateson, P., Barker, D., Clutton-Brock, T., Deb, D., D’Udine, Bacigalupe, L.D. and Bozinovic, F. (2002). Design limita- B., Foley, R.A., Gluckman, P., Godfrey, K., Kirkwood, tions and sustained metabolic rate: Lessons from small T., Lahr, M.M., McNamara, J., Metcalfe, N.B., Monaghan, mammals. Journal of Experimental Biology 205 , 2963–70. P., Spencer, H.G., and Sultan, S.E. (2004). Developmental Badyaev, A.V. (2009). Evolutionary signifi cance of pheno- plasticity and human health. Nature 430 , 419–21. typic accommodation in novel environments: An empir- Battley, P.F. and Piersma, T. (2005). Adaptive interplay ical test of the Baldwin effect. Philosophical Transactions of between feeding ecology and features of the digestive the Royal Society, London, B 364 , 1125–41. tract in birds. In Physiological and ecological adaptations to Bailey, E. (1998). Odds on the FAST gene. Genome Research feeding in vertebrates, eds J.M. Starck and T. Wang, 8 , 569–71. pp. 201–28. Science Publishers, Enfi eld. Baker, A.J., Piersma, T., and Greenslade, A.D. (1999). Battley, P.F., Piersma, T., Dietz, M.W., Tang, S., Dekinga, A., Molecular versus phenotypic sexing in red knots. Condor and Hulsman, K. (2000). Empirical evidence for differ- 101 , 887–93. ential organ reductions during trans-oceanic bird fl ight. Baker, A.J., González, P.M., Piersma, T., Niles, L.J., de Lima Proceedings of the Royal Society, London, B 267 , 191–96. Serrano do Nascimento, I., Atkinson, P.W., Clark, N.A., Battley, P.F., Dietz, M.W., Piersma, T., Dekinga, A., Tang, Minton, C.D.T., Peck, M.K., and Aarts, G. (2004). Rapid S., and Hulsman, K. (2001). Is long-distance bird fl ight population decline in red knots: Fitness consequences of equivalent to a high-energy fast? Body composition decreased refuelling rates and late arrival in Delaware changes in freely migrating and captive fasting great Bay. Proceedings of the Royal Society, London, B 271 , knots. Physiological and Biochemical Zoology 74 , 875–82. 435–49. Bakken, G.S. (1976). A heat transfer analysis of animals: Battley, P.F., Rogers, D.I., Piersma, T., and Koolhaas, A. Unifying concepts and the application of metabolism (2003). Behavioural evidence for heat-load problems in chamber data to fi eld ecology. Journal of Theoretical great knots in tropical Australia fuelling for long-dis- Biology 60 , 337–84. tance fl ight. Emu 103 , 97–103. Bakken, G.S. and Gates, D.M. (1974). Linearized heat Battley, P.F., Rogers, D.I., van Gils, J.A., Piersma, T., Hassell, transfer relations in biology. Science 183 , 976–77. C.J., Boyle, A., and Yang, H-Y. (2005). How do red knots Baldwin, J.M. (1896). A new factor in evolution. American Calidris canutus leave Northwest Australia in May and Naturalist 30 , 354–451. reach the breeding grounds in June? Predictions of stop- Banavar, J.R., Damuth, J., Maritan, A., and Rinaldo, A. over times, fuelling rates, and prey quality in the Yellow (2003). Allometric cascades. Nature 421 , 713–14. Sea. Journal of Avian Biology 36 , 494–500. Barja, G., Cadenas, S., Rojas, C., Perez-Campo, R., and Bauchinger, U. and McWilliams, S. (2009). Carbon turno- López-Torres, M. (1994). Low mitochondrial free radical ver in tissues of a passerine bird: Allometry, isotopic

production per unit O 2 consumption can explain the clocks, and phenotypic fl exibility of organ size. simultaneous presence of high longevity and high aero- Physiological and Biochemical Zoology 82 , 787–97. bic metabolic rate in birds. Free Radical Research 21 , Bauchinger, U., Kolb, H., Afi k, D., Pinshow, B., and 317–28. Biebach, H. (2009). Blackcap warblers maintain diges- Barnes, M. (1992). The reproductive periods and condition tive effi ciency by increasing digesta retention time on of the penis in several species of common Cirripedes. the fi rst day of migratory stopover. Physiological and Oceanography and Marine Biology: an Annual Review , 30 Biochemical Zoology 82 , 541–48. 483–525. Bech, C., Langseth, I., and Gabrielsen, G.W. (1999). Barter, M. (2002). Shorebirds of the Yellow Sea: impor- Repeatability of basal metabolism in breeding female tance, threats and conservation. International Wader kittiwakes Rissa tridactyla . Proceedings of the Royal Society, Studies 12 , 1–104. London, B 266 , 2161–67. Bartholomew, G.A. (1958). The role of physiology in the Beck, C. (2004). Personal portrait Bart Kempenaers— distribution of terrestrial vertebrates. In Zoogeography , Ornithology. Max Planck Research 2 /2004, 64–67. ed. C.L. Hubbs, pp. 81–95. Publ. No. 51, American Bednekoff, P.A. and Lima, S.L. (1998). Re-examining safety Association for the Advancement of Science, in numbers: Interactions between risk dilution and col- Washington, DC. lective detection depend upon predator targeting Bartholomew, G.A. and Cade, T.J. (1963). The water econ- behaviour. Proceedings of the Royal Society, London, B 265 , omy of land birds. Auk 80 , 504–39. 2021–26. REFERENCES 187

Beintema, A.J., Beintema-Hietbrink, R.J., and Müskens, heart of the matter. Proceedings of the Royal Society, G.J.D.M. (1985). A shift in the timing of breeding in London, B 266 , 2275–81. meadow birds. Ardea 73 , 83–89. Black, A.E., Coward, W.A., Cole, T.J., and Prentice, A.M. Bennett, A.F. and Ruben, J.A. (1979). Endothermy and (1996). Human energy expenditure in affl uent societies: activity in vertebrates. Science 206 , 649–54. An analysis of 574 doubly-labelled water measure- Bennett, G.F., Montgomerie, R., and Seutin, G. (1992). ments. European Journal of Clinical Nutrition 50 , 72–92. Scarcity of haematozoa in birds breeding on the arctic Black, J.L., Mullan, B.P., Lorschy, M.L., and Giles, L.R. tundra of North America. Condor 94 , 289–92. (1993). Lactation in the sow during heat-stress. Livestock Berger, J. (2007). Fear, human shields and the redistribu- Production Science 35 , 153–70. tion of prey and predators in protected areas. Biology Blaxter, K. (1989). Energy metabolism in animals and man. Letters 3 , 620–23. Cambridge University Press, Cambridge. Bergmann, C. (1847). Über die Verhältnisse der Blumberg, M.S. (2009). Freaks of nature: What anomalies tell Wärmeökonomie der Thiere zu ihrer Grösse. Göttinger us about development and evolution. Oxford University Studien 3 , 595–708. Press, New York. Bergmann, O., Bhardwaj, R.D., Bernard, S., Zdunek, S., Bohning-Gaese, K., Halbe, B., Lemoine, N., and Barnabé-Heider, F., Walsh, S., Zupinich, J., Alkass, K., Oberrath, R. (2000). Factors infl uencing clutch size, Druid, H., Jivinge, S., and Frisén, J. (2009). Evidence number of broods and annual fecundity of North for cariomycete renewal in humans. Science 325 , American and European land birds. Evolutionary Ecology 98–102. Research 2 , 8230839. Berry, H. (1990). From LA to New York, from New York to LA . Bonner, J.T. (2006). Why size matters: From bacteria to blue Self-published, Chorley, UK. whales . Princeton University Press, Princeton, NJ. Bibby, R., Cleall-Harding, P., Rundle, S., Widdicombe, S., Bonnet, X., Bradshaw, A.D., and Shine, R. (1998). Capital and Spicer, J. (2007). Ocean acidifi cation disrupts versus income breeding: An ectothermic perspective. induced defences in the intertidal gastropod Littorina Oikos 83, 333–42. littorea . Biology Letters 3 , 699–701. Bonnie, K.E. and Earley, R.L. (2007). Expanding the scope Biebach, H. (1990). Strategies of trans-Sahara migrants. In for social information use. Animal Behaviour 74 , 171–81. Bird migration, ed. E. Gwinner, pp. 352–67. Springer- Bonser, R.H.C. (1996). The mechanical properties of feather Verlag, Berlin. keratins. Journal of Zoology 239 , 477–84. Biebach, H., Biebach, I., Friedrich, W., Heine, G., Partecke, Both, C. (2010). Flexibility of timing of avian migration to J., and Schmidl, D. (2000). Strategies of passerine migra- climate change masked by environmental conditions en tion across the Mediterranean Sea and the Sahara desert: route. Current Biology 20 , 243–48. a radar study. Ibis 142 , 623–34. Both, C. and Visser, M.E. (2005). The effect of climate Biernaskie, J.M., Walker, S.C., and Gegear, R.J. (2009). change on the correlation between life-history traits. Bumblebees learn to forage like Bayesians. American Global Change Biology 11 , 1606–13. Naturalist 174 , 413–23. Both, C., Artemyev, V.A., Blaauw, B., Cowie, R.J., Biggerstaff, S. and Mann, M. (1992). Consummatory Dekhuijzen, A.J., Eeva, T., Enemar, A., Gustafsson, L., behaviors and weight regulation in pregnant, lactating, Ivankina, E.V., Järvinen, A., Metcalfe, N.B., Nyholm, and pregnant-lactating mice. Physiology and Behavior 52 , N.E.I., Potti, J., Ravussin, P-A., Sanz, J.J., Silverin, B., 485–91. Slater, F.M., Sokolov, L.V., Török, J., Winkel, W., Wright, Bigland-Ritchie, B. and Vollestadt, N. (1988). Hypoxia and J., Zang, H., and Visser, M.E. (2004). Large-scale geo- fatigue: how are they related? In Hypoxia: the tolerable graphical variation confi rms that climate change causes limits , eds J.R. Sutton, C.S. Houston, and G. Coates, birds to lay earlier. Proceedings of the Royal Society, pp. 315–25. Benchmark, Indianapolis, Ill. London, B 271 , 1657–62. Bijleveld, A.I., Egas, M., van Gils, J.A., and Piersma, T. Both, C., Piersma, T., and Roodbergen, S. P. (2005). Climatic (2010). Beyond the information centre hypothesis: change explains much of the 20 th century advance in Communal roosting for information on food, predators, laying date of northern lapwing Vanellus vanellus in The travel companions and mates? Oikos 119 , 277–85. Netherlands. Ardea 93 , 79–88. Bilz, S., Ninnis, R., and Keller, U. (1999). Effects of hypoos- Both, C., Bouwhuis, S., Lessells, C.M., and Visser, C.M. molality on whole-body lipolysis in man. Metabolism - (2006). Climate change and population declines in a Clinical and Experimental 48 , 472–76. long-distance migratory bird. Nature 441 , 81–83. Bishop, C.M. (1999). The maximum oxygen consumption Both, C., van Asch, M., Bijlsma, R.G., van den Burg, A.B., and and aerobic scope of birds and mammals: Getting to the Visser, M.E. (2009). Climate change and unequal pheno- 188 REFERENCES

logical changes across four trophic levels: Constraints or Bryant, D.M. and Westerterp, K.R. (1980). The energy adaptations? Journal of Animal Ecology 78, 73–83. budget of the house martin ( Delichon urbica ). Ardea 68 , Bouwhuis, S., Sheldon, B.C., Verhulst, S., and Charmantier, 91–102. A. (2009). Great tits growing old: Selective disappear- Bubenik, G.A. and Bubenik, A.B. eds (1990). Horns, prong- ance and the partitioning of senescence to stages within horns, and antlers: Evolution, morphology, physiology, and the breeding cycle. Proceedings of the Royal Society, social signifi cance . Springer-Verlag, New York. London, B 276 , 2769–77. Buehler, D.M. and Baker, A.J. (2005). Population diver- Boyd, I.L. (2000). State-dependent fertility in pinnipeds: gence times and historical demography in red knots Contrasting capital and income breeders. Functional and dunlins. Condor 107 , 497–513. Ecology 14, 623–30. Buehler, D.M. and Piersma, T. (2008). Travelling on a Brakefi eld, P.M. and Reitsma, N. (1991). Phenotypic plas- budget: Predictions and ecological evidence for bottle- ticity, seasonal climate and the population biology of necks in the annual cycle of long-distance migrants. Bicyclus butterfl ies (Satyridae) in Malawi. Ecological Philosophical Transactions of the Royal Society, London, B Entomology 16 , 291–303. 363 , 247–66. Brakefi eld, P.M., Pijpe, J., and Zwaan, B.J. (2007). Buehler, D.M., Baker, A.J., and Piersma, T. (2006). Developmental plasticity and acclimation both contrib- Reconstructing palaeofl yways of the late Pleistocene ute to adaptive responses to alternating seasons of and early Holocene red knot (Calidris canutus ). Ardea 94 , plenty and of stress in Bicyclus butterfl ies. Journal of 485–98. Biosciences 32 , 465–75. Buehler, D.M., Matson, K.D., Piersma, T., and Tieleman, B.I. Briffa, M. and Sneddon, L.U. (2007). Physiological constraints (2008a). Seasonal redistribution of immune function in a on contest behaviour. Functional Ecology 21 , 627–37. shorebird: Annual-cycle effects override adjustments to Brisbin, I.L. (1969). Bioenergetics of the breeding cycle of thermal regime. American Naturalist 172 , 783–96. the ring dove. Auk 86 , 54–74. Buehler, D.M., Piersma, T., and Tieleman, B.I. (2008b). Brody, S. (1945). Bioenergetics and growth . Reinhold Captive and free-living red knots Calidris canutus exhibit (reprinted Hafner Publishing Company), New York. differences in non-induced immunity that suggest dif- Broggi, J., Hohtola, E., Orell, M., and Nilsson, J-Å. (2005). ferent immune strategies in different environments. Local adaptation to winter conditions in a passerine Journal of Avian Biology 39 , 560–66. spreading north: A common-garden approach. Evolution Buehler, D.M., Tieleman, B.I., and Piersma, T. (2009a). Age 59 , 1600–1603. and environment affect constitutive immune function in Brown, J.S. (1988). Patch use as an indicator of habitat red knots ( Calidris canutus ). Journal of Ornithology 150 , preference, predation risk and competition. Behavioral 815–25. Ecology and Sociobiology 22 , 37–48. Buehler, D.M., Encinas-Viso, F., Petit, M., Vézina, F., Brown, J.S. and Kotler, B.P. (2004). Hazardous duty pay Tieleman, B.I., and Piersma T. (2009b). Limited access to and the foraging cost of predation. Ecology Letters 7 , food and physiological trade-offs in a long-distance 999–1014. migrant shorebird. II. Constitutive immune function Brown, J.S., Laundré, J.W., and Gurung, M. (1999). The and the acute-phase response. Physiological and ecology of fear: Optimal foraging, game theory, and Biochemical Zoology 82 , 561–71. trophic interactions. Journal of Mammalogy 80 , 385–99. Buehler, D.M., Koolhaas, A., Van’t Hof, T.J., Schwabl, I., Bruinzeel, L.W., Piersma, T., and Kersten, M. (1999). Low Dekinga, A., Piersma, T., and Tieleman, B.I. (2009c). No costs of terrestrial locomotion in waders. Ardea 87 , evidence for melatonin-linked immunoenhancement 199–205. over the annual cycle of an avian species. Journal of Brunner, D., Kacelnik, A., and Gibbon, J. (1992). Optimal Comparative Physiology A 195 , 445–51. foraging and timing processes in the starling, Sturnus Bundle, M.W., Hoppeler, H., Vock, R., Tester, J.M., and vulgaris: Effect of inter-capture interval. Animal Behaviour Weyand, P.G. (1999). High metabolic rates in running 44 , 597–613. birds. Nature 397 , 31–32. Brunner, D., Kacelnik, A., and Gibbon, J. (1996). Memory Burness, G.P. (2002). Elephants, mice and red herrings. for inter-reinforcement interval variability and patch Science 296 , 1245–46. departure decisions in the starling, Sturnus vulgaris . Buttemer, W.A., Battam, H., and Hulbert, A.H. (2008). Animal Behaviour 51 , 1025–45. Fowl play and the price of petrel: Long-living Bryant, D.M. and Tatner, P. (1991). Intraspecies variation Procellariiformes have peroxidation-resistant mem- in energy expenditure: Correlates and constraints. Ibis brane composition compared with short-living 133 , 236–45. Galliformes. Biology Letters 4 , 351–54. REFERENCES 189

Cadée, N., Piersma, T., and Daan, S. (1996). Endogenous Chappell, M.A., Zuk, M., and Johnsen, T.S. (1996). circannual rhythmicity in a non-passerine migrant, the Repeatability of aerobic performance in red junglefowl: knot Calidris canutus . Ardea 84 , 75–84. Effects of ontogeny and nematode infection. Functional Calder, W.A., III (1984). Size, function, and life history . Ecology 10 , 578–85. Harvard University Press, Cambridge, Mass. Chappell, M.A., Bech, C., and Buttemer, W.A. (1999). The Calder, W.A., III and King, J.R. (1974). Thermal and caloric relationship of central and peripheral organ masses to relations of birds. In Avian biology , eds D. S. Farner aerobic performance variation in house sparrows. and J. R. King, Vol. 4 , pp. 260–415. Academic Press, Journal of Experimental Biology 202 , 2269–79. New York. Chappell, M.A., Rezende, E.L., and Hammond, K.A. Camazine, S., Deneubourg, J-L., Franks, N.R., Sneyd, J., (2003). Age and aerobic performance of deer mice. Theraulaz, G., and Bonabeau, E. (2001). Self-organization Journal of Experimental Biology 206 , 1221–31. in biological systems . Princeton University Press, Chappell, M.A., Garland, T.Jr, Robertson, G.F., and Princeton, NJ. Saltzman, W. (2007). Relationships among running per- Campbell, T.W. (1995). Avian hematology and cytology . Iowa formance, aerobic physiology and organ mass in male State University Press, Ames. Mongolian gerbils. Journal of Experimental Biology 210 , Campbell, I.T. and Donaldson, J. (1981). Energy require- 4179–97. ments of Antarctic sledge dogs. British Journal of Nutrition Charles, J.D. and Bejan, A. (2009). The evolution of speed, 45 , 95–98. size and shape in modern athletics. Journal of Experimental Camphuysen, C.J., Ens, B.J., Heg, D., Hulscher, J.B., van Biology 212 , 2419–25. der Meer, J., and Smit, C.J. (1996). Oystercatcher Charmantier, A., McCleery, R.H., Cole, L.R., Perrins, C., Haematopus ostralegus winter mortality in The Kruuk, L.E.B., and Sheldon, B.C. (2008). Adaptive phe- Netherlands: the effect of severe weather and food sup- notypic plasticity in response to climate change in a ply. Ardea 84A , 469–92. wild bird population. Science 320 , 800–803. Camphuysen, C.J., Berrevoets, C.M., Cremers, H.J.W.M., Charnov, E.L. (1976a). Optimal foraging: The marginal Dekinga, A., Dekker, R., Ens, B.J., van der Have, T.M., value theorem. Theoretical Population Biology 9 , 129–36. Kats, R.K.H., Kuiken, T., Leopold, M.F., van der Meer, J., Charnov, E.L. (1976b). Optimal foraging: attack strategy of and Piersma, T. (2002). Mass mortality of Common a mantid. American Naturalist 110 , 141–51. Eiders (Somateria mollissima ) in the Dutch Wadden Sea, Charnov, E.L. (1993). Life history invariants: Some explora- winter 1999/2000: starvation in a commercially tions of symmetry in evolutionary ecology . Oxford exploited wetland of international importance. Biological University Press, Oxford. Conservation 106 , 303–17. Charnov, E.L. and Orians, G.H. (1973). Optimal foraging: Cannon, B. and Nedergaard, J. (2004). Brown adipose tis- some theoretical explorations. Unpublished manuscript, sue: Function and physiological signifi cance. http://hdl.handle.net/1928/1649 (last accessed 15 June Physiological Reviews 84 , 277–359. 2010). Caramujo, M-J. and Boavida, M.J. (2000). Induction and Cherel, Y., Robin, J-P., and Le Maho, Y. (1988). Physiology cost of tail spine elongation in Daphnia hyalina x galeata : and biochemistry of long-term fasting in birds. Canadian Reduction of susceptibility to copepod predation. Journal of Zoology 66 , 159–66. Freshwater Biology 45 , 413–23. Chew, R.M. (1951). The water exchange of some small Carmi, N., Pinshow, B., Porter, W.P., and Jaeger, J. (1992). mammals. Ecological Monographs 21 , 215–25. Water and energy limitations on fl ight duration in small Chew, R.M. (1961). Water metabolism of desert-inhabiting migrating birds. Auk 109 , 268–76. vertebrates. Biological Reviews 36 , 1–31. Carroll, L. (1960). The annotated Alice: Alice’s adventures in Clark, C.W. (1994). Antipredation and behavior and the Wonderland and through the looking-glass . New American asset-protection principle. Behavioral Ecology 5 , 159–70. Library, New York. Cohen, A.A., McGraw, K.J., Wiersma, P., Williams, J.B., Cartar, R.V. and Morrison, R.I.G. (2005). Metabolic corre- Robinson, W.D., Robinson, T.R., Brawn, J.D., and lates of leg length in breeding arctic shorebirds: The cost Ricklefs, R.E. (2008). Interspecifi c associations between of getting high. Journal of Biogeography 32 , 377–82. circulating antioxidant levels and life-history variation Castro, G., Myers, J.P., and Ricklefs, R.E. (1992). Ecology in birds. American Naturalist 172 , 178–93. and energetics of sanderlings migrating to four lati- Constanzo, K. and Monteiro, A. (2007). The use of chemi- tudes. Ecology 73 , 833–44. cal and visual cues in female choice in the butterfl y Chambers, W.H. (1952). Max Rubner (June 2, 1854–April Bicyclus anyana . Pr oceedings of the Royal Society, London, B 27, 1932). Journal of Nutrition 48 , 1–12. 274 , 845–51. 190 REFERENCES

Convertino, V.A. (1997). Cardiovascular consequences of Cronin, H. (1991). The ant and the peacock: Altruism and bed rest: effect on maximal oxygen uptake. Medicine and sexual selection from Darwin to today . Cambridge Science in Sports and Exercise 29 , 191–96. University Press, Cambridge. Cook, S.A. and Johnson, M.P. (1968). Adaptation to heter- Croze, H. (1970). Searching image in carrion crows: hunt- ogenous environments. I. Variation in heterophylly in ing strategy in a predator and some anti-predator Ranunculus fl ammula L. Evolution 22 , 496–516. devices in camoufl aged prey. Zeitschrift für Tierpsychologie Cooper, S.J. (1999). The thermal and energetic signifi cance Suppl. 5, 1–86. of cavity roosting in mountain chickadees and juniper Crucian, B.E., Stowe, R.P., Pierson, D.L., and Sams, C.F. titmice. Condor 101 , 863–66. (2008). Immune system dysregulation following short- Cooper, H.M., Herbin, M., and Nevo, E. (1993). Visual sys- vs long-duration spacefl ight. Aviation Space and tem of a naturally microphthalmic mammal: the blind Environmental Medicine 79 , 835–43. mole rat, Spalax eherenbergi . Journal of Comparative Currey, J.D. ed. (1984). The mechanical adaptations of bones . Neurology 328 , 313–50. Princeton University Press, Princeton, NJ. Cooper, W.E., Perez-Mellado, V., and Vitt, L.J. (2004). Ease Cuthill, I. and Kacelnik, A. (1990). Central place foraging: and effectiveness of costly autotomy vary with preda- a reappraisal of the ‘loading effect’. Animal Behaviour 40 , tion intensity among lizard populations. Journal of 1087–1101. Zoology 262 , 243–55. Cuthill, I.C., Haccou, P., and Kacelnik, A. (1994). Starlings Cooper, C.B., Hochachka, W.M., Butcher, G., and Dhondt, (Sturnus vulgaris ) exploiting patches: response to long- A.A. (2005). Seasonal and latitudinal trends in clutch term changes in travel time. Behavioral Ecology 5 , 81–90. size: thermal constraints during laying and incubation. Daan, S. and Tinbergen, J.M. (1997). Adaptation in life his- Ecology 86 , 2018–31. tories. In Behavioural ecology: An evolutionary approach , Cornell, H. (1974). Parasitism and distributional gaps eds J.R. Krebs and N.B. Davies, 4th edition, pp. 311–33. between allopatric species. American Naturalist 108 , Blackwell Science, Oxford. 880–83. Daan, S., Dijkstra, C., and Tinbergen, J.M. (1990a). Family Corning, W.R. and Biewener, A.A. (1998). In vivo strains in planning in the kestrel (Falco tinnunculus ): the ultimate pigeon fl ight feather shafts: Implications for structural control of covariation of laying date and clutch size. design. Journal of Experimental Biology 201 , 3057–66. Behaviour 114 , 83–116. Costantini, D. (2008). Oxidative stress in ecology and evo- Daan, S., Masman, D., and Groenewold, A. (1990b). Avian lution: lessons from avian studies. Ecology Letters 11 , basal metabolic rates: their association with body com- 1238–51. position and energy expenditure in nature. American Costantini, D., Dell’Ariccia, G., and Lipp, H-P. (2008). Journal of Physiology 259 , R333–40. Long fl ights and age affect oxidative status of homing Daan, S., Deeerenberg, C., and Dijkstra, C. (1996). Increased pigeons (Columba livia ). Journal of Experimental Biology daily work precipitates death in the kestrel. Journal of 211 , 377–81. Animal Ecology 65 , 539–44. Cott, H.B. (1957). Adaptive coloration in animals . Methuen, Dall, S.R.X., Giraldeau, L-A., Olsson, O., McNamara, J.M., London. and Stephens, D.W. (2005). Information and its use by Cox, D.R. (1972). Regression models and life tables. Journal animals in evolutionary ecology. Trends in Ecology and of the Royal Statistical Society, B 34 , 187–220. Evolution 20 , 187–93. Coyne, J.A. (2009a). Why evolution is true. Viking, Daly, M. and Wilson, M. (1983) . Sex, evolution and behavior , New York. 2nd edition. Wadsworth, Belmont, CA. Coyne, J.A. (2009b). Evolution’s challenge to genetics. Danchin, É., Giraldeau, L-A., Valone, T.J., and Wagner, Nature 457 , 382–83. R.H. (2004). Public information: From nosy neighbors to Creel, S., Christianson, D., Liley, S., and Winnie, J.A. (2007). cultural evolution. Science 305 , 487–91. Predation risk affects reproductive physiology and Danchin, É., Giraldeau, L-A., and Cézilly, F. (2008). demography of elk. Science 315 , 960. Behavioural ecology . Oxford University Press, Oxford. Cresswell, W. (1994). Age-dependent choice of redshank Danks, H.V. (1999). Life cycles in polar arthropods - fl exi- (Tringa totanus ) feeding location: Profi tability or risk? ble or programmed? European Journal of Entomology 96 , Journal of Animal Ecology 63 , 589–600. 83–103. Crispo, E. (2007). The Baldwin effect and genetic assimila- Darveau, C-A., Suarez, R.K., Andrews, R.D., and tion: Revisiting two mechanisms of evolutionary change Hochachka, P.W. (2002). Allometric cascade as a unify- mediated by phenotypic plasticity. Evolution 61 , ing principle of body mass effects on metabolism. Nature 2469–79. 417 , 166–70. REFERENCES 191

Darveau, C-A., Suarez, R.K., Andrews, R.D., and Bay, British Columbia. Canadian Field-Naturalist 112 , Hochachka, P.W. (2003). Allometric cascades: reply to 694–97. West et al ., and Banavar et al . Nature 421 , 714. Dekker, D. and Ydenberg, R. (2004). Raptor predation on Darwin, C. (1854). The balanidae (or sessile cirripedes); the wintering dunlins in relation to the tidal cycle. Condor verrucidae, etc: A monograph of the sub-class cirripedia, with 106 , 415–19. fi gures on all the species . The Ray Society, London. de Nie, G.J. (2002). Schrikrui bij wespendieven Pernis Darwin, C. (1859). The origin of species. John Murray, apivorus . De Takkeling 10 , 107–16. London. Dennett, D.C. (1995). Darwin´s dangerous idea: Evolution and Dasgupta, S., Singh, R.P., and Kafatos, M. (2009). the meanings of life . Allen Lane/Penguin Press, London. Comparison of global chlorophyll concentrations using Denny, M.W. (2008). Limits to running speeds in dogs, MODIS data. Advances in Space Research 43 , 1090–1100. horses and humans. Journal of Experimental Biology 211 , Davis, M.S., Willard, M.D., Williamson, K.K., Steiner, J.M., 3836–49. and Williams, D.A. (2005). Sustained strenuous exercise DeWitt, T.J. and Scheiner, S.M. eds (2004) Phenotypic plas- increases intestinal permeability in racing Alaskan sled ticity: Functional and conceptual approaches . Oxford dogs. Journal of Veterinary Internal Medicine 19 , 34–39. University Press, New York. Davis, M.S., Willard, M.D., Williamson, K.K., Royer, C., DeWitt, T.J., Sih, A., and Wilson, D.S. (1998). Costs and Payton, M., Steiner, J.M., Hinchcliff, K., McKenzie, E., limits of phenotypic fl exibility. Trends in Ecology and and Nelson, S. Jr (2006). Temporal relationship between Evolution 13 , 77–81. gastrointestinal protein loss, gastric ulceration or ero- Dial, K.P., Greene, E., and Irschick, D.J. (2008). Allometry sion, and strenuous exercise in racing Alaskan sled of behaviour. Trends in Ecology and Evolution 23 , dogs. Journal of Veterinary Internal Medicine 20 , 835–39. 394–401. Dawkins, R. (1995). River out of Eden: A Darwinian view of Diamond, J.M.(1993). Evolutionary physiology. In The life . Weidenfeld and Nicolson, London. logic of life: The challenge of integrative physiology , eds Dawkins, R. (1996). Climbing Mount Improbable . Viking, C.A.R. Boyd and D. Noble, pp. 89–111. Oxford University London. Press, Oxford. Dawkins, R. (2005). The ancestor’s tale: A pilgrimage to the Diamond, J.M. (1998). Evolution of biological safety fac- dawn of evolution . Marine Books, New York. tors: a cost/benefi t analysis. InPrinciples of animal design: Dawson, W.R. and Bartholomew, G.A. (1968). Temperature The optimization and symmorphosis debate, eds E.R. Weibel, regulation and water economy of desert birds. In Desert C.R. Taylor, and L. Bolis, pp. 21–27. Cambridge biology, ed. G.W. Brown, pp. 357–94. Academic Press, University Press, Cambridge. New York. Diamond, J.M. (2002). Quantitative evolutionary design. Dawson, W.R. and Marsh, R.L. (1989). Metabolic adapta- Journal of Physiology 542 , 337–45. tions to cold and season in birds. In Physiology of cold Diamond, J.M. and Hammond, K.A. (1992). The matches, adaptation in birds , eds C. Bech and R.E. Reinertsen, achieved by natural selection, between biological capac- pp. 83–94. Plenum Press, New York. ities and their natural loads. Experientia 48 , 551–57. Dawson, W.R., Carey, C., Adkisson, C.S., and Ohmart, Dierschke, V. (2003). Predation hazard during migratory R.D. (1979). Responses of Brewer’s and chipping spar- stopover: are light or heavy birds under risk? Journal of rows to water restriction. Physiological Zoology 42 , Avian Biology 34 , 24–29. 529–41. Dietz, M.W. and Piersma, T. (2007). Red knots give up Dayton, G.H., Saenz, D., Baum, K.A., Langerhans, R.B., fl ight capacity and defend food processing capacity and DeWitt, T.J. (2005). Body shape, burst speed and during winter starvation. Functional Ecology 21, escape behavior of larval anurans. Oikos 111 , 582–91. 899–904. Deerenberg, C., Pen, I., Dijkstra, C., Arkies, B.J., Visser, Dietz, M.W., Dekinga, A., Piersma, T., and Verhulst, S. G.H., and Daan, S. (1995). Parental energy expenditure (1999). Estimating organ size in small migrating shore- in relation to manipulated brood size in the European birds with ultrasonography: an intercalibration exercise. kestrel Falco tinnunculus . Zoology 99 , 39–48. Physiological and Biochemical Zoology 72 , 28–37. Dekinga, A., Dietz, M.W., Koolhaas, A., and Piersma, T. Dietz, M.W., Piersma, T., Hedenström, A., and Brugge, M. (2001). Time course and reversibility of changes in the (2007). Intraspecifi c variation in avian pectoral muscle gizzards of red knots alternately eating hard and soft mass: Constraints on maintaining manoeuvrability with food. Journal of Experimental Biology 204 , 2167–73. increasing body mass. Functional Ecology 21, 317–26. Dekker, D. (1998). Over-ocean fl ocking by dunlins, Calidris Dietz, M.W., Spaans, B., Dekinga, A., Klaassen, M., alpina, and the effect of raptor predation at Boundary Korthals, H., van Leeuwen, C., and Piersma, T. (2010). 192 REFERENCES

Do red knots ( Calidris canutus islandica) routinely skip Dudley, R. and Gans, C. (1991). A critique of symmorpho- Iceland during southward migration? Condor 112, sis and optimality models in physiology. Physiological 48–55. Zoology 64 , 627–37. DiMagno, E.P., Go, G.L.W., and Summerskill, W.H.J. Duijns, S., van Dijk, J.G.B., Spaans, B., Jukema, J., de Boer, (1973). Relations between pancreatic enzyme outputs W.F., and Piersma, T. (2009). Foraging site selection of and malabsorption in severe pancreatic insuffi ciency. two subspecies of bar-tailed godwit Limosa lapponica : New England Journal of Medicine 288 , 813–15. Time minimizers accept greater predation danger than di Pamprero, P.E. (1985). Metabolic and circulatory limita- energy minimizers. Ardea 97 , 51–59.

tions to VO2max at the whole animal level. Journal of Dykhuizen, D. (1978). Selection for tryptophan auxotrophs Experimental Biology 115 , 319–31. of Escherichia coli in glucose-limited chemostat as a test Dobson, A. (1995). The ecology and epidemiology of rind- of the energy conservation hypothesis of evolution. erpest virus in Serengeti and Ngorongoro conservation Evolution 32 , 627–37. area. In Serengeti II: dynamics, management, and conserva- Dzialowski, E.M. (2005). Use of operative temperature tion of an ecosystem , eds A.R.E. Sinclair and P. Arcese, and standard operative temperature models in thermal pp. 485–505. University of Chicago Press, Chicago. biology. Journal of Thermal Biology 30 , 317–34. Dobzhansky, T. (1973). Nothing in biology makes sense Eccard, J.A. and Liesenjohann, T. (2008). Foraging deci- except in the light of evolution. American Biology Teacher sions in risk-uniform landscapes. Public Library of Science 35 , 125–29. One 3 , e3438. Doherty, P.A., Wassersug, R.J., and Lee, J.M. (1998). Eckert, R., Randall, D., and Augustine, G. (1988). Animal Mechnical properties of the tadpole tail fi n. Journal of physiology: Mechanisms and Adaptations . W.H. Freeman, Experimental Biology 201 , 2691–99. New York. Dorfman, T.A., Levine, B.D., Tillery, T., Peschock, R.M., Edelman, G.M. (2004). Wider than the sky. The phenomenal Hastings, J.L., Schneider, S.M., Macias, B.R., Biolo, G., gift of consciousness . Yale University Press, Newhaven. and Hargens, A.R. (2007). Cardiac atrophy in women Edgell, T.C., Lynch, B.R., Trussell, G.C., and Palmer, A.R. following bed rest. Journal of Applied Physiology 103 , (2009). Experimental evidence for the rapid evolution of 8–16. behavioral canalization in natural populations. American Dowling, D.K. and Simmons, L.W. (2009). Reactive oxy- Naturalist 174 , 434–40. gen species as universal constraints in life-history evo- Eisma, D. (1998). Intertidal deposits: River mouths, tidal fl ats, lution. Proceedings of the Royal Society, London, B 276 , and coastal lagoons . CRC Press, Boca Raton. 1737–45. Elgar, M.A. and Harvey, P.H. (1987). Basal metabolic rates Drent, J. (2004). Life history variation of a marine bivalve in mammals: allometry, phylogeny and ecology. ( Macoma balthica) in a changing world. PhD thesis, Functional Ecology 1 , 25–36. University of Groningen, The Netherlands. Elser, J.J. and Hamilton, A. (2007). Stoichiometry and the Drent, R.H. and Daan, S. (1980). The prudent parent: new biology: the future is now. Public Library of Science — Energetic adjustments in avian breeding. Ardea 68 , Biology 5 , 1403–1405. 225–52. Elser, J.J., Dobberfuhl, D.R., MacKay, N.A., and Schampel, Drent, R. and Piersma, T. (1990). An exploration of the J.H. (1996). Organism size, life history, and N:P stoichi- energetics of leap-frog migration in arctic breeding ometry. Bioscience 46 , 674–84. waders. In Bird migration: Physiology and Ecophysiology, Emlen, D.J. (2008). The evolution of animal weapons. ed. E. Gwinner, pp. 399–412. Springer-Verlag, Berlin. Annual Review of Ecology, Evolution, and Systematics 39 , Drent, R., Ebbinge, B., and Weijand, B. (1978). Balancing 387–413. the energy budgets of arctic-breeding geese throughout Emlen, D.J., Lavine, L.C., and Ewen-Campen, B. (2007). the annual cycle: a progress report. Verhandlungen der On the origin and evolutionary diversifi cation of beetle Ornithologischen Gesellschaft in Bayern 23 , 239–64. horns. Proceedings of the National Academy of Sciences, Drent, R., Both, C., Green, M., Madsen, J., and Piersma, T. USA 104 (Suppl. 1), 8661–68. (2003). Pay-offs and penalties of competing migratory Endler, J.A. (1986). Natural selection in the wild . Princeton schedules. Oikos 103 , 274–92. University Press, Princeton. Du Bois, E.F. (1936). Basal metabolism in health and disease . Engel, S., Biebach, H., and Visser, G.H. (2006). Metabolic Lea & Febiger, Philadelphia. costs of avian fl ight in relation to fl ight velocity: a Duckworth, R.A. (2009). The role of behavior in evolution: study in rose coloured starlings (Sturnus roseus , a search for mechanism. Evolutionary Ecology 23 , Linnaeus). Journal of Comparative Physiology B 176 , 513–31. 415–27. REFERENCES 193

Estourgie-van Burk, G.F., Bartels, M., Hoekstra, R.A., Furness, L.J. and Speakman, J.R. (2008). Energetics and Polderman, T.J.C., Delemarre-van der Wal, H.A., and longevity in birds. Age 30 , 75–87. Boomsma, D.I. (2009). A twin study of cognitive costs of Gaffney, B. and Cunningham, E.P. (1988). Estimation of low birth weight and catch-up growth. Journal of genetic trend in racing performance of thoroughbred Pediatrics 154 , 29–32. horses. Nature 332 , 722–24. Feder, M.E. and Watt, W.B. (1992). Functional biology of Galilei, G. (1637). Dialogues concerning two new sciences adaptation. In Genes in ecology, eds R.J. Berry, T.J. (translated by H. Crew and A. De Salvio). Macmillan, Crawford, and G.M. Hewitt, pp. 365–92. Blackwell New York, 1914. Scientifi c Publications, Oxford. Garland, T. Jr (1998). Conceptual and methodological Fiennes, R. (2003). Captain Scott. Hodder and Stoughton, issues in testing predictions of symmorphosis. In London. Principles of animal design: The optimization and symmor- Figuerola, J. (1999). Effects of salinity on rates of infestation phosis debate, eds E.R. Weibel, C.R. Taylor, and L. Bolis, of waterbirds by haematozoa. Ecography 22 , 681–85. pp. 40–47. Cambridge University Press, Cambridge. Fisher, M.O., Nager, R.G., and Monaghan, P. (2006). Garland, T. Jr and Huey, R.B. (1987). Testing symmorpho- Compensatory growth impairs adult cognitive function. sis: does structure match functional requirements? Public Library of Science —Biology 4 , 1462–66. Evolution 41 , 1404–1409. Fitts, R.H., Riley, D.R., and Widrick, J.J. (2001) Functional Garland, T. Jr. and Kelly, S.A. (2006). Phenotypic plasticity and structural adaptations of skeletal muscle to micro- and experimental evolution. Journal of Experimental gravity. Journal of Experimental Biology 204 , 3201–3208. Biology 209 , 2344–61. Forbes-Ewan, C.H., Morrissey, B.L.L., Gregg, G.C., and Geen, R.G. (1995). Human motivation: A social psychological Waters, D.R. (1989). Use of the doubly labelled water approach . Wadsworth Publishing, New York. technique in soldiers training for jungle warfare. Journal Geist, V. (1998). Deer of the world. Stackpole Books, of Applied Physiology 67 , 14–18. Mechanicsburg, PA. Ford, H. and Crowther, S. (1922). My life and work . Gerth, N., Sum, S., Jackson, S., and Starck, J.M. (2009). Electronic edition: www.gutenberg.org/etext/7213 (last Muscle plasticity of Inuit sled dogs in Greenland. Journal accessed 16 June 2010). of Experimental Biology 212 , 1131–39. Foster, C., Hoyos, J. Earnest, C., and Lucia, A. (2005). Gervasi, S.S. and Foufopoulos, J. (2008). Costs of plasticity: Regulation of energy expenditure during prolonged responses to dessication decrease post-metamorphic athletic competition. Medicine and Science in Sports and immune function in a pond-breeding amphibian. Exercise 37 , 670–75. Functional Ecology 22 , 100–108. Freckleton, R.P., Harvey, P.H., and Pagel, M. (2003). Ghalambor, C.K., McKay, C.K., Carroll, S.P., and Reznick, Bergmann’s rule and body size in mammals. American D.N. (2007). Adaptive versus non-adaptive plasticity Naturalist 161 , 821–25. and the potential for contemporary adaptation in new Fretwell, S.D. and Lucas, H.L. (1970). On territorial behav- environments. Functional Ecology 21 , 394–407. iour and other factors infl uencing habitat distribution in Gibbon, J. (1977). Scalar expectancy theory and Weber’s birds. Acta Biotheoretica 19 , 16–36. law in animal timing. Psychological Review 84 , Froehle, A.W. and Schoeninger, M.J. (2006). Intraspecies 279–325. variation in BMR does not affect estimates of early hom- Gienapp, P., Teplitsky, C., Alho, J.S., Mills, J.A., and inin total daily energy expenditure. American Journal of Merilä, J. (2008). Climate change and evolution: disen- Physical Anthropology 131 , 552–59. tangling environmental and genetic responses. Molecular Fry, C.H., Ferguson, I.J., and Dowsett, R.J. (1972). Flight Ecology 17 , 167–78. muscle hypertrophy and ecophysiological variation of Gilbert, S.F. and Epel, D. (2009). Ecological developmental yellow wagtail Motacilla fl ava races at Lake Chad. Journal biology: Integrating epigenetics, medicine, and ecolution . of Zoology 167 , 293–306. Sinauer, Sunderland, Mass. Fuglesteg, B.N, Haga, Ø.E., Folkow, L.P., Fuglei, E., and Gill, J., Sutherland, W.J., Watkinson, A.R. (1996). A method Schytte Blix, A. (2006). Seasonal variations in basal met- to quantify the effects of human disturbance on animal abolic rate, lower critical temperature and responses to populations. Journal of Applied Ecology 33 , 786–92. temporary starvation in the arctic fox ( Alopex lagopus ) Gill, R.E. Jr, Piersma, T., Hufford, G., Servranckx, R., and from Svalbard. Polar Biology 29 , 308–19. Riegen, A. (2005). Crossing the ultimate ecological bar- Fuller, A., Carter, R.N., and Mitchell, D. (1998). Brain and rier: evidence for an 11 000-km-long nonstop fl ight from abdominal temperatures at fatigue in rats exercising in Alaska to New Zealand and eastern Australia by bar- the heat. Journal of Applied Physiology 84 , 877–83. tailed godwits. Condor 107 , 1–20. 194 REFERENCES

Gill, R.E. Jr, Tibbitts, T.L., Douglas, D.C., Mulcahy, D.M., Gould, S.J. (1977). Ontogeny and phylogeny . Harvard Handel, C.M., Gottschalck, J.C., Warnock, N., McCaffery, University Press, Cambridge, Mass. B.J., Battley, P.F., and Piersma, T. (2009). Extreme endur- Gould, S.J. and Lewontin, R.C. (1979). The spandrels of ance fl ights by landbirds crossing the Pacifi c Ocean: San Marcos and the Panglossian paradigm: a critique of Ecological corridor rather than barrier? Proceedings of the the adaptationist program. Proceedings of the Royal Royal Society, London, B 276 , 447–57. Society,London, B 205 , 581–98. Gillooly, J.F. and Allen, A.O. (2007). Changes in body tem- Grant, P.R. (1986). Ecology and evolution of Darwin’s fi nches .

perature infl uence the scaling of VO2max and aerobic Princeton University Press, Princeton, NJ. scope in mammals. Biology Letters 3 , 99–102. Grant, P.R. and Grant, B.R. (2008). How and why species Gillooly, J.F., Brown, J.H., West, G.B., Savage, V.M., and multiply: The radiation of Darwin’s fi nches . Princeton Charnov, E.L. (2001). Effects of size and temperature on University Press, Princeton, NJ. metabolic rate. Science 293 , 2248–51. Graveland, J., van der Wal, R., van Balen, J.H., and van Gingerich, P.D. (1993). Quantifi cation and comparison of Noordwijk, A.J. (1994). Poor reproduction in forest pas- evolutionary rates. American Journal of Science 293A , serines from decline of snail abundance on acidifi ed 453–78. soils. Nature 368 , 446–48. Gitschier, J. (2009). You say you want a revolution: an Green, R.F. (1984). Stopping rules for optimal foragers. interview with Pat Brown. Public Library of Science - American Naturalist 123 , 30–43. Genetics 4 , e1000560. Green, R.F. (2006). A simpler, more general method of Gloutney, M.L. and Clark, R.G. (1997). Nest site selection fi nding the optimal foraging strategy for Bayesian birds. by mallards and blue winged teal in relation to microcli- Oikos 112 , 274–84. mate. Auk 114 , 381–95. Green, M., Alerstam, T., Clausen, P., Drent, R., and Ebbinge, Gluckman, P.D., Hanson, M.A., Bateson, P., Beedle, A.S., B.S. (2002). Dark-bellied brent geese Branta bernicla ber- Law, C.M., Bhutta, Z.A., Anokhin, K.V., Bougnères, P., nicla, as recorded by satellite telemetry, do not minimize Chandak, G.T., Dasgupta, P., Davey Smith, G., Ellison, fl ight distance during spring migration. Ibis 144 , P.T., Forrester, T.E., Gilbert, S.F., Jablonka, E., Kaplan, 106–21. H., Prentice, A.M., Simpson, S.J., Uauy, R., and West- Greene, H.W. (1997). Snakes: The evolution of mystery in Eberhard, M.J. (2009). Towards a new developmental nature . University of California Press, Berkeley. synthesis: adaptive developmental plasticity and human Grimshaw, H.M., Ovington, J.D., Betts, M.M., and Gibb, disease. Lancet 373 , 1654–57. J.A. (1958). The mineral content of birds and insects in González-Alonso, J., Teller, C., Amdersen, S.L., Jensen, plantations of Pinus sylvestris L. Oikos 9 , 26–34. F.B., Hyldig, T and Nielsen, B. (1999). Infl uence of body Groothuis, T.G.G., Muller, W., von Engelhardt, N., Carere, temperature on the development of fatigue during pro- C., and Eising, C. (2005). Maternal hormones as a tool to longed exercise in the heat. Journal of Applied Physiology adjust offspring phenotype in avian species. Neuroscience 86 , 1032–39. and Biobehavioral Reviews 29 , 329–52. Gordon, J.E. (1978). Structures, or why things don’t fall down . Groscolas, R. (1986). Changes in body mass, body tem- Plenum Press, New York. perature and plasma fuel levels during the natural Gordon, M.S. (1998). Evolution of optimal systems: overview. breeding fast in male and female emperor penguins In Principles of animal design: The optimization and symmor- Aptenodytes forsteri . Journal of Comparative Physiology B phosis debate , eds E.R. Weibel, C.R. Taylor, and L. Bolis, 156 , 521–27. pp. 37–39. Cambridge University Press, Cambridge. Gudmundsson, G.A., Lindström, Å., and Alerstam, T. Gosler, A.G. (1987). Pattern and process in the bill mor- (1991). Optimal fat loads and long-distance fl ights by phology of the great tit Parus major . Ibis 129 , 451–76. migrating knots, Calidris canutus, sanderlings, C. alba , Goss, R.J. (1983). Deer antlers: Regeneration, function, and and turnstones, Arenaria interpres . Ibis 133 , 140–52. evolution . Academic Press, New York. Guezennec, C.Y., Satabin, P., Legrand, H., and Bigard, A.X. Goss, R.J. (1995). Future-directions in antler research. (1994). Physical performance and metabolic changes Anatomical Record 241 , 291–302. induced by combined prolonged exercise and different Gotanda, K., Turgeon, K., and Kramer, D. (2009). Body size energy intakes in humans. European Journal of Applied and reserve protection affect fl ight initiation distance in Physiology 68 , 525–30. parrotfi sh. Behavioral Ecology and Sociobiology 63 , 1563–72. Guglielmo, C.G. and Williams, T.D. (2003). Phenotypic Gottlieb, G. (1992). Individual development and evolution: fl exibility of body composition in relation to migratory The genesis of novel behavior. Oxford University Press, state, age, and sex in the western sandpiper ( Calidris New York. mauri ). Physiological and Biochemical Zoology 76 , 84–98. REFERENCES 195

Guglielmo, C.G., Piersma, T., and Williams, T.D. (2001). A Harwood, J. (1996). Weimar culture and biological theory: sport-physiological perspective on bird migration: a study of Richard Woltereck (1877–1944). History of Evidence for fl ight-induced muscle damage. Journal of Science 34 , 347–77. Experimental Biology 204 , 2683–90. Hatch, D.E. (1970). Energy conserving and heat dissipat- Gwinner, E. (1986). Circannual rhythms: Endogenous annual ing mechanisms of Turkey vulture. Auk 87 , 111–24. clocks in the organization of seasonal processes . Springer- Haugen, M., Williams, J.B., Wertz, P., and Tieleman, B.I. Verlag, Berlin. (2003a). Lipids of the stratum corneum vary with cuta- Hacklander, K., Tataruch, F., and Ruf, T. (2002). The effect neous water loss among larks along a temperature- of dietary fat content on lactation energetics in the moisture gradient. Physiological and Biochemical Zoology European hare (Lepus europaeus ). Physiological and 76 , 907–17. Biochemical Zoology 75 , 19–28. Haugen, M., Tieleman, B.I., and Williams, J.B. (2003b). Hails, C.J. (1983). The metabolic rate of tropical birds. Phenotypic fl exibility in cutaneous water loss and lipids Condor 85 , 61–65. of the stratum corneum. Journal of Experimental Biology Haldane, J.B.S. (1949). Suggestions as to quantitative 206 , 3581–88. measurement of rates of evolution. Evolution 3 , 51–56. Hawking, S. (1988). A brief history of time: From the big bang Haldane, J.B.S. (1957). The cost of natural selection. Journal to black holes . Bantam Books, London. of Genetics 55 , 511–24. Hayes, J.P. and Garland, T. Jr. (1995). The evolution of Hammill, E., Rogers, A., and Beckerman, A.P. (2008). endothermy: testing the aerobic capacity model. Costs, benefi ts and the evolution of inducible defences: Evolution 49 , 836–47. a case study with Daphnia pulex . Journal of Evolutionary Hayes, J.P., Bible, C.A., and Boone, J.D. (1998). Repeatability Biology 21 , 705–15. of mammalian physiology: evaporative water loss and Hammond, K.A. and Diamond, J.M. (1992). An experi- oxygen consumption of Dipodomys merriami . Journal of mental test for a ceiling on sustained metabolic rate in Mammalogy 79 , 475–85. lactating mice. Physiological Zoology 65 , 952–77. Hedenström, A. and Alerstam, T. (1995). Optimal fl ight Hammond, K.A. and Diamond, J.M. (1994). Limits to die- speeds of birds. Philosophical Transactions of the Royal tary nutrient intake and intestinal nutrient uptake in Society, London, B 348 , 471–87. lactating mice. Physiological Zoology 67 , 282–303. Heinrich, B. (1997). The trees in my forest . HarperCollins, Hammond, K.A. and Diamond, J.M. (1997). Maximal sus- New York. tained energy budgets in humans and animals. Nature Heinrich, B. (2001). Racing the antelope: What animals can 386 , 457–62.1 teach us about running and life. HarperCollins, New York. Hammond, K.A. and Kristan, D.M. (2000). Responses to Hendry, A.P. and Kinnison, M.T. (1999). The pace of mod- lactation and cold exposure by deer mice ( Peromyscus ern life: measuring of contemporary microevolution. maniculatus ). Physiological and Biochemical Zoology 73 , Evolution 53 , 1637–53. 547–56. Hendry, A.P., Farrugia, T.J., and Kinnison, M.T. (2008). Hammond, K.A., Konarzewski, M., Torres, R., and Human infl uences on rates of phenotypic change in Diamond, J.M. (1994). Metabolic ceilings under a combi- wild animal populations. Molecular Ecology 17 , 20–29. nation of peak energy demands. Physiological Zoology 67 , Henry, C.J.K. (2005). Basal metabolic rate studies in 1479–1506. humans: Measurement and development of new equa- Hammond, K.A., Lam, M., Lloyd, K.C.K., and Diamond, tions. Public Health Nutrition 8 , 1133–52. J.M. (1996a). Simultaneous manipulation of intestinal Hessen, D.O. and Anderson, T.R. (2008). Excess carbon in capacities and nutrient loads in mice. American Journal of aquatic organisms and ecosystems: physiological, eco- Physiology 271 , G969–79. logical and evolutionary implications. Limnology and Hammond, K.A., Lloyd, K.C.K and Diamond, J.M. (1996b). Oceanography , 53 , 1685–96. Is mammary output capacity limiting to lactational per- Hiddink, J.G. (2003). Effects of suction-dredging for cock- formance in mice? Journal of Experimental Biology 199 , les on non-target fauna in the Wadden Sea. Journal of Sea 337–49. Research 50 , 315–23. Hammond, K.A., Chappell, M.A., Cardullo, R.A., Lin, Higashi, M., Abe, T., and Burns, T.P. (1992). Carbon- R-S., and Johnsen, T.S. (2000). The mechanistic basis of nitrogen balance and termite ecology. Proceedings of the aerobic performance in red junglefowl. Journal of Royal Society,London, B 249 , 303–308. Experimental Biology 203 , 2053–64. Hilborn, R. and Mangel, M. (1997). The ecological detective: Harman, D. (1956). Aging: a theory based on free radical Confronting models with data . Princeton University Press, radiation chemistry. Journal of Gerontology 11 , 298–300. Princeton, NJ. 196 REFERENCES

Hill, J.A. and Olson, E.N. (2008). Cardiac plasticity. New Holick, M.F. (1998). Perspectives on the impact of weight- England Journal of Medicine 358 , 1370–80. lessness on calcium and bone metabolism. Bone 22 , Hill, R.W., Wyse, G.A., and Anderson, M. (2004). Animal 105S–111S. physiology . Sinauer Associates, Sunderland, MA. Holick, M.F. (2000). Microgravity–induced bone loss— Hilton, G.M., Houston, D.C., Barton, N.W.H., and Furness, will it limit human space exploration? Lancet 355 , R.W. (1999). Digestion strategies of meat- and fi sh-eat- 1569–70. ing birds. In Proceedings of the 22nd International Holling, C.S. (1959). Some characteristics of simple types Ornithological Congress, eds N. J. Adams and R.J. Slotow, of predation and parasitism. Canadian Entomologist 91 , pp. 2184–97. BirdLife South Africa, Johannesburg. 385–98. Hinchcliff, K.W., Reinhart, G.A., Burr, J.R., Schreier, C.J., Holloszy, J.O. and Smith, E.K. (1986). Longevity of cold- and Swenson, R.A. (1997). Metabolizable energy intake exposed rats: A reevaluation of the ‘rate-of-living the- and sustained energy expenditure of Alaskan sled dogs ory’. Journal of Applied Physiology 61 , 1656–60. during heavy exertion in the cold. American Journal of Holmes, D.J. and Austad, S.N. (1995a). Birds as animal Veterinary Research 58 , 1457–62. models for the comparative biology of aging: A prospec- Hinchcliff, K.W., Constable, P.D., and DiSilvestro, R.A. tus. Journal of Gerontology A 50 , B59–B66. (2004). Muscle injury and antioxidant status in sled dogs Holmes, D.J. and Austad, S.N. (1995b). The evolution of competing in a long-distance sled dog race. Equine and avian senescence patterns: Implications for understanding Comparative Exercise Physiology 1 , 81–85. primary aging processes. American Zoologist 35, 307–17. Hinds, D.S., Baudinette, R.V., MacMillen, R.E., and Holmes, D.J. and Martin, K. (2009). A bird’s-eye view of Halpern, E.A. (1993). Maximum metabolism and the aging: What’s in it for ornithologists? Auk 126 , 1–23. aerobic factorial scope of endotherms. Journal of Holmgren, N.M.A. and Olsson, O. (2000). A three-neuron Experimental Biology 182 , 41–56. model of information processing during Bayesian for- Hirakawa, H. (1995). Diet optimization with a nutrient or aging. In Artifi cial neural networks in medicine and biology: toxin constraint. Theoretical Population Biology 47 , 331–46. Proceedings of the ANNIMAB-1 conference, Göteborg, Ho, C-K., Pennings, S.C., and Carefoot, T.H. (2010). Is diet Sweden, 13 – 16 May 2000, eds H. Malmgren, M. Borga, quality an overlooked mechanism for Bergmann’s rule? and L. Niklasson, pp. 265–70. Springer-Verlag, London. American Naturalist 175 , 269–76. Honkoop, P.J.C. and van der Meer, J. (1997). Reproductive Hoch, J.M. (2008). Variation in penis morphology and mat- output of Macoma balthica populations in relation to ing ability in the acorn barnacle, Semibalanus balanoides . winter-temperature and intertidal-height mediated Journal of Experimental Marine Biology and Ecology 359 , changes of body mass. Marine Ecology Progress Series 126–30. 149 , 155–62. Hochachka, P.W. and Beatty, C.L. (2003). Patterns of con- Hoppeler, H. and Flück, M. (2002). Normal mammalian trol of maximum metabolic rate in humans. Comparative skeletal muscle and its phenotypic plasticity. Journal of Biochemistry and Physiology A 136 , 215–25. Experimental Biology 205 , 2143–52. Hochachka, P.W., Beatty, C.L., Burelle, Y., Trump, M.E., Hoppeler, H., Altpeter, E., Wagner, M., Turner, D.L., McKenzie, D.C., and Matheson, G.O. (2002). The lactate Hokanson, J., König, M., Stalder-Navarro, V.P., and paradox in human high altitude physiological perform- Weibel, E.R. (1995). Cold acclimation and endurance ance. News in Physiological Sciences 17 , 122–26. training in guinea pigs: changes in lung, muscle and Hochachka, P.W., Darveau, C-A., Andrews, R.D., and brown fat tissue. Respiratory Physiology 101 , 189–98. Suarez, R.K. (2003). Allometric cascade: A model for Hoppema, M. and Goeyens, L. (1999). Redfi eld behavior resolving body mass effects on metabolism. Comparative of carbon, nitrogen and phosphorus depletions in Biochemistry and Physiology A 134 , 675–91. Antarctic surface water. Limnology and Oceanography 44 , Hodgson, W.R., McCutcheon, J., Byrd, S.K., Brown, W.S., 220–24. Bayly, W.M., Brengelmann, G.L., and Gollnick, P.D. Horak, P., Saks, L., Ots, I., and Kollist, H. (2002). (1993). Dissipation of metaboic heat in the horse during Repeatability of condition indices in captive green- exercise. Journal of Applied Physiology 74 , 1161–70. fi nches ( Carduelis chloris ). Canadian Journal of Zoology 80 , Hoffman, R.J. (1978). Environmental uncertainty and evo- 636–43. lution of physiological adaptation in Colias butterfl ies. Houston, A.I. (1995). Energetic constraints and foraging American Naturalist 112 , 999–1015. effi ciency. Behavioral Ecology 6 , 393–96. Hofman, M.A. and Swaab, D.F. (1992). Seasonal changes Houston, A.I., Stevens, M., and Cuthill, I.C. (2007). Animal in the suprachiasmatic nucleus of man. Neuroscience camoufl age: compromise or specialize in a 2 patch-type Letters 139 , 257–60. environment? Behavioral Ecology 18, 769–75. REFERENCES 197

Hoverman, J.T. and Relyea, R.A. (2007). How fl exible is Janeway, C.A., Travers, P., Walport, M., and Shlomchik, M. phenotypic plasticity? Developmental windows for trait (2004). Immunobiology: The immune system in health and induction and reversal. Ecology 88 , 693–705. disease , 6th edition. Garland Publishing, New York. Hoyt, D.F. and Taylor, C.R. (1981). Gait and the energetics Jarrett, J.N. (2008). Inter-population variation in shell mor- of locomotion in horses. Nature 292 , 239–40. phology of the barnacle Chthamalus fi ssus . Journal of Hoyt, R.W., Jones, T.E., Stein, T.P., McAnich, G.W., Crustacean Biology 28 , 16–20. Lieberman, H.R., Askew, E.W., and Cymerman, A. Jenni, L. and Jenni-Eiermann, S. (1998). Fuel supply and (1991). Doubly labelled water measurement of human metabolic constraints in migrating birds. Journal of Avian energy expenditure during strenuous exercise. Journal of Biology 29 , 521–28. Applied Physiology 71 , 16–22. Jenni-Eiermann, S., Jenni, L., Kvist, A., Lindström, Å., Hoyt, R.W., Opstad, P.K., Haugen, A–H., DeLany, J.P., Piersma, T., and Visser, G.H. (2002). Fuel use and meta- Cymerman, A., and Friedl, K.E. (2006). Negative energy bolic response to endurance exercise: a wind tunnel balance in male and female rangers: Effects of 7 days of study of a long-distance migrant shorebird. Journal of sustained exercise and food deprivation. American Experimental Biology 205 , 2453–60. Journal of Clinical Nutrition 83 , 1068–75. Jensma, G. (2009). Van professie naar traditie. Het ‘ljipaai- Hughes, M.R., Roberts, J.R., and Thomas, B.R. (1987). Total sykjen’ als uitgevonden traditie. In Friese sport: Tussen body water and its turnover in free-living nestling glau- traditie en professie, eds G. Jensma and P. Breuker, cous-winged gulls with a comparison of body water pp. 18–62. Bornmeer, Gorredijk. and water fl ux in avian species with and without salt Jeschke, J.M. (2007). When carnivores are ‘full and lazy’. glands. Physiological Zoology 60 , 481–91. Oecologia 152 , 357–64. Hulscher, J.B. (1985). Growth and abrasion of the oyster- Jeschke, J.M., Kopp, M., and Tollrian, R. (2002). Predator actcher bill in relation to dietary switches. Netherlands functional responses: discriminating between handling Journal of Zoology 35 , 124–54. and digesting prey. Ecological Monographs 72 , 95–112. Humphrey, N.K. (1983). Consciousness regained . Oxford Johannsen, W. (1911). The genotype conception of hered- University Press, Oxford. ity. American Naturalist 45 , 129–59. Hunte, W., Meyers, R.A., and Doyle, R.W. (1985). Bayesian Johnson, M.S. and Speakman, J.R. (2001). Limits to sus- mating decisions in an amphipod, Gammarus lawren- tained intake. V. Effect of cold-exposure during lactation cianus Bousfi eld. Animal Behaviour 33 , 366–72. in Mus musculus . Journal of Experimental Biology 204 , Huntford, R. (1985). The last place on Earth . Pan Books, 1967–77. London. Johnson, M.S., Thomson, S.C., and Speakman, J.R. (2001a). Iwasa, Y., Highashi, M., and Yamamura, N. (1981). Prey Limits to sustained intake. I. Lactation in the laboratory distribution as a factor determining the choice of opti- mouse Mus musculus . Journal of Experimental Biology 204 , mal foraging strategy. American Naturalist 117 , 710–23. 1925–35. Jablonka, E. (2006). Genes as followers in evolution—a Johnson, M.S., Thomson, S.C., and Speakman, J.R. (2001b). post-synthesis synthesis? Biology and Philosophy 21 , Limits to sustained intake. II. Interrelationships between 143–54. resting metabolic rate, life-history traits and morphol- Jablonka, E. and Lamb, M.J. (2005). Evolution in four dimen- ogy in Mus musculus . Journal of Experimental Biology 204 , sions: Genetic, epigenetic, behavioral and symbolic variation 1937–46. in the history of life . MIT Press, Cambridge, Mass. Johnson, M.S., Thomson, S.C., and Speakman, J.R. (2001c). Jablonka, E. and Lamb, M.J. (2007). Precis of Evolution in Limits to sustained intake. III. Effects of concurrent four dimensions . Behavioral and Brain Sciences 30 , pregnancy and lactation Mus musculus . Journal of 353–65. Experimental Biology 204 , 1947–56. Jablonka, E. and Raz, G. (2009). Transgenerational epige- Johnson, J.B., Burt, D.B., and DeWitt, T.J. (2008). Form, netic inheritance: prevalence, mechanisms, and implica- function, and fi tness: pathways to survival. Evolution 62 , tions for the study of heredity and evolution. Quarterly 1243–51. Review of Biology 84 , 131–76. Johnstone, A.M., Rance, K.A., Murison, S.D., Duncan, J.S., Jackson, S. and Diamond, J.M. (1996). Metabolic and diges- and Speakman, J.R. (2006). Additional anthropometric tive responses to artifi cial selection in chickens. Evolution measures may improve the predictability of basal meta- 50 , 1638–50. bolic rate in adult subjects. European Journal of Clinical Jacobs, J.D. and Wingfield, J.C. (2000). Endocrine control Nutrition 60 , 1437–44. of lifecycle stages: a constraint on response to the envi- Jones, J.H. (1998a). Symmorphosis and the mammalian res- ronment? Condor 102 , 35–51. piratory system: what is optimal design and does it exist? 198 REFERENCES

In Principles of animal design: The optimization and symmor- Ketterson, E.D. and Nolan, V. Jr (1999). Adaptation, exap- phosis debate, eds E.R. Weibel, C.R. Taylor, and L. Bolis, tation, and constraint: A hormonal perspective. American pp. 241–48. Cambridge University Press, Cambridge. Naturalist , 154 , S4–S25. Jones, J.H. (1998b). Optimization of the mammalian respi- Ketterson, E.D., Nolan, V. Jr, Wolf, L., Ziegenfus, C., Dufty, ratory system: symmorphosis versus single species A.M., Ball, G.F., and Johnsen, T.S. (1991). Testosterone adaptation. Comparative Biochemistry and Physiology B and avian life histories: the effect of experimentally ele- 120 , 125–38. vated testosterone on corticosterone and body-mass in Jones, P.H., Jacobs, I., Morris, A., and Ducharme, M.B. dark-eyed juncos. Hormones and Behavior 25 , 489–503. (1993). Adequacy of food rations in soldiers during an Ketterson, E.D., Nolan, V. Jr, Cawthorn, M.J., Parker, P.G., arctic exercise measured by doubly labelled water. and Ziegenfus, C. (1996). Phenotypic engineering: using Journal of Applied Physiology 75 , 1790–97. hormones to explore the mechanistic bases of pheno- Jonsson, K.I. (1997). Capital and income breeding as alter- type variation in nature. Ibis 138 , 70–86. native tactics of resource use in reproduction. Oikos 78 , Kierdorf, U., Kierdorf, H., and Szuwart, T. (2007). Deer 57–66. antler regeneration: cells, concepts and controversies. Jovani, R., Tella, J.L., Forero, M.G., Bertelotti, M., Blanco, Journal of Morphology 268 , 726–38. G., Ceballos, O., and Donázar, J.A. (2001). Apparent King, J.M., Parsons, D.J., Turnpenny, J.R., Nyangaga, J., absence of blood parasites in the Patagonian seabird Bakari, P., and Wathes, C.M. (2006). Modelling energy community: is it related to the marine environment? metabolism of Friesians in Kenya smallholdings shows Waterbirds 24 , 430–33. how heat stress and energy defi cit constrain milk yield Juanes, F. and Hartwick, E.B. (1990). Prey size selection in and cow replacement rate. Animal Science 82 , 705–16. Dungeness crabs: the effect of claw damage. Ecology 71 , King, J.R. (1964). Oxygen consumption and body tempera- 744–58. ture in relation to ambient temperature in the white- Kaandorp, J.A. (1999). Morphological analysis of growth crowned sparrow. Comparative Biochemistry and forms of branching marine sessile organisms along Physiology 12 , 13–24. environmental gradients. Marine Biology 134, 295–306. King, J.R (1974). Seasonal allocation of time and energy Kacelnik, A. (1984). Central-place foraging in starlings resources in birds. In Avian energetics, ed. R.A. Paynter Jr, (Sturnus vulgaris). I. Patch residence time. Journal of pp. 4–85. Nuttall Onithological Club, Cambridge, Mass. Animal Ecology 53 , 283–99. Kingsland, S.E. (1995). Modeling nature: Episodes in the his- Kahl, P.M. Jr. (1963). Thermoregulation in the wood stork, tory of population ecology, 2nd edition. University of with special reference to the role of the legs. Physiological Chicago Press, Chicago. Zoology 36 , 141–51. Kingsolver, J.G. (1988). Thermoregulation, fl ight and the Kamermans, P. and Smaal, A.C. (2002). Mussel culture and evolution of wing pattern in pierid butterfl ies: the cockle fi sheries in The Netherlands: Finding a balance topography of adaptive landscapes. American Zoologist between economy and ecology. Journal of Shellfi sh 28 , 899–912. Research 21 , 509–17. Kingsolver, J.G. (1995a). Viability selection on seasonally Karasov, W.H. (1986). Energetics, physiology and vertebrate polyphenic traits: wing melanin patterns in western ecology. Trends in Ecology and Evolution 4 , 101–104. white butterfl ies. Evolution 49 , 932–41. Karasov, W.H. and Martínez del Rio, C. (2007). Physiological Kingsolver, J.G. (1995b). Fitness consequences of seasonal ecology: How animals process energy, nutrients, and toxins . polyphenism in western white butterfl ies. Evolution 49 , Princeton University Press, Princeton, NJ. 942–54. Kendeigh, S.C. (1970). Energy requirements for existence Kirkwood, J.K. (1983). A limit to metabolisable energy in relation to size of bird. Condor 72 , 60–65. intake in mammals and birds. Comparative Biochemistry Kersten, M. and Piersma, T. (1987). High levels of energy and Physiology A 75 , 1–3. expenditure in shorebirds: Metabolic adaptations to an Kirschner, M.W. and Gerhart, J.C. (2005). The plausibility of energetically expensive way of life. Ardea 75 , 175–87. life: Resolving Darwin´s dilemma. Yale University Press, Kersten, M. and Visser, W. (1996). The rate of food provi- New Haven. sioning in the oystercatcher: Food intake and energy Kishida, O. and Nishimura, K. (2006). Flexible architecture expenditure constrained by a digestive bottleneck. of inducible morphological plasticity. Journal of Animal Functional Ecology 10 , 440–48. Ecology 75 , 705–12. Kersten, M., Bruinzeel, L.W., Wiersma, P., and Piersma, T. Kishida, O., Mizuta, Y., and Nishimura, K. (2006). Reciprocal (1998). Reduced basal metabolic rate of migratory wad- phenotypic plasticity in predator-prey interaction ers wintering in coastal Africa. Ardea 86 , 71–80. between larval amphibians. Ecology 87 , 1599–1604. REFERENCES 199

Kishida, O, Trussell, G.C., and Nishimura, K. (2009). Top- Koteja, P. (1995). Maximum cold-induced energy assimila- down effects on antagonistic inducible defense and tion in a rodent, Apodemus fl avicollis . Comparative offense. Ecology 90 , 1217–26. Biochemistry and Physiology A 112 , 479–85. Klaassen, M. (1995). Water and energy limitations on fl ight Koteja, P. (1996). Limits to the energy budget in a rodent, range. Auk 112 , 260–62. Peromyscus maniculatus : Does gut capacity set the limit? Klaassen, M. (1996). Metabolic constraints on long-dis- Physiological Zoology 69 , 994–1020. tance migration in birds. Journal of Experimental Biology Koteja, P., Swallow, J.G., Carter, P.A., and Garland, T. 199 , 57–64. (2001). Maximum cold-induced food consumption in Klaassen, M. (2004). May dehydration risk govern long- mice selected for high locomotor activity: implications distance migratory behaviour? Journal of Avian Biology for the evolution of endotherm energy budgets. Journal 35 , 4–6. of Experimental Biology 204 , 1177–90. Klaassen, M. and Ens, B.J. (1990). Is salt stress a problem Kraan, C., Piersma, T., Dekinga, A., Koolhaas, A., and Van for waders wintering on the Banc d’Arguin, Mauritania? der Meer, J. (2007). Dredging for edible cockles Ardea 78 , 67–74. Cerastoderma edule on intertidal fl ats: short-term conse- Klaassen, M. and Nolet, B.A. (2008). Stoichiometry of quences of fi shermen’s patch-choice decisions for target endothermy: shifting the quest from nitrogen to carbon. and non-target benthic fauna. ICES Journal of Marine Ecology Letters 11 , 785–92. Science 64 , 1735–42. Klaassen, M., Kvist, A., and Lindström, Å. (2000). Flight Kraan, C., van Gils, J.A., Spaans, B., Dekinga, A., Bijleveld, costs and fuel composition of a bird migrating in a wind A.I., van Roomen, M., Kleefstra, R., and Piersma, T. tunnel. Condor 102 , 444–51. (2009). Landscape-scale experiment demonstrates that Klasing K.C. (1998). Comparative avian nutrition . CAB Wadden Sea intertidal fl ats are used to capacity by mol- International, Wallingford. luscivore migrant shorebirds. Journal of Animal Ecology Kleiber, M. (1961). The fi re of life . John Wiley, New York. 78 , 1259–68. Klepal, W. (1990). The fundamentals of insemination in Kraan, C., van Gils, J.A., Spaans, B., Dekinga, A., and Cirripedes. Oceanography and Marine Biology: an Annual Piersma, T. (2010). Why Afro-Siberian red knots Calidris Review 28 , 353–79. canutus canutus have stopped staging in the western Klepal, W., Barnes, H., and Munn, E.A. (1972). The Dutch Wadden Sea during southward migration. Ardea morphology and histology of the cirripede penis. 97 , in press. Journal of Experimental Marine Biology and Ecology 10 , Kramer, D.L. and Nowell, W. (1980). Central place forag- 243–65. ing in the eastern chipmunk, Tamias striatus . Animal Knaus, W. (2009). Dairy cows trapped between perform- Behaviour 28 , 772–78. ance demands and adaptability. Journal of the Science of Krebs, H.A. (1950). Body size and tissue respiration. Food and Agriculture 89 , 1107–14. Biochimica et Biophysica Acta 4 , 249–69. Kodandaramaiah, U., Vallin, A., and Wiklund, C. (2009). Krebs, J.R. and Davies, N.B. (1978, 1984, 1991, 1997). Fixed eyespot display in a butterfl y thwarts attacking Behavioural ecology: An evolutionary approach . Blackwell birds. Animal Behaviour 77 , 1415–19. Science, Oxford. Koiter, T.R., Moes, H., Valkhof, N., and Wijkstra, S. (1999). Krebs, J.R. and Davies, N.B. (1987). An introduction to Interaction of late pregnancy and lactation in rats. behavioural ecology . Blackwell Science, Oxford. Journal of Reproduction and Fertility 115 , 341–47. Krebs, J.R., Ryan, J.C., and Charnov, E.L. (1974). Hunting Konarzewski, M. and Diamond, J.M. (1995). Evolution of by expectation or optimal foraging? A study of patch basal metabolic rate and organ masses in laboratory use by chickadees. Animal Behaviour 22 , 953–64. mice. Evolution 49 , 1239–48. Krebs, J.R., Erichsen, J.T., Webber, M.I., and Charnov, E.L. Kooi, R.E. and Brakefi eld, P.M. (1999). The critical period (1977). Optimal prey selection in the great tit (Parus for wing pattern induction in the polyphenic tropical major ). Animal Behaviour 25 , 30–38. butterfl y Bicyclus anynana (Satyrinae). Journal of Insect Krogh, A. (1916). The respiratory exchange of animals and Physiology 45 , 201–12. man . Longmans-Green, London. Kopp, M. and Tollrian, R. (2003). Trophic size polyphen- Król, E. and Speakman, J.R. (2003a). Limits to sustained ism in Lembadion bullinum : costs and benefi ts of an intake. VI. Energetics of lactation in laboratory mice at inducible offense. Ecology 84 , 641–51. thermoneutrality. Journal of Experimental Biology 206 , Koteja, P. (1991). On the relation between basal and fi eld 4255–66. metabolic rates in birds and mammals. Functional Król, E. and Speakman, J.R. (2003b). Limits to sustained Ecology 5 , 56–64. intake. VII. Milk energy ouput in laboratory mice at 200 REFERENCES

thermoneutrality. Journal of Experimental Biology 206 , Lam, M., O’Connor, T.P., and Diamond, J.M. (2002). Loads, 4267–81. capacities, and safety factors of maltase and the glucose Król, E., Johnson, M.S., and Speakman, J.R. (2003). Limits transporter SGLT1 in mouse intestinal brush border. to sustained intake. VIII. Resting metabolic rate and Journal of Physiology 542 , 493–500. organ morphology of laboratory mice lactating at ther- Lampert, W., Tollrian, R., and Stibor, H. (1994). moneutrality. Journal of Experimental Biology 206 , Chemische Induktion von Verteidigungsmechanismen 4283–91. bei Süßwassertieren. Naturwissenschaften 81 , 375–82. Król, E., Murphy, M., and Speakman, J.R. (2007). Limits to Lande, R. (2009). Adaptation to an extraordinary environ- sustained intake. X. Effects of fur removal on reproduc- ment by evolution of phenotypic plasticity and genetic tive performance in laboratory mice. Journal of assimilation. Journal of Evolutionary Biology 22 , 1435–46. Experimental Biology 210 , 4233–43. Landys, M.M., Piersma, T., Visser, G.H., Jukema, J., and Książek, A., Konarzewski, M., and Łapo, I.B. (2004). Wijker, A. (2000). Water balance during real and simu- Anatomic and energetic correlates of divergent selec- lated long-distance migratory fl ight in the bar-tailed tion for basal metabolic rate in laboratory mice. godwit. Condor 102 , 645–52. Physiological and Biochemical Zoology 77 , 890–99. Landys-Ciannelli, M.M., Jukema, J., and Piersma, T. (2002). Książek, A., Czerniecki, J., and Konarzewski, M. (2009). Blood parameter changes during stopover in a long- Phenotypic fl exibility of traits related to energy acquisi- distance migratory shorebird, the bar-tailed godwit tion in mice divergently selected for basal metabolic Limosa lapponica taymyrensis . Journal of Avian Biology 33 , rate (BMR). Journal of Experimental Biology 212 , 808–14. 451–55. Künkele, J. (2000). Effects of litter size on the energetics of Lane, N. (2005). Power, sex, suicide: Mitochondria and the reproduction in a highly precocial rodent, the guinea meaning of life . Oxford University Press, Oxford. pig. Journal of Mammalogy 81 , 691–700. Lane, N. (2009a). Life ascending: the ten great inventions of Kvist, A. and Lindström, Å. (2000). Maximum daily energy evolution . Norton, New York. intake: It takes time to lift the metabolic ceiling. Lane, N. (2009b). The furnace within. New Scientist 2694 , Physiological and Biochemical Zoology 73 , 30–36. 42–45. Kvist, A. and Lindström, Å. (2001). Basal metabolic rate in Lank, D.B. and Ydenberg, R.C. (2003). Death and danger migratory waders: Intra-individual, intraspecifi c, inter- at migratory stopovers: Problems with ‘predation risk’. specifi c and seasonal variation. Functional Ecology 15 , Journal of Avian Biology 34 , 225–28. 465–73. Lasiewski, R.C. and Dawson, W.R. (1967). A re-examination Kvist, A. and Lindström, Å. (2003). Gluttony in migratory of the relation between standard metabolic rate and waders – unprecedented energy assimilation rates in body weight in birds. Condor 69 , 13–23. vertebrates. Oikos 103 , 397–402. Laundré, J.W., Hernández, L., and Altendorf, K.B. (2001). Kvist, A., Lindström, Å., Green, M., Piersma, T., and Visser, Wolves, elk, and bison: Reestablishing the ‘landscape of G.H. (2001). Carrying large fuel loads during sustained fear’ in Yellowstone National Park, USA. Canadian bird fl ight is cheaper than expected. Nature 413 , 730–32. Journal of Zoology 79 , 1401–1409. Labocha, M.K., Sadowska, E.T., Baliga, K., Semer, A.K., Laurien-Kehnen, C. and Trillmich, F. (2003). Lactation per- and Koteja, P. (2004). Individual variation and repeata- formance of guinea pigs (Cavia porcellus) does not bility of basal metabolism in the bank vole, Clethrionomys respond to experimental manipulation of pup demands. glareolus . Proceedings of the Royal Society,London, B 271 , Behavioral Ecology and Sociobiology 53 , 145–52. 367–72. Lawton, J.H. and May, R.M. eds (1995). Extinction rates . Lack, D. (1947). Darwin’s fi nches . Cambridge University Oxford University Press, Oxford. Press, Cambridge. Le Maho, Y., Vu Van Kha, H., Koubi, H., Dewasmes, G., Lackner, J.R., and DiZio, P. (2000). Artifi cial gravity as a Girard, J., Ferre, P., and Cagnard, M. (1981). Body com- countermeasure in long-duration space fl ight. Journal of position, energy expenditure, and plasma metabolites Neuroscience Research 62 , 169–76. in long-term fasting geese. American Journal of Physiology Laland, K.N., Odling-Smee, F.J., and Feldman, M.W. 241 , E342–54. (1999). Evolutionary consequences of niche construc- Lenihan, H.S. and Micheli, F. (2001). Soft-sediment com- tion and their implications for ecology. Proceedings of the munities. In, Marine community ecology , eds M.D. National Academy of Sciences, USA 96 , 10242–47. Bertness and S.D. Gaines, pp. 253–88. Sinauer Publishers, Laland, K.N., Odling-Smee, J., and Gilbert, S.F. (2008). Sunderland. EvoDevo and niche construction: Building bridges. Lessells, C.M. (1991). The evolution of life histories. In Journal of Experimental Zoology B 310B , 549–66. Behavioural ecology: An evolutionary approach, eds J.R. REFERENCES 201

Krebs and N.B. Davies, 3rd edition, pp. 32–68. Blackwell Lindström, Å. (2003). Fuel deposition rates in migrating Science, Oxford. birds: causes, constraints and consequences. In Avian Lessells, C.M. (2007). Neuroendocrine control of life histo- migration , eds P. Berthold, E. Gwinner, and E. Sonnenschein, ries: what do we need to know to understand the evolu- pp. 307–20. Springer-Verlag, Berlin. tion of phenotypic plasticity? Philosophical Transactions Lindström, Å. and Kvist, A. (1995). Maximum energy of the Royal Society, London, B 363 , 1589–98. intake rate is proportional to basal metabolic rate in pas- Levey, D.J. and Karasov, W.H. (1989). Digestive responses serine birds. Proceedings of the Royal Society, London, B of temperate birds switched to fruit or insect diets. Auk 261 , 337–43. 106 , 675–86. Lindström, Å. and Nilsson, J.Å. (1988). Birds doing it the Levins, R. (1968). Evolution in changing environments . octopus way—fright moulting and distraction of preda- Princeton University Press, Princeton, NJ. tors. Ornis Scandinavica 19 , 165–66. Levitan, D.R. (1989). Density-dependent size regulation in Lindström, Å., Klaassen, M., and Kvist, A. (1999). Variation Diadema antillarum : Effects on fecundity and survivor- in energy intake and basal metabolic rate of a bird ship. Ecology 70 , 1414–24. migrating in a wind-tunnel. Functional Ecology 13 , Levitan, D.R. (1991). Skeletal changes in the test and jaws 352–59. of the sea urchin Diadema antillarum in response to food Liu, Z., Jaitner, J., Reinhardt, F., Pasman, E., Rensing, S., limitation. Marine Biology 111 , 431–35. and Reents, R. (2008). Genetic evaluation of fertility Lewontin, R.C. (2000). The triple helix: Gene, organism, and traits of dairy cattle using a multiple-trait animal model. environment . Harvard University Press, Cambridge, Mass. Journal of Dairy Science 91 , 4333–43. Lewontin, R.C. and Levins, R. (2007). Biology under infl u- Lively, C.M. (1986). Predator-induced shell dimorphism in ence: Dialectical essays on ecology, agriculture, and health . the acorn barnacle Chthamalus fi ssus . Evolution 40 , Monthly Review Press, New York. 232–42. Leyrer J., Pruiksma S., and Piersma T. (2009). On 4 June Lively, C.M. (1999). Developmental strategies in spatially 2008 Siberian red knots at Elbe Mouth kissed the canon- variable environments: Barnacle shell dimorphism and ical evening migration departure rule goodbye. Ardea strategic models of selection. In The ecology and evolution 97 , 71–79. of inducible defenses, eds R. Tollrian and C.D. Harvell, Li, N.K. and Denny, M.W. (2004). Limits to phenotypic pp. 245–58. Princeton University Press, Princeton, NJ. plasticity: fl ow effects on barnacle feeding appendages. Lively, C.M., Hazel, W.N., Schellenberger, M.J., and Biological Bulletin 206 , 121–24. Michelson, K.S. (2000). Predator-induced defense: Lignot, J-H., Helmstetter, C., and Secor, S.M. (2005). Variation for inducibility in an intertidal barnacle. Postprandial morphological response of the epithelium Ecology 81 , 1240–47. of the Burmese python ( Python molurus ). Comparative Lotze, H.K., Reise, K., Worm, B., Van Beusekom, J., Busch, Biochemsitry and Physiology A 141 , 280–91. M., Ehlers, A., Heinrich, D., Hoffmann, R.C., Holm, P., Lima, S.L. (1993). Ecological and evolutionary perspec- Jensen, C., Knottnerus, O.S., Langhanki, N., Prummel, tives on escape from predatory attack – a survey of W., Vollmer, M., and Wolff, W.J. (2005) Human transfor- North-American birds. Wilson Bulletin 105 , 1–47. mations of the Wadden Sea ecosystem through time: a Lincoln, G.A. (1992). Biology of antlers. Journal of Zoology synthesis. Helgolander Marine Research 59 , 84–95. 226 , 517–28. Lovegrove, B.G. (2000). The zoogeography of mammalian Lind, M.I. and Johansson, F. (2009). Costs and limits of basal metabolic rate. American Naturalist 156 , 201–19. phenotypic plasticity in island populations of the com- Lucia, A., Hoyos, J., Santalla, A., Earnest, C., and Chicharro, mon frog Rana temporaria under divergent selection J.L. (2003). Tour de France versus Vuelta a España: pressures. Evolution 63 , 1508–18. which is harder? Medicine and Science in Sports and Lindstedt, S.L. and Conley, K.E. (2001). Human aerobic Exercise 35 , 872–78.

performance: too much ado about limits to VO 2max . Lucy, M.C. (2001). Reproductive loss in high-producing Journal of Experimental Biology 204 , 3195–99. dairy cattle: where will it end? Journal of Dairy Science Lindstedt, S.L. and Jones, J.H. (1987). Symmorphosis and 84 , 1277–93. the concept of optimal design. In New directions in eco- Luttbeg, B. and Warner, R.R. (1999). Reproductive deci- logical physiology, eds M.E. Feder, A.F. Bennett, W.W. sion-making by female peacock wrasses: fl exible versus Burggren, and R.B. Huey, pp. 289–309. Cambridge fi xed behavioural rules in a variable environment. University Press, Cambridge. Behavioral Ecology 10 , 666–74. Lindström, Å. (1991). Maximum fat deposition rates in Luttikhuizen, P.C., Honkoop, P.J.C., Drent, J., and van der migrating birds. Ornis Scandinavica 22 , 12–19. Meer, J. (2004). A general solution for optimal egg size 202 REFERENCES

during external fertilization, extended scope for inter- McFarland, D.J. (1977). Decision making in animals. Nature mediate optimal egg size and the introduction of Don 269 , 15–21. Ottavio ‘tango’. Journal of Theoretical Biology 231 , McKechnie, A.E., Freckleton, R.O., and Jetz, W. (2006). 333–43. Phenotypic plasticity in the scaling of avian basal meta- MacArthur, R.H. and Pianka, E.R. (1966). On optimal use bolic rate. Proceedings of the Royal Society, London, B 273 , of a patchy environment. American Naturalist 100 , 931–37. 603–609. McLandress, M.R. and Raveling, D.G. (1981). Changes in Machado, A. and Keen, R. (1999). Learning to time (LeT) diet and body composition in Canada geese before or scalar expectancy theory (SET)? A critical test of two spring migration. Auk 98 , 65–79. models of timing. Psychological Science 10 , 285–90. McLaughlin, R.L. and Montgomerie, D.R. (1990). Flight MacLean, S.F. Jr. (1974). Lemming bones as a source of cal- speeds of parent birds feeding nestlings: maximization cium for Arctic sandpipers ( Calidris spp.). Ibis 116 , of foraging effi ciency or food delivery rate?Canadian 552–57. Journal of Zoology 68 , 2269–74. Maclean, G. (1996). Ecophysiology of desert birds . Springer- McNab, B.K. (1986). The infl uence of food habits on the Verlag, Berlin. energetics of eutherian mammals. Ecological Monographs MacLeod, R., Barnett, P., Clark, J.A., and Cresswell, W. 56 , 1–19. (2005). Body mass change strategies in blackbirds Turdus McNab, B.K. (1988). Food habits and the basal rate of merula : the starvation–predation trade-off. Journal of metabolism in birds. Oecologia 77 , 343–49. Animal Ecology 74 , 292–302. McNab, B.K. (1994). Energy conservation and the evolu- Malapani, C. and Fairhurst, S. (2002). Scalar timing in ani- tion of fl ightlessness in birds.American Naturalist 144 , mals and humans. Learning and Motivation 33 , 156–76. 628–42. Marchinko, K.B. (2003). Dramatic phenotypic plasticity in McNab, B.K. (1996). Metabolism and temperature regula- barnacle legs (Balanus glandula Darwin): magnitude, age tion of kiwis (Apterygidae). Auk 113 , 687–92. dependence, and speed of response. Evolution 57, 1281–90. McNab, B.K. (2002). The physiological ecology of vertebrates: Marchinko, K.B. (2007). Feeding behavior reveals the A view from energetics . Cornell University Press, Ithaca. adaptive nature of plasticity in barnacle feeding limbs. McNair, J.N. (1982). Optimal giving-up times and the mar- Biological Bulletin , 213 , 12–15. ginal value theorem. American Naturalist 119 , 511–29. Marchinko, K.B. and Palmer, A.R. (2003). Feeding in fl ow McNamara, J.M. (1982). Optimal patch use in a stochastic extremes: dependence of cirrus form on wave-exposure environment. Theoretical Population Biology 21 , 269–88. in four barnacle species. Zoology 106 , 127–41. McNamara, J.M. and Houston, A.I. (1997). Currencies for Masman, D., Daan, S., and Dijkstra, C. (1988). Time alloca- foraging based on energetic gain. American Naturalist tion in the kestrel (Falco tinnunculus ), and the principle 150 , 603–17. of energy minimization. Journal of Animal Ecology , 57 , McNamara, J.M., Green, R.F., and Olsson, O. (2006). Bayes’ 411–32. theorem and its application in animal behaviour. Oikos Matson, K.D., Ricklefs, R.E., and Klasing, K.C. (2005). A 112 , 243–51. hemolysis-hemagglutination assay for characterizing McNamara, J.M., Barta, Z., Wikelsi, M., and Houston, A.I. constitutive innate humoral immunity in wild and (2008). A theoretical investigation of the effect of lati- domestic birds. Developmental and Comparative tude on avian life histories. American Naturalist 172 , Immunology 29 , 275–86. 331–45. Mauroy, B., Filoche, M., Weibel, E.R., and Sapoval, B. McNeill, W.H. (1976). Plagues and people. Doubleday, (2004). An optimal bronchial tree may be dangerous. New York. Nature 427 , 633–36. McWhirter, N. (1980). Guinness book of world records , 41st Mayr, E. (1963). Animal species and evolution . Belknap Press edition. Sterling Publishing Co., New York. of Harvard University Press, Cambridge, Mass. Meck, J.V., Reyes, C.J., Perez, S.A., Goldberger, A.L., and Mayr, E. (1982). The growth of biological thought: Diversity, Ziegler, M.G. (2001). Marked exacerbation of orthostatic evolution, and inheritance. Harvard University Press, intolerance after long- vs. short-duration spacefl ight in Cambridge, Mass. veteran astronauts. Psychosomatic Medicine 63 , 865–73. Mayr, E. (2002). What evolution is. Weidenfeld and Meijer, T. and Drent, R. (1999). Re-examination of the capi- Nicholson, London. tal and income dichotomy in breeding birds. Ibis 141, McEwen, B.S. and J.C. Wingfi eld (2003). The concept of 399–414. allostasis in biology and biomedicine. Hormones and Meiri, S. and Dayan, T. (2003). On the validity of Behavior 43 , 2–15. Bergmann’s rule. Journal of Biogeography 30 , 331–51. REFERENCES 203

Mendes, L., Piersma, T., Lecoq, M., Spaans, B., and Ricklefs, Monaghan, P. (2008). Early growth conditions, phenotypic R.E. (2005). Disease-limited distributions? Contrasts in development and environmental change. Philosophical the prevalence of avian malaria in shorebird species Transactions of the Royal Society, London, B 363 , 1635–45. using marine and freshwater habitats. Oikos 109 , Monaghan, P., Charmantier, A., Nussey, D.H., and Ricklefs, 396–404. R.E. (2008). The evolutionary ecology of senescence. Mendes, L., Piersma, T., Hasselquist, D., Matson, K.D., Functional Ecology 22 , 371–78. and Ricklefs, R.E. (2006a). Variation in the innate and Monaghan, P., Metcalfe, N.B., and Torres, R. (2009). acquired arms of the immune system among fi ve shore- Oxidative stress as a mediator of life history trade-offs: bird species. Journal of Experimental Biology 209 , 284–91. Mechanisms, measurements and interpretation. Ecology Mendes, L., Piersma, T., and Hasselquist, D. (2006b). Two Letters 12 , 75–92. estimates of the metabolic costs of antibody production Montgomerie, R., Lyon, B., and Holder, K. (2001). Dirty in migratory shorebirds: low costs, internal reallocation, ptarmigan: behavioral modifi cation of conspicuous or both? Journal of Ornithology 147 , 274–300. male plumage. Behavioral Ecology 12 , 429–38. Messner, R. (1979). Everest: Expedition to the ultimate . Kay Moran, N.A. (1992). The evolutionary maintenance of and Ward, London. alternative phenotypes. American Naturalist 139 , Metcalfe, N.B. and Monaghan, P. (2001). Compensation 971–89. for a bad start: grow now, pay later? Trends in Ecology Morgan, C. (1896). Of modifi cation and variation. Science and Evolution 16 , 254–60. 4 , 733–40. Meyers, L.A. and Bull, J.J. (2002). Fighting change with Morrison, R.I.G., Davidson, N.C., and Piersma, T. (2005). change: adaptive variation in an uncertain world. Trends Transformation at high latitudes: Why do red knots in Ecology and Evolution 17 , 551–57. bring body stores to the breeding grounds? Condor 107 , Michimae, H. (2006). Differentiated phenotypic plasticity 449–57. in larvae of the cannibalistic salamander Hynobius Morrison, S.F., Nakamura, K., and Madden, C.J. (2008). retardatus . Behavioral Ecology and Sociobiology 60 , 205–11. Central control of thermogenesis in mammals. Michimae, H., Nishimura, K., and Wakahara, M. (2005). Experimental Physiology 93 , 773–97. Mechanical vibrations from tadpoles’ fl apping tails Mousseau, T.A. and Fox, C.W. (1998). The adaptive signifi - transform salamander’s carnivorous morphology. cance of maternal effects. Trends in Ecology and Evolution Biology Letters 1 , 75–77. 13 , 403–407. Midgley, J.J., Midgley, G., and Bond, W.J. (2002). Why were Munday, P.L., Buston, P.M., and Warner, R.R. (2005). dinosaurs so large? A food quality hypothesis. Diversity and fl exibility of sex-change strategies in ani- Evolutionary Ecology Research 4 , 1093–95. mals. Trends in Ecology and Evolution 21 , 89–95. Millenium Ecosystem Assessment (2005). Ecosystems and Munday, P.L., White, J.W., and Warner, R.R. (2006). A social human well-being: Synthesis. Island Press, Washington, basis for the development of primary males in a sex- DC. changing fi sh. Proceedings of the Royal Society, London, B Miller, G. (2001). The mating mind: How sexual choice shaped 283 , 2845–51. the evolution of human nature . Vintage, London. Murata, A., Imafuku, M., and Abe, M. (2001). Copulation Miller, L.P. (2007). Feeding in extreme fl ows: behavior by the barnacle Tetraclita japonica under natural condi- compensates for mechanical constraints in barnacle tions. Journal of Zoology 253 , 275–80. cirri. Marine Ecology Progress Series 349 , 227–34. Murton, R.K. and Westwood, N.J. (1977). Avian breeding Millet, S., Bennet, J., Lee, K.A., Hau, M., and Klasing, K.C. cycles . Clarendon Press, Oxford. (2007). Quantifying and comparing constitutive immu- Nager, R.G., Keller, L.F., and van Noordwijk, A.J. (2000). nity across avian species. Developmental and Comparative Understanding natural selection on traits that are infl u- Immunology 31 , 188–201. enced by environmental conditions. In Adaptive genetic Miner, B.G., Sultan, S.E., Morgan, S.G., Padilla, D.K., and variation in the wild, eds T.A. Mousseau, B. Sinervo, Relyea, R.A. (2005). Ecological consequences of pheno- and J.A. Endler, pp. 95–115. Oxford University Press, typic plasticity. Trends in Ecology and Evolution 20 , New York. 685–91. Nassuna-Musoke, G.M., Kabasa, J.D., and King, M.J. Mitchell, A.C. (1901). On the convection of heat by air cur- (2007). Microclimate dynamics in smallholder Friesian rents. Transactions of the Royal Society of Edinburgh 40 , 39. dairies of the tropical warm-humid central Uganda. Møller, A.P., Nielsen, J.T., and Erritzoe, J. (2006). Losing Journal of Animal and Veterinary Advances 6 , 907–11. the last feather: feather loss as an antipredator adapta- Navarro, A., Gomez, C., Lopez-Cepero, J.M., and Boveris, tion in birds. Behavioral Ecology 17 , 1046–56. A. (2004). Benefi cial effects of moderate exercise on mice 204 REFERENCES

aging: Survival, behavior, oxidative stress, and mito- Nolet, B.A. and Klaassen, M. (2009). Retrodicting patch chondrial electron transfer. American Journal of Physiology use by foraging swans in a heterogeneous environment 286 , R505–11. using a set of functional responses. Oikos 118 , 431–39. Naya, D.E., Karasov, W.H., and Bozinovic, F. (2007). Nolet, B.A., Langevoord, O., Bevan, R.M., Engelaar, K.R., Phenotypic plasticity in laboratory mice and rats: a Klaassen, M., Mulder, R.J.W., and van Dijk, S. (2001). meta-analysis of current ideas on gut size fl exibility. Spatial variation in tuber depletion by swans explained Evolutionary Ecology Research 9 , 1363–74. by differences in net intake rates. Ecology 82 , 1655–67. Naya, D.E., Veloso, C., Dabat, P., and Bozinovic, F. (2009). Nottebohm, F. (1981). A brain for all seasons: cyclical ana- The effect of short- and long-term fasting on digestive tomical changes in song control. Science 214 , 1368–70. and metabolic fl exibility in the Andean toad,Bufo spinu- Nunney, L. (2003). The cost of natural selection revisited. losus . Journal of Experimental Biology 212 , 2167–75. Annales Zoologici Fennici 40 , 185–94. Neufeld, C.J. and Palmer, A.R. (2008). Precisely propor- Nussey, D.H., Postma, E., Gienapp, P., and Visser, M.E. tioned: Intertidal barnacles alter penis form to suit (2005). Selection on heritable phenotypic plasticity in a coastal wave action. Proceedings of the Royal Society, wild bird population. Science 310 , 304–306. London, B 275 , 1081–87. Nüsslein-Volhard, C. (2006). Coming to life: How genes drive Nevill, A.M. and Whyte, G. (2005). Are there limits to run- development . Kales Press, Carlsbad, CA. ning world records? Medical Science Sports and Exercise Oaten, A. (1977). Optimal foraging in patches: a case for 37 , 1785–88. stochasticity. Theoretical Population Biology 12 , 263–85. Newman, R.A. (1992). Adaptive plasticity in amphibian Ochsenbein, A.F. and Zinkernagel, R.M. (2000). Natural metamorphosis. BioScience 32 , 671–78. antibodies and complement link innate and acquired Newman, S.A. and Müller, G.B. (2000). Epigenetic mecha- immunity. Immunology Today 21 , 624–30. nisms of character origination. Journal of Experimental O’Connor, T. and Diamond, J.M. (1999). Ontogeny of Zoology B , 288B , 304–17. intestinal safety factors: Lactase capacities and lactose Newton, I. (1701). Scala graduum caloris: Calorum loads. American Journal of Physiology 276 , R753–65. descriptiones & fi gna. Philosophical Transactions of the O’Connor, T., Lam, M., and Diamond, J.M. (1999). The Royal Society , London 22 , 824–29. magnitude of functional adaptation after intestinal Newton, I. (2008). The migration ecology of birds . Elsevier/ resection. American Journal of Physiology 276 , R1265–75. Academic Press, London. Odling-Smee, F.J., Laland, K.N., and Feldman, M.W. (1996). Nijhout, H.F. (1999). Control mechanisms of polyphenic Niche construction. American Naturalist 147 , 641–48. development in insects. BioScience 49 , 181–92. Odling-Smee, F.J., Laland, K.N., and Feldman, M.W. Noakes, T.D. (1992). Lore of running: Discover the science (2003). Niche construction: The negelected process in evolu- and the spirit of running . 3rd Ed. Leisure Press, tion . Princeton University Press, Princeton, NJ. Champaign, Ill. Odum, E.P., Rogers, D.T., and Hicks, D.L. (1964). Noakes, T.D. (2006). The limits of endurance exercise. Basic Homeostasis of the non-fat components of migrating Research in Cardiology 101 , 408–17. birds. Science 143 , 1037–39. Noakes, T.D. (2007). The limits of human endurance: what Olff, H., Alonso, D., Berg, M.P., Eriksson, B.K., Loreau, M., is the greatest endurance performance of all time? Piersma, T., and Rooney, N. (2009). Parallel ecological Which factors regulate performance at extreme altitude? networks in ecosystems. Philosophical Transactions of the In Hypoxia and the circulation , ed. R.C. Roach, pp. 259–80. Royal Society, London, B 364 , 1755–79. Springer, New York. Oliver, J.C., Robertson, K.A., and Monteiro, A. (2009). Noakes, T.D., Peltonen, J.E., and Rusko, H.K. (2001). Accomodating natural and sexual selection in butterfl y Evidence that a central governor regulates exercise per- wing pattern evolution. Proceedings of the Royal Society, formance during acute hypoxia and hyperoxia. Journal London, B 276 , 2369–75. of Experimental Biology 201 , 3225–34. Olsson, O. and Brown, J.S. (2006). The foraging benefi ts of Noakes, T.D., St Clair Gibson, A., and Lambert, E.V. information and the penalty of ignorance. Oikos 112 , (2004). From catastrophe to complexity: a novel model 260–73. of integrative central neural regulation of effort and Olsson, O. and Brown, J.S. (2010). Smart, smarter, smart- fatigue during exercise in humans: Summary and est: foraging information states and coexistence. Oikos conclusions. British Journal of Sports Medicine 39 , 119 , 292–303. 120–24. Olsson, O. and Holmgren, N.M.A. (1998). The survival- Noble, D. (2006). The music of life: Biology beyond the genome . rate-maximizing policy for Bayesian foragers: wait for Oxford University Press, Oxford. good news. Behavioral Ecology 9 , 345–53. REFERENCES 205

Olsson, O., Wiktander, U., Holmgren, N.M.A., and Nilsson, Partridge, L. and Gems, D. (2007). Benchmarks for ageing S. (1999). Gaining ecological information about Bayesian studies. Nature 450 , 165–67. foragers through their behaviour. II. A fi eld test with Pauly, D. (2007). The Sea Around Us Project: documenting woodpeckers. Oikos 87 , 264–76. and communicating global fi sheries impacts on marine Olsson, O., Brown, J.S., and Helf, K.L. (2008). A guide to ecosystems. Ambio 36 , 290–95. central place effects in foraging. Theoretical Population Payne, W.M.C., Williams, D.R., and Trudel, G. (2007). Biology 74 , 22–33. Space fl ight rehabilitation.American Journal of Physical Orians, G.H. and Pearson, N.E. (1979). On the theory of Medicine & Rehabilitation 86 , 583–91. central place foraging. In Analysis of ecological systems , Peaker, M. and Linzell, J.L. (1975). Salt glands in birds and eds D.J. Horn, R.D. Mitchell, and G.R. Stairs, pp. 154–77. reptiles . Cambridge University Press, Cambridge. Ohio State University Press, Columbus. Pearl, R. (1922). The biology of death . J.B. Lippincott, Orr, J.C., Fabry, V.J., Aumont, O., Bopp, L., Doney, S.C., Philadelphia. Feely, R.A., Gnanadesikan, A., Gruber, N., Ishida, A., Pearl, R. (1928). The rate of living. Albert Knopf, New York. Joos, F., Key, R.M., Lindsay, K., Maier-Reimer, E., Matear, Pennisi, E. (2005). The inner tube of life—the dynamic gut. R., Monfray, P., Mouchet, A., Najjar, R.G., Plattner, G-K., Science 307 , 1896–99. Rodgers, K.B., Sabine, C.L., Sarmiento, J.L., Schlitzer, R., Pennycuick, C.J. and Battley, P.F. (2003). Burning the Slater, R.D., Totterdell, I.J., Weirig, M-F., Yamanaka, Y., engine: A time-marching computation of fat and protein and Yool, A. (2005). Anthropogenic ocean acidifi cation consumption in a 5420 km non-stop fl ight by great knots over the twenty-fi rst century and its impact on calcify- Calidris tenuirostris . Oikos 103 , 323–32. ing organisms. Nature 437 , 681–86. Pennycuick, C.J., Einarsson, O., Bradbury, T.A.M., and Orr, N.W.M. (1966). The feeding of sledge dogs on Antarctic Owen, M. (1996). Migrating whooper swans Cygnus expeditions. British Journal of Nutrition 20 , 1–12. cygnus : satellite tracks and fl ight performance calcula- Osborn, H.F. (1897). The limits of organic selection. tions. Journal of Avian Biology 27 , 118–34. American Naturalist 31 , 944–51. Pennycuick, C.J., Alerstam, T., and Hedenström, A. (1997). Ott, B.D. and Secor, S.M. (2007). Adaptive regulation of A new low turbulence wind tunnel for bird fl ight exper- digestive performance in the genus Python . Journal of iments at Lund University, Sweden. Journal of Experimental Biology 210 , 340–56. Experimental Biology 200 , 1441–49. Oufi ero, C.E. and Garland, T. Jr (2007). Evaluating per- Perez-Campo, R., López-Torres, M., Cadenas, S., Rojas, C., formance costs of sexually selected traits. Functional and Barja, G. (1998). The rate of free radical production Ecology 21 , 676–89. as a determinant of the rate of aging: evidence from the Owen, M. (1981). Abdominal profi le—a condition index comparative approach. Journal of Comparative Physiology for wild geese in the fi eld. Journal of Wildlife Management B 168 , 149–58. 45 , 227–30. Perhonen, M.A., Franco, F., Lane, L.D., Buckey, J.C., Padilla, D.K. (2001). Food and environmental cues trigger Blomqvist, C.G., Zerwekh, J.E., Peschock, R.M., an inducible offense. Evolutionary Ecology Research 3 , Weatherall, P.T., and Levine, B.D. (2001). Cardiac atro- 15–25. phy after bed rest and spacefl ight. Journal of Applied Padilla, D.K. and Adolph, S.C. (1996). Plastic inducible Physiology 91 , 645–53. morphologies are not always adaptive: the importance Perrigo, G. (1987). Breeding and feeding strategies in of time delays in a stochastic environment. Evolutionary deer mice and house mice when females are chal- Ecology 10 , 105–107. lenged to work for their food. Animal Behaviour 35 , Painter, R.C., Osmond, C., Gluckman, P., Hanson, M., Phillips, 1298–316. D.I.W., and Roseboom, T.J. (2008). Transgenerational Perry, G. and Pianka, E.R. (1997). Animal foraging: past, effects of prenatal exposure to the Dutch famine on neona- present and future. Trends in Ecology and Evolution 12 , tal adiposity and health in later life. British Journal of 360–64. Obstetrics and Gynaecology 115 , 1243–49. Peters, R.H. (1983). The ecological implications of body size . Parmesan, C. and Yohe, G (2003). A globally coherent fi n- Cambridge University Press, Cambridge. gerprint of climate change impacts across natural sys- Peterson, C.C., Nagy, K.A., and Diamond, J.M.(1990). tems. Nature 421 , 37–42. Sustained metabolic scope. Proceedings of the National Parter, M., Kashtan, N., and Alon, U. (2008). Facilitated Academy of Sciences, USA 87 , 2324–28. variation: how evolution learns from past environments Petrusek, A., Tollrian, R., Schwenk, K., Haas, A., and to generalize to new environments. Public Library of Laforsch, C. (2009). A ‘crown of thorns’ is an inducible Science — Computational Biology 4 , e1000206. defense that protects Daphnia against an ancient preda- 206 REFERENCES

tor. Proceedings of the National Academy of Sciences, USA Piersma, T. and Gill, R.E. Jr (1998). Guts don’t fl y: small 106 , 2248–52. digestive organs in obese bar-tailed godwits. Auk 115 , Pfennig, D.W. (1992). Polyphenism in spadefoot toad tad- 196–203. poles as a locally adjusted evolutionarily stable strategy. Piersma, T. and Jukema, J. (1990). Budgeting the fl ight of a Evolution 46, 1408–20. long-distance migrant: changes in nutrient reserve lev- Pfennig, D.W. and Frankino, W.A. (1997). Kind-mediated els of bar-tailed godwits at successive spring staging morphogenesis in facultatively cannibalistic tadpoles. sites. Ardea 78 , 315–37. Evolution 51 , 1993–99. Piersma, T. and Jukema, J. (2002). Contrast in adaptive mass Pierre, J.S. and Green, R.F. (2007). A Bayesian approach to gains: Eurasian golden plovers store fat before midwinter optimal foraging in parasitoids. In Behavioral ecology of and protein before prebreeding fl ight. Proceedings of the insects parasitoids: From theoretical approaches to fi eld appli- Royal Society, London, B 269 , 1101–1105. cations , eds É. Wajnberg, C. Bernstein, and K. van Piersma, T., de Goeij, P., and Tulp, I. (1993a). An evaluation Alphen, pp. 357–83. Wiley–Blackwell, Oxford. of intertidal feeding habitats from a shorebird perspec- Piersma, T. (1987). Hop, skip or jump? Constraints on tive: towards relevant comparisons between temperate migration of arctic waders by feeding, fattening, and and tropical mudfl ats. Netherlands Journal of Sea Research fl ight speed.Limosa 60 , 185–94. 31, 503–12. Piersma, T. (1990). Pre-migratory ‘fattening’ usually Piersma, T., Koolhaas, A., and Dekinga, A. (1993b). involves more than the deposition of fat alone. Ringing Interactions between stomach structure and diet choice & Migration 11 , 113–15. in shorebirds. Auk 110 , 552–64. Piersma, T. (1997). Do global patterns of habitat use and Piersma, T., Hoekstra, R., Dekinga, A., Koolhaas, A., Wolf, migration strategies co-evolve with relative investments P., Battley, P.F., and Wiersma, P. (1993c). Scale and inten- in immunocompetence due to spatial variation in para- sity of intertidal habitat use by knots Calidris canutus in site pressure? Oikos 80 , 623–31. the western Wadden Sea in relation to food, friends and Piersma, T. (1998). Phenotypic fl exibility during migra- foes. Netherlands Journal of Sea Research 31 , 331–57. tion: optimization of organ size contingent on the risks Piersma, T., Tulp, I., and Schekkerman, H. (1994a). Final and rewards of fueling and fl ight? Journal of Avian countdown of waders during starvation: Terminal use Biology , 29 511–20. of nutrients in relation to structural size and concurrent Piersma, T. (2002). Energetic bottlenecks and other design energy expenditure. In Close to the edge: Energetic bottle- constraints in avian annual cycles. Integrative and necks and the evolution of migratory pathways in knots , ed. Comparative Biology 42 , 51–67. T. Piersma, pp. 220–27. Uitgeverij Het Open Boek, Den Piersma, T. (2003). ‘Coastal’ versus ‘inland’ shorebird spe- Burg. cies: Interlinked fundamental dichotomies between Piersma, T., Verkuil, Y., and Tulp, I. (1994b). Resources for their life- and demographic histories? Wader Study long-distance migration of knots Calidris canutus island- Group Bulletin 100 , 5–9. ica and C. c. canutus: how broad is the temporal exploita- Piersma, T. (2007). Using the power of comparison to explain tion window of benthic prey in the western and the habitat use and migration strategies of shorebirds world- eastern Wadden Sea? Oikos 71 , 393–407. wide. Journal of Ornithology 148 (Suppl. 1), S45–S59. Piersma, T., Cadée, N., and Daan, S. (1995a). Seasonality in Piersma, T. (2009). Threats to intertidal soft-sediment eco- basal metabolic rate and thermal conductance in a long- systems. In Water policy in The Netherlands: Integrated distance migrant shorebird, the knot (Calidris canutus ). management in a densely populated delta, eds S. Reinhard Journal of Comparative Physiology B 165 , 37–45. and H. Folmer, pp. 57–69. Resources for the Future, Piersma, T., van Gils, J., de Goeij, P., and van der Meer, J. Washington, DC. (1995b). Holling’s functional response model as a tool to Piersma, T. and Davidson, N.C. (1992). The migrations link the food-fi nding mechanism of a probing shorebird and annual cycles of fi ve subspecies of knots in perspec- with its spatial distribution. Journal of Animal Ecology 64 , tive. Wader Study Group Bulletin , 64 (Suppl.), 187–97. 493–504. Piersma, T. and Dietz, M.W. (2007). Twofold seasonal vari- Piersma, T., Bruinzeel, L., Drent, R., Kersten, M., van der ation in the supposedly constant, species-specifi c, ratio Meer, J., and Wiersma, P. (1996a). Variability in basal of upstroke to downstroke fl ight muscles in red knots metabolic rate of a long-distance migrant shorebird (red Calidris canutus . Journal of Avian Biology 38 , 536–40. knot, Calidris canutus ) refl ects shifts in organ sizes. Piersma, T. and Drent, J. (2003). Phenotypic fl exibility and Physiological Zoology 69 , 191–217. the evolution of organismal design. T rends in Ecology and Piersma, T., Gudmundsson, G.A., Davidson, N.C., and Evolution 18, 228–33. Morrison, R.I.G. (1996b). Do arctic-breeding Red Knots REFERENCES 207

(Calidris canutus) accumulate skeletal calcium before Porter, W.P. and Gates, D.M. (1969). Thermodynamic equi- egg laying? Canadian Journal of Zoology 74 , 2257–61. libria of animals with environment. Ecological Monographs Piersma, T., van Aelst, R., Kurk, K., Berkhoudt, H., and 39 , 227–44. Maas, L.R.M. (1998). A new pressure sensory mecha- Price, P.W. (1980). Evolutionary biology of parasites . Princeton nism for prey detection in birds: the use of principles of University Press, Princeton, NJ. seabed dynamics? Proceedings of the Royal Society, London, Price, T.D., Qvarnström, A., and Irwin, D.E. (2003). The B 265 , 1377–83. role of phenotypic plasticity in driving genetic evolu- Piersma, T., Dietz, M.W., Dekinga, A., Nebel, S., van Gils, tion. Proceedings of the Royal Society, London, B 270 , J., Battley, P.F., and Spaans, B. (1999a). Reversible size- 1433–40. changes in stomachs of shorebirds: when, to what Pugh, L.G. (1972). The logistics of the polar journeys of extent, and why? Acta Ornithologica 34 , 175–81. Scott, Shackleton and Amundsen. Proceedings of the Piersma, T., Gudmundsson, G.A., and Lilliendahl, K. Royal Society of Medicine 65 , 42–47. (1999b). Rapid changes in the size of different functional Pulliam, H.R. (1974). On the theory of optimal diets. organ and muscle groups during refueling in a long- American Naturalist 108 , 59–75. distance migrating shorebird. Physiological and Quaintenne, G., van Gils, J.A., Bocher, P., Dekinga, A., and Biochemical Zoology 72 , 405–15. Piersma, T. (2010). Diet selection in a molluscivore Piersma, T., Reneerkens, J., and Ramenofsky, M. (2000). shorebird across Western Europe: does it show short- or Baseline corticosterone peaks in shorebirds with maxi- long-term rate-maximization? Journal of Animal Ecology mal energy stores for migration: a general preparatory 79 , 53–62. mechanism for rapid behavioral and metabolic transi- Quammen, D. (2006). The reluctant Mr. Darwin: An intimate tions? General and Comparative Endocrinology 120 , portrait of Charles Darwin and the making of his theory of 118–26. evolution . Norton, New York. Piersma, T., Koolhaas, A., Dekinga, A., Beukema, J.J., Dekker, Radford, A.N., Hollén, L.I., and Bell, M.B.V. (2009). The R., and Essink, K. (2001). Long-term indirect effects of higher the better: Sentinel height infl uences foraging mechanical cockle-dredging on intertidal bivalve stocks in success in a social bird. Proceeding of the Royal Society, the Wadden Sea. Journal of Applied Ecology 38 , 976–90. London, B 276 , 2437–42. Piersma, T., Lindström, Å., Drent, R., Tulp, I., Jukema, J., Ramos, M., Irschick, D.J., and Christenson, T.E. (2004). Morrison, R.I.G., Reneerkens, J., Schekkerman, H., and Overcoming an evolutionary confl ict: Removal of a Visser, G.H. (2003). High daily energy expenditure of reproductive organ greatly increases locomotor per- incubating shorebirds on High Arctic tundra: A cir- formance. Proceedings of the National Academy of Sciences, cumpolar study. Functional Ecology 17 , 356–62. USA 101 , 4883–87. Piersma, T., Rogers, D.I., González, P.M., Zwarts, L., Niles, Ranta, E., Lundberg, P., and Kaitala, V. (2006). Ecology of L.J., de Lima Serrano do Nascimento, I., Minton, C.D.T., populations . Cambridge University Press, Cambridge. and Baker, A.J. (2005). Fuel storage rates in red knots Rauw, W.M., Kanis, E., Noordhuizen-Stassen, E.N., and worldwide: Facing the severest ecological constraint in Grommers, F.J. (1998). Undesirable side effects of selec- tropical intertidal conditions? In Birds of two worlds: The tion for high production effi ciency in farm animals: a ecology and evolution of migratory birds , eds R. Greenberg review. Lifestock Production Science 56 , 15–33. and P.P. Marra, pp 262–74. Johns Hopkins University Redfi eld, A.C. (1934). On the proportions of organic deri- Press, Baltimore. vations in sea water and their relation to the composi- Pigliucci, M. (2001a). Phenotypic plasticity: Beyond nature tion of plankton. In James Johnstone memorial volume , ed. and nurture . Johns Hopkins University Press, Baltimore. R. J. Daniel, pp. 177–92. University Press of Liverpool, Pigliucci, M. (2001b). Phenotypic plasticity. In Evolutionary Liverpool. ecology: Concepts and case studies, eds C.W. Fox, D.A. Redfi eld, A.C. (1958). The biological control of chemical fac- Roff, and D.J. Fairbairn, pp. 58–69. Oxford University tors in the environment. American Scientist 46 , 205–21. Press, New York. Reid, J.M., Bignal, E.M., Bignal, S., McCracken, D.I., and Pigliucci, M., Murren, C.J., and Schlichting, C.D. (2006). Monaghan, P. (2003). Environmental variability, life- Phenotypic plasticity and evolution by genetic assimila- history covariation and cohort effects in the red-billed tion. Journal of Experimental Biology 209 , 2362–67. chough Pyrrhocorax pyrrhocorax . Journal of Animal Ecology Polikansky, D. (1982). Sex change in plants and animals. 72 , 36–46. Annual Review of Ecology and Systematics 13 , 471–95. Relyea, R.A. (2001a). Morphological and behavioral plas- Popper, K. (1959). The logic of scientifi c discovery . Basic ticity of larval anurans in response to different preda- Books, New York. tors. Ecology 82 , 523–40. 208 REFERENCES

Relyea, R.A. (2001b). The lasting effects of adaptive plas- Ricklefs, R.E. and Wikelski, M. (2002). The physiology–life ticity: Predator-induced tadpoles become long-legged history nexus. Trends in Ecology and Evolution 17 , 462–68. frogs. Ecology 82 , 1947–55. Ricklefs, R.E., Konarzewski, M., and Daan, S. (1996). The Relyea, R.A. (2002a). Costs of phenotypic plasticity. relationship between basal metabolic rate and daily American Naturalist 159 , 272–82. energy expenditure in birds and mammals. American Relyea, R.A. (2002b). Competitor-induced plasticity in tad- Naturalist 147 , 1047–71. poles: Consequences, cues, and connections to predator- Ridley, M. (1994). The Red Queen: Sex and the evolution of induced plasticity. Ecological Monographs 72 , 523–40. human nature. Penguin Books, London. Relyea, R.A. (2003a). How prey respond to combined Riebesell, U., Zondervan, I., Rost, B., Tortell, P.D., Zeebe, predators: a review and an empirical test. Ecology 84 , R.E., and Morel, F.M.M. (2000). Reduced calcifi cation of 1827–39. marine plankton in response to increased atmospheric 407 Relyea, R.A. (2003b). Predators come and predators go: CO 2 . Nature , 364–67. the reversibility of predator-induced traits. Ecology 84 , Riessen, H.P. and Trevett-Smith, J.B. (2009). Turning induc- 1840–48. ible defenses on and off: Adaptive responses of Daphnia Relyea, R.A. (2005). The heritability of inducible defenses to a gape-limited predator. Ecology 90 , 3455–69. in tadpoles. Journal of Evolutionary Biology 18 , 856–66. Ripple, W.J. and Beschta, R.L. (2004). Wolves and the ecol- Relyea, R.A. (2007). Getting out alive: how predators affect ogy of fear: can predation risk structure ecosystems? the decision to metamorphose. Oecologia 152 , 389–400. Bioscience 54 , 755–66. Relyea, R.A. and Auld, J.R. (2004). Having the guts to com- Ripple, W.J. and Beschta, R.L. (2005). Linking wolves and pete: how intestinal plasticity explains the costs of plants: Aldo Leopold on trophic cascades. Bioscience 55 , inducible defences. Ecology Letters 7 , 869–75. 613–21. Relyea, R.A. and Auld, J.R. (2005). Predator- and competi- Rittweger, J. and Felsenberg, D. (2009). Recovery of mus- tor-induced plasticity: how changes in foraging morphol- cle atrophy and bone loss from 90 days bed rest: results ogy affect phenotypic trade-offs. Ecology 86 , 1723–29. from a one-year follow-up. Bone 44 , 214–24. Renaudeau, D. and Noblet, J. (2001). Effects of exposure to Robbins, C.T. (2001). Wildlife feeding and nutrition . Academic high ambient temperature and dietary protein level on Press, San Diego. sow milk production and performance of piglets. Journal Robert, V.A. and Casadevall, A. (2009). Vertebrate endo- of Animal Science 79 , 1540–48. thermy restricts most fungi as potential pathogens. Reneerkens, J., Piersma, T., and Sinninghe Damsté, J.S. Journal of Infectious Diseases 200 , 1623–26. (2002). Sandpipers (Scolopacidae) switch from mono to Roberts, W.M. (1996). Hummingbirds’ nectar concentra- diester preen waxes during courtship and incubation, tion preferences at low volume: the importance of time but why? Proceedings of the Royal Society, London, B 269 , scale. Animal Behaviour 52 , 361–70. 2135–39. Robin, J-P., Frain, M., Sardet, C., Groscolas, R., and Le Reneerkens, J., Piersma, T., and Sinninghe Damste, J.S. Maho, Y. (1988). Protein and lipid utilization during (2007). Expression of annual cycles in preen wax com- long-term fasting in emperor penguins. American Journal position in red knots: Constraints on the changing phe- of Physiology —Regulatory, Integrative and Comparative notype. Journal of Experimental Zoology A 307A, 127–39. Physiology 254 , 61–68. Ricklefs, R.E. (1969). Preliminary models for growth rates Robling, A.G., Hinant, F.M., Burr, D.B., and Turner, C.H. in altricial birds. Ecology 50 , 1031–39. (2002). Improved bone structure and strength after long- Ricklefs, R.E. (1991). Structures and transformations of life term mechanical loading is greatest if loading is sepa- histories. Functional Ecology 5 , 174–83. rated in short bouts. Journal of Bone and Mineral Research Ricklefs, R.E. (1998). Evolutionary theories of aging: con- 17 , 1545–54. fi rmation of a fundamental prediction, with implica- Rogers, D.I., Battley, P.F., Piersma, T., van Gils, J.A., and tions for the genetic basis and evolution of life span. Rogers, K.G. (2006). High tide habitat choice: Insights American Naturalist 152 , 22–44. from modelling roost selection by shorebirds around a Ricklefs, R.E. (2000). Intrinsic aging-related mortality in tropical bay. Animal Behaviour 72 , 563–75. birds. Journal of Avian Biology 31 , 103–11. Rogowitz, G.L. (1998). Limits to milk fl ow and energy allo- Ricklefs, R.E. (2008). The evolution of senescence from a cation during lactation of the hispid cotton rat (Sigmodon comparative perspective. Functional Ecology 22 , 379–92. hispidus ). Physiological Zoology 71 , 312–20. Ricklefs, R.E. and Scheuerlein, A. (2001). Comparison of Roll-Hansen, N. (1979). The genotype theory of Wilhelm aging-related mortality among birds and mammals. Johannsen and its relation to plant breeding and the Experimental Gerontology 36 , 845–57. study of evolution. Centaurus 22 , 201–35. REFERENCES 209

Rollo, C.D. (1995). Phenotypes: Their epigenetics, ecology and Sapolsky, R.M. (2004). Why zebras don´t get ulcers: The evolution . Chapman and Hall, London. acclaimed guide to stress, stress/related diseases, and coping , Rønning, B., Moe, B., and Bech, C. (2005). Long-term 3rd edition. Owl Books, New York. repeatability makes basal metabolic rate a likely herita- Sarkar, S. (2004). From the Reaktionsnorm to the evolution ble trait in the zebra fi nch Taeniopygia guttata . Journal of of adaptive plasticity. In Phenotypic plasticity: Functional Experimental Biology 208 , 4663–69. and conceptual approaches, eds T.J. DeWitt and S.M. Roseboom, T., de Rooij, S., and Painter, R. (2006). The Scheiner, pp. 10–30. Oxford University Press, New Dutch famine and its long-term consequences for adult York. health. Early Human Development 82 , 485–91. Savage, V.M., Gillooly, J.F., Woodruff, W.H., West, G.B., Rowland, J.M. and Emlen, D.J. (2009). Two thresholds, Allen, A.P., Enquist, B.J., and Brown, J.H. (2004). The three male forms result in facultative male trimorphism predominance of quarter-power scaling in biology. in beetles. Science 323 , 773–76. Functional Ecology 18 , 257–82. Royama, T. (1966). Factors governing feeding rate, food Schaeffer, P.J., Hokanson, J.F., Wells, D.J., and Lindstedt,

requirement and brood size of nestling great tits Parus S.L. (2001). Cold exposure increases running VO 2max and major . Ibis 108 , 313–47. cost of transport in goats. American Journal of Physiology Rubenstein, D.M. (2007). Stress hormones and sociality: 280 , R42–R47. Integrating social and environmental stressors. Schaeffer, P.J., Villarin, J.J., and Lindstedt, S.L. (2003). Proceedings of the Royal Sciety, London, B 274 , 967–75. Chronic cold exposure increases skeletal oxidative Rubner, M. (1908). Das Problem der Lebensdauer und seine structure and function in Monodelphis domestica, a mar- Beziehungen zum Wachstum und Ernahrung . Oldenburg, supial lacking brown adipose tissue. Physiological and München. Biochemical Zoology 76 , 877–87. Ruff, C., Holt, B., and Trinkaus, E. (2006). Who’s afraid of Schaeffer, P.J., Villarin, J.J., Pierotti, D.J., Kelly, D.P., and the big bad Wolff?: ‘Wolff’s Law’ and bone functional Lindstedt, S.L. (2005). Cost of transport is increased adaptation. American Journal of Physical Anthropology after cold exposure in Monodelphis domestica : Training 129 , 484–98. for ineffeiciency. Journal of Experimental Biology 208 , Rutten, A.L., Oosterbeek, K., Ens, B.J., and Verhulst, S. 3159–67. (2006). Optimal foraging on perilous prey: Risk of bill Schekkerman, H., Tulp, I., Piersma, T., and Visser, G.H. damage reduces optimal prey size in oystercatchers. (2003). Mechanisms promoting higher growth rates in Behavioral Ecology 17 , 297–302. arctic than temperate shorebirds. Oecologia 134 , 332–42. Ryan, R.T. and Semlitsch, R.D. (1998). Intraspecifi c hetero- Schekkerman, H., Teunissen, W., and Oosterveld, E. (2008). chrony and life history evolution: Decoupling somatic Mortality of black-tailed godwit Limosa limosa and and sexual development in a facultatively paedomor- northern lapwing Vanellus vanellus chicks in wet grass- phic salamander. Proceedings of the National Academy of lands: infl uence of predation and agriculture.Journal of Sciences, USA 95 , 5643–48. Ornithology 150 , 133–45. Sadowska, E.T., Baliga-Klimczyk, K., Chrząścik, K.M., and Schew, W.A. and Ricklefs, R.A. (1998). Developmental plas- Koteja, P. (2008). Laboratory model of adaptive radia- ticity. In Avian growth and development: evolution within the tion: a selection experiment in the bank vole. Physiological altricial-precocial spectrum, eds J.M. Starck and R.E. and Biochemical Zoology 81 , 627–40. Ricklefs, pp. 288–304. Oxford University Press, Oxford. Salvador, A. and Savageau, M.A. (2003). Quantitative evo- Schlichting, C.D. and Pigliucci, M. (1998). Phenotypic evolu- lutionary design of glucose 6-phosphate dehydrogenase tion: A reaction norm perspective . Sinauer, Sunderland. expression in human erythrocytes. Proceedings of the Schluter, D. (1981). Does the theory of optimal diets apply National Academy of Sciences, USA 100 , 14463–68. in complex environments? American Naturalist 118 , Sandoval, C.P. (1994). Plasticity in web design in the spi- 139–47. der Parawixia bistriata: a response to variable prey type. Schluter, D. (1993). Adaptive radiation in sticklebacks: Functional Ecology 8 , 701–707. size, shape and habitat use effi ciency. Ecology 74 , Sapir, N., Tsurim, I., Gal, B., and Abramsky, Z. (2004). The 699–709. effect of water availability on fuel deposition of two Schluter, D. (2000). The ecology of adaptive radiation . Oxford staging Sylvia warblers. Journal of Avian Biology 35 , University Press, Oxford. 25–32. Schmaljohann, H., Liechti, F., and Bruderer, B. (2007). Sapolsky, R.M. (1997). The trouble with testosterone, and other Songbird migration across the Sahara: The non-stop essays on the biology of the human predicament. Simon and hypothesis rejected! Proceedings of the Royal Society, Schuster, New York. London, B 274 , 735–39. 210 REFERENCES

Schmaljohann, H., Bruderer, B., and Liechti, F. (2008). rary selection leading to a decrease in sexual plumage Sustained bird fl ights occur at temperatures beyond dimorphism in a grassland-breeding shorebird. expected limits of water loss rates. Animal Behaviour 76 , Behavioral Ecology 20 , 797–807. 1133–38. Schubert, K.A. (2009). Synthesis: environmental quality, Schmaljohann, H., Liechti, F., and Bruderer, B. (2009). energy allocation and the cost of reproduction. In Trans-Sahara migrants select fl ight altitudes to mini- Breeding on a budget: Fundamental links between energy mize energy costs rather than water loss. Behavioral metabolism and mammalian life history trade-offs , pp. 120–32. Ecology and Sociobiology 63 , 1609–19. PhD thesis, University of Groningen, The Netherlands. Schmid-Hempel, P. and Ebert, D. 2003. On the evolution- Schubert, K.A., de Vries, G., Vaanholt, L.M., Meijer, H., ary ecology of specifi c immune defence. Trends in Ecology Daan, S., and Verhulst, S. (2009). Maternal energy allo- and Evolution 18 , 27–32. cation to offspring increases with environmental quality Schmid-Hempel, P., Kacelnik, A., and Houston, A.I. (1985). in house mice. American Naturalist 173 , 831–40. Honeybees maximize effi ciency by not fi lling their crop. Schwilch, R., Piersma, T., Holmgren, N.M.A., and Jenni, L. Behavioral Ecology and Sociobiology 17 , 61–66. (2002). Do migratory birds need a nap after a long non- Schmidt-Nielsen, K. (1984). Scaling: Why is animal size so stop fl ight? Ardea 90 , 149–54. important? Cambridge University Press, Cambridge. Scott, I. and Evans, P.R. (1992). The metabolic output of Schmidt-Wellenburg, C.A., Biebach, H., Daan, S., and avian ( Sturnus vulgaris , Calidris alpina) adipose tissue, Visser, G.H. (2006). Energy expenditure and wing beat liver and skeletal muscle: Implications for BMR/body frequency in relation to body mass in free fl ying swal- mass relationships. Comparative Biochemistry and lows ( Hirundo rustica ). Journal of Comparative Physiology Physiology A 103 , 329–32. B 177 , 327–37. Scott, G.R., Egginton, S., Richards, J.G., and Milsom, W.K. Schmidt-Wellenburg, C., Engel, S., and Visser, G.H. (2008). (2009). Evolution of muscle phenotype for extreme high Energy expenditure during fl ight in relation to body altitude fl ight in the bar-headed goose.Proceedings of the mass: Effects of natural increases in mass and artifi cial Royal Society, London, B 276 , 3645–53. load in rose coloured starlings. Journal of Comparative Secor, S.M. (2008). Digestive physiology of the Burmese Physiology B 178 , 767–77. python: broad regulation of integrated performance. Schmitz, O.J., Krivan, V., and Ovadia, O. (2004). Trophic Journal of Experimental Biology 211 , 3767–74. cascades: The primacy of trait-mediated indirect inter- Secor, S.M. and Diamond, J.M. (1995). Adaptive responses actions. Ecology Letters 7 , 153–63. to feeding in Burmese pythons: pay before pumping. Schoeppner, N.M. and Relyea, R.A. (2005). Damage, diges- Journal of Experimental Biology 198 , 1313–25. tion, and defence: the role of alarm cues and kairomones Secor, S.M., Stein, E.D., and Diamond, J.M. (1994). Rapid for inducing prey defences. Ecology Letters 8 , 505–12. up-regulation of snake intestine in response to feeding: Schoeppner, N.M. and Relyea, R.A. (2008). Detecting small a new model of intestinal adaptation. American Journal of environmental differences: Risk-response curves for Physiology 266 , G695–G705. predator-induced behavior and morphology. Oecologia Sedinger, J.S. (1997). Adaptations to and consequences of 154 , 743–54. an herbivorous diet in grouse and waterfowl. Condor 99, Scholander, P.F. (1990). Enjoying a life in science: The autobi- 314–26. ography of P.F. Scholander. University of Alaska Press, Selman, C., McLaren, J.S., Collins, A.R., Duthie, G.G., and Fairbanks. Speakman, J.R. (2008). The impact of experimentally Scholander, P.F., Hock, R., Walters, V., Johnson, F., and Irving, elevated energy expenditure on oxidative stress and L. (1950a). Heat regulation in some Arctic and tropical lifespan in the short-tailed fi eld vole Microtus agrestis . mammals and birds. Biological Bulletin 99 , 237–58. Proceedings of the Royal Society, London, B 275 , 1907–16. Scholander, P.F., Walters, V., Hock, R., and Irving, L. Selvin, S. (1975). On the Monty Hall problem. American (1950b). Body insulation of some Arctic and tropical Statistician 29 , 134. mammals and birds. Biological Bulletin 99 , 225–36. Seppänen, J-T., Forsman, J.T., Mönkkönen, M., and Scholander, P.F., Hock, R., Walters, V., and Irving, L. Thomson, R.L. (2007). Social information use is a proc- (1950c). Adaptation to cold in arctic and tropical mam- ess across time, space, and ecology, reaching heterospe- mals and birds in relation to body temperature, insula- cifi cs. Ecology 88 , 1622–33. tion and basal metabolic rate. Biological Bulletin 99 , Shapiro, A.M. (1976). Seasonal polyphenism. Evolutionary 259–71. Biology 9 , 259–333. Schroeder, J., Lourenço, P.M., Hooijmeijer, J.C.E.W., Both, Sheriff, M.J., Krebs, C.J., and Boonstra, R. (2009). The sen- C., and Piersma, T. (2009). A possible case of contempo- sitive hare: sublethal effects of predator stress on repro- REFERENCES 211

duction in snowshoe hares. Journal of Animal Ecology 78 , Speakman, J.R. (2000). The cost of living: fi eld metabolic 1249–58. rates of small mammals. Advances in Ecological Research Shettleworth, S.J. (1998). Cognition, evolution and behaviour . 30 , 177–297. Oxford University Press, Oxford. Speakman, J.R. (2005). Body size, energy metabolism and Shubin, N. (2008). Your inner fi sh: A journey into the 3.5-bil- lifespan. Journal of Experimental Biology 208 , 1717–30. lion-year history of the human body. Vintage Books, Speakman, J.R. (2008). The physiological costs of repro- New York. duction in small mammals. Philosophical Transactions of Sih, A. and Christensen, B. (2001). Optimal diet theory: the Royal Society, London, B 363 , 375–98. when does it work, and when and why does it fail? Speakman, J.R. and Johnson, M.S. (2000). Relationships Animal Behaviour 61 , 379–90. between resting metabolic rate and morphology in lactat- Simmons, L.W. and Emlen, D.J. (2006). Evolutionary trade- ing mice: what tissues are the major contributors to rest- off between weapons and testes. Proceedings of the ing metabolism? In Living in the cold, eds G. Heldmaier National Academy of Sciences, USA 103 , 16346–51. and M. Klingenspor, Vol 3, pp. 479–86. Springer, Berlin. Sinervo, B. (1999). Mechanistic analysis of natural selec- Speakman, J.R. and Król, E. (2005). Limits to sustained tion and a refi nement of Lack’s and Williams’s princi- intake. IX. A review of hypotheses. Journal of Comparative ples. American Naturalist 154 , S26–S42. Physiology B 175 , 375–94. Sinervo, B. and Basolo, A.L. (1996). Testing adaptation Speakman, J.R., Gidney, A., Bett, J., Mitchell, I.P., and using phenotypic manipulations. In Adaptation, eds Johnson, M.S. (2001). Limits to sustained intake. IV. M.R. Rose and G.V. Lauder, pp. 149–85. Academic Press, Effect of variation in food quality on lactating mice Mus San Diego. musculus . Journal of Experimental Biology 204 , 1957–65. Sjodin, A.M., Anderson, A.B., Hogberg, J.M., and Speakman, J.R., Selman, C., McLaren, J., and Harper, E.J. Westerterp, K.R. (1994). Energy balance in cross-country (2002). Living fast, dying when? The link between aging skiers: a study using doubly labelled water. Medicine and energetics. Journal of Nutrition 132 , 1583–97. and Science in Sports and Exercise 26 , 720–24. Speakman, J.R., Król, E., and Johnson, M.S. (2004a). The Smith, G.T., Brenowitz, E.A., and Wingfi eld, F.C. (1997). functional signifi cance of individual variation in basal Roles of photoperiod and testosterone in seasonal plas- metabolic rate. Physiological and Biochemical Zoology 77 , ticity of the avian song control system. Journal of 900–15. Neurobiology 32 , 426–42. Speakman, J.R., Talbot, D.A., Selman, C., Snart, S., Smith, J.W., Benkman, C.W., and Coffey, K. (1999). The use McLaren, J.S., Redman, P., Król, E., Jackson, D.M., and misuse of public information by foraging red cross- Johnson, M.S., and Brand, M.D. (2004b). Uncoupling bills. Behavioral Ecology 10 , 54–62. and surviving: Individual mice with higher metabolism Solomon, S. (2001). The coldest march: Scott’s fatal Antarctic have higher mitochondrial uncoupling and live longer. expedition . Yale University Press, London. Aging Cell 3 , 87–95. Sonnenfeld, G. (2005). The immune system in space, Staaland, H. (1967). Anatomical and physiological adapta- including earth-based benefi ts of space-based research. tions of nasal glands in Charadriiformes birds. Current Pharmaceutical Biotechnology 6 , 343–49. Comparative Biochemistry and Physiology 23 , 933–44. Spalding, K.L., Bhardwaj, R.D., Buchholz, B.A., Druid, H., Stamper, C.E., Downie, J.T., Stevens, D.J., and Monaghan, and Frisén, J. (2005a). Retrospective birth dating of cells P. (2009). The effects of perceived predation risk on pre- in humans. Cell 122 , 133–43. and post-metamorphic phenotypes in the common frog. Spalding, K.L., Buchholz, B.A., Bergman, L-E., Druid, H., Journal of Zoology 277 , 205–13. and Frisén, J. (2005b). Age written in teeth by nuclear Stanley, S.M. (1970). Relation of shell form to life habits of tests. Nature 437 , 333–34. the Bivalvia (Mollusca). Geological Society of America Spalding, K.L., Arner, E., Westermark, P.O., Bernard, S., emoirs 125 , 1–296. Buchholz, B.A., Bergmann, O., Blomqvist, L., Hoffstedt, Starck, J.M. (2003). Shaping up: how vertebrates adjust J., Näslund, E., Britton, T., Concha, H., Hassan, M., their digestive system to changing environmental con- Rydén, M., Frisén, J., and Arner, P. (2008). Dynamics of ditions. Animal Biology 53 , 245–57. fat cell turnover in humans. Nature 453 , 783–87. Starck, J.M. (2009). Functional morphology and patterns Speakman, J.R. (1997a). Doubly labelled water: Theory and of blood flow in the heart of Python regius . Journal of practice . Chapman and Hall, London. Morphology 270 , 673–87. Speakman, J.R. (1997b). Factors infl uencing the daily Starck, J.M. and Beese, K. (2001). Structural fl exibility of energy expenditure of small mammals. Proceedings of the the intestine of Burmese python in response to feeding. Nutrition Society 56 , 1119–36. Journal of Experimental Biology 201 , 325–35. 212 REFERENCES

Starck, J.M. and Wimmer, C. (2005). Patterns of blood fl ow inadequate energy intake. European Journal of Applied during the postprandial response in ball pythons Python Physiology 76 , 243–52. regius . Journal of Experimental Biology 208 , 881–89. Stuart-Fox, D., Moussalli, A., and Whiting, M.J. (2008). Starck, J.M., Dietz, M.W., and Piersma, T. (2001). The Predator-specifi c camoufl age in chameleons. Biology assessment of body composition and other parameters Letters 4 , 326–29. by ultrasound scanning. In Body composition analysis of Suarez, R.K. (1992). Hummingbird fl ight: sustaining the animals: A handbook of non-destructive methods, ed. J.R. highest mass-specifi c metabolic rates among verte- Speakman, pp. 188–210. Cambridge University Press, brates. Experientia 48 , 565–70. Cambridge. Suarez, R.K. (1998). Oxygen and the upper limits to ani- Starck, J.M., Cruz-Neto, A.P., and Abe, A.S. (2007). mal design and performance. Journal of Experimental Physiological and morphological responses to feeding Biology 201 , 1065–72. in broad-nosed caiman ( Caiman latirostris ). Journal of Suarez, R.K. and Darveau, C-A. (2005). Multi-level regula- Experimental Biology 210 , 2033–45. tion and metabolic scaling. Journal of Experimental Biology Stearns, S.C. (1989). Trade-offs in life-history evolution. 208 , 1627–34. Functional Ecology 3 , 259–68. Suarez, R.K., Darveau, C-A., and Childress, J.J. (2004). Stearns, S.C. (1992). The evolution of life histories . Oxford Metabolic scaling: a many-splendoured thing. University Press, New York. Comparative Biochemistry and Physiology B 139 , 531–41. Stein, T.P. and Wade, C.E. (2005). Metabolic consequences Sutherland, W.J. (1987). Why do animals specialize. Nature of muscle disuse atrophy. Journal of Nutrition 135 , 325 , 483–84. 1824S–1828S. Sutherland, W.J. and Anderson, C.W. (1993). Predicting Stephens, D.W. and Krebs, J.R. (1986). Foraging theory . the distribution of individuals and the consequences of Princeton University Press, Princeton, NJ. habitat loss: the role of prey depletion. Journal of Sterner, R.W. and Elser, J.J. (2002). Ecological stoichiometry: Theoretical Biology 160 , 223–30. The biology of elements from molecules to the biosphere . Sutherland, W.J., Ens, B.J., Goss-Custard, J.D., and Princeton University Press, Princeton, NJ. Hulscher, J.B. (1996). Specialization. In The oystercatcher: Sterner, R.W. and Hessen, D.O. (1994). Algal nutrient limi- From individuals to populations, ed. J.D. Goss-Custard, tation and the nutrition of aquatic herbivores. Annual pp. 56–76. Oxford University Press, Oxford. Review of Ecology and Systematics 25 , 1–29. Sutton, J.R., Reeves, J.T., Wagner, P.D., Groves, B.M., Stevens, C.E. and Hume, I.D. (1995). Comparative physiol- Cymerman, A., Malconian, M.K., Rock, P.B., Young, P.M., ogy of the vertebrate digestive system , 2nd edition. Walter, S.D., and Houston, C.S. (1988). Operation Everest Cambridge University Press, Cambridge. II: Oxygen transport during exercise at extreme simu- Stroud, M.A. (1987a). Nutrition and energy balance on the lated altitude. Journal of Applied Physiology 64 , 1309–21. ‘footsteps of Scott’ expedition 1984–86. Human Nutrition: Swanson, D.L. (1993). Cold tolerance and thermogenic Applied Nutrition 41A , 426–33. capacity in dark-eyed juncos in winter: Geographic var- Stroud, M.A. (1987b). Did Scott simply starve? Lessons iation and comparison with American tree sparrows. from ‘The footsteps of the Scott Antarctic expedition’. Journal of Thermal Biology 18 , 275–81. Proceedings of the Nutrition Society 46 , A129. Swanson, D.L. and Liknes, E.T. (2006). A comparative Stroud, M. (1998). The nutritional demands of very pro- analysis of thermogenic capacity and cold tolerance in longed exercise in man. Proceedings of the Nutrition small birds. Journal of Experimental Biology 209 , 466–74. Society 57 , 55–61. Swart, J.A.A. and van Andel, J. (2008). Rethinking the Stroud, M. (2004). Survival of the fi ttest: Anatomy of peak interface between ecology and society: the case of the physical performance . Vintage, London. cockle controversy in the Dutch Wadden Sea. Journal of Stroud, M.A., Coward, W.A., and Sawyer, M.B. (1993). Applied Ecology 45 , 82–90. Measurements of energy expenditure using isotope-la- Swennen, C., Leopold, M.F., and de Bruijn, L.L.M. (1989). 2 18 belled water ( H 2 O) during an Arctic expedition. Time-stressed oystercatchers, Haematopus ostralegus , can European Journal of Applied Physiology 67 , 375–79. increase their intake rate. Animal Behaviour 38 , 8–22. Stroud, M.A., Ritz, P., Coward, W.A., Sawyer, M.B., Tammaru, T. and Haukioja, E. (1996). Capital breeders and Constantin-Teodosiu, D., Greenhaff, P.L. and Macdonald, income breeders among Lepidoptera—consequences to I.A. (1997). Energy expenditure using isotope-labelled population dynamics. Oikos 77 , 561–64. 2 18 water ( H 2 O), exercise performance, skeletal muscle Taylor, C.R. and Weibel, E.R. (1981). Design of the mam- enzyme activities and plasma biochemical parameters malian respiratory system. I. Problem and strategy. in humans during 95 days of endurance exercise with Respiratory Physiology 44 , 1–10. REFERENCES 213

Taylor, C.R., Maloiy, G.M.O., Weibel, E.R., Langman, V.A., Tinbergen, N. (1963). On the aims and methods of ethol- Kamau, J.M., Seeherman, H.J., and Heglund, N.C. ogy. Zeitschrift für Tierpsychologie 20 , 410–33. (1981). Design of the mammalian respiratory system. III. Tinbergen, J.M. (1981). Foraging decisions in starlings Scaling maximum aerobic capacity to body mass: Wild (Sturnus vulgaris L.). Ardea 69 , 1–67. and domestic animals. Respiratory Physiology 44 , 25–37. Tinbergen, J.M. and Dietz, M.W. (1994). Parental energy Taylor, C.R., Weibel, E.R., Weber, J-M., Vock, R., Hoppeler, expenditure during brood rearing in the great tit (Parus H., Roberts, T.J., and Bricon, G. (1996). Design of the major ) in relation to body mass, temperature, food avail- oxygen and substrate pathways. I. Model and strategy ability and clutch size. Functional Ecology 8 , 563–72. to test symmorphosis in a network structure. Journal of Tinbergen, J.M. and Verhulst, S. (2000). A fi xed energetic Experimental Biology 199 , 1643–49. ceiling to parental effort in the great tit? Journal of Animal Templeton, J.J. and Giraldeau, L-A. (1996). Vicarious sam- Ecology 69 , 323–34. pling: the use of personal and public information by Tinbergen, J.M. and Williams, J.B. (2002). Energetics of starlings foraging in a simple patchy environment. incubation. In Avian incubation: Behaviour, environment Behavioral Ecology and Sociobiology 38 , 105–14. and evolution , ed. D.C. Deeming, pp. 299–313. Oxford Teplitsky, C. and Laurila, A. (2007). Flexible defense strate- University Press, Oxford. gies: competition modifi es investment in behavioral vs. Tinbergen, N., Impekoven, M., and Franck, C.K. (1967). morphological defenses. Ecology 88 , 1641–46. An experiment on spacing-out as a defence against pre- Teplitsky, C., Plénet, S., Léna, J-P., Mermet, N., Malet, E., dation. Behaviour 28 , 307–20. and Joly, P. (2005a). Escape behaviour and ultimate Todd, P.A. (2008). Morphological plasticity in scleractinian causes of specfi c induced defences in an anuran tadpole. corals. Biological Reviews 83 , 315–37. Journal of Evolutionary Biology 18 , 180–90. Tollrian, R. and Dodson, S.I. (1999). Inducible defenses in Teplitsky, C., Plénet, S., and Joly, P. (2005b). Costs and lim- Cladocera: Constraints, costs, and multipredator envi- its of dosage response to predation risk: To what extent ronments. In The ecology and evolution of inducible defenses, can tadpoles invest in anti-predator morphology? eds R. Tollrian and C.D. Harvell, pp. 177–202. Princeton Oecologia 145 , 364–70. University Press, Princeton, NJ. Teplitsky, C., Mills, J.A., Alho, J.S., Yarrall, J.W., and Tollrian, R. and Harvell, C.D. eds (1999). The ecology and Merilä, J. (2008). Bergmann’s rule and climate change evolution of inducible defenses . Princeton University Press, revisited: disentangling environmental and genetic Princeton, NJ. responses in a wild bird population. Proceedings of the Toloza, E.M., Lam, M., and Diamond, J.M. (1991). Nutrient National Academy of Sciences, USA 105 , 13492–96. extraction by cold-exposed mice: A test of digestive safety Tesch, P.A., Trieschmann, J.T., and Ekberg, A. (2004). margins. American Journal of Physiology 261 , G608–20. Hypertrophy of chronically unloaded muscle subjected Tomkovich, P.S. (2001). A new subspecies of red knot to resistance exercise. Journal of Applied Physiology 96 , Calidris canutus from the New Siberian Islands. Bulletin 1451–58. British Ornithologists’ Club 121 , 257–63. Thompson, G.E. (1973). Review of the progress of dairy Tomkovich, P.S. and Soloviev, M.Y. (1996). Distribution, sciences: climatic physiology of cattle. Journal of Dairy migrations and biometrics of knots Calidris canutus Research 40 , 441–73. canutus on Taimyr, Siberia. Ardea 84 , 85–98. Tieleman, B.I. and Williams, J.B. (1999). The role of hyper- Touchon, J.C. and Warkentin, K.M. (2008). Fish and drag- thermia in the water economy of desert birds. onfl y nymph predators induce opposite shifts in color Physiological and Biochemical Zoology 72 , 87–100. and morphology of tadpoles. Oikos 117 , 634–40. Tieleman, B.I. and Williams, J.B. (2000). The adjustment of Tramontin, A.D., Hartman, V.N., and Brenowitz, E.A. avian metabolic rates and water fl uxes to desert envi- (2000). Breeding conditions induce rapid and sequential ronments. Physiological and Biochemical Zoology 73 , growth in adult avian song control circuits: a model of 461–79. seasonal plasticity in the brain. Journal of Neuroscience Tieleman, B.I., Williams, J.B., and Buschur, M.E. (2002). 20 , 854–61. Physiological adjustments to arid and mesic environ- Trappe, S., Costill, D., Gallagher, P., Creer, A., Peters, J.R., ments in larks (Alaudidae). Physiological and Biochemical Evans, H., Riley, D.A., and Fitts, R.H. (2009). Exercise in Zoology 75 , 305–13. space: human skeletal muscle after 6 months aboard the Tieleman, B.I., Williams, J.B., Ricklefs, R.E., and Klasing, International Space Station. Journal of Applied Physiology K.C. (2005). Constitutive innate immunity is a compo- 106 , 1159–68. nent of the pace-of-life syndrome in tropical birds. Trussell, G.C. and Smith, L.D. (2000). Induced defenses in Proceedings of the Royal Society, London, B 272 , 1715–20. response to an invading crab predator: An explanation 214 REFERENCES

of historical and geographic phenotypic change. Van Buskirk, J. and Relyea, R.A. (1998). Selection for phe- Proceedings of the National Academy of Sciences, USA 97 , notypic plasticity in Rana sylvatica tadpoles. Biological 2123–27. Journal of the Linnean Society 65 , 301–28. Tsurim, I., Sapir, N., Belmaker, J., Shanni, I., Izhaki, I., van de Kam, J., Ens, B. J., Piersma, T., and Zwarts, L. (2004). Wojciechowski, M.S., Karasov, W.H., and Pinshow, B. Shorebirds: An illustrated behavioural ecology . KNNV (2008). Drinking water boosts food intake rate, body Publishers, Utrecht. mass increase and fat accumulation in migratory black- van de Kam, J., Battley, P.F., McCaffery, B.J., Rogers, D.I., caps (Sylvia atricapilla ). Oecologia 156 , 21–30. Jae-Sang, Hong, Moores, N., Ju-Yong, Ki, Lewis, J., and Tulp, I., McChesney, S., and de Goeij, P. (1994). Migratory Piersma, T. (2008). Invisible connections: Why migrating departures of waders from north-western Australia: shorebirds need the Yellow Sea . Wetlands International, behaviour, timing and possible migration routes. Ardea Wageningen. 82 , 201–21. van den Hout, P.J., Piersma, T., Dekinga, A., Lubbe, S.K., Tulp, I., Schekkerman, H., Bruinzeel, L.W., Jukema, J., and Visser, G.H. (2006). Ruddy turnstones Arenaria Visser, G.H., and Piersma, T. (2009). Energetic demands interpres rapidly build pectoral muscle after raptor during incubation and chick-rearing in a uniparental and scares. Journal of Avian Biology 37 , 425–30. a biparental arctic breeding shorebird. Auk 126 , 155–64. van den Hout, P.J., Spaans, B., and Piersma, T. (2008). Turner, J.S. (2007). The tinkerer’s accomplice: How design Differential mortality of wintering shorebirds on the emerges from life itself . Harvard University Press, Banc d’Arguin, Mauritania, due to predation by large Cambridge, Mass. falcons. Ibis 150 (Suppl. 1), 219–30. Underwood, E.J. (1977). Trace elements in human and animal van den Hout, P.J., Mathot, K.J., Maas, L.R.M., and nutrition . Academic Press, New York. Piersma, T. (2010). Predator escape tactics in birds: Urabe, J., Clasen, J., and Sterner, R.W. (1997). Phosphorus Linking ecology and aerodynamics. Behavioral Ecology limitation of Daphnia growth: Is it real? Limnology and 21 , 16–25. Oceanography 42 , 1436–43. van de Pol, M., Ens, B.J., Oosterbeek, K., Brouwer, L., Vaanholt, L.M., Daan, S., Schubert, K.A., and Visser, G.H. Verhulst, S., Tinbergen, J.M., Rutten, A.L. and de Jong, (2009). Metabolism and aging: effects of cold exposure M. (2009). Oystercatchers’ bill shapes as a proxy for diet on metabolic rate, body composition and longevity in specialization: More differentiation than meets the eye. mice. Physiological and Biochemical Zoology 82 , 314–24. Ardea 97 , 335–47. Vahl, W.K., van der Meer, J., Weissing, F.J., van Dullemen, van der Meer, J. and Ens, B.J. (1997). Models of interfer- D., and Piersma, T. (2005). The mechanisms of interference and their consequences for the spatial distribution ence competition: two experiments on foraging waders. of ideal and free predators. Journal of Animal Ecology 66 , Behavioral Ecology 16 , 845–55. 846–58. Valencak, T.G. and Ruf, T. (2009). Energy turnover in Vandermeer, J.H. and Goldberg, D.E. (2003). Population European hares is centrally limited during early, but not ecology: First principles . Princeton University Press, during peak lactation. Journal of Comparative Physiology Princeton, NJ. B 179 , 933–43. van der Meer, J. and Piersma, T. (1994). Physiologically Valencak, T.G., Tataruch, F., and Ruf, T. (2009). Peak energy inspired regression models for estimating and predict- turnover in lactating European hares: the role of fat ing nutrient stores and their composition in birds. reserves. Journal of Experimental Biology 212 , 231–37. Physiological Zoology 67 , 305–29. Valladares, F., Gianoli, E., and Gomez, J.M. (2007). van der Merwe, M. and Brown, J.S. (2008). Mapping the Ecological limits to plant phenotypic plasticity. New landscape of fear of the Cape ground squirrel. Journal of Phytologist 176 , 749–63. Mammalogy 89 , 1162–69. Valone, T.J. (1993). Patch information and estimation: a van Gils, J.A. (2010). State-dependent Bayesian foraging cost of group foraging. Oikos 68 , 258–66. on spatially autocorrelated food distributions. Oikos Valone, T.J. (2006). Are animals capable of Bayesian updat- 119 , 237–44. ing? An empirical review. Oikos 112 , 252–59. van Gils, J.A. and Piersma, T. (2004). Digestively con- Valone, T.J. (2007). From eavesdropping on performance to strained predators evade the cost of interference compe- copying the behaviour of others: A review of public infor- tition. Journal of Animal Ecology 73 , 386–98. mation use. Behavioral Ecology and Sociobiology 62 , 1–14. van Gils, J.A. and Tijsen, W. (2007). Short-term foraging Valone, T.J. and Giraldeau, L-A. (1993). Patch estimation costs and long-term fueling rates in central-place forag- by group foragers: what information is used? Animal ing swans revealed by giving-up exploitation times. Behaviour 45 , 721–28. American Naturalist 169 , 609–20. REFERENCES 215 van Gils, J.A., Piersma, T., Dekinga, A., and Dietz, M.W. and de Vries, P.P. (2008). Longer guts and higher food (2003a). Cost–benefi t analysis of mollusc-eating in a quality increase energy intake in migratory swans. shorebird. II. Optimizing gizzard size in the face of sea- Journal of Animal Ecology 77 , 1234–41. sonal demands. Journal of Experimental Biology 206 , van Gils, J.A., Kraan, C., Dekinga, A., Koolhaas, A., 3369–80. Drent, J., de Goeij, P., and Piersma, T. (2009). Reversed van Gils, J.A., Schenk, I.W., Bos, O., and Piersma, T. optimality and predictive ecology: burrowing depth (2003b). Incompletely informed shorebirds that face a forecasts population change in a bivalve. Biology Letters digestive constraint maximize net energy gain when 5 , 5–8. exploiting patches. American Naturalist 161 , 777–93. van Roomen, M., van Turnhout, C., van Winden, E., Koks, van Gils, J.A., Edelaar, P., Escudero, G., and Piersma, T. B., Goedhart, P., Leopold, M., and Smit, C. (2005). Trends (2004). Carrying capacity models should not use fi xed in benthivorous waterbirds in the Dutch Wadden Sea prey density thresholds: a plea for using more tools of 1975-2002: large differences between shellfi sh-eaters behavioural ecology. Oikos 104 , 197–204. and worm-eaters. Limosa 78 , 21–38. van Gils, J.A., de Rooij, S.R., van Belle, J., van der Meer, J., Van Valen, L. (1973). A new evolutionary law. Evolutionary Dekinga, A., Piersma, T., and Drent, R. (2005a). Digestive Theory 1 , 1–30. bottleneck affects foraging decisions in red knots Calidris Vermeij, G.J. (1987). Evolution and escalation: An ecological canutus. I. Prey choice. Journal of Animal Ecology 74 , history of life . Princeton University Press, Princeton, NJ. 105–19. Vermeij, G.J. (1993). A natural history of shells . Princeton van Gils, J.A., Dekinga, A., Spaans, B., Vahl, W.K., and University Press, Princeton, NJ. Piersma, T. (2005b). Digestive bottleneck affects forag- Vézina, F. and Williams, T.D. (2005). The metabolic cost of ing decisions in red knots Calidris canutus. II. Patch egg production is repeatable. Journal of Experimental choice and length of working day. Journal of Animal Biology 208 , 2533–38. Ecology 74 , 120–30. Vézina, F., Jalvingh, K.M., Dekinga, A., and Piersma, T. van Gils, J.A., Battley, P.F., Piersma, T., and Drent, R. (2006). Acclimation to different thermal conditions in a (2005c). Reinterpretation of gizzard sizes of red knots northerly wintering shorebird is driven by body mass- world-wide emphasises overriding importance of prey related changes in organ size. Journal of Experimental quality at migratory stopover sites. Proceedings of the Biology 209 , 3141–54. Royal Society, London, B 272 , 2609–18. Vézina, F., Jalvingh, K.M., Dekinga, A., and Piersma, T. van Gils, J.A., Piersma, T., Dekinga, A., and Battley, P.F. (2007). Thermogenic side-effects to migratory predispo- (2006a). Modelling phenotypic fl exibility: An optimality sition in shorebirds. American Journal of Physiology 292, analysis of gizzard size in red knots (Calidris canutus ). R1287–97. Ardea 94, 409–20. Vézina, F., Dekinga, A., and Piersma, T. (2010). Phenotypic van Gils, J.A., Piersma, T., Dekinga, A., Spaans, B., and compromise in the face of confl icting ecological Kraan, C. (2006b). Shellfi sh-dredging pushes a flexible demands: an example in red knots Calidris canutus . avian top predator out of a protected marine ecosystem. Journal of Avian Biology 41, 88–93. Public Library of Science —Biology 4 , 2399–2404. Via, S. and Lande, R. (1985). Genotype-environment inter- van Gils, J.A., Spaans, B., Dekinga, A., and Piersma, T. action and the evolution of phenotypic plasticity. (2006c). Foraging in a tidally structured environment by Evolution 39 , 505–22. red knots (Calidris canutus ): Ideal, but not free. Ecology Via, S., Gomulkiewicz, R., de Jong, G., Scheiner, S.M., 87 , 1189–1202. Schlichting, C.D., and van Tienderen, P.H. (1995). van Gils, J.A., Dekinga, A., van den Hout, P.J., Spaans, B., Adaptive phenotypic plasticity: concensus and contro- and Piersma, T. (2007a). Digestive organ size and behav- versy. Trends in Ecology and Evolution 10 , 212–17. ior of red knots (Calidris canutus ) indicate the quality of Vico, L., Lafage-Proust, M-H., and Alexandre, C. (1998). their benthic food stocks. Israel Journal of Ecology and Effects of gravitational changes on the bone system in Evolution 53 , 329–46. vitro and in vivo . Bone 22 , 95S–100S. van Gils, J.A., Munster, V.J., Radersma, R., Liefhebber, D., Vico, L., Collet, P., Guignandon, A., Lafage-Proust, M-H., Fouchier, R.A.M., and Klaassen, M. (2007b). Hampered Thomas, T., Rehailia, M., and Alexandre, C. (2000). foraging and migratory performance in swans infected Effects of long-term microgravity exposure on cancel- with low-pathogenic avian infl uenza A virus. Public lous and cortical weight-bearing bones of cosmonauts. Library of Science One 2 , e184. Lancet 355 , 1607–11. van Gils, J.A., Beekman, J.H., Coehoorn, P., Corporaal, E., Videler, J.J. (2005). Avian fl ight. Oxford University Press, Dekkers, T., Klaassen, M., van Kraaij, R., de Leeuw, R., Oxford. 216 REFERENCES

Vince, G. (2006). Your amazing regenerating body/The Ward, P. and Zahavi, A. (1973). The importance of certain many ages of man. New Scientist 2556 , 15. assemblages of birds as ‘information centres’ for fi nding Visser, G.H., Dekinga, A., Achterkamp, B., and Piersma, T. food. Ibis 115 , 517–34. (2000). Ingested water equilibrates isotopically with the Ward, S., Moller, U., Rayner, J.M.V., Jackson, D.M., body water pool of a shorebird with unrivaled water Nachtigall, W., and Speakman, J.R. (2004). Metabolic fl uxes. American Journal of Physiology —Regulatory, power of European starlings Sturnus vulgaris during Integrative and Comparative Physiology 279 , R1795–1804. fl ight in a wind tunnel, estimated from heat transfer Visser, M.E., Adriaensen, F., van Balen, J.H., Blondel, J., modelling, doubly labelled water and mask respirome- Dhondt, A.A., van Dongen, S., du Feau, C., Ivankina, E.V., try. Journal of Experimental Biology 207 , 4291–98. Kerimov, A.B., De Laet, J., Matthysen, E., McCleery, R.H., Waters, W.W., Ziegler, M.G., and Meck, J.V. (2002). Orell, M., and Thomson, D.L. (2003). Variable responses to Postspacefl ight orthostatic hypotension occurs mostly large-scale climate change in European Parus populations. in women and is predicted by low vascular resistance. Proceedings of the Royal Society, London, B 270 , 367–72. Journal of Applied Physiology 92 , 586–94. Visser, M.E., Both, C., and Lambrechts, M.M. (2004). Global Watling, L. and Norse, E.A. (1998). Disturbance of the sea- climate change leads to mistimed avian reproduction. bed by mobile fi shing gear: a comparison to forest clear- Advances in Ecological Research 35 , 89–110. cutting. Conservation Biology 12 , 1180–97. Vitousek, P.M., Mooney, H.A., Lubchenko, J., and Melillo, Watt, W.B. (2000). Avoiding paradigm-based limits to J.M. (1997). Human domination of earth’s ecosystems. knowledge of evolution. Evolutionary Biology 32 , 73–96. Science 277 , 494–99. Weathers, W.W. (1979). Climatic adaptation in avian stand- Vleck, C.M. and Vleck, D. (1979). Metabolic rate in 5 bird ard metabolic rate. Oecologia 42 , 81–89. species. Condor 81 , 89–91. Weathers, W.W. (1997). Energetics and thermoregulation Vogel, S. (2001). Prime mover: A natural history of muscle . by small passerines of the humid, lowland tropics. Auk Norton, New York. 114 , 341–53. Vogel, S. (2003). Comparative biomechanics: Life’s physical Weber, T.P., Bergudd, J.J., Hedenström, A., Persson, K., world . Princeton University Press, Princeton, NJ. and Sandberg, G. (2005). Resistance of fl ight feathers to Vogt, G., Huber, M., Thiemann, M., van den Boogaart, G., mechanical fatigue covaries with moult strategy in two Schmitz, O.J., and Schubart, C.D. (2008). Production of warbler species. Biology Letters 1 , 27–30. different phenotypes from the same genotype in the Weibel, E.R. (1985). Design and performance of muscular same environment by developmental variation. Journal systems: an overview. Journal of Experimental Biology 115 : of Experimental Biology 211 , 510–23. 405–12. Vollrath, F. (1998). Dwarf males. Trends in Ecology and Weibel, E.R. (2000). Symmorphosis: On form and function in Evolution 13 , 159–63. shaping life . Harvard University Press, Cambridge, Mass. Waage, J.K. (1979). Foraging for patchily distributed hosts Weibel, E.R. (2002). The pitfalls of power laws. Nature 417 , by the parasitoid, Nemeritis canescens . Journal of Animal 131–32. Ecology 48 , 353–71. Weibel, E.R. and Bolis, L. (1998). In memory of Charles Waddington, C.H. (1942). Canalization of development Richard Taylor (1939–1995). In Principles of animal design: and the inheritance of acquired characters. Nature 150 , The optimization and symmorphosis debate, eds E.R. Weibel, 563–65. C.R. Taylor, and L. Bolis, pp. xv–xvii. Cambridge Wagner, P.D. (1993). Algebraic analysis of the determi- University Press, Cambridge. 93 nants of VO2 max. Respiratory Physiology , 221–37. Weibel, E.R. and Hoppeler, H. (2005). Exercise-induced Wajnberg, É. (2006). Time allocation strategies in insect maximal metabolic rates scales with muscle aerobic parasitoids: from ultimate predictions to proximate capacity. Journal of Experimental Biology 208 , 1635–44. behavioural mechanisms. Behavioral Ecology and Weibel, E.R., Taylor, C.R., and Hoppeler, H. (1991). The Sociobiology 60 , 589–611. concept of symmorphosis: a testable hypothesis of Walsberg, G.E. (1993). Thermal consequences of diurnal structure-function relationship. Proceedings of the microhabitat selection in a small bird. Ornis Scandinavica National Academy of Sciences, USA , 88 , 10357–61. 24 , 174–82. Weibel, E.R., Taylor, C.R., and Hoppeler, H. (1992). Walsberg, G.E. (2000). Small mammals in hot deserts: Variations in function and design: testing symmorpho- some generalizations revisited. Bioscience 50 , 109–20. sis in the respiratory system. Respiratory Physiology 87 , Walther, G.R., Berger, S., and Sykes, M.T. (2005). An eco- 325–48. logical ‘footprint’ of climate change. Proceedings of the Weibel, E.R., Taylor, C.R., Weber, J-M., Vock, R., Roberts, Royal Society, London, B 272 , 1427–32. T.J., and Hoppeler, H. (1996). Design of the oxygen and REFERENCES 217

substrate pathways. VII. Different structural limits for human studies from Skylab to the present. Acta oxygen and substrate supply to muscle mitochondria. Astronautica 58 , 59–81. Journal of Experimental Biology 199 , 1699–1709. White, C.R. and Seymour, R.S. (2004). Does basal meta- Weibel, E.R., Taylor, C.R., and Bolis, L. eds (1998). Principles bolic rate contain a useful signal? Mammalian BMR of animal design: The optimization and symmorphosis debate . allometry and correlations with a selection of physio- Cambridge University Press, Cambridge. logical, ecological, and life-history variables. Physiological Weibel, E.R., Bacigalupe, L.D., Schmitt, B., and Hoppeler, and Biochemical Zoology 77 , 929–41. H. (2004). Allometric scaling of maximal metabolic rate White, C.R. and Seymour, R.S. (2005). Allometric scaling in mammals: muscle aerobic capacity as determinant of mammalian metabolism. Journal of Experimental factor. Respiratory Physiology and Neurobiology 140 , Biology 208 , 1611–19. 115–52. White, C.R., Blackburn, T.M., Martin, G.R., and Butler, P.J. Weiner, J. (1992). Physiological limits to sustainable energy (2007). Basal metabolic rate of birds is associated with budgets in birds and mammals: Ecological implications. habitat temperature and precipitation, not primary pro- Trends in Ecology and Evolution 7 , 384–88. ductivity. Proceedings of the Royal Society, London, B 274 , Weiner, J. (1994). The beak of the fi nch: A story of evolution in 287–93. our time . Jonathan Cape, London. White, C.R., Terblanche, J.S., Kabat, A.P., Blackburn, T.M., Weishaupt, M.A., Staernpfl i, H., Billeter, R., and Straub, R. Chown, S.L., and Butler, P.J. (2008). Allometric scaling of (1996). Temperature changes during strenuous exercise maximum metabolic rate: the infl uence of temperature. in different body compartments of the horse. Functional Ecology 22 , 616–23. Pferdeheilkunde 12 , 450–54. White, T.C.R. (1993). The inadequate environment: Nitrogen Weiss, S.L., Lee, E.A., and Diamond, J.M. (1998). and the abundance of animals . Springer-Verlag, Berlin. Evolutionary matches of enzyme and transporter capac- Wiersma, P. and Piersma, T. (1994). Effects of microhabitat, ities to dietary substrate loads in the intestinal brush fl ocking, climate and migratory goal on energy expend- border. Proceedings of the National Academy of Sciences, iture in the annual cycle of red knots. Condor 96 , USA 95 , 2117–21. 257–79. Welham, C.V.J. and Ydenberg, R.C. (1993). Effi ciency- Wiersma, P. and Piersma, T. (1995). Scoring abdominal maximizing fl ight speeds in parent black terns. Ecology profi les to characterize migratory cohorts of shorebirds: 74 , 1893–1901. an example with red knots. Journal of Field Ornithology Werner, E.E. and Hall, D.J. (1974). Optimal foraging and 66 , 88–98. the size selection of prey by the bluegill sunfi sh (Lepomis Wiersma, P., Bruinzeel, L., and Piersma, T. (1993). Energy macrochirus ). Ecology 55 , 1042–52. savings in waders: studies on the insulation of knots. Werner, E.E. and Peacor, S.D. (2003). A review of trait-me- Limosa 66 , 41–52. diated indirect interactions in ecological communities. Wiersma, P., Selman, C., Speakman, J.R., and Verhulst, S. Ecology 84 , 1083–1100. (2004). Birds sacrifi ce oxidative protection for reproduc- West, G.B., Savage, V.M., Gillooly, J., Enquist, B.J., tion. Proceedings of the Royal Society, London, B 271 , Woodruff, W.H., and Brown, J.H. (2003). Why does met- S360–63. abolic rate scale with body size? Nature 421 , 713. Wiersma, P., Muñoz-Garcia, A., Walker, A., and Williams, J.B. West-Eberhard, M.J. (1992). Adaptation: current usages. In (2007). Tropical birds have a slow pace of life. Proceedings Keywords in evolutionary biology, eds E.F. Keller and E.A. of the National Academy of Sciences, USA 104 , 9340–45. Lloyd, pp. 13–18. Harvard University Press, Cambridge, Wijnandts, H. (1984). Ecological energetics of the long- Mass. eared owl (Asio otus ). Ardea 72 , 1–92. West-Eberhard, M.J. (2003). Developmental plasticity and Wilbur, H.M. (1990). Coping with chaos: toads in ephem- evolution . Oxford University Press, New York. eral ponds. Trends in Ecology and Evolution 5 , 37. West-Eberhard, M.J. (2005). Developmental plasticity and Wilbur, H.M. and Collins, J.P. (1973). Ecological aspects of the origin of species differences. Proceedings of the amphibian metamorphosis. Science 182 , 1305–14. National Academy of Sciences, USA 102 , 6543–49. Williams, G.C. (1966a). Natural selection, the costs of Westerterp, K.R., Saris, W.H.M., van Es, M., and ten Hoor, reproduction, and a refi nement of Lack’s principle. F. (1986). Use of the doubly labeled water technique in American Naturalist 100 , 687–90. humans during heavy sustained exercise. Journal of Williams, G.C. (1966b). Adaptation and natural selection. Applied Physiology 61 , 2162–67. Princeton University Press, Princeton, NJ. Whedon, G.D. and Rambaut, P.C. (2006). Effects of long- Williams, J.B. and Tieleman, B.I. (2000). Flexibility in basal duration space fl ight on calcium metabolism: review of metabolic rate and evaporative water loss among 218 REFERENCES

hoopoe larks exposed to different environmental tem- Wund, M.A., Baker, J.A., Clancy, B., Golub, J.L., and Foster, peratures. Journal of Experimental Biology 203 , 3153–59. S.A. (2008). A test of the ‘fl exible stem’ model of evolu- Williams, J.B. and Tieleman, B.I. (2005). Physiological tion: ancestral plasticity, genetic accommodation, and adaptation in desert birds. BioScience 55 , 416–25. morphological divergence in the threespine stickleback Willmer, P., Stone, G., and Johnston, I. (2000). Environmental radiation. American Naturalist 172 , 449–62. physiology of animals . Blackwell Science, Oxford. Yan, L.J. and Sohal, R.S.(2000). Prevention of fl ight activity Wilson, D.S. (1976). Deducing the energy available in the prolongs the life span of the housefl y, Musca domestica , environment: an application of optimal foraging theory. and attenuates the age-associated oxidative damage to Biotropica 8 , 96–103. specifi c mitochondrial proteins. Free Radical Biology and Wilson, J.S., Kraft, P.G., and van Damme, R. (2005). Medicine 29 , 1143–50. Predator-specfi c changes in the morphology and swim- Yang, H-Y., Chen, B., and Zhang, Z-W. (2008). Seasonal ming performance of larval Rana lessonae . Functional changes in numbers and species composition of migra- Ecology 19 , 238–44. tory shorebirds in northern Bohai Bay, China. Wader Windig, J.J., Brakefi eld, P.M., Reitsma, N., and Wilson, Study Group Bulletin 115 , 133–39. J.G.M. (1994). Seasonal polyphenism in the wild: Survey Ydenberg, R.C. and Hurd, P. (1998). Simple models of of wing patterns in fi ve species ofBicyclus butterfl ies in feeding with time and energy constraints. Behavioral Malawi. Ecological Entomology 19 , 285–98. Ecology 9 , 49–53. Wingfi eld, J.C. (2005). The concept of allostasis: Coping Ydenberg, R.C., Welham, C.V.J., Schmid-Hempel, R., with a capricious environment. Journal of Mammalogy 86 , Schmid-Hempel, P., and Beauchamp, G. (1994). Time 248–54. and energy constraints and the relationships between Wingfi eld, J.C. (2008). Organziation of vertebrate annual currencies in foraging theory. Behavioral Ecology 5 , cycles: implications for control mechanisms. 28–34. Philosophical Transactions of the Royal Society, London, B Ydenberg, R.C., Butler, R.W., Lank, D.B., Smith, B.D., and 363 , 425–41. Ireland, J. (2004). Western sandpipers have altered Wingfi eld, J.C. and Kenagy, G.J. (1991). Natural regulation migration tactics as peregrine falcon populations have of reproductive cycles. In Vertebrate endocrinology: Fun- recovered. Proceedings of the Royal Society, London, B 271 , damentals and biomedical implications, eds M. Schreibman 1263–69. and R.E. Jones, Vol. 4B, pp. 181–241 . Academic Press, Ydenberg, R.C., Butler, R.W., and Lank, D.B. (2007). Effects San Diego. of predator landscapes on the evolutionary ecology of Winslow, C.E.A. and Herrington, L.P. (1949). Temperature routing, timing and molt by long-distance migrants. and human life . Princeton University Press, Princeton, Journal of Avian Biology 38 , 523–29. NJ. Yoder, J.A., McClain, C.R., Feldman, G.C., and Esaias, W.E. Wolf, T.J. and Schmid-Hempel, P. (1989). Extra loads and (1993). Annual cycles of phytoplankton chlorophyll foraging life span in honeybee workers. Journal of Animal concentrations in the global ocean: a satellite view. Ecology 58 , 943–54. Global Biochemical Cycles 7 , 181–93. Wolff, J. (1892). Das Gesetz der Transformation der Knochen . Zuk, M. (2007). Riddled with life: Friendly worms, ladybug A. Hirchwild, Berlin (published in translation in 1986 as sex, and the parasites that make us who we are . Harcourt, The law of bone remodelling . Springer-Verlag, Berlin). Orlando. Wolff, W.J. and Smit, C.J. (1990). The Banc d’Arguin, Zwarts, L. and Wanink, J.H. (1991). The macrobenthos Mauritania, as an environment for coastal birds. Ardea fraction accessible to waders may represent marginal 78 , 17–38. prey. Oecologia 87 , 581–87. Woltereck, R. (1909). Weitere experimentelle Untersuchungen Zwarts, L., Blomert, A-M., and Wanink, J.H. (1992). Annual über Artveränderung, speziell über das Wesen quantita- and seasonal variation in the food supply harvestable tiver Artunterschiede bei Daphniden. Versuche Deutsche by knot Calidris canutus staging in the Wadden Sea in Zoologisches Gesellschaft 19, 110–72. late summer. Marine Ecology Progress Series 83 , 129–39. Wouters-Adriaans, M.P.E. and Westerterp, K.R. (2006). Zwarts, L., Ens, B.J., Goss-Custard, J.D., Hulscher, J.B., and Basal metabolic rate as a proxy for overnight energy Kersten, M. (1996). Why oystercatchers Haematopus ost- expenditure: the effect of age. British Journal of Nutrition ralegus cannot meet their daily energy requirements in a 95 , 1166–70. single low water period. Ardea 84A , 269–90. Wright, J., Stone, R.E., and Brown, N. (2003). Communal Zwarts, L., Bijlsma, R.G., van der Kamp, J., and Wymenga, roosts as structured information centres in the raven, E. (2009). Living on the edge: Wetlands and birds in a chang- Corvus corax . Journal of Animal Ecology 72 , 1003–14. ing Sahel . KNNV Publishing, Zeist. Name index

Adams, G.R. 83 Crowther, S. 34 , 36 Fiennes, R. 55–6 , 58 Alberts, H. 44 Croze, H. 121 Ford, H. 34–6 Alerstam, T. 117 Currey, J.D. 41 Alexander, R.M. 41 , 43 Cuthill, I. 113–14 , 121 Galama, Y. 12 , 15 Alice in Wonderland, 147–8 Gans, C. 44 Allen, A.O. 66 Daan, S. 58–60 , 67–8 , 110 Garland, T. 18 , 182–3 Amundsen, R. 55–7 Dall, S.R.X. 122 Gates, D.M. 21 Anderson, C.W. 158–9 Daly, M. 148 Gerth, N. 95 Arsenault, D.J. 80 Darveau, C-A. 50–1 Gibbon, J. 127 Aubret, F. 178 Darwin, C. 76 , 79 , 175 Gienapp, P. 168 Avital, E. 179 , 181 , 183 Darwin, E. 175 Gilbert, S.F. 89 Davidson, N.C. 10 , 30 Gill, R.E. 12 , 60 , 75 Badyaev, A.V. 12 , 177–8 Davies, N.B. 112–13 Gillooly, J.F. 66 Bakken, G. 20–1 Dawkins, R. 34 Gould, S.J. 101 Banavar, J.R. 51 Dawson, W.R. 60 Grant, B.R. 103 Barnes, M. 80 Deerenberg, C. 70 Grant, P.R. 103 Bartholomew, G.A. 1 , 2 , 24 de Goeij, P. 11 Green, R.F. 126 Battley, P.F. 49 , 104 , 106 , 168 de Jonge, M. 163 Bauchinger, U. 105 Dekinga, A. 154 Haldane, J.B.S. 177 Bayes, T. 123 Dennett, D. 34 Hall, M. 123–4 Bergmann, C. 18 de Vlas, J. 163 Hammond, K.A. 48 , 60 , 64 , 67 Bergmann, O. 95 DeWitt, T.J. 88 , 94 , 98 Handel, C. 12 Bibby, R. 109–10 Diamond, J.M. 40–2 , 47–8 , Harvell, C.D. 88 Biernaskie, J.M. 125 60 , 67 Haugen, M. 24 Bijlsma, R. 12 Dierschke, V. 156 Hawking, S. 8 Blumberg, M.S. 88 Dietz, M.W. 12 , 152–4 Hedenström, A. 117 Both, C. 12 , 170–1 DiZio, P. 83 Heinrich, B. 12 Boyd, H. 12 , 55 Dobzhansky, T. 2 , 183 Hendry, A.P. 167–9 Brakefi eld, P.M. 12 , 92–3 Dorfman, T.A. 83 Henry, C.J.K. 58 Brown, P. 174 Drent, R. v, 3 , 11 , 58–60 , 90 , 94 , Hill, J.A. 82 Brown, J.S. 125 , 155 137 , 174 Hilton, G.M. 105 Brugge, M. 151–2 Dudley, R. 44 Hochachka, P.W. 50–1 Brunner, D. 127 Duijns, S. 157 Hogan, J. 12 Buehler, D.M. 10 , 149–50 Dykhuizen, D. 35 Holling, C.S. 138 Bull, J.J. 94 Hoppeler, H. 12 Eaton, H. 12 Houston, A.I. 115 , 117 Carroll, L. 147 Eckert, R. 16 Hoyt, D.F. 36 Chappell, M.A. 60 Edgell, T.C. 178 Humphrey, N. 35 Charmantier, A. 170 Elser, J.J. 27 Hurd, P. 118–19 Charnov, E.L. 70 , 71 Emlen, D.J. 91 , 97 Christensen, B. 132 Epel, D. 89 Irschick, D. 12

219 220 NAME INDEX

Jablonka, E. 12 , 176 , 179–81 , 183 Meyers, L.A. 94 Scholander, P. 17–21 Jacobs, J.D. 4 Midgley, J. 27 Schubert, K.A. 12 , 63 Jansen, R. 12 Monaghan, P. 172 Scott, P. 55 Johannsen, W. 87 , 175 Montgomerie, R. 90 Scott, R.F. 55–8 , 70 Jones, J.H. 37 , 45 Mulcahy, D. 12 Secor, S.M. 86 Jonkers, B. 12 Sedinger, J.S. 105 Nagy, K.A. 60 Shackleton, E. 55 Kaandorp, J.A. 87 Naya, D.E. 96 Sherman, I. 12 Kacelnik, A. 111–14 , 121 , 127 Neufeld, C.J. 12 , 79–81 Shettleworth, S.J. 112–13 Karasov, W.H. 7 , 174 Nijhout, H.F. 87 , 106 Shine, R. 178 Kelly, S.A. 182–3 Noakes, T. 12 , 45 , 55–7 , 68 , 70 Sih, A. 132 Kempenaers, B. 8 Noble, D. 7 , 182–3 Simmons, L.W. 97 Kersten, M. 73 , 76 Nolan, V. 175 Sjodin, A.M. 58 Ketterson, E.D. 175 Nolet, B. 26–7 Spalding, K.L. 95 King, J.R. 19 Speakman, J.R. 12 , 62–5 , 68 , 71–2 Kinnison, M.T. 167–8 Oaten, A. 123 Staaland, H. 26 Kirkwood, J.K. 59–60 , 73 Olson, E.N. 82 Starck, J.M. 12 , 104 Klaassen, M. 25–7 Olsson, O. 12 , 120 , 125–6 Stroud, M. 12 , 55–8 Klasing, K.C. 104 Suarez, R. 33–4 Kleiber, M. 46 , 50 Palmer, A.R. 79–81 Sutherland, W.J. 158–9 Kraan, C. 163–5 Pearl, R. 70 Swennen, C. 117–19 Krebs, J.R. 112–13 , 121 , 131–2 Perrigo, G. 63 , 68 Krogh, A. 57 Perry, G. 132 Taylor, C.R. 33 , 35–7 , 39 Król, E. 62–5 Peterson, C.C. 60 Teplitsky, C. 100 , 169 Kvist, A. 74 Pianka, E.R. 131–3 Tesch, P.A. 83 Piersma, T. passim Tieleman, B.I. 12 , 19 , 20 , 24 Lackner, J.R. 83 Pigliucci, M. 12 , 87–8 , 94 , 98 , Tijsen, W. 120 Lamarck, J-B. 175 176 , 178 Tinbergen, J.M. 12 Lamb, M.J. 179–81 , 183 Porter, W.P. 21 Tinbergen, N. 110–11 , 121 Lampert, W. 87 Postma, F.M. 12 , 181 Tollrian, R. 88 Lane, N. 72 Tomkovich, P.S. 166 Lank, D.B. 156 Quammen, D. 76 Turner, J.S. 88 , 89 , 101–2 Le Maho, Y. 29 Lessells, C.M. 94 Ramos, M. 12 , 97 Valladares, F. 94 Levey, D.J. 174 Redfi eld, A.C. 26 Valone, T.J. 123 Levins, R. 175–6 Reid, J.M. 172 van de Kam, J. 9 , 12 , 167 Levitan, D.R. 87 , 89 Reitsma, N. 92 van den Hout, P.J. 12 , 153 Lewontin, R. C. 101 , 175–6 , 183 Relyea, R. 12 , 99–101 van der Meer, J. 29 Lignot, J-H. 86 Ricklefs, R.E. 5 , 12 , 76 van Gestel, J. 12 , 181 Lindstedt, S.L. 37 Ridley, M. 147–8 van Gils, J.A. passim Lindström, Å. 74 Robling, A.G. 84 Veltman, I. 12 Lively, C.M. 87 Rollo, C.D. 88 Vézina, F. 60 , 76 , 94 Royama, T. 58 , 67 Vico, L. 84 MacArthur, R.H. 131 , 133 Rubner, M. 15 , 70–1 Videler, J.J. 60 Marchinko, K.B. 79–80 Ruff, C. 84 Vince, G. 95 Marsh, R.L. 60 Rydell, J. 42 Visser, D. 12 Martin, M-A. 12 , 71 Vogel, S. 37 , 41 Mayr, E. 182 Sandoval, C.P. 87 McNair, J.N. 122 Scheiner, S.M. 88 Waage, J.K. 126–7 McNamara, J.M. 12 , 115 , 117 Schlichting, C.D. 87–8 Wagner, P.D. 50 McWilliams, S. 105 Schluter, D. 8 Wanink, J.H. 156 Messner, R. 69–70 Schoeppner, N.M. 100 Weibel, E. 12 , 33–4 , 36–40 , 43 NAME INDEX 221

Weiner, J. 46–7 , 61–2 Wiersma, P. 22 , 23 Woltereck, R. 85 , 87 , 89 Wertz, P. 24 Williams, J. 24 Wund, M.A. 179 West, G.B. 51 Wilson, D.S. 125 West-Eberhard, M.J. 3 , 88 , Wilson, M. 148 Ydenberg, R.C. 115 , 117–19 , 156 176–7 Wingfi eld, J.C. 5 Westerterp, K. 12 Winkler, D. 12 Zeligowski, A. 180 White, C.R. 66 Wolff, J. 82–3 Zwarts, L. 156 Subject index

In the hope that this index will be as fl exible as the subject matter on which it throws light, the page numbersin bold represent chapters or major sub-sections of text. Readers can therefore concentrate on these, for a speedy appraisal of a topic, or delve deeper into the sub-headings for further enlightenment. The extensive cross-referencing should guide you to any related ideas—phenotypes are variable, and so is the terminology used to describe them!

Abdominal Profi le Index 151 , 157 symmorphosis 39 red knots 1 , 9–10 , 20 , 22 , 140 , 147 , absorption tarbutniks 179 165–8 , 175 energy/food 46–8 , 56 , 62–3 , to temperature 20 shorebirds 6 , 42 , 74–5 , 66 , 104 to water loss 24 147–9 , 160 heat 21–2 see also phenotypic fl exibility ; sled dogs 95 information 119 phenotypic plasticity temperature 19–20 water 104 adipose tissue 28 , 61–4 , 105 see also Antarctica ; polar activity 93 , 114 aerodynamic issues 44 , 151–3 explorers ; tundra butterfl ies 92–3 Africa 4 , 71 Aristotle’s lantern 89 costs of 22 , 115 East 37 , 39 , 45 , 92 , 150 arthropods 147 , 165–6 foraging 1 , 22 , 93 , 118–19 , 132 , 139 South 31 , 55 asset protection principle 157–8 metabolic rates 37 , 57–61 , 66 , 78 West 1 , 9 , 30–2 , 143 , 157 , 162 , 171 assimilation organ size 96 age factors 71 , 105 , 157 , 174 energy/food 46–7 , 61 , 66 , 74 , 103 red knot teams 153 , 155 airplanes 28–9 genetic 176–9 and resources 128 Alaska 17 , 56 , 61 astronauts 81–3 tadpoles 99–100 allocation problems 96 , 103 athletes see also exercise ; performance ; allometry 11 , 38–9 , 50–1 , 59–60 , arachnid 96 work 65–6 human 39–40 , 45–6 , 55 , 60 , adaptation 2–5 , 174–5 , 183 allostatic loads 69 67–9 , 75 adaptive behaviour 179–80 altitudinal factors 24 , 65 , 69 see also running adaptive design 101 , 175 amphibians 98–102 ATP 33–4 , 37–8 , 46 , 60 adaptive peaks 5 , 44–6 amphipods 123 Australia 9 , 11 , 105 , 143 , 166–8 , adaptive plasticity 97–8 amputation 42 , 96–7 179–80 adaptive radiation 177 annelids 89 avian infl uenza 151 arm-waving 79–81 Antarctica 9 , 16 , 29 , 55–8 , 70 axolotl, Mexican 99 bone functional adaptation antibodies 149 model 84 antlers 91–2 , 101–2 bacteria 35 , 149–50 colour change 44 , 92 arctic zones balances, see physical balances egg-laying dates 171 bar-tailed godwits 157 Ballastplaat 129–30 , 139 evolutionary 175–6 cold climates 6 , 16–20 , 55–6 , 60 , bank voles 158 maladaptive phenotypic 63–6 , 69 barnacles 79–81 , 85 , 87 , 96 , 172 change 94 explorers 56–7 , 75 bar-tailed godwits 28 , 61 , 75 , 104 , partial 176 geese 143 147 , 157 power of phenotypic mammals 17–20 Barro Colorado Island, Panama 18 plasticity 172 polar 16–18 , 148 basal metabolic rate (BMR) 76–8 random 94 ptarmigan 90 allometric relationships 50–1

222 SUBJECT INDEX 223

birds 19 , 59 , 67 , 73 , 76 vigilance 155–6 tropical 18–19 evaporative water loss 24 water balance 24 wild versus captive, see wild mammals 19 , 58–9 , 66 see also coping strategies ; versus captive animals metabolic ceilings 61–2 , ethotypes ; fear ; information ; wing bones 41–2 66–8 , 71–7 integrative views ; learning ; see also fl ight ; foraging ; seasonal changes 104 performance ; psychology ; gizzards ; migration ; thermoregulation 18–22 selection processes ; trade-offs molluscivores ; moult ; as yardstick for activity 57–60 , behavioural ecology 58 , 110–12 , phenotypes ; plumage ; 76–8 125–6 , 131 , 136 predation ; reproduction ; bats 41–2 , 71 benefi ts and costs, see costs and shorebirds ; songbirds ; Bayesian updating 117 , 122–6 , 128 benefi ts waterbirds/fowl ; names of bears 16–18 , 159 bill-tip organ 10 individual species/groups beavers 18 , 155 biology of the future 183 bivalves 9–11 , 114 , 134 , 137 , bees 68 , 114 , 123–5 biome conversions 169 156–8 , 161–2 behaviour 3–5 , 107–44 biosphere 167 black-capped chickadees 121 adaptive 179–80 birds blackcaps 25 avoidance 109–10 acceleration 153 black-tailed godwit 147 and brain evolution 70 arctic, see arctic zones blood as buffer against change 183 aviary, see captive animals carbon isotope retention 105 choice of, see choice and Bayesian 117 , 124 , 128 cell turnover 95 decision-making beak/bill 4 , 10 , 102–3 , 112 cold-blooded animals 16 , 27 constraints/currencies BMR 19 , 59 , 67 , 73 , 76 during exercise 33 , 36 , 38 , 40 109–30 , 133–5 body mass/size 18 , 103–5 fl ow through the legs 22–3 cues 110 , 128 candle burning at both ends 70–2 fl ow to the brain 38 , 69 , 82 and dehydration 24–5 captive, see captive animals fl ow to the gut 85 escape 132 , 152–4 , 164 clutch sizes 65 haematocrit 104 and environment 6 , 109 , 120–4 , dehydration 24–5 parasites 148 , 150 130 , 178 depredated 156 pH regulation 43 and evolution 70 , 182–3 desert 19 , 20 , 24 , 169 proteins 27 farm animals 71 digestive system 10 , 49 , 103–6 stoichiometry 27 and fi tness 110–11 , 125–6 ecological cascade events 155 viscous 76 fl exible 109–10 , 124 , 130 , 159 eggs in spring 170–1 warm-blooded animals 16–20 , 26–7 fl ocking 153 , 155 , 160 escape abilities 152–4 , 164 see also leukocytes ; lymphocytes ; and fl ight performance 152 factorial scopes 61 monocytes ; phagocytes foraging 111–30 , 131–44 , 153 food supplementation 62–3 bodies 26–8 and function 110 , 130–1 , 135–6 free-living, see wild versus captive are earth 8 , 184 gene expression 93 animals balanced, see physical balances

hiding 156 habitat preferences 149 14 C in 95 inheritance system 181–2 immune system 103 , 148 changes in, see phenotypic lazy 19 , 68–9 , 70 , 114–15 lack of sweat glands 23 fl exibility ; phenotypic migration 1 , 9–10 , 139–44 , 166–8 life-cycle staging 93 plasticity optimality 109–30 , 131 , 144 , 154 lifetime energy expenditure 70–2 constant 6 plasticity 89 , 91 , 100 manoeuvrability 105 , 152–4 , 164 core temperature 19–20 , 22 , 60 and predation 151–2 , 155–6 , nitrogen requirements 26–7 design of, see design 160 , 178 not airplanes 28–9 development of, see prey 132 , 155 organ adjustment/size 6 , 103 developmental biology risky 156 parent 58–9 , 67 and ecology 2–4 , 32 , 174 roost 128 , 154 phenotypic fl exibility in 102–6 enigma of 1 sensible 70 predation 7 and environment, see environment spring temperature changes 171 scavenging 128 evolutionary aspects, see evolution thermoregulatory 22 sleeping together 128–30 and exercise, see exercise time factors 109 , 124 , 130 temperature changes 171 fi ne-tuning of 36 224 SUBJECT INDEX bodies (cont.) metabolic ceilings 58–9 , 77 ranges 1 , 10 , 28 , 147–8 , 157 fl exible, see phenotypic fl exibility and migration 167 red knots 9–10 , 28 , 144 , 147 , 165–6 and genotype, see genetic factors oxygen consumption 38–40 , 50 selective 45 , 71 and gravity 8 , 81–5 and pectoral muscle mass 152–3 success, see fi tness hot and cold 16–20 and plasticity 98 timing 171 identical twins 174 , 179 and predation 154 , 156 see also arctic zones ; insulation 18–19 , 66 spiders 96 reproduction life span 68 , 70–2 , 92–3 starvation 30–1 , 153 buttercups 87 , 89 machinery of, see body machinery tadpoles 100 butterfl ies 5 , 44 , 87 , 89 ,90–93 , 155 maintenance of, see body thermoregulation 18–19 maintenance variability 70 , 103 caeca 104–6 mass, see body mass and work 31–2 caiman 85 morphological features 92 , 100 , body parts, see organs calcium 27–8 , 103 103 , 106 , 109 body size 18 , 27 , 103 , 135 camoufl age 44 , 90–1 not static machines 32 see also body mass Canada 12 , 17 , 55 parts and whole organisms 106 body tissue, see tissues ptarmigan 90 plasticity, see phenotypic plasticity body weight, see weight red knots 1 , 4 , 22 , 28 , 165 size of, see body size Bohai Bay 167 shorebirds 42 starved, see starvation bones cannibalism 100 tired 58 , 114–15 antlers 91–2 , 101–2 captive animals weightless 8 , 81–5 density 83–4 Bayesian birds 117 , 124 see also death and dying ; digestion by snakes 86 body mass 30 digestive processes ; disease ; DNA turnover 95 cage-existence metabolism 73 food and diet ; heat ; leg 41–3 diet 133–4 , 136 hormones ; immune system ; and space travel 82–4 digestive organs 106 metabolism ; organisms ; stoichiometry 27–8 energetic balances 2 organs ; phenotypes ; proteins ; wing 41–2 foraging 131 respiration ; symmorphosis ; book writers 57 gizzard size 2 , 136–7 , 140–1 , 165–6 tissues ; water ; work brain immunity 150 body composition 160 antler formation 102 manoeuvrability 151–2 body condition 158 and exercise 38 , 69–70 metabolic ceilings 63–9 , 73–4 body fat, see fat carbon isotope retention 105 as models for free-living animals 2 body machinery 32 energy fl ow 62 patch behaviour 115 , 121 , 127 developmental 94 evolutionary behaviour 70 thermoregulation 17 , 20–1 digestive 103 information processing 119 see also wild versus captive metabolic 66 , 68 , 75–7 oldest tissue 95–6 animals protein 182 seasonal changes 103 carbohydrates 16 , 47 body maintenance song nuclei 103 carbon 26–7 and BMR 66 , 73 , 77 space travel 82 dioxide 33 , 43 , 109 costs of 22 , 76 , 89 , 94 , 97 , 105 , 141 stoichiometry 27 14 C 95 body mass see also behaviour isotopes 105 and aerodynamic performance 152 breeding carrying capacity 158–9 birds 18 , 103–5 bar-tailed godwits 157 cars 34–5 , 123–4 bivalves 156 Calidrinae 147–8 cats 41 , 70 , 156 and energy 66 , 70–4 colonies 128 cell turnover 95–6 , 105 and excess carbon 27 digestive organs 106 cellular epigenetic inheritance 181–2 fasting 29 extra season 110 chameleons 44 and feeding patch choice 135 nest-box species 171 cheetahs 37 and fl ight costs 139 phenotypic changes 93 chicken and egg problems 136 , 140 and food availability 29–32 , 89 plumage 22 choice and decision-making 125–8 , and functional capacity 105 , 136 prey delivery 111 131–44 and gizzard mass 11 , 166 prudent parents 58 Bayesian 126 SUBJECT INDEX 225

behavioural ecology 110–12 , 125–6 energetic 6 , 56 , 98 , 115 , 141 , 164 , fasting 29 , 31 , 136 , 153 defi nitions 58 , 110 172 and fl ight ability 72 , 75 distributional 103 , 161 fi tness 100 , 172 , 178 hard work 6 , 56–8 , 67–9 food/foraging 120–1 , 123 , fl ight 61 , 105 , 129–30 , 139 life and death decisions 119 , 166 131–44 , 161 foraging 114–15 , 119–20 , 131 , 133 , microbial 149–50 gizzard size 139–40 139 , 158 mortality risk factors 57–8 , 72 , 75 habitat 39 , 131 , 136 , 147–9 , 155 , 166 genetic 94 , 98 oystercatchers 162 life and death 119 , 166 gizzard size 139–42 , 164 predation 151 , 154 manipulation of 136–7 gregarious living 128 red knots 162 , 165–6 optimal solutions 113 , 154 immunity 149–50 starvation 29 , 31 , 136 , 153 physiological 137 induced defences 101 state of entropy 15 predation 152 information 94 sudden-death feeding patches 113 selective breeding 45 , 71 larval adjustments 100 thermoregulation 19–20 short-term 110–11 maintenance 22 , 76 , 89 , 94 , 97 , water restriction 25 starvation 153 105 , 141 see also life span ; survival stopover site 143–4 metabolic exploitation 158 decision-making, see choice and work levels 68–9 muscles 105 decision-making see also coping strategies ; natural selection 177 deer 90–3 , 101–2 learning ; trade-offs phenotypic plasticity 6 , 94 , 176–8 defi nitions, see terminology circannual factors 5 , 93 physical 6 dehydration 24–5 cirral nets 79–80 , 96 plasticity itself 97–8 demotype 4 , 5 , 7 , 178 , 183 climate predation 151 , 154–9 see also fi tness change 168–72 production 94 density factors cold 6 , 16–20 , 55–6 , 60 , 63–6 , 69 thermoregulatory 16 , 17 , 22 , bones 83–4 dangerous/extreme 5 , 17 73 , 140 conspecifi cs 93 , 100–1 hot 16 , 20 time 115 , 131 , 133 fl ocking behaviour 153 , 155 , 160 see also arctic zones ; deserts transport 36 , 43 foraging 130 , 132 , 134–6 , and desert birds ; temperate work 67–9 , 172 138–9 zones ; tropical zones ; see also death and dying ; Giving-Up Density (GUD) 125–6 , weather conditions physical balances ; survival ; 158–9 coastal habitats 147 trade-offs mitochondria 95 coatis 18 counting abilities 126 patch feeding 116 , 122 , 125–6 , cockles 11 , 118, 133–6 , 138 , 162–3 cows 15–16 , 41 , 65 , 70–1 157–9 conservation issues 160–2 , 172 Cox’s proportional hazards pathogens 149–50 coping strategies 172–3 model 126 phenotypic plasticity 87 , 89 danger 152 , 154–8 crabs 109–10 , 114 , 135 , 162 , 178 population issues 164–6 delayed negative effects 173 cranefl ies 42 deserts and desert birds 19 , 20 , environmental variations 2 , 109 , crustaceans 89 , 133–6 , 138 , 162 24 , 169 167 , 172 crypsis 44 , 90–2 , 121 design 13–51 , 174–5 predation threats 151 current value assessment rule 125 adaptive 101 , 175 prey quality variations 166 cuttlefi sh 44 constraints 6 thermodynamics 15 cyclists 45–6 , 55–8 , 60 , 62 , 68 ecological aspects 3 see also choice and decision- economic 33–51 making danger 5 , 17 , 100 , 109 , 152 , 154–9 intelligent 34 coral 87 , 89 Daphnia 26 , 85 , 87 , 89 , 97 modular 176 cosmonauts 81–4 dark-eyed juncos 175 multiple criteria 43–4 costs and benefi ts 96–8 Darwinian demons 97–8 rules of 5 activity 22 , 115 Darwin’s fi nches 102–3 safe, see safety factors barnacle penis length 80 , 96 Darwins 168 and selection processes 44 developmental plasticity 93 death and dying 31–2 , 151–4 see also symmorphosis digestive system 103–5 cyclists 45 developmental biology 89 , compensatory growth 173 DNA mortality 67 176–7 , 181–3 226 SUBJECT INDEX

developmental plasticity 3–5 , 85–8 , dreamcows 15–16 bottlenecks/ceilings 117 , 119 92–4 , 98 , 103 , 176–7 drinking 23 , 24 and carrying capacity 159 see also reaction norms ducks 147 , 162 conservation of 15 developmental-range limit 94 , 98 dumb waiters 41–2 constraints 117 diet quality dung beetles 97 costs and benefi ts 6 , 56 , 98 , 115 , coping with variations 166 dunlin 147 , 152 141 , 164 , 172 feeding patches 121–2 , 133–4 , currencies 115–19 , 155 138–44 , 154 earth 8 , 82 , 85 , 167–9 , 184 daily expenditure (DEE) 63 , 68 , metabolic ceilings 63 see also global factors 70–3 , 141 , 143 nitrogen requirements 27 eating 15–16 , 46 , 69 , 89 , 135–40 , 147 effi ciency 115–19 organ adjustment 6 , 103 see also digestive processes ; and endurance 55–7 and survival 158 , 164–8 food and diet expenditure, see physical balances see also food and diet echinoderms 89 fi tness value of one unit 156–7 digestive processes 2 ecological stoichiometry 26–8 fl ows 62 in birds 10 , 49 , 103–6 ecology 2–4 , 5–8, 55–78 gains and losses 5 , 132 bottlenecks/constraints 105 , behavioural 58 , 110–12 , 125–6 , harvesting versus processing 65 116–17 , 119 , 133–9 , 151 131 , 136 intake, see physical balances digestive rate models 135 , 138 cascading events 155 lifetime expenditure 70–2 and disease 151 demand functions, see and metabolic rates 43 , 58 hypertrophy 66 , 137 performance and migration 1–2 , 25 , 61 , 75 , 141 , and immune function 151 and design 3 148 , 157 and metabolic rates 61–3 , 66 , evolutionary 6–7 , 183 plateau expenditure 60–1 74 , 77 expressed in bodies 2–4 , and predation risk 156 plasticity 103–6, 144 , 175 32 , 174 of prey items 7 , 131–4 and predation 153 extrinsic mortality 72 and resource investment 97 shellfi sh, see shells and shellfi sh of fear 155 spreading out of 15 snakes eating bones 86 and gizzard size 141 stored resources 58 , 62 , 164–7 in space 84 global changes 161 stress 150 symmorphosis 46–9 integrative views 183 supply and demand 31 , 36 , 46 , 51 time factors 74 , 114–15 , 141 , 157 , see also environment ; and survival 67 , 69–71 , 75 , 172 167–8 integrative views ; organisms ; symmorphosis 34–5 upregulating capacity 63 seasonal factors trade-offs 153–4 , 156–7 see also eating ; excretion ; economic imperative 33–51 and weather conditions 164 food and diet ; gizzards ; guts ; ecosystem services 167 , 169 see also exercise ; food and diet ; intestines ; stomach ectotherms 17 , 26 foraging ; fuel and fuelling ; dinosaurs 27 egg dates 170–1 heat ; oxygen ; performance ; disease 67 , 69 , 82 , 147–60 , 172–3 eland 37 radiation ; respiration ; see also infection ; pathogens electron leakage 71 thermodynamics ; DNA elephants 18 thermoregulation ; water ; work cell turnover 95, 105 elevator cables 40–2 engineering 34 , 40–1 , 44 , 175 gene expression/sequencing elks 102 , 155 see also cars 182–3 Ellesmere Island, Canada 28 entropy 15–16 genotype 4 endotherms 16–17 , 26–7 environment 2–7 , 53–106 metabolic rates 27 endurance 55–8 , 60–1 , 67 , 70 , 95 , 97 and behaviour 6 , 109 , 120–4 , 130 , mortality 67 energetics 1–2 , 5 , 7 , 72 , 75 178 red knots 9 energy 5–6 , 15–32 clean/dirty 67 tarbutniks 179–80 absorption 46–8 , 56 , 62–3 , 66 , 104 constant 6 dogs 15 , 17 , 39–40 , 55–7 , 70 , 95 assimilation 46–7 , 61 , 66 , 74 , 103 coping with, see coping strategies downward causation 182–3 balance, see physical balances cues 94 , 97 dragonfl y larvae 99–101 barrel model 46 , 61–2 and development 85 , 183 dredging intertidal fl ats 161–2 , basal requirements 105 disease pathogens in 69 , 148–9 163 , 172 and body mass 66 , 70–4 dynamic, see variability SUBJECT INDEX 227

ecological/social 97 micro-evolutionary processes 173 feeding patches 131–44 energy in 62 modifi cation by descent 175 behaviour in 115–28 and fi tness 172 phenotypic plasticity 3–4 , 88 , costly 158 fl uctuating, see variability 93–4 , 97 , 103 , 173 , 175–9 dangerous 158–9 and genotypes, see genetic factors of sex 147 density factors 116 , 122 , 125–6 , gravity-free 8 , 81–5 excretion 157–9 information about 93 , 98 , 119–21 , faeces/urine 23, 62–3 , 104 and ecology of fear 155 123–4 , 128 salt 25 Giving-Up Density (GUD) 125–6 , integrative view 183 exercise 158–9 metabolic ceilings 6 , 77 and blood 33 , 36 , 38 , 40 Giving-Up Time (GUT) 121–2 mis-matched phenotypes 93 and brain 38 , 69–70 loss of 162 niche construction 179–80 , 181 and heart 76 , 82 , 95 patch choice versus prey and optimal phenotypes 97–8 metabolic ceilings 56–61 , 64–6 , 70 , choice 134 and organisms 4 , 6–7 , 85 , 97 , 105 , 76–7 quality of 121–2 , 133–4 , 138–44 , 154 167 , 174–6 , 180–4 and muscles 33–4 , 36–8 , 43–4 , red knots 115–17 , 124–5 , 129 , 131 parasites/pathogens in 150 56 , 61 and roosts 154 phenotypic plasticity 5 , 79–106 , phenotypic plasticity 83 safe 156 , 158–9 169 , 172 , 174–6 symmorphosis 37–8 sudden-death 113 predation danger 155 see also activity ; endurance ; work switching 115 , 160 , 162 , 165–6 , predictability 30 , 93–4 extinction rates 167 172 , 183 quality of 125 , 131 eyes larger than stomachs 133 time factors 112–16 , 126–7 seasonal, see seasonal factors see also food and diet ; foraging tadpoles sensing changes in 100 factorial scopes 59–61 , 66–8 feelings 20–3 as ‘theatre’ 183 fantasy creatures 15–16 , 97–8 , ferrets 71 variability 2–3 , 6–7 , 77 , 85 , 87 , 179–81 , 183 fi eld versus laboratory studies, see 93–4 , 97 , 109 , 128 , 167 , 172 , 178 fasting 29–30 , 96 wild versus captive animals visual scanning of 160 fat fi sh 44 , 89 , 100 , 123 , 156–8 , 177 , 179 see also altitudinal factors ; in abdominal cavity 105 fi shing nets for barnacles, climate ; earth ; ecology ; absorption 47 see cirral nets habitats ; latitudinal gradients ; blubber 16 fi shing pressure 162 , 167 temperature brown adipose 61 , 64 fi tness 3–7 , 58 enzymes 34 , 42 , 56 , 104 catabolism 29 , 32 and behavioural choice 110–11 , eosinophils 149–50 and disease 151 125–6 epigenetic inheritance 182 DNA turnover 95 costs 100 , 172 , 178 epiphenotype problem 94 , 98 around gizzard 105 currencies 6 , 109–30 erythropoietin (EPO) 45–6 and migration 25 , 28–32 , 61 , and disease 148 , 172 ethotypes 4–6 , 183 105 , 151 and early life strategies 172–3 see also behaviour oxidation in space 82 economy of design 34 evaporative cooling 21 , 23–4 in skin 16 , 24 environmental change 172 evolution 174–84 as social information 128 genetic assimilation 178 arms race 101 , 147 storage 28 , 62 , 74 , 104 measurable surrogates for 111 behaviour 70 , 182–3 see also body mass ; fuel and and plasticity 98 climate change 168–9 fuelling and predation 156–8 creativity of 6 fatigue 58 , 114–15 protecting long-term assets 66–8 and ecology 6–7 , 183 fear 98–101 , 151, 155–7 , 158–9 survival of the fi ttest 175 of endothermy 27 feathers value of food 156 genetic factors 174–7 , 179 , 182 colour 4 value of one unit of energy 156–7 and inheritance 180–2 mass 104 see also demotype ; survival and intelligent design 34 ptarmigan 90 fl ies 68 , 150 of laziness 68–9 safety factors 41–2 fl ight of metabolic ceilings 58 , 69–70 , thermoregulation 18 , 20 , 22–3 costs of 61 , 105 , 129–30 , 139 71 , 75 see also plumage and drinking 24 228 SUBJECT INDEX fl ight (cont.) fuelling ; nutrients ; prey ; galliformes 105–6 erratic display 153 starvation ; trade-offs gastrointestinal organs 105 feathers 42 foraging see also digestive processes and foraging/travel time activity 1 , 22 , 93 , 118–19 , 132 , 139 geese 41 , 69 , 143 , 159 111–14 , 116 barrel model 46 genetic factors 3–4 , 6–7 metabolic ceiling 77 Bayesian 123–6 accommodation 176–7 and mortality 72 , 75 behaviour 111–30 , 131–44 , 153 antler size and shape 102 muscles 29 , 61 , 75 , 104–5 central-place 111–12 , 120 assimilation 176–9 non-stop 61 , 74 , 139 choice 120–1 , 123 , 131–44 , 161 butterfl ies 92 performance 44 , 105 , 151–4 , 164 costs 114–15 , 119–20 , 131 , 133 , climate change 171 and water production 24 , 29 139 , 158 constraints/costs 94 , 98 see also migration ; space travel ; decrease of suitable area 164 disease defence 147 wings distributions 157 egg-laying dates 171 food and diet 131–44 gain ratio 117–19 and environment 3 , 106 , 168–9 , abundance, see density factors interference with 80 176–9 barnacles 80 metabolic factors 77 , 112–13 , 115–17 and evolution 174–7 , 179 , 182 and body mass 29–32 , 89 optimal 123 , 125–6 , 131–44 gene expression 93 , 182 breadth of 132 parasitoid 126 mutation 175 , 177 captive animals 133–4 , 136 past success 128 parthenogenetic reproduction 148 change of 128 , 164 , 183 and predation 156 , 159 plasticity 89 , 94 , 97–8 , 101 choice 120–1 , 123 , 131–44 , 161 profi tability 131–3 , 135 proteins 147 , 182 cues 101 , 158 sit-and-wait-strategies 84 reaction norms 87 in early life 173 solitary versus group 128 red-knot population structure 10 elemental composition 26–7 tadpoles 100 selection processes 101 , 174 and exercise 56 time factors 111–14 , 115–19 , 121–39 tarbutniks 179–80 , 183 fi tness value 156 trade-offs 111 , 113 , 132 traits 93 , 97 foodscape 130 Wadden Sea 129–30 , 134–5 , variability 4 , 181–3 food to fuel pathway 46 139–42 , 164–6 see also ATP ; DNA ; inheritance ; and gizzard mass, see gizzards wild versus captive animals 115 , mitochondria guided tours 128 121 , 134–8 gevoelstemperatuur 20 human intake reporting 57 see also density factors ; feeding giraffes 18 ignoring 131–2 , 135 patches ; food and diet ; prey gizzards intake, see physical balances foxes 17–18 captive birds 2 , 136–7 , migrant birds, see migration free-radical production 71 140–1 , 165–6 models 131–4 , 144 freshwater habitats/species 147–9 costs of enlarging 139–42 , 164 optimal 6 , 132 , 136 frogs 33–4 , 98–101 fat around 105 phenotypic plasticity 6 , 93 , Fryslân 170 fl exibile 49 , 104–6 , 135–44 , 154 , 100 , 131 fuel and fuelling 46–7 164–8 , 175 processing, see digestive processes deposition rates 74 infl exible 166–8 quality, see diet quality from fuel to force 33–4 magnitude of change 105–6 quantity versus quality 139 , 165 gizzard size 141–4 mass and body mass 11 , 166 seasonal factors 25 , 129 , 134–5 , metabolic ceilings 65 , 74 , and migration staging-sites 140–4 , 147 migration 28–32 , 48 , 65 , 114–15 , 142–3 , 165–7 soft/hard 2 , 9–11 , 48–9 , 134–141 , 144 , 151–2 , 157 , 168 optimal size 140–4 , 165 174–5 optimal working capacity 5 , 58 , processing capacity 164 , 175 storage 28 , 62–5 , 74 , 76 , 104–5 , 166 66–8 , 69 rate-mazimizing 140–3 supplementation 62–3 storage 74 , 76 , 105 , 166 red knots 10–11 , 48–9 , 117 , 133–5 , water content 23 , 25 time factors 74 , 114–15 , 141 , 157 , 142–4 , 153–4 , 164–6 see also absorption ; assimilation ; 167–8 satisfi cing 140–3 , 153 digestive processes ; eating ; see also energy ; fat ; food and seasonal factors 48–9 , 136 , 140–2 , energy ; fasting ; fat ; feeding diet ; nutrients 166 patches ; foraging ; fuel and fur 18–20 , 63–5 starvation 136 , 165 SUBJECT INDEX 229 global factors 7–10 , 142 , 161 , temperate, see temperate zones effort over time 68–9 167–9 , 173 tropical, see tropical zones endurance 55–8 see also earth wetland 147 exhaustion 61 glucose 47–8 see also environment ; feeding food intake reporting 57 glycogen 62 patches ; microhabitats ; tundra global change 167–9 , 173 goats 39 , 123–4 Haldanes 168–9 and gravity 8 , 81–5 godwits 28 , 61 , 75 , 104 , 147 , 157 hard nuts 133–5 insulation 19 gravity 8 , 81–5 hard times 109–10 , 172 landscapes of fear 159 great knots 48–9 , 104–5 hard work, see work lifetime energy expenditure 70 great tits 28 , 131–2 , 170–1 hares 63–4 , 155 low-altitude 69 Greenland 1 , 28 , 95 , 165 health 173 motivation 72 grey plover 147 see also disease pacing 68 Griend 129 , 134 , 139 , 154 heart performance levels 39–40 , 45 guinea pigs 40 , 64 , 70 atrophy 82 seasonal brain changes 103 gulls 17–18 , 59 , 147 , 152 , 169 carbon isotope retention 105 sweating 65 guts 84–5 , 131–44 fi xed number of beats 70 teeth 41 , 95 DNA/epithelium turnover 95 hypertrophy 82 , 85 , 96 temperature 19–20 fuel processing 34 , 47 muscle 29 , 95 trained 33–4 , 45 gut feelings 136 oldest tissue 95 hummingbirds 33–4 , 121 magnitude of change 105–6 oxygen 33 , 37 , 40 hypothetical creatures 15–16 , 97–8 , and manoeuvrability 105 , 164 phenotypic change 104 179–81 , 183 red knot 11 , 166 rates at altitude 69 and shellfi sh 135–40 , 166 stoichiometry 27 Iceland 22 , 28 , 48–9 size/mass 103 , 105–6 , 131–44 , 166 strenuous exercise/workload iguanas 27 in space 84 76 , 82 , 95 immune system 149–51 and storage organs 62 , 65 heat 15–32 , 44 birds 103 , 148 see also digestive processes ; absorption 21–2 farm animals 71 gizzards ; intestines dissipation 64–7 fl exibility 6 , 149–51 , 159–60 gyrfalcons 90 energy balance 46 , 62 space travel 82 generation in mammals 61 symmorphosis 49 habitats sense of loss 20 induced defences/offence 87 , 101 arctic, see arctic zones see also energy ; temperature infection 58 , 67 , 149–51 aviary, see captive animals heredity, see inheritance infl ammation 150 coastal 147 heterophyls 87 , 89 , 149–50 information 119–22 dangerous/risky 155 , 157 hibernation 93 carry-over between degradation 167–70 , 172–3 Himalayas 69 generations 180–1 desert, see desert and desert birds Holy Grail 161 costs of acquiring 94 and disease/immunity 147–51 homeostasis 101 , 178 , 183 environmental 93 , 98 , 119–21 , forest 91 , 155 , 161 , 169 homeotherms 16–20 , 27 123–4 , 128 freshwater 147–9 hormones 92–4 , 98 , 101 , 175 , 177–8 exchange at roosts 128 , 130 grassland 169–70 horses 21 , 35–7 , 39–41 , 43–5 , 65 , 70 genotype accommodation 176–9 laboratory, see captive animals house fi nches 177–8 reliability of 94 marine 27 , 109–10 , 147–9 , 177 , 179 house martins 59 sampling 125 , 130 models of habitat use 144 , 151 humans see also learning polar 16–18 , 148 athletes 39–40 , 45–6 , 55 , 60 , ingestion, see eating safe 155 67–9 , 75 inheritance 4 , 94 , 173–8 , 180–4 saline 147–9 BMR 58 see also genetic factors savannah 44 , 150 , 169 Bayesian updating 123 insects 68 , 93 selection of 39 , 131 , 136 , 147–9 , culture 181 see also butterfl ies and names of 155 , 166 disease 150 , 173 other species switching 115 , 160 , 162 , 165–6 , downstream effects of early insulation 18–19 , 66 172 , 183 life 172–3 integrative views 2–3 , 6–8 , 106 , 145–84 230 SUBJECT INDEX

interference competition 128 lemmings 17–19 maximal metabolic rate 66 intertidal mudfl ats 161–2 , 163 , 172 leukocytes 104 , 149–50 milk 48 , 63–6 , 68 , 71 bivalves 11 life and death 119 , 166 nitrogen requirements 27 China 167 see also danger ; death and dying ; optimal work levels 68–9 foraging time 118 fear ; predation pathogens in 150 scarring 162–3 life history pelt, see fur South Korea 167 characteristics 175 predation 159 thermoregulation 22 cycle staging 3–5 , 93–4 pregnancy 63–4 , 82 , 173 Wadden Sea 129 , 161–2 , 165 decisions 110 respiration 33 , 37–40 , 43 Yellow Sea 166 early life strategies 172–3 sweat glands 23 intestines invariant 70–1 tropical 18 , 20 fl exibility of 95 knock-on effects 155 water loss adaptations 24 large 96 lifetime energy expenditure 70–2 see also humans ; names of small 47–9 , 96 , 104–6 strategy/structure 4 individual mammals snake 85–6 versus metabolic machinery 75 marginal value theorem see also guts life span 68 , 70–2 , 92–3 (MVT) 119–21 , 125 , 155 isometric scaling 50 see also survival marine habitats/species 27 , 109–10 , lions 37 147–9 , 177 , 179 jays 17–18 lipids, see fat see also shells and shellfi sh; Jurassic Park 27 liver Wadden Sea ; Yellow Sea ; metabolism 62 , 66 , 104 names of individual species kairomones 85 plasticity 85 , 96 , 104–6 Mars bar 83 kangaroo 41–2 starvation 153 mass/size factors, see body mass ; kestrels 68 , 70 , 110 stoichiometry 27 body size ; feathers ; organs ; kidneys 27 , 38 , 41 , 85 , 96 , 105 symmorphosis 38 , 47–9 shells and shellfi sh kiwis 76 livestock 71 , 150 Mauritania 9 , 31 , 143 knots, see great knots ; red knots see also cows ; horses ; pigs ; sheep ; metabolic ceilings 6 , 22 , 45–6 , 55–78 steers metabolic rates laboratory animals, see captive lizards 16–17 , 42 activity 37 , 57–61 , 66 , 78 animals load-distance correlations 120–1 allometric relationships 50–1 lactation 48 , 63–6 , 68 , 71 loading effect 111 basal, see basal metabolic lactic acid 60 , 69 locomotion 61–2 rate (BMR) Lapland longspurs 114 long-eared owls 59 and body size 27 lapwings 170–1 long-tailed skuas 73 butterfl ies 92–3 larks 19 , 20 , 24 lumberjacks 58–9 cold-induced 55–6 , 60 larvae 80 , 92–3 , 97–102 lungs 23 , 33 , 37–44 , 39 , 96 , 103 constraints on 61–5 latitudinal gradients 18 , 20 , 65 , 73 , lymphocytes 149–50 digestive processes 61–3 , 66 , 74 , 77 144 , 149 , 166 and energy 43 , 58 see also seasonal factors Malawi 92 exercise-induced 56–60 laziness 19 , 68–9 , 70 , 114–15 Mali 147 foraging 77 , 112–13 , 115–17 learning mammals and gizzard size 141 evolution/tarbutniks 179–80 , 183 aquatic 16 of lungs 43 by mistakes 114 arctic 17–20 maintenance processes 37 speed of 121 , 124 , 173 BMR 19 , 58–9 , 66 maximal 50 , 60 , 66 to time (LeT) 128 body size increases 18 maximum sustained/peak, see see also information burrowing 159 metabolic ceilings leatherjackets 110–14 as experimental models 63 minimum 63 legs factorial scopes 61 seasonal factors 142 birds 22–3 , 43 heat generation 61 summit 60 bones 41–3 life-cycle staging 93 sustained metabolic scope, see losing 42–3 lifetime energy expenditure 71–2 factorial scope muscles 69 , 82–3 , 105 , 153 mammary glands 63–4 , 66 thermal conditions 2 , 17–18 SUBJECT INDEX 231

variability of 60 , 76 digestive organ size 140 Modern Synthesis 174 , 177 , 182 see also performance distances 148 mole rats 35 metabolism energy 1–2 , 25 , 61 , 75 , 141 , 148 , 157 molluscivores 9–11 , 161 , 162–5 body machinery 66 , 68 , 75–7 and fat 25 , 28–32 , 61 , 105 , 151 see also red knots by-products of 62 fl exibility of 105 , 139–44 , molluscs 11 , 28 , 89 , 132–5 , 139 , 143 cage-existence 73 166–7 , 168 see also names of individual species changes in space 82 fl ight phase, see fl ight money in the bank 165–7 limits of, see metabolic ceilings food, see food and diet mongooses 37 , 39 liver 62 , 66 , 104 fuelling phase, see fuel and fuelling monkeys 18 maintenance 22 , 31 gizzards at staging-sites monocytes 149–50 oxidative 33 142–3 , 165–7 Monty Hall problem 123–4 rates of, see metabolic rates at high altitude 24 moose 102 , 159 snakes 85 hotspots 143 more haste less speed 115–17 water production in fl ight 24 , 29 at night 24 , 65 morphological features 92 , 100 , 103 , metamorphosis 92 , 99 and protein 25 , 29 , 32 , 61 , 75 106 , 109 methodological issues speed of 166–7 mortality, see death and dying Abdominal Profi le Index 151 , 157 stopover sites in N/S motivation 57 , 72 , 126 allometry 11 , 38–9 , 50–1 , hemispheres 144 moult 42 , 89–91 , 93 59–60 , 65–6 survival 75 Mount Everest 69–70 Bayesian updating 117 , 122–6 , 128 timing of 75 , 140 , 143 mudfl ats, see intertidal mudfl ats BMR as yardstick 76–8 see also bar-tailed godwits ; red mussels 11 , 137 , 141 , 161–2 causation/correlation 171 , 180–4 knots and names of other muscles currencies 6–7 , 109–30 , 133–5 , 155 species ; seasonal factors ; atrophy 75 current value assessment rule 125 shorebirds cardiac 95 direct versus indirect methods 73 Millennium Ecosystem in the cold 64 , 69 fl ight performance 152 Assessment 167 , 169 contraction 62 immunity 149 minerals 26 costs of 105 isometric scaling 50 mitochondria 33–4 , 36–40 , 46 , 71–2 , 95 DNA turnover 95 load-distance correlations 120–1 models during exercise 33–4 , 36–8 , 43–4 , marginal value theorem allometric cascade 50 56 , 61 (MVT) 119–21 , 125 , 155 barrel/energy balance 46 , 61–2 fast-twitch fi bres 153 measurement versus insight 183 Bayesian foraging 123 fatigue 68–70 phenotypic design 174–5 bone functional adaptation 84 fl ight 29 , 61 , 75 , 104–5 reciprocal causation captive for wild birds 2 heart 29 , 95 feedbacks 180–4 carrying capacity 158 intercostal 95 rate-of-living theory 70–1 Cox’s proportional hazards 126 leg 69 , 82–3 , 105 , 153 scalar expectancy theory (SET) 127 diet/foraging 123 , 131–4 , 144 muscular stomach, see gizzards stoichiometry 26–8 digestive rate 135 , 138 pectoral 76 , 151–4 ultrasonography 135–7 ecology of fear 155 sled dogs 95 use-disuse principle 81–4 , 180 genetic 169 snakes 85 see also design ; models ; reaction gizzard optimality 142–3 in space 82–4 norms ; terminology ; habitat use 144 , 151 stoichiometry 27 variability ; wild versus Holling’s disc 138 tadpole tail 99–102 captive animals ideal free-distribution 128–9 thermogenesis 66 mice 37 , 39 , 41 , 47–8 , 63–8 , 180 mammals as experimental 63 mutation 175 , 177 microbes, see pathogens optimality 67 , 69 , 114 , 131–2 , 142 microhabitats 22 , 155 , 167 sparrowhawk 153 natural selection, see selection micronutrients 15 stochastic dynamic processes microvilli 85–6 programming 130 Netherlands migration 1–2 , 6–7 , 9–11 symmorphosis 36 Ardea 58 aerodynamic performance 151 taxidermic 21 Dutch dreamcows 15–16 dehydration/water 24–5 , 29 trade-off 97 Dutch postwar famine 173 232 SUBJECT INDEX

Netherlands (cont.) whole and body parts 106 assessing 50–1 gevoelstemperatuur 20 see also behaviour ; birds ; bodies ; endurance 55–8, 60–1 , 67 , 70 , infl uence on authors 1 , 169 ecology ; energy ; fi tness; 95 , 97 kestrels 70 humans ; mammals ; failure 72 larks 24 morphological features ; genetic factors 182 oystercatchers 162 performance ; reproduction ; and life history variables 71 red knots 30–1 , 153 , 172 sex ; survival ; traits limits to 45–6 , 58 , 63–6 , 75 Royal Netherlands Institute for organs maximal/peak 45 , 60–6 Sea Research 151 atrophy/shrinkage 22 , 29 , 75 , migrating shorebirds 75 Schiermonnikoog 111 , 113 82 , 85 and plastic responses 94 , 98 , spring temperatures 170–1 bill-tip 10 105 , 166 starling parents 113 capacity 3 , 5–7 , 40–1 , 77–8 , race horses 45 see also Wadden Sea 164 , 174–5 regulation of 61–5 , 70 neuroendocrine control system 94 , 98 digestive, see digestive processes selection processes 46 newts 98 , 100 failure 72 symmorphosis 36–41 , 45 , 63 New Zealand 9 , 61 , 76 , 169 fl exibility 6 , 47 , 85–6 , 105 , 135 , 166 see also activity ; energy ; niche construction 179–80 , 181 hypertrophy 66–7 , 96 metabolic ceilings ; metabolic nitrogen 16 , 26–7 , 29 , 104 miniaturization 79 rates ; work norm of reaction, see reaction norms nutritional, see digestive processes periwinkles 109–10 nutrients 6 , 15–32 , 47 as orchestra 7 , 182–4 phagocytes 149–50 see also fat ; food and diet ; fuel severing of 42 , 95–6 phenotypes and fuelling ; solutes sex 79–81 , 96–7 accommodation 176 size/mass 77–8 , 96 , 103–6 , 139 , adaptive, see adaptation optimality 6 144 , 153–4 adjustments 6–7 , 93–7 , 100 , behaviour 109–30 , 131 , 144 , 154 stoichiometry 27 154–5 , 172–7 foraging 111–12 , 123 , 125–6 , 131–44 storage 62–3 , 65 see also phenotypic fl exibility ; gizzard size 140–4 , 165 systems in series 33 phenotypic plasticity models 67 , 69 , 114 , 131–2 , 142 see also names of individual organs behavioural dimension of, phenotypes 97–8 , 109–30 , 131 , 144 osteoblasts 101 see behaviour working capacity 5 , 58 , 66–8 , 69 ostriches 41–2l breakdown of 172–3 organisms Oxford, UK 12 , 113 , 170–1 change, see phenotypic fl exibility ; and adversaries 147 oxygen 33–40 , 43 , 46–7 , 50–1 , 60 , 69 phenotypic plasticity assessing changes in 50–1 oystercatchers 31 , 103 , 117–19 , 162 defi nition of 4–5 coping techniques, see coping developmental processes 6 , strategies pancreas 41 , 85 , 104–5 105 , 174 defi nition of 4–5 parasites 17 , 75 , 147–8 , 150 see also developmental plasticity design of, see design parasitoid wasp 126–7 engineering 175 diseases, see pathogens parrots 71 environmental factors, distribution, see population issues patches, see feeding patches see environment and environment, see environment pathogens 69 , 75 , 147–51 evolutionary factors, see evolution feedback loops 6 , 181–3 see also bacteria ; disease ; of fear 98–101 , 151 , 155–7 , 58–9 immobile 172 infection fi ttest, see fi tness integrative studies of 2–3 , 6–8 , pathologies 172 fi xed traits/robustness 4 , 94 , 97 106 , 145–84 see also disease fl exible, see phenotypic fl exibility intra-generational stages 4 paying the carpenter 165 and genotype, see genetic factors niche construction 179–80 , 181 pedipalps 96–7 larval 80 , 92–3 , 97–102 phenotypic variation, see penguins 16 , 29 and life history cycles, see life phenotypes penis length in barnacles 79–81 , 96 history physical laws, see physical peregrine falcons 151–2 maladaptive change 94 balances ; physiology performance 3 , 5 , 61–5 and niches 179–80 , 181 predated, see predation aerodynamic/fl ight 44 , 105 , modular organization of 176 ‘theatre’ of 182–4 151–4 , 164 optimal 109–30 , 131 , 144 SUBJECT INDEX 233

partially adapted 176 limits to 94–8 , 100 , 103 , 105 , pied fl ycatchers 171 and pathologies, see disease 166 , 172–3 pigeons 24 , 41 plastic, see phenotypic plasticity patterns of 167–8 pigs 65 and predation, see predation phylogenetic constraints 94 , 98 plankton 26 , 80 , 96 reversible changes 3–6 , 79–80 , 89 , in plants 87 , 89 , 94 , 98 plants 87 , 89 , 94 , 98 93–4 , 101 , 174 , 176 red knots 104 , 136 plasticity, see phenotypic plasticity see also phenotypic fl exibility reproductive, see reproduction plumage 22 , 90–1 , 128 seasonal changes, see seasonal reversible, see phenotypic see also feathers factors fl exibility poikilotherms 16–17 and selection, see selection seasonal, see seasonal factors Point Barrow, Alaska 17 , 18 processes shells, see shells and shellfi sh polar explorers 55–8 , 75 struggling with 1 and survival, see survival polar habitats/species 16–18 , 148 time factors, see time and timing tadpoles 98–102 polyphenism 3 , 5 , 87 , 89 , 92–3 trade-offs, see trade-offs through time, see time and timing population issues 7 , 9–10 , 161–73 variable, see variability tinkerer’s accomplishment bivalves 158 see also bodies ; organisms ; 88 , 101–2 and individual constraint 98 organs versus no plasticity at all 97 and phenotypic fl exibility 161 phenotypic fl exibility see also adaptation ; see also survival avian, see birds developmental plasticity ; potential value assessment rule 125 and climate change 169–72 phenotypic fl exibility ; predation 147–60 core concept 7 , 182 self-organization ; variability aerial 164 developmental 101 phosphorus 26–7 and behaviour 151–2 , 155–6 , see also developmental plasticity physical balances 5–6 , 15–32 160 , 178 and evolution, see evolution barrel model 46 costs of 151 , 154–9 immune system 6 , 149–51 , 159–60 disease 148 crabs 109–10 , 114 , 178 and infl exibility 2–4 , 94–8 , 161 , gizzard mass 11 , 166 fear 98–101 , 151 , 155–7 , 158–9 166–8 metabolic ceilings 57–76 metabolic ceilings 58 , 60 , 72 , 75 population consequences, see optimal behaviour/foraging predatory snails 85 , 87 population issues 111–25 , 131 , 144 ptarmigan 90 reversible changes 3–6 , 79–80 , 89 , population consequences 164–5 risk factors 58 , 100 , 152 , 155–60 93–4 , 101 , 174 , 176 predation 151 , 153 , 155–9 shorebirds/insects 42 , 151 , 156 speed of 124 symmorphosis 36 , 48 tadpoles 99–100 tissue-specifi c 96 physical challenge, see endurance ; see also molluscivores ; prey ; see also phenotypic plasticity ; work reaction norms ; snakes variability physiology 3 , 5 , 7 , 55–78 pressure sensors 9–10 phenotypic plasticity 79–106 acclimatization 93 prey behavioural, see behaviour dehydration 24–5 behaviour 132 , 155 books about 88–9 and environment 178 and bill shape 103 categories/defi nition of 3 , 85–9 extreme conditions 17 buried 9–11 , 134 , 156 , 158 classes/degrees of 94–6 farm animals 71 changes in 128 classical 5 , 85–9 , 93 fasting 29 cryptic 121 and environment 5 , 79–106 , 169 , food-processing capacity 144 delivery decisions 111 172 , 174–6 gene expression 93 density, see density factors and evolution 3–4 , 88 , 93–4 , 97 , integrative views 2–3 , 6–8 , digestion of, see digestive 103 , 173 , 175–9 106 , 145–84 processes generational factors 4 , 94 , manipulated `choices’ 137 distribution 125 , 128 , 134 , 154 173–8 , 180–4 metabolic ceilings 64 , 68–70 , 72 dynamic 130 genetic-related theories 89 , 94 , prey capture by snakes 85–6 energetics 7 , 131–4 98 , 101 sled dogs 95 exposure time 130 and global change 7–10 , 142 , 167–9 see also integrative views ; handling 116 , 131–3 Haldane metric for 168–9 metabolism ; physical immobile/mobile 132 larval stages 80 , 92–3 , 97–102 balances marginal 156 234 SUBJECT INDEX prey (cont.) diet, see food and diet reproductive success/value, optimal number of 111 digestive system, see digestive see fi tness pecking intensity 127 processes reptiles 16 , 26–7 , 85 preferences 133–4 DNA 9 see also names of individual species quality, see diet quality farshore foragers 153 resources 58 , 62 , 94 , 97–8 , 128 rates of capture 47 foraging behaviour, see foraging see also energy ; food and diet ; remote detection of 9–10 genetic population structure 10 heat ; water sampling 121–2 , 132 gizzard size/mass, see gizzards respiration searching for 112 , 116 , 125–6 , ideal but not free 130 evaporative water loss 23–4 130–1, 133–4 immunity in, see immune system metabolic ceilings 63 , 76 with shells, see shells and shellfi sh life-cycle stages 4 symmorphosis 33 , 36–40 , of snakes 85–6 migrant, see migration 43–6 , 61–2 of squid 41 mortality 162 , 165–6 see also heart ; lungs whole ingestion 133 nasal salt glands 25–6 rhinoceros 16 , 18 see also feeding patches ; food phenotypic change, see Richel 129–30 , 135 , 139 , 154 and diet ; foraging ; predation ; phenotypic fl exibility ; rinderpest 150 names of types of predators phenotypic plasticity risk and risk factors and prey population issues 10 , 161–2 , behaviour 156 Prince Patrick Island 17 164 farm animals 71 proteins predation, see predation ; prey fatigue 58 in blood 27 roost behaviour 128 , 154 infection 58 catabolism 29 , 32 , 153 seasonal changes, see seasonal mortality 57–8 , 72 , 75 energy release 16 factors organ failure 72 genetic aspects 147 , 182 shell-handling, see shells and predation 58 , 100 , 152 , 155–60 immune 149 shellfi sh starvation 31 , 155 during migration 25 , 29 , 32 , 61 , 75 starvation 30–2 , 153–4 see also danger ; safety factors storage 62 stopover site selection 143–4 roads 159 proventriculus 49 , 104 subspecies 9–10 , 22 , 143 rodents 63–6 prudent parents 58 canutus 22 , 30 , 143 roosts 128 , 130 , 139 , 154 psychology 55 , 125–8 islandica 10 , 22 , 129 , 141 , 143 Rubner’s legacy 70–2 ptarmigan 89–93 , 109 piersmai 10 , 22 , 143 , 166 ruffs 147 ptiloerection 23 rogersi 143 running 33 , 36 , 43 , 45 , 55–6 , 64 , pythons 84–6 rufa 143 147–9 survival, see survival see also athletes raccoons 18–19 team activity 153 , 155 radiation, thermal 20–1 thermoregulation 20 , 22–3 safety factors 40–3 rainfall 93 water ingestion 25 in breeding areas 166 raptors 152–4 , 156 wintering areas 10 , 30 , 147 energy/safety trade-off rate-of-living theory 70–1 Red Queen hypothesis 147–9 153–4 , 156–7 rats 18–19 , 64 , 68 , 71 , 84 , 86 Redfi eld ratio 26 feathers 41–2 ravens 128 redshanks 157 feeding patches 156 , 158–9 reaction norms Reaktionsnorm 4 , reproduction food intake rates 47 85–9 , 176 , 178 and ageing 157 habitats 155 reciprocal causation barnacles 79 lung capacity 40 feedbacks 180–4 lactation 48 , 63–6 , 68 , 71 predation 154 , 156–7 red-billed choughs 172 metabolic ceilings 67–9 , 71 at roosts 154 red knots passim parthenogenetic 148 see also risk and risk factors age factors 157 phenotypic change 93 salamanders 99–101 as Bayesians 124 sea urchins 89 saline habitats 147–9 breeding ranges 10 , 147 and temperature change 170–1 see also marine habitats/species calcium dynamics 27–8 see also breeding ; prudent salt 25 , 26 captive, see captive animals parents sanderling 147 SUBJECT INDEX 235 sandpipers 48–9 , 147 natural/sexual 5 , 35 , 98 , 168 , population issues 161 scalar expectancy theory (SET) 127 175–8 , 181–3 predators 151 , 156 Schiermonnikoog 111 , 113 organ size/capacity 78 running 43 Scotland 31 on performance 46 tropically wintering 22 Scott’s polar exploration 55–7 and phenotype plasticity 98 , wetland wintering 147 sealife, see marine habitats/species ; 171 , 184 see also names of individual species Wadden Sea tinkerer’s apprentice 101 shrimps 133–5 seals 16 and wear and tear on tissues 71–2 skin sea urchins 87 , 89 self-amputation 42 , 96–7 antler velvet 92 , 101 seasonal factors self-organization 6 , 98 , 101 , 106 , 175 , carbon isotope retention Antarctica 29 , 55 180 , 183 times 105 BMR 104 sex distasteful toads 99 brain changes in humans 103 appeal of cars 35 DNA turnover 95 diet 25 , 129 , 134–5 , 140–4 , 147 changes 89 fat/lipids 16 , 24 disease 151 , 160 evolution of 147 intra-individual change 104 Dutch winter famine 173 gender 4 , 96–7 , 135 , 148 stoichiometry 27 European wintering 1 , 22 , 30–2 hormones 175 stuffed 1 foraging time 118 organs 79–81 , 96–7 sweating 23 fuelling rates 144 and pathogens 148 sleeping sickness 150 gizzard mass 48–9 , 136 , 140–2 , 166 sexual advertisement 44 , 90–1 sloths 18–19 habitat choice 147–8 sexual selection 182–3 snails 28 , 85 , 87 , 109 , 178 homeothermy 17–18 see also reproduction see also periwinkles immunity 149 sheep 65 snakes 84–6 , 178–9 lazy kestrels 68 shells and shellfi sh 135–40 snow buntings 17 metabolic requirements 142 in acidifi ed seas 109–10 sodium 26 organ size 96 crushing 10 , 48 , 104 , 117 , 133 solutes 15 phenotypic plasticity 3–5 , 87 , fl esh-to-shell ratio 165 songbirds 73 , 103 , 105–6 , 140 , 156 89–93 , 94 , 103–4 hard/soft 2 , 9–11 , 48–9 , 134–141 , space travel 8 , 81–5 polyphenism 87 , 89 , 92–4 174–5 spandrels 101 red knots 9–10 , 20 , 25 , 30 , 103 , mass 133 , 137–8 , 141 , 143 , 164 sparrowhawks 152–3 147 , 154 , 162–6 processing 134 , 137–8 , 141 , 165–6 spiders 41–3 , 87 , 96–7 shorebirds 22 , 31 , 74 , 147 red knots 9–11 , 25 , 117 , 132–7 , 166 sport, see athletes ; cyclists sled dogs 95 retention time 133 spring, advancing 144 , 169–72 song nuclei 103 thickness 109–10 squid shell 41 spring, advancing 144 , 169–72 in Wadden Sea 161–2 squirrels 17–18 starvation 153 see also names of individual starlings 59 , 111–14 , 119–21 , 127–8 tropical wintering 1–2 , 22 , 30–2 species starvation see also breeding ; latitudinal shivering 61 , 66 body mass 30–1 , 153 gradients ; migration shorebirds fi shing pressure 162 sediment characteristics 162 basal metabolic rate 22 gizzard size 136 , 165 selection criteria, see choice and digestive organs/system 49 , 105–6 mortality 29 , 31 , 136 , 153 decision-making distribution 7 polar explorers 56 selection processes energy balance/expenditure 73–4 protein catabolization 153 artifi cial 45 , 71 fat store dynamics 28 versus predation 155 , 159 as `audience/spectators’ 183 and fi shermen 167 steers 37 , 39 and BMR 77–8 food-processing 47–9 sticklebacks 177 , 179 and design processes 44 habitat preferences 147–8 stoichiometry 26–8 and functional responses 176 hard-working 6 , 72–5 stomach 47 , 49 , 85–6 , 96 , 104–5 and genes 101 , 174 incubating 73 muscular, see gizzards and metabolic ceilings 69 insect predation 42 storks 23 and mutation 177 insurance strategies 30–1 sugars 47 migrant, see migration summer, see seasonal factors 236 SUBJECT INDEX

surfbird 48–9 metabolic rates 2 , 17–18 and effort 68–9 survival operative 20–2 energy balance 5 diet quality 158 , 164–8 phenotypic changes 93 and evolutionary change 178 and early life 172–3 regulation, see thermoregulation food processing/fuelling 74 , and energy expenditure 67 , reproduction/spring 114–15 , 141 , 157 , 167–8 69–71, 75 , 172 changes 170–1 foraging 111–14 , 115–19 , 121–39 of the fi ttest 175 upper critical 19–20 Giving–Up Time (GUT) 121 optimal behaviour 110 , 126 , 128 see also heat gizzard mass change 137 , 167 and phenotypic plasticity 89 , 93 , terminology information acquisition 98 98 , 100 , 173 behavioural ecology slang 58 lagtime on phenotypic predation 157–8 choice 58 , 110 response 94 , 98 red knots 162 , 165–6 , 172 cold-blooded 16 learning to time (LeT) 128 and work 67–9 , 172 currencies 7 , 111 lifetime energy expenditure 70–2 see also death and dying ; decision 110 loss of plasticity 178 fi tness ; life span ; population ecology 7 migration 75 , 140 , 143 issues evolution 3 mutational load 177 swans 120 , 151 , 158–9 inconsistent 16 phenotypic plasticity 6 , 84 , 94 , symbolic inheritance 181–2 metabolic ceiling 56 168 , 178 symmorphosis 33–51 , 66 motivational variables 126 predation 153 , 156 defi nition of 33 , 36 optimality 6 prey exposure 130 food-processing chain 46–9 organism 4–5 prey quality 165–6 null hypothesis 49 , 63 phenotype 4–5 reproduction 93 , 171 performance limitation 36–41 , phenotypic plasticity 3 , 85–9 resource investments 97 45 , 63 symmorphosis 33 , 36 roost arrival 128 respiratory chain 33 , 36–40 , tarbutniks 179 and selection pressures 44 43–6 , 61–2 warm-blooded 16 spring arrival 170 see also design willingness 58 survival 93 , 175 sympatric speciation 177 termites 26 , 87 time of year, see seasonal factors terns 114 , 147 wasting time 132 tadpoles 98–102 testosterone 175 see also endurance ; evolution tarbutniks 179–81 , 183 thermodynamics 5 , 15 , 23 tinker er’s accomplishment 88 , 101–2 tellinids 11 , 133 , 158 , 162 thermogenesis 66 tiredness 58 , 114–15 Temminck’s stint 147 thermoregulation 16–23 , 44 , 60 , 62 , tissues temperate zones 73 , 140 adipose 28 , 61–4 , 105 disease pressure 148 see also ectotherms ; endotherms ; burning up 29 , 61 energy factors 2 , 19 , 22 , 31–2 homeotherms ; poikilotherms cell turnover 95 , 105 habitat loss 169 threat detection 100 damage and death 67 resource quality 166 see also danger fl exibility differences 95–6 , 105 temperature 20–3 thrush nightingale 61 fuel availability 47 air/ambient 20–2 , 25 , 66 tidal factors 118–19 , 139 , 154 , 161 genetic factors 182 and allometric scaling see also intertidal mudfl ats growth 46 constant 65–6 time and timing maintenance processes 37 and butterfl y phenotypes 92 allocation problems 103 oldest 95–6 and caterpillar phenology 171 animals’ assessment of 126–8 oxygen in 33 core body 19–20 , 22 , 60 behavioural change 109 , 124 , 130 pH 26 and egg-laying dates 170 body changes 105 protein 29 and exercise 64–6 carbon isotope retention 105 , 133 snails 178 feelings/gevoelstemperatuur 20 costs and benefi ts 115 , 131 , 133 specifi c plasticity 96 and food-processing 48 diet quality 142 wear and tear 71–2 and immunity 150 duration of metabolic see also blood ; bones ; brain ; lower critical 17–20 intensity 75 muscles ; skin SUBJECT INDEX 237 toads 96 , 98–100 ultrasonography 135–7 cockle-dredging 163 tough going 109–10 , 172 updating animals 122–5 foraging areas 129–30 , 134–5 , Tour de France 45–6 , 55–8 , 60 , 62 , 68 urohydrosis 23 139–42 , 164–6 trade-offs 93 , 96–8 use-disuse principle 81–4 , 180 intertidal mudfl ats 129 , 161–2 , bivalve burying versus food 165 intake 156 Vancouver Island 79 predation 154 , 157 energy/food/safety 153–4 , 156–7 variable seedeaters 20 red knot immunity 150 foraging 111 , 113 , 132 variability wintering area 22 , 30–2 immune function and antler size 91 water 23–6 digestion 151 behavioural 125 , 130 , 159 , 178 absorption 104 insemination/speed and body mass 70 , 103 barnacle fi shing nets/penises endurance 97 bone loss in space 83 79–81 predation/starvation 155 cell types 95 coral structure 87 , 89 roost safety/travel distances 154 climate change 171 doubly-labelled 64 , 73 work/survival 67 cost of fl ight 61 fl eas 26 , 85 , 87 , 89 , 97 see also choice and cyclic 93 and fl ight 24 , 29 decision-making ; costs energy balance/expenditure 28 , 70 in food 23 , 25 and benefi ts environmental 2–3 , 6–7 , 77 , 85 , 87 , loss 24–5 , 44 traits 93–4 , 97 , 109 , 128 , 167 , 172 , 178 and migration 24–5 , 29 BMR 78 experimental/induced 175 , 178 oxidative 23 butterfl ies 92 fi rst egg dates 170–1 periwinkles in dangerous 109 development 98 genetic 4 , 181–3 restriction and death 25 energy balance 32 gizzard size 49 , 104–6 , 135–44 , thermoregulation 22 and environmental conditions 87 154 , 164–8 , 175 see also freshwater habitats/ fi xed versus fl exible/ and heredity 177 species ; marine habitats/ reversible 3–4 , 94 , 97–8 immunity 149 species ; wetlands genetic factors 93 , 97 inter-individual 94 , 140 , 174–5 waterbirds/fowl 1 , 24 , 105–6 , 140 inheritance 174–5 , 178 , 184 intra-individual 3 , 90 , 94 , 98 , see also shorebirds phenotypic compromise 94 103–4 , 174 , 176 weasels 17–19 red knots 11 life history 71 weather conditions 20–1 , shorebirds 104 metabolic rates 60 , 76 164 , 170 sled dogs 95 motivational 126 see also climate ; radiation, thermal ; treadmills 36–7 , 60 , 83 organ size/mass 47 , 103–5 rainfall ; wind tropical zones phenotypes 176–8 weight butterfl ies 92 population sizes 161 loss in polar explorers 56 disease 148 predictable 93–4 and predation 156 energy factors 1–2 , 18–20 , 22 , prey quality 166 see also body mass 30–2 , 65 , 144 reversible 3 , 79–80 , 89 weightlessness 8 , 81–5 habitat degradation 166–7 , 169 self-organized pre-existing 175 Westwad 129–30 tryptophan 35 snail plasticity 178 wetlands 147 tsetse fl ies 150 and survival of the fi ttest 175 white-crowned sparrows 19 , 103 tundra trait 174 wild versus captive animals habitat loss 169 water fl eas 85 , 87 digestive organs 106 predation pressure 42 see also phenotypic fl exibility ; energetic balances 1–2 ptarmigan 90 , 109 phenotypic plasticity ; foraging 115 , 121 , 134–8 red knots 1 , 9 , 28 , 135 , 147 , 166 , seasonal factors gizzard size 2 , 136–7 , 140–1 , 172 , 175 viruses 149–50 165–6 shorebirds 73 vultures 23 , 128 immunity 150 see also arctic zones seasonal plasticity 92–3 turnstones 152–3 waders, see shorebirds thermal conditions 21 turtle doves 59 Wadden Sea 1 , 161–2 see also captive animals 238 SUBJECT INDEX

wildebeest 37 foraging 118–19 , 139 wear and tear 42 , 72 wind 20–2 , 61 , 118 marginal value of 67 willingness to 68 wings 41–2 , 44 , 92 metabolic ceilings 55–7 , 58–9 see also activity ; see also fl ight mortality 6 , 56–8 , 67–9 endurance ; energy ; exercise ; winter, see seasonal factors motivation for 57 , 72 muscles ; performance ; wolves 18 , 155 optimal working capacity 5 , 58 , running woodpeckers 125–6 66–8 , 69 work resistance to 69 yeasts 149–50 and body mass 31–2 shorebirds 6 , 72–5 Yellow Sea 161 , 166–8 costs and benefi ts 67–9 , 172 sled dogs 95 Yellowstone National Park 155 deskbound 57 and survival 67–9 , 172 and energy 15 , 43 , 46 , 48 , 62 symmorphosis 36–7 zebra fi nches 105 , 173