The Life History of Damaraland Mole-Rats Fukomys Damarensis: Growth, Ageing and Behaviour

The Life History of Damaraland Mole-rats Fukomys damarensis: Growth, Ageing and Behaviour

Jack Thorley

A thesis submitted to the University of Cambridge in application for the degree of Doctor of Philosophy

Jesus College

July 2018

The Life History of Damaraland Mole-rats Fukomys damarensis: Growth, Ageing and Behaviour

Jack Thorley

SUMMARY

The social mole-rats have often been typecast as extreme examples of mammalian sociality.

With their pronounced reproductive skew, status-related contrasts in lifespan and morphology, and the suggestion of a division of labour amongst helpers, mole-rat societies have repeatedly been likened to the structurally complex societies of some eusocial insects.

However, because few studies of mole-rats have quantified individual variation in growth and behaviour across long periods of development, it has remained unclear the extent to which mole-rat societies, and the features of individuals within them, should be considered unique amongst social vertebrates. In this thesis, I examine life history variation in

Damaraland mole-rats Fukomys damarensis from three perspectives- growth, behaviour, and ageing- to explore how individual developmental trajectories contribute to, and are influenced by, the structure of mole-rat societies. First, I use a large longitudinal dataset to test for the presence of behavioural specialisation in non-breeding mole-rat helpers. I find no indication of individual specialisation in cooperative activities. Instead, individual differences in helping behaviour are largely the result of age-related changes in the extent to which individuals commit to all forms of helping (Chapter 3); refuting the notion of helper castes. I then focus on the variation in growth across non-breeders, developing a novel biphasic model to accurately quantify sex differences in growth and explore the influence of social effects on growth trajectories (Chapter 4). Despite the proposition of intense intrasexual competition in mole-rat societies, there was no clear signature of sex-specific competition on helper growth trajectories. A more conspicuous form of socially-mediated growth in mole-rats is the secondary growth spurt displayed by females that have acquired the dominant breeding position, causing them to become larger and more elongated. By experimentally controlling reproduction in age-matched siblings, I show that rather than being stimulated by the removal from reproductive suppression, this adaptive morphological divergence is achieved through a lengthening of the lumbar vertebrae when breeding is commenced (Chapter 5). With

i contrasts in size and shape following the acquisition of the breeding role, this status-related growth pattern shares similarities with growth in naked mole-rats and other social vertebrates. Breeders also show a twofold greater lifespan than non-breeders in Fukomys mole- rats, prompting the suggestion that the transition to dominance also sets individuals onto a slower ageing trajectory. To date, there is little evidence to support a physiological basis to lifespan extension in breeders. This assertion is bolstered by the absence of longer telomeres or slower rates of telomere attrition in breeding females compared to non-breeding females residing in groups (Chapter 6), each of which might be expected if breeders age more slowly.

I argue that previous studies exploring status-related ageing in captive Fukomys mole-rats have overlooked the importance of demographic processes (and associated behavioural influences) on mortality schedules. Irrespective of the proximate basis of the longer lifespan of breeders, at an interspecific level the social mole-rats are unusually long-lived for their size.

A recent large-scale comparative analysis concluded that prolonged lifespan is a general characteristic of all mammalian cooperative breeders, but this conclusion is premature, as in most of the major clades containing both cooperative and non-cooperative species there is no consistent trend towards lifespan extension in cooperative species (Chapter 7). In the case of mole-rats, it seems more likely that their exceptional longevity arises principally from their subterranean habits and related reductions in extrinsic mortality. Overall, these findings demonstrate that cooperative breeding has important consequences for individual life histories, but there is no strong basis for the claim that Damaraland mole-rat societies are markedly different in form than other cooperative breeding societies.

PREFACE

This thesis is the result of my own work, and contains no work done in collaboration except where stated at the commencement of each chapter. The text does not exceed 60,000 words.

No part of this thesis has been submitted to any other university in application for a higher degree.

Jack Thorley

July 2018

iii

PAPERS ARISING FROM THESIS

Thorley, J., Mendonça, R., Vullioud, P., Torrents-Ticó, M, Zöttl, M., Gaynor, D., and T. Clutton-Brock. 2018. No task specialisation among helpers in Damaraland mole-rats. Animal Behaviour 143: 9-24. doi: 10.1098/rspb.2018.0897

Thorley, J., Katlein, N., Goddard, K., Zöttl, M., and T. Clutton-Brock. 2018. Reproduction triggers adaptive increases in body size in female mole-rats. Proceeding of the Royal Society of London B 285: 20180897. doi: 10.1098/rspb.2018.0897

OTHER RELATED CONTRIBUTIONS

Zöttl, M., Thorley, J., Gaynor, D., Bennett, N.C., and T. Clutton-Brock. 2016. Variation in growth of Damaraland mole-rats is explained by competition rather than by functional specialization for different tasks. Biology Letters 12: 20160820. doi: /10.1098/rsbl.2016.0820

Thorley, J., and T. Clutton-Brock. 2017 Kalahari vulture declines, through the eyes of meerkats. Ostrich: Journal of African Ornithology. Ostrich 88(2): 177-181. doi: 10.2989/00306525.2016.1257516

ACKNOWLEDGEMENTS

My first thanks go to my supervisor, Professor Tim Clutton-Brock, for providing me with a stimulating environment in which to conduct my PhD, in Cambridge and in the Kalahari. I admire Tim a great deal for the intellectual contributions he has made to behavioural ecology, and it has been a pleasure to be a member of the ‘Large Animal Research Group’. From him I have learnt a great deal about the method of scientific enquiry into the natural world, and the importance of clarity and precision in formulating hypotheses and constructing arguments. Above all, he has instilled in me the importance of iterations to the writing process. A result I hope to have eradicated the “streams of consciousness” that pervaded earlier versions of my thesis to leave an insightful body of work.

In LARG, I have been fortunate to work as part of a close-knit community of researchers- the LARGies. It would be over-indulgent to write something about each of them, but I am nonetheless grateful to AJ, Arik, Chris, Corina, Constance, Dieter, Dom, Mark, Monica, Peter and Teja for the insightful chats at lunch, and the less coherent chats in pubs throughout my tenure. All have had an instructive role in my learning and contributed to my enjoyment in Cambridge. I reserve special mention for Markus and Philippe, members of the mole-rat team with whom I have worked closely. Along with Rute Mendonça, they have invested much time and energy into the mole-rat project to ensure that is runs as smoothly as possible. Coordinating a project from 9,000km away is always a challenging and thankless task, but where it can be thanked it should.

For a brief period in my thesis I was based in Glasgow to undertake telomere work in the lab of Pat Monaghan. Pat was a wonderful host and I am thankful to her lab for the warm environment they provided to me. While there, Rob Gillespie was a fantastic mentor and source of much entertainment through the hours of pipetting. My telomeres would have shortened exceptionally fast without his help.

Whilst Cambridge has been a wonderful base for a PhD, it is the time spent in the South Africa that I will remember most vividly and cherish most fondly. As Niko Tinbergen notes in his Curious Naturalists “it is only natural for a man to have occasional doubts about the value of what he is doing”. Such doubts have certainly surfaced in me from time to time throughout my studies. But to return to the Kalahari each time was to feel refreshed and reinvigorated. It is a mesmerising place, beautiful in its harshness and deeply affecting in its enormity. Each visit to the Kalahari would reaffirm my decision to pursue an ecologically-oriented career. The intellectual gains provided some validation, but more fundamental is that feeling of intense kinship with a place that comes with extended fieldwork, living not with the landscape, but in it. The nights trapping mole-rats under a quicksilver milky way; the early morning hoar frosts; the metallic chattering of the sparrow-weavers; the blood-stained sunsets; the calms before the storms. I have been immensely lucky and will miss it all.

Yet for all its remoteness, the desert was a place of friendships, and there are many people at the Kalahari Meerkat Project who I am fortunate to have met. Many afforded me their time and patience, their skills, their jokes. Some indulged me with their cookery skills, and one even gave me their affection. In particular, I would like to thank Dave Gaynor for his hospitality and infectious enthusiasm for the natural world. He was a great sounding board and I would always enjoy nipping up to the ‘dog-house’ for a coffee. Nanine, for her warmth, and Bracken for his optimism and boundless energy- I’m sorry I never found that Lego piece you were missing. Tim Vink was always of great help with his technological and mechanical wizardry, second only to his skills at Braaing. And to the scores of meerkat volunteers with which I shared the place, Baie Danke. Looking back at the earliest photos now it is alarming just how many people I met across my PhD. If I had the remaining funds to buy them all a copy of this thesis I would, though I think most would prefer a crate of Savanna Dry.

Which brings me to everyone at the mole-rat project. Being forced to work in a lab whilst surrounded by a desert is a character-building experience and might go some way to explaining the insanities of all who work there. Whether causal or correlative I am still uncertain, but it made for a fantastic place to work. I thank Adam, Katy, Andrea and Niklas, for commandeering the mole-rat ship with aplomb when I was there. Francesco, Holly, Rachel, Sean, and Tash - chief nut cases-, for all the bizarre stories, games and crosswords. Nathan, for putting up with me as a boss despite my aptitude for indecisiveness. Aurora and Sofia for showing an Englishman how to be [more] direct. Ijen, Christina, Stephen, Dirkie and JP for putting in the hard work in the background with little fuss. Kyle, for being my mentor in the field and provider of bread. And Miquel, for being a close friend and enlightening me daily with his life advice. I have often quipped about the frequent and infrequent workers in the lab, but in truth, all worked far beyond their remit to ensure the data was collected fastidiously and the animals were well cared for. I know many of us will be in touch for a long time, and I look forward to seeing where we all end up.

My final and most heartfelt words go to my family and friends. Spending long periods in the field has always been easy in the knowledge that I have a safe and caring home to which I am anchored and can always return. My fascination with the natural world has often been met with curiosity by my brothers, but their indifference to my academic progress and interest principally in my personal contentment is a continual source of reassurance. My biggest thanks of all go to my Mum and Dad, who have always given me the unfailing support and freedom to fuel my curiosity and chase my ambitions, whatever they might be. I have come a long way from the boy that endlessly watched the Animals of Farthing Wood on VHS and caught newts in a pond with my bare hands. All that I have done, and continue to do is built upon the enriching environment they raised me in.

Purple-black and darker than the night the supple dunes slept beside the ancient river. The sky sparkled with the points of starlight and meteors streaked through the atmosphere. Below, the grasses, dry and tan before the dry season, reflected the celestial light, as if the river moved again.

Mark Owens, Cry of the Kalahari

vii

viii

TABLE OF CONTENTS

Summary i Preface iii Papers arising from thesis iv Acknowledgements v Table of Contents ix

Chapter 1: Introduction 1 The evolution of cooperation 2 Cooperative breeding life histories 3 Social mole-rats and the case for eusociality 7 Aims and thesis structure 13

Chapter 2: The Study System and General Methods 17 Study Site 18 The Damaraland mole-rat 22 Taxonomy 22 Geographic distribution and the aridity-food distribution hypothesis 22 Group dynamics, reproductive suppression, and skew 24 The wild population 25 The laboratory population 27

Chapter 3: No Evidence of Task Specialisation Among DMR Helpers 31 Abstract 32 Introduction 33 Methods 38 Animal housing and data collection 38 Models: multilevel multinomial behaviour models 38 Models: fitted models 40 Results 43 Individual trade-offs 43 General effects on mole-rat behaviour 47 Discussion 52

Chapter 4: The Shape of Growth in Damaraland Mole-rats: Sex and Social Effects 57 Abstract 58 Introduction 59 Methods 63 The shape of Damaraland mole-rat growth 63

A biphasic extension 64 Mass at birth 64 Group and litter-level effects on growth 65 Results 67 Discussion 75

Chapter 5: Reproduction Triggers Adaptive Increases in Size in Female DMRs 80 Abstract 81 Introduction 82 Methods 85 X-ray methodology 85 Morphological divergence 86 Skeletal changes 86 Body length and fitness 88 Results 89 Morphological divergence 92 Skeletal changes in breeders 92 Body length and fitness 94 Discussion 96

Chapter 6: The Long-lived Queen: Telomere Dynamics in the Wild and the Case for Lifespan Extension in Breeders 101 Abstract 102 Introduction 103 Methods 110 Blood sample collection 110 DNA Extraction and qPCR 113 Dataset 115 Statistics 115 Results 116 Discussion 119

Chapter 7: The Association Between Mammalian Cooperative Breeding and Lifespan 125 Abstract 126 Introduction 127 Methods 130 Global Dataset 130 Global Analysis 131 Reduced Data Analysis 131 Results 133 Discussion 136

Chapter 8: General Discussion 139 Overview 140 Growth 142 Ageing 143 Behaviour 145

References 147

Appendix A: Supporting Information for Chapter 3 164 Appendix B: Supporting Information for Chapter 4 174 Appendix C: Supporting Information for Chapter 5 181 Appendix D: Species Names mentioned in Thesis 186

xii

1 CHAPTER Introduction

______

INTRODUCTION

THE EVOLUTION OF COOPERATION

The evolution of cooperation has provided a fundamental and persistent problem for biologists since Charles Darwin (Herbers 2009). For Darwin himself, it was the reproductive division of labour he had observed in some social insects that posed the major challenge to his theory of evolution by natural selection. Writing in The Origin of Species, he states:

“I… will confine myself to one special difficulty, which at first appeared to me insuperable, and actually fatal to my whole theory. I allude to the neuters or sterile females in insect communities: for these neuters often differ widely in instinct and in structure from both male and fertile females, and yet from being sterile they cannot propagate their kind.” (Darwin 1859)

Darwin’s special difficulty can be generalised by considering any action where an individual helps another at some personal cost to themselves. This apparent paradox of cooperation has motivated a large body of theoretical and empirical research which has sought to explain the conditions under which cooperation [or ‘altruism’] can evolve, and thereafter be maintained (Kurzban et al. 2015; Lehmann and Keller 2006; Nowak 2006; Nowak and Sigmund 2005). Bill Hamilton provided the major conceptual advance in this area with two papers in the 1960’s where he outlined kin selection and inclusive fitness theory

(Hamilton 1964a, 1964b). Hamilton’s mathematically-driven approach highlighted that because genes are shared by relatives, altruistic traits can be selected for because of the benefit they bring to relatives. Hamiltonian theory has since been used to explain a diverse range of phenomena related to cooperation (Abbot et al. 2011; Bourke 2011), such as the general association in birds and mammals between group kinship and the propensity to direct care towards kin (Cornwallis et al. 2009). However, kinship does not always predict helping behaviour at an intraspecific level (Griffin and West 2002). In meerkats Suricata suricatta for example, an individual’s propensity to babysit and provision dependent pups is unaffected by relatedness (Clutton-Brock et al. 2000), and kin recognition is often undetected in numerous other social species where it might be expected (Gardner and West 2007). When this is the case, it is important to consider other means by which helping might be affecting direct fitness, and in this vein, a range of theoretical models have shown that cooperation can be maintained between non-kin via mutualistic benefits accrued through processes such as group augmentation (Kokko et al. 2001) or ‘paying-to-stay’ (Kokko 2002). In seeking broader

INTRODUCTION

empirical support for the manifold theoretical treatments related to cooperation, studies of cooperatively breeding societies, where individuals provide care to non-descendant kin in the form of protection and rearing, have proven particularly fertile testing grounds (Clutton-

Brock 2016; Koenig and Dickinson 2016; Solomon and French 1997).

COOPERATIVE BREEDING LIFE HISTORIES

Cooperative breeders are found in birds, mammals and fish, though their taxonomic scope within each of these major clades is limited: some 9% of bird species are cooperative breeders (Cockburn 2006), 1% of mammals (Lukas and Clutton-Brock 2012a), and a small fraction of fishes, all from within the Lamprologini cichlid tribe (Heg and Bachar 2006).

The idea that certain life history characteristics are involved in the evolution of cooperative breeding has a long history. Early bird-focussed studies suggested that low mortality or high clutch size could create a surplus of individuals which would favour prolonged offspring dependence, delayed reproduction, and helping (Brown 1974, 1987;

Ricklefs 1974). That life history traits were highly conserved across avian evolution lent credence to this argument (Owens and Bennett 1995), and the subsequent application of comparative phylogenetic approaches has confirmed the importance of life history parameters to avian cooperative breeding (Arnold and Owens 1998; Cornwallis et al. 2009;

Downing et al. 2015. Notably, by reconstructing the ancestral states of cooperative breeding across birds Downing et al. (2015) found that the transition from non-cooperative to cooperative breeding was facilitated by long lifespan; this transition likely taking place through an intermediate stage of family-living (Griesser et al. 2017).

Evidence from mammals also supports the role of life history on the evolution of cooperative breeding. Cooperative breeding in mammals has been restricted to lineages with monogamous ancestors where females produced multiple offspring per birth (Lukas and

Clutton-Brock 2012a, 2012b), as in the Canids, the Callitrichid primates and the Rodents. Even so, only a small number of monogamous mammals have made the transition to cooperative breeding, and this is probably because ecological factors also play a mediating role. This is evidenced by the geographical restriction of mammalian cooperative breeders to areas of low and unpredictable rainfall (Lukas and Clutton-Brock 2017), such that alloparental care might

INTRODUCTION

have evolved to buffer the challenge of raising multiple offspring in harsh environments

(Lukas and Clutton-Brock 2017). Avian cooperative breeders are similarly clustered in areas marked by low and sporadic rainfall (Jetz and Rubenstein 2011), so it seems that in vertebrates generally, cooperative breeding has been driven by the mutual interplay of life history traits and ecological factors (Griesser et al. 2017; Hatchwell and Komdeur 2000).

Although life history parameters have been causally involved in the evolution of cooperative breeding, once evolved, cooperative rearing can drive further selection on life history traits. The incidence of extreme litter sizes and high rates of reproduction in cooperative breeding mammals implies that this has been the case with respect to fecundity.

This is best illustrated by the exceptional fecundity of the naked mole-rat Heterocephalus glaber, where females produce litters of 20-30 pups (mean = 11.4, Sherman et al. 1999), as compared to the 2-4 pups that are more typical of solitary mole-rats. Likewise, the cooperatively breeding marmosets and tamarins exhibit unusually high fecundity within the higher primates (Harris et al. 2014), being the only representatives of this clade to show consistent production of twins. The link between cooperative breeding and increases in fecundity is further supported by unusual patterns of secondary growth in reproductive females in various species (Dengler-Crish and Catania 2007; Young and Bennett 2010b), as well as observations of reduced inter-birth intervals with increases in group size (Creel and Creel

2002; Russell et al. 2003). These findings are presumably explained by alloparents reducing the energetic burden on mothers. A similar argument has been made to explain the positive correlation between the amount of help provided by non-mothers and brain size in most mammalian groups (Isler and van Schaik, 2012), though this view, encapsulated within the wider ‘cooperative breeding hypothesis’ of Burkart and van Schaik (2016; 2009) has been heavily challenged on both conceptual and empirical grounds (Thornton et al. 2016).

Cooperative breeding has also been implicated in longer lifespan. In birds, long lifespan is a precursor of cooperative breeding, whilst in mammals, extended longevity is often documented to be a consequence of cooperative rearing (Healy 2015; Williams and

Shattuck 2015); group-living reduces extrinsic mortality and selects for long life. Whilst it is undoubtedly true that some cooperative mammals are exceptionally long-lived for their size, as in the social mole-rats where individuals can live for two to three decades in captivity despite weighting 40-300g (Dammann and Burda 2006; Dammann 2011; Ruby 2018), support

INTRODUCTION

for a consistent association between cooperative rearing and lifespan (i.e. in all the major groups containing cooperative breeders) is equivocal; a point I raise in my thesis.

If we look beyond its influence on demographic parameters, cooperative breeding is also related to atypical patterns of reproductive competition and behaviour. In most obligate cooperative breeders, reproduction is restricted to a few individuals of either sex within a group, generating intense intrasexual competition in both males and females over access to mating opportunities. This is manifest in high frequencies of aggressive encounters, but because the variance in reproductive success in some cooperative breeders is higher in females than males (Hauber and Lacey 2005), it is often in females that the most conspicuous forms of aggression occur (Clarke and Faulkes 2001; Kutsukake and Clutton-Brock 2006; Snowdon and

Pickhard 1999). For example, in cooperatively breeding acorn woodpeckers Melanerpes fomicivorus, 25% of groups contain two or more joint-nesting females that cooperate in various forms of communal behaviour. When joint-nesting occurs, breeding females frequently destroy the eggs of any females with which they are breeding synchronously, resulting in approximately 40% of eggs laid in groups with joint-nesters being destroyed (Koenig et al.

1995). In strictly singular cooperative breeders, stress-induced abortion, eviction and infanticide are all used by dominant females to suppress the reproductive axis of subordinates

(Clutton-Brock and Manser, 2016), as seen in African wild dogs Lycaon pictus, common marmosets Callithrix jacchus, dwarf mongooses Helogale parvula, and meerkats (Creel and

Waser 1997; Creel et al. 1997; Saltzman et al. 2009). Sometimes, evictions are temporary, as evicted females are allowed back in the group once the offspring of the dominant have reached a stage of development when they are no longer threatened by infanticide: (as in meerkats, Young et al. 2006). At other times, eviction is permanent, and when this occurs it is presumably to reduce the likelihood that dominant females will be challenged and overthrown by the oldest and most competitive subordinates (albeit usurpation is rare;

Clutton-Brock 2016). From these examples the inference can be made that subordinate reproduction is routinely under dominance control. In many cases incest avoidance may also contribute to the low reproductive rates of subordinates, because individuals frequently lack access to unrelated partners within their group (Cooney and Bennett 2000; Nelson-Flower et al. 2011).

INTRODUCTION

The social environment can exert pronounced effects on individual developmental trajectories, particularly in species where access to breeding rights is tightly coupled to age and size. Typically, the effect of the social environment on growth has been viewed as a direct consequence of the competitive or cooperative interactions between conspecifics: with more helpers the growth rate of dependent offspring is increased (Russell et al. 2002; Solomon and

French 1997; Sparkman et al. 2011), whilst increases in group size can also intensify feeding competition and suppress growth during independence (Young 2015). Recent work in cooperative breeders has shown that individuals might also play an active role in in adjusting their growth according to current or future breeding opportunities. In both the social mole- rats and meerkats the acquisition of dominance prompts a secondary period of growth that may serve to enhance fecundity or consolidate status (O’Riain and Jarvis 1998; Russell et al.

2004; Young and Bennett 2010). The transition from subordinance to dominance is also met with pronounced endocrinological, behavioural, and neuroanatomical changes (Clutton-

Brock 2016; Creel 2001; Davies et al. 2016; Holmes et al. 2009). Socially-responsive growth plasticity has also been detected in subordinate meerkats. In this species, subordinate helpers increase their food intake and rate of growth when the mass of their closest same-sex rival converges upon their own, a pattern which can be replicated experimentally through targeted provisioning (Huchard et al. 2016). In addition, female meerkat helpers have been observed to display short-term increases in growth rate following the immigration of an unrelated male to the group (Dubuc and Clutton-Brock, In press).

As contributions to cooperative activities are typically state- and age-dependent, individual variation in developmental trajectories can have far-reaching consequences for the distribution of cooperative behaviour. In general, helpers perform a range of alloparental behaviours and variation in helping effort reflects the relative costs and benefits involved: younger individuals help less as they are still investing in growth; individuals in better condition help more; and sex modifies the costs and benefits of specific behaviours

(Bergmüller et al. 2010; Clutton-Brock 2016; Clutton-Brock et al. 2002; Heinsohn and Cockburn

1994; Koenig and Dickinson 2004). By extension, when groups of cooperative vertebrates are composed of individuals of different ages, sexes, and states, the distribution of cooperative effort integrated across the group might bear the characteristics of a division of labour (Arnold

2005). This is somewhat different to the labour divisions observed in many social insects,

INTRODUCTION

where nonreproductive individuals are assorted into discrete castes (temporary or permanent) that are associated with behavioural and morphological specialisation for a given role. This form of labour division is thought to arise from the much greater size of insect societies: increasing group size can reduce the reproductive potential of workers to low enough levels to select for role specialisation in the form of caste differentiation (Bourke 1999).

Specialisation of this form is regarded as a pinnacle of social evolution (Maynard Smith and

Szathmáry 1995) and is one of the major features of the so-called eusocial societies. Societies of comparable organisational complexity are absent in social vertebrates, possibly excepting the social mole-rats, whose social organisation has often drawn parallels to eusocial insects.

SOCIAL MOLE-RATS AND THE CASE FOR EUSOCIALITY

The eusociality concept of social organisation can trace its origins to a 1966 paper written by Suzanne Batra in the Indian Journal of Entomology (Batra 1966). In the paper, Batra uses the term eusocial (literally “good social”) to describe the nesting behaviour she observed in halictine bees, wherein “the nest-founding parent survives to cooperate with a group of her mature daughters, with division of labour”. Shortly afterwards, Michener (1969) and Wilson

(1971, 1975) formalised what is now commonly invoked as the three-part definition of eusociality, defining as eusocial species that have i) cooperative care of offspring, ii) a reproductive division of labour, and iii) overlap of adult generations. By defining societies on these grounds, early proponents of the term sought to set eusocial societies apart as the most advanced form of social organisation, because societies so-defined contained individuals whose behaviour [and morphology] appeared specifically adapted for the performance of selfless tasks - the clearest examples being the sterile worker castes of ants.

In the intervening years there has been much disagreement over the appropriate criteria for defining eusociality. Some authors have claimed that the requirement for a reproductive division of labour within the three-part definition presented above is too vague and should be restricted to societies where nonreproductive individuals are irreversibly sterile and reproductive individuals are permanently modified for their role (the “narrow” definition, Crespi and Yanega 1995). In contrast, others have argued that eusociality should be considered as a continuum with species delineated according to reproductive skew or their

INTRODUCTION

dependency on cooperation (the “broad” sense definition, Gadagkar 1994; Sherman et al.

1995). Whilst the former definition is restricted to invertebrates, the latter broad definition provides scope to include cooperatively breeding vertebrates, and according to some, humans

(Foster and Ratnieks 2005), within the definition of eusociality.

That such discrepancies exist in the application of the eusociality concept detracts from the goal of identifying convergence in the evolution of social complexity. It could be maintained that where behavioural classification is required, it should be tailored to address specific hypotheses related to social organisation (Wcislo 1997). However, introducing a degree of flexibility around any classification scheme is sure to further complicate the semantic issues that pervade comparative approaches to social organisation (Boomsma and

Gawne 2018). A more constructive approach might instead aspire to identify the similarities and contrasts in specific, biologically relevant features of societies, whether that be between invertebrates and vertebrates, or within smaller taxonomic groupings. For example, Beekman et al. (2006) argued that a consideration of developmental divergence is pivotal to any comparisons of reproductive skew across social taxa, and that by specifically seeking to quantify the extent to which individuals within cooperative societies are committed to specific roles, as measurable in behavioural and morphological traits, one can begin to understand the operation of selection at differing levels of organisation, or if one wishes, to ‘unambiguously assess the point of no return clicks of the evolutionary ratchet that constitute major transitions’

(Boomsma and Gawne 2018).

The value of a trait-based approach to the understanding of social complexity is illustrated by a recent study of mammals by Lukas and Clutton-Brock (2018). In this study, it is shown that the expression of traits associated with organisational complexity (a reproductive division of labour, female infanticide, reproductive suppression) tends to preclude the expression of traits associated with relational complexity (a developed dominance hierarchy, rates of aggression, grooming reciprocity). In addition, the two axes of social variation are related to the kinship structure of female groups: societies with complex organisational structure usually display high levels of average kinship between females living in the same group, as in cooperatively breeding species such as African wild dogs or dwarf mongooses, whereas females living in relationally complex societies typically live in groups

INTRODUCTION

with few close relatives, as exemplified by spotted hyenas Crocuta crocutta, plains zebra Equus quagga, and most simian primates (Figure 1.1).

Figure 1.1. Social complexity and kinship in mammalian societies. Data for 43 mammal species were taken from Lukas and Clutton-Brock (2018) and re-plotted to demonstrate the association between two forms of social complexity and kinship. Average kinship refers to the average relatedness between females group members. The axes for relational and organisational complexity display the proportional presence of key traits in either each form of social organisation; for relational: high rates of aggression, a strict dominance hierarchy, reciprocity in grooming interactions; the presence of coalitions; for organisational: alloparental provisioning, infanticide, reproductive division of labour, and reproductive suppression. Points are coloured according to kinship. The regression plane displays the predicted relationship between kinship and social complexity, taken from a linear regression of the form: kinship ~ relational complexity + organisational complexity. Note the absence of species with high levels of both relational and organisational complexity.

INTRODUCTION

While the above cases illustrate the limitations of the eusociality concept as a criterion for social classification, its use has been central to discussions of mole-rat sociality, and specifically, to the notion that mole-rat societies are fundamentally different in structure to other social vertebrates. The aim of this thesis is not to ask whether mole-rats are eusocial.

This question has been addressed many times. Instead, I intend my thesis to focus on distinctly quantifiable features of mole-rat life history that can be placed in the context of other social taxa. However, because of the historic importance that has been placed on the eusociality concept in mole-rat research, I will first outline the historical case for mole-rat eusociality, before noting some important caveats of past studies of mole-rat social organisation that might have distorted perceptions of their uniqueness.

In 1976, the entomologist Richard Alexander laid out a hypothetical model for a eusocial vertebrate, predicting that if found, it would likely have a nest that was safe (1) and expandable (2), and near to an abundant food source (3) that could be acquired with little risk

(4). Alexander was presenting his ideas to an audience that included the mammologist Terry

Vaughan, who was immediately struck by the resemblance of Alexander’s description to naked mole-rats, a subterranean rodent whose ecology and physiology was being studied by

Jennifer Jarvis in Cape Town. Thereafter commenced a dialogue between Jarvis and

Alexander that culminated in the first examination of the naked-mole rat social structure, as presented in the influential 1981 paper in Science, Eusociality in a mammal: cooperative breeding in naked mole-rat colonies.

Jarvis (1981) likened the social organisation of naked mole-rats to certain species of termite (the insect group with which Alexander which most familiar). The initial attribution of eusociality in this species was supported by the clear reproductive division of labour, for female reproduction was restricted to a single ‘queen’; the overlap of generations; and evidence of direct and indirect offspring care in the form of pup thermoregulation, group defence, allocoprophagy, and communal foraging (together meeting the general criteria outlined by the three-point definition of Wilson). The similarity to termites was bolstered by the suggestion that helpers could be stratified into ‘castes’ that differed in their size and frequency of work, and it was proposed that small-bodied individuals might remain in a nonreproductive caste for life. Analogous claims later followed in the Damaraland mole-rat

Fukomys damarensis, a distantly related sister species that evolved its organisational

INTRODUCTION

complexity independently of the naked mole-rat (Faulkes et al. 1997), and from these initial studies, the description of these two subterranean rodents as eusocial in primary research papers thereafter has been ubiquitous, totalling 34.4% and 64.3% in naked mole-rats and

Damaraland mole-rats respectively (Figure 1.2).

a b

Figure 1.2. The temporal use of the term “eusocial” to describe a) Damaraland and b) naked mole- rats in primary research papers since the original 1981 Jarvis Science paper. Papers that focused on each species were identified via an initial literature search using the ‘bibliometrix’ package in R; the title, abstracts and introductions of these papers were then scanned to determine whether the species were being described as eusocial. The literature search was performed on the 8th June 2018.

Additional features have contributed to the apparent resemblance between mole-rats and some social insects. There is good evidence that breeders are morphologically distinct from non-breeders (Dengler-Crish and Catania 2007; O’Riain et al. 2000; Young and Bennett,

2010a) and that breeders live much longer lives than non-breeders in both captive and natural settings, showing few signs of senescence in later life (Dammann and Burda, 2006, 2007;

Dammann et al. 2011; Edrey et al. 2011; Hochberg et al. 2016; Ruby et al. 2018; Schmidt et al.

2013). Furthermore, early studies reported exceptionally high levels of relatedness within groups of naked mole-rats that indicated frequent inbreeding (Faulkes et al. 1997; Reeve et al.

INTRODUCTION

1990), and also identified a putative disperser caste containing individuals that were morphologically, behaviourally and physiologically distinct from other group members

(O’Riain et al. 1996). Finally, it has been repeatedly suggested that the reproductive suppression of non-breeders is maintained by contact with the breeding female, as is known to be the case in many social insects (Bourke and Franks 1995; Oster and Wilson 1978).

Despite the pervasive comparison of mole-rats to insects, historical studies of the mole- rats have exclusively been cross-sectional in design, which has made it difficult to evaluate the contribution of individual developmental to patterns of social organisation observed at the group level. In other cooperative breeding vertebrates, a large proportion of the variation in helping behaviour can be explained by differences in age and state (Clutton-Brock 2016) and in vertebrates more generally, both within-individual and between-individual processes are known to contribute to changes in phenotypic traits (van de Pol and Verhulst 2006). The extent to which this is also true of the social mole-rats is unclear. To fully understand mole- rat life history requires the adoption of a longitudinal approach that tracks the morphology, behaviour and physiology of individuals across their development. Only then can the structure of mole-rat societies be compared to the structure of other societies in the animal kingdom.

INTRODUCTION

AIM AND THESIS STRUCTURE

In this thesis I aim to critically examine variation in life history trajectories of

Damaraland mole-rats. By considering how patterns of growth, behaviour and ageing contribute to the structuring of mole-rat societies, I attempt to overcome the relative impasse that has developed in the debate about whether the African mole-rats are eusocial. I argue, as others have before, that strict adherence to this eusocial – non-eusocial dichotomy is not only misleading, but also hinders attempts to understand how common evolutionary processes have shaped the gradients of sociality apparent in ‘complex’ societies across the vertebrates and invertebrates.

Central to this understanding is the need to quantify the extent to which individuals in social groups are specialised to certain roles, and whether such specialisations are phenotypically inflexible during development (Beekman et al. 2006). Statistical methods in ecology are now eminently well suited to this task. In my first data chapter (Chapter 3), I investigate the development of behaviour in Damaraland mole-rat helpers. Using multilevel models that capture the multinomial character of the behavioural response, I explore the degree to which the behavioural repertoire of mole-rats varies as a function of age, relative mass (‘state’) and group size, and examine behavioural trade-offs operating within individuals across their development; do individuals that frequently exhibit one behaviour

(e.g. work) also display relatively more of less of another behaviour (e.g. ‘food carrying’). In considering such trade-offs, this work has clear implications for how we should view division of labour in mole-rat societies.

In my second data chapter (Chapter 4) I characterise the shape of growth in

Damaraland mole-rat helpers, before investigating the influence of social factors on growth trajectories. In part, this analysis reveals whether patterns of growth in Damaraland mole-rats resemble other cooperative breeders or are more akin to highly specialised insects. More fundamentally, it provides insights into the male-biased sexual size dimorphism of this species. The male dimorphism of Damaraland mole-rats remains difficult to reconcile against the pronounced reproductive skew in females, which would classically be expected to select for enhanced female growth.

INTRODUCTION

Irrespective of whether variation in size and growth among non-breeding individuals is consistent with functional divergence in behaviour, there is good evidence that breeding females in societies of both naked and Damaraland mole-rats are morphologically divergent from non-breeding females, being larger and more elongated. Data from the wild supports the idea that this divergence arises from a modification of skeletal growth following the attainment of dominance (Young and Bennett 2010). However, insightful as this previous treatment was, it could not exclude the possibility that individuals that became dominant were of a different age to the in-group females from which they diverged. Since all females that became dominant in this study were by proxy no longer subordinate to a breeding female themselves, nor could it distinguish whether the growth adjustment of newly dominant females is a response to reproduction itself, or a response to the removal from reproductive suppression they no longer experience. Finally, the study used morphometric information taken in the field to inform their work, so a multivariate description of morphological divergence is currently lacking. In Chapter 5 I overcome these limitations by experimentally manipulating the breeding status of age-matched sibling females – who either remained in the group as a non-breeder, were paired with an unrelated male, or were placed solitarily - and tracked their resultant skeletal morphology using X-rays. In combination with additional

X-ray data from females in captivity and the wild, this work reveals the developmental basis for a remarkable trait.

Another conspicuous trait present in eusocial insects and the social mole-rats is the extended longevity of breeders compared to non-breeders. In social insects, this apparent contradiction to life history theory can be explained by the low extrinsic mortality of the nest- bound queens (Keller and Genoud 1997), and in some instances, proteins present in the sperm of males may also contribute to lifespan extension (Schrempf et al. 2005). The cause of breeder longevity in social mole-rats is likely to have a different basis. One explanation could be that breeders are lazy and devote little effort to energetically expensive work. Alternatively, nonbreeders might be experiencing chronic stress, and chronic stress could impinge on survival.

This second possibility would be consistent with the general tendency for subordinate helpers to disperse away from their natal group in many cooperative breeders; when dispersal is prevented in captivity, this might elicit chronic stress, if for example, subordinates then face increased levels of aggression from the dominant individuals. Nevertheless, Dammann and

INTRODUCTION

Burda (2006) dismissed each of these explanations for the bimodal ageing pattern of Fukomys mole-rats and instead claimed that status-related differences in lifespan are driven by sexual activity, i.e. the act of breeding moves individuals onto a slower ageing trajectory. To date, there is little evidence to support a physiological basis to lifespan extension in breeders. In

Chapter 6, I back-up this with an examination of telomere dynamics in female Damaraland mole-rats in the wild. Telomere length and telomere shortening serve as informative biomarkers of lifespan and ageing rates, but changes in telomere dynamics and their relationship to ageing profiles has not yet been explored in mole-rats. In this chapter I also re- assess the alternative explanations for breeder lifespan extension in the Fukomys mole-rats.

Reconciling the paradox of cooperation requires the application of both ecological and evolutionary perspectives. The above chapters all take a largely ecological perspective, founded on quantifying and describing the distribution of cooperative activities and life history traits between individuals. In contrast, an evolutionary perspective is necessary for elucidating the ancestral conditions leading to sociality, and the evolutionary consequences of sociality. One such consequence I have already mentioned is the alleged relationship between cooperative breeding and increases in maximum lifespan (Healy 2015). If this relationship holds, then a clear signature of cooperative breeding on lifespan should be present in each of the mammalian families that contain both cooperatively and non- cooperatively breeding species. In my final data chapter (Chapter 7), I show that the evidence for a consistent association of cooperative breeding on lifespan is weak. I suggest that previous analyses investigating this topic have not fully accounted for the strong positive covariation between fossoriality (making extensive use of the underground environment) and cooperative breeding: very few mammals are cooperative breeders, and in those that are, amongst the longest-lived of all for their size are the social mole-rats, which live underground and are cooperative breeders, making it difficult to disentangle the relative contribution of these traits to longevity.

2 CHAPTER The Study System and General Methods

______

GENERAL METHODS

STUDY SITE

The data underlying the core of this thesis was collected at the Kuruman River Reserve in the Northern Cape of South Africa (26°48’S, 21° 49’E, Figure 2.1a). The site is the location of a long-running and highly fruitful research project investigating the behavioural ecology of the meerkat Suricata suricatta (overview by Clutton-Brock and Manser 2016), an obligate cooperatively breeding mongoose that exhibits a striking array of alloparental behaviours including babysitting, allolactation, prolonged pup feeding and sentinelling. In 2013, this core research stream was expanded upon to include Damaraland mole-rats, with the great benefit that the newly established laboratory population could be studied alongside a field-based capture-mark-recapture study. I will provide details on both facets of the mole-rat study system, but first I will mention the habitat, climatic and faunal features that characterise the environment they inhabit.

The landscape surrounding the Kuruman River Reserve is principally arid thornveld, typified by loose coverts of scrub (e.g. Acacia mellifera, Grewia flava, Lycium cinereu), dense thickets of driedoring Rhigozum trichotum), and more open plains dominated by grasses

(Eragrostis, Aristida, Stipagrostis and Schmidita spp.). Some trees scatter the open areas (e.g.

Acacia erioloba, Acaia haematoxylon, and Boscia albitrunca), but any large trees are confined to the dry riverbeds. Relief is provided by red sand dunes that seam across the Kalahari landscape, and the red arenosols of which they are composed support a diversity of geophytes including the gemsbok cucumber Acanthosicyos naudinianus, whose starchy tubers form the main component of the Damaraland mole-rat diet.

The year is demarcated into a cold and dry winter (May – September) and a very hot summer (October – April), although large fluctuations in air temperature occur throughout the calendar (Figure 2.1d). The region suffers episodic droughts, but when rain does fall, it is usually concentrated into short summer downpours (Figure 2.1e) that drive the phenology of the vegetation (Figure 2.1b and c) and the reproductive schedules of the fauna. Beneath the surface, soil temperatures reflect the annual variation in ambient environmental temperature, but the variation in temperature declines quickly with increasing depth (Figure 2.2) and so it is of little surprise that many of the local animals escape the extremes of the climate by spending a substantial portion of their time underground.

GENERAL METHODS

A variety of other taxa are studied at the Kuruman River Reserve, including bat-eared foxes Otocyon megalotis, Cape ground squirrels Xerus inauris, fork-tailed drongos Dicrurus adsimilis, southern pied babblers Turdoides bicolor and southern yellow-billed hornbills Tockus leucomelas. Aside from these researched species, the most conspicuous mammals of the region are the large diurnal ungulates – springbok Antidorcas marsupialis, gemsbok Oryx gazella, eland

Taurotragus oryx, red hartebeest Alcephalus buselaphus caama, and blue wildebeest Connochaetes taurus- while many of the more elusive mammals are nocturnal or crepuscular, species such as aardvark Orcteropus afer, aardwolf Proteles cristata, Cape porcupine Hystrix cristata and ground pangolin Smutsia temminckii, all of which are present on the reserve. The predator guild is locally absent because of the conflicts with the livestock and game-hunting industry that dominates local livelihoods, and one must therefore go to the nearby Kgalagadi

Transfrontier Park to see the lions Panthera leo melanochaita, leopards Panthera pardus, cheetahs

Acinonyx jubatus, hyenae (brown – Parahyaena brunnea, spotted- Crocuta crocuta) and black- backed jackals Canis mesomelas that would have historically spanned the region.

GENERAL METHODS

Figure 2.1 a) The distribution of the Damaraland mole-rat in sub-Saharan Africa, with the Kuruman River Reserve positioned at the star [distribution taken from IUCN, 14th Nov 2017]. b) A mole-rat colony during a spell of low rainfall, and c) the same landscape in the weeks following heavy rain. d) Air temperature and e) Rainfall at the Kuruman River Reserve between 1998 and 2017. Points represent the mean monthly maximum temperature (red), minimum temperature (blue) and rainfall (black, ± 1 s.e.m) across the period. Points are scaled to the median within-month variance of each climatic variable. Dashes on the temperature graph show the absolute maximum and minimum temperature across the period.

GENERAL METHODS

Figure 2.2. a) Subterranean temperatures of the red arenosols of the Kuruman River Reserve in 2016. Lines display the average daily soil temperature measured by a temperature probe at a depth of 10, 40 and 80cm beneath the surface. Although mean temperature varies little with changing depth across the range of depths considered, temporal variability is greatly reduced at lower depths.

GENERAL METHODS

THE DAMARALAND MOLE-RAT (Fukomys damarensis)

Taxonomy

The Damaraland mole-rat Fukomys damarensis (formerly Cryptomys damarensis: Ogilby

1838) sits within the Bathyergid family of African mole-rats, a subterranean clade of rodents restricted to sub-Saharan Africa. Historically, whilst there was broad consensus on the placement of the Bathyergidae within the Old World phiomorphs alongside the Petromuridae

(rock rats) and the Thyronomyidae (cane rats: Nedbal 1994), systematics within the family were complicated by high levels of morphological convergence that arose as adaptations to subterranean living. The application of molecular methods over the ensuing two decades has now clarified the situation (Allard and Honeycutt 1992; Faulkes et al. 1997; Van Deale et al.

2007), suggesting that the African mole-rat radiation is separable into six genera: the monotypic and highly social Heterocephalus (H. glaber the naked mole-rat), the solitary-living

Bathyergus, Heliophobius and Georhycus, and the social Cryptomys and Fukomys. Some 30 or so species have currently been described across these genera, though new taxa are being described episodically (Faulkes et al. 2017) and anecdotal evidence suggests further splits are probable, such as the proposition that the widely dispersed Somalian and Kenyan-Ethiopian populations of naked mole-rats represent distinct species (Faulkes and Bennett 2000, p27.).

The Fukomys genus is particularly species-rich (Kock et al. 2006), which Faulkes et al. (2010) attributes to changes in river drainage patterns in the East African Rift over the past 10 million years, promoting allopatry of remnant subpopulations, and speciation. The monotypic genus of Heterocephalus (the naked mole-rat) is ancestral within the family and diverged from its common Bathyergid ancestor more than 30 million years ago. Even so, given that several of the Bathyergid family are solitary and asocial, it appears that the highly derived social systems of Damaraland mole-rats and naked mole-rats arose independently in two disjunct regions of

Africa (Jarvis and Bennett 1993).

Geographic distribution and the aridity-food distribution hypothesis

The distribution of Damaraland mole-rats closely maps onto the arid thornveld biome of southern Africa (Figure 2.1a). Damaraland mole-rats are thus geographically constrained to an area of low and unpredictable rainfall, in keeping with the general close association

GENERAL METHODS between cooperative breeding and aridity in mammals (Lukas and Clutton-Brock 2017) and birds (Jetz and Rubenstein 2011).

The link between cooperative breeding and aridity was explored within the mole-rats

(Faulkes et al. 2004) long before Lukas and Clutton-Brock (2017) highlighted a more general relationship across cooperatively breeding mammals. The mole-rat specific work revealed a strong positive correlation between maximum group size and the coefficient of variability in rainfall (the cooperative giant Zambian mole-rat Fukomys mechowi provides an interesting exception, frequenting mesic areas, Bennett and Aguilar 1995), and importantly, also demonstrated an inverse relationship between group size and geophyte density, indicating that food resources for mole-rats are distributed more patchily in unstable environments.

These interspecific comparisons supported the “aridity food distribution hypothesis” (AFDH) initially proposed by Jarvis (1978) as an extension of ecological constraints hypothesis (Emlen

1982). The hypothesis posits that the high dispersion of food resources in arid environments acts in conjunction with the high energetic costs associated with burrowing for subterranean mole-rats to promote natal philopatry, group-living, and sociality (reviewed in Jarvis et al.

1994). Patterns of recruitment and dispersal in two populations of common mole-rat

Cryptomys hottentotus experiencing contrasting levels of climatic predictability provided further backing for the AFDH (Spinks et al. 2000), which places emphasis on foraging as the key cooperative behaviour of Damaraland mole-rats.

Group dynamics, reproductive suppression and skew

Reproduction in Damaraland mole-rat groups is restricted to a single female and one, or rarely two, males. Unlike in naked mole-rats, where inbreeding is facultatively tolerated

(Reeve et al. 1990; though outbreeding is preferred, Ciszek 2000), Damaraland mole-rats are obligate outbreeders and all breeding events take place between unrelated individuals. In male Damaraland mole-rats, there is little evidence for reproductive suppression aside from minor status-related differences in sperm composition (Maswanganye et al. 1999; Molteno et al. 2004). In contrast, reproductive suppression in females manifests itself in a complete blocking of ovulation in non-breeders (Molteno and Bennett 2000), though changing social circumstances can modify the strength of suppression. For example, the introduction of an

GENERAL METHODS unrelated male can stimulate the recrudescence of ovarian activity in non-breeders (Rickard and Bennett 1997; Clarke et al. 2001), a scenario prompting high levels of female-female aggression that sometimes leads to the usurpation of the incumbent breeder. In addition, the reproductive readiness of nonbreeding females in the wild - measured by the downstream production of luteinising hormone following injection of pituitary gonadotrophin-releasing hormone - is elevated during periods of high rainfall when the likelihood of meeting dispersing males or of dispersing oneself is higher (Young and Bennett 2010). Together, this would suggest that dominance control is not sufficient as the sole proximate mechanism maintaining female reproductive monopoly, and physiological suppression in Damaraland mole-rats is therefore more consistent with “self-restraint” models of reproductive skew theory (Snowdon 1996; Clutton-Brock 1998) founded on incest avoidance. That females remain relatively plastic to changes in mating opportunities bears striking similarities to common marmosets Callithrix jacchus (Abbott 1984; Saltzman et al. 2009) and dwarf mongooses Helogale parvula (Creel and Waser 1997), but deviates from other cooperative breeders such as alpine marmots Marmota marmota (Hӓcklander et al. 2003) and meerkats

(Kutsukake and Clutton-Brock 2006) where high rates of dominant female aggression towards subordinate females are more suggestive of a direct role of dominants on subordinate breeding rates.

Damaraland mole-rat groups in the wild typically contain 10 individuals (in our study population, the largest group contained 26 individuals, mean = 9.47 ± 5.44, median = 8), although a group containing 41 individuals was captured in Namibia (Jarvis and Bennett

1993). The strong reproductive division of labour across groups of this size generates a considerable reproductive skew at the population-level that is further exacerbated by long dominance tenures (Schmidt et al. 2013). Precise estimates of skew in males are elusive, but in females, a five-year mark-recapture study in Namibia found that 92% (370/403) of captured females had never reproduced. Skew in females is thought to be higher than that in males.

High levels of reproductive skew is typically related to intense intrasexual competition and selection for traits that enhance competitive ability such as body size and weaponry

(Andersson 1994). In this context, Damaraland mole-rats pose something of a paradox as males are considerably larger than females (Young and Bennett 2013), and in both sexes in captivity, aggressive interactions are conspicuously rare.

GENERAL METHODS

THE WILD POPULATION

Groups of mole-rats in the wild are revealed by the lines of mounds that they extrude when excavating their tunnels (Figure 2.3a). In our study population, groups are trapped periodically (6 or 12-month intervals) using modified Hickman traps (Hickman 1979) as part of an ongoing capture-mark-recapture study. Traps are baited with sweet potato and are positioned into tunnel systems that are opened by digging (Figure 2.3b and c). After trap setting, traps are checked every 2-3 hours throughout the day and night. On capture, animals are placed into a closed, sand-filled box with other group members, and provided food and shelter (Figure 2.3d). Intermittently, individuals are transported back to the laboratory where they are weighed, measured and blood sampled (see methods below). When transporting animals from the field to the lab, traps are temporarily disabled to prevent individuals being captured and held in traps for long time periods. All individuals captured at the study sites are marked with passive integrated transponder (PIT) tags on first capture to allow individual recognition. After sampling, groups are housed temporarily in semi-natural tunnel systems in the laboratory (see methods below), and once a whole group has been captured, as evidenced by an absence of activity for 24hrs, the animals are all returned to their natural tunnel system. Since trapping began in 2013, more than 70 separate groups have been captured and sampled, facilitating the recapture of many individuals at multiple timepoints.

GENERAL METHODS

Figure 2.3. a) Gemsbok cucumber. b) A tunnel entrance located by digging a trench between two adjacent mounds. c) The placement of the traps. d) A group of mole-rats prior to being transported back to the laboratory for sampling.

GENERAL METHODS

THE LABORATORY POPULATION

Set-up and husbandry

Animals were captured from the local wild population in late 2013 and either maintained in their original group, or unrelated males and females were paired to create new groups. The lab therefore contains a mix of known-age, lab-born individuals, and unknown- age, lab-introduced individuals. Animals can be recognised individually via a unique coloured dye mark applied to their white head patch (Figure 2.4), and secondarily via a PIT tag that is implanted in early life. All groups are housed in self-contained tunnel systems made of polyvinyl-chloride (PVC) pipes that are modified to have transparent plastic ‘windows’ through which behaviour can be observed. Within each tunnel system, pipes connect various compartments that serve as a nest box, a toilet, a food store and a large waste box (Figure 2.5 and 2.6). Depending on the group size, one to three vertical pipes are incorporated into the tunnel design through which clean sand from the surrounding area can be added. Animals clear the sand from the vertical pipes and move it through the tunnel system to the peripheral waste box, thereby gaining access to food placed behind the previously sand-filled tunnel.

Animals are provisioned twice daily (ad libitum) on a diet of sweet potatoes and cucumbers.

Pieces of tissue paper are introduced into the tunnel system periodically, which the mole-rats readily use for nesting material. Tunnel systems are cleaned briefly every day and more thoroughly once a week.

Behavioural and weights sampling

All behavioural data appearing in the thesis is derived from all-occurrence scan sampling (Altmann 1974). Each scan is performed on a group for 12 hours, and individual behaviours are recorded at 4-minute sampling intervals and inputted onto a handheld

Android device using the Pocket Observer software (Noldus Information Technology); generating approximately 180 sampling events per individual per scan session. During observation periods, sand is added to tunnel systems at 2-hour intervals to increase the expression of ‘work’ behaviours. The full ethogram for the Damaraland mole-rat is presented in Appendix A Table 1.

GENERAL METHODS

Weights are acquired by manually removing individuals from their tunnel system and placing them onto an electronic scale (Sartorius TE4100). All individuals are weighed approximately every week until the age of 90 days, and fortnightly thereafter, producing high resolution weights curves for individuals whose age is known.

Figure 2.4. Damaraland mole-rats in the nest, with unique dye marks illustrated.

GENERAL METHODS

Figure 2.5. Typical set-up of a medium-sized tunnel system. Boxes locate key features.

Nest Boxes

Figure 2.6. Schematic of a large-sized tunnel system.

No evidence of task specialisation among

3 Damaraland mole-rat helpers CHAPTER ______

This chapter was written for Animal Behaviour.

Thorley, J., Mendonça, R., Vullioud, P., Torrents-Ticó, M., Zöttl, M., Gaynor, D., and T. Clutton-Brock. 2018. No task specialisation among helpers in Damaraland mole-rats. Animal Behaviour 143: 9-24.

TASK SPECIALISATION

ABSTRACT

The specialisation of individuals in specific behavioural tasks is often attributed either to irreversible differences in development which generate functionally divergent cooperative phenotypes, or to age-related changes in the relative frequency with which individuals perform different cooperative activities; both of which are common in many insect caste systems. However, contrasts in cooperative behaviour can take other forms and, to date, few studies of cooperative behaviour in vertebrates have explored the effects of age, adult phenotype and early development on individual differences in cooperative behaviour in sufficient detail to discriminate between these alternatives. Here, I use multinomial models to quantify the extent of behavioural specialisation within subordinate, nonreproductive

Damaraland mole-rats Fukomys damarensis at different ages. I show that, although there are large differences between individuals in their contribution to cooperative activities, there is no evidence of individual specialisation in cooperative activities that resembles the differences found in insect societies with distinct castes where individual contributions to different activities are negatively related to each other. Instead, individual differences in helping behaviour appear to be the result of age-related changes in the extent to which individuals commit to all forms of helping. A similar pattern is observed in cooperatively breeding meerkats Suricata suricatta, and there is no unequivocal evidence of caste differentiation in any cooperative vertebrate. The multinomial models I employ offer a powerful heuristic tool to explore task specialisation and developmental divergence across social taxa and provide an analytical approach that may be useful in exploring the distribution of different forms of helping behaviour in other cooperative species.

TASK SPECIALISATION

INTRODUCTION

The morphological and behavioural specialisation of individuals to specific tasks is a common feature of complex insect societies (Wilson 1971; Maynard Smith and Szathmáry

1995; Boomsma and Gawne 2018). To infer specialisation it is necessary to show that investment in one cooperative behaviour trades off against investment other forms of cooperative behaviour. Species differ in the extent to which individuals become irreversibly committed to specific roles (Beekman et al. 2006; English et al. 2015) and the extent to which they do so is commonly regarded as an indicator of the complexity of their society on the basis that increased division of labour improves efficiency (Bourke 1999; Oster and Wilson 1978; but see Dornhaus 2008). Some of the most extreme examples are provided by species of ant and termite where discrete and permanent phenotypic differences exist between functionally sterile workers that focus on different tasks, such as brood care, colony defence or foraging

(Michener 1969; Hölldobler and Wilson 1990; Bourke and Franks 1995; Roisin and Korb 2010).

In contrast, in some other social insects, behavioural specialisation is more labile and takes the form of temporal castes where task allocation varies with age as nonreproductive individuals shift from one role to another; as in honey bees, Apis mellifera (Seeley 1982), some lower termites (Noirot and Pasteels 1987; Korb and Hartfelder 2008), and fungus-cultivating ambrosia beetles (Biedermann and Taborsky 2011). Evidence of behavioural specialisation is rare outside of the social insects, but studies of some cooperative mammals have argued that in some species that breed cooperatively, nonreproductive helpers display forms of task specialisation analogous to those of castes in social insects.

The case for behavioural specialisation in cooperatively breeding mammals has been most strongly advanced for several of the social African mole-rats, including the naked mole- rat, Heterocephalus glaber and the Damaraland mole-rat, Fukomys damarensis. In these two species it has been suggested that individuals can be separated into discrete functional groups that differ in their relative contributions to different cooperative activities (Jarvis 1981, Bennett and Jarvis 1988; Bennett and Faulkes, 2000; Scantlebury et al. 2006) and their probability of dispersing (O’Riain et al. 1996), as well as in related aspects of their size and shape (Bennett and Faulkes 2000). However, other studies of the distribution of cooperative behaviour in social mole-rats found continuous rather than discrete differences between individuals in their cooperative contributions (Lacey and Sherman 1991), and a recent study in Damaraland

TASK SPECIALISATION

mole-rats has suggested that helpers do not specialise in specific tasks but rather vary in overall helpfulness (Zöttl, Vullioud et al. 2016).

Determining whether individuals within cooperative societies are behaviourally specialised is more complex than initially appears for the expression of cooperative behaviour can vary between and within individuals in many ways. For example, individuals may differ either in their general contribution to all cooperative activities or in their relative contributions to specific activities. In addition, relative differences in behaviour may be (1) largely driven by age, (2) unrelated to either age or adult phenotype, or (3) associated with contrasts both in adult phenotype and early development, as in the caste systems of many eusocial insects (see

Table 3.1). There may also be many different combinations and sub-divisions of the six distributions of cooperative behaviour shown in Table 3.1. Without longitudinal studies of the behaviour of individuals at different ages, it is often impossible to distinguish between the developmental processes leading to individual differences in behaviour or to allocate societies to different categories. With this information, it is possible to examine the extent to which cooperative behaviours are correlated within individuals, the temporal stability of any correlations across development, and other phenotypic determinants of behaviour, which together underpin the distribution of behaviour across individuals in cooperative societies.

Although earlier studies of social mole-rats have described contrasts in cooperative behaviour between individuals and suggested that they are a consequence of variation in development (Bennett 1988; Burda 1990; Lacey and Sherman 1991), the absence of longitudinal data for individuals has made it impossible to tell whether or not individual differences are a consequence of permanent contrasts in development analogous to those found in insect societies with distinct castes. More recently, Mooney et al. (2015) used a combination of in-group observations and out-of-group tests of pup care and colony defence in naked mole-rats and showed that contributions to different cooperative tasks (work-related behaviour, pup care and colony defence) varied across nonbreeding group members in naked mole-rats, and that the expression of these behaviours was stable across time and across litters.

They also showed that there was a trade-off between pup care and both colony defence and working behaviour that is suggestive of task specialisation. In contrast, recent research on

Damaraland mole-rats has shown that individual contributions to cooperative effort are partly a consequence of differences in age and growth, and partly of variation in contributions

TASK SPECIALISATION

to all forms of cooperative behaviour (including digging, nest building and food carrying:

Zöttl, Vullioud, et al. 2016).

Despite these two previous studies using longitudinal data, it is still not fully clear whether or not there is specialisation in the relative contributions of individuals to different cooperative activities in either species. In the study of naked mole-rats, specific estimates for individual trade-offs were derived from aggregated observational data collected across a period of days rather than months (30 minutes), and each observation period on groups was short in the context of naked mole-rat activity periods (see Riccio and Goldman 2000). In the study of Damaraland mole-rats, behavioural data was similarly aggregated for each individual, and as individuals in the dataset were sampled heterogeneously across development, the estimated correlations did not control for variation in age, sex, size or group conditions, all of which are implicated in the expression of cooperative behaviour in other societies (fish: Bruintjes and Taborsky 2011, Tanaka et al. 2017; mammals: Clutton-Brock 2016; insects: Field et al. 2006, Thomas and Elgar 2003; birds: Koenig and Dickinson 2004).

Consequently, it remains unclear what form the distribution of cooperative activity takes in mole-rats and whether or not individuals specialise in particular tasks, as has been suggested

(Table 3.1).

In this chapter, I analyse longitudinal records of the development of behaviour in individually marked nonreproductive Damaraland mole-rats to examine individual differences in behaviour and quantify individual correlations across cooperative behaviours to determine whether or not these are negative. I do so using multilevel, multinomial logistic regressions. These statistical models (a form of generalised linear mixed model) are well suited to the structure of observational data but have seldom been used in the context of animal behaviour (see Koster and McElreath 2017). By treating behaviour as a multinomial response, they overcome the need to aggregate across behavioural categories or across observations within individuals when quantifying individual variation in behaviour (e.g.,

Clutton-Brock et al. 2003; Arnold et al. 2005; Zöttl, Vullioud et al. 2016), and therefore allow the estimation of individual-level variance and individual correlations all within the framework of a single model. Trade-offs between different forms of behaviour take the form of correlated random effects, which are here used to elucidate whether mole-rats that

TASK SPECIALISATION

regularly engage in one behaviour (e.g. work) also express relatively more or less of other behaviours (e.g. food carrying).

In addition to asking whether Damaraland mole-rats are behaviourally specialised, I investigate the role of age, group size, relative body mass, and sex on cooperative behaviour.

I also test whether the presence of pups affects the expression of care behaviour in nonreproductive mole-rats, either through direct contributions to nest building, or through increased time spent time in the nest. As Damaraland mole-rat pups are highly altricial and hairless, social thermoregulatory benefits derived from huddling might therefore constitute an important form of social, and arguably, cooperative behaviour (Arnold 1990; Kotze et al.

2008).

TASK SPECIALISATION

Table 3.1. Forms of individual variation in cooperative behaviour across cooperative societies.

Description of variation in cooperative Trade- Early Adult Age Examples behaviour across individuals offs Development Phenotype†

Meerkat Suricata suricattaa No specialisation (temporary or permanent). Differences in all forms of White-winged chough Corcorax melanorhamphosb ✗ ✗ ✓ ✗ cooperative behaviour associated with Social spider Anelosimus eximiusc age. Damaraland mole-rat, Fukomys damarensisd Social spider Anelosimus studiosuse Specialisation in cooperative behaviour ✓ ✗ ✗ ✗ Lion panthera leof independent of age or adult phenotype Chimpanzee Pan troglodyteg Princess of Burundi cichlid Neolamprologus pulcherh Specialisation in cooperative behaviour Honey bee Apis melliferai ✓ ✗ ✓ ✗ associated with age Paper wasp Polistes canadensisj Ambrosia beetle Xyleborinus saxenseniik Leafcutter ant Acromyrmex echinatiorl Specialisation in cooperative behaviour Big-headed ant Pheidole megacephalam associated both with contrasts in adult ✓ ✓ ✗ ✓ n phenotype and in early development Nasute termite Velocitermes barrocoloradensis Aphid Tuberaphis styracio a- Clutton-Brock et al. 2003; b- Heinsohn and Cockburn 1994; c- Settepani et al. 2013; d- Zöttl, Vullioud et al. 2016, this study; e- Wright et al. 2014; f- Stander 1992; g- Boesch 2002; h- Bruintjes and Taborsky 2011; i- Seeley 1982; j- Giray et al. 2005; k- Biedermann and Taborsky 2011; l- Hughes et al. 2003; m- Sameshima et al. 2004; n- Roisin 1996; o- Shibao et al. 2010; † Qualitative non-behavioural differences in adult phenotype.

TASK SPECIALISATION

METHODS

Animal Housing and Data Collection

Data were collected from a captive population of Damaraland mole-rats maintained between October 2013 and January 2017 at the Kuruman River Reserve in the Northern Cape of South Africa. All individuals in this chapter were born in captivity and housed in tunnel systems as outlined in Chapter 2.

Behavioural data were derived from instantaneous scan sampling (Chapter 2) conducted on individuals throughout their development. Intact breeding groups were observed for 12 h in each observation period, with individual behaviour recorded at 4 min sampling intervals (hereafter we refer to a single 12-hour observation period as a scan) and inputted onto a handheld Android device using the Pocket Observer software (Noldus

Information Technology, Wageningen, Netherlands). In this way, 180 sampling events were generated per individual per scan. As our study is concerned with the behaviour of nonreproductive individuals, information from breeding males and females was removed.

The analyses were restricted to 10 scan sessions per individual as a compromise between data coverage and computing requirements. The first and last scan were included for all individuals to ensure maximum age coverage, in addition to eight further randomly chosen scans (mean time between scans per individual = 63.82 ± 2.10 days). The total dataset considered 60 nonreproductive females and 56 nonreproductive males in 35 groups (mean age of first scan = 136.1 ± 0.9 days; mean age at last scan = 716.5 ± 14.3 days, mean age span across scans = 580.4 ± 13.4 days). The ethogram covers 16 behaviours (Appendix A Table 1), which were collapsed into 6 categories for the multinomial modelling; active non-helping, eating, food carrying, nest building, resting, and working. All active non-helping behaviours were grouped together so that a distinction could be made between time allocated to helping

(food carrying, nest building and working) versus more general patterns of activity.

Models: Multilevel multinomial behaviour models

I present a general statistical description of multilevel multinomial behaviour models

(MMBMs, following Koster and McElreath 2017), before outlining the structure of the models

TASK SPECIALISATION

fitted in this study. Being multinomial, discrete behavioural categories follow a categorical

(generalized Bernoulli) distribution, where the probability of observing each behaviour k is separately specified as πk. Of the total K behaviours, one serves as a reference category against which the odds of the other K – 1 behaviours are contrasted in K – 1 ‘sub-equations’.

The multinomial model can be readily extended to include random effects and thereby allow the probability of exhibiting behaviour k to vary across clusters within the data, such as across individuals and groups, or across temporally autocorrelated sampling periods.

Consider a simple scenario in which three behaviours (k = 1,2,3) are modelled as varying across a population of individuals as a function of age. If behaviours recorded at time t are temporally independent, and k = ‘3’ serves as the reference category, then the log-odds of individual i displaying behaviour ‘1’ or’ 2’ instead of the reference behaviour ’3’ is given by:

휋1푖푡 log ( ) = 훽1푖푡 + 훽a1 + 훾1푖 휋3푖푡 휋2푖푡 log ( ) = 훽2푖푡 + 훽a2 + 훾2푖 휋3푖푡 2 훾1푖 휎훾1 ⌊ ⌋ ~ 푁표푟푚푎푙(0, Ω훾) ∶ Ω훾 = [ ] 훾2푖 2 휎훾1,2 휎훾2

휋1 + 휋2 + 휋3 = 1

where β1it and β2it are the intercepts that contrast the first two behaviours against the reference category, βa1 and βa2 are fixed effects for age at the level of each of the first two behaviours relative to the reference category, and γ1i and γ2i are the individual-level random effects, taken to be multivariate normal distributed with zero means. When individual-level intercepts are positive, γki > 0, an individual is more likely than average to exhibit behaviour

K instead of the reference category, and vice versa for negative individual-level intercepts.

The variance-covariance matrix of the individual-level random effects also facilitates the estimation of within-individual correlations across the K - 1 behavioural categories: ρ1,2 = σγ1,2

/ (σγ1.σγ2). Positive correlations indicate that individuals who do more of the first behaviour also do more of the second behaviour (each relative to the reference), whilst negative correlations indicate the opposite. Through these correlations between random effects,

MMBMs explicitly enable the estimation of individual trade-offs in behavioural time budgets.

TASK SPECIALISATION

Models: Fitted Models

I specified three MMBMs for each sex which differ in the specification of the random effects and the presence or absence of fixed covariates. Each sex was analysed separately so that estimated variance components and behavioural trajectories are sex-specific, and by extension, discussions of sex differences in behaviour are qualitatively informed rather than quantitatively informed. I used the Widely Applicable Information Criterion (WAIC) to assess the relative fit of models, but its relevance in the current chapter is somewhat limited as each model provides uniquely important information about the structuring of behavioural variation in mole-rat societies. WAIC is therefore used as a general indicator rather than a model selection tool, and prominence is instead placed on comparisons of the model outputs and changes in the apportionment of variance with increasing model complexity.

Model 1 was limited to the intercepts and random effects at the level of the individual, and therefore reveals i) the extent to which individual-level variance is partitioned across behavioural responses and ii) estimates the [among-individual] correlations across these responses. Since individuals were all been measured repeatedly for over a year of their life, the within-individual random effects correlations here represent individual behavioural correlations across their development (recall that all individuals have been observed an equal number of times). As this chapter focused on individual trade-offs in time allocation during nonresting periods, resting behaviour was set as the reference category throughout modelling

(i.e. correlations between resting and nonresting behaviours are not estimated).

Model 2 retained the random effects at the level of the individual and incorporated several fixed covariates that were hypothesized to be important ecological predictors of behaviour in mole-rats. In differing from Model 1 only in the specification of fixed effects, comparing the first two models provides some information about how much individual-level variance in behavioural categories is accounted for by the fixed effects (notwithstanding some caveats: Koster and McElreath 2017). Because the expression of behaviour in cooperative breeders is often age-dependent, age was included as a first-, second- and third-order polynomial (Zöttl, Vullioud et al. 2016). It was anticipated that group-level processes could also mediate behavioural time budgets and contributions to cooperation, so group size was specified as a first- and second-order polynomial. A categorical covariate for the presence of

TASK SPECIALISATION

pups was included to test the hypothesis that individuals spent more time in the nest when pups were present; we classed pups as animals less than 40 days old. Lastly, as it is common in cooperative societies for individual state to influence contributions to cooperative behaviour, a term for ‘relative mass’ was included that uses the mass of the animal closest to the observation period. In order that mass was estimated relative to other same-sex, same-age group members, ‘relative mass’ represents the residuals from sex-specific linear mixed models that fitted log(mass) as a function of log(age), in the presence of a random term for group identity (Appendix A, Figure 1). All continuous covariates were z-score transformed before model fitting.

Model 3 retained the structure of Model 2 and included further random effects at the level of the scan, the litter, and the group. The inclusion of scan-level random effects controlled for the temporal pseudoreplication introduced by using repeated observations from individuals within a single scan. There was also clustering in the data at the level of the group and the litter; the number of observations at each of these levels was modest, which presumably places low confidence around the estimation of their variances. Their inclusion should nonetheless refine the estimation of the fixed effects. The addition of further random effects also changes the interpretation of the individual-level variances and the within- individual random effects correlations. Notably, the individual-level random effects no longer represent the deviations from the population-level average, making this model unsuited to the estimation of individual trade-offs. Instead, Model 3 was used to describe general effects on the distribution of cooperative behaviour between the sexes and across individuals.

Models were fitted and assessed using the RStan and rethinking packages in R respectively, under a Bayesian framework. In comparison to traditional Markov chain Monte

Carlo approaches, RStan makes use of a Hamiltonian Monte Carlo algorithm for model estimation that requires many-fold fewer iterations before posterior distributions are mixed.

We specified three chains of 1000 iterations for every model, half of which were allocated to the warm-up. As per Koster and McElreath (2017), a non-centred parameterisation of the random effects was specified, using a Cholesky factorization of the variance - covariance matrices. Weakly informative priors were set for the fixed effects parameters and the variance

- covariance matrices and were chosen so that the data influence the posterior values as much

TASK SPECIALISATION

as possible (i.e. priors have only a weak influence on the posterior distribution). Model diagnostics highlighted sufficient mixing of chains for all models.

The significance of the correlated random effects in Models 1 and 2 was evaluated from the credible intervals of their posterior distributions, such that a biologically important effect was inferred in cases where the 95% credible intervals did not overlap zero. For the continuous fixed effects in Models 2 and 3, the predicted probabilities are emphasised above the raw model coefficients for the posterior means, as the latter are difficult to interpret directly because of their relationship to the reference category. The predicted probabilities were only calculated from the fixed effects. For the single categorical fixed effect (presence of pups), we followed the advice of Koster and McElreath (2017) and used the distribution of the contrasts from each posterior sample to test significance, rather than prediction intervals; the intervals incorporate uncertainty from the other fixed covariates and therefore reduce the confidence with which differences among categorical factors can be assessed.

TASK SPECIALISATION

RESULTS

In presenting the results, I first present the evidence for within-individual trade-offs before dealing with general effects on the distribution of cooperative behaviour between the sexes and across individuals. As expected WAIC comparisons highlighted a successively better fit with increasing model complexity (Table 3.2), so the presentation of general effects of age, relative body mass and group size is restricted to Model 3 for each sex.

Table 3.2. Model comparisons for the three models fitted to female and male datasets.

Fixed Model Random Effects WAIC (SE) ΔWAIC Weight Effects

Female 242504.2 1 individual N 7466.6 0 (636.83) 241569.3 2 individual Y 6531.7 0 (635.96) individual, scan, 235037.7 3 Y 0 1 group, litter (637.42)

Male 222472.9 1 individual N 6529.6 0 (613.72) 221382.8 2 individual Y 5439.5 0 (612.78) individual, scan, 215943.3 3 Y 0 1 group, litter (615.10)

TASK SPECIALISATION

Individual Trade-offs

Our analyses provided no evidence for task specialisation. The presence of task specialisation is predicated on negative correlations between different cooperative behaviours at the individual level (trade-offs), but in no case do we detect a significant negative correlation between two behaviours in (excluding the reference category of resting). Instead, nonresting behaviours are positively correlated across development (Appendix A Table 2 for random effects correlations from all models), which suggests that individuals that frequently exhibit one nonresting behaviour also tend to have a high probability of engaging in other nonresting behaviours (Table 3.3). This trend extends to cooperative behaviours: females that work relatively more across their development than the population average were also more frequently observed nest building (ρ4,5 = 0.31 ± 0.12) and food carrying (ρ3,5 = 0.16 ± 0.11), and males that were more frequently working also engaged more often in food carrying (ρ4,8 = 0.24±

0.11). Most of the correlations are strengthened by the addition of fixed effects (Table 3.3 lower), so that after having controlled for general factors affecting behaviour, positive associations between cooperative behaviours predominated (Figure 3.1, from Model 2).

TASK SPECIALISATION

a b

Figure 3.1. Individual-level random effects correlations from Model 2 for a) female b) males. Note that the values presented in the lower half of the matrix represent the correlations between the median individual-level intercept in the posterior samples for each behaviour; they are therefore larger than the correlations presented in Table 3, which are taken directly from the variance-covariance matrices of the posterior samples.

TASK SPECIALISATION

Table 3.3. Correlations of random effects across the responses in each of tested models. The upper half of each matrix denotes the correlations from model 1 for each sex, the lower half from model 2 for each sex. Estimates represent the means from the posterior samples (standard deviations in the parentheses). Parameters in bold indicate estimates where the 95% credible intervals do not span zero.

Active non- Nest Eat Food Carry Work helping Building Females Active non-helping 0.49 (0.12) 0.34 (0.12) 0.28 (0.14) 0.60 (0.09) Eat 0.54 (0.11) 0.10 (0.15) 0.07 (0.17) 0.14 (0.12) Food Carry 0.43 (0.12) 0.07 (0.15) -0.13 (0.15) 0.16 (0.11) Nest Building 0.19 (0.14) 0.17 (0.17) -0.13 (0.15) 0.31 (0.12) Work 0.63 (0.08) 0.28 (0.12) 0.22 (0.10) 0.30 (0.11)

Males Active non-helping 0.62 (0.10) 0.30 (0.12) 0.30 (0.17) 0.68 (0.08) Eat 0.55 (0.11) 0.19 (0.16) -0.31 (0.21) 0.11 (0.12) Food Carry 0.24 (0.13) 0.35 (0.15) -0.04 (0.20) 0.24 (0.11) -0.31 Nest Building 0.27 (0.16) 0.08 (0.19) 0.00 (0.16) (0.20) Work 0.61 (0.09) 0.08 (0.13) 0.22 (0.12) 0.02 (0.16)

TASK SPECIALISATION

General Effects on Mole-rat behaviour

Sex differences in overall time budgets were minimal, with males and females allocating similar amounts of time to each behaviour (coefficients of intercepts- Appendix A

Table 3). The individual variances associated with the behaviours also showed parity between the sexes (Appendix A Table 4 contains all random effects variances). Behaviours with low variance characterised activities that were distributed relatively evenly across individuals, such as eating and active non-helping behaviour, whilst some of the less common activities – nest building and food carrying– displayed high variances and were therefore less consistently expressed across individuals. Since work behaviour was expressed often in males and females but displayed a relatively modest individual-level variance, this suggests that all individuals engaged in appreciable levels of work behaviour.

Age and relative body mass were both major determinants of cooperative contributions in Damaraland mole-rats (Figure 3.2, Figure 3.3, Appendix A Table 5). With respect to age, most behaviours displayed non-linear patterns (Figure 2). Total activity is reflected in the inverse of the predicted curve for rest, indicating that total activity increases until one year of life, before thereafter declining. This general trend in activity was mirrored by analogous age-related changes in cooperative behaviour, with nest building, food carry and work behaviours all being expressed increasingly frequently in the first year of life. Nest building behaviour peaked particularly early, at around 9 months and shows steep declines after this point. The degree to which time allocation to work decline in mid-life seem to be sex- dependent, as marked declines in this behaviour are only apparent in females. With respect to relative body mass, increases in body mass were associated with reductions in nest building in females (invariant in males), but after fixing age to the mean value across the dataset (400 days for females, 396 days for males), a larger relative body mass was associated with increases in both food carrying and work behaviour (Figure 3.3), the latter effect being stronger in males.

Individual behaviour was also influenced by group size (Figure 3.4), and in most cases, the visualisation of quadratic trends suggests that these effects manifest themselves at the upper and lower boundaries of group sizes, where confidence surrounding the estimates is weaker. Nevertheless, the models suggest that the effect of group size on work is sex-

TASK SPECIALISATION

dependent (Figure 3.4, Appendix A Table 5), with increases in group size raising workload in females and reducing workload in males in a quadratic fashion. Beyond this, several behaviours display linear relationships with group size, most notable being the reduction in resting behaviour and food carrying behaviour in females and males respectively.

Males and females did not spend more time in the nest when pups were present

(Appendix A Figure 2), and other aspects of cooperative behaviour were similarly unaffected by the presence of pups (Appendix A Table 5).

TASK SPECIALISATION

Figure 3.2. Model predictions of response behaviours with changing age, a) females, b) males. All other fixed covariates are held at sample mean, with shaded regions specifying the 89% percentile intervals, calculated from the posterior samples of the model 3 for each sex.

TASK SPECIALISATION a

Figure 3.3. Model predictions of response behaviours with changing relative size, upper- females, lower- males. All other fixed covariates are held at the sample mean, with shaded regions specifying the 89% percentile intervals, calculated from the posterior samples of Model 3 for each sex.

TASK SPECIALISATION

Figure 3.4. Model predictions of response behaviours with changing group size, a) females, b) males. All other fixed covariates are held at the sample mean, with shaded regions specifying the 89% percentile intervals, calculated from the posterior samples of Model 3 for each sex.

TASK SPECIALISATION

DISCUSSION

My analysis found no evidence for task specialisation in nonreproductive Damaraland mole-rats. If present, task specialisation should be detectable in the form of individual trade- offs between functionally divergent behaviours. Instead, I found that individual correlations across non-resting behaviours are consistently positive, indicating that individual mole-rats that are more active and spend more time away from their nest tend to engage more in all forms of cooperative behaviour: food carrying, nest building, and work behaviour.

Any division of labour over workload that has previously been suggested in

Damaraland mole-rats from direct observations in captivity (Bennett and Jarvis 1988; Bennett

1990) or indirect measures of activity in the wild (Scantlebury et al. 2006) probably stems from variation in the cooperative contributions of cohorts of animals at different developmental stages and thus sizes (see also Zöttl, Thorley et al. 2016), each of which will affect the relative energetic costs of helping (McNamara and Houston 1992; Clutton-Brock 2016). The absence of longitudinal sampling from known-age individuals in earlier studies made it impossible to determine whether the cooperative contributions of individuals were due to age, or to divergent developmental trajectories in the sense of permanent castes. By incorporating information from known-aged individuals it has become clear that age is a key determinant of cooperative behaviour (Zöttl, Vullioud et al. 2016; this chapter) and the case for permanent castes has been refuted on the basis that all cooperative behaviours show the same trajectory; increasing during ontogeny and decreasing after reaching asymptotic mass. However, it remained possible that individuals could nonetheless be specialised in their cooperative contributions as they age in a manner that might mirror the temporal castes of honey bees

(Seeley 1982); doing relatively more or less of different cooperative activities as they age. Here, in failing to find any evidence of specialisation (trade-offs) across cooperative behaviours in this chapter, I also refute the case for temporal castes, and consequently, it seems that the behavioural differentiation of individuals in Damaraland mole-rat groups is fundamentally different to that observed in eusocial insects, where labour division amongst nonreproductives are associated with behavioural and/or morphological specialisation

(Boomsma and Gawne 2018).

TASK SPECIALISATION

My results add further information regarding the factors affecting behavioural expression in Damaraland mole-rats. I found that the ontogenetic trajectories of behaviour of nonreproductive males and females are extremely similar in both shape and magnitude. Nest building behaviour peaked particularly early, at around 9 months, and shows steep declines in individuals after this point. This relatively infrequent behaviour was therefore mostly performed by young non-breeders of both sexes, perhaps reflecting the lower energetic requirements of nest building compared to working and food carrying behaviour. The lack of overall sex differences deviates somewhat from other cooperative breeders where a substantial component of variation in behaviour is due to sex (Clutton-Brock et al. 2002;

Hodge 2007; Clutton-Brock 2016) and might reflect the similarly negligible opportunities for independent breeding in subordinates of each sex in Damaraland mole-rats, which would be expected to minimise sex-specific divergence in helping strategies (Holmes et al. 2009). The extreme reproductive suppression of subordinate females (Molteno and Bennett 2000) presumably also prevents the evolution of allolactation in the social mole-rats, with females effectively entering into a state of suspended development until reproduction stimulates a secondary burst of ‘puberty-like’ growth and the onset of sexual characteristics (Dengler-

Crish and Catania 2007). One exception where sex differences in helping are apparent in mole- rats is pup care in the form of pup carrying, which has previously been shown to be more frequently performed by females (Zöttl, Vullioud, et al. 2016a; Zöttl et al. 2018). I could not investigate this association in this chapter because I excluded pup carrying from the analysis as it was extremely rarely observed. This decision was based on statistical grounds as rare behaviours are not well accommodated in the modelling framework (Koster and McElreath

2017); incorporating additional information from standardised behavioural assays could be particularly informative when this occurs (e.g. Mooney et al. 2015).

Sex differences aside, the distribution of cooperative behaviour among individuals in

Damaraland mole-rats resembles that in meerkats. Meerkats show a more diverse array of cooperative behaviours than mole-rats, including allolactation, babysitting and pup feeding as well as burrow digging and group defence. Males contribute more to sentinel duty than females who contribute more to babysitting and pup feeding (Clutton-Brock 2002) but, as in

Damaraland mole-rats, all meerkat helpers engage in the full range of activities, and show no evidence of individual specialisation in specific forms of cooperation (Clutton-Brock et al.

TASK SPECIALISATION

2003): relatively heavy female helpers contribute more to most cooperative activities in their first year of life, and in adulthood cooperative contributions are instead driven by increases in daily weight gain, an index of foraging success (Clutton-Brock et al. 2001). The general commitment of different individuals to all forms of cooperative behaviour initially increases up to the second year of life and shows a tendency to decline in older helpers, shortly before they will disperse (Clutton-Brock et al. 2003). Similar processes are likely to explain the age- related declines in helping seen in mole-rats. The precise timing of dispersal in mole-rats in the wild is hard to determine because of the difficulties of ageing wild mole-rats. Nevertheless, loss of individuals from intact groups and recaptures of dispersive individuals suggests that individuals of both sexes remain philopatric for 12-18 months before dispersing (Torrents-

Ticó et al., In review). This timing matches the decline in helping behaviour seen in captivity.

However, if anything, the declines in helping behaviour are more prominent in females, which is at odds with the evidence that males disperse earlier, and more frequently, than females (Hazell et al. 2000; Torrents Ticó et al., In review)

That two species of cooperative breeding mammal fail to show evidence of task specialisation raises important questions about its presence in naked mole-rats. Naked mole- rats remain one of the strongest candidates for task specialisation in the vertebrates, displaying high reproductive skew, extreme group sizes (up to 295 individuals Brett 1991;

Jarvis and Bennett 1993), and socially-induced infertility in nonbreeders (Faulkes et al. 1990;

Faulkes et al. 1991), which together would be expected to enhance selection for a nonreproductive division of labour with task allocation (Bourke 1999). As I have described,

Mooney et al. (2015) suggested that task specialisation occurs in nonreproductive naked-mole- rats, based on evidence of individual consistency in relative contributions to different cooperative activities. However, although they show that contributions to pup care are negatively related to work (digging and colony maintenance) and defensive behaviour, these trade-offs are based upon observations conducted over a period of days rather than the period of months that they used in the same study to demonstrate behavioural consistency within individuals. Their inference of specialisation is therefore indirect and relies on the combined presence of short-term trade-offs and longer-term consistency and at no point are trade-offs measured throughout the development of individuals as is necessary when testing for specialisation. In addition, the ages of individuals included in their analyses are not clear. As

TASK SPECIALISATION

a result, I feel that it remains uncertain whether specialisation and caste formation occurs in non-reproductive naked mole-rats and that further longitudinal data are necessary (i.e. to ascertain where they should fit in Table 3.1).

Firm evidence of task specialisation in other nonhuman social vertebrates is also scarce. Some of most frequently cited examples of specialisation refer to societies engaged in coordinated hunts where individuals repeatedly adopt specific roles, as has been reported in

African lions Panthera leo (Stander 1992), bottlenose dolphins Tursiops truncatus (Gazda et al.

2005), and chimpanzees Pan troglodytes (Boesch 2002). In the case of bottlenose dolphins in

Florida, ‘drivers’ consistently herd fish towards other barrier-forming group members, corralling them into tight shoals that improves the hunting efficiency of the group, whilst in

African lionesses, increases in hunting success were achieved by females repeatedly adopting either a peripheral stalking role or a central attacking role. Presumably such coordinated hunting relies on relatively stable groups where individuals recognise one another and interact repeatedly, allowing individuals to practice and perfect the specific motor controls for their role within what could be defined as a team (Anderson and Franks 2001; albeit many social animals do not have such defined roles when hunting in groups, Lang and Farine 2017).

Other putative examples of specialisation have been presented outside the context of group hunting, and these cases refer more strictly to individual-level trade-offs across cooperative tasks. In cooperatively breeding noisy miners Manorina melanocephala, Arnold et al. (2005) found a negative correlation between helper investment in chick provisioning and predator defence that is indicative of specialisation if maintained across multiple breeding attempts, and in the mound building mouse Mus spicilegus task-related consistency was apparent when collective mound building was induced in captivity (Hurtado et al. 2013). These aside, other cases are limited. This might in part reflect research effort, as few studies appear to have set out with the aim of testing for individual trade-offs within or across cooperative behaviours throughout development. However, given that its quantification falls into the wider and highly topical agenda in behavioural ecology to quantify individual variation in behaviour

(often in the context of ‘animal personality’, ‘behavioural syndromes’, or ‘social niche specialisation’: Bergmüller et al. 2010; Jandt et al. 2014; Montiglio et al. 2013; Wright et al. 2014;

Walton and Toth 2016), it seems probable that task specialisation is uncommon outside of the insects, and that where it does occur in vertebrates it will more often involve cognitively

TASK SPECIALISATION

demanding tasks requiring multiple individuals to cooperate in teams, rather than largely individual tasks where group members receive benefits indirectly.

Studies of the structure of animal societies commonly need to ask questions about the extent and distributions of individual differences in behaviour. Do individuals follow different social trajectories? Do they specialise in certain roles across development? Are specialisations transient, sequential, or irreversible? Are contrasts in development related to changes in gene function or in genotype? The multinomial models that I have employed in this chapter are well suited to address questions of this kind in many different taxa and can provide a common quantitative framework which will make it possible to discriminate the different ways in which individual differences in behaviour develop.

The shape of growth in Damaraland

4 mole-rats: sex and social effects CHAPTER ______

HELPER GROWTH

ABSTRACT

The social environment can impose strong selective pressures on individual growth trajectories in societies where size strongly dictates reproductive opportunities, as is the case in cooperative breeders. Yet few studies in cooperative breeders have examined the influence of social effects on individual growth trajectories beyond the effect of helpers on early life growth. Here, I investigate the effects of sex, group size, and litter size on patterns of growth in nonbreeding Damaraland mole-rats Fukomys damarensis in captivity. A biphasic model outperformed conventional monophasic models and showed that the larger size of males arises from an increased growth rate across two identified phases of development, rather than through any sex differences in the duration of growth. This dimorphism might reflect the different means by which males and females acquire and maintain breeding positions. Despite the suggestion of intense intrasexual competition in mole-rat societies, modification of the biphasic models found that the total number of conspecifics better predicts variation in growth than the number of same-sex conspecifics. This includes a trend for individuals in large groups to grow more slowly (under ad libitum feeding conditions), but slow growth in large groups does not go on to compromise adult body mass.

HELPER GROWTH

INTRODUCTION

The pace at which animals grow and reach reproductive maturity is tightly coupled to a broad suite of traits that defines their life history, including body size, survival, and reproduction (Roff 1992), and as such, identifying the factors underlying variation in growth rates is a central goal of many ecological studies (Arendt 1997; Gaillard et al. 1997; Starck and

Ricklefs 1998). Both environmental and social factors are important in shaping patterns of growth over the course of individual lifespans. Poor environmental conditions and high population density in early life can limit food availability, inhibiting early growth and reducing adult body mass (Festa-Bianchet et al. 2000; Lummaa and Clutton-Brock 2002; Ozgul et al. 2010; Douhard et al. 2013), whilst social parameters can also affect the availability of food through the positive (provisioning) and negative (competition) actions of conspecifics (Wood-

Gush 1971; Clutton-Brock 1991a; Barton 1993; Hudson and Trillmich 2008; Pusey and

Schroepfer-Walker 2013).

In cooperatively breeding vertebrates, where multiple group members participate in the rearing of young, offspring reared in large groups typically grow faster and go on to achieve higher reproductive success than offspring reared in small groups (Emlen and Wrege

1991; Russell et al. 2002; Hodge 2005; Doerr and Doerr 2007; English et al. 2013). The social mole-rats provide an exception to this general trend, with increasing group size being met with reduced growth (Bennett and Navarro 1997; Young et al. 2015; Zöttl, Thorley et al. 2016), but why this is the case is unclear. Unlike in other cooperative breeders, there is a general lack of alloparental care and direct feeding by helpers in mole-rats, and without the direct influence of helpers on offspring development, it is possible that individuals developing in large groups may be food limited and thereby grow more slowly. There is little support that this is the case though, as reduced growth in large groups occurs in captivity despite food being provided ad libitum (Bennett and Navarro 1997; Zöttl, Thorley et al. 2016), and in wild mole-rats, large groups occupy resource-rich sites with higher densities of underground tubers (the principal food source) than smaller groups (Jarvis at al. 1998); large groups persist for longer (naked mole-rats: Jarvis et al. 1994; Damaraland mole-rats: Jarvis et al. 1998); and large groups also have higher rates of offspring recruitment (Young et al. 2015). In the absence of food limitation, competition amongst nonbreeding helpers in large groups might act as the major social factor reducing growth in Damaraland mole-rats, with competitive interactions

HELPER GROWTH

with conspecifics either suppressing or restraining growth. However, aside from this very general influence of group size on growth in mole-rats, the extent to which social parameters influence growth trajectories has been little explored.

Social effects on growth may be particularly strong in cooperative breeders like the social mole-rats because the high variance in reproductive success found in these societies is expected to generate strong intrasexual competition over breeding rights and promote the development of competitive traits (Hauber and Lacey 2005; Clutton-Brock et al. 2006). Several lines of evidence support the role of intrasexual competition on growth in singular cooperative breeders. Dominant breeding individuals are often larger than subordinate helpers and increased early-life growth and resultant size in adulthood raises the likelihood of acquiring and maintaining a dominant position (Hodge et al. 2008; Lardy et al. 2012; English et al. 2013). In several species, the increased size of breeders is aided by a period of secondary growth that increases their relative size advantage over their competitors (Russell et al. 2004;

Dengler-Crish and Catania 2007), and similar episodes of strategic growth adjustment have been observed in other contexts in subordinate individuals. For example, the death of the dominant breeder in naked mole-rats prompts protracted fighting between group members competing for the newly available alpha position (Clarke and Faulkes 1997), and this scenario stimulates a period of accelerated growth in subordinates that presumably serves to enhance their competitive ability (Dengler-Crish and Catania 2007). In addition, recent work in meerkats has shown that subordinate helpers increase their rate of food intake and rate of growth if the mass of a same-sex individual immediately below them in the social hierarchy converges upon their own mass (Huchard et al. 2016), or if an unrelated male immigrates into the group (Dubuc and Clutton-Brock, In press).

As intrasexual competition is likely to scale with group size (because increasing group size is correlated with increasing numbers of same-sex group members), general effects of group size or intrasexual competition on growth could be conflated if they are analysed in isolation, group-level effects being attributed to intrasexual competition, and vice versa.

Taking the social mole-rats, it has been proposed that reduced growth in large groups is a result of general increases in resource competition, but it is feasible that the same pattern could be driven by heightened intrasexual competition in large groups. If the latter predominates, then one could expect the number of same-sex competitors to exert a stronger influence on

HELPER GROWTH

growth trajectories than total group size. To date, no attempt has been made to tease apart group-dependent versus sex-dependent influences on Damaraland mole-rat growth.

As skew is often pronounced in both sexes in cooperative breeders, intrasexual competition tends to be high in males and females (Clutton-Brock 2016), and in several species where reproductive monopolisation in females is particularly high, such as meerkats (Clutton-

Brock et al. 2006) and mole-rats (Lacey and Sherman 1997; Young and Bennett 2013), intrasexual competition may even be higher in females. This view is supported by observations of intolerance between breeding females in cooperative breeders, with dominant breeders frequently expressing hyper-aggressive behaviours such as the killing of non- descendant offspring and the eviction of other female group members (Creel and Waser 1997;

Clutton-Brock et al. 2001; Gilchrist 2006). Dominant females also show elevated levels of testosterone (Clarke and Faulkes 1997; Lutermann et al. 2013; Davies et al. 2016), a hormone implicated in aggression. Yet despite these traits, the larger variance in reproductive success in females in cooperative societies has not led females to be of larger size or sport relatively larger ornaments or armaments as sexual selection theory predicts. The absence of female- biased size dimorphism is conspicuous and difficult to reconcile, and whilst various possible explanations have been put forward by Young and Bennett (2013), testing between the alternatives is difficult as no studies have taken the first step of examining whether the mediation of growth by the social environment varies between the sexes.

Here, I investigate patterns of growth in captive Damaraland mole-rats by asking 1) if there are social effects on growth, 2) at what stage of development do these effects occur, and

3) do they differ between the sexes. Before addressing these questions, it is important that sex differences in growth have been accurately quantified. In Damaraland mole-rats, nonbreeding males are estimated as being 1.4 times heavier than non-breeding females, and whilst previous analyses have stated that this dimorphism arises from an increased rate and a greater growth duration in males (Young and Bennett 2013), the validity of this interpretation is dependent upon the reliability of Gompertz and Logistic models to describe mole-rat growth (Bennett et al. 1991; Bennett and Navarro 1997; O’Riain and Jarvis 1998; Bennett and Faulkes 2000; Young and Bennett 2013; Zöttl et al. 2016b). These models have, and continue, to be applied frequently to mammalian growth data, but given the wide variation in the shape of growth across mammals (Gaillard et al. 1997; Vinicius and Mumby 2013), it is timely that mole-rat

HELPER GROWTH

growth is characterised with more flexible forms of growth. As there is the potential for biological interpretations of growth to be greatly affected by the choice of growth model , the incentive for doing so is clear. I illustrate the importance of model choice by fitting a variety of growth functions to an extensive dataset of Damaraland mole-rats weights taken in captivity. Part of this modelling exercise involves extending the monophasic Chapman-

Richards model into a biphasic form that incorporates two growth parameters. Though multiphasic models are commonly employed in the context of animal breeding (Grossman and Koops 1988; Koops and Grossman 1991) and fisheries research (Alós et al. 2010), they have received little to no attention in less applied ecological areas excepting an analysis of meerkat growth (English et al. 2012). This omission is unfortunate as allowing the growth constant to vary at different developmental stages is biologically intuitive, especially in species like cooperative breeders that often show a relatively prolonged nutritional dependence (i.e. are altricial).

Having characterised the shape of Damaraland mole-rat growth, I go on to examine how social effects influence the growth trajectories of males and females. Specifically, given the proposed role of intrasexual competition in cooperative breeders, I investigate whether sex-dependent or sex-independent social parameters are better able to explain individual variation in growth. To achieve this goal, I modified the parameters of the biphasic growth curve to incorporate both group and litter-level effects that either consider sex-specific terms

(e.g. total number of females, total number of female siblings) or sex-independent terms (total group size, total number of siblings). If intrasexual competition is a key driver of mole-rat growth, then one would expect that the number of same-sex conspecifics will leave a stronger signature on individual growth trajectories than total group size. Similarly, if the main form of competition in cooperative breeders comes from same-aged siblings, as the demonstration of competitive growth in meerkats would imply (Huchard et al. 2016), then one might also expect the number of same-aged siblings (litter size) to impact growth more strongly than overall group size. Directly modifying the curve parameters enables me to determine how each covariate affects each aspect of the growth process: by increasing or decreasing early-life growth, by affecting asymptotic mass, and so forth, thereby pinpointing where in development social effects exert their influence. The overall approach is built upon a mixed- effects modelling framework that is implicitly well equipped to deal with the hierarchical

HELPER GROWTH

clustering inherent in growth data (Sofaer et al. 2013). In so doing I provide an in-depth examination of variation in mole-rat growth which yields new insights about the size structure of mole-rat societies.

METHODS

Study animals

Weights data were collected from the captive population of Damaraland mole-rats maintained between October 2013 and June 2016 at the Kuruman River Reserve in the

Northern Cape of South Africa. All individuals used in this study were of known age, being weighed approximately every week until the age of 90 days, and fortnightly thereafter. The total dataset used in this study comprised 26600 weights taken across 194 females (mean =

66.79 ± 2.79 weights) and 197 males (mean = 69.29 ± 2.81 weights) born into 57 different groups.

For females, only weights taken when individuals were nonbreeders were used (i.e. not a dominant female), as weight curves of reproductive females would be complicated by pregnancy events and possible changes in body size on acquiring dominance status (O’Riain and Jarvis 1998; Chapter 5). As similar status-related changes have not been reported in males, a similar exclusion is not made for this sex.

The shape of Damaraland mole-rat growth

All statistical analyses of growth used non-linear mixed effects models using the nlme package in R, version 3.3.0. Males and females were modelled separately throughout. A

Chapman-Richards model was fitted for each sex. This is a flexible growth function that captures variability in the form of sigmoidal growth through the parameter m (parameterised in Table 4.1 as per Gaillard et al. 1997 and hereafter referred to as ‘Richards’); m = 2 when growth follows the logistic form, approaches 1 when growth is approximately Gompertz, and equals 0 when growth is monomolecular. Four conventional monophasic growth models were also tested to serve as a formal comparison: Gompertz, logistic, von Bertalanffy and monomolecular (Table 4.1, parameterised according to English et al. 2012). For each of the monophasic models, a random effect was specified for the asymptote (A) and the ‘inflexion

HELPER GROWTH

point’ (t0- although for monomolecular growth this parameter is better thought of as the age of onset of growth) at the level of the individual and litter. Additional random effect terms for the growth rate constant were not specified, as doing so caused models to fail to converge, or estimated random effects correlations above 0.9, a sign of overparameterisation (Pinheiro and

Bates 2000). For a similar reason, the correlation between the random effects of the asymptote and the inflexion point at the level of the litter was constrained to 0.

Models were compared using Akaike’s information criterion, the model with lowest

AIC taken as providing the best fit. To minimise heteroscedasticity a power variance function was also consistently applied as per English et al. (2012), and significantly improved the fit of models. Specifically, nlme’s varPower function was set such that observations were assumed to vary normally about a mean given the expected weight of an individual (μ), with a standard deviation parametrised according to ϒ and ρ.

휌 푊푡 ~ N(μ = E(푊푡), σ = ϒ∙휇 )

A biphasic extension

The monophasic form of the Chapman-Richards model was extended to a biphasic model, enabling the estimation of two growth rates around the transition from nutritional dependence to independence. Likelihood profiling was used to determine the most plausible timing of nutritional independence, under the assumption that growth rates would be expected to drop at weaning when highly calorific milk is completely replaced by solid foods.

I ran a series of biphasic curves (parameterised in Table 4.1) where the changepoint between the two phases of the growth was varied for each day between 20 and 100 days of age. The model with the highest log likelihood for each sex was taken to provide the best estimate for the changepoint in growth, which is assumed to reflect the best estimate of nutritional independence.

Mass at birth

To compare the mass of females and males at birth for individuals in our dataset whose exact birth date was known (n = 43 females and 36 males), a general linear model was fitted with weight as the dependent variable and sex and litter size as the independent covariates.

HELPER GROWTH

Group and litter-level effects on growth

The biphasic model was extended by the incorporation of additional fixed effects to test for the influence of social factors on growth. As stated in the introduction, the purpose for doing so was two-fold. Firstly, I wanted to differentiate between group-level and litter-level effects. Secondly, as intrasexual competition is predicted to be high in cooperative breeding mammals, I wanted to examine whether, at each of the two levels, the number of same-sex conspecifics exerted a stronger influence on growth parameters than the total number of conspecifics. Four models were tested for each sex:

M1: Group size + Number of same-age sibs.

M2: Group Size + Number of same-sex same-age sibs.

M3: Number of same-sex group mates + Number of same-age sibs.

M4: Number of same-sex group mates + Number of same-sex same-age sibs

For each model, weight was thus modelled as:

푨 + 푨푮풙푮 + 푨푳풚푳 ퟏ ; 풕 < 풕ퟏ ( ) (ퟏ − (풎 − ퟏ)풆−(풌 + 풌품풙품+ 풌풍풚풍)(풕−풕ퟎ)) 풎−ퟏ

푨 + 푨푮풙푮 + 푨푳풚푳 ퟏ ; 풕 ≥ 풕ퟏ ( ) (ퟏ − (풎 − ퟏ)풆−(풌 + 풌품풙품+ 풌풍풚풍)(풕−풕ퟎ) +(풌ퟐ + 풌ퟐ푮풙푮+ 풌ퟐ푳풚푳)((풕−풕ퟎ)) 풎−ퟏ

, where, as in Table 4.1, A, k, k2, t0, and m refer to the asymptote, the growth rate in the first developmental period (t1 = 50 days for males and 53 days for females), the growth rate in the second developmental period, the inflexion point, and the shape parameter. The additional parameters denote group-level (AG, kg, k2G) and litter-level (AL, kl, k2L) deviations from the population-level for the asymptote or growth rate constant parameter (i.e. a typical individual in a typical litter), with indicator variables (xg, xG, yl, yL) specifying the average group-level or litter level effect in either period of development (lower case if first phase of development, upper case if second phase), scaled to the mean with one standard deviation. For example, for female model 1, kg estimates the change in the first-phase growth rate constant induced by a

HELPER GROWTH

unit deviation increase in the average number of group members in that phase (xg), kG estimates the change in the second-phase growth rate constant induced by a unit deviation increase in the average number of group members in the second phase (xG), AG estimates the change in asymptote induced by a unit deviation increase in the average number of group members in the second phase of growth (xG), and so forth. All random effects were specified as in previous models. Model fit of the four competing models were assessed by AIC values.

Note that although both group- and litter-level metrics were taken as averages in each phase of growth, early life conditions typically reflect later-life conditions in the laboratory set-up

(Appendix B Figure 4), making the fixed effects representative across the growth trajectory.

HELPER GROWTH

RESULTS

A comparison of standard monophasic curves demonstrated that Damaraland mole- rat growth is best described by a monomolecular function (Table 4.1, Figure 4.1), which provided a notably better fit to the data than the Gompertz-like functions that have been employed in previous studies of social mole-rats (Appendix B Figure 1). By extension, the shape parameter of the flexible Chapman-Richards function converged close to zero in males and females, (female, m = -0.11 ± 0.01; male, m = 0.17 ± 0.01). This suggests that the post-natal growth of this species lacks a clear inflexion point, however, a closer examination of growth through parameterisation of a biphasic formulation of the Chapman-Richards function refutes this notion, as the shape parameter for each sex (female, m = 0.98 ± 0.03; male, m = 1.19 ± 0.02) indicates a minor inflexion in growth that is apparent in the first few weeks post-parturition

(Figure 4.2). The biphasic curve identified a clear changepoint in individual growth rate approximately 7-8 weeks following parturition (50 days for males and 53 days for females), after which point the growth rate of males and females decreases markedly (Table 4.2,

Appendix B Figure 2).

Forcing growth to take one of the four specified monophasic forms leads to very different estimates of size-related traits in Damaraland mole-rats (Table 4.3). Whilst the monophasic monomolecular model appears to circumvent the overestimation of early life body mass that occurs when more convex growth curves are fitted (Appendix B Figure 1), it results in spuriously high estimates of growth duration and sexual size dimorphism; if the biphasic function is taken to most accurately describe male and female growth. From the biphasic model, Damaraland mole-rats display a sexual size dimorphism of 1.30. This dimorphism is apparent at birth (male mean = 10.58 ± 0.31, female mean = 9.77 ± 0.28, GLM: β

= 0.80 ± 0.40, test statistic = 1.98, p = 0.051) and is amplified throughout development not from differences in the duration of growth, but rather from a higher growth rate in males across the two identified phases of development (Table 4.2). The parameter estimates of the biphasic models proved robust to major reductions in the number of records per individual

(resampling approach described in Appendix B Figure 3).

Table 4.1. Comparison of growth curves of female and male Damaraland mole-rats. Equations are parameterised as in English et al. (2012). E(Wt) is the predicted weight at time t, A is the asymptotic mass, k is the growth rate constant, and t0 is the age of onset of growth (monomolecular model) or the inflexion point (other models), m a growth-form parameter of the Chapman-Richard’s model. The biphasic Richard’s model contains an additional parameter (k2) that allows the estimation of a separate growth rate constant either side of nutritional independence. -ln[L] is the negative log likelihood.

Growth E(Wt) Females Males Curves -ln[L] AIC ΔAIC -ln[L] AIC ΔAIC

Monophasic 푨 Logistic ퟏ + 풆−풌(풕−풕ퟎ) 48740.3 97500.7 7838.8 51420.9 102861.8 6200.5 −풌(풕−풕 ) Gompertz 푨풆−풆 ퟎ 47384.8 94789.6 5127.7 50264.6 100549.2 3887.9 ퟏ Von 푨(ퟏ − 풆−풌(풕−풕ퟎ))ퟑ Bertalanffy ퟑ 46810.5 93641.1 3979.2 49874.0 99768.1 3106.8 Monomolecul 푨(ퟏ − 풆−풌(풕−풕ퟎ)) ar 45701.1 91422.2 1760.3 49658.2 99336.9 2675.6 푨 Richards ퟏ ( ) (ퟏ − (풎 − ퟏ)풆−풌(풕−풕ퟎ)) 풎−ퟏ 45664.0 91351.5 1689.6 49516.8 99055.5 2394.2 Biphasic 푨 ퟏ ; 풕 < 풕ퟏ ( ) Richards (ퟏ − (풎 − ퟏ)풆−풌(풕−풕ퟎ)) 풎−ퟏ 푨 44831.4 89661.9 0.00 48318.6 96661.3 0.00 ퟏ ; 풕 ≥ 풕ퟏ ( ) (ퟏ − (풎 − ퟏ)풆−풌(풕−풕ퟎ) +풌ퟐ(풕−풕ퟎ)) 풎−ퟏ

Figure 4.1. Four monophasic growth curves fitted to female (a) and male (b) Damaraland mole-rats.

HELPER GROWTH

Table 4.2. Model output for biphasic Richards growth model in female and male mole- rats.

Male Fixed Random Std. Cor t0,ind- Est SE Effect Effect dev. Aind

A 195.73 3.46 Aind 35.42 0.849

k 0.0145 0.0004 Alitter 21.91

k2 0.0049 0.000007 t0,ind 9.50

t0 104.10 0.95 t0,litter 3.52 m 1.27 0.03

Female Fixed Random Std. Cor t0,ind- Est SE Effect Effect dev. Aind

A 150.99 2.51 Aind 28.35 0.822

k 0.0117 0.0004 Alitter 13.74

k2 0.0044 0.00007 t0,ind 8.47

t0 83.67 1.00 t0,litter 4.77 m 1.00 0.03

Aind, Alitter, t0,ind, t0,litter refer to a random term for asymptote and ‘inflexion point’ at the levels of the individual and the litter.

Table 4.3. The influence of the choice of growth form on size traits.

Male: time to Female: time 95% to 95% Male Female Sexual Size asymptotic asymptotic Model Asymptote Asymptote Dimorphism mass / years mass / years Monophasic Logistic 190.37 143.22 1.33 1.77 1.77 Gompertz 201.57 151.26 1.33 2.21 1.60 Von Bertalanffy 205.92 153.21 1.34 2.49 2.13 Monomolecular 235.78 162.82 1.45 2.49 3.00 Richards 220.24 169.55 1.30 3.00 3.00 Biphasic Richards 195.73 150.99 1.30 2.24 2.21

HELPER GROWTH

Figure 4.2. The shape of Damaraland mole-rat growth predicted from biphasic monomolecular models. Points represent raw data for females (blue) and males (red); lines display predicted weight for females (lower) and males (upper). The inset subplot is concentrated upon the period of early development (30 – 150 days), with the vertical line indicating the changepoint between phases in the growth curve; 53 and 50 days for females and males respectively.

HELPER GROWTH

Adjustment of the biphasic model highlighted an influence of social factors on

Damaraland mole-rat growth (Figure 4.3). Model fitting indicated that the total group size and the total number of same-age siblings (‘litter size’) outperformed the number of same-sex group mates and the number of same-sex same-age siblings respectively as variables linked to growth (Table 4.4). Thus, the aggregate interactions with conspecifics seem to supersede sex-specific interactions in shaping growth trajectories. Closer inspection of model parameters

(Appendix B Table 2) revealed a strong negative relationship between group size and the second-phase growth rate in male (k2G = -0.00061 ± 0.00003) and female mole-rats (k2G = -0.00080

± 0.00003, p < 0.001). In contrast, there is no evidence that group size influenced growth rate in the first developmental phase, the time when young are nutritionally dependent. Nor did it affect asymptotic mass, such that any dampening effect of group size on growth does not appear to constrain adult size.

The association between litter size and growth differed according to sex. An effect in early life was only detected in males, with those individuals reared in large litters showing heightened first-phase growth (kl = 0.00061 ± 0.00008, p < 0.001). After nutritional independence, these same males in larger litters experienced reduced growth rates (k2L = -

0.00018 ± 0.00003, p < 0.001), but went on to attain a higher asymptotic mass (AL = 6.05 ± 2.09, p = 0.037). The reverse pattern held for females, the faster-growing individuals in large litters

(k2L = 0.00031 ± 0.00003, p < 0.001) being smaller in adulthood (AL = -6.44 ± 2.08g, p = 0.019). As above, the parameter estimates of the biphasic models proved robust to randomised dataset reductions (Appendix B Table 2); the significance of the added fixed effects is therefore unlikely to represent an artefact of large sample size.

HELPER GROWTH

Figure 4.3. The effect of social factors on male (red, upper) and female (blue, lower) Damaraland mole-rat growth; lines display predicted weight based on parameters from the best-fitting models (Table 4.1).

HELPER GROWTH

Table 4.4. Model comparisons for biphasic Chapman-Richards growth models that specified different group-level and litter-level effects on parameters of the growth curve.

Model Total Total same-sex Total Group group Total same-sex Size members siblings siblings -ln[LN] AIC ΔAIC 1 Y Y 44508.1 89052.2 403.9 2 Y Y 44555.1 89146.2 309.9 Females 3 Y Y 44627.5 89291.0 165.1 4 Y Y 44616.9 89456.1 0.0 1 Y Y 47866.7 95769.4 380.1 2 Y Y 47903.9 95843.5 306.0 Males 3 Y Y 47974.6 95985.2 164.3 4 Y Y 48056.7 96149.5 0.0

HELPER GROWTH

DISCUSSION

This chapter examines sex differences in the growth of Damaraland mole-rats and investigates whether social effects influence growth trajectories in either sex. Although sigmoidal curves are often thought to be a good general descriptor of mammalian growth

(Zullinger et al. 1984; Ricklefs 2010), I show that for mole-rats greater inferential insights can be gained by allowing growth to follow a biphasic trajectory. The biphasic model identified an increased rate of male growth across development, and further adjustment to this model found that group and litter-level attributes influence individual growth variation. In explaining this variation, there was little support for a strong role of intrasexual competition on growth as might be expected for a cooperative breeder with large variance in reproductive success between individuals.

Whilst previous studies have assumed sigmoidal growth in mole-rats, a much better fit to the data was achieved by allowing growth to be biphasic. The biphasic model tested was a novel extension of the Chapman-Richards growth curve and incorporated two growth rate constants and two shape parameters that could vary either side of a changepoint identified through likelihood profiling. This highly flexible approach partitioned growth into an initial phase of high growth in the first 7-8 weeks post-parturition (50 days for males and 53 days for females), and a slower second phase of growth thereafter. It is plausible that this changepoint reflects the transition of individuals to nutritional independence, which places the timing of weaning later than the four weeks previously noted by Bennett and Jarvis (2004) and aligns with observations of female lactation in captivity. Because of its flexibility, the biphasic model also provided far more accurate estimates of early mass, growth duration and asymptotic mass than standard growth curves. For example, the Gompertz and logistic curves that have routinely been applied to mole-rats yielded poor estimates of early-life mass and underestimated both growth duration and asymptotic mass; if the biphasic model is taken to most accurately describe growth. In contrast, if growth is taken to be monomolecular-like as comparisons of the monophasic curves would advocate, then growth duration and asymptotic mass would be overestimated, and the degree of sexual dimorphism overinflated.

Monomolecular growth also implies the absence of an inflexion point, a feature that is more typical of precocial species (Gaillard et al. 1997) and would therefore be unusual in

HELPER GROWTH

Damaraland mole-rats where offspring are altricial. That the biphasic model identifies a subtle inflexion in early life is therefore more consistent with broader comparative work.

The models also highlighted that the male-biased size dimorphism is not due to sex differences in growth duration as previously documented for Damaraland mole-rats (Young and Bennett 2013), but instead arises from the higher growth rate of males across development. At birth, males are already 1.08 times larger than females and the subsequent divergence in growth rates during the first and second phase of growth generates a size dimorphism in adulthood of 1.30. As Damaraland mole-rats are polytocous it seems unlikely that the faster growth of males in the first phase of growth stems from biased maternal investment towards sons during lactation, as there is little support for sex-biased provisioning in even the most size-dimorphic mammals producing single offspring (e.g. California sea lion

Zalophus californianus, Ono and Boness 1996; southern elephant seals Mirounga leonina,

Wilkinson and van Aarde 2001). It is therefore more likely that any sex differences in nursing behaviour, if indeed they exist, are a result of contrasts in offspring demand (Clutton-Brock

1991b). Male mole-rats may also be more sensitive to conditions during development and this could facilitate their capacity to allocate resources to growth; as is seen in dimorphic polygynous mammals (Badyaev 2002).

Sex differences in growth and size could also be a consequence of sex differences in reproductive competition, but examination of patterns of growth in this study do not provide clear evidence that this is the case in Damaraland mole-rats. On the contrary, adjustments of the biphasic growth model showed that at the level of the group and the litter, the total number of conspecifics explained more variation in growth than the number of same-sex conspecifics, implying that aggregate interactions with group members supersede interactions with conspecifics in shaping individual growth. At the group level, these social effects are manifest in growth reductions in large groups, recovering the pattern documented elsewhere in captive (Bennett and Navarro 1997; Zöttl, Thorley et al. 2016) and wild mole-rats

(Young et al. 2015). Notwithstanding the general effects of resource competition, Young and

Bennett (2015) postulated that lowered growth rate in large groups may reflect i) socially induced growth suppression imposed by the greater number of larger, more dominant individuals in bigger groups, or ii) active growth restraint on the part of subordinates, if, for example, growth poses a threat to larger and higher-ranking individuals (as in some group-

HELPER GROWTH

living fish:). It is difficult to distinguish between these two processes without the incorporation of behavioural data and targeted experiments. Nevertheless, discussions of socially induced growth suppression usually invoke aggression from dominant individuals towards subordinate individuals, and such aggressive encounters are conspicuously rare in

Damaraland mole-rats (although the threat of punishment may alone be sufficient: Cant 2011;

Cant et al. 2014). Moreover, this study shows that the slower growing individuals in larger groups do not go on to attain a lower asymptotic mass, as would be anticipated if growth suppression were operating. This leaves open the possibility that individuals might restrain their growth in large groups. This is an intriguing possibility, for fast growth is commonly costly (Metcalfe and Monaghan 2001), and whereas mole-rats developing in large groups can benefit from the combined actions of a large workforce, principally to extend foraging tunnels, small groups cannot. Individuals in small groups might therefore be required to grow quickly to provide a greater contribution to cooperative tasks, which are positively related to body mass (Chapter 3). This could in part explain why slow-growing individuals in large groups show higher rates of recruitment (Young et al. 2015).

It is possible that the absence of strong intrasexual effects on variation in mole-rat growth might be due to the composition of social groups in the study. As Damaraland mole- rats actively avoid inbreeding, non-breeding females only provide an imminent threat to breeders when they have access to an unrelated male, and as such, it is only on the introduction of foreign males that sexual activity in non-breeders is elicited, levels of female- female aggression are elevated, and dominance is challenged (Cooney and Bennett 2000). Very few, if any, of the females in this study were living in a group with an unrelated male (groups were either captured from the wild or created through pairing), and this might reduce the scope for social conflict to influence growth patterns. Nevertheless, immigrant males are often captured in wild mole-rat groups (Burland et al. 2004) and it therefore seems plausible that individuals in captive groups might modulate their own growth or the growth of others in anticipation of a hypothetical immigration event. It is also conceivable that female-female competition in mole-rat societies mostly takes the form of dominant female suppression (of reproduction and/or growth), in which event the total number of female group members or the total number of female siblings does not provide a useful metric of intrasexual competition. Although breeding females are behaviourally dominant over non-breeding

HELPER GROWTH

females (Jacobs et al. 1991), it is not currently clear whether breeders actively suppress the growth of subordinates in any cooperative breeding mammal. Further work that manipulates the social context of individuals in mole-rat societies will be needed to further examine the role of intrasexual competition on growth.

In addition to the question of how males get bigger than females, there is the question of why they are bigger, because the large variance in reproductive success in female mole-rats is predicted to generate strong selection for increased size in females. Young and Bennett

(2013) propose several factors that may explain male-biased size dimorphism in Damaraland mole-rats and cooperative breeders in general, but perhaps the main selective pressure shaping dimorphism is the differential means by which males and females acquire and maintain a breeding position. As obligate outbreeders, male mole-rats must disperse from their natal group and then immigrate into established breeding groups to acquire breeding opportunities. To do so they must challenge the incumbent dominant male who is highly xenophobic towards intruders (Cooney 2002). This requirement to fight to acquire and later maintain breeding rights in the face of immigrants might select for increases in male size and explain the shorter philopatry and shorter tenures of males (Torrents Ticó et al., In press;

Young and Bennett 2013) when compared to females, whose route to reproduction is of a different form. Although some females may inherit breeding rights locally should the dominant female die (Torrents Ticó et al., In press), the low mortality and prolonged lifespan of the single breeding females (Schmidt et al. 2013; see also Dammann et al. 2011) makes natal inheritance extremely rare, and even in the event it is achieved, it is not necessarily the case that an unrelated male will be present. Instead, capture-mark-recapture studies in the

Kalahari suggest that females routinely disperse and settle solitarily, and thereafter wait to be found by dispersing males. This scenario reduces the role of female-female competition in the attainment of breeding status and may temper selection on female size.

Overall, this chapter extends our understanding of mole-rat growth. Firstly, it accurately describes sex differences in growth and identifies how male-biased size dimorphism arises. Secondly, it shows that reductions in growth experienced in large groups are not clearly driven by intrasexual competition between conspecific group members and are instead more likely to represent a general effect of resource competition or active growth restraint. Targeted experiments in captive animals offer a clear avenue to explore the role of

HELPER GROWTH

the social environment on growth further. Mole-rats, like other cooperative breeders, should be particularly interesting in this area, because contrasts in size can have large implications, which is presumably why plastic responses to the risk competition have already been shown in these taxa. Finally, this chapter has highlighted the advantages of non-linear mixed effects models for understanding mammalian growth, with random effects enabling the estimation of variation in growth parameters at different levels of hierarchical organisation, and the modification of fixed effects permitting the statistical testing of growth-related hypotheses.

These models offer a powerful statistical framework to advance our understanding of the ecological processes affecting growth and the shaping of growth patterns on evolutionary timescales.

5 Reproduction triggers adaptive increases

CHAPTER in size in female Damaraland mole-rats

______

This chapter has been published by Proceedings of the Royal Society of London B: Biological Sciences.

Thorley, J., Katlein, N., Goddard, K., Zöttl, M., and T. Clutton-Brock. 2018. Reproduction triggers adaptive increases in body size in female mole-rats. Proc. R. Soc. B. 285: 20180897

MORPHOLOGICAL DIVERGENCE

ABSTRACT

In social mole-rats, breeding females are larger and more elongated than nonbreeding female helpers. This status-related morphological divergence is thought to arise from modifications of skeletal growth following the death or removal of the previous breeder and the transition of their successors from a nonbreeding to a breeding role. However, it is not clear what changes in growth are involved, whether they are stimulated by the relaxation of reproductive suppression or by changes in breeding status, or whether they are associated with fecundity increases. Here, I show that, in captive Damaraland mole-rats (Fukomys damarensis) where breeding was experimentally controlled in age-matched siblings, individuals changed in size and shape through a lengthening of the lumbar vertebrae when they began breeding. This skeletal remodelling results from changes in breeding status since i) females removed from a group setting and placed solitarily showed no increases in growth, and ii) females dispersing from natural groups that have not yet bred do not differ in size and shape from helpers in established groups. Growth patterns in Damaraland mole-rats cannot therefore be explained by early-life developmental divergence or by status-specific age differences, as in many eusocial insects, and instead resemble those of naked mole-rats Heterocephalus glaber and other social vertebrates, where contrasts in size and shape follow the acquisition of the breeding role. The results also suggest that the increases in female size provide fecundity benefits as longer-bodied females had larger litters and produced relatively heavier pups (after the effects of litter size were controlled for) than shorter-bodied individuals.

MORPHOLOGICAL DIVERGENCE

INTRODUCTION

In several social vertebrates where a single dominant female monopolizes reproduction in each group, subordinates that acquire a dominant breeding position display an increase in their growth rate (fish: Hoffman et al. 1999; Heg et al. 2004; Buston 2003; meerkats Suricata suricatta: Russell et al. 2004; naked mole-rats: Dengler-Crish and Catania

2007). For example, in the social mole-rats (including the naked mole-rat and the Damaraland mole-rat), dominant breeding females are both larger and more elongated than nonbreeding subordinates (O’Riain et al. 2000; Young and Bennett 2010) and longitudinal studies of individuals show that subordinate females removed from established breeding groups and paired with novel partners increase in size and weight (Dengler-Crish and Catania 2007). As body size and weight typically confer competitive advantages, increases in growth in newly dominant individuals may help to consolidate their position and to increase their fecundity, with parallels being drawn between the elongated phenotype of female breeders in mole-rats and the physogastry (enlargement of the abdomen through increasing numbers of ovarioles) observed in queens of some eusocial insect societies (Jarvis et al. 1991; O’Riain et al. 2000;

Young and Bennett 2010).

While the presence of status-related changes in growth in social mole-rats is well established, their immediate causes are still uncertain, with studies implicating different growth patterns in naked mole-rats and Damaraland mole-rats. In naked mole-rats, longitudinal X-ray sampling of captive newly created breeders revealed upregulated growth of the lumbar vertebrae (O’Riain et al. 2000; Dengler-Crish and Catania 2007) and episodic bursts of vertebral growth in successive periods of pregnancy (Henry et al. 2007). This causes breeding females to exhibit a longer body length relative to their skull width. The same status- related morphological difference is seen in Damaraland mole-rats, but repeated measurements of wild females suggested that breeder elongation is not driven by upregulated vertebral growth (Young and Bennett 2010) like in naked mole-rats. Instead it appeared to originate from a decrease in the relative growth of the skull (zygomatic arch width) compared to growth towards total body length (Young and Bennett 2010). This result prompted the idea that female Damaraland mole-rats reallocate resources from growth towards reproduction as they become dominant; during this reallocation, skull growth is reduced but growth towards body length is maintained because of the inherent fitness benefits of increased body length

MORPHOLOGICAL DIVERGENCE

(Young and Bennett 2010). The focus on growth reductions in skull size is particularly pertinent in this argument as the skull of mole-rats is home to their prominent buccal incisors that are so important for soil excavation (involved in foraging and burrow maintenance), and as a result, reproductive investment in the form of elongation can be argued to trade off directly against investment into work.

As sociality arose independently in these two subterranean rodents (Faulkes et al.

1997), it is plausible that the attainment of elongation (which is assumed to be directly related to their sociality) occurs through alternate developmental routes. However, as the prior analysis of Damaraland mole-rats was based on human measurements of morphological traits taken with callipers and a tape measure, a formal characterisation of morphological divergence at a skeletal level is currently missing. Moreover, although Young and Bennett

(2010) go to great lengths to remove the possibility that status-related age differences underpin skeletal divergence - because changes in morphology only occurred in females after transitioning to dominance, and because the opportunity to acquire dominance is somewhat stochastic- a contribution of age to breeder elongation has not been definitively ruled out as individuals in the wild were of unknown age.

It is also unclear whether the changes in growth in female Damaraland mole-rats that acquire the breeding position in their group are stimulated by the relaxation of reproductive suppression by the previous dominant female or by the onset of reproduction itself. The degree of reproductive suppression in the social mole-rats is extreme, manifesting itself in a complete blocking of ovulation in nonbreeding Damaraland mole-rats (Molteno and Bennett

2000). Even so, the introduction of an unrelated male results in a recrudescence of ovarian activity in nonbreeders (Rickard and Bennett 1997; Clarke et al. 2001) and stimulates high levels of aggression between females that sometimes leads to the usurpation of the incumbent breeder (Cooney and Bennett 2000; see Faulkes and Abbott 1993; Clarke and Faulkes 1997, for comparable results in naked mole-rats). Nonbreeding females also start ovulating in the absence of breeding females (Molteno and Bennett 2000; Snyman 2006), and in the wild, the reproductive readiness of nonbreeders - measured by the downstream production of luteinising hormone following injection of pituitary gonadotrophin-releasing hormone - is elevated during periods of high rainfall when the likelihood of meeting dispersing males or of dispersing oneself is higher (Young and Bennett 2010). These physiological changes in

MORPHOLOGICAL DIVERGENCE nonbreeders in anticipation of reproductive opportunities share similarities with the onset of puberty in other mammals, where sex steroid secretion is a major driver of skeletal longitudinal and radial growth (Clarke and Khosla 2001), and as such, the removal from reproductive suppression and resultant hormonal changes may stimulate the morphological divergence of nonbreeding Damaraland mole-rats towards a more elongated phenotype. On the other hand, studies of classical rodent lab models highlight that many of the most pronounced changes in skeletal remodelling occur during pregnancy (Bowman and Miller

2001; Qing et al. 2012), and in naked mole-rats, housing females in isolation did not cause them to develop elongated vertebrae, whereas pairing them with a receptive male (with resultant pregnancy) did (O’Riain et al. 2000). Based on these studies, reproduction itself may provide the necessary cue for the skeletal remodelling of Damaraland mole-rat females.

In this chapter, I used information from X-rays to characterise the morphological divergence of breeders and nonbreeders in Damaraland mole-rats according to three principal aims: 1) to identify the skeletal changes that lead to increases in female size, 2) to identify the precise circumstances that stimulate growth adjustment in females, and 3) to investigate whether growth adjustments are associated with increases in fecundity. To identify the skeletal differences between breeders and nonbreeders, I first carried out cross-sectional comparisons of morphology in captive and wild Damaraland mole-rats. In addition, I experimentally manipulated the life history trajectories of female siblings in captivity to track the longitudinal development of elongation within age-matched individuals. By keeping some females as nonbreeders within their natal group, isolating others by placing them in their own tunnel system, and pairing others with an unrelated male to initiate reproduction,

I was able to determine whether removal from reproductive suppression is involved in female elongation, whilst controlling for the influence of age on development. If females placed in isolation display growth patterns analogous to newly reproductive females, this would provide strong evidence that the elongation of subordinate females in breeding groups is hindered by reproductive suppression. A similar argument extends to wild females that have dispersed and settled solitarily but have yet to reproduce, so I also compared the morphology of solitary females to breeders and in-group nonbreeders from a wild population of mole-rats.

Lastly, I investigated the fecundity implications of increased body size using correlative data from litters born in captivity.

MORPHOLOGICAL DIVERGENCE

METHODS

X-ray methodology

This chapter used information from X-rays taken on both captive and wild

Damaraland mole-rats. The general details of the captive study system and field work is detailed in Chapter 2, so the methods instead focus on how specific data sources were used to meet the main aims, and the methodology of X-raying individuals. Note that when taking X- rays on wild animals that had been transported temporarily back to the laboratory, individuals were also weighed and had their total body length measured manually; two people measured body length from the front of the snout to the tip of the tail to an accuracy of 1mm using a tape measure. Total body length was taken as the average of the two measures, and the human measurement of body length is referred to as ‘Total body length’ to distinguish it from the ‘Skeletal body length’ measured from X-rays (below). As mentioned in Chapter 2, after sampling, groups were housed temporarily in tunnel systems in the laboratory, and once a whole group was captured, as evidenced by an absence of activity for 24hrs, the animals were all returned to their natural burrow system. The data from captive animals were collected between November 2015 and July 2017, and from wild animals between February

2015 and June 2017.

X-rays were taken using the Gierth TR 90/20 battery-operated generator unit with portable Leonardo DR Mini plate (OR Technology, Rostock, Germany) under protocols approved by the University of Pretoria ethics committee. For each X-ray, mole-rats were immobilised under isoflurane anaesthesia and gently positioned in a dorsoventral position with straightened spine and splayed limbs. 9 skeletal traits were measured from each X-ray

(Figure 5.1) using ImageJ software. As in studies that have examined elongation in naked mole-rats, the length of a lumbar vertebra (L5 herein) served as an index of lumbar length (see

Henry et al. 2007; Dengler-Crish and Catania 2007, 2009). Similarly, the log ratio of the L5 vertebra length to zygomatic arch width served as a metric of elongation. Hereafter I refer to zygomatic arch width as skull width. All animals were weighed after anaesthesia. All statistical tests were performed in R version 3.2.3 (R Core Team, 2017), and standard model assumptions (normal errors and homogeneity of residual variance) were checked throughout.

MORPHOLOGICAL DIVERGENCE

Morphological divergence

To quantify the morphological divergence of breeders and nonbreeders I analysed a cross-sectional dataset of X-rays taken in captivity. The cross-sectional analysis from captive females was restricted to individuals larger than 100g, as the smallest reproductive female in the captive population was 100g. This resulted in data from 32 breeders and 79 in-group nonbreeders spanning 40 different groups, where group refers either to the group of original capture or group of birth. All 9 skeletal traits were included in a standard principal component analysis (PCA), and a multivariate analysis of variance (MANOVA) was performed on the resulting principal components to test for broad morphological differences between the two classes of female. Follow-up univariate ANOVAs on each of the first five principal components were used to determine on which of the principal components reproductive status was exerting its influence.

To examine whether the bivariate scaling relationships of morphological traits differed with breeding status, I fitted a series of linear models of the form loge(trait1) ~ loge(trait2)*Status, where trait2 represents either the skull width or skeletal body length, two metrics of size. In these models, a significant interaction term denotes significantly different slope between breeders and in-group nonbreeders. When the slopes did not differ, the difference in intercept was tested by removing the interaction term from each model: loge(trait1) ~ loge(trait2) + Status.

Skeletal changes

To further identify the skeletal changes that lead to increases in female size and to determine the circumstances leading to growth adjustment, I manipulated the life history trajectory of 30 natal sisters born in captivity by altering their social status, and tracked their development longitudinally (originally 32 females, but 2 died shortly after pairing and were excluded throughout). These females came from fourteen litters and were of known age, which also removed any possible confounding effect of age that could have been present in the cross-sectional analysis of captive mole-rats. Females were randomly allocated to one of three treatments: remaining in their natal group as a nonbreeder, placed in a new artificial tunnel system as a solitary nonbreeder, or paired with an unfamiliar male in a new tunnel

MORPHOLOGICAL DIVERGENCE system to become a breeder. X-rays were taken on these females at 2-month intervals from the point of treatment initiation for 12 months (except in a few cases where it was deemed that females were too heavily pregnant to be anaesthetised and X-rayed). Of the females that survived to eight months (n = 28), the mean age at manipulation was 526.8 ± 34.18, 536.45 ±

30.91 and 526.9 ± 36.78 days for in-group nonbreeders, solitary nonbreeders, and breeders respectively. Of the 11 females that were paired with a male, only two had failed to produce a litter by 6 months of age, and only one individual by 10 months. The median time to first parturition after pairing was 101.5 days. All females in the “breeder” treatment were included in all analyses.

General linear models were used to investigate skeletal growth trajectories across the three social treatments. The within-individual change in L5 lumbar vertebra length, skull width, and the elongation factor were each fit as a response, with treatment and the initial trait value specified as fixed covariates. Models were fitted for each response variable at every

2-month sampling interval to determine the point at which skeletal morphological divergence occurred. Pairwise comparisons of significant treatment effects were assessed with Tukey’s multiple comparisons (multcomp package, Hothorn et al. 2008), and similarities/differences between breeders and nonbreeders were used to assess whether growth adjustment is a result of breeding itself (in which case breeder morphology would differ from both classes of nonbreeding female) or rather due to the relaxation of reproductive suppression (in which case breeder morphology would only differ from females remaining in their natal group).

To control for a possible artefact of captivity on morphological patterns, X-rays from wild mole-rats were also examined. X-rays from wild animals only included females heavier than 101g, the weight of the smallest breeder that was captured for X-ray sampling. This produced a dataset of 56 females captured between February 2016 and June 2017, including

21 breeders, 12 in-group nonbreeders, and 23 solitary nonbreeders. The presence of an unperforated vagina confirmed that these solitary females had not previously engaged in sexual activities, and previous trapping records indicated that many of these solitary nonbreeding females had been solitary for at least two years (and in a few cases four years).

The relative elongation (logeL5 vertebra / logeSkull width) of the three classes of female was fitted in a linear mixed effects model with normal errors, with group identity set as a random

MORPHOLOGICAL DIVERGENCE effect to control for possible morphological similarities within groups. Pairwise comparisons were treated as above.

Body length and fitness

To investigate whether increases in growth are associated with increases in fecundity I used an extensive dataset of birth events in captivity. The total body length of female breeders is measured during routine sampling in the lab, which made it possible to investigate whether body length was related to three measures of fecundity among a large cohort of breeders: litter size at birth, the total mass of pups at birth (total neonate mass), and individual pup mass at birth. Information was only included from females whose total body length was measured within 90 days of the birth of a litter to ensure that it reflects size around parturition. Further, to remove any uncertainty around birth date, litters were only used if checks of the nest box indicated that the litter must have been born on the day of the check or the day preceding it.

The total dataset included 186 litters born to 58 mothers that produced 587 pups. Pups were measured to the nearest gram on an electronic balance. The three traits were fitted to linear mixed effects models, where total neonate mass and individual pup mass were fitted to a normal error structure, and litter size was fitted to a Poisson error structure. For total neonate mass and litter size, a single model was fitted in each case that included breeder body length and whether it was the females first reproductive episode (primiparity) as fixed terms, and maternal identity as a random term. For individual pup mass, the model contained breeder body length, primiparity and litter size as fixed terms, and maternal identity and litter identity as random terms; by including litter size I could test whether longer mothers produced relatively larger pups; partial residuals of body length were extracted using the remef package

(Hohenstein and Kliegl 2013) before being plotted. The significance of fixed terms was assessed by likelihood ratio tests.

To check that any morphological pattern in the results could not be driven by a bias in the

X-ray annotation by the lead author, a random subset of 150 X-ray images that formed part of the study were annotated blind by a second person unrelated to the study. The correlation between the measurements across skeletal traits was consistently high (r > 0.93, except for pelvis length, where r = 0.497), and so it was highly improbable that measurer bias affected the results.

MORPHOLOGICAL DIVERGENCE

RESULTS

Morphological divergence

In captive Damaraland mole-rats, breeding females are both larger and relatively longer than subordinate nonbreeders residing in groups, after controlling for the weight of individuals (Figure 5.1b; breeder mean = 143.47g ± 3.56 vs in-group nonbreeder mean 137.91g

± 2.78; Welch’s t-test, t70.1 = -1.23, p = 0.22): principal component analysis revealed a morphometric separation of breeders and in-group nonbreeders according to both size and shape (Figure 5.1c; trait loadings for the first 5 PCs, which together explain 91.8% of the variance, are in Appendix C Table 1). There was considerable overlap in the morphological space occupied by either class. PC1 revealed positive loading for all traits and is indicative of general size, breeders being generally larger than in-group females. PC2 separated breeders and in-group nonbreeders by shape, and reflects the relatively longer L5 lumbar vertebra, longer skeletal body length, wider pelvis, and shorter femurs of breeders. The MANOVA performed on PCs 1 – 5 revealed a significant effect of social status (F5,104 = 26.48, p < 0.001), with follow-up univariate ANOVAs on each PC 1 – 5 suggesting the status effect is most prominent in the first three components (PC1: F1 = 11.04, p < 0.001; PC2: F1 = 57.1, p < 0.001;

PC3: F1 = 7.43, p = 0.007; PC4: F1 = 2.80, p = 0.097; PC5: F1 = 2.87, p = 0.093).

Breeders in captivity were more elongated than subordinate nonbreeders, as shown by their relatively longer lumbar vertebrae for a given size (skull width, Figure 5.2a, or skeletal body length, Appendix C Table 2), as well as being longer overall (Total body length: breeder mean = 18.51 ± 0.14cm, in-group nonbreeder mean 17.72 ± 0.11cm, Welch’s t-test, t64.3 = -4.36, p

< 0.001). Breeders and in-group nonbreeders also showed different bivariate scaling relationships in several other skeletal traits (all bivariate relationships in Appendix C Table

2), most notable being the wider pelvic girdle of breeders.

MORPHOLOGICAL DIVERGENCE

b c

Figure 5.1. Morphological divergence of breeders and nonbreeders a) Dorsoventral X- ray annotated with the 9 morphological traits used in multivariate analysis. A: rostrum length, B: skull width, C: ulna, D: L5 vertebra length, E: pelvic girdle width, F: pelvis length, G: femur length, H: tibia length, I: skeletal body length. b) Weight distribution of captive breeding and nonbreeding in-group females in multivariate analysis. c) PC1 and PC2 separates captive females according to their social status.

MORPHOLOGICAL DIVERGENCE

Figure 5.2. a) The bivariate relationship between the length of the lumbar vertebra and the width of the skull in captive female mole-rats indicates that breeders (triangles) are relatively more elongated than in- group nonbreeders (circles). b) Breeding females are also more elongated than both in-group females and solitary females in the wild, the latter having dispersed and settled in isolation. Numbers within the plot refer to sample size.

MORPHOLOGICAL DIVERGENCE

Skeletal changes in breeders

Experimental manipulation of the social status of natal sisters of known age in captive animals showed that the elongation of breeding females was caused by an increased lengthening of the skeletal vertebrae (Figure 5.3). These increases in growth are a consequence of breeding rather than of the relaxation of reproductive suppression, since females removed from a group setting and placed solitarily displayed a similar growth pattern to nonbreeding females residing in groups. Females paired with unfamiliar males to become breeders were already more elongated than the two classes of nonbreeding female four months into the experiment (Figure 5.3a; F2 = 11.31, both pairwise contrasts p < 0.001). This coincided with the point at which the L5 vertebra of breeders were significantly longer (Figure 5.3b displays the contrasts in L5 vertebra length at 12 months (LMM contrast = 0.251 mm ± 0.05 relative to in- group breeders, p < 0.001, contrast = 0.014mm ± 0.06, p = 0.025), an effect that persisted thereafter (Appendix C Figure 1a displays contrasts at each sampling interval). The lengthening effect in breeders was also shown by other lumbar vertebrae (see Appendix C

Figure 2 for growth in L4 and L6 vertebrae). In contrast, there was no evidence that changes in skull width contributed to the relative elongation of breeders (Figure 5.3c and Figure

Appendix C Figure 1b; p > 0.10 for all pairwise contrasts), although there was a trend in the direction of reduced growth in the skull of breeders.

Evidence from wild mole-rats agreed with the role of reproduction in breeder elongation. Here too, breeders displayed a more elongated phenotype than female nonbreeders residing in groups (Figure 2b; elongation factor contrast = 0.013 ± 0.005, p = 0.027) and solitary nonbreeders (contrast = 0.014 ± 0.006, p = 0.036), whereas the two classes of nonbreeder did not differ from one another (contrast = -0.0001 ± 0.006, p = 0.89). Breeders

(mean = 18.81 ± 0.17cm) also exhibited a longer total body length than both solitary nonbreeders (mean 17.86cm ± 0.36cm; contrast p value < 0.001) and in-group nonbreeders

(mean 17.81cm ± 0.27cm; contrast p value = 0.052).

MORPHOLOGICAL DIVERGENCE

b c

Figure 5.3. a) Change in the relative elongation b) L5 lumbar vertebra length and c) skull width of captive females experimentally manipulated to follow different social trajectories. Numbers within the plot refer to sample sizes at each timepoint of the experiment.

MORPHOLOGICAL DIVERGENCE

Body Length and fitness

Across breeding females increased body length is associated with higher fecundity

(Appendix C Table 3) as longer females have larger litters (Figure 5.4a; GLMM: β = 0.12 ± 0.05,

χ21 = 5.81, p = 0.016). Since litter size is highly correlated with total neonate mass (r = 0.92), longer females also produced litters with larger total mass (LMM: β = 3.83 ± 1.23, χ21 = 8.89, p

= 0.003). However, longer females do not trade off increasing offspring quantity for reduced quality offspring. On the contrary, controlling for litter size, longer females invest proportionally more per pup (Figure 5.4b, LMM: β = 0.36 ± 0.17, χ21 = 4.12, p = 0.040), which suggests that increases in body length may facilitate increases in both offspring quality and quantity.

MORPHOLOGICAL DIVERGENCE a

Figure 5.4. Fitness consequences of body length. Increases in body length are associated with increases in a) litter size and b) individual pup mass across 186 litters born in captivity. a) Points represent raw data scaled to according to sample size, and lines show predicted litter size ± 95 confidence intervals from linear mixed effects models. b) Partial residuals of individual pup mass at birth, corrected for litter size and primiparity. Points represent partial residuals estimated from linear mixed effects model. Line displays the partial effect of body length on individual pup mass, but confidence intervals are not provided as the generation of a partial effect removes variation from the fitted model. The partial residuals were standardised as they do not provide meaningful values.

MORPHOLOGICAL DIVERGENCE

DISCUSSION

The morphological divergence of reproductive and nonreproductive individuals has only been documented in two vertebrate societies, those of the naked mole-rat and the

Damaraland mole-rat (O’Riain et al. 2000; Young and Bennett 2010). In this chapter, by altering the life history trajectories of age-matched females in captivity, I show that the lengthening of breeding females in Damaraland mole-rats is caused by the upregulated growth of the skeletal vertebrae in breeders, causing breeders to be more elongated than nonbreeders. I also show that it is the onset of breeding, rather than the removal from reproductive suppression, that acts as the driver of this skeletal remodelling. Two lines of evidence support this view. First, I found that reproductively naive females in the wild that dispersed from their natal group and settled solitarily were morphologically equivalent to female nonbreeders still resident in their natal group, despite many of the former being isolated for multiple years. Secondly, I found that individuals in captivity only changed shape when they were paired with an unrelated male, whereas no such change took place in females that were housed solitarily. The increase in elongation demonstrated by breeding females presumably serves to enhance fecundity by allowing females to be larger without gaining extra girth. It is probably also highly advantageous for a species that occupies a system of narrow subterranean tunnels where increases in girth must impose strong constraints on mobility, as unlike in eusocial insects, reproductive female mole-rats are not bound to the nest (e.g. Ansell’s mole-rat Fukomys anselli:

Šklíba et al. 2016). The fecundity benefits of lengthening were confirmed by analyses of birth events in captivity which showed that longer breeders produced larger litters and invested more prenatally in each pup after the effect of litter size was statistically controlled for.

In finding that upregulated vertebral growth is central to the elongation of breeding females, the results of this chapter deviate from a previous result in wild Damaraland mole- rats which suggested that breeder elongation is achieved through a relative reduction in growth of the skull compared to growth towards total body length (Young and Bennett 2010).

The ad libitum feeding of individuals in captivity might also have generated different patterns of growth in captivity compared to the wild. This possibility can only be addressed with longitudinal X-rays of wild animals. Without this information, this study implicates vertebral growth as the key process involved in elongation. As such, similar developmental routes

MORPHOLOGICAL DIVERGENCE appear to underpin the morphological divergence of naked and Damaraland mole-rats, albeit the extent of divergence is greater in naked mole-rats (O’Riain et al. 2000). This may be an indirect result of the much larger group sizes of naked mole-rats (up to 295 individuals, Brett

1991); with increasing group size, breeder fecundity will be stronger (larger workforces can rear more offspring; mean and maximum litter sizes in the field = 11.3 ± 6.2 and 28 individuals respectively, Sherman et al. 1999) and the reproductive potential of helpers declines (because the likelihood of inheriting the dominance position is reduced), which together would favour increased elongation in breeders and developmental arrest in nonbreeders. However, although naked and Damaraland mole-rats display the largest group sizes and most extreme forms of sociality in the mole-rats, other members of the Bathyergidae family exhibit comparable social features (group-living, a reproductive division of labour and delayed dispersal: Bennett and Faulkes 2000), and one could speculate that if the social environment is important in morphological divergence then morphological divergence may also be present in other mole-rats. Alternatively, if skeletal elongation were principally a consequence of subterranean living and associated constraints on abdominal width, then even the solitary species of mole-rat may undergo vertebral lengthening across reproductive episodes (e.g. the

Cape dune mole-rat Bathyergus suillus, the Cape mole-rat Georychus capensis, and the Silvery mole-rat Heliophobius argenteocinereus, amongst others); though group-living and sociality could still increase the magnitude of this effect trait. A broader examination of rank-related growth would reveal whether skeletal elongation is a unique adaptation associated with the highly social mole-rats or a more general feature of the Bathyergidae family and their subterranean habits.

This chapter indicates that the vertebral elongation of formerly subordinate helpers is stimulated by the onset of reproductive activities. The alternative possibility that relaxation of reproductive suppression is sufficient to induce divergence towards an elongated phenotype can be ruled out, since nonbreeding females removed from their natal group and housed solitarily for a year exhibited a growth trajectory equivalent to nonbreeders retained in complete groups (see O’Riain et al. 2000 for an analogous result in naked mole-rats).

Likewise, solitary females in the wild were morphologically indistinguishable from in-group nonbreeders. Various endocrinological changes that take place throughout pregnancy could be implicated in this parity-driven bone growth. Sex steroids such as estradiol and

MORPHOLOGICAL DIVERGENCE progesterone increase during pregnancy and are heavily involved in bone formation (Syed and Khosla 2005; see Clarke et al. 2001 for progesterone in DMRs), often by mediating levels of Growth hormone and insulin-like growth factor 1 (Locatelli and Bianchi 2014). Testosterone also promotes bone growth (Clarke and Khosla 2009), and although commonly thought of as a male hormone, the heightened levels of testosterone measured in breeding female

Damaraland mole-rats (Lutermann et al. 2013) – a pattern also seen in other cooperative breeders (Davies et al. 2016) – raises the possibility that testosterone is also involved in vertebral lengthening. The action of such hormones sets up a skeletal mineral reserve that is subsequently resorbed and used in offspring development and milk production (Kovacs and

Kronenburg 1997; Bowman and Miller 2001; Kovacs 2016). Yet, despite skeletal demineralisation during lactation, reproduction has been shown to lead to a net increase in total skeletal size. For example, in mice it has been shown that reproduction generates permanent increases in total body length (Schutz et al. 2009), a result bearing obvious relevance to mole-rats, whilst in humans, high parity has been associated with increases in bone size (Specker and Binkley 2012; Wiklund et al. 2012). Similar physiological mechanisms might therefore operate to drive the elongation and increased size of female mole-rats, whose unusual physiology could yet offer important biomedical insights concerning skeletal development (Pinto et al. 2010).

The metabolic challenges of reproduction (Clutton-Brock et al. 1989; Speakman 2008) can be expected to be particularly large in cooperative breeding mammals like Damaraland mole-rats because dominant females within cooperative societies often breed multiple times per year, and it is not uncommon for a female to conceive during the period that she is lactating for her current litter (meerkats: Barrette et al. 2012; Russell et al 2003; naked mole- rats: Jarvis 1991). In this context, continued skeletal growth in mole-rats is particularly remarkable because short inter-birth intervals must necessarily reduce the opportunity to recover calcium lost during lactation. In some cooperative breeders such as the mongooses and the canids, some of the maternal burden of lactation is offset by helpers lactating for non- descendent pups (e.g. African wild dog Lycaon pictus: Creel and Creel 2002; dwarf mongoose

Helogale parvula: Creel et al. 1991; meerkat: Macleod et al. 2013), but allo-lactation is absent in mole-rats. In the absence of allo-lactation, the atypically long gestation period of Damaraland mole-rats could be important by allowing more time for pregnant females to accrue mineral

MORPHOLOGICAL DIVERGENCE reserves prior to lactation (78-92 days, Bennett and Faulkes 2000). In meeting their calcium requirement, breeding females may also rely on the exceptionally high calcium content of their primary food source - the subterranean tubers of the gemsbok cucumber Acanthosicyos naudininianus - and a highly efficient mode of calcium uptake to drive their skeletal growth

(Pitcher et al. 1992; see also Buffenstein 2000). In fact, the calcium content of gemsbok cucumber is four to five times higher per unit mass than the sweet potato and cucumber diet given to the animals in our captive population, and this lowered calcium diet may even have led to an underestimation of the degree of vertebral lengthening documented in this study.

Overall, this chapter confirms the morphological divergence of breeding female

Damaraland mole-rats at the skeletal level. By examining the growth trajectories of age- matched individuals, it provides the first definitive evidence that the growth patterns underlying skeletal dimorphism cannot be explained by early-life developmental divergence or by status-specific age differences like those that underpin phenotypic development in some eusocial ants and honeybees (Volny et al. 2002; Kucharski et al. 2008; Schwander et al. 2016).

Instead, growth patterns resemble other cooperatively breeding vertebrates (meerkats,

Russell et al. 2004; naked mole-rats, Dengler-Crish and Catania 2007; Haplochromis burtoni cichlid, Hofmann et al. 1999) and other social insects (e.g Ponerine ants, Peeters and Ito 2001; termites, Noirot 1990) where contrasts in size and shape between breeders and nonbreeders are the result of changes that occur on or around the acquisition of a dominant breeding position. Viewed more broadly, the secondary growth of mole-rat breeders provides a clear example of socially responsive growth adjustment, or what might be termed ‘strategic growth’

(Heg et al. 2004; Huchard et al. 2016). Similar forms of adaptive, socially responsive growth might be more prevalent in mammals than is currently recognised, but the extent to which this is the case, and the implications for the structuring of mammalian dominance hierarchies is as yet poorly understood. Of the few cases documented in mammals, emphasis has been placed on the status-related upregulation of growth (Russell et al. 2004; Dengler-Crish and

Catania 2009; this study), but an equally interesting perspective will be to understand the circumstances and mechanisms that cause development to be delayed or arrested in subordinate individuals in the first place. For example, in adult male orang-utans Pongo abelii, some individuals develop conspicuous, sexually-selected cheek flanges soon after reaching sexual maturity, whereas others may reach sexual maturity and remain unflanged for 20 years

MORPHOLOGICAL DIVERGENCE before developing this conspicuous secondary trait (Utami et al. 2002), though the causes of this individual variation in development are largely unknown. Societies with marked reproductive skew and suppression of subordinate reproduction provide an obvious place to investigate socially-responsive growth further, but as subordinate group members often enjoy a small share of reproduction in even heavily-skewed societies (Russell 2004; Clutton-Brock et al. 2016), it seems that if rank-related divergence in size and shape is identified in other mammals, it is unlikely to be of a magnitude comparable to the social mole-rats where subordinates never reproduce.

100

The long-lived queen: telomere dynamics

6 in the wild and the case for lifespan CHAPTER extension in breeders

______

101

TELOMERE DYNAMICS

ABSTRACT

In the Fukomys mole-rats, lifespan is heavily dependent on social status, with breeders displaying a twofold greater lifespan than nonbreeders. It has been argued that this bimodal ageing pattern cannot be explained by contrasts in social rank, intrinsic quality or time budgeting. Instead, it has been proposed that the transition to a dominant breeding role sets individuals onto a slower ageing trajectory, but so far, there is little evidence to support a physiological basis to lifespan extension in breeders. Telomere length and telomere shortening serve as informative biomarkers of lifespan and ageing rates, but changes in telomere dynamics and their relationship to ageing profiles has not yet been explored in mole-rats. In this chapter, I examined status-related telomere dynamics of female Damaraland mole-rats

Fukomys damarensis from the wild. There was no evidence for longer telomeres or slower rates of telomere attrition in breeding females compared to non-breeding females in groups, each of which might be expected if breeders are ageing more slowly. However, there was a tendency for higher rates of telomere attrition in solitary non-breeding females (who have dispersed), which may reflect the higher energetic costs of foraging alone. In some mammals with similar breeding systems, differences in longevity between dominant breeders and subordinate helpers occur because dominants kill or evict older subordinates. It is not currently clear if this is also the case in Damaraland mole-rats, but I suggest that previous studies exploring status-related ageing in captive Fukomys mole-rats have overlooked the importance of demographic processes (and associated behavioural influences) on mortality schedules. Future studies should synthesize behavioural and endocrinological information to explore the possibility that the shortened lifespan of Fukomys nonbreeders is driven by agonistic interactions and social stress shortly before death.

102

TELOMERE DYNAMICS

INTRODUCTION

The African mole-rats (family Bathyergidae) are amongst the longest-lived species for their size of all terrestrial mammals (Dammann and Burda 2006; Healy 2015; Sherman and

Jarvis 2002). The naked mole-rat Heterocephalus glaber (~40g) provides the most extreme example within the family, with individuals living for three decades in captivity and showing no clear age-related changes in physiology or mortality rate across their long life (Buffenstein

2008, Ruby et al. 2018). Extended longevity is also a feature of the Fukomys clade, as captive

Ansell’s mole-rats Fukomys anselli (~90g) and giant mole-rats Fukomys mechowii (250-400g) have both been known to live beyond twenty years (Dammann and Burda 2006; Dammann et al. 2011).

Fukomys mole-rats also show marked intraspecific variation in lifespan that is directly related to breeding status. Unlike in naked mole-rats where variation in lifespan is apparently independent of breeding status (Sherman and Jarvis 2002; Buffenstein 2005), in Fukomys species, the longest-lived individuals are exclusively the single breeder of each sex; the lifespan of breeders being twofold greater than nonbreeders (Dammann and Burda 2006;

Dammann et al. 2011; Schmidt et al. 2013).

In other vertebrate societies where reproductive individuals outlive non-reproductive individuals, social or demographic processes can account for the positive association between reproduction and lifespan (Novikov et al. 2015; Cram et al. 2018). Nonbreeding individuals in such societies are invariably subordinate and must disperse away from their natal group to acquire reproductive opportunities. Dispersal brings with it a heightened mortality risk that is not faced by group-bound breeders, generating divergent aging profiles between the two social classes. Even so, the prolonged lifespan of breeding Fukomys mole-rats is maintained in captivity where dispersal is prevented entirely. In this setting other factors may contribute the observed longevity differences: the shortened lifespan of nonbreeders may arise from social rank effects related to subordinacy (Sapolsky et al. 1997; Beehner and Lu 2013; Novikov et al.

2015), from variation in the intrinsic quality of breeders versus non-breeders, or from differences in activity patterns such that nonbreeders are more active than breeders. Each of these possibilities have previously been dismissed with support from behavioural data, and

103

TELOMERE DYNAMICS

it has instead been argued that the extended longevity of breeders is due to the act of reproduction itself (Dammann and Burda 2006; Dammann et al 2011).

The suggestion that reproductive activity can delay ageing in mole-rats reverses the classical prediction from life history theory, backed by empirical evidence, of a trade-off between reproduction and survival (Williams 1957; Kirkwood and Rose 1991; Nussey et al.

2006; Hayward et al. 2014). Outside of vertebrates, there are nonetheless many examples in eusocial insects of slowed ageing profiles in reproductive ‘queens’ compared to workers

(Schrempf et al. 2005; Keller and Jemielty 2006; Heinze and Schrempf 2008; Kramer and

Schaible 2013), and various studies have identified molecular pathways involved in queen lifespan extension, such as via vitellogenin signalling (Corona et al. 2007) or the insulin/insulin-like growth factor signalling pathway (Münch et al. 2008). In many cases, the physiological changes underpinning divergent ageing trajectories in social insects are epigenetically programmed and phenotypically plastic, but whether analogous physiological changes drive the extended longevity of mole-rat breeders is unclear. Since the acquisition of a breeding role is met with a suite of physiological (Bennett et al. 1996; Dengler-Crish and

Catania 2007; Schmidt et al. 2014; Schielke et al. 2017) and neuroanatomical changes (Anyan et al. 2011; Holmes et al. 2009) in Fukomys mole-rats, it remains possible that phenotypic changes around breeding also promotes the adoption of a slower ageing profile in mole-rats, despite the expected costs of reproduction.

To date, only a handful of studies have examined status-related ageing in Fukomys mole-rats at a proximate level, and those that have done so provide only indirect or no support for a physiological basis to lifespan extension in breeders. For example, in Ansell’s mole-rats, breeders show higher resting metabolic rates than nonbreeding helpers (Schielke et al. 2017).

This pattern would typically be assumed to reflect the increased energetic costs of reproduction (Thompson 1992), and if anything, be associated with reductions in lifespan, not extensions (Speakman 2005). Perhaps the strongest case for a physiological link between breeding and lifespan extension is a study by Schmidt et al. (2014)showing that breeding female Damaraland mole-rats Fukomys damarensis have lower levels of oxidative damage markers compared to nonbreeders in some tissues (see Blount et al. 2016 for a meta-analysis investigating the link between reproduction and oxidative stress). As oxidative damage has often been thought to reflect the deterioration of bodily function commensurate with ageing

104

TELOMERE DYNAMICS

(Finkel and Holbrook 2000; Kregel and Zhang 2007), the seemingly reduced susceptibility to oxidative damage of breeders was highlighted as a possible mechanism for prolonged lifespan. In contrast, a study of Ansell’s mole-rats found heightened levels of advanced glycoxidation end products, a source of reactive oxygen species and thus a promoter of oxidative damage, in the skin of breeders compared to nonbreeders (Dammann et al. 2012).

Paradoxically, this result was still used as support for physiological mediation of breeder longevity; the argument being made that breeders must have evolved powerful antioxidant defences to combat the oxidative stress that would otherwise have been induced by the reactive oxygen species. This assertion is difficult to reconcile without any information on oxidative damage (Selman et al. 2012). In addition, there is currently a lack of consensus on whether oxidative damage acts as a consistent mediator of life-history trade-offs (Selman et al. 2012; Speakman and Garratt 2014; Blount et al. 2016), raising questions about its utility of as a metric of cellular ageing, or what might more accurately be termed ‘somatic redundancy’

(Boonekamp et al. 2013). An investigation of better-established hallmarks of ageing (López-

Otín et al. 2013) can therefore clarify the extent to which the extended lifespan of breeding

Fukomys mole-rats is driven by physiological processes.

In this chapter I examine status-related changes in telomere dynamics in female

Damaraland mole-rats in the wild. Telomeres are widely regarded as a powerful and informative indicator of the cellular deterioration that routinely accompanies organismal ageing (López-Otín et al. 2013). Telomeres are highly conserved DNA-protein complexes located at the ends of chromosomes which serve to protect genomic integrity (O’Sullivan and

Karlsreder 2010). In most cell types, telomeric DNA shortens with each round of cellular replication (Olovnikov 1996), eventually reaching a threshold length that causes cells to undergo replicative senescence (Campisi 2003). The build-up of such cells compromises tissue performance and thereby drives physiological declines. As such, although a direct causal role of telomeres on ageing is debated (Simons 2015; Young 2018), there is a large body evidence highlighting that both absolute telomere length and the rate of telomere shortening predict mortality and lifespan inter- and intra-specifically (Gomes et al. 2011; Olsson et al. 2017;

Tricola 2018; Young 2018). Telomeres may therefore be implicated in the lifespan extension of breeders. Specifically, if differences in lifespan between breeders and non-breeders arise from divergent ageing profiles, one would predict that telomere shortening is reduced or absent in

105

TELOMERE DYNAMICS

breeders compared to nonbreeders. I test this prediction using repeat blood samples taken from wild Damaraland mole-rats (the mean sampling interval was approximately 9 months).

By conducting sampling in the wild, the study overcomes possible biases introduced by captivity. For example, the provisioning of food ad libitum in captivity might limit the costs of work or reproduction, and as captive animal cannot disperse, the age-structure of captive groups might differ from wild groups unless carefully managed. In addition to exploring status-related telomere dynamics, I critically assess the broader evidence base surrounding intraspecific variation in aging rates in Fukomys mole-rats (Table 1). This chapter focuses on females as the lack of paternity information meant that the identity of breeding males was unknown.

106

TELOMERE DYNAMICS

Table 6.1. A summary of the putative factors explaining status-related differences in ageing rate in Fukomys mole-rats.

Putative Factor Prediction Evidence For Evidence Against Comments

Oosthuizen and Bennett (2015) used wild caught animals so age might have confounded trends (breeders No difference between social would have been older). Activity classes in time budgeting or was also measured in an unnatural Breeders show marked total activity in captive F. anselli setting, as individuals were housed reductions in total activity, Breeding female F. damarensis were less (Dammann and Burda 2006; individually (measured via infrared Activity or dedicate less time to active than female non-breeders in captivity Schielke et al. 2012) motion detectors). Patterns energetically expensive (Oosthuizen and Bennett 2015) behaviours, compared to Schielke et al. (2012) measured total nonbreeders. locomotor activity using radio No consistent differences frequency identification from 6 between social classes in total groups, over a period of one week. locomotor activity in captive F. mechowii (Dammann et al. 2011). Dammann et al. (2011) and Dammann and Burda (2006) observed mole-rat groups for 72 hours to derive time budgets.

Not clear how activity patterns in captivity relate to activity patterns in the wild. Dispersal of nonbreeders Our capture-mark-recapture of F. damarensis from natal groups explains indicates dispersal of nonbreeders between 1 Individuals that are first captured at divergent ageing rates. and 4 years of age (Thorley pers obs; Hazell low weight are known to be et al. 2000) juveniles and their age can therefore None be approximated. The dispersal age Demography In captivity, the prevention In captivity, older F. damarensis female of nonbreeders can therefore be of dispersal upregulates the nonbreeders are evicted (Thorley pers obs). inferred from the rate such ‘known- stress axis in nonbreeders, They also show increasing cortisol levels age’ individuals are lost from raising their mortality rate. with age (Novikov et al. 2015); consistent existing groups (extrinsic mortality is thought to be low); many of these

107

TELOMERE DYNAMICS

with a role of stress on shortened lifespan of dispersers have subsequently been nonbreeders. recaptured in other groups (Thorley pers obs). New groups in captivity are Intrinsic created through the pairing of Breeders are of higher None unrelated males and females. Quality intrinsic quality. The decision to pair an animal Differences or not is random, removing any scope for quality effects. No difference in rate of telomere shortening between breeders and non-breeders in wild F. damarensis (this study). Physiologically- Physiological traits Female F. damarensis breeders show reduced mediated associated with energetic oxidative stress levels in some somatic F. anselli breeders display higher divergence in costs/increased somatic tissues (Schmidt et al. 2014). resting metabolic rate than investment should be higher nonbreeders (Schielke et al. life history in nonbreeders than 2017), whilst in wild F. trajectories breeders. damarensis there is no significant difference between breeders and nonbreeders in resting metabolic rates or daily energy expenditure in wet season, the period of peak digging (Scantlebury et al. 2006).

No status-related differences in thyroid hormone metabolism in F. anselli (Henning et al. 2014).

108

TELOMERE DYNAMICS

The assessment of social rank effects on survival has been based purely Dammann and Burda (2006) and on comparisons of nonbreeders. All Dammann et al. (2011) else being equal, it would be As with the demography information, In compared the survivorship of preferable to compare breeders and Social Rank Individuals of higher captivity, older F. damarensis female ‘dominant’ versus ‘subordinate’ nonbreeders, which requires dominance rank should nonbreeders are often evicted (Thorley pers nonbreeders, finding no maintaining social rank differences Effects outlive more subordinate obs). They also show increasing cortisol difference in rates of mortality. between breeders and nonbreeders individuals. levels with age (Novikov et al. 2015); The assignment of dominance but preventing reproduction in the consistent with a role of stress on shortened was based on litter order breeding female. One means of lifespan of nonbreeders. (individuals from earlier litters doing so would be to castrate the being identified as dominant breeding male in the group. In over later-born litters). doing so, the reproductive axis of the breeding females and associated behaviour should be unaffected, but reproduction itself is prevented. Note that blocking reproduction in the females using contraceptives would be unsatisfactory, as changes to reproductive physiology could influence the social interactions between breeders and nonbreeders.

109

TELOMERE DYNAMICS

METHODS

Blood sample collection

Damaraland mole-rats are subterranean, social rodents distributed throughout the semi-arid zone of Southern Africa. For this study, mole-rats were captured between February

2015 and June 2016 at the Kuruman River Reserve (KRR, 26°58’S 21°49’E) and the Tswalu

Reserve (27°26’S 22°16’E) in South Africa. Groups were trapped episodically (range 6 – 13 months) using modified Hickman live traps baited with sweet potato, which were placed into tunnel systems located via mole hills visible at the surface. Traps were checked every 2-3 hours throughout the day and night, and captured animals were placed in a closed, sand- filled box alongside other group members, and provided food and shelter. Intermittently, individuals were transported back to the laboratory where they were weighed, measured and blood sampled. When transporting animals from the field to the lab, traps were temporarily disabled to prevent individuals being kept in the traps for long periods. All individuals captured at the study sites are marked with PIT tags (Trovan®) at first capture to allow individual recognition. After sampling, groups were housed in semi-natural tunnel systems in the laboratory, and once a whole group had been captured, indicated by an absence of trap activity for 24hrs, the animals are returned to their tunnel system in the wild.

For blood sampling, individuals were anaesthetised via isoflurane inhalation. Whole blood was collected from a vein in the foot using an EDTA-lined capillary tube and stored in an EDTA-lined Eppendorf tube. At the Kuruman River Reserve, whole blood was placed straight into a -80°C freezer, whilst at Tswalu they were placed into a -20°C freezer before being transferred to the Kuruman River Reserve periodically. All sampling was conducted under ethical permits from the University of Pretoria.

110

TELOMERE DYNAMICS

Figure 6.1. A wild Damaraland mole-rat (Fukomys damarensis) shortly after capture

111

TELOMERE DYNAMICS

Figure 6.2. Blood sample collection from an anaesthetized Damaraland mole-rat.

112

TELOMERE DYNAMICS

DNA extraction and qPCR

DNA was extracted using Gentra Puregene Blood Kits (QIAGEN Ltd, Manchester,

United Kingdom), broadly following standard protocols for extraction of genomic DNA from

300µl whole blood but with the following volume modifications: 600µl of Cell Lysis buffer and 5µl Proteinase K (20mg/ml) were used for lysis, 230µl of Protein Precipitation Solution was added to lysate, 1.5µl glycogen solution (QIAGEN Ltd, Manchester, United Kingdom) was added to supernatant before DNA precipitation step, 700µl isopropanol was used for

DNA precipitation, 700µl of 70% ethanol was used for pellet washing, and DNA was rehydrated with 25µl DNA Hydration Solution. Samples were incubated at 56°C for 1-3 hours until completely lysed, and DNA was left to rehydrate at 4°C overnight before being mixed with 175µl PBS to be purified using MACHEREY-NAGEL NucleoMag® Blood 200µL kits

(MACHEREY-NAGEL GmbH and Co. KG, Düren, Germany) alongside the KingFisher™ Flex

Purification System (Thermo Scientific, Wilmington DE, USA), following kit protocols. DNA was eluted in kit elution buffer MBL5 (5 mM Tris, pH 8.5) and stored at -20°C until further use. The concentration and purity of DNA was assessed using a Nanodrop-8000

Spectrophotometer (Thermo Scientific, Wilmington DE, USA). Average DNA concentration was 64.5 ± 24.2 ng/ul (mean ± SD) and average 260/280 and 260/230 ratios were 1.94 ± 0.08

(mean ± SD) and 2.24 ± 0.22 (mean ± SD). DNA integrity was assessed by running 25ng of

DNA in a 0.7% agarose gel at 110V for 25 minutes and was deemed to be acceptable for telomere measurement.

Telomere length was measured via quantitative PCR (Cawthorn 2002). This measure represents the average telomere length across cells in a sample and is reported as the abundance of telomeric sequence relative to a non-variable copy number gene. Here, I used

RAG1 as a reference gene, by virtue of its known status as a single copy gene in vertebrates

(Crottino et al. 2012; Groth and Barrowclough 1999) and used a primer pair designed from

Accession XM_010612551.1 selected for their good performance, and their lack of non-specific binding or primer-dimer formation (confirmed by melt curve analysis and gel electrophoresis during optimisation). DNA samples (2.5ng) were assayed in triplicate and on separate plates for telomere and single-copy targets. Reactions were conducted using 1X Absolute blue qPCR

SYBR green Low Rox master mix (Thermo Scientific, Wilmington DE, USA) with RAG1

113

TELOMERE DYNAMICS

forward (5’- GCAAGCCATTGTGCCTTATG -3’) and reverse (5’-

CTCCCATCTCCAGCAGTAATTC -3’) primers at 150nM and telomere primers Tel1b (5’-

CGGTTTGTT TGGGTTTGGGTTTGGGTTTGGGTTTGGGTT-3’) and Tel2b (5’-

GGCTTGCCTTACCCTTACCCTTACCCTTAC CCTTACCCT-3’) (Epel et al. 2004) at 500nM, bringing reaction volumes up to 25µl with water. Mx3000P 96-well skirted plates (Agilent,

Santa Clara, United States) were manually loaded, sealed with 8x strip optical caps (Agilent,

Santa Clara, United States) and run in an Agilent Technologies Stratagene Mx3005P real-time

PCR machine. RAG1 thermal profile was 15 min at 95°C, followed by 40 cycles of 15 s at 95°C,

30 s at 60°C, 30 s at 72°C. Telomere thermal profile was 15 min at 95°C, followed by 26 cycles of 15 s at 95°C and 60s at 58°C. Both assays were followed by melt curve analysis of (58–95°C

1°C/5 s ramp). Dissociation curves showed a single peak for both assays in all reactions. A pooled aliquot of DNA samples was serially diluted (10ng to 0.6ng) to generate a 5-point standard curve for each plate, which was used to calculate plate efficiencies (1.982-1.991 for

RAG1 plates; 2.073-2.144 for telomere plates). The r² for all plates was >0.985. Efficiency controlled relative telomere lengths (T/S ratios) were calculated for each sample using the

2.5ng point of the standard curve as the “gold” control sample for each plate following Pfaffl

(2001) using the following equation:

T/S = (ETELO^(CtTELO[GOLD] – CtTELO[SAMPLE]))/ (ERAG1^(CtRAG1[GOLD] – CtRAG1[SAMPLE]))

ETELO and ERAG1 are the reaction efficiencies of each target plate on which a sample was run, CtTELO[GOLD] and CtTELO[SAMPLE] are the mean Cts of the gold and experimental sample on the telomere plate, respectively and, similarly, CtRAG1[GOLD] and

CtRAG1[SAMPLE] are the mean Cts of the gold and experimental sample on the telomere plate, respectively. Technical replicates falling outside 0.5Cts were excluded or repeated, as were samples that fell out of the standard curve. Samples were assigned to plates randomly, but all samples for a particular individual were run on the sample plate to minimise the effect of interplate variation. The average intraplate variation of the Ct values for RAG1 and telomere plates was 0.48% and 0.57%, respectively. Intraplate variation was calculated as the coefficient of variation of the replicates of the gold sample (2.5ng standard curve point) within a given plate. Interplate variation, calculated as the coefficient of variation of the ΔCt for the gold sample on all plates, was 2.46%.

114

TELOMERE DYNAMICS

Dataset

Repeated blood samples were taken from 132 individuals of varying sex and size and assayed for telomere length. As the lightest breeding female was 84g, the focus of this study is on the 45 adult females that weighed 84g or more. As in Chapter 5, females were separated according to social context, being either a breeder, a nonbreeder resident in a group (in-group nonbreeder), or a nonbreeder found in isolation (solitary nonbreeder). Nonbreeding residents will with high probability be natal females that have yet to disperse, whereas solitary nonbreeders represent females that have dispersed and settled but have yet to be encountered by a male (some for more than two years). Reproductive activity or inactivity can be easily assessed by visual inspection, with prominent nipples and a perforated vagina identifying breeding females. Three females changed status between their two sampling events from an in-group non-breeder to a breeder. As the results were qualitatively equivalent regardless of how these females were classed, the presented results treat these females as in-group nonbreeders. The dataset therefore comprises 18 breeders, 16 in-group nonbreeders and 11 solitary nonbreeders. The mean time between repeat blood samples was 273.11 ± 11.92 days

(range 187 – 419 days); the mean group size of females (excluding solitary females) across sampling events was 8.21 ± 0.63.

Statistics

The rate of change in relative telomere length, RTL, was modelled in a linear mixed effects model with a Gaussian error distribution. Social status (three-level factor), initial RTL, the change in body mass, and the time difference between two sampling events were entered as fixed effects, and location was entered as a random effect. A random effect of group was not specified as in only a few instances were multiple females sampled from one group.

Interactions were also not specified to avoid model overfitting, and continuous covariates were scaled to a mean of zero and unit standard deviation. The results of the full model are presented (p < 0.05) with the significance of fixed effects determined through likelihood ratio testing.

115

TELOMERE DYNAMICS

RESULTS

Wild female Damaraland mole-rats of differing social status (dominant breeder versus in-group nonbreeder versus solitary nonbreeder) did not differ in their initial relative telomere length (ANOVA, F2,42 = 0.210, p = 0.81) or their body weight (ANOVA, F2,42 = 0.292, p

= 0.75) at first telomere sampling. Several factors affected the change in relative telomere length across the two sampling events (Table 6.2). Faster attrition rates were associated with longer telomeres, whilst increasing time between sampling events also increased attrition rate.

Having controlled for these two factors, the model suggested a significant role of social status on rates of change in telomere length (Figure 6.3; χ22 = 6.01, p = 0.049). Solitary non-breeders showed a tendency to experience higher telomere attrition than both breeders and in-group non-breeders (without quite attaining significance in post-hoc tests: solitary nonbreeder – breeder contrast: -0.36 ± 0.19, p = 0.12; solitary nonbreeder – in-group nonbreeder contrast: -

0.38 ± 0.16, p = 0.12; in-group nonbreeder – breeder contrast: 0.02 ± 0.17, p = 0.91). The telomere length of breeders and in-group nonbreeders were predicted to change little over the mean sampling interval of 9 months (273 days).

116

TELOMERE DYNAMICS

Figure 6.3. Estimated change in relative telomere length (RTL) in wild female Damaraland mole-rats. Points estimate mean predicted change from the full model when all other covariates were held at their mean value. Error bars provide 95% confidence intervals based upon the fixed effects only. Predictions were made over a 273-day period, the mean sampling interval in our dataset.

117

TELOMERE DYNAMICS

Table 6.2. Model output for linear mixed effects model examining factors affecting change in relative telomere length. *Specific comparisons between categories of female (status) documented in main text; breeding female forms the reference category in the intercept. RTL- relative telomere length. The location term explained a small amount of the random effects variation (0.033; residual variation 0.229).

Fixed covariate Estimate Standard Wald p value Error Statistic (χ2), df Intercept -0.027 0.174 Initial RTL -0.512 0.081 30.99, 1 < 0.001 Time Between Sampling -0.158 0.079 5.34, 1 0.021 Change in Weight 0.117 0.085 3.08, 1 0.079 Status*: In-group Nonbreeder 0.020 0.172 6.01, 2 0.049 Solitary Nonbreeder -0.360 0.188

118

TELOMERE DYNAMICS

DISCUSSION

The results of this study do not support the proposition that reproduction delays ageing in Fukomys mole-rats. Breeding females do not possess longer telomeres than group- living nonbreeders, nor do they show reduced rates of telomere attrition, each of which might be expected if divergent ageing trajectories were principally driven by physiological changes associated with reproduction. The only social class of female to show a tendency of telomere shortening over the sampling period were solitary nonbreeding females. Given the benefits of communal foraging to group-living in the African mole-rats (Faulkes et al. 1997), telomere attrition in these solitary females likely reflects the high energetic costs of solitary subterranean foraging. Solitary females may also face higher metabolic costs in the absence of communal huddling in the nest (see Kotze et al. 2008), which could also contribute to telomere shortening.

The absence of an association between telomere dynamics and lifespan extension in breeding Fukomys mole-rats raises questions about the proximate mechanisms underpinning the bimodal ageing patterns of mole-rats in captivity. I suggest that previous studies invested in this issue have overlooked the importance of demographic processes on mortality schedules, for although dispersal is negated by captivity, its prevention could influence behavioural patterns in groups, and potentially also induce chronic stress in nonbreeders.

Low social rank has been shown to be associated with stress levels in various other rodents in captivity (naked mole-rat: Faulkes and Abbott, 1997; rat Rattus norvegicus Popova and

Naumenko, 1972) and in cooperative breeders generally, increases in the age and size of subordinate individuals provokes increasing aggression from behaviourally dominant individuals, often escalating to severe injury or expulsion from the group (Kutsukake and

Clutton-Brock 2006; Young et al. 2006; Bruintjes and Taborsky 2008; Cant et al. 2010;).

Repeated exposure to agonistic interactions with age might raise stress profiles and curtail lifespan, but chronic stress might also stem directly from an inability to disperse. Either or both of these scenarios might explain the age-related increases in cortisol concentrations in a laboratory population of nonbreeding female Ansell’s mole-rats (Novikov et al. 2015). If these factors are operating, then the claim that breeding delays ageing per se is incorrect (Table 6.1).

It would be more accurate to state that the lifespan of nonbreeders is curtailed by demographic

119

TELOMERE DYNAMICS

processes and related influences of social rank; the status-related position within the dominance hierarchy.

Other factors have been raised as possible contributors to the differential survivorship of breeders and nonbreeders (Table 6.1). Differences in activity patterns or time budgets were hypothesized as one such factor, but there is no consistent trend across Fukomys mole-rats to suggest that reproductive females are less active than nonbreeders (Dammann and Burda

2006; Dammann et al. 2011; Schielke et al. 2012; Ooosthuizen and Bennett 2015). Another possible contributor to differential survivorship is variation in individual quality. Selective disappearance of lower quality individuals is widespread in wild vertebrate populations (van de Pol and Wright 2009), and in cooperative breeders too, larger, higher quality individuals are more likely to acquire dominance, with high quality also bringing survival benefits

(Clutton-Brock 2016). However, although such a phenomenon might contribute in part to variation in mole-rat lifespan in the wild, it cannot be invoked to explain bimodal ageing captivity as the decision to make individuals a breeder in captivity is performed randomly.

Previous studies have dismissed a role of social rank on breeder longevity in captive

Fukomys mole-rats (Dammann and Burda 2006; Dammann et al. 2011), but whether this is the case is currently unclear. The argument against social rank effects has been based on two pieces of information. The first piece of information used to dismiss social ranks effects has been the observation that sociopositive behaviours are predominant in mole-rats and aggressive behaviours are relatively rare, which was argued to limit the scope for social stress to drive differential longevity. I do not dispute these observations, as they align with general observations in our own captive system, and with data detailing equivalent cortisol levels in breeding and nonbreeding Damaraland mole-rats (Clarke et al. 2001). However, it is also the case that aggression from dominants to subordinates in captivity can sometimes be overt and prompt the forcible eviction or killing of subordinate nonbreeders, as in other singular cooperative breeders. Because of their relative rarity, such aggressive behavioural episodes may not be reflected in general baseline stress profiles, but they could nonetheless drive mortality differences in captivity, if, as we have seen in our own system, overt aggression is temporally sustained (and sometimes fatal). There is an indirect suggestion that this is the case in another subterranean cooperative breeding rodent, the northern mole vole Ellobius talpinus.

In this species in captivity, the glucocorticoid levels of nonbreeders were higher than breeders

120

TELOMERE DYNAMICS

only in the final few months of their life, though it is not mentioned whether this coincides with high rates of received aggression (Novikov et al. 2015). Generally, it is difficult to assess the role of social rank effects on lifespan in captive mole-rats as none of the studies that have followed nonbreeders longitudinally (and beyond the age at which they might normally disperse in the wild) have reported high resolution behavioural data throughout the lifespan of individuals. As a result, it is not known to what extent the social interactions of nonbreeders contribute to ageing rates.

The second piece of evidence used to dismiss social rank effects on mole-rat ageing profiles was the absence of contrasts in survivorship between dominant nonbreeders versus subordinate nonbreeders, where dominance was equated with litter order (i.e. older nonbreeders were by default classed as dominant). However, such an analysis reveals nothing about rank effects between breeders and nonbreeders and cannot be used to inform the discussion surrounding breeder lifespan. To do so, one would need to distinguish social rank effects between breeders and nonbreeders from other more general effects related to reproduction. One means of doing so could be to surgically or chemically sterilise the partners of the dominant individuals of choice, and then compare the survivorship of the dominant breeder (who can no longer reproduce) against subordinate nonbreeders. By blocking reproduction through the partner rather than the focal dominant, the physiology of breeders would be maintained (and presumably associated social behaviours largely unchanged) excepting the now absent opportunity for reproduction; leaving any sex differences in lifespan as down to social rank effects. Overall then, the evidence for social rank effects is weak.

Recent work in wild meerkats Suricata suricatta supports the view that the extended lifespan of dominants in cooperatively breeding vertebrates is driven by behavioural differences rather than physiology. Despite living longer, dominant meerkats show faster rates of telomere attrition than subordinates, and the rate of telomere attrition is associated with mortality in dominants only (Cram et al. 2018). In conjunction with their higher parasite burden (Smyth et al. 2016), this suggests a cost of dominance that is also observed in other cooperative breeders (Creel 2001; Bell et al. 2012). That nonbreeders live shorter lives in the absence of such costs is principally due to the elevated mortality risk associated with dispersal

(Cram et al. 2018). Subordinate meerkats are evicted (females) or disperse voluntarily (males),

121

TELOMERE DYNAMICS

and only if dispersal is successful will individuals acquire dominant breeding rights (Hodge et al. 2008; Spong et al. 2008), invariably then remaining dominant and group-bound for the rest of their lives (Sharp and Clutton-Brock 2011). This study does not detect a similar cost of dominance in mole-rats, as telomere length is predicted to change little in breeding females

(as in other longitudinal studies in relatively long-lived species: Magellanic penguin

Spheniscus magellanicus, Cerchiara et al. 2017; wandering albatross Diomedea exulans, Hall et al.

2004; barnacle goose Branta leucopsis, Paulinyet al. 2012; leatherback turtle Dermochelys coriacea,

Plot et al. 2012). However, the parity in telomere dynamics between breeders and in-group nonbreeders also refutes the case for physiologically-mediated lifespan extension in breeders.

Of course, telomeres represent only one means by which lifespan extension might be mediated in mole-rat breeders and the sample size of this study is modest. It would therefore be fruitful to examine other hallmarks of ageing (such as changes in genome stability, epigenetic alterations or proteostasis; López-Otín et al. 2013) in the context of mole-rat dominance to further investigate the possibility of physiologically mediated divergence in ageing profiles.

Combining the evidence from studies of Fukomys mole-rats with general information about the structure of cooperative breeding societies, there is currently no good evidence supporting a physiological basis to the extended lifespan of mole-rat breeders. This does not mean that such a physiological mechanism does not exist, but given current knowledge about ageing biology, variation in telomere dynamics would have been earmarked as one the key candidates associated with lifespan variation. Instead, the differential longevity of breeders and nonbreeders observed in wild mole-rats is more consistent with a behavioural explanation, and future studies in mole-rats should aim to better determine the consequences of preventing dispersal in captivity on social interactions and stress profiles. In addition, future attempts to dissect the factors responsible for the extended lifespan of mole-rats should account for the lack of sex differences in breeder survivorship. Finally, it is worth emphasising that although this study does not find support for a role of telomeres in the status-related contrasts in mole-rat lifespan, this does not mean that telomeres are not more generally involved in the comparatively high maximum lifespan of African mole-rats compared to other similarly-sized mammals. The application of genomic techniques has revealed positive selection on various genes associated with telomere and telomerase functioning across the

122

TELOMERE DYNAMICS

African mole-rat family (Kim et al. 2011; Davies et al. 2015), but whether telomeres are causally involved in the long life of mole-rats awaits further investigation.

123

124

The association between mammalian

7 cooperative breeding and lifespan CHAPTER ______

125

ABSTRACT

In keeping with previous work in birds and insects, a recent study has suggested that cooperative breeding in mammals is associated with increases in maximum lifespan. In this chapter, I analysed an expanded dataset of longevity records across terrestrial mammals to re-examine whether cooperatively breeding mammals are relatively long-lived, but after controlling for body mass, data quality and mode of life, there was little evidence that this was the case. To confirm that this null result was not influenced by competing predictor variables present within a global phylogenetic model, lifespan variation was also analysed within each of the main mammalian families that contain cooperative breeders (i.e. the

Bathyergidae, Callitrichidae, Canidae, Cricetidae, and Herpestidae), as each of these clades also contain non-cooperatively breeding sister species that can be used in direct comparison.

These secondary analyses also found no consistent relationship between cooperative breeding and increased lifespan. The only exception to this general pattern was in the Bathyergidae, the

African mole-rats, where the social, cooperatively breeding species appear to be much longer- lived than the solitary species. However, even in the mole-rats it is difficult to attribute the long lifespans of the social species to cooperative breeding per se because solitary species have rarely been kept in captivity, which will lead to underestimates of their potential lifespan. It therefore remains possible that extended longevity is a general feature of the African mole- rats and stems directly from their subterranean lifestyle.

126

COMPARATIVE ANALYSIS OF LIFESPAN

INTRODUCTION

Comparative studies of birds and insects have shown that cooperative and eusocial breeding systems are associated with extended lifespans (Arnold and Owens 1998;

Beauchamp 2014; Downing et al. 2015; Keller and Genoud 1997). One explanation for the association between these breeding systems and lifespan is the reduction in extrinsic mortality in breeding individuals that is brought about by group-living, which selects for longer life.

This is thought to be the case in eusocial insect societies, with the ‘queens’ of many ant, bee and termite species living in a sheltered nest that is defended against predators by a large workforce (Carey 2001; Hölldobler and Wilson 1990). An alternative possibility is that long life facilitates the evolution of cooperative breeding. Ancestral reconstructions of cooperative breeding favour this explanation in birds (Downing et al. 2015), supporting earlier work that emphasised the importance of demographic processes to the evolution of avian cooperative breeding: by increasing the number of breeding individuals in a saturated habitat, high annual survival increases local breeding competition, leading to delayed dispersal, family-living, and helping (Brown 1987; Griesser et al. 2017).

Observations from certain mammals are also suggestive of a relationship between cooperative breeding and lifespan in this clade. Some of the longest lived of all mammals for their size are the cooperatively breeding social mole-rats (Dammann et al. 2011; Sherman and

Jarvis 2002). The most extreme case is the naked mole-rat Heterocephalus glaber, a 40g species that lives for three decades in captivity and shows no apparent age-related changes in mortality rate or physiology (Ruby et al. 2018; Sherman and Jarvis 2002), and for these reasons, the biology of mole-rat ageing has sparked much biomedical interest (Buffenstein 2005; Fang et al. 2014; Kim et al. 2011; Ruby et al. 2018). However, as well as being long-lived, members of this family are completely subterranean, and as individual mole-rats will spend virtually their entire life cut-off from terrestrial predators in closed tunnel systems, this raises the question of whether it is their subterranean lifestyle, their cooperative breeding, or both that contribute to their longevity. As various other cooperative breeders display fossorial tendencies, such as the meerkat and the dwarf mongoose, this question might be extrapolated to mammals generally. To this end, a recent large-scale comparative analysis by Healy (2015) examined the role of fossoriality (making extensive use of the underground strata) and cooperative breeding on lifespan across terrestrial mammals and found that when both traits

127

COMPARATIVE ANALYSIS OF LIFESPAN

were considered together in a phylogenetically controlled analysis, only cooperative breeding was significantly associated with longer lifespan.

Not all studies of mammals have detected a positive relationship between these cooperative breeding and lifespan. Using phylogenetically independent contrasts, Lukas and

Clutton-Brock (2012b) compared cooperatively breeding mammals to socially monogamous mammals and found no clear difference in maximum longevity between these two mating systems. With cooperative breeding being so sparsely represented across the mammalian tree

(less than 1% of mammals, Lukas and Clutton-Brock 2012a), slight differences in data availability or analytical method could contribute to these inconsistent results. For example, although large-scale comparative analyses allow the examination of multiple predictors within a single framework, they are also prone to the same biases as any other regression technique, and their assumptions should be scrutinised (Mundry 2014). As both cooperative breeding and subterranean living are rarely represented in longevity records, this could cause species possessing either trait to exert undue influence on model estimates. This effect could be further compounded by the presence of taxa displaying both traits, as with social mole-rats and the northern mole vole Ellobius talpinus.

In this chapter, I re-analysed the variation in maximum lifespan in terrestrial mammals to explore whether cooperative breeding is associated with longer lifespan. Although the use of maximum lifespan records as a proxy for lifespan is not without criticism (e.g. Baylis et al.

2014), the absence of detailed life table information for many species usually precludes the use of alternative ageing metrics in a comparative setting. I first carried out a large-scale phylogenetic analysis of lifespan across 719 species, controlling for subterranean living, data quality, and body mass. The application of different modelling techniques on this dataset suggested that the distribution of data across categorical predictors might have affected the ability of the models to discriminate between the effect of cooperative breeding and subterranean living on lifespan. To overcome this issue, lifespan variation was also analysed at the phylogenetic level of the family, because the five main mammalian families that contain cooperatively breeding species also contain non-cooperatively breeding species with which they could be compared (the bathyergid mole-rats, the callitrichid primates, the canids, the cricetid rodents, and the herpestid mongooses). If cooperative breeding is consistently

128

COMPARATIVE ANALYSIS OF LIFESPAN

associated with increased lifespan, one would expect a signature of cooperative breeding on lifespan in each family.

Figure 7.1. The relationship between loge adult body mass and loge maximum lifespan for 719 terrestrial mammals. Species are coloured according to Order, and cooperative breeders have been given larger, highlighted points. An ordinary least-squares regression is fitted through each Order to indicate phylogenetic differences in the scaling relationship.

129

COMPARATIVE ANALYSIS OF LIFESPAN

METHODS

Global dataset

Maximum lifespan data was taken from the AnAGE database for 719 terrestrial mammal species (Figure 7.1; De Magalhães and Costa 2009). This dataset provides estimates of the sample size for each longevity record, reflecting orders of magnitude in the number of specimens that contributed to the record. Species with fewer than 10 specimens contributing to their record (‘tiny’) were excluded. The only exception to this data restriction was the Cape mole-rat Georychus capensis, which was retained because of the general lack of longevity data for the African mole-rats, family Bathyergidae; for model fitting purposes this species was reclassified as having ‘small’ sample size. This left 323 species with a ‘small’ sample size (10-

100 specimens), 272 species with a medium sample size (100-1000), and 123 species with a

‘large sample size’ (over 1000 specimens). Sampling was not evenly distributed across cooperative and non-cooperative species: for cooperative species, large, medium and small lifespan sample sizes reflected 39.3%, 39.3% and 21.4% of species, as compared to 16.2%, 37.8% and 46.0% for non-cooperative species (Χ22 = 12.01, p = 0.002). All else being equal, longevity records for cooperative breeders are likely to be higher simply due to sampling effort.

Cooperatively breeding species were defined as those species where a proportion of females do not breed regularly and have been shown to perform alloparental care in the form of offspring provisioning (as per Lukas and Clutton-Brock, 2012, 2017; Solomon and French,

1997; n = 28 cooperative breeders). For each species further information was added on adult body mass, fossoriality and lifestyle. Fossoriality demarcated species as subterranean or non- subterranean using information from several sources (Begall et al. 2007; Mittermeier et al.

2018; n = 13 subterranean species). For lifestyle, species were defined as arboreal (n = 150), semi-arboreal (n = 70) or terrestrial (n = 499), using information from Walker’s Mammals of the

World (Nowak 1999) and the Handbook of the Mammals of the World (Mittermeier et al. 2018). I do not control for group size, as previous comparative analyses failed to detect an effect of group size on mammalian longevity (Kamilar et al. 2010). The updated mammalian phylogeny from Fritz et al. (2009) was used across all analyses.

130

COMPARATIVE ANALYSIS OF LIFESPAN

Global analysis

To investigate the influence of the chosen predictors on maximum lifespan, a global model was fitted with two different phylogenetic comparative models.

ln(maximum lifespan) ~ ln(adult body mass) + cooperative breeding +

fossoriality + lifestyle + lifespan sample size

Adult body mass was z-score transformed before model fitting. Method 1 took a phylogenetic generalised least squares (PGLS) approach using the R package phylolm (Tung

Ho and Ané 2014). PGLS is a modification of the generalised least squares approach, which incorporates information from a phylogeny to generate parameter estimates that account for the expected covariance of traits that is due to shared ancestry. The phylogenetic nonindependence is calculated through lambda (λ), a multiplier of the off-diagonal elements of the expected variance-covariance matrix. When λ = 0, the off-diagonal elements are set to zero and phylogenetic dependency is completely absent, whereas λ = 1 implies a strong phylogenetic signal in the data that is structured according to a Brownian motion model of trait evolution. A PGLS model was also fitted in the caper package (Orme et al. 2018) to serve as an additional comparison. Method 2 implemented a Bayesian phylogenetic mixed model

(PLMM) approach using the R package MCMCglmm (Hadfield 2010). PLMMs account for nonindependence between species by incorporating the phylogenetic tree as a random effect

(specific details are given in Hadfield and Nakagawa 2010). The model was run for 115000 iterations with a burn-in of 15000 and a thinning interval of 100. An inverse Wishart prior was chosen for the variance components (V = 1, nu = 0.02). Diagnostics confirmed an adequate mixing of chains for these settings.

Reduced data analysis

Maximum lifespan variation was also analysed at the level of the family for five mammalian families containing both cooperatively breeding and non-cooperatively breeding species: the bathyergid mole-rats, the callitrichid primates, the canids, the cricetid rodents, and the herpestid mongooses. For the Callitrichidae, Canidae, Cricetidae, and Herpestidae, a general linear model was fitted of the form: ln(maximum lifespan) ~ ln(adult body mass) + lifespan sample size. For the Bathyergidae, the lifespan sample size was excluded as only six

131

COMPARATIVE ANALYSIS OF LIFESPAN

species were present in the dataset. The residuals from each analysis represent the residual lifespan after controlling body mass and sample size. Emphasis is placed on visualisation of the residuals for the Bathyergidae and the Herpestidae for sample size reasons. For the remaining three families, t-tests were carried out on the residuals to compare cooperatively breeding and non-cooperatively breeding species.

132

COMPARATIVE ANALYSIS OF LIFESPAN

RESULTS

Whether cooperative breeding was associated with significant increases in lifespan was model-dependent (Table 7.1). When models were fitted using pgls or MCMCglmm, cooperative breeding was not estimated to significantly increase lifespan (pgls, p = 0.091;

MCMCglmm, p = 0.089), whereas subterranean living was associated with an 8.9% increase in lifespan: pgls, p = 0.019; MCMCglmm: p = 0.015). The significance of cooperative breeding and subterranean living and was reversed when lifespan variation was analysed using phylolm

(subterranean living, p = 0.096; cooperative breeding, p = 0.011) and reflected a 2.8% increase in lifespan. That three models disagreed in their attribution of significance to model terms despite similar effect sizes suggests that it was difficult for the models to apportion variance in lifespan to either cooperative breeding or subterranean living.

The lack of clear relationship between cooperative breeding and maximum lifespan was confirmed by analyses conducted on specific mammalian families (Figure 7.2). After correcting for body mass and sample size, residual lifespan did not vary with cooperative breeding status in the callitrichid primates (Welch’s t-test, t = 0.22, p = 0.83), the canids (t-test, t = 1.26, p = 0.24), the cricetid rodents (t-test, t = -0.46, p = 0.66), or the herpestid mongooses

(visual inspection of Figure 7.2). The only exception where cooperatively breeding species live longer are in Bathyergid mole-rats. Here the social taxa such as the naked mole-rat (31- year maximum lifespan) and Damaraland mole-rat (15.5 years) live noticeably longer than the non-social, solitary members of the family such as the silvery mole-rat Heliophobius argentocinereus (7.5 years) or the Cape mole-rat Georychus capensis (11.2 years); although very few solitary species have been maintained in captivity.

133

COMPARATIVE ANALYSIS OF LIFESPAN

Table 7.1. Global analyses of maximum lifespan across terrestrial mammals. Three models were fitted using two phylogenetic methods. Significant model terms are highlighted in bold; *** p < 0.001, ** p < 0.01, * p < 0.05, Ɨ p < 0.1.

Model Term PGLS, caper PGLS, phylolm PGLMM, MCMCglmm Estimate (SE) Estimate (SE) Estimate (95% credible interval) Intercept 3.098 (0.327) dd 3.117 (0.459) dd 3.113 (2.49 – 3.75) dd Adult Body Mass 0.314 (0.026) *** 0.249 (0.029) *** 0.316 (0.27 – 0.36) *** Cooperative Breeder: non-cooperative -0.069 (0.041) Ɨ* -0.088 (0.034) * -0.069 (-0.14 – 0.01) Ɨ Fossoriality: non-subterranean -0.299 (0.127) *** -0.279 (0.169) Ɨ -0.301 (-0.55 – - 0.06) * Lifestyle: semi-arboreal 0.081 (0.045) Ɨ 0.071 (0.054) dd 0.082 (0.00 – 0.17) Ɨ Lifestyle: arboreal 0.121 (0.051) * 0.105 (0.062) d Ɨ 0.124 (0.02 – 0.22) * Sample size: medium 0.125 (0.019) *** 0.092 (0.017) *** 0.124 (0.09 – 0.16) *** Sample size: large 0.205 (0.025) *** 0.183 (0.024) *** 0.205 (0.16 – 0.26) ***

λ = 0.971 σ2 = 0.0044 Residual variance = 0.012 (0.965-0.981, 95% CI) (0.009 – 0.016, 95% CI) Phylogenetic variance = 0.378 (0.327 – 0.440, 95% CI)

134

COMPARATIVE ANALYSIS OF LIFESPAN

Figure 7.3. The residual lifespan of cooperative breeders compared to non-cooperative breeders in the families containing several cooperative breeders. Residuals represent the remaining variation in maximum lifespan after the effects of body mass and lifespan sample size have been accounted for in a general linear model. (The bathyergidae model only included body mass as a predictor). Black points represent raw standardised residuals, while red points detail the mean ± 1 sem.

135

DISCUSSION

My study found no support for a consistent relationship between mammalian cooperative breeding and increased lifespan. This result differs from a previous treatment of the same topic by Healy (2015), but the approach taken in this chapter differed from that of

Healy (2015) in several important respects. Firstly, whereas my study defined fossoriality according to whether a species was subterranean or not, Healy (2015) decided to group semifossorial species (such as armadillos, aardvark Orcyteropus afer, or Eurasian badger Meles meles) and subterranean species (such as moles, mole-rats, pocket gophers, or cururo

Spalacopus cyanus) into a single fossorial class for analyses. But semifossorial and subterranean animals differ in their use of the underground environment and in their exposure to predators: whilst subterranean species spend virtually their entire life underground, almost all species that could be classed as semifossorial forage above ground for a large part of their activity period. As a result, significant reductions in extrinsic mortality might only be expected in subterranean species (Healy et al. 2014) and lumping the two classes could remove the influence of subterranean living on lifespan. Second, whilst our own and other studies have controlled for sample size in models of lifespan (Kamilar et al. 2010; Minias and Podlaszczuk,

2017), this factor was not accounted for by Healy. Since cooperative breeders are charismatic, group-living species that are frequently kept in captivity, their longevity estimates are likely to be upwardly inflated compared to mammals in general (see methods), and estimates should be corrected accordingly. Thirdly, by focussing on ground-dwelling (i.e. non-arboreal) mammals, Healy disregards longevity data from the callitrichid primates (the marmosets and tamarins), which represent a sizeable fraction of the total pool of cooperatively breeding species.

Different phylogenetic approaches differed in their attribution of significance to cooperative breeding or subterranean living. It is likely that the failure of different models to generate consistent inferences is due to the distribution of the categorical covariates: both subterranean species and cooperatively breeding species are rarely represented in the global dataset, and when present, they are also found in combination. For this reason, it was important to also analyse variation in maximum longevity at the level of the family, which reiterated the absence of a general trend between cooperative breeding and lifespan. A consideration of the social organisation of cooperative societies can make sense of this null

136

COMPARATIVE ANALYSIS OF LIFESPAN

result. Cooperative mammal societies are characterised by high rates of reproduction

(dominant females can breed repeatedly throughout the year) and intense intrasexual competition (Barrette et al. 2012; Clutton-Brock 2016; Clutton-Brock et al. 2006), both features which would classically be expected to reduce lifespan (Stearns 1989). Only dominant individuals survive to older ages in cooperative societies, and physiological parameters attest to the high physiological cost of dominance. For example, dominant individuals often display higher basal glucocorticoid levels (Creel 2001), and in meerkats Suricata suricatta, dominant females show elevated parasite burdens (Smyth and Drea 2016) and accelerated rates of telomere shortening compared to subordinate helpers (Cram et al. 2018). In addition, given that a previous study found no link between group size and lifespan in mammals (Kamilar et al. 2010), nor is it clear that the large group size of cooperative breeders necessarily contributes to their pace of life.

The only mammalian family where an association between cooperative breeding and lifespan is noticeable is in the African mole-rats, where social species comfortably outlive solitary species. However, this inference must be treated with caution, as unlike the social species, solitary mole-rats have attracted much less interest and are notoriously difficult to maintain in captivity, which will lead to large underestimates of the longevity of solitary species. Similarly, subterranean mammals have rarely been kept in captivity or been the focus of long-term individual-based studies, as reflected in the low preponderance of subterranean taxa in the AnAge database. It therefore seems premature to place judgement on the role of fossoriality on lifespan or other ageing parameters until more data is collated. For example, it is not particularly useful to claim that placental moles are on average characteristically short- lived in the context of all mammals (as per Williams and Shattuck 2015). For one, moles are rarely studied, and a cursory glance at an alternative ageing database (Jones et al. 2009) indicates that one of the few well-studied species, the European mole Talpa europaea (not present in this study), can live for seven years, a relatively long time for an 80g mammal.

Furthermore, it would seem to make sense to compare longevity between species that are ecologically similar apart from their fossorial habits; in the case of the placental moles, a comparison to shrews, which are also small, insectivorous, and largely non-fossorial (Hutterer

1985), would seem appropriate. Future studies assessing the role of sociality or mode of life

137

COMPARATIVE ANALYSIS OF LIFESPAN

on lifespan variation should take care to consider the availability and quality of data before conducting large-scale comparative analyses.

138

8 CHAPTER General Discussion

______

139

GENERAL DISCUSSION

OVERVIEW

The early attribution of eusociality to naked and Damaraland mole-rats has perpetuated a general perception of social mole-rats as being unique amongst vertebrates and more akin to insects in aspects of their life history and social organisation. But can the social mole-rats really be compared to social insects? I would argue that such a comparison is unwarranted, and if mole-rat societies are to be placed in a wider comparative context, then it would be more fruitful to understand to what extent social mole-rats differ from, or provide extreme examples of, mammalian cooperative breeding.

In the most vital respect, the expression of cooperative behaviours, the social mole-rats are not exceptional. Unlike other cooperatively breeding mammals (Solomon and French

1997; Clutton-Brock 2016), mole-rats show little, if any, direct provisioning of offspring, and it is the finding of tubers, stocking of communal food stores, and maintenance and defence of the tunnel system that constitute their core cooperative actions. As such, cooperation in mole- rats is largely indirect. As in other species, the expression of cooperative behaviour is related to age and individual state (Chapter 3; Zöttl, Vullioud et al. 2016), and contrary to previous ideas that individuals display role specialisation (like the caste systems of eusocial insects), I have shown that commitments to cooperation are general; individuals differing in the extent to which they commit to all forms of cooperative activities (Chapter 3).

Aspects of Damaraland mole-rat ageing have also been presented as atypical amongst mammals. Specifically, the extended lifespan of breeders compared to nonbreeders has been suggested to be a consequence of earlier and more rapid ageing in nonbreeders, but to date, no studies in mole-rats have measured intrinsic ageing rates using an informative biomarker.

Here, by measuring telomere dynamics in wild Damaraland mole-rats, I find no support for differential ageing rates among breeders and nonbreeders (Chapter 6). Instead, I suggest that the most parsimonious explanation for the bimodal ageing pattern observed in social mole- rats is related to demography, as it is in meerkats (Cram et al. 2018). Being unable to disperse in captivity, the curtailed life expectancy of helpers is probably driven by increasing rates of aggression directed at them by breeders, prompting eviction and physiological stress.

One case where the social mole-rats can be said to represent outliers amongst mammals concerns the magnitude of reproductive suppression experienced by helpers. In all

140

GENERAL DISCUSSION

other cooperatively breeding mammals, female helpers enjoy a small fraction of reproductive opportunities. In the social mole-rats, reproduction within a group is entirely limited to a single breeding female. The proximate causes of this suppression are well understood

(occurring at the level of the pituitary, Bennett and Faulkes 2000), the social factors underlying this suppression less so. It seems likely that limited reproductive opportunities are involved, but a role of dominance control can also not be ruled out. Either way, the suppression is such that helpers appear to delay puberty and sit in a state of arrested development (see Pinto et al. 2009 for analogous interpretation in naked mole-rats). When afforded the opportunity to transition to a dominant breeding role, individuals are released from this state and undergo a secondary growth spurt that presumably helps to consolidate status and enhance fecundity; in the process driving the morphological divergence of breeders and nonbreeders. I have shown that for Damaraland mole-rats, this morphological divergence is driven largely by increases in lumbar vertebral growth across episodes of reproductive activity (Chapter 5), as it is in naked mole-rats. The physiological basis of this growth plasticity is at present uncertain, but collaborative efforts with other laboratories are ongoing to better understand the epigenetic and hormonal changes associated with vertebral lengthening in breeders. It is also unclear to what extent the phenomenon should be regarded as unique to social mole-rats, or a general feature of the bathyergid family and their subterranean habits.

Overall, my thesis adds to the growing evidence base indicating that the uniqueness of mole-rat social organisation has been overstated. Despite this fact, the life histories of cooperatively breeding mammals provide interesting deviations from the life histories of classically studied mammalian systems where polygyny is the norm. In the following sections,

I will talk more specifically about aspects of growth, ageing and behaviour in Damaraland mole-rats, briefly summarising how my thesis has contributed to what is known, and offering possible avenues for future work that could help to better understand what is not known.

141

GENERAL DISCUSSION

GROWTH

In societies where size strongly dictates reproductive opportunities, one might expect selection to favour the evolution of highly flexible growth strategies. In keeping with this prediction, recent work suggests that vertebrates may control their own growth as well as the growth of their offspring in a strategic fashion. Studies of several group-living fish show that dominant individuals direct frequent aggression at subordinates that approach them in body mass and that subordinates constrict their food intake and rate of growth as their mass approaches that of the individual immediately above them in the hierarchy (Buston 2003;

Wong et al. 2008). Fish have served as an invaluable litmus test of the work in this area, with the lack of relatedness of within-group individuals and the patchy nature of resources across the landscapes simplifying ecological complexity. Similar insights have more recently been shown in mammals, with a study of meerkats showing that subordinate helpers will increase their food intake and rate of growth when the mass of the individual immediately below them in the hierarchy converges upon their own mass (Huchard and Clutton-Brock 2016).

The secondary growth of Damaraland mole-rats that acquire a breeding position

(Chapter 5) can also be thought of as a form of strategic growth, because the growth spurt displayed by females appears to drive fecundity increases. There is some anecdotal evidence from naked mole-rats that helpers might also modulate their own growth according to reproductive opportunity. For example, the death of a breeding male or older group members was followed by short-term growth acceleration in other group members, and the magnitude of this effect was greatest in older individuals (O’Riain and Jarvis 1998). Given the anticipated importance of such growth adjustments for fitness, the further investigation of growth plasticity in mole-rats could form an important avenue of ongoing research.

142

GENERAL DISCUSSION

AGEING

In insects, cooperative rearing has important consequences for female longevity

(Keller and Genoud 1997). In the social species where queens are nest-bound and provisioned by workers, reproductive females have been known to live beyond twenty years of age, whilst workers are lucky to survive beyond a few months (Keller 1998). This pattern shows superficial similarities to the extended lifespan of Damaraland mole-rat breeders compared to nonbreeders, but whereas part of the rank-related divergence in lifespan in insects can be explained by differences in intrinsic rate of ageing, it is not clear that that the transition to a breeding position in Damaraland mole-rats likewise leads to slowed ageing (Chapter 6). This inference is based partly from my analysis of telomere dynamics, as breeding females in the wild do not seem to possess longer telomeres or experience faster rates of telomere attrition than in-group nonbreeders, and partly from information about the demography of cooperative breeding mammals.

To more fully address the causes of breeder lifespan extension, it will be necessary to combine behavioural and physiological information from individuals tracked longitudinally.

Ideally one would want to follow individuals across their entire lifespan, but it is unrealistic to expect research institutions to follow such long-lived mammals for two decades. This being the case, zoos might provide particularly useful repositories of long-lived animals that could be sampled non-invasively. Alternatively, experiments that manipulate the social status of known-age individuals can act as a useful surrogate to study ageing processes within feasible time-scales. There are intentions to carry out such experiments in our own laboratory, mimicking the experimental paradigm undertaken in Chapter 5 that I carried out in the context of morphology, and doubtless other institutions are carrying out similar investigations.

I focused my ageing analysis on telomeres because they have been shown to have useful predictive capacity of morbidity and mortality in the wild (Young 2018). However, longitudinal studies in cooperatively and non-cooperatively breeding species have also shown that telomere dynamics in the wild can be complex because telomeres do not always shorten in a simple linear fashion (e.g. Soay sheep Ovis aries, Fairlie et al. 2016; banded mongooses Mungos mungos, Hares et al. 2017). With this is mind, it would be useful to extend

143

GENERAL DISCUSSION

the timeframe of my study of telomere dynamics in wild mole-rats to incoporate more repeats per individual. In addition, it would be preferable to use the Telomere Restriction Fragment method (TRF: Harley et al. 1990) to measure telomere length, as this techniques is known to be higher resolution and more reliable than the qPCR method I employed herein. On the downside, TRF is very labour-intensive and requires large volumes of blood, which makes it not particularly tractable for mole-rats.

Telomeres are just one of several possible hallmarks of ageing that could be assayed in Damaraland mole-rats (López-Otín et al. 2013). One particularly exciting development in recent years is the use of epigenetic clocks to measure biological ageing (Horvath 2013). This method relies on the strong correlation between age and methylation of CpG sites in the genome. Once calibrated with known age individuals, deviations from the expected levels of

DNA methylation for a given age reflect relative health status: individuals with a relatively hypermethylated epigenome for their age being in poorer health, and vice versa. It would be desirable to apply this method to status-related ageing in mole-rats, but this would rely on the sampling of known-age individuals to ‘build the clock’. Morevoer, the epigenetic clock has only been applied to humans Homo sapiens (Horvath 2013), chimpanzees Pan troglodytes

(Ito et al. 2018), dogs (Ito et al. 2017), the wolf Canis lupus (Thompson et al. 2017), and the humpback whale Megaptera novaeangliae (Polanowski et al. 2014) to date, so applying the method to mole-rats would require considerable bioinformatic validation.

Irrespective of the causes of bimodal ageing patterns in Fukomys mole-rats, the social mole-rats are exceptionally long-lived for their size. There has been the suggestion that long lifespan is a general feature of cooperatively breeding mammals (Healy 2015), but I find that this conclusion is premature. In none of the callitrichid primates, the canids, the cricetid rodents, or the herpestid mongooses, do cooperatively breeding species significantly outlive their non-cooperative sister taxa (Chapter 7). The only clade in which this might be the case is the African mole-rats, but as solitary mole-rats are rarely maintained in captivity it is not currently possible to infer a causal association between sociality and lifespan in the social mole-rats; despite a recent study suggesting that naked mole-rats mortality rates ‘defy

Gompertzian laws by not increasing with age’ (Ruby et al. 2018).

144

GENERAL DISCUSSION

BEHAVIOUR

In virtually all cooperatively breeding species, the distribution of cooperative behaviour varies widely across individuals (Koenig and Dickinson 2016). As with other behaviours, there is growing recognition that cooperative behaviour is not always plastic but may show high levels of consistency across contexts (Bergmüller et al. 2010). In Damaraland mole-rats, the idea that behaviour might be highly consistent across contexts was taken to its extreme with the suggestion that individuals could be assorted into discrete castes that differed in their relative contributions to cooperative behaviour (Bennett and Faulkes 2000).

My own work (Chapter 3), in combination with previous studies (Zöttl, Vullioud et al. 2016), has falsified this claim: individuals differ in the amount they contribute to all forms of cooperation, and this contribution is heavily influenced by age and relative mass (or state).

This finding relied on longitudinal behavioural data collected from multiple individuals throughout their development, alongside multilevel models that could partition the behavioural variation.

To date few studies of mole-rats have made use of longitudinal behavioural data, which could be used to address numerous other questions concerning the social organisation of Damaraland mole-rats. At the simplest level, it remains unclear how dominance hierarchies are structured in Damaraland mole-rats, or how perturbations to group structure influence social dynamics. Of greater relevance to this thesis is the realisation that individual variation in growth and ageing trajectories are sure to be inextricably tied to behavioural interactions with other group members. I have already stressed the possible importance of breeder aggression on nonbreeding lifespan (Chapter 6); I have suggested that mole-rats can modify their growth patterns according to social circumstance (Chapter 5); and I have shown that group structure might in turn impose constraints on growth (Chapter 4).

Our laboratory set-up in the Kalahari is ideally suited to study the interplay of behaviour, growth and ageing in Damaraland mole-rats. Individuals can be observed and readily weighed; blood sampling is straightforward, and experiments can easily be set up that manipulate group structure, food quality or workload. One must always consider the possible role of captivity on the life history of a study species (e.g. Künzl et al 2003). We have attempted to design our tunnel systems to be as ecologically relevant as possible, but there will always

145

GENERAL DISCUSSION

remain the possibility that mole-rats in the wild behave slightly differently. The clearest difference is the size of the tunnel system. Mole-rats tunnel systems in the wild can extend for hundreds of metres. In captivity we are limited to 16m of tunnel. The implications of this difference are hard to foresee, but mole-rat researchers have been resigned to the fact that by choosing a subterranean mammal as the focus of the of their studies, they relinquished the ability to say anything about behaviour in a natural setting. However, with advances in technological capacity there is now the potential for aspects of mole-rat behaviour to be studied in the wild. Radiotelemetry can be used to determine the locations of individuals below ground, and information from accelerometers might be employed to trace specific behaviour patterns related to workload. The application of these technologies will be hugely useful in furthering understanding of mole-rat social organisation.

Tell a member of the public that you work on mole-rats and you will be met with one of two questions: “is it true they don’t get cancer?” or “don’t they live forever?”. Along these lines the social mole-rats have often been viewed as subterranean oddities, and in certain aspects of their life history they are undoubtedly extreme cases in the mammalian realm. But extremes, not exceptions. To date, the life histories of the social mole-rats have often been set apart from other organisationally complex mammal societies. From my own studies of growth, ageing and behaviour in Damaraland mole-rats, it seems that this species should be viewed no differently from other cooperatively breeding mammals. The question remains as to whether the same can be said of its distant cousin, the naked mole-rat.

146

REFERENCES

References ______

Abbot P., Abe J., Alcock J., Alizon S., Alpedrinha J.A.C., Andersson M., Andre J-B., van Baalen M., Balloux F., Balshine S., et al. 2011. Inclusive fitness theory and eusociality. Nature 471(7339): E1-4-10.

Abbott D.H. 1984. Behavioral and physiological suppression of fertility in subordinate marmoset monkeys. American Journal of Primatology 6: 169-186.

Allard M.W. and R.L. Honeycutt. 1992. Nucelotide sequence variation in the mitochondrial 12S rRNA gene and the phylogeny of African mole-rats (Rodentia: Bathyergidae). Molecular Biology and Evolution 9: 27–40.

Alós J., Palmer M., Alonso-Fernández A., and B. Morales-Nin. 2010. Individual variability and sex-related differences in the growth of Diplodus annularis (Linnaeus , 1758). Fisheries Research 101: 60–69.

Altmann, J. 1974. Observational study of behaviour: sampling methods. Behaviour 49: 227–267.

Anderson C., and N.R. Franks. 2001. Teams in animal societies. Behavioral Ecology 12: 534-540.

Andersson, M. 1994. Sexual Selection. Princeton University Press, Princeton, USA.

Anyan J.J., Seney M.L., Holley A., Bengston L., Goldman B.D., Forger N.G., and M.M. Holmes. 2011. Social status and sex effects on neural morphology in Damaraland mole-rats, Fukomys damarensis. Brain, Behavior and Evolution 77(4): 291–298.

Arendt J.D. 1997. Adaptive intrinsic growth rates: an integration across taxa. The Quarterly Review of Biology 72(2): 149-177.

Arnold K.E., and I.P.F. Owens. 1998. Cooperative breeding in birds: a comparative test of the life history hypothesis. Proceedings of the Royal Society of London B 265(1398): 739–745.

Arnold K.E., Owens, I.P.F., and A.W. Goldizen. 2005. Division of labour within cooperatively breeding groups. Behaviour 142: 1577–1590.

Arnold W. 1990. The evolution of marmot sociality: II. Costs and benefits of joint hibernation. Behavioral Ecology and Sociobiology 27(4): 239-246.

Badyaev A.V. 2002. Growing apart: an ontogenetic perspective on the evolution of sexual size dimorphism’, Trends in Ecology and Evolution 17(8): 369–378.

Barette M., Monfort S.L., Festa-Bianchet M., Clutton-Brock T.H., and A.F. Russell. 2012. Reproductive rate, not dominance status, affects fecal glucocorticoid levels in breeding female meerkats. Hormones and Behavior 61: 463- 471.

Barton R.A. 1993 Sociospatial mechanisms of feeding competition in female olive baboons Papio anubis. Animal Behaviour 46: 791–802.

Batra S.W.T. 1966. Nests and social behaviour of halictine bees of India. Indian Journal of Entomology 28:375 393.

Baylis, S., de Lisle, M., Hauber, M. 2014. Inferring maximum lifespan from maximum recorded longevity in the wild carries substantial risk of estimation bias. Ecography 37: 770-780.

Beauchamp, G. 2014. Do avian cooperative breeders live longer? Proceedings of the Royal Society of London B 281(1787): 6–10.

Beehner J. C., and A. Lu. 2013. Reproductive suppression in female primates: A review. Evolutionary Anthropology 22(5): 226–238.

Beekman M., Peeters C., and M.J. O’Riain. 2006. Developmental divergence: Neglected variable in understanding the evolution of reproductive skew in social animals. Behavioral Ecology 17(4): 622–627.

Begall, S., Burda, H., and C.E. Schliech. 2007. Subterranean rodents: News from underground. Springer Verlag, Berlin Heidelberg.

Bell M.B.V., Nichols H.J., Gilchrist J.S., Cant M.A., and S.J. Hodge. 2012. The cost of dominance: suppressing subordinate reproduction affects the reproductive success of dominant female banded mongooses. Proceedings of

147

REFERENCES the Royal Society of London B 279: 619–624.

Bennett N.C. 1990. Behaviour and social organization in a colony of the Damaraland mole-rat Cryptomys damarensis. Journal of Zoology 220: 225-248.

Bennett N.C., and C.G. Faulkes. 2000. African mole-rats: Ecology and Eusociality. Cambridge Univ. Press, Cambridge, UK.

Bennett N.C., and G.H. Aguilar. 1995 The reproductive biology of the giant Zambian mole-rat, Cryptomys mechowi (Rodentia: Bathyergidae). South African Journal of Zoology 30: 1 -4.

Bennett N.C., and J.U.M. Jarvis. 1988. The social structure and reproductive biology of colonies of the mole-rat Cryptomys damarensis (Rodentia: Bathyergidae). Journal of Mammalogy 69: 293-302.

Bennett N.C., and J.U.M. Jarvis. 2004. Cryptomys damarensis. Mammalian Species 756: 1–5.

Bennett N.C., and R. Navarro. 1997. Differential growth patterns between successive litters of the eusocial Damaraland mole-rat, Cryptornys durnarensis, from Namibia. Journal of Zoology 241: 465–473.

Bennett N.C., Faulkes C., and A. Molteno. 1996. Reproductive suppression in subordinate, non-breeding female Damaraland mole-rats: two components to a lifetime of socially induced infertility. Proceedings of the Royal Society of London B 263: 1599–1603.

Bennett N.C., Jarvis J.U.M., Aguilar G.H. and E.J. McDaid. 1991. Growth and development in six species of African mole-rats (Rodentia: Bathyergidae). Journal of Zoology 225: 13–26.

Bergmüller R., Schürch R., and I.M. Hamilton. 2010. Evolutionary causes and consequences of consistent individual variation in cooperative behaviour. Philosophical Transactions of the Royal Society of London B 365(1553): 2751–64.

Biedermann P.H.W., and M. Taborsky. 2011. Larval helpers and age polyethism in ambrosia beetles. Proceedings of the National Academy of Sciences USA 108(41): 17064-17069.

Blount J.D., Vitikainen E.I.K., Stott I., and M.A. Cant. 2016. Oxidative shielding and the cost of reproduction. Biological Reviews 91(2): 483–497.

Boesch, C. 2002. Cooperative hunting roles among Taï chimpanzees. Human Nature 13: 27-46.

Boomsma J.J., and R. Gawne. 2018. Superorganismality and caste differentiation as points of no return: how the major evolutionary transitions were lost in translation. Biological Reviews 93: 28-54.

Boonekamp J.J., Simons M.J.P., Hemerik L., and S. Verhulst. 2013. Telomere length behaves as biomarker of somatic redundancy rather than biological age. Aging Cell 12: 330–332.

Bourke A.F.G. 1999. Colony size, social complexity and reproductive conflict in social insects. Journal of Evolutionary Biology 12(2): 245–257.

Bourke A.F.G. 2011. The validity and value of inclusive fitness theory. Proceedings of the Royal Society of London B 278(1723): 3313–3320.

Bourke A.F.G., and N.R. Franks. 1995. Social evolution in ants. Princeton Univ. Press, Princeton, USA.

Bourke, A.F.G. 1999. Colony size, social complexity and reproductive conflict in social insects. Journal of Evolutionary Biology. 12: 245-257.

Bowman B.M., and Miller S.C. 2001. Skeletal adaptations during mammalian reproduction. Journal of musculoskeletal and neuronal interactions 1(4): 347-355.

Brett R.A. 1991. The population structure of naked mole-rat colonies. In The Biology of the Naked Mole Rat (eds Sherman, P.W., Jarvis, J.U.M., and R.D. Alexander), pp. 97-137. Princeton Univ. Press, Princeton, USA.

Brown J.L. 1974. Alternate routes to sociality in Jays-With a theory for the evolution of altrusim and communal breeding. American Zoologist 14: 63–80.

Brown J.L. 1987. Helping and communal breeding in birds. Princeton, New Jersey, USA.

Bruintjes R., and M. Taborsky. 2008. Helpers in a cooperative breeder pay a high price to stay: effects of demand, helper size and sex. Animal Behaviour 75: 1843–1850.

148

REFERENCES

Bruintjes R., and M. Taborsky. 2011. Size-dependent task specialization in a cooperatively cichlid in response to experimental variation of demand. Animal Behaviour 81(2): 387-394.

Buffenstein R. 2000. Ecophysiological responses of subterranean rodents to underground habits. In Life underground, the biology of subterranean rodents (eds Lacey, E.A., Patton, J.L., and G.N. Cameron). Chicago University Press, Chicago.

Buffenstein R. 2005. The naked mole-rat: a new long-living model for human aging research. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences 60(11): 1369–1377.

Buffenstein R. 2008. Negligible senescence in the longest living rodent, the naked mole-rat: insights from a successfully aging species. Journal of Comparative Physiology. B, Biochemical, Systemic, and Environmental Physiology, 178(4), 439–45.

Burda H. 1990. Constraints of pregnancy and the evolution of sociality in mole-rats. With special reference to reproductive and social patterns in Cryptomys hottentotus (Bathyergidae, Rodentia). Journal of Zoology Systematics and Evolutionary Research. 28: 26-39.

Burkart J. M., and van Schaik C.P. 2009. Cognitive consequences of cooperative breeding in primates? Animal Cognition 13(1): 1–19.

Burkart J.M., and van Schaik C.P. 2016. Revisiting the consequences of cooperative breeding. Journal of Zoology 299(2): 77–83.

Burland T.M., Bennett N.C., Jarvis J.U.M. and C.G. Faulkes. 2004. Colony structure and parentage in wild colonies of co-operatively breeding Damaraland mole-rats suggest incest avoidance alone may not maintain reproductive skew, Molecular Ecology 13(8): 2371–2379.

Buston P. 2003. Size and growth modification in clownfish. Nature 424: 145-146.

Campisi J. 2003. Cellular senescence and apoptosis: how cellular responses might in uence aging phenotypes. Experimental Gerontology 38: 5–12.

Cant M.A, Hodge S. J., Bell M.B.V, Gilchrist J.S., and H.J. Nichols. 2010. Reproductive control via eviction (but not the threat of eviction) in banded mongooses. Proceedings of the Royal Society of London B 277(1691): 2219–2226.

Cant M.A. 2011. The role of threats in animal cooperation. Proceedings of the Royal Society of London B 278(1703): 170– 178.

Cant M.A., Nichols H.J., Johnstone R.A. and S.J. Hodge. 2014. Policing of reproduction by hidden threats in a cooperative mammal. Proceedings of the National Academy of Sciences USA 111: 326–330.

Carey J. 2001. Demographic mechanisms for the evolution of long life in social insects. Experimental gerontology 36: 713-722.

Cawthorn R.M. 2002. Telomere measurement by quantitative PCR. Nucleic Acid Research 30: 1–6.

Cerchiara J.A., Risques R.A., Prunkard D., Smith J.R., Kane O.J., and P.D. Boersma, P. D. 2017. Magellanic penguin telomeres do not shorten with age with increased reproductive effort, investment, and basal corticosterone. Ecology and Evolution 7: 5682–5691.

Ciszek D. 2000. New colony formation in the “highly inbred” eusocial naked mole-rat: outbreeding is preferred. Behavioural Ecology 11: 1-6.

Clarke B.L., and S. Khosla. 2009. Androgens and bone. Steroids 74(3): 296-305.

Clarke B.L., and S. Khosla. 2010. Female Reproductive System and Bone. Archives of Biochemistry and Biophysics 503(1): 118-128.

Clarke F.M., and C.G. Faulkes. 1997. Dominance and queen succession in captive colonies of the eusocial naked mole-rat, Heterocephalus glaber. Proceedings of the Royal Society of London B 264(1384): 993–1000.

Clarke F.M., and C.G. Faulkes. 2001. Intracolony aggression in the eusocial naked mole-rat, Heterocephalus glaber. Animal Behaviour 61(2): 311–324.

Clarke F.M., Miethe G.H. and N.C. Bennett. 2001. Reproductive suppression in female Damaraland mole-rats Cryptomys damarensis: dominant control or self-restraint. Proceedings of the Royal Society of London B 268: 899-909.

149

REFERENCES

Clutton-Brock T. 1991a. The Evolution of Parental Care. Princeton Univ. Press, Princeton, USA.

Clutton-Brock T. 1991b. The evolution of sex differences and the consequences of polygyny. In The Development and Integration of Behaviour: Essays in Honour of Robert Hinde (eds P. Bateson), pp. 229-253. Cambridge Univ. Press, Cambridge University Press.

Clutton-Brock T. 1998. Reproductive skew, concessions and limited control. Trends in Ecology and Evolution 13: 288- 292.

Clutton-Brock T. 2016. Cooperative Breeding. In Mammalian Societies (T Clutton-Brock), pp. 557–604. John Wiley and Sons Ltd, Chichester, UK.

Clutton-Brock T., and M. Manser. 2016. Meerkats: Cooperative breeding in the Kalahari. In Cooperative Breeding in Vertebrates (eds Koenig W.D. and J.L. Dickinson) pp. 294-317. Cambridge, UK. Cambridge University Press.

Clutton-Brock T., Russell A.F., and L.L. Sharpe. 2003. Meerkat helpers do not specialize in particular activities. Animal Behaviour 66(3): 531-540.

Clutton-Brock T., Russell A.F., Sharpe L.L., Brotherton P.N., McIlrath G.M., White S., and E.Z Cameron. 2001. Effects of helpers on juvenile development and survival in meerkats. Science 293: 2446-2449.

Clutton-Brock T., Russell A.F., Sharpe L.L., Young, A.J., Balmforth, Z., and G.M. McIlrath. 2002. Evolution and development of sex differences in cooperative behaviour in meerkats. Science 297: 253-256.

Clutton-Brock T.H., and M.B. Manser. 2016. Meerkats: Cooperative breeding in the Kalahari. In Cooperative Breeding in Vertebrates: Studies of Ecology, Evolution, and Behavior (eds W.D. Koenig and J.L. Dickinson), pp. 294–317). Cambridge Univ. Press, Cambridge, UK.

Clutton-Brock T.H., Brotherton P.N., O’Riain M.J., Griffin A.S., Gaynor D., Sharpe L., Kansky, R., Manser M.B., and G.M. McIlrath. 2000. Individual contributions to babysitting in a cooperative mongoose, Suricata suricatta. Proceedings of the Royal Society of London B (1440): 301–5.

Clutton-Brock T.H., Brotherton P.N., Russell A.F., O’Riain M.J., Gaynor D., Kansky R., Griffin A., Manser M., Sharpe L., McIlrath G.M. et al. 2001. Cooperation, control, and concession in meerkat groups. Science 291(5503): pp. 478–81.

Clutton-Brock T.H., Hodge S.J., Spong G., Russell A.F., Jordan N.R., Bennett N.C., Sharpe L.L. and M.B. Manser. 2006. Intrasexual competition and sexual selection in cooperative mammals. Nature, 444(7122): pp. 1065–1068.

Clutton-Brock, T.H., Albon, S.D., and F.E. Guinness. 1989. Fitness costs of gestation and lactation in wild mammals. Nature 337: 260-262.

Cockburn A. 2006. Prevalence of different modes of parental care in birds. Proceedings of the Royal Society of London B 273(1592): 1375–1383.

Cooney R. 2002. Colony defense in Damaraland mole-rats, Cryptomys damarensis. Behavioral Ecology 13(2): 160–162.

Cooney R., and N.C. Bennett. 2000. Inbreeding avoidance and reproductive skew in a cooperative mammal. Proceedings of the Royal Society of London B 267: 802–806.

Cornwallis C.K., West S.A. and A.S. Griffin. 2009. Routes to indirect fitness in cooperatively breeding vertebrates: kin discrimination and limited dispersal. Journal of Evolutionary Biology 22(12): 2445–57.

Corona M., Velarde R.A., Remolina S., Moran-lauter A., Wang Y., Hughes K.A., and G.E. Robinson 2007. Vitellogenin, juvenile hormone, insulin signaling, and queen honey bee longevity. Proceedings of the National Academy of Sciences USA 104: 7128–7133.

Cram D.L., Monaghan P., Gillespie R., Dantzer B., Duncan C., Spence-Jones H., and T. Clutton-Brock. 2018. Rank- related contrasts in longevity arise from extra-group excusrions not delayed senescence in a cooperative mammals. Current Biology 28: 2934-2939.

Creel S. 2001. Social dominance and stress hormones. Trends in Ecology and Evolution 16(9): 491–497.

Creel S., and N.M. Creel 2002. The African Wild Dog: Behaviour, Ecology and Evolution. Princeton Univ. Press, Princeton, New Jersey, USA.

Creel S., Creel N.M., Mills M.G.L., and S.L. Monfort. 1997. Rank and reproduction in cooperatively breeding African wild dogs: Behavioral and endorcrine correlates. Behavioral Ecology 8(3): 298–306.

150

REFERENCES

Creel S.R., and P.M. Waser. 1997. Variation in reproductive suppression among dwarf mongooses: interplay between mechanisms and evoltuion. In Cooperative Breeding in Mammals (eds N.G. Solomon and J.A. French), pp. 150–170. Cambridge, UK: Cambridge University Press.

Creel S.R., Monfort S.L., Wildt D.E., and P.M. Waser. 1991. Spontaneous lactation is an adaptive result of pseudopregnancy. Nature 351: 660-662.

Crespi B.J., and D. Yanega. 1995. The definition of eusociality. Behavioral Ecology 6: 109–115.

Crottini A., Madsen O., Poux C., Strauss A., Vieites D.R., and M. Vences. 2012. Vertebrate time-tree elucidates the biogeographic pattern of a major biotic change around the K-T boundary in Madagascar. Proceedings of the National Academy of Sciences USA. 109(14): 5358–5363.

Dammann P., and H. Burda. 2006. Sexual activity and reproduction delay ageing in a mammal. Current Biology 16: 117–118.

Dammann P., and H. Burda. 2007. Senescence patterns in African Mole-rats (Bathyergidae, Rodentia). Subterranean Rodents: News from Underground (eds Begall S., Burda H., and C.E. Schliech), pp. 251–263. Springer-Verlag, Berlin Heidelberg.

Dammann P., Sell D.R., Begall S., Strauch C., and V.M. Monnier. 2012. Advanced glycation end-products as markers of aging and longevity in the long-lived Ansell’s mole-rat (Fukomys anselli). Journals of Gerontology - Series A Biological Sciences and Medical Sciences 67 A(6): 573–583.

Dammann P., Sumbera R., Massmann C., Scherag A., and H. Burda. 2011. Extended longevity of reproductives appears to be common in Fukomys mole-rats (Rodentia, Bathyergidae). PLoS ONE 6(4): 2–8.

Darwin, C. 1859. On the origin of species by means of natural selection, 1st end. John Murray, London, UK.

Davies C.S., Smyth K.N., Greene L.K., Walsh D.A., Mitchell J., Clutton-Brock T., and C.M. Drea. 2016. Exceptional endocrine profiles characterise the meerkat: sex, status, and reproductive patterns. Scientific Reports 6: 1–9.

Davies K.T.J., Bennett N.C., Tsagkogeorga G., Rossiter S.J., and C.G. Faulkes. 2015. Family Wide Molecular Adaptations to Underground Life in African Mole-Rats Revealed by Phylogenomic Analysis. Molecular Biology and Evolution 32: 3089–3107.

De Magalhães, J.P., and J. Costa. 2009. A database of vertebrate longevity records and their relation to other life- history traits. Journal of Evolutionary Biology 22(8): 1770–1774.

Dengler-Crish C., and K.C. Catania. 2007. Phenotypic plasticity in female naked mole-rats after removal from reproductive suppression. Journal of Experimental Biology 210: 4351-4358.

Dengler-Crish C., and K.C. Catania. 2009. Cessation of reproduction-related spine elongation after multiple breeding cycles in female naked mole-rats. Anatomical record 292: 131-137.

Doerr E.D., and V.A.J. Doerr. 2007. Positive effects of helpers on reproductive success in the brown treecreeper and the general importance of future benefits. Journal of Animal Ecology 76(5): 966–976.

Dornhaus, A. 2008. Specialization does not predict individual efficiency in an ant. PLoS Biology 6(11): 285.

Douhard M., Gaillard J.-M., Delorme D., Capron G., Duncan P., Klein F. and C. Bonenfant. 2013. Variation in adult body mass of roe deer: early environmental conditions influence early and late body growth of females. Ecology 94(8): 1805–1814.

Downing P.A., Cornwallis C.K., and A.S. Griffin. 2015. Sex, long life and the evolutionary transition to cooperative breeding in birds. Proceedings of the Royal Society of London B 282: 1–7.

Dubuc C., and T.H. Clutton-Brock. In Press. Male immigration triggers increased growth in subordinate female meerkats.

Edrey Y.H., Hanes M., Pinto M., Mele J., and R. Buffenstein. 2011. Successful Aging and sustained good health in the Naked Mole Rat: a long-lived mammalian model for biogerontology and biomedical research. ILAR journal 52(1): 41–53.

Emlen S. and P.H. Wrege. 1991. Breeding Biology of White-Fronted Bee-Eaters at Nakuru: The Influence of Helpers on Breeder Fitness. Journal of Animal Ecology 60: 309–326.

Emlen S.T. 1982. The evolution of helping. I. An ecological constraints models. American Naturalist 119: 29-39.

151

REFERENCES

English S., Bateman A. and T. Clutton-Brock. 2012. Lifetime growth in wild meerkats: incorporating life history and environmental factors into a standard growth model. Oecologia 169: 143–153.

English S., Browning L.E., and N.J. Raihani. 2015. Developmental plasticity and social specialization in cooperative societies. Animal Behaviour 106: 37-42.

English S., Huchard E., Nielsen J.F. and T.H. Clutton-Brock. 2013. Early growth, dominance acquisition and lifetime reproductive success in male and female cooperative meerkats. Ecology and Evolution 3(13): 4401–4407.

Epel E.S., Blackburn E.H., Lin J., Dhabhar F.S., Adler N.E., Morrow J.D., and R.M. Cawthon. 2004. Accelerated telomere shortening in response to life stress. Proceedings of the National Academy of Sciences USA. 101(49): 17312– 17315.

Fairlie J., Holland R., Pilkington J.G., Pemberton J.M., Harrington L., and Nussey D.H. 2016. Lifelong leukocyte dynamics in a free-living mammal. Aging Cell 15: 140-148.

Fang, X., Seim, I., Huang, Z., Gerashchenko, M.V, Xiong, Z., Turanov, A.A., Zhu, Y., Lobanov, A.V., Fan, D., Hee Yim, S. et al. 2014. Adaptations to a subterranean environment and longevity revealed by the analysis of mole rat genomes. Cell Reports 8(5): 1354–64.

Faulkes C.G., Abbott D.H., and J.U.M. Jarvis. 1990. Social suppression of ovarian cyclicity in captive and wild colonies of naked mole-rats, Heterocephalus glaber. Journal of Reproduction and Fertility 88: 559-568.

Faulkes C.G., Abbott D.H., and J.U.M. Jarvis. 1991. Social suppression of reproduction in male naked mole-rats, Heterocephalus glaber. Journal of Reproduction and Fertility 91: 593-604.

Faulkes C.G., Abbott D.H., O’Brien, H.P., Lau L., Roy M.R., Wayne R.K., and M.W. Bruford. 1997. Micro- and macrogeographical genetic structure of colonies of naked mole-rats Heterocephalus glaber. Molecular Ecology 6(7): 615–628.

Faulkes C.G., and D.H. Abbott. 1993. Evidence that primer pheromones do not cause social suppression of reproduction in male and female naked mole-rats (Heterocephalus glaber). Journal of Reproduction and Fertility 99: 225-230.

Faulkes C.G., and D.H. Abbott. 1997. The physiology of a reproductive dictatorship: regulation of male and female reproduction in colonies of naked mole-rats. In Cooperative Breeding in Mammals (eds. N. Solomon and J. French), pp. 302–334. Cambridge University Press, Cambridge, UK.

Faulkes C.G., Bennett N.C., Bruford M.W., O’Brien H.P., Aguilar G.H., and J.U.M. Jarvis. 1997. Ecological constraints drive social evolution in the African mole-rats. Proceedings of the Royal Society of London B 264(1388), 1619–27.

Faulkes C.G., Mgode D.F., Le Comber S.C., and N.C. Bennett. 2010. Cladogenesis and endemism in Tanzanian mole-rats, genus Fukomys: (Rodentia Bathyergidae): a role for tectonics? Biological Journal of the Linnaean Society 100: 337-352.

Faulkes C.G., Mgode G.F., Archer E.K. and N.C. Bennett. 2017. Relic populations of Fukomys mole-rats in Tanzania: description of two new species F. livingstoni sp. nov. and F. hanagensis sp. nov. PeerJ 5, e3124.

Faulkes C.G., Verheyen E., Verheyen W., Jarvis J.U.M., and N.C. Bennett. 2004. Phylogeographical patterns of genetic divergence and speciation in African mole-rats (Family: Bathyergidae). Molecular Ecology 13: 613-629.

Festa-bianchet, M., Jorgenson, J.T. and D. Réale. 2000. Early development, adult mass, and reproductive success in bighorn sheep. Behavioral Ecology 11(6): 633–639.

Field J., Cronin A. and C. Bridge. 2006. Future fitness and helping in social queues. Nature 441(7090): 214-217.

Finkel T., and N. Holbrook. 2000. Oxidants, oxidative stress and the biology of ageing. Nature 408, 239–247.

Foster K., and F.L.W. Ratnieks. 2005. A new eusocial vertebrate? Trends in Ecology and Evolution 20(7): 363–364.

Fritz, S.A., Bininda-Emonds, O.R.P., and A. Purvis. 2009. Geographical variation in predictors of mammalian extinction risk: Big is bad, but only in the tropics. Ecology Letters 12(6): 538–549.

Gadagkar R. 1994. Why the Definition of Eusociality Is Not Helpful to Understand Its Evolution and What Should We Do about It. Oikos 70(3): 485–488.

Gaillard J-M., Pontier D., Allaine D., Loison A., Herve J. C. and A. Heizmann. 1997. Variation in growth form and

152

REFERENCES precocity at birth in eutherian mammals. Proceedings of the Royal Society of London B 264(1383): 859–68.

Gardner A., and S.A. West. 2007. Social evolution: the decline and fall of genetic kin recognition. Current Biology 17(18): 810–812.

Gazda S.K., Connor R.C., Edgar R.K. and F. Cox. 2005. A division of labour with role specialization in group- hunting bottlenose dolphins (Tursiops truncates) off Cedar key, Florida. Proceedings of the Royal Society of London B 272: 135-140.

Gilchrist J. S. 2006. Female eviction, abortion, and infanticide in banded mongooses (Mungos mungo): Implications for social control of reproduction and synchronized parturition’, Behavioral Ecology 17(4): 664–669.

Giray T., Giovanetti M., and M.J. West-Eberhard. 2005. Juvenile hormone, reproduction, and worker behaviour in the netropical social wasp Polistes canadensis. Proceedings of the National Academy of Sciences USA 102(9): 3330-3335.

Gomes N.M.V, Ryder O.A, Houck M.L., Charter S.J., Walker W., Forsyth N.R., Austad S.N., Venditti C., Pagel M., Shay W. and W.E. Wright. 2011. Comparative biology of mammalian telomeres: hypotheses on ancestral states and the roles of telomeres in longevity determination. Aging Cell 10(5): 761–8.

Griesser, M., Drobniak, S. M., Nakagawa, S., and C.A. Botero. 2017. Family living sets the stage for cooperative breeding and ecological resilience in birds. PLoS Biology 15(6): 1–17.

Griffin A.S., and S.A. West. 2002. Kin selection : fact and fiction. Trends in Ecology and Evolution 17: 15–21.

Grossman M., and W.J. Koops. 1988. Multiphasic analysis of growth curves in chickens. Poultry Science 67: 33–42.

Groth J. G., and G.F. Barrowclough. 1999. Basal Divergences in Birds and the Phylogenetic Utility of the Nuclear RAG-1 Gene. Molecular Phylogenetics and Evolution 12(2): 115–123.

Hӓcklander K., Möstl E., and W. Arnold. 2003. Reproductive suppression in female Alpine marmots, Marmota marmota. Animal Behaviour 65: 1133-1400.

Hadfield, J. 2010. MCMC methods for Multi-Response Generalized Linear Mixed Models: The MCMCglmm R Package. Journal of Statistical Software 33(2): 1–22.

Hadfield, J.D., and S. Nakagawa. 2010. General quantitative genetic methods for comparative biology: Phylogenies, taxonomies and multi-trait models for continuous and categorical characters. Journal of Evolutionary Biology 23(3): 494–508.

Hall M.E., Nasir L., Daunt F., Gault E.A., Croxall J.P., Wanless S., and P. Monaghan. 2004. Telomere loss in relation to age and early environment in long-lived birds. Proceedings of the Royal Society of London B 271(1548): 1571–6.

Hamilton W.D. 1964a. The genetical evolution of social behaviour. I. Journal of Theoretical Biology 7: 1–16.

Hamilton W.D. 1964b. The genetical evolution of social behaviour. II. Journal of Theoretical Biology 7: 17–52.

Hares M.C., Vitikainen E.I.K., Marshall H.H., Thompson F.J., Blount J.D., and M.A. Cant. 2018. Telomere dynamics in wild banded mongooses: evaluating longitudinal and quasi-longitudinal markers of senescnece. Experimental Gerontology 107: 67-73.

Harley C.B., Futcher A.B., and C.W. Greider. 1990. Telomeres shorten during ageing of human fibroblasts. Nature 345: 458-460.

Harris R.A., Tardif S.D., Vinar T., Wildman D.E., Rutherford J.N., Rogers J., Worley, K.C., and K.M. Aagaard. 2014. Evolutionary genetics and implications of small size and twinning in callitrichine primates. Proceedings of the National Academy of Sciences USA 111(4): 1467–1472.

Hatchwell B.J., and J. Komdeur. 2000. Ecological constraints, life history traits and the evolution of cooperative breeding. Animal Behaviour 59(6): 1079–1086.

Hauber M.E., and E.A. Lacey. 2005. Bateman’s Principle in Cooperatively Breeding Vertebrates: The Effects of Non- Breeding Alloparents on Variability in Female and Male Reproductive Success. Integrative and Comparative Biology 45(5): 903–914.

Hayward A.D., Mar K.U., Lahdenperä M., and V. Lummaa. 2014. Early reproductive investment, senescence and lifetime reproductive success in female Asian elephants. Journal of Evolutionary Biology 27(4): 772–783.

153

REFERENCES

Hazell R.W.A., Bennett N.C., Jarvis J.U.M. and M. Grif. 2000. Adult dispersal in the co-operatively breeding Damaraland mole-rat (Cryptomys damarensis): a case study from the Waterberg region of Namibia. Journal of Zoology 252: 19-25.

Healy K. 2015. Eusociality but not fossoriality drives longevity in small mammals. Proceedings of the Royal Society of London B 282: 20142917.

Healy, K., Guillerme, T., Finlay, S., Kane, A., Kelly, S. B., McClean, D., Kelly, D.J., Donohue, I., Jackson, A.L., and N. Cooper. 2014. Ecology and mode-of-life explain lifespan variation in birds and mammals. Proceedings of the Royal Society of London B 281: 20140298.

Heg D., and Z. Bachar. 2006. Cooperative breeding in the Lake Tanganyika cichlid Julidochromis ornatus. Environmental Biology of Fishes 76(2–4): 265–281.

Heg D., Bender, N., and I. Hamilton. 2004. Strategic growth decisions in helper cichlids. Proceedings of the Royal Society of London B 271: S505-S508.

Heinsohn R., and A. Cockburn. 1994. Helping is costly to young birds in cooperatively breeding white-winged choughs. Proceedings of the Royal Society of London B 256(1347): 293–298.

Heinze J., and A. Schrempf. 2008. Aging and reproduction in social insects - A mini-review. Gerontology 54(3): 160– 167.

Henning Y., Vole C., Begall S., Bens M., Broecker-Preuss M., Sahm A., Szafranski K., Burda H., and P. Dammann. 2014. Unusual ratio between free thyroxine and free triiodothyronine in a long-lived mole-rat species with bimodal ageing. PLoS ONE 9(11): e113698.

Henry E.C., Dengler-Crish, C.M. and K.C. Catania. 2007. Growing out of caste- reproduction and the making of the queen mole-rat. Journal of Experimental Biology 210: 261-268.

Herbers J. M. 2009. Darwin’s “one special difficulty”: Celebrating Darwin 200. Biology Letters 5(2): 214–217.

Hickman, G.C. 1979 A live-trapping technique for fossorial mammals. Suid-Afrikaanse Tydskrif vir Dierkunde 14, 9– 12.

Hochberg M.E., Noble, R.J., and S. Braude. 2016. A Hypothesis to Explain Cancers in Confined Colonies of Naked Mole Rats. bioRxiv 1–13. doi: https://doi.org/10.1101/079012.

Hodge S. J., Manica A., Flower T.P. and T.H. Clutton-Brock. 2008. Determinants of reproductive success in dominant female meerkats. The Journal of Animal Ecology 77: 92–102.

Hodge S.J. 2005. Helpers benefit offspring in both the short and long-term in the cooperatively breeding banded mongoose. Proceedings of the Royal Society of London B 272(1580): 2479–84.

Hodge S.J. 2007. Counting the cost: the evolution of male-biased care in the cooperatively breeding banded mongoose. Animal Behaviour 74: 911-919.

Hoffman H.A., Benson M.E., and R.D. Fernald. 1999. Social status regulates growth rate: consequences for life history strategies. Proceedings of the National Academy of Sciences USA 96: 14171-14176.

Hohenstein S., and R. Kliegl. 2013. Remef: Remove Partial Effects. R package version 1.0.6.9000.

Hölldobler, B., and E.O. Wilson. 1990. The Ants. Springer, Berlin, Germany.

Holmes M.M., Goldman B.D., Goldman S.L., Seney M.L., and N.G. Forger. 2009. Neuroendocrinology and sexual differentiation in eusocial mammals. Frontiers in Neuroendocrinology 30(4): 519–533.

Horvath S. 2013. DNA methylation age of human tissues and cell types. Genome Biology 14: R115.

Hothorn T., Bretz F., and Westfall P. 2008. Simulation inferences in general parametric models. Biometric Journal 50(3): 346-363.

Huchard E., English S., Bell M.B.V., Tavarajah N., and T. Clutton-Brock. 2016. Competitive growth in a cooperative mammal. Nature 533: 532-534.

Hudson R., and F. Trillmich. 2008. Sibling competition and cooperation in mammals: Challenges, developments and prospects. Behavioral Ecology and Sociobiology 62(3): 299–307.

154

REFERENCES

Hughes W.O.H., Sumner S., Van Borm S., and J.J. Boomsma. 2003. Worker caste polymorphism has a genetic basis in Acromyrmex leaf-cutting ants. Proceedings of the National Academy of Sciences USA 100(16): 9394-9397.

Hurtado M.J., Fénéron R., and P. Gouat. 2013. Specialization in building tasks in the mound-building mouse, Mus spicegus. Animal Behaviour 85: 1153-1160.

Hutterer, R. 1985. Anatomical adaptations of shrews. Mammal Review 15(1): 43–55.

Isler K. and C.P. van Schaik. 2012. Allomaternal care, life history and brain size evolution in mammals. Journal of Human Evolution 63: 52–63.

Ito H., Udono T., Hirata S., and M. Inoue-Murayama. 2018. Estimation of chimpanzee age based on DNA methylation. Scientific Reports 8: 9998.

Ito G., Yoshimura K., and Y. Momoi. 2017. Analysis of DNA methlation of potential age-related methylation sites in canine peripheral blood leukocytes. Journal of Veterinary Medical Science 79: 745-750.

Jacobs D., Bennett N., Jarvis J.U.M. and T. Crowe. 1991. The colony structure and dominance hierarchy of the Damaraland mole-rat, Cryptomys damarensis (Rodentia: Bathyergidae), from Namibia, Journal of Zoology 224: 553– 576.

Jandt J.M., Bengston S., Pinter-Wollman N., Pruitt J.N., Raine N.E., Dornhaus A., and A. Sih. 2014. Behavioural syndromes and social insects: personality at multiple levels. Biological Reviews. 89: 48-67.

Jarvis J.U.M. 1978. Energetics of survival Heterocephalus glaber, the naked mole-rat (Rodentia: Bathyergidae). Bulletin of the Carnegie Museum of Natural History 6: 81-87.

Jarvis J.U.M. 1981. Eusociality in a mammal: cooperative breeding in naked mole-rat colonies. Science 212(4494): 571–573.

Jarvis J.U.M. 1991. Reproduction of naked mole-rats. In The Biology of the Naked Mole Rat (eds Sherman, P.W., Jarvis, J.U.M., and R.D. Alexander), pp. 384-425. Princeton University Press, Princeton, USA.

Jarvis J.U.M., and N.C. Bennett. 1993. Eusociality has evolved independently in two genera of bathyergid mole- rats – but occurs in no other subterranean mammal. Behavioral Ecology and Sociobiology 33: 253-260.

Jarvis J.U.M., Bennett N.C. and A.C. Spinks. 1998. Food availability and foraging by wild colonies of Damaraland mole-rats (Cryptomys damarensis): implications for sociality. Oecologia 113(2): 290–298.

Jarvis J.U.M., O’Riain M.J., and E. McDaid. 1991. Growth and factors affecting body size in naked mole-rats. In The Biology of the Naked Mole Rat (eds Sherman, P.W., Jarvis, J.U.M., and R.D. Alexander), pp. 358-383. Princeton University Press, Princeton, USA.

Jarvis J.U.M., O’Riain,M.J., Bennett N.C., and P.W. Sherman. 1994. Mammalian eusociality: a family affair. Trends in Ecology and Evolution 9(2): 1–5.

Jetz W., and D.R. Rubenstein. 2011. Environmental uncertainty and the global biogeography of cooperative breeding in birds. Current Biology 21: 72–78.

Jones, K.E., Bielby, J., Cardillo, M., Fritz, S.A., O’Dell, J., Orme, C.D.L., Safi, K., Sechrest, W., Boakes, E.H., Carbone, C., et al. 2009. PanTHERIA: a species-level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology 90(9): 2648–2648.

Kamilar, J.M., Bribiescas, R.G., and B.J. Bradley. 2010. Is group size related to longevity in mammals? Biology Letters, 6(6): 736–739.

Keller L., and M. Genoud. 1997. Extraordinary lifespans in ants: a test of evolutionary theories of ageing. Nature 389: 958–960.

Keller L. 1998. Queen lifespan and colony characteristics in ants and termites. Insectes Sociaux 45: 235-246.

Keller L., S. Jemielity. 2006. Social insects as a model to study the molecular basis of ageing. Experimental Gerontology 41(6): 553–556.

Kim E.B., Fang X., Fushan A.A., Huang Z., Lobanov A.V, Han L., Marino S.M., Sun X., Turanov A.A., Yang P., et al. 2011. Genome sequencing reveals insights into physiology and longevity of the naked mole rat. Nature 479: 223– 7.

155

REFERENCES

Kim, E.B., Fang, X., Fushan, A.A., Huang, Z., Lobanov, A.V., Han, L., Marino, S.M., Sun, X., Turanov, A.A., Yang, P., Hee Yim, S., et al. 2011. Genome sequencing reveals insights into physiology and longevity of the naked mole rat. Nature 479(7372): 223–7.

Kirkwood T., and M. Rose. 1991. Evolution of senescence: late survival sacrificed for reproduction. Philosophical Transactions of the Royal Society of London B, Biological Sciences 332: 15–24.

Kock D., Ingram C.M., Frabotta L.J., and H. Burda. 2006. On the nomenclature of Bathyergidae and Fukomys. n. gen. (Mammalia: Rodentia). Zootaxa 1142: 51–55.

Koenig W.D., and J. Dickinson. 2004. Ecology and evolution of cooperative breeding in birds. Cambridge Univ. Press, Cambridge, UK.

Koenig W.D., Mumme R.L. Stanback M.T., and F.A. Pitelka. 1995. Patterns and consequences of egg destruction among joint-nesting acorn woodpeckers. Animal Behaviour 50(3): 607–621.

Kokko H. 2002. The evolution of parental and alloparental effort in cooperatively breeding groups: when should helpers pay to stay? Behavioral Ecology 13(3): 291–300.

Kokko H., Johnstone R., and T. Clutton-Brock. 2001. The evolution of cooperative breeding through group augmentation. Proceedings of the Royal Society of London B 268(1463): 187–196.

Koops W.J., and M. Grossman. 1991. Applications of a multiphasic growth functions to body composition in pigs. Journal of Animal Science 69: 3265–3273.

Korb, J., and K. Hartfelder. 2008. Life history and development – a framework for understanding development plasticity in lower termites. Biological Reviews 83(3): 295-313.

Koster J., and R. McElreath. 2017. Multinomial analysis of behaviour: statistical methods. Behavioral Ecology and Sociobiology 71: 138.

Kotze, J., Bennett, N.C., and M. Scantlebury. 2008. The energetics of huddling in two species of mole-rat (Rodentia: Bathyergidae). Physiology and Behavior 93(1-2): 215-221.

Kovacs C.S. 2016. Maternal mineral and bone metabolism during pregnancy, lactation, and post-weaning recovery. Physiological Reviews 96: 449-547.

Kovacs C.S., and H.M. Kronenberg, 2007. Maternal-fetal calcium and bone metabolism during pregnancy, puerperium, and lactation. Endocrine Reviews 18: 832-872.

Kramer B. H., and R. Schaible. 2013. Colony size explains the lifespan differences between queens and workers in eusocial Hymenoptera. Biological Journal of the Linnean Society 109(3): 710–724.

Kregel K.C., and H.J. Zhang. 2007. An integrated view of oxidative stress in aging: basic mechanisms, functional effects, and pathological considerations. American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 292: R18-36.

Kucharski R., Maleszka, S., Foret, R., and R. Maleszka. 2008. Nutritional control of reproductive status in honeybees via DNA methylation. Science 319: 1827-1830.

Künzl C, Kaiser S, Meier E, and N. Sachser. 2003. Is a wild mammal kept and reared in captivity still a wild mammal? Hormones and Behaviour 43: 187-196.

Kurzban R., Burton-Chellew M.N., and S.A. West. 2015. The Evolution of Altruism in Humans. Annual Review of Psychology 66: 575–599.

Kutsukake N., and T.H. Clutton-Brock. 2006. Aggression and submission reflect reproductive conflict between females in cooperatively breeding meerkats Suricata suricatta. Behavioral Ecology and Sociobiology 59(4): 541–548.

Lacey E.A., and P.W. Sherman. 1991. Social organisation of naked mole-rats: Evidence of division of labor. In The Biology of the Naked Mole-Rat (eds Sherman, P.W., Jarvis, J.U.M., and R.D. Alexander), pp. 275-336. Princeton University Press, Princeton.

Lacey E.A., and P.W. Sherman. 1997. Cooperative breeding in naked mole-rats: implications for vertebrate and invertebrate sociality. In Cooperative Breeding in Mammals (eds Solomon N.G. and J.A. French), pp. 267-334. Cambridge Univ Press, Cambridge, UK.

156

REFERENCES

Lang S.D.J., and D.R. Farine 2017. A multidimensional framework for studying social predation strategies. Nature Ecology and Evolution 1: 230-239.

Lardy S., Cohas A., Desouhant E., Tafani M. and D. Allainé, D. 2012. Paternity and dominance loss in male breeders: the cost of helpers in a cooperatively breeding mammal. PloS one 7(1): e29508.

Lehmann L., and L. Keller. 2006. The evolution of cooperation and altruism - A general framework and a classification of models. Journal of Evolutionary Biology 19(5): 1365–1376.

Locatelli V., and V.E. Bianchi. 2014. Effect of GH/IGF-1 on bone metabolism and osteoporosis. International Journal of Endocrinology 2014: 235060.

López-Otín C., Blasco M.A., Partridge L., Serrano M., and G. Kroemer. 2013. The hallmarks of aging. Cell 153(6): 1194-1217.

Lukas D., and T. Clutton-Brock. 2012a. Cooperative breeding and monogamy in mammalian societies. Proceedings of the Royal Society of London B 279(1736): 2151–2156.

Lukas D., and T. Clutton-Brock. 2012b. Life histories and the evolution of cooperative breeding in mammals. Proceedings of the Royal Society of London B 279(1744): 4065–70.

Lukas D., and T. Clutton-Brock. 2017. Climate and the distribution of cooperative breeding in mammals. Royal Society Open Science 4: 160897.

Lukas D., and T. Clutton-Brock. 2018. Social complexity and kinship in animal societies. Ecology Letters 21(8): 1129- 1134.

Lummaa V., and T.H. Clutton-Brock. 2002. Early development, survival and reproduction in humans. Trends in Ecology and Evolution 17(3): 141–147.

Lutermann H., Young A.J. and N.C. Bennett. 2013. Reproductive status and testosterone among females in cooperative mole-rat societies. General and comparative endocrinology 187: 60–65.

Macleod K.J., Nielsen J.F., and T.H. Clutton-Brock. 2013. Factors predicting the frequency, likelihood and duration of allonursing in cooperative meerkat. Animal Behaviour 86: 1059-1067.

Maswanganye, K.A., Bennett, N.C., Brinders, J. and R. Cooney. 1999. Oligospermia and azoospermia in nonreproductive male Damaraland mole-rats Cryptomys damarensis (Rodentia: Bathyergidae). Journal of Zoology 248: 411-418.

Maynard Smith J.M. and E. Szathmáry. 1995. The major transitions in evolution. Oxford Univ. Press, Oxford, UK.

McNamara, J.M., and A.I. Houston. 1996. State-dependent life histories. Nature 380(6571): 215-221.

Metcalfe N.B. and P. Monaghan. 2001. Compensation for a bad start: grow now, pay later?’, Trends in Ecology and Evolution 16(5): 254–260.

Michener C. 1969. Comparative Social Behavior of Bees. Annual Review of Entomology 14: 299–342.

Minias, P., and P. Podlaszczuk, P. 2017. Longevity is associated with relative brain size in birds. Ecology and Evolution 7(10): 3558–3566.

Mittermeier, R.A., Rylands, A.B., and D.E. Wilson. 2018. Handbook of the Mammals of the World. Lynx Edicions, Barcelona, Spain.

Molteno A.J., and N.C. Bennett. 2000. Anovulation in non-reproductive female Damaraland mole-rats (Cryptomys damarensis). Journal of Reproduction and Fertility 119: 35-41.

Molteno A.J., Kallo I., Bennett N.C., King J.A. and C.W. Coen. 2004. A neuroanatomical and neuroendocrinological study into the relationship between social status and the GnRH system in cooperatively breeding female Damaraland mole-rats, Cryptomys damarensis. Reproduction 127: 13-21.

Montiglio P.O., Ferrari C., and D. Réale 2013. Social niche specialization under constraints: personality, social interactions and environmental heterogeneity. Philosophical Transactions of the Royal Society of London B, Biological Sciences 368: 20120343.

Mooney S.J., Filice D.C., Douglas N.R., and M.M. Holmes. 2015. Task specialisation and task switching in eusocial mammals. Animal Behaviour 109: 227-233.

157

REFERENCES

Münch D., Amdam G.V., and F. Wolschin. 2008. Ageing in a eusocial insect: Molecular and physiological characteristics of life span plasticity in the honey bee. Functional Ecology 22(3): 407–421.

Mundry, R. 2014. Statistical issues and assumptions of phylogenetic generalized least squares. In Modern comparative phylogenetic methods and their application in evolutionary biology (eds L.Z. Garamszegi ), pp. 131–153. Springer, Heidelberg, Germany.

Nedbal M.A., Allard M.W., and R.L. Honeycutt. 1994. Molecular systematics of the hystricognath rodents: evidence from the mitochondrial 12S rRNA gene. Molecular Phylogenetics and Evolution 3: 206-220.

Nelson-Flower M.J., Hockey P.A.R., O’Ryan, C., Raihani, N.J., Du Plessis, M.A., and A.R. Ridley. 2011. Monogamous dominant pairs monopolize reproduction in the cooperatively breeding pied babbler. Behavioral Ecology 22(3): 559–565.

Noirot C., and J.M. Pasteels. 1987. Ontogenetic development and the evolution of the worker caste in termites. Experientia 43(8): 851-860.

Noirot, C. 1990. Sexual castes and reproductive strategies in termites. In Social Insects. An Evolutionary Approach to Castes and Reproduction (eds. W. Engels), pp. 5-35. Springer-Verlag.

Novikov E., Kondratyuk E., Petrovski D., Titova T., Zadubrovskaya I., Zadubrovskiy P., and M. Moshkin. 2015. Reproduction, aging and mortality rate in social subterranean mole voles (Ellobius talpinus Pall.). Biogerontology 16(6): 723–732.

Novikov E., Rondatyuk E., and H. Burda. 2015. Age-related increases of urine cortisol in non-breeding individuals from Fukomys anselli (Rodentia, Bathyergidae) from a laboratory colony. Zoologicheskiĭ Zhurnal 94: 119–124. [In Russian].

Nowak M.A. 2006. Five rules for the evolution of cooperation, Science 314: 1560–1563.

Nowak, M. A., and K. Sigmund. 2005. Evolution of indirect reciprocity. Nature 437(7063): 1291–1298.

Nowak, R.M. 1999. Walker’s mammals of the world. John Hopkins University Press, Baltimore, USA.

Nussey D.H., Kruuk L.E.B., Donald A., Fowlie M., and T.H. Clutton-Brock. 2006. The rate of senescence in maternal performance increases with early-life fecundity in red deer. Ecology Letters, 9(12): 1342–50.

O’Riain M.J. and J.U.M. Jarvis. 1998. The dynamics of growth in naked mole-rats: the effects of litter order and changes in social structure. Journal of Zoology 246: 49–60.

O’Riain M.J., Jarvis J.U.M., Alexander R., Buffenstein R., and C. Peeters. 2000. Morphological castes in a vertebrate. Proceedings of the National Academy of Sciences USA 97: 13194-13197.

O’Riain M.J., Jarvis, J.U.M., and C.G. Faulkes. 1996. A dispersive morph in the naked mole-rat. Nature 380: 619– 621.

O’Sullivan R. J., and J. Karlseder. 2010. Telomeres: Protecting chromosomes against genome instability. Nature Reviews Molecular Cell Biology 11(3): 171–181.

Ogilby W. 1838. On a collection of Mammalia procured by Captain Alexander during his journey into the country of the Damaras. Proceedings of the Zoological Society of London 1838, 5-15.

Olovnikov A. M. 1996. Telomeres, telomerase, and aging: Origin of the theory. Experimental Gerontology 31(4): 443– 448.

Olsson M., Wapstra E., and C.R. Friesen. 2017. Evolutionary ecology of telomeres: a review. Annals of the New York Academy of Sciences 350: 569.

Ono K.A. and D.J. Boness. 1996. Sexual Dimorphism in Sea Lion Pups: Differential Maternal Investment, or Sex- Specific Differences in Energy Allocation?. Behavioral Ecology and Sociobiology 38: 31–41.

Oosthuizen M.K., and N.C. Bennett. 2015. The effect of ambient temperature on locomotor activity patterns in reproductive and non-reproductive female Damaraland mole-rats. Journal of Zoology, 297(1): 1–8.

Orme, D., Freckleton, R., Thomas, G., Petzoldt, T., Fritz, S., Isaac, N., and W. Pearse. 2018. caper: Comparative Analyses of Phylogenetics and evolution in R. R package version 1.0.1.

158

REFERENCES

Oster G.F., and E.O. Wilson. 1978. Caste and Ecology in the Social Insects. Princeton University Press, Princeton, New Jersey.

Owens I.P.F., and P.M. Bennett. 1995. Ancient ecological diversification explains life-history variation among living birds. Proceedings of the Royal Society of London B 261(1361): 227–232.

Ozgul, A., Childs, D.Z., Oli, M.K., Armitage, K.B., Blumstein, D.T., Olson, L.E., Tuljapurkar, S. and T. Coulson. 2010. Coupled dynamics of body mass and population growth in response to environmental change, Nature 466(7305): 482–485.

Pauliny A., Larsson K., and D. Blomqvist. 2012. Telomere dynamics in a long-lived bird, the barnacle goose. BMC Evolutionary Biology 12(1): 257.

Peeters C., and F. Ito. 2001. Colony dispersal and the evolution of queen morphology in social hymenopetera. Annual Reviews in Entomology 46: 601-630.

Pfaffl M.W. 2001. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acid Research 29(9): 16–21.

Pinheiro, J. and D. Bates. 2000. Mixed-effects models in S and S-Plus. Springer, New York, USA.

Pinto M., Jepsen K.J., Terranova C.J., and R. Buffenstein. 2010. Lack of sexual dimorphism in femora of the eusocial and hypogonadic naked mole-rat: a novel animal model for the study of delayed puberty on the skeletal system.Bone 46: 112-120.

Pitcher T., Buffenstein R, Keegan J.D., Modley G.P. and S. Yahav. 1992. Dietary calcium content, calcium balance and mode of uptake in a subterranean mammal, the damara mole-rat. Journal of Nutrition 122: 108-114.

Plot V., Criscuolo F., Zahn S., and J-Y. Georges. 2012. Telomeres, age and reproduction in a long-lived reptile. PloS One 7(7): e40855.

Polanowski A.M., Robbins J., Chandler D., and S.N. Jarman. 2014. Epigenetic estimation of age in humpback whales. Molecular Ecology Resources 14: 967-987.

Popova N.K., and E.V. Naumenko. 1972. Dominance Relations and the Pituitary Adrenal System in Rats. Animal Behaviour 20(1): 108–111.

Pusey A.E. and K. Schroepfer-Walker. 2013. Female competition in chimpanzees. Philosophical Transactions of the Royal Society of London B, Biological Sciences 368(1631): 20130077.

Qing H., Ardeshipour L., Pajevic P.D., Dusevich V., Jahn K., Kato S., Wysolmerski J., and L.F. Bonewald. 2012. Demonstration of osteocytic perilacunar/canicular remodelling in mice during lactation. Journal of Bone and Mineral Research 27: 1018-1029.

R Core Team. 2018. R: A language and environment for statistical computing. R Foundation for statistical computing. Vienna, Austria. URL https://www.R-project.com

Reeve H.K., Westneat D.F., Noon W.A., Sherman P.W. and C.F. Aquadro. 1990. DNA “fingerprinting” reveals high levels of inbreeding in colonies of the eusocial naked mole-rat. Proceedings of the National Academy of Sciences USA 87: 2496-2500.

Riccio A.P., and B.D. Goldman. 2000. Circadian rhythms of locomotor activity in naked mole-rats (Heterocephalus glaber). Physiology and Behavior 71: 1-13.

Rickard C.A., and N.C. Bennett. 1997 Recrudescence of sexual activity in a reproductively quiescent colony of Damaraland mole-rat (Cryptomys damarensis), by the introduction of an unfamiliar and genetically unrelated male- a case of incest avoidance in ‘queenless’ colonies. Journal of Zoology 241: 185 – 202.

Ricklefs R.E. 1974. The evolution of cooperative breeding in birds. Ibis 117: 531–534.

Ricklefs R.E. 2010. Embryo growth rates in birds and mammals. Functional Ecology 24(3): 588–596.

Roff D. A. 1992. The Evolution of Life Histories: Theory and Analysis. Chapman and Hall, New York, USA.

Roisin Y. 1996. Castes in humivorous and litter-dwelling neotropical nasute termites (Isoptera, Termitidae). Insectes Sociaux 43, 375-389.

159

REFERENCES

Roisin Y., and K. Korb. 2010. Social organisation and the status of workers in termites. In Biology of Termites: A Modern Synthesis (eds Bignell, D.E., Roisin, Y. and N. Lo), pp. 133-164. Springer Verlag, Heidelberg, Germany.

Ruby J.G., Smith M., and R. Buffenstein. 2018. Naked mole-rat mortality rates defy gompertzian laws by not increasing with age. eLife 7: 1–18.

Russell A. F. 2004. Mammals: comparisons and contrasts. In Ecology and evolution of cooperative breeding in birds (eds Koenig, W.D., and J. L. Dickinson), pp. 210–227. Cambridge University Press, Cambridge.

Russell A.F., Brotherton P., McIlrath G., Sharpe L., and T. Clutton-Brock. 2003. Breeding success in cooperative meerkats: effects of helper number and maternal state. Behavioral Ecology 14(4): 486–492.

Russell A.F., Carlson A.A., McIlrath G.M., Jordan N.R. and T. Clutton-Brock. 2004. Adaptive size modification by dominant female meerkats. Evolution 58(7): 1600–7.

Russell A.F., Clutton-Brock T.H., Brotherton P.N.M., Sharpe L.L., McIlrath G.M., Dalerum F.D., Cameron E.Z. and J.A. Barnard. 2002. Factors affecting pup growth and survival in co-operatively breeding meerkats Suricata suricatta. Journal of Animal Ecology 71(4): 700–709.

Saltzman W., Digby L.J., and D.H. Abbott. 2009. Reproductive skew in female common marmosets: what can proximate mechanisms tell us about ultimate causes? Proceedings of the Royal Society of London B 276(1656): 389–399.

Sameshima S.Y., Miura T., and T. Matsmumoto. 2004. Wing disc development during caste differentiations in the ant Pheidole megacephala (Hymenoptera: Formicidae). Evolution and Development 6(5): 336-341.

Sapolsky R.M., Alberts S.C., and J. Altmann. 1997. Hypercortisolism associated with social subordinance or social isolation among wild baboons. Archives of General Psychiatry 54: 1137–1143.

Scantlebury M., Speakman J.R., Oosthuizen M.K., Roper T.J., and N.C. Bennett. 2006. Energetics reveals physiologically distinct castes in a eusocial mammal. Nature 440: 795–7.

Schielke C.K.M., Begall S., and H. Burda. 2012. Reproductive state does not influence activity budgets of eusocial Ansell’s mole-rats, Fukomys anselli (Rodentia, Bathyergidae): A study of locomotor activity by means of RFID. Mammalian Biology 77(1): 1–5.

Schielke C.K.M., Burda H., Henning Y., Okrouhlík, J., and S. Begall. 2017. Higher resting metabolic rate in long- lived breeding Ansell’s mole-rats (Fukomys anselli). Frontiers in Zoology 14: 1–7.

Schmidt C.M., Blount J.D., and N.C. Bennett. 2014. Reproduction is associated with a tissue-dependent reduction of oxidative stress in eusocial female Damaraland mole-rats (Fukomys damarensis). PloS One 9(7): e103286.

Schmidt C.M., Jarvis J.U.M. and N.C. Bennett. 2013. The long-lived queen: reproduction and longevity in female eusocial Damraland mole-rats (Fukomys damarensis). Journal of African Zoology 48: 193-196.

Schrempf A., Heinze J., and S. Cremer. 2005. Sexual cooperation: mating increases longevity in ant queens. Current Biology 15: 267–270.

Schutz H., Donovan E.D., and J.P. Hayes. 2009. Effects of parity on pelvic size and shape dimorphism in Mus. Journal of Morphology 270: 834-842.

Schwander, T., Lo, N., Beekman, M., Oldroyd, B.P., and L. Keller. 2010. Nature versus nurture in social insect caste differentiation. Trends in Ecology and Evolution 25: 275 – 282.

Seeley T.D. 1982. Adaptive significance of age polyethism schedules in honeybee colonies. Behavioral Ecology and Sociobiology 11: 287-293.

Selman C., Blount J. D., Nussey D. H., and J.R. Speakman. 2012. Oxidative damage, ageing, and life-history evolution: Where now? Trends in Ecology and Evolution 27(10): 570–577.

Settepani, V., Grinsted, L., Granfeld, J., Jensen, J.L., and T. Bilde. 2012. Task specialisation in two social spiders, Stegodyphus sarasinorum (Eresidae) and Anelosimus eximius (Theridiidae). Journal of Evolutionary Biology 26: 51-62.

Sharp S.P., and T.H. Clutton-Brock. 2011. Reluctant challengers: why do subordinate female meerkats rarely displace their dominant mothers? Behavioral Ecology 22(6): 1337–1343.

Sherman P., and J.U.M. Jarvis. 2002. Extraordinary life spans of naked mole rats (Heterocephalus glaber). Journal of Zoology 258: 307–311.

160

REFERENCES

Sherman P.W., Braude S., and J.U.M. Jarvis. 1999. Litter Sizes and Mammary Numbers of Naked Mole-Rats: Breaking the One-Half Rule. Journal of Mammalogy 80(3): 720–733.

Sherman P.W., Braude, S., and J.U.M. Jarvis. 1999. Litter sizes and mammary numbers of naked mole-rats: breaking the one-half rule. Journal of Mammalogy 80(3): 720-733.

Sherman P.W., Lacey E.A., Reeve, H.K., and L. Keller. 1995. The eusociality continuum. Behavioral Ecology 6: 102– 108.

Shibao H., Kutuskake M., Matsusyama S., Fukatsu T., and M. Shimada. 2010. Mechanisms regulating caste differentiation in an aphid social system. Communicative and Integrative Biology 3: 1-5.

Simons M.J.P. 2015. Questioning causal involvement of telomeres in aging. Ageing Research Reviews 24: 191–196.

Šklíba, J, Lövy, M., Burda, H. and R. Šumbera. 2016. Variability of space-use patterns in a free-living eusocial rodent, Ansell’s mole-rat indicates age-based rather than caste polyethism. Scientific Reports 6: 37497.

Smyth K.N., Greene L.K., Clutton-Brock T.H., and C.M. Drea. 2016. Androgens predict parasitism in female meerkats: a new perspective on a classic trade-off. Biology Letters 12: 20160660.

Smyth, K.N., and C.M. Drea. 2016. Patterns of parasitism in the cooperatively breeding meerkat: A cost of dominance for females. Behavioral Ecology 27(1): 148–157.

Snowdon C.T. 1996. Infant care in cooperatively breeding species. Advances in the Study of Behaviour 25: 643-689.

Snowdon C.T. and J.J. Pickhard. 1999. Family Feuds: Severe Aggression antong Cooperatively Breeding Cotton- Top Tamarins. International Journal of Primatology 20: 651–663.

Snyman P.C., Jackson C.R., and N.C. Bennett. 2006. Do dispersing non-reproductive Damaraland mole-rats Cryptomys damarensis (Rodentia: Bathyergidae) exhibit spontaneous or induced ovulation? Physiology and Behavior 87: 88-94.

Sofaer H.R., Chapman P.L., Sillett T.S. and C.K. Ghalambor. 2013. Advantages of nonlinear mixed models for fitting avian growth curves. Journal of Avian Biology 44: 469–478.

Solomon N.G. and J.A. French (eds). 1997. Cooperative breeding in mammals. Cambridge University Press, Cambridge, UK.

Sparkman A.M., Adams J., Beyer A., Steury T. D., Waits L., and D.L. Murray. 2011. Helper effects on pup lifetime fitness in the cooperatively breeding red wolf (Canis rufus). Proceedings of the Royal Society of London B 278(1710): 1381–1389.

Speakman J.R. 2005. Body size, energy metabolism and lifespan. Journal of Experimental Biology 208(9): 1717–1730.

Speakman J.R. 2008. The physiological costs of reproduction in small mammals. Philosophical Transactions of the Royal Society of London B, Biological Sciences 363(1490): 375-398.

Speakman J.R., and M. Garratt. 2014. Oxidative stress as a cost of reproduction: Beyond the simplistic trade-off model. BioEssays 36: 93–106.

Specker B., and T. Binkley. 2005. High parity is associated with increased bone size and strength. Osteoporosis International 16: 1969-1974.

Spinks A.C., Jarvis J.U.M, and N.C. Bennett. 2000. Comparative patterns of philopatry and dispersal in two common mole-rat populations: implications for the evolution of mole-rat sociality. Journal of Animal Ecology 69: 224–234.

Spong G.F., Hodge S.J., Young A.J., and T.H. Clutton-Brock. 2008. Factors affecting the reproductive success of dominant male meerkats. Molecular Ecology 17(9): 2287–99.

Stacey P.B., and J.D. Ligon. 1987. Territory quality and dispersal options in the acorn woodpecker, and a challenge to the habitat-saturation model of cooperative breeding. American Naturalist 130: 654–676.

Stander P.E. 1992. Cooperative hunting in lions: the role of the individual. Behavioral Ecology and Sociobiology 29: 445-454.

Starck J. and R.E. Ricklefs. 1998. Avian growth and development: evolution within the altricial-precocial spectrum. Oxford Univ. Press, Oxford, UK.

161

REFERENCES

Stearns, S. 1992. The evolution of life histories. Oxford University Press, New York.

Syed F., and S. Khosla. 2005. Mechanisms of sex steroid effects on bone. Biochemical and Biophysical Research Communications 328: 699-696.

Tanaka H., Frommen J.G., Engqvist L and M. Kohda 2017. Task-dependent workload adjustment of female breeders in a cooperatively breeding fish. Behavioral Ecology 29: 221-229.

Thomas M., and M.A. Elgar. 2003. Colony size affects division of labour in the ponerine ant Rhytidoptera metallica. Naturwissenschaften 90: 88-92.

Thompson M.J., von Holdt B., Horvath S., and M. Pellegrini. 2017. An epigenetic clock for dogs and wolves. Aging 9: 1055-1058.

Thompson S.D. 1992. Gestation and lactation in mammals: basal metabolic rates and the limits of energy use. In Mammalian Energetics: Inter-disciplinary Views of Metabolism and Reproduction (eds. T.E. Tomasi and T.H. Horton), pp. 213–259. Cornell University Press, Ithaca, New York.

Thornton A. Mcauliffe K., Dall S.R.X., Fernandez-Duque E., Garber P.A., and A.J. Young, 2016. Fundamental problems with the cooperative breeding hypothesis. A reply to Burkart and van Schaik. Journal of Zoology 299(2): 84–88.

Torrents-Ticó M., Bennett N.C., Jarvis J.U.M. and M. Zöttl. In Review. ‘Sex-differences in timing and context of dispersal in Damaraland mole-rats (Fukomys damarensis).

Tricola G.M., Simons M.J.P., Atema E., Boughton R.K., Brown J.L., Dearborn D.C., Divoky D., Eimes J.A., Huntington C.E., Kitaysky A.S., et al. 2018. The rate of telomere loss is related to maximum lifespan in birds. Philosophical Transacations of the Royal Society of London B 373(1741): 20160445.

Tung Ho, L.S., and C. Ané. 2014. A linear-time algorithm for gaussian and non-gaussian trait evolution models. Systematic Biology 63(3): 397–408.

Utami S., Goossens M., Bruford M.W., de Ruiter J.R., and J.A.R.A.M. van Hooff. 2002. Male bimaturism and reproductive success in Sumatran orang-utans. Behavioral Ecology 5: 643-542.

Van Daele P.A.A.G., Faulkes C.G., Verheyen E., and D. Adriaens. 2007. Cytochrome b sequence analysis reveals differential molecular evolution in the African mole-rats of the chromosomally hyperdiverse genus Fukomys (Batherygidae, Rodentia) from the Zambezian region. Molecular Phylogenetics and Evolution 45: 142–147. van de Pol M., and J. Wright. 2009. A simple method for distinguishing within- versus between-subject effects using mixed models. Animal Behaviour 77(3): 753–758. van de Pol M., and S. Verhulst. 2006. Age-dependent traits: a new statistical model to separate within- and between- individual effects. The American Naturalist 167(5): 766–73.

Vinicius L. and H.S. Mumby. 2013. Comparative analysis of animal growth: a primate continuum revealed by a new dimensionless growth rate coefficient. Evolution 67(5): 1485–1492.

Volny V., and D.M. Gordon. 2002. Genetic basis for queen-worker polymorphism in a social insect. Proceedings of the National Academy of Sciences USA 99: 6108-6111.

Walton A., and Toth, A.L. 2016. Variation in individual worker honey bee behaviour shows hallmarks of personality. Behavioral Ecology and Sociobiology 70: 999-1010.

Wcislo W.T. 1997. Are behavioral classifications blinders to studying natural variation? In The evolution and social behavior in insects and arachnids (eds J.C. Choe and B.J. Crespi). Cambridge Univ. Press, Cambridge, UK.

Wiklund P.K., Xu L., Wang, Q., Mikkola T., Lyytikäinen A., Völgyi E., Munukka E., Cheng S.M., Alen M., Keinänen- Kiukaanniemi S., and S. Cheng. 2012. Lactation is associated with greater maternal bone size and bone strength later in life. Osteoporosis International 23: 1939-1945.

Wilkinson I.S., and R.J. van Aarde. 2001. Investment in sons and daughters by southern elephant seals, Mirounga leonina, at Marion Island. Marine Mammal Science 17: 873–887.

Williams G.C. 1957. Pleiotropy, Natural Selection, and the Evolution of Senescence. Evolution 11(4): 398–411.

Williams S.A., and M.R. Shattuck. 2015. Ecology, longevity and naked mole-rats: confounding effects of sociality? Proceedings of the Royal Society of London B 282(1802): 7–9.

162

REFERENCES

Wilson E.O. 1971. The Insect Societies. Harvard Univ. Press, Cambridge, USA.

Wilson E.O. 1975. Sociobiology. Belknap, Cambridge, USA.

Wong M.Y.L., Munday P.L., Buston P.M., and G.P. Jones. 2008. Fasting or feasting in a fish social hierarchy. Current Biology 18: 372-373.

Wood-Gush D.G.M. 1971. The behaviour of the Domestic Fowl. Heinemann, London, UK.

Wright C.M., Holbrook C.T., and J.N. Pruitt. 2014. Animal personalities aligns task specialization and task proficiency in a spider society. Proceedings of the National Academy of Sciences USA 111(26): 9533-9537.

Young A.J. 2018. The role of telomeres in the mechanisms and evolution of life-history trade-offs and ageing. Philosophical Transactions of the Royal Society of London B, Biological Sciences 373(1741): 20160452.

Young A.J., and N.C. Bennett. 2010. Morphological divergence of breeders and helpers in wild Damaraland mole- rat societies. Evolution 64: 3190-3197.

Young A.J., and N.C. Bennett. 2010. Physiological suppression eases in Damaraland mole-rat societies when ecological constraints on dispersal are relaxed. Hormones and Behavior 57: 177-183.

Young A.J., and N.C. Bennett. 2013. Intra-sexual selection in cooperative mammals and birds: why are females not bigger and better armed? Philosophical Transactions of the Royal Society of London B, Biological Sciences 368: 20130075.

Young A.J., Carlson A.A, Monfort S.L., Russell A.F., Bennett N.C., and T. Clutton-Brock. 2006. Stress and the suppression of subordinate reproduction in cooperatively breeding meerkats. Proceedings of the National Academy of Sciences USA 103(32): 12005–12010.

Young A.J., Jarvis J.U.M., Barnaville J. and N.C. Bennett. 2015. Workforce Effects and the Evolution of Complex Sociality in Wild Damaraland Mole Rats. The American Naturalist 186(2): 302-311.

Zöttl M., Thorley J., Gaynor D., Bennett N.C., and T. Clutton-Brock. 2016b. Variation in growth of Damaraland mole-rats is explained by competition rather than functional specialization for different tasks. Biology Letters 12: 20160820.

Zöttl M., Vuilloud, P., Mendonça, R., Torrents-Ticó, M., Gaynor, D., Mitchell, A., and T. Clutton-Brock. 2016a. Differences in cooperative behaviour among Damaraland mole-rats are consequences of an age-related polyethism. Proceedings of the National Academy of Sciences USA 113(37): 10382-10387.

Zöttl M., Vullioud P., Goddard K., Torrents Ticó M., Gaynor D., Bennett N.C. and T. Clutton-Brock. 2018. Allo- parental care in Damaraland mole-rats is female biased and age dependent, though indpendent of testosterone levels. Physiology and Behavior 193: 149-153.

Zullinger E.M., Ricklefs R.E., Redford K.H. and G.M. Mace. 1984. Fitting Sigmoidal Equations to Mammalian Growth Curves. Journal of Mammalogy 65(4): 607–636.

163

APPENDIX A

SUPPORTING INFORMATION FOR CHAPTER 3: No Evidence of Task Specialisation among Damaraland mole-rat helpers

Table 1. Damaraland mole-rat ethogram.

Response Category* Sub-categories Description of behaviour

Gnaw Gnawing on the tunnel walls with teeth. Moving through the tunnel system, but not Locomotion engaged in obvious work Characteristic, repetitive up and down Pump movement of the rear body part of the Active non-helping individual. behaviours Behaviours not assignable to the above Other categories. Self-groom Self-directed grooming. Sniff Investigating something with the nose. Sparring with incisors, tail pulling, biting, Social Interaction copulation, dominance interaction. Eat Eat Eating food. Transporting food either by pushing it along Food Carry (helping) Food Carry the tunnels (forward) or by dragging it while moving backwards. Individual engaged with paper. Either Nest Building dragging in the direction of the nest, or chewing Nesting Material (helping) it into small pieces or trying to pull it out of a certain location. Resting in body contact with other individuals Huddling Rest in the tunnel (in sight) Rest Individuals resting in the nest (out of sight) Using extra-buccal teeth and front paws to dig Dig in the sand or blockage of paper. Pushing sand into tunnel gaps or other Kick locations with the hind legs or with the nose. Work (helping) Often used to block feeders or tunnel gaps Sweeping sand with hindpaws through the Sweep tunnel system, often to the waste box. Locomotion Between Moving between episodes of the above Work behaviours *Pup carrying- the retrieval of pups that have left the nest - has been excluded from the ethogram as it was observed on very few occasions across all the scans in the dataset.

164

APPENDIX A

Table 2. Correlations of random effects across the responses in each of tested models. Estimates represent the means from the posterior samples (standard deviations in the parentheses). Parameters in bold indicate estimates where the 95% credible intervals do not span zero.

FEMALES, random effects correlations

Model 1, individual-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.49 (0.12) 0.34 (0.12) 0.28 (0.14) 0.60 (0.09) Eat - 0.10 (0.15) 0.07 (0.17) 0.14 (0.12) Food Carry - - -0.13 (0.15) 0.16 (0.11) Nest Building - - - 0.31 (0.12) Work - - - -

Model 2, individual-level Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.54 (0.11) 0.43 (0.12) 0.19 (0.14) 0.63 (0.08) Eat - 0.07 (0.15) 0.17 (0.17) 0.28 (0.12) Food Carry - - -0.14 (0.15) 0.22 (0.30) Nest Building - - - 0.58 (0.33) Work - - - -

Model 3, individual-level Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.76 (0.12) 0.44 (0.17) 0.46 (0.20) 0.64 (0.10) Eat - 0.04 (0.28) 0.22 (0.29) 0.12 (0.23) Food Carry - - -0.21 (0.25) 0.25 (0.17) Nest Building - - - 0.29 (0.22) Work - - - -

, scan-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.86 (0.03) 0.66 (0.05) 0.38 (0.07) 0.84 (0.02) Eat - 0.20 (0.10) 0.07 (0.13) 0.11 (0.07) Food Carry - - 0.08 (0.11) 0.13 (0.05) Nest Building - - - 0.09 (0.05) Work - - - -

165

APPENDIX A

, group-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping -0.02 (0.35) 0.09 (0.34) -0.04 (0.34) 0.04 (0.36) Eat - -0.08 (0.33) 0.04 (0.36) 0.05 (0.34) Food Carry - - -0.01 (0.36) 0.01 (0.36) Nest Building - - - 0.03 (0.35) Work - - - -

, litter-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.02 (0.35) 0.05 (0.36) -0.00 (0.35) 0.05 (0.35) Eat - 0.05 (0.35) 0.06 (0.34) 0.02 (0.34) Food Carry - - 0.01 (0.35) 0.10 (0.35) Nest Building - - - 0.05 (0.36) Work - - - -

MALES, random effects correlations

Model 1, individual-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.62 (0.10) 0.30 (0.12) 0.30 (0.17) 0.68 (0.08) Eat - 0.19 (0.16) -0.31 (0.21) 0.11 (0.12) Food Carry - - -0.04 (0.20) 0.24 (0.11) Nest Building - - - -0.01 (0.16) Work - - - -

Model 2, individual-level Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.55 (0.11) 0.24 (0.13) 0.27 (0.16) 0.61 (0.09) Eat - 0.35 (0.15) -0.31 (0.20) 0.08 (0.13) Food Carry - - 0.08 (0.19) 0.22 (0.12) Nest Building - - - 0.02 (0.16) Work - - - -

166

APPENDIX A

Model 3, individual-level Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.63 (0.16) 0.32 (0.26) 0.07 (0.33) 0.66 (0.15) Eat - 0.24 (0.26) -0.26 (0.33) 0.25 (0.21) Food Carry - - 0.11 (0.33) 0.26 (0.19) Nest Building - - - -0.04 (0.24) Work - - - - , scan-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.83 (0.04) 0.51 (0.06) 0.42 (0.09) 0.80 (0.03) Eat - 0.39 (0.12) 0.18 (0.16) 0.14 (0.08) Food Carry - - -0.08 (0.15) 0.10 (0.08) Nest Building - - - -0.10 (0.08) Work - - - -

, group-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.09 (0.34) -0.02 (0.34) 0.04 (0.36) 0.18 (0.36) Eat - 0.04 (0.35) 0.07 (0.34) -0.16 (0.33) Food Carry - - -0.03 (0.35) 0.04 (0.35) Nest Building - - - 0.07 (0.33) Work - - - -

, litter-level

Active Non- Eat Food Nest Work helping Carry Building Active Non-helping 0.24 (0.36) 0.26 (0.33) 0.24 (0.33) 0.45 (0.36) Eat - 0.22 (0.30) -0.13 (0.33) -0.06 (0.29) Food Carry - - 0.11 (0.30) 0.16 (0.28) Nest Building - - - 0.11 (0.29) Work - - - -

167

APPENDIX A

Table 3. Posterior means (standard deviations in parentheses) of the intercepts in each model. i.e. expression of behaviour relative to resting.

Active Food Nest Non- Eat Work Carry Building Helping Model 1, -1.22 -2.43 -4.43 -5.12 -1.94 Female (0.03) (0.03) (0.09) (0.09) (0.05) Model 2, -1.15 -2.49 -4.34 -5.20 -1.75 Female (0.04) (0.04) (0.10) (0.12) (0.05) Model 3, -1.08 -2.45 -4.37 -5.47 -1.70 Female (0.06) (0.05) (0.14) (0.16) (0.08) Model 1, -1.25 -2.42 -4.34 -5.36 -1.96 Male (0.04) (0.03) (0.10) (0.08) (0.06) Model 2, -1.18 -2.46 -4.22 -5.34 -1.79 Male (0.04) (0.04) (0.11) (0.11) (0.06) Model 3, -1.21 -2.45 -4.42 -5.56 -1.90 Male (0.07) (0.05) (0.15) (0.17) (0.09)

168

APPENDIX A

Table 4. Variance estimates of the random effects in the six models tested. Estimates represent the standard deviations of the random effects (values in parentheses are the standard deviations of these estimates in the posterior distributions).

FEMALE MALE RE Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Individual-Level Active Non- 0.43 (0.63) 0.27 (0.03) 0.20 (0.02) 0.29 (0.03) 0.29 (0.03) 0.19 (0.05) helping Eat 0.24 (0.25) 0.21 (0.02) 0.13 (0.03) 0.19 (0.02) 0.19 (0.02) 0.15 (0.03) Food Carry 0.63 (0.07) 0.62 (0.07) 0.41 (0.09) 0.72 (0.08) 0.62 (0.07) 0.37 (0.11) Nest Building 0.58 (0.09) 0.60 (0.09) 0.34 (0.10) 0.38 (0.08) 0.41 (0.09) 0.20 (0.13) Work 0.38 (0.04) 0.38 (0.04) 0.36 (0.04) 0.40 (0.04) 0.40 (0.04) 0.32 (0.06)

Scan-level Active Non- 0.56 (0.02) 0.53 (0.02) helping Eat 0.34 (0.02) 0.27 (0.02) Food Carry 0.98 (0.07) 0.91 (0.06) Nest Building 1.23 (0.09) 0.85 (0.10) Work 0.75 (0.03) 0.69 (0.03)

Group-level Active Non- 0.11 (0.06) 0.11 (0.06) helping Eat 0.07 (0.04) 0.10 (0.05) Food Carry 0.26 (0.15) 0.15 (0.12) Nest Building 0.17 (0.12) 0.22 (0.16) Work 0.08 (0.06) 0.14 (0.09)

Litter-level Active Non- 0.05 (0.04) 0.12 (0.07) helping Eat 0.10 (0.04) 0.08 (0.05) Food Carry 0.19 (0.12) 0.48 (0.14) Nest Building 0.13 (0.10) 0.33 (0.16) Work 0.06 (0.05) 0.18 (0.10)

169

APPENDIX A

Table 5. Posterior means of fixed effects in model 3 for each sex (standard deviations in parentheses). Parameters in bold indicate estimates where the 95% credible intervals do not span zero.

Model Active Non-helping Eat Food Carry Nest Building Work 3, Females Age -0.06 (0.04) -0.17 (0.03) 0.08 (0.09) -0.28 (0.14) 0.06 (0.05) Age2 -0.15 (0.02) 0.00 (0.02) -0.24 (0.07) -0.44 (0.10) -0.40 (0.30) Age3 0.03 (0.01) 0.01 (0.03) 0.04 (0.03) 0.10 (0.06) 0.09 (0.02) Group Size 0.13 (0.04) 0.09 (0.02) -0.00 (0.09) -0.15 (0.11) 0.10 (0.06) Group Size2 -0.03 (0.03) 0.01 (0.02) -0.04 (0.05) 0.18 (0.06) -0.01 (0.03) Relative Mass 0.06 (0.02) 0.10 (0.02) 0.19 (0.06) -0.06 (0.08) 0.04 (0.03) Pups Present (Y) -0.07 (0.08) -0.03 (0.06) -0.02 (0.18) 0.10 (0.25) 0.01 (0.11)

3, Males Age 0.01 (0.03) -0.15 (0.03) 0.16 (0.09) -0.20 (0.13) 0.16 (0.05) Age2 -0.08 (0.02) 0.01 (0.02) -0.20 (0.08) -0.28 (0.10) -0.30 (0.03) Age3 0.02 (0.01) -0.01 (0.01) 0.04 (0.03) 0.09 (0.05) 0.09 (0.01) Group Size 0.05 (0.05) -0.01 (0.03) -0.25 (0.09) -0.12 (0.11) -0.03 (0.06) Group Size2 0.03 (0.03) 0.05 (0.02) 0.02 (0.06) 0.18 (0.07) 0.06 (0.04) Relative Mass 0.09 (0.02) 0.04 (0.02) 0.17 (0.07) 0.03 (0.08) 0.09 (0.03) Pups Present (Y) -0.18 (0.09) -0.15 (0.06) -0.08 (0.19) -0.17 (0.24) 0.03 (0.12)

170

APPENDIX A

a b

Figure 1. Log-log plots of body mass ~ age, from which residual mass was extracted, a) males b) females. The line represents the slope from a simple linear regression, but the residual points for each weight measure are taken from linear mixed effects models that include a random term for the group; they therefore represent individual mass relative to other same-sex, same-age group mates. To maximise the power of this analysis, all known-age individuals in the lab population were used, regardless of whether they had enough behavioural information to be included in the multinomial behavioural models.

171

APPENDIX A

Figure 2. Distributions of posterior contrasts for pup presence; the contrast from each sample in the posterior, a) females b) males. This method is preferred over the prediction intervals as per the continuous covariates, as the latter incorporate uncertainty from all the parameters in the model 4 and therefore offer less confidence in assessing differences among categorical covariates.

172

173

APPENDIX B

SUPPORTING INFORMATION FOR CHAPTER 4: The shape of growth in Damaraland mole-rats: sex and social effects

a b

Figure 1. Model fits for male growth curves; a) Richards monophasic, b) Gompertz monophasic, c) Richards biphasic. Inset sub-plots are concentrated upon early-life development (< 30g).

174

APPENDIX B

Figure 2. Log-likelihood of biphasic monomolecular growth models that varied the age at which the two phases of the growth curve met, which we take to indicate nutritional independence. Female nutritional independence (a) occurring at 53 days, male independence (b) at 50 days.

175

APPENDIX B

The effect of dataset size restrictions on biphasic growth curves

The weights dataset used throughout the modelling was extensive, having large numbers of records per individual. To examine the influence of the data structure on parameter estimates. To this end, the biphasic model that included both sexes was remodelled with subsets of the initial dataset in which 75% of the weights from every individual’s total set of weights were removed. An animal with 60 weights across ontogeny would therefore yield 15 weights for the first restriction, or 6 weights for the second. The process was iterated 100 times in each case, the female dataset containing 3017 rows (mean 7.72 records/individual), the male dataset 3092 (mean 7.91 records/individual). The resultant curves were extremely similar the curves produced using the whole dataset. The model was parameterised exactly as the model in the main text.

Figure 3. Biphasic growth curves for female (blue, lower) and male (red, upper) Damaraland mole-rats. Each line represents a curve derived from the model being fit to one hundred datasets that were restricted in size.

176

APPENDIX B a

Figure 4. The consistency of group size (a) and the total number of siblings (b) across different stages of development. Inset correlation coefficients derived from parametric linear regressions

177

APPENDIX B

Table 1. Model output for biphasic Richard’s growth model in female and male mole-rats.

Aind, Alitter, t0,ind, t0,litter refer to a random term for asymptote and ‘inflexion point’ at the levels of the individual and the litter.

Females

Fixed Random Std. Cor t0,ind- Est SE p value Effect Effect dev. Aind

A 151.29 2.62 -- Aind, 34.73 0.822

AL -6.44 2.08 0.019 Alitter 0.02

AG 2.70 2.10 0.20 t0,ind 9.10

k 0.013 0.0003 -- t0,litter 5.77

kl -0.00014 0.00007 0.052

kg 0.00009 0.00007 0.17

K2 0.00458 0.00007 --

K2L 0.00031 0.00003 < 0.001

K2G -0.0008 0.00003 < 0.001

t0 88.58 1.04 -- m 1.09 0.03 --

Males Fixed Random Std. Cor t0,ind- Est SE p value Effect Effect dev. Aind

A 203.39 3.97 -- Aind 39.77 0.890

AL 6.05 2.90 0.037 Alitter 25.27

AG 3.27 2.84 0.25 t0,ind 10.63

k 0.0153 0.0004 -- t0,litter 1.81

kl 0.00061 0.00008 < 0.001

kg -0.00010 0.00008 0.20

k2 0.00488 0.00007 --

k2L -0.00018 0.00003 < 0.001

k2G -0.00061 0.00003 < 0.001

t0 106.20 1.00 -- m 1.32 0.03 --

178

APPENDIX B

Table 2. The influence of data restrictions on model estimates when investigating group and litter size effects on growth, as per the biphasic Richard’s model.

I repeated the resampling exercise described before Appendix B Figure 3, remodelling using 100 subsets of the initial dataset in which 75% of the weights from every individual’s total set of weights were removed. The resultant curves were extremely similar the curves produced using the whole dataset. The model was parameterised exactly as the model in the main text.

Full Dataset Reduced Dataset Female Fixed Median p value Est Median Est (IQR) Effects (IQR) A 151.29 148.68 (148.02 – 149.61) – AL -6.44 -6.35 (-6.78 – -5.60) 0.002 (0.009 – 0.006)

AG 2.70 2.57 (2.07 – 3.40) 0.228 (0.107 – 318) k 0.013 0.0127 (0.0123 – 0.0133) kl -0.00014 0.00015 (0.00007 – 0.00022) 0.185 (0.075 – 0.552)

Kg 0.00009 0.00017 (0.00009 – 0.00023) 0.054 (0.016 – 0.416)

k2 0.00458 0.00458 (0.00449 – 0.00468) –

k2L 0.00031 0.00027 (0.00025 – 0.00032) < 0.001

k2G -0.00080 -0.00067 (-0.00071 – -0.00064) < 0.001

t0 88.58 83.53 (82.97 – 84.15) – m 1.09 1.03 (1.00 – 1.08) –

Male Fixed Median p value Est Median Est (IQR) Effects (IQR) A 203.39 200.16 (199.06 – 201.65) – AL 6.05 6.40 (5.41 – 7.67) 0.52 (0.40 – 0.69)

AG 3.27 1.90 (1.08 – 2.61) 0.052 (0.016 – 0.08) k 0.0153 0.0163 (0.0154 – 0.0172) – kl 0.00061 0.00058 (0.00053 – 0.00065) < 0.001

kg -0.00010 0.00015 (0.00007 – 0.00024) 0.06 (0.01 – 0.15)

k2 0.00488 0.0051 (0.0049 – 0.0052) –

k2L -0.00018 -0.00021 (-0.00025 – -0.00019) p < 0.001

k2G -0.00061 -0.00057 (-0.00060 – -0.00054) p < 0.001

t0 106.20 105.38 (104.63 – 105.96) – m 1.32 1.37 (1.32 – 1.41) –

179

180

APPENDIX C

SUPPORTING INFORMATION FOR CHAPTER 5: Reproduction Triggers Adaptive Increases in Body Size in Female Mole-rats

Table 1. Proportional trait contributions (a) and trait loadings (b) to the first five principal components from a PCA of skeletal measures in captive adult female mole-rats (> 100g)

a) Proportional Contribution

Trait PC1 PC2 PC3 PC4 PC5 Rostrum 3.4 0.42 11.94 0.98 20.6 Ulna 6.99 1.86 2.69 8.86 8.14 L5 Vertebra 15.74 26.35 1.08 12.66 27.97 Pelvic Girdle 10.75 22 18.39 24.72 20.21 Pelvis Length 6.02 0.04 40.93 39.57 11.19 Femur Length 31.43 41.89 16.31 2.12 3.81 Tibia Length 6.54 2.3 1.55 5.45 0.18 Skeletal Body Length 10.41 4.83 1.99 4.31 0.00 Skull Width 8.72 0.3 5.14 1.33 7.88

b) Trait Loadings

Trait PC1 PC2 PC3 PC4 PC5 Rostrum 0.0271 -0.0058 0.0252 0.0065 0.0247 Ulna 0.0389 -0.0123 0.0119 0.0195 0.0153 L5 Vertebra 0.0584 0.0463 -0.0075 0.0232 -0.0285 Pelvic Girdle 0.0483 0.0423 -0.0313 -0.0325 0.0242 Pelvis Length 0.0361 0.0019 0.0466 -0.0411 -0.018 Femur Length 0.0825 -0.0583 -0.0294 -0.0095 -0.0105 Tibia Length 0.0377 -0.0137 0.0091 0.0153 0.0023 Skeletal Body Length 0.0475 0.0198 0.0102 0.0135 0.0005 Skull Width 0.0435 0.0049 0.0166 0.0076 0.0146

181

APPENDIX C

Table 2. Bivariate scaling relationships of skeletal size measures in captive female mole- rats; SBL = Skeletal Body Length. SW = Skull Width. All linear traits were log-transformed. Bold, underlined terms represent slopes that differ significantly at α = 0.05 (*), 0.01 (**), 0.001

(*). Significantly different slopes were determined by the interaction between Trait2 and Class. Significantly different intercepts are taken from the model including the interaction if it was significant, otherwise they derive from a simpler model in which the interaction term was removed. Note that the difference in intercepts for relationships with an interaction are estimated at length zero, so they lie outside the bounds of the data. Difference in intercepts represent an in-group nonbreeder relative to the queen. Note that total head width here refers to the zygomatic arch as measured on the anaesthetised animal with digital callipers (i.e. not from X-rays).

diff Trait1 Trait2 β Queen β nonbreeder diff Slopes Intercept Rostrum Length SBL 0.561 ± 0.161 0.594 ± 0.189 0.033, n.s. 0.013, n.s. Ulna SBL 0.314 ± 0.154 0.774 ± 0.181 0.460, * -2.26, * L5 Vertebra SBL 0.586 ± 0.160 0.912 ± 0.189 0.326, n.s. -0.112, *** Pelvic Girdle Width SBL 0.205 ± 0.238 0.566 ± 0.281 0.362, n.s. -0.076, *** Pelvis Length SBL 0.619 ± 0.234 0.627 ± 0.276 0.008, n.s. 0.018, n.s. Femur Length SBL 0.490 ± 0.334 1.039 ± 0.394 0.549, n.s. 0.017, n.s. Tibia Length SBL 0.449 ± 0.148 0.625 ± 0.174 0.176, n.s. 0.001, n.s. Skull Width SBL 0.510 ± 0.116 0.914 ± 0.137 0.404, ** -1.99, ** Head Width SBL 0.402 ± 0.216 1.281 ± 0.254 0.880, *** -4.31, *** Weight SBL 0.424 ± 0.120 0.951 ± 0.014 0.527, *** -2.60, *** Rostrum Length SW 0.566 ± 0.172 0.557 ± 0.196 -0.009, n.s. 0.012, n.s. Ulna SW 0.730 ± 0.144 0.689 ± 0.163 0.041, n.s. 0.012, n.s. L5 Vertebra SW 0.228 ± 0.193 0.756 ± 0.219 0.530, * -1.89, * Pelvic Girdle Width SW 0.420 ± 0.251 0.548 ± 0.285 0.129, n.s. -0.076, *** Pelvis Length SW 0.381 ± 0.248 0.681 ± 0.282 0.300, n.s. 0.018, n.s. Femur Length SW 0.457 ± 0.354 1.015 ± 0.402 0.559, n.s. 0.017, n.s. Tibia Length SW 0.482 ± 0.155 0.601 ± 0.177 0.119, n.s. 0.001, n.s. Head Width SW 1.257 ± 0.187 1.261 ± 0.212 0.004, n.s 0.056, *** Weight SW 0.749 ± 0.115 0.883 ± 0.131 0.134, n.s. 0.001, n.s.

182

APPENDIX C

Table 3: Full model outputs for linear mixed effects models exploring the influence of body length on three fitness metrics: a) Litter size, Poisson errors) b) Total neonate mass, normal errors c) Individual pup mass, normal errors. Significance of fixed covariates were estimated from likelihood ratio tests comparing models with and without the fixed effect of interest.

Fixed Terms Random Terms Standard LRT Standard Term Estimate Error (χ21) p value Term Variance Deviation a) Litter Size Intercept -1.06 0.92 Mother 0 Total Body Length 0.12 0.05 5.81 0.016 Primiparity (YES) -0.3 0.17 3.57 0.059

b) Total Neonate Mass Intercept -39.52 23.02 Mother 37.7 6.1 Total Body Length 3.83 1.23 8.89 0.003 Residual 82.49 9.1 Primiparity (YES) -5.65 2.52 5.05 0.025

c) Individual Pup 0.81 0.9 Mass Intercept 5.25 3.17 Mother Total Body Length 0.36 0.17 4.12 0.040 Litter 0.84 0.92 Primiparity (Y) 0.45 0.33 1.77 0.183 Residual 1.24 1.11 Litter Size -0.6 0.08 47.76 < 0.001

183

APPENDIX C

a b

Figure 1. Bimonthly change in a) L5 lumbar vertebra and b) skull width of captive females experimentally manipulated to follow different social trajectories. a b

Figure 2. Growth of L4 (a) and L6 (b) vertebra of captive female mole-rats experimentally manipulated to follow different social trajectories.

184

185

APPENDIX D- SPECIES NAMES

APPENDIX D

SPECIES COMMON AND SCIENTIFIC NAMES Kingdom Animalia Phylum Arthropoda Class Arachnida Order Araneae Social spider Anelosimus eximius Social spider Anelosimus studiosus

Class Insecta Order Coleoptera Fruit-tree pinhole borer Xyleborinus saxesnii

Order Hemiptera Aphid Tuberaphis styraci

Order Hymenoptera Big-headed ant Pheidole megacephala Honey bee Apis mellifera Nasute termite Velocitermes barrocoloradensis Panamanian leaf-cutter ant Acromyrmex echinatior Red paper wasp Polistes canadensis

Phylum Chordata Class Actinopterygii Order Cichliformes Princess of Burundi cichlid Neolamprologus pulcher Burton’s mouthbrooding cichlid Haplochromis burtoni

Class Aves Order Anseriforms Barnacle goose Branta leucopsis

Order Bucerotiformes Southern yellow-billed hornbill Tockus leucomelas

Order Passeriformes Fork-tailed drongo Dicrurus adsimilis Noisy miner Manorina melanocephala Southern pied babbler Turdoides bicolor White-winged chough Corcorax melanorhamphos

186

APPENDIX D- SPECIES NAMES

Order Procellariiformes Wandering albatross Diomedea exulans

Order Sphenisciformes Magellanic penguin Spheniscus magellanicus

Class Mammalia Order Artiodactyla Blue wildebeest Connochaetes taurinus Common bottlenose dolphin Tursiops truncatus Gemsbok Oryx gazelle Humpback whale Megaptera novaeangliae Red hartebeest Alcelaphus buselaphus Soay sheep Ovis aries Springbok Antidorcas marsupialis

Order Carnivora Aardwolf Proteles cristata African lion Panthera leo African wild dog Lycaon pictus Banded mongoose Mungos mungos Bat-eared fox Otocyon megalotis Black-backed jackal Canis mesomelas Brown hyena Hyaena brunnea California sea lion Zalophus californianus Cheetah Acinonyx jubatus Dwarf mongoose Helogale parvula Eurasian badger Meles meles Leopard Panthera pardus Meerkat Suricata suricatta Southern elephant seal Mirounga leonina Spotted hyena Crocuta crocutta Wolf Canis lupus

Order Eulipotypla European mole Talpa europaea

Order Perissodactyla Plains zebra Equus quagga

Order Pholidota Ground pangolin Smutsia temminckii

Order Primates

187

APPENDIX D- SPECIES NAMES

Chimpanzee Pan troglodytes Common marmoset Callithrix jacchus Human Homo sapiens Sumatran orang-utan Pongo abelii

Order Rodentia Alpine marmot Marmota marmota Ansell’s mole-rat Fukomys anselli Brown rat Rattus norvegicus Cape dune mole-rat Bathyergus suillus Cape ground squirrel Xerus inauris Cape mole-rat Georychus capensis Cape porcupine Hystrix africaustralis Common mole-rat Cryptomys hottentotus Coruro Spalacopus cyanus Damaraland mole-rat Fukomys damarensis House mouse Mus musculus Mound-building mouse Mus spicilegus Naked mole-rat Heterocephalus glaber Northern mole vole Ellobius talpinus Silvery mole-rat Heliophobius argenteocinereus

Order Tubulidentata Aardvark Orycteropus afer

Class Reptilia Order Testudines Leatherback turtle Dermochelys coriacea

Kingdom Plantae Phylum Angiosperms Class Rosids Order Cucurbitales Gemsbok cucumber Acanthosicyos naudinianus

188

189