Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Ecological influences and the biogeographic distribution of sociality in Anelosimus spiders Purcell, Jessica 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_fall_purcell_jessica.pdf [ 51.31MB ]
Metadata
JSON: 24-1.0067207.json
JSON-LD: 24-1.0067207-ld.json
RDF/XML (Pretty): 24-1.0067207-rdf.xml
RDF/JSON: 24-1.0067207-rdf.json
Turtle: 24-1.0067207-turtle.txt
N-Triples: 24-1.0067207-rdf-ntriples.txt
Original Record: 24-1.0067207-source.json
Full Text
24-1.0067207-fulltext.txt
Citation
24-1.0067207.ris

Full Text

ECOLOGICAL INFLUENCES AND THE BIOGEOGRAPHIC DISTRIBUTION OF SOCIALITY IN ANELOSIMUS SPIDERS by JESSICA PURCELL Hon.B.A Williams College, 2002  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES  (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2009 © Jessica Purcell, 2009  ABSTRACT The puzzle of how complex and costly social behaviours have evolved in so many diverse organisms has challenged many generations of biologists. This thesis focuses on interactions between sociality and ecology. My empirical work investigates South American Anelosimus spiders. This genus provides an ideal system for investigating the ecology of social evolution because the species are easy to manipulate and possess social behaviours ranging from nearly solitary to highly social. Social species cooperate to build communal nests, capture prey, and raise young, and groups may persist for many generations. Most Anelosimus species exhibit subsocial behaviours, in which siblings cooperate for a portion of their life cycle, but disperse each generation prior to sexual maturity. I investigate four distinct questions regarding the role of ecology in spider sociality and more generally. First, I ask whether sociality varies between populations of a social species along an altitudinal gradient. I then experimentally transplant small subsocial groups across this altitudinal gradient to investigate the ecological factors that may contribute to this pattern. Third, I examine how sociality may shape community structure in an area where social and subsocial Anelosimus species coexist. Finally, I explore the co-evolutionary dynamics between different social behaviours and dispersal in an individual-based simulation model. I document an intraspecific gradient of decreasing sociality with increasing elevation within the social spider Anelosimus eximius in Ecuador. Through a transplant experiment, I demonstrate that ecological factors including intense rainfall and predator abundance likely contribute to the absence of small groups or solitary Anelosimus spiders from the lowland tropical rainforest. In one area containing at least five sympatric Anelosimus species, I find that social and subsocial species utilize different local habitats. Within those habitats, cooccurring species show different phenologies and construct nests in different positions on a common plant substrate. My modelling study shows that less costly social traits are less sensitive to selection on dispersal than more costly ones, thus extending previous research emphasizing the interplay between dispersal and costly altruistic behaviours. Overall, this thesis shows that ecological factors can influence the origin and maintenance of sociality, both in current communities and over evolutionary time. ii  TABLE OF CONTENTS Abstract .................................................................................................................... ii Table of contents...................................................................................................... iii List of tables ............................................................................................................ vi List of figures.......................................................................................................... vii Acknowledgements .................................................................................................. ix Co-authorship statement............................................................................................ x CHAPTER 1: General Introduction ...................................................................... 1 1.1 Introduction...................................................................................................... 1 1.2 Geographic patterns in the distribution of social systems in terrestrial arthropods......................................................................................... 1 1.2.1 Methods................................................................................................... 3 1.2.2 Global patterns in the geographic distribution of sociality ........................ 4 1.2.3 Local gradients in sociality ...................................................................... 7 1.2.4 Ecological factors that vary along altitudinal and latitudinal axes ......................................................................................................... 8 1.2.5 Conclusions and future directions .......................................................... 15 1.3 Species coexistence and sociality.................................................................... 17 1.4 Modelling the evolution of sociality and dispersal .......................................... 18 1.5 Background on Anelosimus spiders................................................................. 20 1.6 Themes and hypotheses investigated in the thesis ........................................... 20 1.7 References...................................................................................................... 25 CHAPTER 2: Smaller colonies and more solitary living mark higher elevation populations of a social spider ................................................................ 38 2.1 Introduction.................................................................................................... 38 2.2 Methods ......................................................................................................... 40 2.3 Results ........................................................................................................... 43 2.3.1 Upper versus lower elevations ............................................................... 43 2.3.2 Forest edge versus forest interior............................................................ 43 2.4 Discussion ...................................................................................................... 44 2.5 References...................................................................................................... 55  iii  CHAPTER 3: Gradients of precipitation and ant abundance may contribute to the altitudinal range limit of subsocial spiders: insights from a transplant experiment............................................................................... 59 3.1 Introduction.................................................................................................... 59 3.2 Methods ......................................................................................................... 62 3.2.1 Species description ................................................................................ 62 3.2.2 Habitat description................................................................................. 62 3.2.3 Transplant methods................................................................................ 63 3.2.4 Rain exclosure ....................................................................................... 65 3.2.5 Analysis................................................................................................. 65 3.3 Results ........................................................................................................... 66 3.3.1 Survival and dispersal ............................................................................ 66 3.3.2 Mechanisms........................................................................................... 66 3.4 Discussion ...................................................................................................... 67 3.4.1 Survival and dispersal in native versus foreign habitat ........................... 67 3.4.2 Mechanisms preventing subsocial species from colonizing the lowland rainforest .................................................................................. 69 3.4.3 Conclusions ........................................................................................... 71 3.5 References...................................................................................................... 79 CHAPTER 4: Different functional strategies applied by five coexisting social and subsocial Anelosimus spider species .................................................... 83 4.1 Introduction.................................................................................................... 83 4.2 Methods ......................................................................................................... 86 4.2.1 Species description ................................................................................ 86 4.2.2 Habitat description................................................................................. 88 4.2.3 Sampling ............................................................................................... 88 4.2.4 Analysis................................................................................................. 90 4.3 Results ........................................................................................................... 91 4.4 Discussion ...................................................................................................... 94 4.4.1 Conclusions ........................................................................................... 98 4.5 References.....................................................................................................110 CHAPTER 5: Are all social traits created equal? Co-evolutionary dynamics between three social traits and dispersal ............................................114 5.1 Introduction...................................................................................................114 5.2 The Model.....................................................................................................117 5.2.1 Model structure.....................................................................................117 5.2.2 Global population control scenarios ......................................................118 5.2.3 Tolerance model ...................................................................................118 5.2.4 Cooperation and altruism models ..........................................................119 5.2.5 Simulations and baseline conditions......................................................119 5.3 Results ..........................................................................................................120 iv  5.4 Discussion .....................................................................................................121 5.4.1 Conclusions ..........................................................................................127 5.5 References.....................................................................................................135 CHAPTER 6: General Conclusions ....................................................................138 6.1 General overview of the thesis.......................................................................138 6.2 Future direction .............................................................................................142 6.3 Significance...................................................................................................144 6.4 References.....................................................................................................144 Appendix 1: Regressions of Anelosimus eximius nest cross-section area to number of females per colony for Cuyabeno River (a), Cuyabeno Forest (b), Jatun Sacha (c), and Via a Loreto (d)......................................................................149 Appendix 2: Summary of the experimental design of each section of the transplant experiment (chapter 3), including the number of spiders and duration of each treatment ......................................................................................150 Appendix 3: Additional principal components analysis tables and figures for chapter 4 ................................................................................................................152 Appendix 4: Full statistical results for nest size and spatial variables discussed in chapter 4 ............................................................................................................156 Appendix 5: Detailed description of model functions and parameter sensitivity test results for chapter 5..........................................................................................158 Appendix 6: Long term evolutionary dynamics of the model (chapter 5) ...............165  v  LIST OF TABLES Table 1.1: Summary and definitions of the social behaviours compared in this review......................................................................................................... 22 Table 1.2: Summary of the sociality gradients described in the text with citations............................................................................................................. 23 Table 2.1: Study site locations and descriptions with average annual temperature and rainfall estimates ..................................................................... 48 Table 2.2: Statistical test results for predictions that colonies from higher elevations are smaller than those at lower elevations and that solitary females occupy a greater proportion of the nests in the higher elevation sites................................................................................................................... 49 Table 3.1: Summary of statistical method and result for each comparison in the transplant experiment................................................................................... 72 Table 4.1: Principal component analysis results for comparison of spatial and temporal variables ............................................................................................100 Table 4.2: Summary of two-way statistical analyses for spatial comparisons .........101 Table 5.1: Descriptions of model parameters and list of baseline values ................128  vi  LIST OF FIGURES Figure 2.1: Diagram of measurements taken for each nest within each transect on a side view (a) and a cross section view (b)................................................... 51 Figure 2.2: In general, upper elevation sites have a significantly higher proportion of solitary females than any of the lowland sites............................... 52 Figure 2.3: Boxplots showing variation in colony size distributions (log number of adult and subadult females) between lowland versus upper elevation sites as well as lowland edge (CR) and interior (CF and JS) sites............................... 53 Figure 2.4: An analysis of variance (ANOVA) comparison of nest (a) and individual (b) density showed that there is no significant difference between upper elevation and lowland habitat................................................................... 54 Figure 3.1: (a) The greatest colony survival (circles, cloud forest; squares, lower montane; triangles, lowland rainforest) and (b) number of individuals remaining occurred in the native cloud forest habitat ....................... 74 Figure 3.2: (a) Transplanted subsocial spider groups yielded more propagules (new colonies founded by dispersers) per nest and (b) those propagules survived longer in their native cloud forest (circles, 2100 m) habitat compared to the lower montane (squares, 1000 m) transplant habitat................. 75 Figure 3.3: (a) Average daily rainfall did not decrease linearly with elevation. However (b) rainfall intensity increased with descreasing elevation as expected ............................................................................................................ 76 Figure 3.4: In rain-sheltered nests (grey) in the lowland rainforest, small groups of both the foreign ((a, c) A. baeza (subsocial)) and native ((b, d) A. eximius (social)) species showed improved survival and a greater amount of webbing compared to nests exposed to the rain (white).................................................... 77 Figure 3.5: (a) The presence of ants in transplanted spider nests was highest in the lowland rainforest, while (b) the reverse pattern was observed for salticid spiders................................................................................................... 78 Figure 4.1: Mean and 95% confidence intervals of each species along principal components axes 1-3 ........................................................................................103 Figure 4.2: Comparison of nest sheet area (a), nest prey capture height (b), and broad-scale habitat measurements (c-e) across all five sympatric species..........104  vii  Figure 4.3: Local scale spatial comparisons of forest edge species (left panel) and forest interior species (right panel) show that species in both habitats exhibit some differences in nest position ..........................................................106 Figure 4.4: Number of nests observed for each species and the most common instar present in each ........................................................................................107 Figure 4.5: Webs of A. jabaquara (a. scale bar 10 cm) A. baeza (b), A. studiosus (c), A. nigrescens (d), and A. dubiosus (e) (b-e scale bar 5 cm) ..........................................................................................109 Figure 5.1: Equilibrium values for the last 5000 generations, averaged over at least 100 replicates, are shown for each parameter combination........................129 Figure 5.2: A time series of 1000 generations, averaged over multiple iterations for each model type (top row: tolerance; middle row: cooperation; bottom row: altruism) for each ecological scenario.......................................................131 Figure 5.3: Group and fecundity frequencies are compared with the model generated by equation 1 (Avilés 1999; Appendix 5), which predicts the per capita growth rate based on the group size........................................................133  viii  ACKNOWLEDGEMENTS I would first like to thank my supervisor, Leticia Avilés, for her mentorship and support during the last five years. I am truly grateful for the opportunity to study such fascinating organisms, and for the freedom to explore my own ideas, both of which were provided in the Avilés lab. I also thank my supervisory committee, B. Crespi, J. Myers, D. Srivastava, and M. Vellend, whose suggestions and insights aided in the development of these ideas, from the proposal stage through to publication. My lab mates, I. Agnarsson, J. Fletcher, J. Guevara, P. Salazar, M. Salomon, and K. Samuk, and members of the Maddison lab, M. Bodner, G. Blackburn, D. Elias, W. Maddison and J. Zhang have also provided a wonderful sounding board over the years, and I thank them all for their thoughtful discussion and encouragement. For their assistance with field work and logistics, I want to thank G. Iturralde, L. Neame, A. Leung, K. Zeron, Yanayacu Biological Station, Jatun Sacha Biological Station, Cuyabeno River Lodge, the Vasquez family in Hollin, all in Ecuador, and the staff at Serra do Japi in Brazil. Thanks also to NSERC and NSF for funding, and to the Museo Ecuatoriano de Ciencias Naturales and SIMBIOE (Sociedad para la Investigación y Monitoreo de la Biodiversidad Ecuatoriana) for sponsoring our research in Ecuador. I could not have gotten to this point without the intellectual and emotional support of my friends, both in the department and back home. You are too many to list, but thanks especially to M. Franklin, J. Ngai, L. White, S. Kolitz, and A. Shui for your friendship and your humour over the years. Thanks also to the folks at Dania for providing a great place to relax during my time in Vancouver. Finally, I have to thank my family, who I can always count for unconditional love and support. My husband Alan has provided his advice, his insight, and his patience, and has probably read every word of this thesis at least twice. My parents, Betsy and Arthur, have always been my biggest supporters, and their confidence in me has always been a huge source of comfort and motivation. Thanks also to my sisters, my grandparents, and the rest of the family for always being there for me and for believing in me.  ix  CO-AUTHORSHIP STATEMENT I designed the studies presented in chapters 2 and 3 in consultation with L. Avilés. I performed the research, analyzed the data, and wrote the manuscripts with editorial and statistical advice from my co-author. The project in chapter 4 was planned in collaboration with L. Avilés and J. Fletcher. I collected the data with the help of J. Fletcher and J. Vasconcellos-Neto. J. VasconcellosNeto and M. Gonzaga contributed their expertise on local natural history to identify spider species and plant families. M. Gonzaga also contributed figure 4.5. I then carried out the statistical analyses and wrote the article and the appendices, with editorial comments from each of my co-authors. I designed the modelling study presented in chapter 5, and developed the model with the help of A. Brelsford. L. Avilés contributed her expertise in shaping the questions investigated and providing advice based on her previous modelling studies. I wrote the manuscript with editorial advice from my co-authors.  x  CHAPTER 1: GENERAL INTRODUCTION 1.1 INTRODUCTION Species exhibiting social behaviours both influence and are influenced by the ecology of their particular environment. This thesis will explore both aspects of this feedback loop through a literature review, several empirical investigations, and a simulation model. In this introductory chapter, I review patterns in the biogeographic distribution of terrestrial social arthropods and link these patterns back to the ecological factors that may shape them (section 1.2). This review will provide a general framework for the empirical studies that I describe in chapters two and three. I then briefly review the literature relevant to the coexistence of similar species sharing a common environment (1.3) and the previous theoretical work exploring how different environmental conditions may impact the evolution of sociality (1.4). Finally, I provide background information on my study organisms, the Anelosimus spiders (1.5), and I describe the themes of this thesis (1.6). 1.2 GEOGRAPHIC PATTERNS IN THE DISTRIBUTION OF SOCIAL SYSTEMS IN TERRESTRIAL ARTHROPODS1 The challenge of explaining the evolution and persistence of diverse and complex social systems has been a major focus among biologists since Darwin’s time. When Hamilton (1964) published his proposed solution to the problem of altruism, the field shifted to focus on intrinsic factors, especially relatedness (r), that may allow specific organisms to become social (Anderson 1984; Costa 2006). For example, many studies of social Hymenoptera have focused on their haplodiploid mode of inheritance as a factor facilitating their social evolution (e.g. Crozier & Pamilo 1996). However, this genetic mechanism neither aids in our understanding of eusociality in other taxa, nor sheds light on why many Hymenoptera are not social (Anderson 1984; Crespi & Choe 1997a). Similarly, Pike and Foster (2008) point out that although aphids are primarily clonal and therefore perfectly related, only 1-2% of described species are considered social. Several recent reviews have 1  A version of section 1.2 has been submitted for publication. Purcell, J. Submitted. Geographic patterns in the distribution of social systems in terrestrial arthropods. Biol. Rev. 1  suggested that researchers should not overlook the other two parameters in Hamilton’s Rule, the costs (c) and benefits (b) of altruistic acts, both of which may be linked to ecology (e.g. Crespi & Choe 1997b; Schwarz et al. 2007; Korb & Heinze 2008). A closer examination of these factors should shed new light on one intriguing problem in sociobiology that studies of relatedness failed to address, namely the incredible diversity of social systems that have evolved repeatedly in a wide variety of organisms (Costa 2006). To answer this question, we must combine our existing understanding of the intrinsic drivers of sociality with new studies of extrinsic factors that may select for greater degrees of sociality (Table 1.1) in some habitats or communities, but not in others. Both social differences across related species and different social behaviours expressed within species should be considered in parallel. Recent taxonomic revisions of lineages containing social species have demonstrated that social traits are more labile than was previously believed (e.g. Brockmann 1997; Danforth & Eickwort 1997; Wcislo & Danforth 1997; Agnarsson et al. 2006; Schwarz et al. 2007; Pike & Foster 2008). These studies show that sociality has evolved independently in many lineages, and that social behaviour can also be lost over evolutionary time (Wcislo & Danforth 1997). Even within a single species, differences in social behaviours exhibited between different populations and individuals can be as great as differences among species (Wcislo 1997a). While many of these studies have focused on taxa with facultative social traits, studies of taxonomic groups with only eusocial members, such as ants and termites, have also started to investigate differences in key social traits, such as the number of nest founders, number of queens, and colony size, both within species and between related species (Table 1.1; Heinze 2008; Korb 2008). Both approaches seek factors that influence the evolution and maintenance of specific social traits, and may share similar ecological influences. Thus, in this review, I compare studies that explore relative differences in social behaviours between related organisms. However, characterizing absolute differences between social systems in different taxonomic groups will also be useful for future studies. Classical studies proposed potential ecological mechanisms that were thought to shape sociality in various social taxa (Evans 1966; Lin & Michener 1972; West-Eberhard 1975). A recent resurgence of studies on the ecology of social evolution has identified and tested several of these factors, including climate variables, defence against predation, and access to a wider breadth of resources, that may play a consistent role in selecting for social 2  behaviours in many organisms (e.g. Kukuk et al. 1998; Hunt & Amdam 2005; Powers & Avilés 2007; Purcell & Avilés 2008; Yip et al. 2008; Zammit et al. 2008). These relevant ecological variables often change along latitudinal and altitudinal gradients. Lower latitudes generally have less seasonal and rainier environments, greater predation rates, and greater food availability, and similar ecological variation may occur along altitudinal gradients in some areas (e.g. Jeanne 1979; Bridgman & Oliver 2006; Powers & Avilés 2007). Here, I review studies of social arthropods to determine whether the degree of sociality is typically greater at (a) lower latitudes and altitudes, (b) higher latitudes and altitudes, or (c) neutral with respect to geography. I also briefly address studies that show local gradients in sociality. In these comparisons, I review studies that have investigated social versus solitary populations or species, as well as studies documenting patterns of variation in social characteristics within a species or taxonomic group with only social members (Table 1.1). The scope of this literature review includes well-studied terrestrial arthropods in taxonomic groups with either eusocial or non-territorial permanent social members (Wilson 1975; Avilés 1997). In effect, this limits my exploration to insects (Hemiptera (Aphidoidea), Hymenoptera, Isoptera, Thysanoptera) and spiders (Araneae). I secondarily review several ecological factors that are known to both influence sociality and vary along latitudinal and altitudinal gradients, since variation in ecology, rather than latitude or altitude per se, probably drives social distribution patterns (e.g. Jeanne 1979; Porter & Hawkins 2001; Avilés et al. 2007). I ask whether these ecological factors play a common role in the social systems of different social arthropods. 1.2.1 Methods In formulating this review, I sought to determine whether any common patterns of social distribution exist across different social arthropod groups. I researched this idea with two complementary approaches. First, I searched the Web of Science database and Google Scholar for literature pertaining to gradients in sociality, or variance in the degree of sociality with altitude and latitude. Search terms used included “social*”, “cooperat*”, and “solitary” with “latitud*”, “altitud*”, and “gradient”. I then expanded my search to include articles that cited these studies or had been cited by them. I used the same approach with the relevant articles in two recent edited volumes that explored the role of ecology in social evolution (Choe & Crespi 1997, Korb & Heinze 2008). Second, I scanned the literature pertaining to 3  each family and genus of ant, bee, wasp, termite, aphid, thrips, and spider for articles that were relevant to gradients in social behaviour as well as ecological mechanisms driving the evolution of sociality or selection for more complex social traits. Here, I focus on more recent papers, but also reference some of the classical papers where many of these ideas were first discussed. I present every paper that I found which explicitly documented a pattern in the distribution of arthropod sociality, and several studies that implicitly showed this pattern, in the following section. In the sections describing local patterns and ecological factors contributing to these social distributions, the available literature is much larger. I therefore discuss a representative subset of the studies in these areas. 1.2.2 Global patterns in the geographic distribution of sociality In this section, I first discuss taxa wherein species or populations with more complex social traits or larger groups tend to be found at lower latitudes and lower altitudes. Then, I describe organisms that show the opposite pattern, with larger groups occurring at more temperate latitudes and higher altitudes. Some relevant studies have investigated a full range of latitudes, from tropical to temperate areas, but many are centred in the tropics or the temperate zone. This is noted in the description of each example. Finally, I discuss taxa that do not show a clear broad scale geographic pattern in the distribution of sociality. Greater degree of social organization at lower latitudes and altitudes Wilson (1975) proposed that ant societies likely evolved in the tropics. He suggested that the less seasonal environment of tropical latitudes would allow colonies to grow to larger sizes, and, in turn, would permit queens to differentiate more extremely from workers in behaviour and morphology. Based on our current knowledge of social arthropods from temperate versus tropical latitudes, many groups seem to follow the pattern of having larger or more complex societies in the tropics. Gaston et al. (1996) showed that the proportion of eusocial Hymenoptera was significantly higher in Costa Rica than in the UK (both countries have a relatively well-studied Hymenoptera fauna). Among the ants, both the most complex and the simplest societies are still found at tropical latitudes (e.g. Fowler 1983; Farji-Brener & Ruggiero 1994; Peeters 1997; Anderson & McShea 2001). Most of the ants found in the temperate zone tend to be socially intermediate, according to the social complexity scale developed by Anderson and McShea (2001). In termites, Porter and Hawkins (2001) demonstrated that colony sizes tend to increase as they approach the equator (Table 1.2). 4  Wilson (1975) similarly noted that primitively eusocial termite groups are found in the temperate zone, while the tropics contain more complex termite species (called ‘higher’ termites by Wilson, family Termitidae). The patterns in these lineages with only eusocial members could arise from either phylogenetic history, or from current selection for or against specific social traits in different habitats (e.g. Danforth & Ji 2001). However, a similar pattern is also present in some lineages wherein sociality has evolved repeatedly. The distribution of social spiders is particularly striking, with at least 20 social spider species from six families occurring at tropical latitudes, while only a few social species are found outside the tropics (Avilés 1997; Agnarsson 2006). In the well-studied genus Anelosimus, social species are restricted to the tropics while congeneric subsocial and non-social species are distributed at both tropical and temperate latitudes around the world (Avilés 1997; Agnarsson 2006). Intraspecific studies of seven halictine bee species follow a similar pattern, with higher latitude populations generally showing more solitary behaviour, while lower latitude populations are usually social (Table 1.2; Packer 1990; Wcislo 1997b; Miyanaga et al. 1999; Richards 2000; Soucy & Danforth 2002; Zayed & Packer 2002; Cronin & Hirata 2003; Richards et al. 2005). Most of these well-documented gradients in halictine bee sociality occur within the temperate zone (Wcislo 1997b), and more studies are needed to determine whether tropical halictine species exhibit similar patterns. Cronin and Schwarz (1999) noted that a greater proportion of colonies produced two broods per year in northern (subtropical) populations of the Australian allodapine bee Exoneura robusta than in southern (temperate) populations, and suggested that this bivoltine lifestyle may present more opportunities for the formation of non-reproductive castes. Among the wasps, Polistes annularis colonies appear to be founded by larger groups of potential queens in lower latitude populations, whereas higher latitude populations have smaller foundress associations (Strassmann 1989). Many of the same patterns observed along latitudinal gradients have been mirrored in comparisons of populations or species along altitudinal gradients. For example, social spiders in the genus Anelosimus are replaced by subsocial congeneric species at higher elevations in Ecuador (Table 1.2; Avilés et al. 2007). Tropical augochlorine bees in the genus Megalopta may show a similar pattern in Costa Rica. Tierney et al. (2008) suggested that lowland species may have a greater proportion of nests formed by groups of cooperative 5  females than the newly described highland species M. atra, but they emphasize that more tests are needed to support this possibility. The same pattern has been observed within the social spider species Anelosimus eximius, which has more solitary females and smaller colony sizes in higher elevation populations in the tropics (Purcell & Avilés 2007). Many halictine bee species from the temperate zone also have solitary populations occurring in high elevation habitats while a eusocial lifestyle is maintained in low elevation populations (Table 1.2; Sakagami & Munakata 1972; Eickwort et al. 1996; Wcislo 1997b; Soucy & Danforth 2002). Greater degree of social organization at higher latitudes and altitudes Some groups of social arthropods do not fit the pattern described above. For example, in comparing six ant subfamilies with temperate and tropical representatives, Kaspari and Vargo (1995) found significantly larger colony sizes at temperate latitudes in two subfamilies, and a similar, though non-significant, trend in three additional subfamilies. In subarctic tundra, ants primarily establish new colonies cooperatively through pleometrosis (i.e. founding by multiple queens), whereas related species in more temperate climates establish nests solitarily (Heinze 1993; Heinze 2008). Within the tropics, Jeanne (1991) reported that, in general, swarm founding wasp species in the genus Polybia formed smaller colonies closer to the equator and larger colonies at higher tropical latitudes. He suggested that a similar pattern may exist for some other swarm founding wasp species, although he pointed out some exceptions as well (Jeanne 1991). Within a species, the subsocial spider Anelosimus studiosus exhibits joint nest founding behaviour more frequently in populations at higher latitudes near the species range edge in Tennessee compared to subtropical populations in Florida (Furey 1998; Jones et al. 2007). Similarly, in temperate populations of the Australian allodapine bee species Exoneura bicolor and Exoneura richardsoni, bees living in heathland habitat tend to found nests alone, while bees in a montane habitat at slightly higher elevations often found colonies in small associations (Schwarz et al. 1997). Fewer examples of these gradients are found in the literature (Table 1.2), even though temperate species tend to be more thoroughly studied than tropical ones.  6  Distribution of sociality neutral with respect to latitude and altitude Unlike the previous examples, social aphid species (characterized by the development of a soldier caste during juvenile instars) appear to be evenly distributed between tropical and temperate habitats (e.g. Stern & Foster 1997; Pike & Foster 2008). However, since new aphid social systems are being described every year, this may either be a case of a neutral distribution with respect to geography or result from a paucity of species descriptions (Stern & Foster 1997). Other groups show variation in the degree of social behaviour both between and within species, but do not show any clear broad-scale biogeographic pattern. These taxa will be discussed in the section on local gradients in social behaviour. 1.2.3 Local gradients in sociality Many social species have been observed to change their basic social behaviour in response to local environmental conditions. In Australian gall thrips, for example, phylogenetic evidence suggests that Oncothrips rodwayi is a solitary species descended from social ancestors (Kranz et al. 2002). This species is found in more temperate habitat than closely related social species, which occur in adjacent arid habitat. In this case, the authors suggest that local conditions in the temperate habitat provide a safer and more predictable dispersal and gall induction window, allowing this species to revert to a solitary lifestyle. Wcislo and Danforth (1997) point out that such reversions from social to solitary behaviour in bee species also generally occur across clear environmental gradients, particularly along gradients from warmer to colder habitats. In other examples, it is more difficult to pinpoint the environmental conditions that result in the observed patterns. In at least two Australian allodapine bee species, Exoneura robusta and Exoneura angophorae, Cronin and Schwarz (1999) found substantial variation in colony size and the number of foundresses that was not explained by position along a latitudinal axis. Among halictine bees, several species show intraspecific variation in colony size and number of founders that is not linked to latitudinal or altitudinal gradients, as in Evylaeus albipes in France (Plateaux-Quenu et al. 2000) and Lasioglossum duplex in Japan (Table 1.2; Hirata et al. 2005). Variation in nest founding behaviour has been studied in many different social organisms. In these species, individuals may establish nests in small groups or remain in groups with multiple queens for an extended period of time in harsher environments, or in places where solitary founders are extremely unlikely to succeed (e.g. Ross & Visscher 1983; 7  Pfennig 1995; Rissing et al. 2000; Matsuura & Nishida 2001; Sanetra & Crozier 2002; Dunn & Richards 2003; Bono & Crespi 2006; Heinze 2008). However, individuals living in more clement environments would exhibit more solitary nest founding. This type of variation often differs between neighbouring habitats, but may also occur along latitudinal or altitudinal gradients. For example, Anelosimus eximius populations in the lowland tropical rainforest rarely have nests established by solitary females (haplometrosis), because solitary founders experience an average per capita reproductive output of less than one in that habitat (Avilés & Tufiño 1998). Solitary nest establishment increases in frequency in populations at higher altitudes where this behaviour is presumably less risky (Purcell & Avilés 2007). Similarly, desert ants such as Messor pergandei may establish nests in small groups of queens in some habitats. With more individuals present to excavate a new nest, they are able to reduce their time above ground, during which desiccation and predation present large risks (Pfennig 1995; Rissing et al. 2000). Michener (2000) proposed that bees may jointly found nests in habitats where soil is difficult to excavate, but would initiate nests alone under more moderate conditions. 1.2.4 Ecological factors that vary along altitudinal and latitudinal axes Latitude and altitude per se are unlikely to drive the observed variation in social behaviour. Instead, ecological factors that vary along these axes probably contribute to the social gradients described above. In this section, I discuss four general ecological variables that have clear gradients along latitudinal and altitudinal axes and which have also been proposed to influence the social systems of social arthropods. Many of these ecological factors may also vary on more local scales. The review of the literature in this section is not meant to be exhaustive, but rather to point out promising future directions that may aid in the identification of the mechanisms underlying the broad-scale geographic patterns discussed above. Many other ecological factors that are not likely to differ along latitudinal and altitudinal axes, including local population density and factors constraining nest sizes, are not considered here, but may also drive some of the variation in sociality that is observed in these organisms (Schwarz et al. 1997; Brockmann 1997; Gunnels et al. 2008; Korb 2008). Climate Seasonality and temperature differ between tropical and temperate latitudes (e.g. Jeanne 1979; Blanckenhorn & Demont 2004). Temperature, at least, also declines with 8  increasing elevation (e.g. Neill & Jørgensen 1999). Wilson (1975) suggested that ant eusociality probably evolved in warm, less seasonal environments, which would have allowed colonies to become perennial. Perenniality, in turn, would permit the co-occurrence of three conditions that should predispose colonies to caste differentiation: longer colony life cycles, overlapping generations, and the concentration of related individuals into larger groups (Oster & Wilson 1978). Along the same lines, Brady et al. (2006) showed that three independent origins of eusociality within the halictid bees probably occurred simultaneously during a warming period 15-25 million years ago, which would have similarly led to conditions favourable to eusocial evolution. Hunt and Amdam (2005) proposed an alternative mechanism for the evolution of sociality in Polistes wasps. They suggested that social behaviour emerged from intrinsic characteristics of the bivoltine life cycle exhibited by species living in seasonal environments, either in the tropics (wet/dry) or in the temperate zone (warm/cold). In solitary species producing two generations per year, one generation would experience a diapause phase during the unfavourable season. They propose that this diapause pathway, which is partially governed by nutrition, may be co-opted to prevent future workers from completing their sexual development as sociality evolves in these species (Hunt & Amdam 2005). Thus, members of the first brood produced would become workers helping to raise their mother’s second brood of future reproductive females. In either case, individuals living in habitats with extremely short growing seasons in which only one brood may be produced, are the least likely to exhibit social behaviours (Yanega 1997; Wcislo & Danforth 1997). Such habitats are most likely to occur at higher elevations (Eichwort et al. 1996) and higher latitudes (Packer 1990) where the unfavourable season is longer. Temperature, as well as seasonality, impacts the ability of bees to produce multiple broods. Hirata and Higashi (2008) showed that differences in social behaviour within a single population of Lasioglossum baleicum may be attributed to local temperature differences. Their results suggested that larvae grew more slowly in nests located in cooler habitat. This explained why bees were solitary (or only able to produce a single brood) in shady habitat, while queens produced a worker brood and a reproductive brood in sunny habitat at the same site. In general, the time that it takes for a brood to develop to maturity relative to the maternal lifespan is thought to influence sociality in a number of organisms. 9  When the mother is likely to die prior to the maturation of her dependent offspring, social behaviour may be favoured as a form of insurance to increase the chances that at least one adult will survive to care for the collective brood (Gadagkar 1990). Offspring development times are often thought to be longer in cooler habitat, as in the example above, which may result in the occurrence of greater degrees of sociality occurring at higher latitudes in some species. For example, in the subsocial spider Anelosimus studiosus, Jones et al. (2007) suggested that females should nest together as a risk aversion strategy in cooler environments where young require more time to develop. Similarly ants in cold, arctic habitats were prone to found nests in small groups of queens, while related species occurring at lower latitudes usually founded nests solitarily (Heinze 1993). The intrinsic extended development time of offspring is also cited as a possible factor maintaining social behaviour in Hover wasps, which occur only in tropical environments (Field 2008) and in maintaining social polymorphism in the tropical bee Megalopta genalis (Smith et al. 2007). Precipitation rates may also play a role in social biology, particularly in taxa that build relatively fragile nests. Precipitation rates tend to vary consistently along latitudinal gradients. Equatorial regions in the intertropical convergence zone tend to receive regular rainfall throughout the year, while subtropical high-pressure cells sitting at the horse latitudes (~30°) tend to result in arid climates. This pattern repeats itself at higher latitudes, with another moist zone at temperate latitudes and a dry zone towards each pole (Bridgman & Oliver 2006). Of course, other factors such as ocean currents and mountain ranges also influence precipitation rates (Bridgman & Oliver 2006). Purcell and Avilés (2008) showed that rainfall intensity, or the amount of rain that falls in a 30-minute window on a rainy day, decreases at higher elevations in the tropics. They suggested that intense rainfall could destroy or degrade the nests of social spiders, and showed that, in the lowland tropical rainforest, small social spider groups that were sheltered from the rain exhibited greater survival and built larger nests, compared to spiders in exposed nests (Purcell & Avilés 2008). The authors suggested that this effect would be reduced in larger colonies, where the ratio of individuals to amount of webbing needed for nest construction and repair would be smaller. Rain has been proposed to impact the distribution or the degree of sociality in other social spider families (Riechert et al. 1986), ants (Murphy & Breed 2007), termites (Picker et al. 2006) and wasps (Barlow et al. 2002). 10  Conversely, drought has also been hypothesized to influence the evolution of sociality by making soil harder and raising the cost of digging for soil nesters in the wasps (Wcislo 1997b) and bees (Michener 2000). In this case, larger groups may form in drier areas, in order to take advantage of unpredictable and short windows of opportunity when nest excavation is possible. A similar mechanism has been proposed as an important ecological driver of sociality in naked mole rats (Jarvis et al. 1994). Arid habitats may also influence the degree of sociality in Australian gall thrips. Kranz et al. (2002) suggested that the solitary life cycle exhibited by Oncothrips rodwayi could result from the shift from the arid habitats occupied by social Oncothrips species, to the more temperate habitat. In the arid habitat, dispersing thrips must time their gall induction carefully to reduce the risk of desiccation, which means that groups may be required to stay in the natal gall for a longer period of time (Chapman et al. 2008). In this case, investing in gall defence (i.e. soldiers) would be a useful strategy, whereas the temperate species would experience longer and more predictable dispersal windows, so the period of cohabitation and, by extension, the need for soldiers would be reduced. So far, there is relatively little evidence to suggest whether differences in social behaviour in response to abiotic conditions result from phenotypic plasticity or from local adaptation (e.g. Wcislo & Danforth 1997). In one exception, Plateaux-Quènu et al. (2000) collected Evylaeus albipes from two populations with different social phenotypes and reared them in a common garden. Although the sample size was small, they showed that bees collected from the eusocial population in western France tended to retain more social traits than those collected from the solitary population in eastern France, suggesting that local adaptation probably plays a role. Similarly, Soucy and Danforth (2002) found that high altitude and high latitude populations of Halictus rubicundus in North America were more closely related to each other than to neighbouring social populations at lower elevations and latitudes. This pattern suggests that solitary populations share a common ancestor and probably did not result from independent colonization events of colder environments by nearby social populations. Predation Defence against predation has been suggested as a major function of social groups in a wide range of organisms (e.g. Lin & Michener 1972; Pike & Foster 2008). Carnivorous ant 11  species appear to be among the most important predators of social arthropods, including other ants (Hölldobler & Wilson 1990). Ants are more abundant and speciose at lower latitudes (Jeanne 1975, 1979; Hölldobler & Wilson 1990) and lower elevations in the tropics (Janzen et al. 1976; Samson et al. 1997; Bruhl et al. 1999; Purcell & Avilés 2008; Guevara & Avilés in press). This, in turn, suggests that predation pressure would be greater in the lowland tropics, an idea that has been suggested in previous studies (e.g. Jeanne 1991; Kaspari & Vargo 1995). Two opposing hypotheses have been posed with regards to the effect of an abundant predator on social behaviour. Jeanne (1991) suggested that higher rates of predation by army ants could cause some swarm-founding wasps to form smaller colonies with a higher intrinsic rate of growth, given that many species are highly susceptible to army ant predation regardless of colony size. Kaspari and Vargo (1995) echoed the same idea in their discussion of the potential causes of smaller colonies in some ant subfamilies in the tropics. Alternatively, many studies have suggested that larger social groups would have an advantage over smaller groups in defence against intrusions by other predators, particularly scouting and recruiting ants (e.g. Lin & Michener 1972; Jeanne 1991; Henschel 1998; London & Jeanne 2003). The latter pattern is more likely to result in larger colonies occurring in the lowland tropics, and is believed to be an important selective force for larger colonies or, in some cases, specialized social defences (including soldier castes and specialized nest structures) in Atta ants (Powell & Clark 2004), aphids (Foster 2002; Pike et al. 2007), thrips (Chapman et al. 2008), termites (e.g. Brandao et al. 2008; Korb 2008), wasps (O’Donnell & Jeanne 1990; Bouwma et al. 2007), bees (Wcislo & Schatz 2003), hornets (e.g. Martin 1992), and social spiders (Henschel 1998; Purcell & Avilés 2008). Other social adaptations also result from predation pressure. For example, Aphaenogaster araneoides ant colonies maintain multiple nests, but occupy only one at any given time. Individuals are then able to retreat to an unoccupied nest in the event of an army ant raid (McGlynn 2007). Predation pressure may also influence colony foundation strategies. Experiments on allodapine bees in Australia demonstrated that queens often found nests in pairs or join established nests in the presence of predators, but usually establish alone when predators are absent (Zammit et al. 2008). In the social spider Stegodyphus dumicola, Henschel (1998) showed a negative linear relationship between the  12  rate of ant attacks and the proportion of spiders nesting solitarily in nine different populations from Namibia. Prey size and availability In many cases, groups of predatory organisms may be able to access larger prey than individuals (e.g. Caraco & Wolf 1975; Rypstra & Tirey 1991). Recent studies suggest that insects tend to be larger in lowland tropical rainforest habitat than at higher elevations in the tropics (Janzen et al. 1976; Guevara & Avilés 2007) and than in the temperate zone (Powers & Avilés 2007). If this is the case, predatory taxa should form larger or more complex social groups in the lowland tropics. Social spiders, for example, work together to capture prey that are many times larger than individuals would be able to subdue alone, and larger colonies need these larger prey items to meet their nutritional needs (Yip et al. 2008). All species of ants are social, and only a subset of ant species are predatory, but cooperative access to resources may drive an increase in the degree of sociality in some tropical ant groups, such as the army ants. Many of the examples of nest mates working together to subdue and transport large prey items occur in the tropics, as in neotropical army ants (O’Donnell et al. 2005), and African ants Leptogenys nitida and Oecophylla longinoda (Duncan & Crewe 1994; Wojtusiak et al. 1995). Non-predatory groups may also cooperate to access food, as in bark beetles (Berryman et al. 1985), where individuals could not overcome the defences of a tree alone. On the other hand, many social groups do not require cooperation to fulfil their dietary needs. In some such cases, however, recruitment of colony mates to food patches may play an important and parallel role in the evolution of cooperation. In comparing sociality in bees versus their ancestral sphecid wasps, Strohm & Liebig (2008) actually suggested that a shift from foraging on large prey items to foraging on pollen and nectar facilitated the repeated evolution of sociality in many bees. In fact, they point out that the only instance of sociality in a sphecid wasp occurs in a species that feed its larvae with many very small prey items. They suggest that collecting enough of a small but abundant resources would require more helpers, and that the helpers could be small but efficient, whereas individuals capturing and transporting large intact prey items must be larger (Strohm & Liebig 2008). Cooperative transport would be more difficult in organisms requiring flight to reach their nests (unlike ants and spiders). Many other social polistine, vespine, and 13  stenogastrine wasps are known to butcher large prey items, which may similarly select for cooperative foraging and provisioning (Strohm & Liebig 2008), but would be more likely to follow the geographic distribution of prey sizes outlined in the previous paragraph. As both examples demonstrate, the ability to access more resources through cooperation and information transfer may be an important selective force in many social arthropods. Parasites and pathogens Pathogens appear to become more common at subtropical and tropical latitudes, but parasitism rates may be greater in the temperate zone (e.g. Hawkins et al. 1997). However, many parasitoid species are restricted to the tropics (e.g. Fernandez-Marin et al. 2006a) or are more prominent in tropical ecosystems (Gaston et al. 1996; Schmid-Hempel 1998). Theory predicts that disease should select against social behaviours, as individuals living in larger groups with more social contact would be more susceptible to infection than those living alone (reviewed by Cremer et al. 2007). We may expect to observe similar disadvantages in social groups with regards to parasite transmission and infection rates (Shykoff & Schmid-Hempel 1991; Schmid-Hempel 1998). Nevertheless, many social organisms have evolved specialized behaviours or other defences that allow large groups to avoid or mitigate infection, which is one reason that the social organisms may still occur in habitats where pathogens and parasites are common. In the termite Coptotermes formosanus, for example, mutual grooming resulted in resistance to entomopathogenic fungi, while solitary individuals were highly susceptible (Yanagawa et al. 2008). Many ant and bee species show similar mutual grooming behaviours, deploying antimicrobial compounds during these activities, in the presence of pathogens (e.g. Fernandez-Marin et al. 2006b; Stow et al. 2007). Baer and Schmid-Hempel (1999) explored the possibility that a queen bumblebee can lower her colony’s susceptibility to parasitism by increasing the genetic diversity of her offspring through polyandry. In that study and subsequent studies, they found that avoidance of pathogens may be a major force selecting against inbreeding and high intracolony relatedness in a range of social organisms (reviewed by Cremer et al. 2007). Other types of parasitism may have different effects on social behaviours. When faced with highly specialized parasites attacking individuals in social groups, the presence of large groups or aggregations of groups may actually be advantageous through the selfish herd effect (Lin & Michener 1972). Social groups may also be able to defend against certain 14  parasites (e.g. Danforth & Eickwort 1997; Brockmann 1997). For example, the rapid colony growth achieved by multiple founders and larger colony sizes in general may deter potential kleptoparasitism, as Bono and Crespi (2006) observed in Australian Dunatothrips. On the other hand, social defences cannot counter Trojan horse parasitism, in which parasites oviposit on prey being transported to the nest (Strohm & Liebig 2008). Ultimately, while infection may be a major obstacle to sociality, more studies of pathogens and parasites of social arthropods are needed to determine how these factors influence the evolution and distribution of social behaviours. 1.2.5 Conclusions and Future Directions Most social arthropods investigated in this survey fell into one of two global patterns of social distribution: the degree of sociality and/or group size increased at lower latitudes and lower altitudes, or colony size increased at higher latitudes and higher altitudes (Table 1.2). While other studies have explored these patterns within a genus or a subfamily, this review seeks patterns that are general to a more diverse subset of social arthropods. Several ecological factors also vary along latitudinal and altitudinal gradients and appear to influence the social systems similarly in some of the organisms compared here. Like previous reviews comparing social and solitary behaviours between related species (e.g. Wcislo & Danforth 1997; Schwarz et al. 2007), I suggest that identifying consistent biogeographic patterns in the distribution of sociality may help us to isolate and test environmental mechanisms that contribute to the evolution of sociality in these diverse taxa. Several future directions follow from these findings. First, it would be interesting to extend the scope of this comparison to include other forms of social behaviours, such as cooperative breeding, communal living and temporary aggregations, and to look at vertebrates as well as arthropods. However, even within the scope of this article, we are still not able to see the full story in many taxa. In particular, the social systems of tropical species and populations have not been investigated thoroughly. Thus, I secondarily suggest that further natural history studies of tropical species be carried out, with a particular focus on characterizing social traits that will be comparable both within the taxon of interest and across other social taxa. Third, I emphasize the need for more broad scale studies seeking patterns in the distribution of social systems across different ecological zones. I suggest using two separate but complementary approaches: niche mapping to describe the 15  distribution of social species as predicted by key environmental variables, and meta-analyses to compare the distribution of social traits, both within and between species. With both approaches, the phylogenetic history of species should be taken into account, since similar distributions could result from shared ancestry or selection favouring a similar suite of behaviours. However, identifying patterns is only the first step. We should apply these common patterns to identify likely mechanisms driving the evolution, persistence, and success of these diverse social systems, with the ultimate goal of moving beyond correlational evidence and experimentally testing the potential ecological drivers of sociality. Future studies will face some specific statistical and empirical challenges that should be acknowledged from the outset. First, patterns in the gradients of sociality along latitudinal and altitudinal axes will always be confounded with overlapping gradients in species richness and abundance. For example, the tropics contain the full range of ant societies, from the simplest and smallest societies that contain fewer than 10 group members (Pachycondyla sublaevis; Peeters et al. 1991) to the largest and most complex societies, such as the leaf cutter and army ants (Anderson & McShea 2001). On the other hand, temperate ants often seem to exhibit intermediate social complexity (Anderson & McShea 2001). This pattern could emerge by chance from the fact that the majority of ant species diversity is located in the tropics (Hölldobler & Wilson 1990), but we would need to know the relative number of ant species in each zone to control for this possibility. This leads to the second statistical problem: How do we account for species that have not been described or for which social descriptions or colony size data are not available? This is especially problematic, since researchers have generally studied temperate fauna in more detail than tropical fauna. There are two potential solutions, which would complement one another. On the one hand, seeking patterns and mechanisms within well-studied genera or sub-families in well-studied areas, as we have for the spider genus Anelosimus (Avilés et al. 2007), might reduce the potential for errors resulting from variable sampling effort. On the other hand, some social organisms including Anelosimus species may represent a biased sample of the species in the tropics, since researchers actively seek social species. To counteract this possibility and mitigate the effects of problem #1, we should also compare the frequency of social species observed during expeditions where local fauna are exhaustively sampled, and social organisms are not specifically targeted, if it is possible to identify social taxa under such circumstances. 16  Additionally, I am concerned that the reference to ‘harsh’ habitats resulting in higher degrees of sociality, tolerance, or joint nest founding could lead to circular logic. For example, Heinze (1993) argued that the tundra environment colonized by a handful of ants is particularly harsh, which is why these ants almost universally found nests through pleometrosis. At the same time, Avilés and Tufiño (1998) identified the tropical rainforest as a harsh environment, requiring cooperation for survival. While both of these environments may pose profound challenges for the organisms living there, these challenges are clearly different. Avilés (1999) suggested that “harsh” environments be defined as habitats where solitary individuals cannot replace themselves, and where group living and cooperation make persistence in those environments possible. To prevent the possibility of circular arguments in the future, I suggest the separation of the definitions of the demographically harsh environments described by Avilés (1999) from the external conditions that make an environment “harsh” for a given organism. In this case, it will be important to identify what specific ecological factors cause demographically harsh environments and, in turn, whether the occurrence of those factors can predict the distribution of sociality in other areas. This review, like other recent reviews and books (e.g. Schwarz et al. 2007; Korb & Heinze 2008), suggests that extrinsic or ecological, factors are at least as important in selecting for social behaviours as the intrinsic factors that have been very well studied during the past 40 years. Here, I compile studies showing that many social arthropods vary in their degree of sociality both between and within species, and that much of this variation occurs along broad-scale biogeographic axes. This type of large-scale study, in turn, can identify ecological factors that seem to consistently influence independently evolved social lineages in similar ways. Ultimately, these gradients of sociality should provide a natural laboratory for future experimental tests of these ecological mechanisms. 1.3 SPECIES COEXISTENCE AND SOCIALITY Classical ecological theory suggests that species sharing the same ecological niche cannot co-occur in a single environment (Gause 1934). Instead, species must differ from one another in their use of resources or in the timing of their activities in order to coexist (Hutchinson 1959; MacArthur & Levins 1967). This “limiting similarity” has been documented in a wide range of natural systems (MacArthur 1958; Schoener 1968; Pianka 17  1969; Albrecht & Gotelli 2001; Silvertown 2004; Behmer & Joern 2008). Species living in the same environment will often differ in one or more key traits, which enables them to utilize a different set of resources. For instance, two mice species overlap in the Negev Desert and prefer the same type of food in a cafeteria experiment. However, one species is diurnal and the other is nocturnal, so they utilize different food resources in their natural environment (Kronfeld-Schor & Dayan 1999). Other species differ in their use of spatial resources, such as sympatric gecko species living on islands in the Indian Ocean, which use perches of different widths and heights (Harmon et al. 2007). Some sympatric species differ in their body size or morphology, which may enhance ecological differentiation along dietary or spatial axes (Schoener 1968, 1974; Basset 1995; Grant & Grant 2006; Murray & Baird 2008). For example, benthic and limnetic forms of three-spine sticklebacks are found in the same lakes, but have developed morphological and body size differences that allow them to efficiently exploit different habitats within each lake (Schluter 1996). A shift from solitary living to sociality in different species should result in differences in the way that a species interacts with its environment that may be analogous to shifts in body size. Social organisms living in nests would require a larger and more complex structure to contain each group, and may have access to food resources not available to otherwise similar solitary organisms (e.g. Powers & Avilés 2007; Guevara & Avilés 2007; Yip et al. 2008). Along the same lines, larger groups may face different habitat requirements than small groups. Thus, similar species with different social systems may be able to coexist because of differences in their use of resources and habitat requirements. This idea has not yet been carefully explored in the literature. 1.4 MODELING THE EVOLUTION OF SOCIALITY AND DISPERSAL Explaining the evolution of sociality in some species is a challenge because social individuals must face the disadvantages of competing with their group mates for resources and increasing their susceptibility to parasites and diseases (e.g. Hughes et al. 2004). Some individuals may cheat in order to gain fitness without paying cooperation costs, thus decreasing the fitness of cooperators. These fitness costs must be balanced by intrinsic benefits of living in a group. Theory suggests that altruism can evolve between relatives (Hamilton 1964), non-relatives who share a tendency to help (Queller 1985), or individuals 18  that help because they expect to receive help in return (Trivers 1971). The fact that altruism can evolve between non-relatives suggests that cooperation may yield synergistic fitness benefits in some organisms (Avilés 1999). This may occur through increased access to resources (e.g. Giraldeau 1984; Yip et al. 2008), better protection from predators and other environmental disturbances (e.g. Purcell & Avilés 2008), decreased energy expenditure in maintaining territory or rearing young, or decreased risk of fitness costs resulting from premature mortality (e.g. Gadagkar 1990). The evolution of sociality probably influences and is influenced by selection on dispersal behaviours. Empirical studies have demonstrated that increasing the predation risk experienced by dispersers may prolong the cooperative period in some species (e.g. Heg et al. 2004), and that increasing the benefits of philopatry through food supplementation can cause subsocial species to delay dispersal (e.g. Korb & Schmidinger 2004). These findings support the idea that there is a positive feedback between increasing social grouping and decreasing dispersal. This idea has received a great deal of theoretical attention in recent years. Classical models of dispersal demonstrated that dispersal and philopatry each carry conflicting costs and benefits (Hamilton & May 1977). Dispersing individuals may face an increased risk of mortality as they seek appropriate habitats (Johnson & Gaines 1990), but they have a small but important chance of founding a new population or colony, which would dramatically increase their fitness (e.g. Muller 2001; Parisod et al. 2005). Philopatry, on the other hand, would result in a more certain fitness payoff, but may subject individuals to greater local competition among relatives (Hamilton & May 1977) and to inbreeding (Bengtsson 1978). Later studies suggested that social traits may arise to alleviate the costs of philopatry, and that this might be most common in environments where mobility is limited. However, theoretical studies showed that local competition exactly cancelled out the benefits of cooperation under simple conditions (Taylor 1992; Wilson et al. 1992). Recent theoretical studies have changed gears to focus on exploring the co-evolutionary dynamics between altruism and dispersal (Le Galliard et al. 2005) and identifying the specific kinship or environmental conditions under which altruism can evolve in the presence of dispersal and kin competition (Perrin & Lehmann 2001; Lehmann & Perrin 2002; Lehmann et al. 2006; Vainstein et al. 2007; Alizon & Taylor 2008). 19  1.5 BACKGROUND ON ANELOSIMUS SPIDERS Spiders in the genus Anelosimus provide an ideal system for exploring the ecology of social evolution, because closely related spider species vary in their social behaviours from nearly solitary to highly social (Avilés 1997; Agnarsson 2006). Species in the genus occur throughout the world and live in a wide variety of environments (Avilés 1997; Furey 1998; Agnarsson 2006; Agnarsson & Zhang 2006; Avilés et al. 2007). Social spiders live together throughout their life cycle and cooperate in prey capture, web maintenance, and brood rearing. Successive generations continue to occupy the same web, with dispersers only rarely leaving to establish new colonies. In subsocial species, on the other hand, colonies consist of a mother and her offspring. Siblings share the natal nest for part of their life cycle, but disperse before maturity to establish independent colonies (Avilés 1997; Avilés & Gelsey 1998; Avilés & Bukowski 2006). The timing of dispersal for subsocials and the maximum group size for socials, provide a more or less continuous axis of ‘social behaviour’ within this group, from nearly non-social species (i.e. young spiders disperse as juveniles) to highly social species (i.e. colony sizes reaching thousands of adult spiders.) Sociality is thought to have evolved at least four times within the genus (Agnarsson 2006), which allows for comparisons across multiple origins of a similar behavioural phenomenon. Moreover, social spiders are thought to have evolved from subsocial-like ancestors through the loss of their dispersal phase (Avilés 1997). In general, social spiders are restricted to lowland tropical areas, while subsocial species tend to occupy highland tropical and temperate zones (Avilés et al. 2007). In intermediate zones, social and subsocial species frequently overlap. This spatial distribution provides the opportunity to explore questions about why social and subsocial species frequently occur in different habitats, and what happens when they overlap. 1.6 THEMES AND HYPOTHESES INVESTIGATED IN THE THESIS In this thesis, I explore how environmental factors influence sociality and how social behaviours, in turn, impact local ecology. In the first two empirical chapters (chapters 2 and 3), I investigate several examples of how ecology can influence social behaviours. I ask whether populations of a social spider centred in the lowland tropical rainforest, but 20  extending to about 1300 m elevation in Ecuador exhibit differences in social behaviour along this altitudinal gradient. Specifically, I test the hypothesis that populations at the higher elevations will be more ‘subsocial-like,’ because subsocial species replace social species at higher elevations in the area. I follow up on the results of this and a related study (Avilés et al. 2007) in my third chapter with an experimental transplant designed to identify factors that simultaneously prevent subsocial Anelosimus species from colonizing the lowland rainforest while favouring species with large social groups in this habitat. In chapter 4, I study how sociality influences community dynamics and resource partitioning in habitat shared by social and subsocial Anelosimus species. I hypothesize that different levels of sociality may result in the use of different local habitats and the timing of dispersal in ways that may reduce competition and thus facilitate coexistence. Finally, I simulate the co-evolutionary dynamics of several distinct social traits with dispersal under two different simple environmental conditions in my theoretical chapter (chapter 5). I test the hypothesis that increasing the cost of dispersal will cause each of the social traits that I investigate to increase. I also expect to find differences in the social dynamics depending on whether stochastic mortality is imposed at the individual level or at the patch level. Overall, each of these chapters investigates one piece of the larger puzzle of how ecological conditions influence the evolution and persistence of sociality in spiders.  21  Table 1.1: This table summarizes and defines the social traits compared in this review. Some of the studies reviewed here focus on shifts between solitary or subsocial systems to social or eusocial life styles, while others investigate more subtle differences in the degree of sociality exhibited within a species or between closely related species. Research on the origin of sociality may focus on shifts from solitary or subsocial behaviour to social behaviour, or to eusocial behaviour, so the order of these words in this table is not important. Each paper generally compares only one or two social traits across different populations or species. Definitions of the terms in the “origin of sociality” section are modified from box 1 in Wcislo and Danforth (1997). The term “social” is also used in the thesis in a more general way to characterize interactions between individuals that benefit one (i.e., altruism) or both (i.e. cooperation) participants. Less Social  Definition  More Social  Definition  social  Groups of multiple adults in which the dominantsubordinate relationships among individuals (if any) are unspecified  Origin of sociality  solitary  Each female occupies her own nest and is solely responsible for rearing her brood  subsocial  Parent(s) feed and guard immature offspring for an extended period during offspring development  eusocial  Social behaviour characterized by the overlapping generations and division of labour in which some individuals are effectively sterile  small colonies  relatively few group members  large colonies  many group members  one colony founder (haplometrosis)  new colonies are founded by a single female or a female and a male  several colony founders (pleometrosis)  new colonies are founded by more than one female  few colony founders  propagules (new colonies) contain relatively few members (either reproductives or workers)  many colony founders  propagules contain many members (either reproductives or workers)  Elaboration of sociality  22  Table 1.2: Summary of the gradients described in the text with citations. social currency compared Increasing Social Behaviour at lower latitude, elevation General Grouping  within or between species  organism  gradient(s)  Termites  latitude  colony size  interspecific  Social spiders  altitude and latitude  degree of sociality  interspecific  citation(s)  Porter & Hawkins 2001; Wilson 1971 Avilés 1997; Avilés et al. 2007  Species Halictine bees Augochlorella striata  latitude  Halictus poeyi  latitude  Halictus rubicundus  altitude and latitude  Lasioglossum apristum Lasioglossum balecium Lasioglossum calceatum Lasioglossum malachurum Nomoiides minutissimus Paper wasps Polistes annularis  latitude latitude altitude latitude latitude  degree of sociality degree of sociality degree of sociality degree of sociality degree of sociality degree of sociality degree of sociality degree of sociality  intraspecific  Packer 1990  intraspecific  Zayed & Packer 2002  intraspecific  Eickwort et al. 1996; Soucy & Danforth 2002  intraspecific  Miyanaga et al. 1999  intraspecific  Cronin & Hirata 2003  intraspecific intraspecific  Sakagami & Munakata 1972; Schwarz et al. 2007 Richards 2000; Richards et al. 2005  intraspecific  Wcislo 1997b  latitude  number of founders  intraspecific  Strassmann 1989  altitude  degree of sociality  intraspecific  Purcell & Avilés 2007  colony size  interspecific  Kaspari & Vargo 1995  colony size  interspecific  Jeanne 1991  number of founders  interspecific  Heinze 1993  Social spiders Anelosimus eximius  Decreasing Social Behaviour at lower latitude, elevation General Grouping Ants (5 subfamilies) Swarm-founding Wasps Genus Arctic ants (Leptothorax species)  latitude latitude (tropics only)  latitude  23  organism  gradient(s)  social currency compared  within or between species  citation(s)  number of founders  intraspecific  Schwarz et al. 1997  Allodapine Bees Exoneura bicolor  altitude  Exoneura richardsoni Social spiders Anelosimus studiosus  altitude  number of founders  Intraspecific  Schwarz et al. 1997  latitude  number of founders  intraspecific  Furey 1998; Jones et al. 2007  interspecific  Stern & Foster 1997; Pike et al. 2007  degree of sociality  intraspecific  Cronin & Schwarz 1999  degree of sociality  interspecific  Kranz et al. 2002  degree of sociality  intraspecific  Plateaux-Quenu et al. 2000  degree of sociality  intraspecific  Hirata & Higashi 2008  degree of sociality  intraspecific  Hirata et al. 2005  Neutral social distribution with respect to geography presence of Aphids soldiers Local Variation Allodapine bees Exoneura robusta, E. angophorae Thrips Oncothrips rodwayi  solitary; occurs in temperate environment  Halictine bees Evylaeus albipes Lasioglossum baleicum Lasioglossum duplex  east versus west of France shady versus sunny nest location two habitats in northern Japan  24  1.7 REFERENCES Agnarsson, I. 2006. A revision of the New World eximius lineage of Anelosimus (Araneae, Theridiidae) and a phylogenetic analysis using worldwide exemplars. Zool. J. Linn. Soc.-Lond. 146, 453-493. Agnarsson, I., Avilés, L., Coddington, J.A. & Maddison, W.P. 2006. Social theridiid spidersrepeated origins of an evolutionary dead-end. Evolution. 60, 2342-2351. Agnarsson, I. & Zhang, J.X. 2006. New species of Anelosimus (Theridiidae, Araneae) from Southeast Asia and Africa, with notes of sociality and color polymorphism. Zootaxa 1147, 1-34. Albrecht, M. & Gotelli, N.J. 2001. Spatial and temporal niche partitioning in grassland ants. Oecologia 126, 134-141. Alizon, S. & Taylor, P. 2008. Empty sites can promote altruistic behavior. Evolution 62, 1335–1344. Anderson. M. 1984. The evolution of eusociality. Ann. Rev. Ecol. Syst. 15, 165-189. Anderson, C. & McShea, D.W. 2001. Individual versus social complexity, with particular reference to ant colonies. Biol. Rev. Camb. Philos. Soc. 76, 211-237. Avilés, L. 1997. Causes and consequences of cooperation and permanent sociality in spiders. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 476-498. Cambridge, UK: Cambridge University Press. Avilés, L. 1999. Cooperation and non-linear dynamics: An ecological perspective on the evolution of sociality. Evol. Ecol. Res. 1, 459-477. Avilés, L. & Bukowski, T.C. 2006. Group living and inbreeding depression in a subsocial spider. Proc. R. Soc. B 270, 157-163. Avilés, L. & Gelsey, G. 1998. Natal dispersal and demography of a subsocial Anelosimus species and its implications for the evolution of sociality in spiders. Can. J. Zoo. 76, 2137-2147. Avilés, L. & Tufiño, P. 1998. Colony size and individual fitness in the social spider Anelosimus eximius. Am. Nat. 152, 403-418. Avilés, L., Agnarsson, I., Salazar, P., Purcell, L., Iturralde, G., Yip, E., Powers, K.S. & Bukowski, T. 2007. Altitudinal patterns of spider sociality and the biology of a new mid-elevation social Anelosimus species in Ecuador. Am. Nat. 170, 783-792. 25  Baer, B. & Schmid-Hempel, P. 1999. Experimental variation in polyandry affects, parasite loads and fitness in a bumblebee. Nature 397, 151-154. Barlow, N.D., Beggs, J.R., & Barron, M.C. 2002. Dynamics of common wasps in New Zealand beech forests: a model of density dependence and weather. J. Anim. Ecol. 71, 663-671. Basset, A. 1995. Body size-related coexistence – an approach through allometric constrains on home-range use. Ecology 76, 1027-1035. Behmer, S.T. & Joern, A. 2008. Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc. Natl. Acad. Sci. U.S.A. 105, 1977-1982. Bengtsson, B.O. 1978. Avoiding inbreeding- at what cost. J. Theor. Biol. 73, 439-444. Bergmann, C. 1847. Über die Verhältnisse der Wärmeökonomie der Thiere zu ihrer Grösse. Göttinger Studien 3, 595-708. Berryman, A.A., Dennis, B., Raffa, K.F., & Stenseth, N.C. 1985. Evolution of optimal group attack, with particular reference to bark beetles (Coleoptera: Scolytidae). Ecology 66, 898-903. Blanckenhorn, W.U. & Demont, M. 2004. Bergmann and converse Bergmann latitudinal clines in arthropods: Two ends of a continuum? Integ. Comp. Biol. 44, 413-424. Bono, J.M. & Crespi, B.J. 2006. Costs and benefits of joint colony founding in Australian Acacia thrips. Insectes Soc. 53, 489-495. Bouwma, A.M., Howard, K.J., & Jeanne, R.L. 2007. Rates of predation by scouting-andrecruiting ants on the brood of a swarm-founding wasp in Costa Rica. Biotropica 39, 719-724. Brandao, I.C.T., Florencio, D.F., & Diehl, E. 2008. Differential predation of soldiers and workers of Cortaritermes fulviceps (Termitidae: Nasutitermitinae) by ants (Hymenoptera: Formicidae). Sociobiology 51, 523-537. Bridgman, H.A. & Oliver, J.E. 2006. The Global Climate System. Cambridge, UK: Cambridge University Press. Brockmann, H.J. 1997. Cooperative breeding in wasps and vertebrates: the role of ecological constraints. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 347-371. Cambridge, UK: Cambridge University Press. 26  Bruhl, C.A., Mohamed, M. & Linsenmair, K.E. 1999. Altitudinal distribution of leaf litter ants along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. J. Trop. Ecol. 15, 265-277. Caraco, T. & Wolf, L.L. 1975. Ecological determinants of group sizes of foraging lions. Am. Nat. 109, 343-352. Chapman, T.W., Crespi, B., & Perry, S.P. 2008. The evolutionary ecology of eusociality in Australian gall thrips: a ‘model clades’ approach. Preface. In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. 57-84. Berlin, Germany: Springer-Verlag. Choe, J.C. & Crespi, B. 1997. The evolution of social behaviour in insects and arachnids. Cambridge, UK: Cambridge University Press. Costa, J.T. 2006. The Other Insect Societies. Cambridge, MA, USA: Belknap Press. Cremer, S., Armitage, S., & Schmid-Hempel, P. 2007. Social immunity. Curr. Biol. 17, R693-R702. Crespi, B. & Choe, J.C. 1997a. Explanation and evolution of social systems. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 499-524. Cambridge, UK: Cambridge University Press. Crespi, B. & Choe, J.C. 1997b. Introduction. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 1-7. Cambridge, UK: Cambridge University Press. Cronin, A.L. & Hirata, M. 2003. Social polymorphism in the sweat bee Lasioglossum (Evylaeus) baleicum (Cockerell) (Hymenoptera, Halictidae) in Hokkaido, northern Japan. Insectes Soc. 50, 379-386. Cronin, A.L. & Schwarz, M.P. 1999. Latitudinal variation in the life cycle of allodapine bees (Hymenoptera; Apidae). Can. J. Zool. 77, 857-864. Crozier, R.H. & Pamilo, P. 1996. Evolution of Social Insect Colonies: Sex Allocation and Kin Selection. Oxford, UK: Oxford University Press. Danforth, B.N. & Eickwort, G.C. 1997. The evolution of social behavior in the augochlorine sweat bees (Hymenoptera: Halictidae) based on a phylogenetic analysis of the genera. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 270-292. Cambridge, UK: Cambridge University Press.  27  Duncan, F.D. & Crewe, R.M. 1994. Group hunting in a ponerine ant, Leptogenys nitida Smith. Oecologia 97, 118-123. Dunn, T. & Richards, M.H. 2003. When to bee social: interactions among environmental constraints, incentives, guarding, and relatedness in a facultatively social carpenter bee. Behav. Ecol. 14, 417-424. Eickwort, G.C., Eickwort, J.M., Gordon, J. & Eickwort, M.A. 1996. Solitary behavior in a high altitude population of the social sweat bee Halictus rubicundus (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 38, 227-233. Evans, H.E. 1966. Behavior patterns of solitary wasps. Ann. Rev. Entomol. 11, 123-154. Farji-Brener, A.G. & Ruggiero, A. 1994. Leaf-cutting ants (Atta and Acromyrmex) inhabiting Argentina: Patterns in species richness and geographical range sizes. J. Biogeogr. 21, 391-399. Fernandez-Marin, H., Zimmerman, J.K., & Wcislo, W.T. 2006a. Acanthopria and Mimopriella parasitoid wasps (Diapriidae) attack Cyphomyrmex fungus-growing ants (Formicidae, Attini). Naturwissenschaften 93, 17-21. Fernandez-Marin, H., Zimmerman, J.K., Rehner, S.A., & Wcislo, W.T. 2006b. Active use of the metapleural glands by ants in controlling fungal infection. Proc. R. Soc. B 273, 1689-1695. Field, J. 2008. The ecology and evolution of helping in hover wasps (Hymenoptera: Stenogastrinae). In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. 85-108. Berlin, Germany: Springer-Verlag. Foster, W.A. 2002. Soldier aphids go cuckoo. Trends Ecol. Evol. 17, 199-200. Fowler, H.G. 1983. Distribution patterns of Paraguayan leaf-cutting ants (Atta and Acromyrmex) (Formicidae: Attini). Stud. Neotrop. Fauna Envi. 18, 121-138. Furey, R.E. 1998. Two cooperatively social populations of the theridiid spider Anelosimus studiosus in a temperate region. Anim. Behav. 55, 727-735. Gadagkar, R. 1990. Evolution of eusociality: The advantage of assured fitness returns. Philos. T. Roy. Soc. B 329, 17-25. Gaston, K.J., Gauld, I.D. & Hanson, P. 1996. The size and composition of the hymenopteran fauna of Costa Rica. J. Biogeogr. 23, 105-113. Gause , G. F.1934. The Struggle for Existence. Williams and Wilkins, Baltimore, MD. 28  Giraldeau, L.A. 1984. Group foraging - the skill pool effect and frequency-dependent learning. Am. Nat. 124, 72-79. Grant, P.R. & Grant, B.R. 2006. Evolution of character displacement in Darwin’s finches. Science 313, 224-226 Guevara, J. & Avilés, L. 2007. Multiple sampling techniques confirm differences in insect size between low and high elevations that may influence levels of spider sociality. Ecology 88, 2015-2023. Guevara, J. & Avilés, L. In press. Elevational changes in insect and other arthropod composition at tropical latitudes: a comparison of multiple sampling methods and social spider diets. Insect Conserv. Biodiv. Gunnels, C.W., Dubrovsky, A. & Avalos, A. 2008. Social interactions as an ecological constraint in a eusocial insect. An. Behav. 75, 681-691. Hamilton, W.D. 1964. The genetical evolution of social behaviour. J. Theor. Biol. 7, 1-52. Hamilton, W.D. & May, R.M. 1977. Dispersal in stable habitats. Nature 269, 578-581. Harmon, L.J., Harmon, L.L., & Jones, C.G. 2007. Competition and community structure in diurnal arboreal geckos (genus Phelsuma) in the Indian Ocean. Oikos 116, 1863-1878. Hawkins, B.A., Cornell, H.V., & Hochberg, M.E. 1997. Predators, parasitoids, and pathogens as mortality agents in phytophagous insect populations. Ecology 78, 2145-2152. Heg, D., Bachar, Z., Brouwer, L., & Taborsky, M. 2004. Predation risk is an ecological constraint for helper dispersal in a cooperatively breeding cichlid. Proc. R. Soc. B 271, 2367-2374. Heinze, J. 1993. Life histories of subarctic ants. Arctic 46, 354-358. Heinze, J. 2008. Social Plasticity: Ecology, Genetics, and the Structure of Ant Societies. In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. 129-150. Berlin, Germany: Springer-Verlag. Henschel, J.R. 1998. Predation on social and solitary individuals of the spider Stegodyphus dumicola (Araneae, Eresidae) J. Arachnol. 26, 61-69. Hirata, M. & Higashi, S. 2008. Degree-day accumulation controlling allopatric and sympatric variations in the sociality of sweat bees, Lasioglossum (Evylaeus) baleicum (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 62, 1239-1247.  29  Hirata, M., Cronin, A.L., Kidokoro, M., & Azuma, N. 2005. Spatio-temporal variation of colony structure and eusociality level of the Japanese sweat bee Lasioglossum (Evylaeus) duplex (Hymenoptera: Halictidae). Ecol. Res. 20, 75-79. Hölldobler, B. & Wilson, E.O. 1990. The Ants. Cambridge, MA: Belknap Press. Hughes, D.P., Kathirithamby, J., & Beani, L. 2004. Prevalence of the parasite Strepsiptera in adult Polistes wasps: field collections and literature overview. Ethol. Ecol. Evol. 16, 363-375. Hutchinson, G.E. 1959 Homage to Santa Rosalia, or why are there so many kinds of animals? Am. Nat. 93, 145-159. Hunt, J.H. & Amdam, G.V. 2005. Bivoltinism as an antecendent to eusociality in the paper wasp genus Polistes. Science 308, 264-267. Janzen, D.H., Ataroff, M., Farinas, M., Reyes, S., Rincon, N., Soler, A., Soriano, P. & Vera, M. 1976. Changes in the arthropod community along an elevational transect in the Venezuelan Andes. Biotropica 8, 193-203. Jarvis, J.U.M., Oriain, M.J., Bennett, N.C. & Sherman, P.W. 1994. Mammalian eusociality – a family affair. Trends Ecol. Evol. 9, 47–51. Jeanne, RL. 1975. The adaptiveness of social wasp nest architecture. Q. Rev. Biol. 50, 267287. Jeanne, RL. 1979. A latitudinal gradient in rates of ant predation. Ecology 60, 1211-1224. Jeanne, RL. 1991. The swarm founding Polistinae. In The Social Biology of Wasps. (eds. K.G. Ross & R.W.), pp. 191-231. Ithaca, NY: Cornell University Press. Johnson, M.L. & Gaines, M.S. 1990. Evolution of dispersal - theoretical models and empirical tests using birds and mammals. Ann. Rev. Ecol. Syst. 21, 449-480. Jones, T.C., Riechert, S.E., Dalrymple, S.E. & Parker, P.G. 2007. Fostering model explains variation in levels of sociality in a spider system. Anim. Behav. 73, 195-204. Kaspari, M. & Vargo, E.L. 1995. Colony size as a buffer against seasonality: Bergmann’s Rule in Social Insects. Am. Nat. 145, 610-632. Korb, J. 2008. The ecology of social evolution in termites. In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. 151-174. Berlin, Germany: Springer-Verlag. Korb, J. & Heinze, J, 2008. Preface. In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. v-vii. Berlin, Germany: Springer-Verlag. 30  Korb, J. & Schmidinger, S. 2004. Help or disperse? Cooperation in termites influenced by food conditions. Behav. Ecol. Sociobiol. 56, 89-95. Kranz, B.D., Schwarz, M.P., Morris, D.C., & Crespi, B.J. 2002. Life history of Klodothrips ellobus and Oncothrips rodwayi: insight into the origin and loss of soldiers in gallinducing thrips. Ecol. Entomol. 27, 49-57. Kronfeld-Schor, N. & Dayan, T. 1999. The dietary basis for temporal partitioning: food habits of coexisting Acomys species. Oecologia 121, 123-128. Kukuk, P.F., Ward, S.A., & Jozwiak A. 1998. Mutualistic benefits generate an unequal distribution of risky activities among unrelated group members. Naturwissenschaften 85, 445-449. Le Galliard, J., Ferrière, R., & Dieckmann, U. 2005. Adaptive evolution of social traits: Origin, trajectories, and correlations of altruism and mobility. Am. Nat. 165, 206-224. Lehmann, L. & Perrin, N. 2002. Altruism, dispersal, and phenotype-matching kin recognition. Am. Nat. 159, 451-468. Lehmann, L., Perrin, N., & Rousset, F. 2006. Population demography and the evolution of helping behaviors. Evolution 60, 1137-1151. Lin, N. & Michener, C.D. 1972. Evolution of sociality in insects. Q. Rev. Bio. 47, 131-159. London, K.B. & Jeanne, R.L. 2003. Effects of colony size and stage of development on defense response by the swarm-founding wasp, Polybia occidentalis. Behav. Ecol. Sociobiol. 57, 445-456. MacArthur, R.H. 1958. Population ecology of some warblers of northeastern coniferous forests. Ecology 39, 599-619. MacArthur, R.H. & Levins, R. 1967. The limiting similarity, convergence and divergence of co-existing species. Am. Nat. 101, 377-385. Martin, S.J. 1992. Colony defence against ants in Vespa. Insectes Soc. 39, 99-111. Matsuura, K. & Nishida, T. 2001. Comparison of colony foundation success between sexual pairs and female asexual units in the termite Reticulitermes speratus (Isoptera: Rhinotermitidae). Popul. Ecol. 43, 119-124. McGlynn, T.P. 2007. Serial monodomy in ants: an antipredator strategy? Ecol. Entomol. 32, 621-626. 31  Michener, C.D. 2000. The Bees of the World. Baltimore, MD: Johns Hopkins University Press. Miyanaga, R., Maeta, Y., & Sakagami, S.F. 1999. Geographical variation of sociality and size-linked color patterns in Lasioglossum (Evylaeus) apristum (Vachal) in Japan (Hymenoptera, Halictidae). Insectes Soc. 46, 224-232. Muller, J. 2001. Invasion history and genetic population structure of riverine macro invertebrates. Zool. Anal. Complex Syst. 104, 346-355. Murphy, C.M & Breed, M.D. 2007. A predictive distribution map for the giant tropical ant, Paraponera clavata. J. Insect Sci. 7, insectscience.org/7.08 Murray, M.G. & Baird, D.R. 2008. Resource-ratio theory applied to large herbivores. Ecology 89, 1445-1456. Neill, D.A. & Jørgensen, P.M. 1999. Climate. In Catalogue of the vascular plants of Ecuador (eds. P. M. Jørgensen & S. León-Yánez), pp. 8-13. St. Louis, MO: Missouri Botanical Gardens. O’Donnell, S. & Jeanne, R.L. 1990. Notes on an army ant (Eciton burchelli Westwood) raid of a social wasp nest (Agelaia yepocapa Richards) in Costa Rica. J. Trop. Ecol. 6, 507509. O’Donnell, S., Kaspari, M. & Lattke, J. 2005. Extraordinary predation by the neotropical army ant Cheliomyrmex andicola: Implications for the evolution of the army ant syndrome. Biotropica 37, 706-709. Oster, G.F. & Wilson, E.O. 1978. Caste and Ecology in the Social Insects. Princeton, NJ: Princeton University Press. Packer, L. 1990. Solitary and eusocial nests in a population of Augochlorella striata (Provancher) (Hymenoptera; Halictidae) at the northern edge of its range. Behav. Ecol. Sociobiol. 27, 339-344. Parisod, C., Trippi, C., & Gallad, N. 2005. Genetic variability and founder effect in the pitcher plant Sarracenia purpurea (Sarraceniaceae) in populations introduced into Switzerland: From inbreeding to invasion. Ann. Bot. 95, 277-286. Peeters, C. 1997. Morphologically ‘primitive’ ants: comparative review of social characteristics, and the importance of queen-worker dimorphism. In The evolution of  32  social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 476-498. Cambridge, UK: Cambridge University Press. Peeters, C., Higashi, S., & Ito, F. 1991. Reproduction in ponerine ants without queens: Monogyny and exceptionally small colonies in the Australian Pachycondyla sublaevis. Ethol. Ecol. Evol. 3, 145-152. Perrin, N. & Lehmann, L. 2001. Is sociality driven by the costs of dispersal or the benefits of philopatry? A role for kin-discrimination mechanisms. Am. Nat. 158, 471-483. Pfennig, D.W. 1995. Absence of joint nesting advantage in desert seed harvester ants: evidence from a field experiment. Anim. Behav. 49, 567-575. Pianka, E. R. 1969. Sympatry of desert lizards (Ctenotus) in Western Australia. Ecology 50, 1012-1030. Picker, M.D., Hoffman, M.T., & Leverton, B. 2006. Density of Microhodotermes viator (Hodotermitidae) mounds in southern Africa in relation to rainfall and vegetative productivity gradients. J. Zool. 271, 37-44. Pike, N. & Foster, W.A. 2008. The ecology of altruism in a clonal insect. In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. 37-56. Berlin, Germany: SpringerVerlag. Pike, N., Whitfield, J.A., & Foster, W.A. 2007. Ecological correlates of sociality in Pemphigus aphids, with a partial phylogeny of the genus. BMC Evol. Biol. 7, 185 (112). Plateaux-Quénu, C., Plateaux, L., & Packer, L. 2000. Population-typical behaviours are retained when eusocial and non-eusocial forms of Evylaeus albipes (F.) (Hymenoptera: Halictidae) are reared simultaneously in the laboratory. Insectes Soc. 47, 263-270. Porter, E.E. & Hawkins, B.A. 2001. Latitudinal gradients in colony size for social insects: termites and ants show different patterns. Am. Nat. 157, 97-106. Powell, S. & Clark, E. 2004. Combat between large derived societies: A subterranean army ant established as a predator of mature leaf-cutting ant colonies. Insectes Soc. 51, 342351. Powers, K.S. & Avilés, L. 2007. The role of prey size and abundance in the geographic distribution of spider sociality. J. Anim. Ecol. 76, 995-1003.  33  Purcell, J. & Avilés, L. 2007. Smaller colonies and more solitary living mark higher elevation populations of a social spider. J. Anim. Ecol. 76, 590-597. Purcell, J. & Avilés, L. 2008. Gradients of precipitation and ant abundance may contribute to the altitudinal range limit of subsocial spiders: insights from a transplant experiment. Proc. R. Soc. B 275, 2617-2625. Queller, D.C. 1985. Kinship, reciprocity and synergism in the evolution of social-behavior. Nature 318, 366-367. Richards, M.H. 2000. Evidence for geographic variation in colony social organization of an obligately social sweat bee, Lasioglossum malachurum Kirby (Hymenoptera; Halictidae). Can. J. Zool. 78, 1259-1266. Richards, M.H., French, D., & Paxton, R.J. 2005. It’s good to be queen: classically eusocial colony structure and low worker fitness in an obligately social sweat bee. Mol. Ecol. 14, 4123-4133. Riechert, S.E., Roeloffs, R. & Echternacht, A.C. 1986. The ecology of the cooperative spider Agelena consociata in equatorial Africa (Araneae, Agelenidae). J. Arachnol. 14, 175191. Rissing, S.W., Johnson, R.A., & Martin, J.W. 2000. Colony founding behavior of some desert ants: geographic variation in metrosis. Psyche 103, 95-101. Ross, K.G. & Visscher, P.K. 1983. Reproductive plasticity in yellowjacket wasps: a polygynous, perennial colony of Vespula maculifrons. Psyche 90, 179-192. Rypstra, A.L. & Tirey, R.S. 1991. Prey size, prey perishability and group foraging in a social spider. Oecologia 86, 25-30. Sakagami, S.F. & Manakata, M. 1972. Distribution and bionomics of a transpalearctic eusocial halictine bee, Lasioglossum (Evylaeus) calceatum, in northern Japan, with reference to its solitary life cycle at high altitude. J. Fac. Sci., Hokkaido Univ. Ser. VI, Zool. 18, 411-439. Samson, D.A., Rickart, E.A. & Gonzales, P.C. 1997. Ant diversity and abundance along an elevational gradient in the Philippines. Biotropica 29, 349-363. Sanetra, M. & Crozier, R.H. 2002. Daughters inherit colonies from mothers in the ‘livingfossil’ ant Nothomyrmecia macrops. Naturwissenschaften 89, 71-74.  34  Schluter, D. 1996. Ecological speciation in postglacial fishes. Philos. T. Roy. Soc. B 351, 807-814. Schmid-Hempel, P. 1998. Parasites in Social Insects. Princeton, NJ: Princeton University Press. Schoener, T.W. 1968. The Anolis lizards of Bimini: Resource partitioning in a complex fauna. Ecology 49, 704-726. Schoener, T.W. 1974. Resource partitioning in ecological communities. Science 185, 27-39. Schwarz, M.P., Silberbauer, L.X., & Hurst, P.S. 1997 Intrinsic and extrinsic factors associated with social evolution in allodapine bees. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 476-498. Cambridge, UK: Cambridge University Press. Schwarz, M.P., Richards, M.H., & Danforth, B.N. 2007. Changing paradigms in insect social evolution: new insights from halictine and allodapine bees. Ann. Rev. Entomol. 52, 127150. Shykoff, J.A. & Schmid-Hempel, P. 1991. Parasites and the advantages of genetic variability within social insect colonies. Proc. R. Soc. B 243, 55-58. Silvertown, J. 2004. Plant coexistence and the niche. Trends Ecol. Evol. 19, 605-611. Smith, A.R., Wcislo, W.T., & O’Donnell, S. 2007. Survival and productivity benefits to social nesting in the sweat bee Megalopta genalis (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 61, 1111-1120. Soucy, S.L. & Danforth, B.N. 2002. Phylogeography of the socially polymorphic sweat bee Halictus rubicundus (Hymenoptera: Halictidae). Evolution 56, 330-341. Stern, D.L. & Foster, W.A. 1997. The evolution of sociality in aphids: a clone’s eye view. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 150-165. Cambridge, UK: Cambridge University Press. Strassman, J. 1989. Group colony foundation in Polistes annularis (Hymenoptera: Vespidae). Psyche 96, 223-236. Strohm, E. & Liebig, J. 2008. Why are so many bees but so few digger wasps social? The effect of provisioning mode and helper efficiency on the distribution of sociality among the Apoidea. In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. 109-128. Berlin, Germany: Springer-Verlag. 35  Taylor, P. D. 1992. Inclusive fitness in a homogeneous environment. Proc. R. Soc. B 249, 299-302. Tierney, S.M., Gonzalez-Ojeda, T. & Wcislo, W.T. 2008. Biology of a nocturnal bee, Megalopta atra (Hymenoptera: Halictidae, Augochlorini) from the Panamanian highlands. J. Nat. Hist. 42, 1841-1847. Trivers, R. L. 1971. The evolution of reciprocal altruism. Q. Rev. Bio. 46, 1-35. Vainstein, M.H., Silva, A.T.C., & Arenzon, J.J. 2007. Does mobility decrease cooperation? J. Theor. Biol. 244, 722-728. Wcislo, W.T. 1997a. Are behavioural classifications blinders to studying natural variation? In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 8-13 Cambridge, UK: Cambridge University Press. Wcislo, W.T. 1997b. Behavioral environments of sweat bees (Halictinae) in relation to variability in social organization. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 316-332. Cambridge, UK: Cambridge University Press. Wcislo, W.T. & Danforth, B.N. 1997. Secondarily solitary: the evolutionary loss of social behavior. Trends Ecol. Evol. 12, 468-474. Wcislo, W.T. & Schatz, B. 2003. Predator recognition and evasive behavior by sweat bees, Lasioglossum umbripenne (Hymenoptera: Halictidae), in response to predation by ants, Ectatomma ruidum (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 53, 182-189. Wilson, E. O. 1975. Sociobiology. Cambridge, MA: Belknap Press. Wilson, D.S., Pollock, G.B., & Dugatkin, L.A. 1992. Can altruism evolve in purely viscous populations? Evol. Ecol. 6, 331-341. Wojtusiak, J., Godzinska, E.J., & Dejean, A. 1995. Capture and retrieval of very large prey by workers of the African weaver ant, Oecophylla longinoda (Latreille 1802). Trop. Zool. 8, 309-318. Yanagawa, A., Yokohari, F., & Shimizu, S. 2008. Defense mechanism of the termite, Coptotermes formosanus Shiraki, to entomopathogenic fungi. J. Invertebr. Pathol. 97, 165-170.  36  Yanega, D. 1997. Demography and sociality in halictine bees (Hymenoptera: Halictidae). In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 293-315. Cambridge, UK: Cambridge University Press. Yip, E. C., Powers, K. C, & Avilés, L. 2008. Cooperative capture of large prey solves the problem of a declining surface area to volume ratio of large social spider colonies. Proc. Natl. Acad. Sci. U.S.A. 105, 11818-11822. Zammit, J., Hogendoorn, K. & Schwarz, M. P. 2008. Strong constraints to independent nesting in a facultatively social bee: quantifying the effects of enemies-at-the-nest. Insect. Soc. 55, 74-78. Zayed, A., & Packer, L. 2002. Genetic differentiation across a behavioural boundary in a primitively eusocial bee, Halictus poeyi Lepeletier (Hymenoptera, Halictidae). Insectes Soc. 49, 282-288.  37  CHAPTER 2: SMALLER COLONIES AND MORE SOLITARY LIVING MARK HIGHER ELEVATION POPULATIONS OF A SOCIAL SPIDER2  2.1 INTRODUCTION Social behaviour can allow an organism to occupy a niche that it may not otherwise be able to access (Wilson 1975; Slobodchikoff 1984; Avilés 1999). Naked mole rats and emperor penguins, for instance, are able to withstand notoriously harsh environmental conditions by either cooperatively searching for new food patches during short and unpredictable windows of time (Jarvis et al. 1994) or by huddling together to maintain warmth during the frigid Antarctic winter (Ancel et al. 1997; Gilbert et al. 2006). Likewise, by foraging in groups, tree-killing bark beetles are able to overcome the defences of live trees (Raffa & Berryman 1987) and mammalian carnivores and social spiders are able to catch prey larger than their body size (Macdonald 1983; Nentwig 1989; Rypstra & Tirey 1991). Sociality may also allow for increased defence from predators, as in spider mites (Mori & Saito 2005), swarm-founding wasps (London & Jeanne 2003), and sweat bees (Smith et al. 2003). On the other hand, sociality may be unnecessary or even unfavoured under some conditions, as reflected in both the inter- and intra-specific variability in social behaviour observed in a range of species in response to various climatic and habitat-related conditions (Lott 1984; Wcislo & Danforth 1997; Cronin 2001). Within halictine bee species, for instance, there is an increase in solitary nests with increasing latitude (Packer 1990) and altitude (Eickwort et al. 1996). In both cases, a shorter foraging season is thought to reduce the number of broods produced, thus decreasing the need for workers. Temperate conditions were also cited as a possible factor in reduced sociality across several closely related species of thrips in Australia (Kranz et al. 2002), while the production of fewer broods per generation was thought to cause an increased number of solitary foundress colonies in a southern California population of paper wasps (Liebert et al. 2005). In a number of bird  2  A version of this chapter has been published. Purcell, J. & Avilés, L. 2007. Smaller colonies and more solitary living mark higher elevation populations of a social spider. J. Anim. Ecol. 76, 590-597. 38  species, flocking behaviour is reduced in the absence of predators (Beauchamp 2004). In colonial orb-weaving (non cooperative) spiders (Uetz & Hieber 1997), Metepeira atascadero individuals occurring in a prey poor environment live solitarily or in small groups, with relatively large spacing among group members. Closely related M. incrassata living in a prey rich environment occur in much larger groups with less spacing between individuals (Uetz & Hodge 1990). All of these findings indicate that, even though sociality may allow expansion into new ecological niches, in some habitats it may be disadvantageous. Thus, sociality may restrict the geographic distribution of social species to areas where their particular social niche can be realized. The cooperative social spiders (also known as non territorial permanent social) form colonies that may contain hundreds to tens of thousands of spiders and are generally restricted to tropical latitudes and, within some regions, to lower elevations (reviewed in Avilés 1997). Related subsocial species, which form colonies containing a single mother and her brood, in contrast, extend into higher latitudes and elevations and may be absent from lowland tropical areas where social species thrive (Agnarsson 2006; L. Avilés unpublished). In the spider genus Anelosimus (Theridiidae), this pattern may be partly due to the availability of larger insect prey in the areas where social species are present relative to where they are absent (Powers 2004; Guevara & Avilés 2007; Yip et al. 2008). Large insects both require more individuals for their capture (e.g. Nentwig 1985) and are a perishable resource that can be shared (Rypstra & Tirey 1991). Other factors that may be correlated with both latitude and altitude are probably also involved in determining the geographical pattern in this and other genera (Avilés 1997), including rainfall (e.g. Riechert 1985; Riechert et al. 1986), temperature (e.g. Jones et al. 2007) and predation rates (e.g. Henschel 1998). Here we explore whether the pattern of decreasing sociality with increasing elevation found across species in the genus Anelosimus may also be applicable to differences between populations of a single species across an altitudinal gradient. The social Anelosimus eximius Simon 1891 is amenable to intraspecific investigation since it is found in both lowland rainforests and on the slope of the Andes up to about 1300 m (Levi 1963; Agnarsson 2006; this paper). If the conditions that allow this species to thrive in the lowland tropical rainforest change with elevation, we would expect A. eximius populations at the altitudinal limits of its range to be less social, or to exhibit characteristics more similar to subsocial 39  species at higher elevations. If this is the case, then we expect the upper elevation populations to have a greater proportion of nests with solitary females (prediction 1), smaller average colony sizes (prediction 2) and greater nest density (prediction 3) due to spiders dispersing from their natal nests more readily, and at smaller colony sizes than in lowland populations. Since the upper elevation populations are near the limit of the A. eximius species range, we also expect that this habitat may be marginal, which may lead to a lower overall population density than lowland sites (prediction 4). In order to test these predictions, we compared A. eximius populations at different elevations in various regions of Ecuador. Secondarily we compare forest edge and interior populations at the lowland sites as earlier studies have suggested that the forest interior may be a marginal habitat relative to forest edge for this species (Pasquet & Krafft 1989; Leborgne et al. 1994). This comparison then allows us to control for the separate effects of elevation and habitat type (edge vs. interior) in our elevation tests, as upper elevation populations only occurred at the forest edge (see Methods). 2.2 METHODS From May-June 2005 and May-July 2006, Anelosimus eximius nests were mapped in three transects within each of six sites in eastern Ecuador (Table 2.1). Additionally, data collected by LA (unpublished) from a seventh site in western Ecuador (Endesa, Table 2.1) were used for one of the tests. The sites studied ranged from lowland rainforest (sites < 700 m) to lower montane rainforest or “cloud forest” (sites > 700 m) (Neill 1999). The Cuyabeno Faunistic Reserve is a primary rainforest site traversed in some areas by trails as well as rivers navigable by motorized canoe. Three transects surveyed the forest interior (CF), while three others mapped the river edge by canoe (CR). The Jatun Sacha Reserve (JS) is a well conserved, though isolated, patch of primary rainforest. Three transects mapped nests in the rainforest interior. Endesa (EN) is an open (disturbed) habitat area in a lumber plantation 6-8 Km north of the Km 113 of the Quito – Pto. Quito highway. All colonies seen along a 1000 m stretch of road were surveyed. Puyo (PU) transects mapped nests along three secondary roads stemming from the highway connecting Puyo and Macas north of the Pastaza River, along a mid elevation ridge. The roadside alternates between areas of farmland and relatively intact cloud forest. In this region, we also explored a forest reserve, but did not 40  find any A. eximius nests along the trail system. The highway to Loreto (VL) follows the base of the Sumaco volcano, remaining around 1000 m above sea level for the first 50 km, where three transects were mapped. The roadside alternates between farmland and open roadsides with natural vegetation for the first 20-30 meters along the road, with cloud forest beyond the road cut. We explored several forested trails in this area, but did not find A. eximius. Macas (MA) transects surveyed nests along two secondary roads stemming from the Puyo-Macas road south of the Pastaza River. The roadside is primarily agricultural land, with shrubby vegetation in areas used for grazing livestock. At all six regular sites, nests were mapped in two random and one non-random transect. Random transects were 600 x 5 meters at all sites except along Cuyabeno River, where they were 1000 x 5 m. These transects were initiated by generating random points (in Excel 10.0) corresponding to distances along trails, roads, or rivers. The T-squared sampling method (Krebs 1999) was then used to find the nearest nest to the nest closest to the random point, thus ensuring that transects were initiated with a random nest. Starting at this random nest, forest transects continued in a fixed direction determined at the start by the direction of the trail near the nest, but diverging thereafter. At the edge sites (CR, VL, PU, and MA), transects followed the road or river edge, and the compass direction was noted at every bend. Non-random transects were placed in areas of known nest abundance. The length and width of the non-random transects varied at each site: 1000 x 5 m along the Cuyabeno River; 150 x 30 m in Cuyabeno Forest; 600 x 5 m and paralleling a trail in Jatun Sacha; 200 x 5 m in Puyo; 300 x 5 m in Macas. In Vía a Loreto, the transect spanned two stretches of roadside habitat interrupted by a heavily degraded construction area, for a total area of 200 x 5 m. In Endesa (EN), both sides of a 1000 m non-random stretch of road were mapped; all nests near the roadside (up to about 10 meters) were included. Additionally, we measured the extent of continuous vegetation at random points along each transect as the maximum length of leafy vegetation with gaps of less than 30 cm. All nests within each transect were measured and sketched (Fig. 2.1). Nest cross section areas (Fig. 2.1b), used to infer colony size (see below and Appendix 1), were then estimated by tracing irregular shapes into Canvas 3.5 (Deneba Systems Inc.) using a writing tablet or by calculating the area of an ellipse when shapes were regular. Nest contents were inspected for the presence of one or multiple females. For single female nests (with or 41  without their brood), the presence of additional subadult or adult females was ruled out by careful inspection. In a few ambiguous cases where careful inspection was not possible (mostly in VL), nests were excluded from the solitary female comparison. The regression of number of individuals to nest size (cross section area) was calculated in Jatun Sacha and both Cuyabeno sites by LA, and in Vía a Loreto by JP (Appendix 1). Entire nests of different sizes, or substantial, well-measured fractions of huge nests, were collected from each site, and the inhabitants were sorted by age and counted. Colony sizes were estimated using these regressions. The Vía a Loreto regression was applied to the Puyo and Macas nests, since these sites were similar in elevation and habitat. Nest density was calculated by summing the total number of nests per transect and dividing by the surveyed area. Individual density was determined using the nest size regressions (Appendix 1) to estimate total population size in each transect, divided by the surveyed area. Only random transects were included in the density analyses, since nonrandom transects targeted areas of high density. Non-parametric tests were used whenever our data did not conform to normal assumptions. JMP IN 5.1, R 1.12, Microsoft Excel 10.0, and the website http://faculty.vassar.edu/lowry/binomialX.html were used for statistical analyses. The Kolmogorov-Smirnov Test was used to compare the distribution of colony sizes (including solitary females, One-sided). Proportions of nests with solitary females were compared using a Binomial Exact One-Tailed Test. Only unambiguous nests where careful inspection of the nest occupants was possible were used in this analysis. In VL, the non-random transect nests were excluded from this analysis because they were not carefully viewed. Median colony sizes (excluding solitary females) were compared using the Wilcoxon Test. For this test, ambiguous nests were categorized as occupied by one or more females and included or excluded based on the colony size estimate from the nest size regression (Appendix 1). Continuous vegetation differences were compared using 95% Confidence Intervals. Population and nest densities were compared with an ANOVA, followed by independent contrasts to compare upper elevation versus lowland sites and edge versus forest interior sites in the lowlands, both planned comparisons. Alpha values were adjusted down using a Bonferroni correction to account for multiple comparisons.  42  2.3 RESULTS 2.3.1 Upper versus lower elevations Upper elevation sites all had colony size distributions consisting of significantly smaller colonies than at least one of the three lower elevation sites (Kolmogorov-Smirnov One-Sided Test, Table 2.2). A large percentage of these small upper elevation colonies contained a solitary female (Fig. 2.2). Thus Endesa (EN, 700 m), Puyo (PU, 900 m), Vía a Loreto (VL, 1000 m), and Macas (MA, 1200 m) each had significantly more solitary females than any of the lowland sites, except for MA versus Jatun Sacha (JS, 400 m, a forest interior site) (Binomial Exact One-Tailed Test, Table 2.2). When solitary females are excluded from the analyses, the median colony sizes in one of the three upper elevation sites remain smaller than those in two of the three lowland sites (Fig. 2.3). In this case, Vía a Loreto (VL, 1000 m) colonies remain significantly smaller than Cuyabeno River (CR, 225 m) and Cuyabeno Forest (CF, 240 m) colonies (Wilcoxon Test, Table 2.2). Puyo appeared to have substantially smaller colony sizes than CR and CF (Fig. 2.3), but the small sample of nests with more than one female at this site may be responsible for the lack of significance in this comparison. Although nest density was lower (Fig. 2.4a) and individual density higher (Fig. 2.4b) when comparing upper elevation sites to lowland sites, these differences were not significant whether we include lowland sites in both edge and interior habitats (ANOVA, Nest Density: F1,10 = 0.51, p = 0.49; Individual Density: F1,10 = 0.30, p = 0.59) or only edge populations (ANOVA, Nest Density: F1,6 = 0.44, p = 0.53; Individual Density: F1,6 = 0.68, p = 0.44). The power of these density tests, however, is low (adjusted power = 0.05 for an a = 0.05 given the observed variance and effect size). Upper elevation sites appeared more variable in both nest and individual density. 2.3.1 Forest edge versus forest interior Along the lowland forest edge (CR), colonies attained larger average sizes than those within the forest (Fig. 3). Vegetation along the edge tended to be significantly more continuous on the horizontal axis than vegetation inside the forest (length of continuous vegetation, mean +/- 95% Confidence Intervals: 26+/-12.16 m along the forest edge, 1.59+/0.63 m in the forest interior), a pattern that may relate to the differences in average nest sizes between these two habitats (see Discussion). Nest density was similar in the three lowland sites (ANOVA: F1,4 = 0.22, p = 0.66, Fig. 2.4a), but individual density was marginally higher 43  along the forest edge (CR) compared with the two sites (CF and JS) in the forest interior (ANOVA: F1,4 = 1.98, p = 0.23, Fig. 2.4b). 2.4 DISCUSSION The data support our expectation that the altitudinal pattern of sociality known to exist between Anelosimus species (Avilés et al. 2007) is mirrored within at least one generalist species. As predicted, colony size distributions are significantly more biased towards smaller colonies at the upper elevations (prediction 2; Fig. 2.3 and Table 2.2), to a large extent because solitary females occupy a much greater proportion of nests in the upper elevation sites (prediction 1; Fig. 2.2). When solitary females are excluded, median colony size in one out of three upper elevation sites remained significantly smaller than two lower elevation sites (prediction 2, Fig. 2.3). Contrary to our expectations, nest density was not significantly greater in upper elevation sites (prediction 3), although this result should be taken with caution given the low power of our tests (see Results). Nonetheless, two upper elevation transects did have much higher nest density than any lowland transects and the overall mean was in the expected direction (Fig. 2.4a). The finding that upper elevation colonies can sometimes grow to large sizes leads us to consider a couple of plausible explanations. First, social species at upper elevations may, in fact, occupy two social niches: one for colonies that tend to remain whole and grow very large and another for small colonies containing one or a few females. In this case, the first niche would mirror the typical social spider strategy from the lowland habitat, while the second would be similar to the lifestyle of subsocial spider species. Alternatively, the attractor (e.g., Avilés 1999) representing the growth trajectory of colonies at the upper elevation sites may have a long tail, thus allowing for the occasional large colony, even though overall it would correspond to smaller colony sizes. The preponderance of solitary females and small colonies at the upper elevation sites also allows for two possible, nonmutually exclusive, explanations: either more females are dispersing and establishing nests solitarily or more solitary dispersers are surviving. Both possibilities may occur simultaneously, with a greater dispersal tendency of females at these sites also partly explaining the paucity of very large colonies at these sites.  44  Unlike the altitudinal and latitudinal gradients of eusociality in bees, where length of the reproductive season may be responsible for lower levels of sociality at higher elevations and latitudes (Packer 1990; Eickwort et al. 1996), we cannot yet ascribe the observed patterns to any one particular factor. Among abiotic factors, temperature is clearly correlated with elevation. The pattern we observe, however, is opposite of that expected under the hypothesis that Jones et al. (2007) proposed to explain the increasing occurrence of multiple female nests at higher latitudes (and lower temperatures) in the subsocial spider Anelosimus studiosus in North America (see also Furey 1998). Jones et al. (2007) suggest that lower temperatures slow the developmental rate of the progeny, making the presence of additional caretakers advantageous in the event of the mother’s death. Rainfall, on the other hand, does not vary linearly with elevation (Table 2.1), but may nonetheless play a role in spider sociality (Riechert 1985; Riechert et al. 1986). In A. eximius, rainfall and rain intensity may, for instance, influence the frequency and success of colony foundation (Venticinque et al. 1993) and may partially explain the paucity of small nests in the lowlands. Among the biotic factors, we have found insect size to be correlated with elevation and latitude, with large insects being more common in the lowland rainforests occupied by social Anelosimus species relative to higher elevations or latitudes where social species are absent (Powers 2004; Guevara & Avilés 2007). We have argued that the presence of an abundant supply of large insects provides an incentive for cooperative prey capture and may be necessary to support large colony sizes. A preference for large prey has also been noted in the social species Stegodyphus mimosarum (Crouch & Lubin 2000). Predation rates may also be lower at higher elevations, as suggested by studies that have documented lower species richness and abundance in a range of potential spider predators at higher elevations, including insectivorous birds (Rahbek 1997) and some ants and spiders (Olson 1994). Henschel (1998) found that spiders living in larger groups were less vulnerable to predation by birds, araneophagus spiders, and ants compared with solitary spiders in the social Stegodyphus dumicola in Namibia. Riechert et al. (1986) found a similar pattern in Agelena consociata in Gabon. If predation rates decrease with increasing elevation, we might expect greater dispersal tendencies and/or greater survival of solitary dispersers in upper elevation spiders  45  Contrary to expectations (prediction 4), upper elevation sites did not appear marginal for this species based on overall spider density (Fig. 2.4b) and the occasional presence of very large nests. If these habitats are not marginal, then what factors prevent A. eximius from expanding to even higher elevations? One possibility is that increased competition with subsocial species for resources or space maintains the species boundary. Interspecific competition, however, has been notoriously difficult to demonstrate in spiders, suggesting that competition for resources may rarely be a limiting factor in these organisms (Wise 1993). Another possibility is that the species may not be plastic enough to lose its tendency to produce large colonies as the conditions favourable to large colony size decline with elevation. In either case, it would be interesting to investigate the relative role of plasticity and local adaptation on the dispersal tendencies and the propensity to establish nests solitarily in different populations of this species. We also found some differences between lowland forest interior and edge habitat, including smaller colonies (Fig. 2.3) and marginally lower individual density (Fig. 2.4b) inside the forest. Greater nest sizes along the forest edge (Fig. 2.3) may result from the more continuous vegetation substrate in this habitat (see Results), while forest interior nests would need to break up once they reach the limit of their current plant substrate. This may also account for the greater individual density in the edge habitat (Fig. 2.4b), since the per capita rate of growth may not reach maximum levels in colonies constrained to smaller sizes (Avilés & Tufiño 1998). The alternative possibility of a greater availability of insects or greater insect sizes accounting for larger colony sizes at the forest edge has been rejected by recent studies that found no differences in these two aspects across the two habitat types (Powers 2004; Guevara & Avilés 2007; Yip et al. 2008). While these data are not sufficient to support or refute the suggestion that forest interior habitat may be less favourable for A. eximius (Pasquet & Krafft 1989; Leborgne et al. 1994), the pattern seems to be stronger at higher latitudes (Pasquet & Krafft 1989; Leborgne et al. 1994) and higher elevations (this study) where A. eximius colonies appear absent from the forest interior. The relatively broad range of habitats and elevations occupied by A. eximius allows us to investigate the efficacy of its social strategy across different environments. Contrary to our expectation, we found that the upper elevation habitat supports a robust population. While the fact that all upper elevation sites were in edge habitat may contribute to this 46  finding, we suggest that the mechanism leading to higher population density may differ between lowland and upper elevation sites. Whereas an abundance of large nests predominate the population along the lowland river edge, upper elevation populations seem to be split more evenly between large and small nests, suggesting a more important role for small nests. Our data suggest that the sites on the altitudinal range edge of this species may host a phenotype with some characteristics resembling the subsocial strategy. As such, this area offers an excellent opportunity to investigate the costs and benefits of several socially intermediate traits. For instance, it would be interesting to explore the fitness of solitary females in this environment, which we might expect to be greater than the fitness of solitary females at lower elevations. Also, we expect dispersal to occur more frequently and at smaller colony sizes in this habitat. Thus it may be interesting to determine how the shape of the per capita reproductive success function (e.g., Avilés & Tufiño 1998) relates to when and how frequently dispersal takes place. Finally, while these observations do not allow us to distinguish the relative influences of phenotypic plasticity and local adaptation in shaping these differences, they do provide an interesting baseline for future investigations.  47  Table 2.1: Site locations and descriptions. Average annual temperature and rainfall estimates taken from the following websites: Cuyabeno: http://pin.primate.wisc.edu/factsheets/entry/pygmy_marmoset, Jatun Sacha: www.jatunsacha.org, Endesa: www.nmnh.si.edu/botany/projects/cpd/sa/sa40.htm, Puyo: www.worldclimate.com, Vía a Loreto: www.amazoniaecuador.net, Macas: www.unep-wcmc.org. Location  Province  Cuyabeno Faunistic Reserve: River Cuyabeno Faunistic Reserve: Forest Jatun Sacha Reserve: Forest Endesa: Road  Sucumbios  Puyo Area  Pastaza  Vía a Loreto: Road  Napo  Macas Area  Morona-Santiago  Sucumbios Napo Pichincha  Coordinates  Elevation  0.028º S, 76.294º W 0.032º S, 76.321º W 1.060º S, 77.617º W 0.08º N, 79.07º W 1.5º S, 77.9º W 0.703º S, 77.736º W 2.3º S, 78.1º W  225 m 240 m 400 m 700 m 900 m 1000 m 1200 m  Habitat Type Forest Edge Forest Interior Forest Interior Forest Edge Forest Edge Forest Edge Forest Edge  Abbreviation CR  Avg. Annual Temperature 22-29º C  Avg. Annual Rainfall 3000 mm  CF  22-29º C  3000 mm  JS  25º C  5000 mm  EN  20° C  2000 mm  PU  20.9º C  4538 mm  VL  18º C  1800 mm  MA  20º C  2414 mm  48  Table 2.2: Tests of the predictions that (1) colonies from higher elevations are smaller than those at lower elevations and that colonies from forest interior are smaller than those from edge sites at the same elevation and (2) that solitary females occupy a greater proportion of nests in the higher elevation sites. Compared are (a) the entire colony size distributions (One-sided KolmogorovSmirnov test), (b) percent solitary females (Binomial One-tailed Test), and (c) median size of colonies containing more than one female (Wilkoxon Test) among A. eximius populations at various locations (see Table 2.1) . Alpha values for each test adjusted down using a Bonferroni correction to account for the multiple comparisons. *Note that for % solitary females, PU vs. MA and VL vs. MA were significantly different from one another (p<0.0001 for both comparisons) in a two-tailed test.  49  Location  Comparison  CR  CF JS EN PU VL MA JS EN PU VL MA EN PU VL MA PU VL MA VL MA MA  CF  JS  EN  PU VL  a. Colony Size Distribution  b. Solitary Females  One-sided KolmogorovSmirnov test α = 0.0033 D P-value 0.33 0.057 0.55 0.01 0.71 <0.0001 0.65 <0.0001 0.47 0.0027 0.54 0.0097 0.59 <0.0001 0.6 <0.0001 0.31 0.053 0.27 0.34 0.22 0.4 0.061 0.94 0.07 0.82 0.03 0.97 0.014 0.99  One-tailed Binomial Exact test α = 0.0024 P-value 0.63 0.38 <0.0001 <0.0001 <0.0001 <0.0001 0.31 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0031 0.0004 0.0061 0.77 0.67 1.00* 1.00*  c. Colony Size Without Solitary Females  Wilcoxon Test α = 0.0033 2 χ P-value 4.22 0.04 6.35 0.012 5.94 0.015 17.23 <0.0001 3.76 0.052 3.48 0.062 1.57 0.21 11.86 0.0006 0.26 0.61 1.93 0.17 0.26 0.61 2.82 0.093 2.35 0.13 0.14 0.71 5.1 0.024  50  Figure 2.1: Diagram of measurements taken for each nest within each transect on a side view (a) and a cross section view (b).  51  Figure 2.2: In general, upper elevation sites have a significantly higher proportion of solitary females than any of the lowland sites. The following letter coding indicates which sites are significantly different from one another based on a pairwise Binomial Exact Test (α=0.0024, table 2.2): CR: A, CF: A, JS: AB, EN: CD, PU: E, VL: DE, and MA: BC.  52  Figure 2.3: Boxplots showing variation in colony size distributions (log number of adult and subadult females) between lowland versus upper elevation sites as well as lowland edge (CR) and interior (CF and JS) sites. The median (black line), quantiles (box), most extreme value up to 1.5 times the interquartile range (whiskers), and outliers (circles) are shown for the overall colony size distribution. The dotted grey line represents the median of the distribution that excludes solitary females. See Table 2.2 and text for results of significance tests.  53  Figure 2.4: An analysis of variance (ANOVA) comparison of nest (a) and individual (b) density showed that there is no significant difference between upper elevation and lowland habitat. The grey line shows the overall mean; mean and 95% confidence intervals are also shown for each column. The following shapes represent density calculations from each random transect in each site: stars for Cuyabeno River (CR), open triangles for Cuyabeno Forest (CF), hexagons for Jatun Sacha (JS), squares for Puyo (PU), circles for Vía a Loreto (VL), and closed triangles for Macas (MA). 54  2.5 REFERENCES Agnarsson, I. 2006. A revision of the New World eximius lineage of Anelosimus (Araneae, Theridiidae) and a phylogenetic analysis using worldwide exemplars. Zool. J. Linn. Soc. 146, 453-493. Ancel, A., Visser, H., Handrich, Y., Masman, D., & Le Maho, Y. 1997. Energy saving in huddling penguins. Nature 385, 304-305. Avilés, L. 1997. Causes and consequences of cooperation and permanent sociality in spiders. In The evolution of social behaviour in insects and arachnids. (eds. J.C. Choe & B. Crespi), pp. 476-498. Cambridge, UK: Cambridge University Press. Avilés, L. 1999. Cooperation and non-linear dynamics: An ecological perspective on the evolution of sociality. Evol. Ecol. Res. 1, 459-477. Avilés, L. & Tufiño, P. 1998. Colony size and individual fitness in the social spider Anelosimus eximius. Am. Nat. 152, 403-418. Avilés, L., Agnarsson, I., Salazar, P., Purcell, L., Iturralde, G., Yip, E., Powers, K.S. & Bukowski, T. 2007. Altitudinal patterns of spider sociality and the biology of a new mid-elevation social Anelosimus species in Ecuador. Am. Nat. 170, 783-792. Beauchamp, G. 2004. Reduced flocking by birds on islands with relaxed predation. Proc. R. Soc. B 271, 1039-1042. Cronin, A.L. 2001. Social flexibility in a primitively social allodapine bee (Hymenoptera : Apidae): results of a translocation experiment. Oikos 94, 337-343. Crouch, T. E, & Y. Lubin. 2000. Effects of climate and prey availability on foraging in a social spider, Stegodyphus mimosarum (Araneae, Eresidae). J. Arachnol. 28, 158-168. Eickwort, G.C., Eickwort, J.M., Gordon, J., & Eickwort, M.A. 1996. Solitary behavior in a high altitude population of the social sweat bee Halictus rubicundus (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 38, 227-233. Furey, F.E. 1998 Two cooperatively social populations of the theridiid spider Anelosimus studiosus in a temperate region. An. Behav. 55, 727-735. Gilbert, C., Robertson, G., Le Maho, Y., Naito, Y., & Ancel, A. 2006. Huddling behavior in emperor penguins: Dynamics of huddling. Physiol. Behav. 88, 479-488.  55  Guevara, J. & Avilés, L. 2007. Multiple sampling techniques confirm differences in insect size between low and high elevations that may influence levels of spider sociality. Ecology 88, 2015-2023. Henschel, J.R. 1998. Predation on social and solitary individuals of the spider Stegodyphus dumicola (Araneae, Eresidae). J. Arachnol. 26, 61-69. Jarvis, J.U.M., Oriain, M.J., Bennett, N.C., & Sherman, P.W. 1994. Mammalian eusociality a family affair. Trends Ecol. Evol. 9, 47-51. Jones, T.C., Riechert, S.E., Dalrymple, S.E. & Parker, P.G. 2007. Fostering model explains variation in levels of sociality in a spider system. Anim. Behav. 73, 195-204. Kranz, B.D., Schwarz, M.P., Morris, D.C., & Crespi, B.J. 2002. Life history of Kladothrips ellobus and Oncothrips rodwayi: insight into the origin and loss of soldiers in gallinducing thrips. Ecol. Entomol. 27, 49-57. Krebs, C.J. 1999. Ecological Methodology, 2nd edn. Menlo Park, CA: Benjamin/Cummings. Leborgne, R., Krafft, B., & Pasquet, A. 1994. Experimental study of foundation and development of Anelosimus eximius colonies in the tropical forest of French-Guiana. Insectes Soc. 41, 179-189. Levi, H.W. 1963. The American spiders of the genus Anelosimus (Araneae, Theridiidae). T. Am. Microsc. Soc. 82, 30-48. Liebert, A.E., Nonacs, P., & Wayne, R.K. 2005. Solitary nesting and reproductive success in the paper wasp Polistes aurifer. Behav. Ecol. Sociobiol. 57, 445-456. London, K.B. & Jeanne, R.L. 2003. Effects of colony size and stage of development on defense response by the swarm-founding wasp Polybia occidentalis. Behav. Ecol. Sociobiol. 54, 539-546. Lott, D.F. 1984. Intraspecific variation in the social-systems of wild vertebrates. Behaviour 88, 266-325. Macdonald, D.W. 1983. The ecology of carnivore social-behavior. Nature 301, 379-384. Mori, K. & Saito, Y. 2005. Variation in social behavior within a spider mite genus, Stigmaeopsis (Acari : Tetranychidae). Behav. Ecol. 16, 232-238. Neill, D.A. 1999. Vegetation. In Catalogue of the vascular plants of Ecuador. (eds. P. M. Jørgensen and S. León-Yánez), pp.13-25. St. Louis, MO: Monogr. Syst. Bot. MO. Bot. Gard. 75. 56  Nentwig, W. 1985. Social spiders catch larger prey: a study of Anelosimus eximius (Araneae: Theridiidae). Behav. Ecol. Sociobiol. 17,79-85. Nentwig, W. 1989. Seasonal and taxonomic aspects of the size of arthropods in the tropics and its possible influence on size-selectivity in the prey of a tropical spider community. Oecologia 78, 35-40. Olson, D.M. 1994. The distribution of leaf-litter invertebrates along a neotropical altitudinal gradient. J. Trop. Ecol. 10, 129-150. Packer, L. 1990. Solitary and eusocial nests in a population of Augochlorella striata (Provancher) (Hymenoptera, Halictidae) at the northern edge of its range. Behav. Ecol. Sociobiol. 27, 339-344. Pasquet, A. & Krafft, B. 1989. Colony distribution of the social spider Anelosimus eximius (Araneae, Theridiidae) in French Guiana. Insectes Soc. 36, 173-182. Powers, K.S. 2004. Prey abundance and the evolution of sociality in Anelosimus (Araneae, Theridiidae). Tuscon, AZ: Ph.D. Thesis, University of Arizona. Raffa, K.F. & Berryman, A.A. 1987. Interacting selective pressures in Conifer-Bark Beetle systems - a basis for reciprocal adaptations. Am. Nat. 129, 234-262. Rahbek, C. 1997 The relationship among area, elevation, and regional species richness in neotropical birds. Am. Nat. 149, 875-902. Riechert, S.E. 1985. Why do some spiders cooperate? Agelena consociata, a case study. Fla. Entomol. 68, 105-116. Riechert, S.E., Roeloffs, R., & Echternacht, A.C. 1986. The ecology of the cooperative spider Agelena consociata in equatorial Africa (Araneae, Agelenidae). J. Arachnol. 14, 175191. Rypstra, A.L. & Tirey, R.S. 1991. Prey size, prey perishability and group foraging in a social spider. Oecologia 86, 25-30. Simon, E. 1891. Observations biologiques sur les Arachnides. I. Araignées sociables. Voyage de M. E. Simon au Venezuela. Ann. Soc. Entomol. Fr. 60, 5-14. Slobodchikoff, C.N. 1984. Resources and the evolution of social behavior. In A New Ecology: Novel Approaches to Interactive Systems (eds P.W. Price, C. N. Slobodchikoff, & W. S. Gaud). New York, NY: John Wiley & Sons.  57  Smith, A.R., Wcislo, W.T., & O'Donnell, S. 2003. Assured fitness returns favor sociality in a mass-provisioning sweat bee, Megalopta genalis (Hymenoptera : Halictidae). Behav. Ecol. Sociobiol. 54, 14-21. Uetz, G. & Hieber, C.S, 1997. Colonial web-building spiders: balancing the costs and benefits of group living. In The evolution of social behavior in insects and arachnids (eds J.C. Choe & B.J. Crespi), pp. 476-498. Cambridge, UK: Cambridge University Press. Uetz, G. W. & Hodge, M.A. 1990. Influence of habitat and prey availability on spatial organization and behavior of colonial web-building spiders. Natl. Geogr. Res. 6, 22-40. Venticinque, E.M., Fowler, H.G., & Silva, C.A. 1993. Modes and frequencies of colonization and its relation to extinctions, habitat and seasonality in the social spider Anelosimus eximius in the Amazon (Araneidae: Theridiidae). Psyche, 100, 35-41. Wcislo, W.T. & Danforth, B.N. 1997. Secondarily solitary: the evolutionary loss of social behavior. Trends Ecol. Evol. 12, 468-474. Wilson, E.O. 1975 Sociobiology. Cambridge, MA: The Belknap Press. Wise, D.H. 1993. Spiders in Ecological Webs. Cambridge, UK: Cambridge University Press. Yip, E. C., Powers, K. C, & Avilés, L. 2008. Cooperative capture of large prey solves the problem of a declining surface area to volume ratio of large social spider colonies. Proc. Natl. Acad. Sci. U.S.A. 105, 11818-11822.  58  CHAPTER 3: GRADIENTS OF PRECIPITATION AND ANT ABUNDANCE MAY CONTRIBUTE TO THE ALTITUDINAL RANGE LIMIT OF SOCIAL SPIDERS: INSIGHTS FROM A TRANSPLANT EXPERIMENT3 3.1 INTRODUCTION Social traits can change the way an organism interacts with its environment, often by allowing groups to access a niche not available to solitary individuals (Wilson 1975; Slobodchikoff 1984; Avilés 1999). As a result, social behaviours may differ across environmental gradients, both within species (e.g. Eickwort et al. 1996; Liebert et al. 2005; Purcell & Avilés 2007) and among related species (e.g. Mori & Saito 2005; Avilés et al. 2007). By investigating factors influencing the distribution of social organisms across such environmental gradients, we simultaneously seek to understand the factors maintaining species ranges and the selective forces favouring particular social phenotypes in different environments. Theoretical studies have demonstrated the importance of both environmental and intrinsic factors in delimiting the geographical range of a species (e.g. Gaston 2003; Holt et al. 2005). Kirkpatrick and Barton (1997), for instance, modelled a hypothetical species with simple quantitative fitness determination along an environmental gradient. Populations in the centre of the range were well adapted to the local environment, but those on the range edge performed poorly, as genotypes from the range centre swamped locally favourable mutations through migration. Empirical studies are still needed, however, to test the hypotheses generated by this and other theoretical models of range boundaries in real systems (Bridle & Vines 2007). Transplant experiments can provide insight into the role of both intrinsic and environmental factors at a range boundary by facilitating a direct comparison of life history traits in native and foreign habitats with measurable environmental differences (e.g. Riechert & Hall 2000; Angert & Schemske 2005; Geber & Eckhart 2005). Here, we use the transplant experimental approach to investigate factors, both intrinsic and extrinsic (environmental), 3  A version of this chapter has been published. Purcell, J. & Avilés, L. 2008. Gradients of precipitation and ant abundance may contribute to the altitudinal range limit of subsocial spiders: insights from a transplant experiment. Proc. R. Soc. B 275, 2617-2625. 59  that may prevent subsocial Anelosimus spiders from colonizing the lowland tropical rainforest in Ecuador (e.g. Avilés et al. 2007). We focus on intrinsic aspects of the biology of these organisms that relate to their social system, such as colony size and dispersal tendencies, although we cannot exclude the possibility that physiological factors (e.g., temperature and humidity preferences) may also play a role in determining the observed patterns. Several studies have noted that social (non-territorial permanently social) spiders are concentrated in tropical regions of the world, and occupy only a subset of the habitats utilized by members of their phylogenetic lineages (Avilés 1997; Avilés et al. 2007). Just as striking, but less frequently noted, is the absence of related subsocial or solitary species from some areas where social species are present. This pattern is especially clear in the genus Anelosimus where subsocial species are absent from lowland tropical rainforest habitats in eastern Ecuador (Agnarsson 2006; Avilés et al. 2007). In subsocial species, colonies typically contain the offspring of a single female, usually up to several dozen individuals, which cooperate early in their life cycle, but disperse prior to sexual maturity. Social species, thought to be derived from subsocial-like ancestors (Avilés 1997; Agnarsson 2006; Lubin & Bilde 2007), cooperate in prey capture, nest maintenance, and brood rearing and extend their group living period for multiple generations, with some colonies containing tens of thousands of spiders. In discussing the mechanisms that may lead to the distinct geographical distribution of social and subsocial Anelosimus species, Avilés et al. (2007) suggested that two separate patterns need to be addressed: first, the absence of social spiders from higher altitudes and higher latitudes; and second, the absence of subsocial spiders from the lowland tropical rainforest. The former may result from an insufficient abundance of large insect prey outside of the lowland rainforest to meet the nutritional requirements of large spider groups (Guevara & Avilés 2007; Powers & Avilés 2007; Yip et al. 2008; see also Uetz & Hodge 1990; Rypstra & Tiery 1991). With both large and small prey available in the lowland rainforest, however, this mechanism cannot explain the absence of subsocial species from this habitat. Using a transplant experiment, here we test the hypotheses, based on Avilés et al. (2007), that subsocial species may instead be excluded from the lowland rainforest by factors that increase mortality during dispersal (H1) or lead to the complete failure of nests with 60  solitary foundresses and their brood (H2). We also explore the possibility that intrinsic benefits of remaining in the natal group may disfavour dispersal in this habitat (H3). To test these hypotheses, we transplanted nests of a subsocial spider from its native upper elevation cloud forest habitat to lower montane and lowland rainforest. We explore the following specific predictions derived from these hypotheses. If subsocial species have not simply failed to disperse to lowland rainforest, but rather are maladapted to this habitat, we expect (i) increased extinction of small groups and (ii) higher individual mortality as distance to the rainforest decreases; moreover, if spiders in the lowland rainforest are better off living in groups than solitarily, we would expect (iii) decreased dispersal tendencies of individuals (i.e., fewer dispersers) and (iv) decreased survival of dispersers at lower elevations compared to the cloud forest control habitat. We simultaneously explore potential candidate factors that may contribute to the failure of solitary dispersers or small groups in the lowland rainforest, namely rainfall intensity (H4) and predation (H5). Anelosimus spiders build three-dimensional webs requiring active maintenance and a large volume of thread, which would be costly to repair when damaged by intense rain showers or falling branches (Avilés 1997). Riechert et al. (1986) observed that larger groups of the social spider Agelena consociata recovered more rapidly from extreme weather disturbance events than small groups. In subsocial species, with fewer individuals present to maintain and rebuild the web, nest damage resulting from intense rain could render some habitats inhospitable. We test this hypothesis by sheltering half of the nests transplanted in the lowland rainforest from rain. We expect (i) greater survival rates and (ii) larger nests in the sheltered compared to the exposed treatment. We also shelter small groups of the native social species and test the same predictions. Predation was suggested as a mechanism selecting for sociality in Stegodyphus dumicola (Henschel 1998) and colonial aggregations in Metepeira incrassata spiders (Uetz et al. 2002), and in other social organisms (London & Jeanne 2003; Smith et al. 2003; Mori & Saito 2005). We hypothesize that small subsocial groups may be excluded from areas where potential predators are abundant, because they are not able to repulse or evade predators as effectively as large social groups. We test our prediction that predators may be more abundant in the lower elevation habitats by comparing the presence/absence of two likely spider predators,  61  ants and jumping spiders (J. Purcell 2006, personal observation), in nests at all three transplant habitats. 3.2 METHODS We transplanted colonies of the subsocial spider species Anelosimus baeza Agnarsson (2006) within the centre (2100 m elevation) and to the edge of its native habitat (1000 m) and to a low elevation habitat beyond the species range boundary (400 m). Along the Andes’ eastern slope, elevation appears to be correlated with several environmental factors that may be relevant in this system, including rainfall intensity and potential predator abundance (Guevara & Avilés in press), so we use elevation as a proxy to describe the position of each habitat along the transition from cloud forest to lowland tropical rainforest. 3.2.1 Species description Anelosimus baeza, a subsocial spider, is widespread in South America, extending from Panama to Peru and Brazil (Agnarsson 2006). Like other subsocial spiders, A. baeza lives in single-family groups— a female and her brood or groups of pre-reproductive siblings— with individuals dispersing to live solitarily in their late juvenile to young adult stages (Powers 2004). On the eastern slopes of the Ecuadorian Andes, this species occurs primarily in cloud forest habitat (900 - 2500 m above sea level), with the occasional nest found down to ~ 600 m elevation. It is absent from elevations below 600 m in the tropical rainforest (Avilés et al. 2007). We used this species for our studies because it has a body size comparable to that of Anelosimus eximius Keyserling (1884), a common social species in the lowland rainforest (adult female total length: 4.0 cm for A. baeza vs. 4.6 cm for A. eximius, Agnarsson 2006) and because, among the Ecuadorian subsocial species, it occurs in habitats closest to the rainforest (Avilés et al. 2007). The latter reduces the possibility of physiological maladaptation to the transplant environments. 3.2.2 Habitat descriptions Transplanted groups were placed at three locations within 65 km of one another along an altitudinal gradient on the eastern slope of the Andes in Ecuador’s Napo province: at its native cloud forest habitat near the Yanayacu Biological Station (0.061° S, 77.893° W, 20002200 m), at an intermediate elevation lower montane rainforest area adjacent to the Hollin River near Sumaco National Park (0.695° S, 77.731° W, 950-1100 m), and at lowland 62  tropical rainforest habitat located in the Jatun Sacha Reserve along the Napo River, east of Tena (1.072° S, 77.617° W, 380-450 m). Both the upper and intermediate elevation habitats have been termed ‘lower montane rainforest’ or ‘cloud forest’ habitats in the literature (Neill 1999a). Here, we distinguish between them by referring to the upper elevation habitat as cloud forest and the intermediate habitat as lower montane rainforest, because the plant composition differs greatly between these two habitats (see Neill 1999a, b for floral and climatic characteristics of all three habitats). The focal subsocial species A. baeza occurred naturally in open and disturbed habitats in the two upper elevation locations. The social A. eximius, the subsocial Anelosimus elegans, and an undescribed subsocial species were also present in the lower montane habitat. Anelosimus eximius and two additional social species, Anelosimus domingo and Anelosimus rupununi, occurred naturally in the lowland rainforest (Avilés et al. 2007). At all three transplant locations, we measured rainfall intensity by collecting rain for 30 minutes during heavy showers with a plastic precipitation gauge placed in an open field (at least three measurements per habitat). We also estimated average daily rainfall by collecting rain for 24-hour periods in a second plastic gauge (at least seven measurements per habitat, throughout the course of the study). 3.2.3 Transplant Methods For transplantation, we collected relatively large A. baeza nests (median 50, range 20325 spiders) from seven distinct sites—separated by 500 m or more—near the Yanayacu Biological Station. We dissected each nest and sorted the occupants by age and sex. At the time of collection, nests in the area contained primarily late-instar juveniles to young adults. We transplanted only female spiders (late-instar juvenile to adult) in order to avoid inadvertently introducing the species into foreign habitat. We placed the spiders from each transplant group together in containers with a small plastic plant to encourage initial nest construction and to provide identical structures to support the nests at all habitats. All groups built webbing on this artificial substrate. The number of spiders transplanted per colony varied from 12 to 100 individuals. We matched colonies by size, and randomly assigned them to treatments to ensure an even distribution of group sizes in each treatment. Three large groups were divided into two or three transplant groups and placed in separate treatments to avoid pseudoreplication. 63  Within each transplant location, we selected four to five sites that were at least 1 km apart, except for two sites at the intermediate elevation that were separated by 500 m on a steep hillside. We transplanted 20 nests to five sites within the source habitat (2100 m); 19 nests to four sites at the intermediate habitat (1000 m); and 50 nests to 5 sites in the lowlands (400 m; Appendix 2). At each of the sites, we placed transplanted nests at least 10 m from each other and at least 5 m from any natural Anelosimus nests. Within the source habitat, we placed transplanted nests on their preferred plant substrate, a Baccharis and a Monochaelum species (Asteraceae and Melastomataceae, respectively; J. Purcell 2005, personal observations). At the other two locations, where these plant species were not common, we sought sites that resembled the characteristics (leaf size, shape and texture, branch architecture, and the openness of the habitat) of the preferred nest locations in the source habitat as much as possible. At all sites we avoided areas where ants were already present on the substrate. To transplant nests, we attached the artificial plant with the incipient nest to the selected branch and enclosed the branch in a mesh tube. We removed the mesh tubes after at least 4 days, when new webbing was observed within the enclosure (the spiders were fed insect prey every 2-3 days while enclosed). After removing the mesh tubes, we closely monitored the incipient nests to prevent any premature dispersal attempts. Although some spiders tried to disperse immediately after the mesh tubes were removed, at the two upper elevation sites most spiders soon settled in their newly transplanted nest. In the lowlands, on the other hand, we were forced to suspend the mesh removal process after it became clear that the spiders would not remain at their prescribed location. We were nonetheless able to obtain partial data for the rain exclosure experiment from the 35 nests that were recollected, without the mesh tubes having been removed, approximately 15 days after the start of the experiment, and from some of the 15 nests for which the mesh removal process was completed (nine nests for number of spiders remaining and one for amount of webbing built; see results and Appendix 2). We then monitored all nests that became established at least once per week for up to two months (Appendix 2), noting the size and condition of the nest, approximate number of occupants and of dead spiders (carcasses), presence within the nests of potential predators (including ants or jumping spiders), and occurrence of dispersal, as evidenced by the presence of newly founded nests (here termed “propagules”) within an 64  approximately 5 m radius of the surveyed nest. Previous studies have demonstrated that the majority of subsocial Anelosimus propagules, easily distinguishable by the fresh webbing and clear appearance, become established within 5 m of the source nest (Avilés & Gelsey 1998; Powers & Avilés 2003). Therefore we feel confident in our ability to detect the presence or absence of dispersal. All propagules were flagged and monitored with the same criteria as the transplanted nests for the rest of the experiment. When no propagules were observed, we ascribed the disappearance of spiders to predation or mortality from other causes. At the end of the experimental period, all nests, both transplants and propagules, were recollected and the spiders sorted and counted. 3.2.4 Rain exclosure In order to test the effect of intense rainfall on the survival and web-building ability of the transplanted nests in the lowland rainforest habitat (400 m), we sheltered 25 of the 50 transplanted A. baeza nests with semipermeable tarpaulin sheets (approx. 150 x 150 cm) placed approximately 1 m above the nest. We repeated this test using small groups (100 spiders) of the locally occurring social species, A. eximius. We established 22 such A. eximius groups at three sites chosen to match the preferred location and substrate characteristics of the species, with half the nests at each location sheltered from the rain and half exposed. For the experimental period, the colony dynamics of a small social group should resemble a subsocial group prior to dispersal. We monitored the nests as above, additionally noting the size of the nest and the number of webbing threads present at the end of the experiment. 3.2.5 Analysis We used hierarchical linear models (HLMs) to investigate differences between our treatment and control groups, taking into account the variation across different transplant sites within a habitat. For presence-absence and non-normally distributed outcome data, we used a generalized linear mixed model via the GLIMMIX macro in SAS v. 9.1 (SAS Institute, Cary, NC) to compare habitat as a fixed effect, with a random intercept clustered around transplant sites to control for within habitat variation. For normally distributed data, we used the MIXED procedure in SAS, with a similar model structure (Singer 1998). For the comparisons of colony survival and propagule survival, we used the survival actuarial analysis in StatView v. 4.1 (Abacus Concepts, Inc., Berkeley, CA), which allows variables to 65  be censored. Censoring a variable enables differentiation between samples that persist until the end of the experimental period versus those that perish from natural causes. We tested both the effect of habitat, and of site within habitat, on colony survival. The remaining tests were performed using JMP v. 5.1 (SAS Institute, Cary, NC). For the comparisons of predator presence in nests across all three habitats, we constructed a nested model with one fixed effect (habitat) and one random effect (site within habitat). The effect of rain exclosure on small social groups was assessed using t-tests for normally distributed data and Wilcoxon rank-sum tests for non-normal data. Rainfall measurements were compared using ANOVA. We transformed proportions and nest size measurements using the arcsine square root and natural log, respectively. 3.3 RESULTS 3.3.1 Survival and Dispersal Following our initial expectation, A. baeza colony survival decreased with decreasing elevation (Fig. 3.1, Table 3.1). In the cloud forest (2100 m), 93% of transplanted colonies survived the initial 30 days following mesh tube removal (13 out of 14 nests), whereas 50% survived in the lower montane rainforest (1000 m, 9 out of 18 nests). In the lowland rainforest, all colonies for which the mesh tubes were removed went extinct or disbanded within four days (400 m, 15 out of 15 nests, Fig. 3.1a). Within the species range, nests in lower montane habitat had a significantly smaller proportion of spiders remaining at the end of the experiment, but also produced fewer propagules, compared to the source habitat (Fig. 3.1b; Fig 3.2a). We found no survival differences between transplant sites within habitats. The propagules that became established at the lower montane habitat (1000 m) went extinct faster than those in the cloud forest (Fig. 3.2b, Table 3.1). We were unable to measure propagule formation and survival in the lowland rainforest since none of the transplanted colonies yielded propagules. 3.3.2 Mechanisms i. Rain exclosure: Rainfall intensity (Fig. 3.3b) tended to decrease with increasing elevation (ANOVA: F13=3.23; p=0.079), but daily rainfall (Fig. 3.3a) did not change linearly with elevation (ANOVA F29=1.08, p=0.35). Our measurements suggest that June-August 2006 received more rainfall than average, based on monthly averages recorded in Baeza, near 66  Yanayacu (approx. 8 mm/day,1900 m) and Tena, near Jatun Sacha (approx. 13 mm/day, 500 m; Grubb & Whitmore 1966; Neill 1999b). Although A. baeza survival was quite low in nests transplanted to the lowland rainforest even when the mesh tube was never removed, six times more spiders remained at the time of mesh removal or nest recollection in nests that were sheltered from the rain compared to those that were exposed (Fig. 3.4a, Table 3.1). Spiders in the rain protection treatment built significantly more webbing than those in the unsheltered treatment (Fig. 3.4c). The native social species, A. eximius, also built significantly larger nests (nest volume compared; Fig. 3.4d, Table 3.1) and showed a nonsignificant trend (1.5-fold increase) towards greater survival (Fig. 3.4b) when protected from the rain. ii. Predator presence: Significantly more nests had ants present within them as we approached the lowland rainforest (Fig. 3.5a, Table 3.1). In the lowlands, at least 62% of the A. baeza nests contained ants, some even before removal of the mesh tubes. Jumping spiders (family Salticidae), another potential spider predator, by contrast, occurred more frequently in nests at higher elevations (Fig. 3.5b, Table 3.2). During this study, we observed predation of Anelosimus spiders by ants six times and by jumping spiders twice. 3.4 DISCUSSION 3.4.1 Survival and dispersal in native versus foreign habitat In some scenarios, sociality may be an adaptation that allows organisms to colonize habitats where solitary individuals are unable to replace themselves (Wilson 1975; Slobodchikoff 1984; Avilés 1999). Along these lines, Avilés and Tufiño (1998) showed that females of the highly social spider A. eximius produce, on average, less than one surviving offspring per capita in lowland tropical rainforest habitats when attempting to raise their brood solitarily. This suggests that the lowland tropical rainforest may be a sink environment for non-social species in the genus, a suggestion further supported by the absence of subsocial Anelosimus species in this habitat (Avilés et al. 2007). The idea that sociality is favoured in environments where solitary foundresses cannot replace themselves underlies the ‘assured fitness returns hypothesis’ developed by Gadagkar (1990) to explain the evolution of worker castes in eusocial organisms. This hypothesis has been generalized to explain selection for sociality in conditions that are unfavourable to solitary living in a range of 67  organisms, including the subsocial spider Anelosimus studiosus (Jones et al. 2007). The present study investigates the hypothesis that solitary dispersers and their broods cannot replace themselves in the lowland tropical rainforest, thus preventing colonization of that habitat by subsocial species. Consistent with our predictions, the subsocial A. baeza suffered a dramatic decrease in colony and individual survival when transplanted to two lower elevation rainforest areas relative to control transplants within the native cloud forest habitat (Fig. 3.1). Mortality increased in severity with decreasing elevation, such that all groups went extinct within four days of mesh removal at the lowest transplant location (400 m). The strength of the effect suggests that lowland rainforest is unsuitable for this species, although we cannot yet tease apart the extent to which this effect may be due to the social phenotype (i.e., small groups containing the offspring of a single female being unable to maintain their web or ward off predators) versus physiological maladaptation to this environment. Some factors, such as higher temperatures (Neill 1999b) or diseases not present in the native range of the species, could lead to greater mortality independent of social phenotype. We found, for instance, a greater number of moulding A. baeza carcasses in the rainforest than at the two higher elevation habitats, demonstrating that predation was not the only source of mortality in the lowlands. The apparent reluctance of A. baeza spiders to settle in incipient nests in the lowland rainforest (see Methods) further suggests that the spiders found local conditions inhospitable. Anelosimus baeza groups yielded significantly fewer propagules in the lower montane habitat, where temperature, rain intensity, and ant abundance were intermediate between cloud forest and lowland habitats (Fig. 3.2a). This finding supports our third prediction that spiders should attempt to remain in groups by dispersing less frequently as we approach the rainforest. This result suggests that at least one key subsocial trait that appears adaptive in the species’ native habitat —early dispersal—may be less so in areas approaching lowland rainforest conditions. Previous studies have demonstrated that subsocial spiders will delay dispersal if greater food abundance or other factors favour remaining in groups (e.g. Krafft et al. 1986; Ruttan 1990). Such behavioural plasticity may explain the decreased dispersal behaviour in the lower montane habitat. Newly established propagules also went extinct faster in the lower montane habitat (1000 m) relative to native cloud forest habitat (2100 m), 68  further supporting the notion that a solitary life style is maladaptive in areas closer to the rainforest. 3.4.2 Mechanisms preventing subsocial species from colonizing the lowland rainforest We identified two environmental factors—rain intensity and incidence of ants— that were positively correlated with environments where greater mortality of solitary dispersers and small groups occurred. While other environmental parameters may also influence the success of the spiders in this system, we suggest that these two factors, in particular, could disadvantage small groups or solitary spiders relative to large groups in lowland rainforest and lower montane habitats. The same factors may have played a historical role selecting for large groups in the lowland rainforest, thus contributing to the repeated evolution of sociality in this genus (Agnarsson 2006; Agnarsson et al. 2006). Rain intensity (mm per 30 min.) increased linearly as we approached the lowland rainforest from higher elevation cloud forest areas (Fig. 3.3b), though average rainfall did not (Fig. 3.3a). Rain sheltered nests of both the social and subsocial species transplanted in the lowlands had increased survival and more webbing than exposed nests (Fig. 3.4). These results support the hypothesis that intense rainfall can adversely affect the web maintenance abilities and survival of spiders with dense and potentially expensive webs. Anelosimus baeza individuals living in smaller groups contribute more webbing per capita to maintain their nests (Powers 2004). By extension, small groups might be particularly susceptible to nest damage. Our experiment demonstrated that a release from intense rainfall allowed groups containing the same number of individuals to build more webbing or larger nests (Fig. 3.4). Historically, intense rainfall in the rainforest may have exerted a strong selective pressure on individuals to remain in the natal nest and share the cost of web maintenance and rebuilding, as was suggested for the social spider Agelena consociata (Riechert et al. 1986). Our finding that shelter from rain had a larger proportional effect on survival in subsocial than in social groups (see results) suggests that social species may be better adapted to cope with heavy rains. An experiment testing whether larger social groups experience a reduced benefit from rain protection compared to the small groups tested here would further support the idea that cooperative web care is an important benefit of group living. However, because the rain-sheltered subsocial groups still went extinct, rain is clearly not the only factor preventing this subsocial species from colonizing the lowland rainforest. 69  Greater predation rates may also contribute to the absence of subsocial species from the lowland rainforest. We observed ants inside nests most frequently in the lowland rainforest (400 m), where A. baeza nests failed to establish (Figs. 3.1, 3.5a). Ants are known to be important predators in terrestrial habitats (Hölldobler & Wilson 1990), and have recently been shown to reduce the survival of solitary allodapine bees, thus driving increased cooperative nesting (Zammit et al. 2008). Our finding matched previous studies of ant abundance, which showed that ants are most common in lowland tropical rainforests in South America (e.g. Janzen et al. 1976; Guevara & Avilés in press), the Phillipines (Samson et al. 1997), and Malaysia (Bruhl et al. 1999). Previous studies of social spiders and our preliminary observations (see results) also suggest that ants can cause rapid reduction in colony size or colony extinction (Henschel 1998). In response to ants, we observed Anelosimus spiders moving away from plant surfaces, attacking small numbers of ants (J. Purcell 2006, personal observation.) or building a web retreat (J. Guevara 2005, personal communication). Through cooperation and greater nest volume, larger groups would have an advantage over small groups and solitary foundresses in all three responses. Thus increased predation or disturbance by ants in the lowland rainforest may exclude A. baeza from this habitat. However, further experiments are needed to directly demonstrate that predation by ants hinders the establishment of Anelosimus species with exclusively small groups and solitary females in the lowland tropical rainforest. By contrast, though spiders may be ecologically important Anelosimus predators (Perkins et al. 2007), our findings suggest that jumping spiders may not threaten the survival of Anelosimus spiders in the lowland rainforest (Fig. 3.5b). Jones et al. (2007) suggested that colder temperatures favour multi-female groups in the otherwise subsocial species A. studiosus by lengthening the juvenile development period. This, in turn, provides cooperative nesters an increased probability of surrogate offspring care in the event of maternal death. Although not driven by temperature, we suggest that a similar mismatch between offspring development and maternal lifespan may occur in the lowland rainforest. Here, the mismatch may result from predation increasing the risk of maternal mortality or intense rains hindering the ability of single individuals or small groups to maintain their prey capture webs and feed their offspring.  70  3.4.3 Conclusion We found higher mortality both of transplanted groups and of propagules at the range boundary of A. baeza compared with the source population, thus matching the findings of theoretical predictions about species ranges (e.g. Kirkpatrick & Barton 1997) and previous empirical studies (e.g. Angert & Schemske 2005). Our example is novel in suggesting social behaviour as an adaptive phenotypic axis. While further work is still required to elucidate the relative influence of individual versus social traits to the adaptive limits of A. baeza, this study has identified at least two factors—intense rainfall and presence of potential predators in nests— that are correlated with an absence of subsocial Anelosimus in the lowland rainforest. These factors, combined with the availability of an abundant supply of large insects in this habitat (Guevara & Avilés 2007; Powers & Avilés 2007; Yip et al. 2008), may have selected for larger social groups during the initial colonization of the lowland rainforest. This system thus presents an ideal opportunity to simultaneously investigate the role of interacting environmental factors driving the distinctive pattern of social distribution in spiders, and to pursue several of the unanswered questions about range margins, as enumerated by Bridle & Vines (2007).  71  Table 3.1: Summary of statistical method and result for each comparison. Habitats are abbreviated according to their relative elevation, so high is cloud forest (2100 m), mid is lower montane rainforest (1000 m), and low is lowland rainforest (400 m). Treatments Species  Compared  # observations Comparison  Approach  per habitat  Test Statistic  Model  Direction  (One-sided unless High, mid, low/ stated)  Closed, open  Logrank (Mantel-  H: 14, M: 18,  Cox) Test  L: 15  HLM (Mixed)  H: 12, M: 9  Test Statistic  p-value  41.37  <0.0001  Survival Results A. baeza  A. baeza  high, mid & low  high & mid  nest extinction  proportion of  χ2  As predicted: high>mid>low  F  11.74  0.007  spiders remaining  As predicted: high>mid  Dispersal Results A. baeza  high & mid  number of  HLM (Mixed)  H: 18, M: 17  F  7.96  0.009  propagules A. baeza  high & mid  As predicted: high>mid  propagule  Logrank (Mantel-  survival  Cox) Test  low, exposed vs.  proportion of  HLM (Glimmix)  covered  spiders remaining  H: 47, M: 18  χ2  2.88  0.045  As predicted: high>mid  Rain Exclosure A. baeza  C: 22, O: 22  F  3.68  0.032  As predicted: covered>exposed  72  Treatments Species  Compared  # observations Comparison  Approach  per habitat  Test Statistic  Model  Direction  (One-sided unless High, mid, low/  A. baeza  A. eximius  A. eximius  low, exposed vs. number of strands covered  on plant  low, exposed vs.  proportion of  covered  spiders remaining  low, exposed vs.  nest volume  stated)  Closed, open  HLM (Mixed)  C: 17, O: 19  F  Test Statistic  p-value  5.58  0.05  As predicted: covered>exposed  Wilcoxon Test  C: 11, O: 10  χ2  0.24  0.31  As predicted: covered>exposed  T-test  C: 11, O: 11  T  1.85  0.04  covered  As predicted: covered>exposed  Predation A. baeza  A. baeza  high, mid & low presence/ absence Two-sided Nested  H: 18, M: 17,  of ants  L: 50  Model  high, mid & low presence/ absence Two-sided Nested  H: 18, M: 17,  of spiders  L: 50  Model  χ2  26  0.038  As predicted: high<mid<low  χ2  42.9  0.0002  Opposite of Predicted: high>mid>low  73  Figure 3.1: (a) The greatest colony survival (circles, cloud forest; squares, lower montane; triangles, lowland rainforest) and (b) number of individuals remaining occurred in the native cloud forest habitat. Subsocial colonies did not persist in the lowland rainforest.  74  Figure 3.2: (a) Transplanted subsocial spider groups yielded more propagules (new colonies founded by dispersers) per nest and (b) those propagules survived longer in their native cloud forest habitat (circles, 2100 m) compared to the lower montane (squares, 1000 m) transplant habitat. Propagule survival was measured from the first day we observed a new propagule, and was censored if the propagule survived until the end of the experiment, or uncensored if the propagule went extinct during the experiment. 75  Figure 3.3: (a) Average daily rainfall did not decrease linearly with elevation. However, (b) average rainfall intensity increased with decreasing elevation as expected.  76  Figure 3.4: In rain-sheltered nests (grey) in the lowland rainforest, small groups of both the foreign ((a, c) A. baeza (subsocial)) and native ((b, d) A. eximius (social)) species showed improved survival and a greater amount of webbing compared to nests exposed to the rain (white). Our subsocial measurements were taken upon mesh tube removal or recollection of nests still enclosed in mesh (see methods). Both comparisons (a, c) were significant for the subsocial species, while only nest volume (d) was significantly greater for the social species (Table 1).  77  Figure 3.5: (a) The presence of ants in transplanted spider nests was highest in the lowland rainforest, while (b) the reverse pattern was observed for salticid spiders. Error bars show 95% CI.  78  3.5 REFERENCES Agnarsson, I. 2006. A revision of the New World eximius lineage of Anelosimus (Araneae, Theridiidae) and a phylogenetic analysis using worldwide exemplars. Zool. J. Linn. Soc.-Lond. 146, 453-493. Agnarsson, I., Avilés, L., Coddington, J. A., & Maddison, W. P. 2006. Sociality in theridiid spiders: repeated origins of an evolutionary dead end. Evolution 60, 2342-2351. Angert, A. L. & Schemske, D. W. 2005. The evolution of species’ distributions: Reciprocal transplants across the elevation ranges of Mimulus cardinalis and M. lewisii. Evolution 59, 1671-1684. Avilés, L. 1997. Causes and consequences of cooperation and permanent sociality in spiders. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 476-498. Cambridge, UK: Cambridge University Press. Avilés, L. 1999. Cooperation and non-linear dynamics: an ecological perspective on the evolution of sociality. Evol. Ecol. Res. 1, 459-477. Avilés, L. & Gelsey, G. 1998. Natal dispersal and demography of a subsocial Anelosimus species and its implications for the evolution of sociality in spiders. Can. J. Zoo. 76, 2137-2147. Avilés, L. & Tufiño, P. 1998. Colony size and individual fitness in the social spider Anelosimus eximius. Am. Nat. 152, 403-418. Avilés, L., Agnarsson, I., Salazar, P., Purcell, L., Iturralde, G., Yip, E., Powers, K. S. & Bukowski, T. 2007. Altitudinal patterns of spider sociality and the biology of a new mid-elevation social Anelosimus species in Ecuador. Am. Nat. 170, 783-792. Bridle, J. R. & Vines, T. H. 2007. Limits to adaptation at range margins: when and why does adaptation fail? Trends Ecol. Evol. 22, 140-147. Bruhl, C. A., Mohamed, M. & Linsenmair, K. E. 1999. Altitudinal distribution of leaf litter ants along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. J. Trop. Ecol. 15, 265-277. Eickwort, G. C., Eickwort, J. M., Gordon, J. & Eickwort, M. A. 1996. Solitary behavior in a high altitude population of the social sweat bee Halictus rubicundus (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 38, 227-233.  79  Gadagkar, R. 1990. Evolution of eusociality: The advantage of assured fitness returns. Philos. T. Roy. Soc. B 329, 17-25. Gaston, K. J. 2003. The structure and dynamics of geographic ranges. Oxford, UK: Oxford University Press. Geber, M. A. & Eckhart, V. M. 2005. Experimental studies of adaptation in Clarkia xantiana. II. Fitness variation across a subspecies border. Evolution 59, 521-531. Grubb, P. J. & Whitmore, T. C. 1966. A comparison of montane and lowland rain forest in Ecuador: II. The climate a its effects on the distribution and physiognomy of the forests. J. Ecol. 54, 303-333. Guevara, J. & Avilés, L. 2007. Multiple sampling techniques confirm differences in insect size between low and high elevations that may influence levels of spider sociality. Ecology 88, 2015-2023. Guevara ,J. & Avilés, L. in press. Elevational changes in insect and other arthropod composition at tropical latitudes: a comparison of multiple sampling methods and social spider diets. Insect Conserv. Biodiv. Henschel, J. R. 1998. Predation on social and solitary individuals of the spider Stegodyphus dumicola (Araneae, Eresidae). J. Arachnol. 26, 61-69. Hölldobler, B. & Wilson, E. O. 1990. The Ants. Cambridge, MA: Belknap Press. Holt, R. D., Keitt, T. H., Lewis, M. A., Maurer, B. A. & Taper, M. L. 2005. Theoretical models of species’ borders: single species approaches. Oikos 108, 18-27. Janzen, D. H., Ataroff, M., Farinas, M., Reyes, S., Rincon, N., Soler, A., Soriano, P. & Vera, M. 1976. Changes in the arthropod community along an elevational transect in the Venezuelan Andes. Biotropica 8, 193-203. Jones, T. C., Riechert, S. E., Dalrymple, S. E. & Parker, P. G. 2007. Fostering model explains variation in levels of sociality in a spider system. Anim. Behav. 73, 195-204. Keyserling, E. 1884. Die Spinnen Amerikas. II. Theridiidae. Nürnberg 1. Kirkpatrick, M. & Barton, N. 1997. Evolution of a species’ range. Am. Nat. 150, 1-23. Krafft, B., Horel, A. & Julita, J. M. 1986. Influence of food supply on the duration of the gregarious phase of a maternal-social spider, Coelotes terrestris (Araneae, Agelenidae). J. Arachnol. 14, 219-226.  80  Liebert, A. E., Nonacs, P. & Wayne, R. K. 2005. Solitary nesting and reproductive success in the paper wasp Polistes aurifer. Behav. Ecol. Sociobiol. 57, 445-456. London, K. B. & Jeanne, R. L. 2003. Effects of colony size and stage of development on defense response by the swarm-founding wasp, Polybia occidentalis. Behav. Ecol. Sociobiol. 57, 445-456. Lubin, Y. & Bilde, T. 2007. The evolution of sociality in spiders. Adv. Study Behav. 37, 83145. Mori, K. & Saito, T. 2005. Variation in social behavior within a spider mite genus, Stigmaeopsis (Acari : Tetranychidae). Behav. Ecol. 16, 232-238. Neill, D. A. 1999a. Vegetation. In Catalogue of the vascular plants of Ecuador. (eds. P. M. Jørgensen & S. León-Yánez), pp.13-25. St. Louis, MO: Monogr. Syst. Bot. MO. Bot. Gard. 75. Neill, D. A. 1999b. Climates. In Catalogue of the vascular plants of Ecuador. (eds. P. M. Jørgensen & S. León-Yánez), pp. 8-13. St. Louis, MO: Monogr. Syst. Bot. MO. Bot. Gard. 75. Perkins, T. A., Riechert, S. E. & Jones, T. C. 2007. Interactions between the social spider Anelosimus studiosus (Araneae, Theridiidae) and foreign spiders that frequent its nests. J. Arachnol. 35, 143-152. Powers, K. S. 2004. Prey abundance and the evolution of sociality in the social spider genus Anelosimus. PhD Thesis, University of Arizona. Powers, K. S. & Avilés, L. 2003. Natal dispersal patterns of a subsocial spider Anelosimus cf. jucundus (Theridiidae). Ethology 109, 725-737. Powers, K. S. & Avilés, L. 2007. The role of prey size and abundance in the geographic distribution of spider sociality. J. Anim. Ecol. 76, 995-1003. Purcell, J. & Avilés, L. 2007. Smaller colonies and more solitary living mark higher elevation populations of a social spider. J. Anim. Ecol. 76, 590-597. Riechert, S. E. & Hall, R. F. 2000. Local population success in heterogeneous habitats: reciprocal transplant experiments completed on a desert spider. J. Evolution. Biol. 13, 541-550.  81  Riechert, S. E., Roeloffs, R. & Echternacht, A. C. 1986. The ecology of the cooperative spider Agelena consociata in equatorial Africa (Araneae, Agelenidae). J. Arachnol. 14, 175-191. Ruttan, L. M. 1990. Experimental manipulations of dispersal in the subsocial spider, Theridion pictum. Behav. Ecol. Sociobiol. 27, 169-173. Rypstra, A. L. & Tirey, R. S. 1991. Prey size, prey perishability and group foraging in a social spider. Oecologia 86, 25-30. Samson, D. A., Rickart, E. A. & Gonzales, P. C. 1997. Ant diversity and abundance along an elevational gradient in the Philippines. Biotropica 29, 349-363. Singer, J. D. 1998. Using SAS PROC MIXED to Fit Multilevel Models, Hierarchical Models, and Individual Growth Models. J. Educ. Behav. Stat. 24, 323-355. Slobodchikoff, C. N. 1984. Resources and the evolution of social behaivor. In A New Ecology: Novel Approaches to Interactive Systems. (eds. P. W. Price, C. N. Slobodchikoff & W. S. Gaud) pp 227-251. New York, NY: John Wiley and Sons. Smith, A. R., Wcislo, W. T. & O’Donnell, S. 2003. Assured fitness returns favor sociality in a mass-provisioning sweat bee, Megalopia genalis (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 54, 14-21. Uetz, G. W. & Hodge, M. A. 1990. Influence of habitat and prey availability on spatial organization and behavior of colonial web-building spiders. Natl. Geogr. Res. 6, 22-40. Uetz, G. W., Boyle, J., Hieber, C. S. & Stimson Wilcox, R. 2002. Antipredator benefits of group living in colonial web-building spiders: the ‘early warning’ effect. Anim. Behav. 63, 445-452. Wilson, E. O. 1975. Sociobiology. Cambridge, MA: Belknap Press. Yip, E. C., Powers, K. C, & Avilés, L. 2008. Cooperative capture of large prey solves the problem of a declining surface area to volume ratio of large social spider colonies. Proc. Natl. Acad. Sci. U.S.A. 105, 11818-11822. Zammit, J., Hogendoorn, K. & Schwarz, M. P. 2008. Strong constraints to independent nesting in a facultatively social bee: quantifying the effects of enemies-at-the-nest. Insect. Soc. 55, 74-78.  82  CHAPTER 4: DIFFERENT FUNCTIONAL STRATEGIES OF FIVE COEXISTING SOCIAL AND SUBSOCIAL ANELOSIMUS SPIDER SPECIES4 4.1 INTRODUCTION The mechanism whereby closely related species coexist in the same habitat remains a current and perplexing question in ecology. Classical theory suggested that no two coexisting species can employ identical ecological strategies (Gause 1934) and predicted that a certain degree of separation, or limiting similarity, would be necessary to allow coexistence (e.g. Hutchinson 1959; MacArthur & Levins 1967). Empirical work has confirmed that coexisting species often differ from one another along at least one of three ecological axes— spatial, temporal, and dietary— in organisms ranging from plants (e.g. Silvertown 2004) to invertebrates (e.g. Albrecht & Gotelli 2001; Kaplan & Denno 2007; Behmer & Joern 2008; Gilbert et al. 2008) and vertebrates (MacArthur 1958; Schoener 1968; Pianka 1969; Kronfield-Schor & Dayan 1999; Harmon et al. 2007). Previous studies have demonstrated that differentiation at different spatial and temporal scales, as well as segregation along multiple ecological axes, may allow greater divergence between otherwise similar species. For example, Gilbert et al. (2008) found that the bromeliad-dwelling larvae of four sympatric mosquito species segregated along two different spatial axes. They demonstrated that the species tended to remain at different depths in a water column. The two species that occupied the most similar position in the water column almost never co-occurred in the same bromeliad wells, while species occupying different depths frequently overlapped in the same wells. Temporally, organisms can vary in phenology on an annual scale, as in sympatric mantid species with offset egg hatching times (Hurd 1988), or in their daily activity patterns, as seen in pollinator species that partition floral resources and visitations over the course of a single day (Stone et al. 1996). Other studies have showed that segregation along one ecological axis can increase differentiation along another. For instance, two species of spiny mice in the Negev Desert foraged at 4  A version of this chapter will be submitted for publication. Purcell, J., Vasconcellos-Neto, J., Gonzaga, M.O., Fletcher, J.A., & Avilés, L. Different functional strategies of five coexisting social and subsocial Anelosimus spider species. 83  different times of the day and night, thus utilizing different food resources despite similar preferences in a cafeteria experiment (Kronfeld-Schor & Dayan 1999). Similarly, ithamiine butterflies in Brazil have been shown to use the same food resources, but at different times (Vasconcellos-Neto 1991). Coexisting species may alternatively specialize on different spatial zones and the dietary resources therein in areas where they overlap, as in the benthic and limnetic forms of the three-spine stickleback (Schluter 1996). More locally, sympatric gecko species living on islands in the Indian Ocean use perches of different heights and diameters in areas where they overlap, and this can influence the type of prey that they capture (Harmon et al. 2007). In spiders, competition for prey has been notoriously difficult to demonstrate in the field (Wise 1993). Despite frequent observations that coexisting spiders tend to differ along spatial and temporal axes on a relatively fine scale (e.g. Greenstone 1980; Brown 1981), manipulative experiments frequently reveal little or no direct competition for prey items between coexisting species (e.g. Schaefer 1978; Wise 1981; but see Spiller 1984). Instead, Wise (1993) suggested that spiders may compete for web positions, or that spider densities may be maintained at relatively low levels by predation or disturbance. Each of these mechanisms can influence the structure of communities in areas where similar species overlap (Leibold 1995; Herberstein 1998; Amarasekare 2007). Recent studies of spiders have suggested that nest location plays an important role not only in foraging success, but also in optimizing microclimatic conditions and predator avoidance for the organism (Harwood et al. 2003). Herberstein (1998) found that competition for nest position occurs in two sympatric Linyphiid spider species, which build their nests on the same plant. If the spiders are released onto an empty tree, they build their nests at an intermediate height, but if another spider nest is present, they will build their nests at higher or lower heights above the ground, respectively. Relatively few studies of spiders have investigated resource and habitat use among many sympatric species in the same genus. One notable exception is the research on Tetragnatha spider community assembly in the Hawaiian Islands. Gillespie and her colleagues demonstrated that communities assembled through adaptive radiation and dispersal on a single island contain ecomorphs that employ different ecological strategies, but that similar communities repeatedly assemble on different islands (e.g. Gillespie 2004). Comparing closely related species reduces the ecological variation that may arise from 84  different phylogenetic histories, thereby allowing a clearer identification of the role of the intrinsic and extrinsic ecological patterns that enable coexistence. Anelosimus (Theridiidae, Araneae) species exhibit social behaviours ranging from subsocial (non-territorial periodic social) to highly social (non-territorial permanent social) (Avilés 1997; Agnarsson 2006). Social species occupy a shared nest for multiple generations, where group members cooperate in brood rearing, prey capture, and nest maintenance. Depending on the habitat, social spider nests may contain up to hundreds of thousands of individuals. In contrast, the nests of subsocial species typically contain a single family group, as adults usually nest solitarily and the offspring remain together and cooperate only for a portion of their life cycle (Avilés 1997). Some subsocial species experience shorter cooperative phases than others, depending on the age at which the offspring disperse from the natal nest (Lubin & Bilde 2008). Aside from the differences in colony size and population dynamics reflecting differences in social behaviour, the ecology and life history traits of many species in this genus are strikingly similar. The three-dimensional web architectures of the species in this genus show a high degree of similarity, and allow the spiders to intercept and consume a variety of flying insects. The nests are occupied for an extended period of time, and also provide the spiders with protection from predators and from the elements (under leaves within the nests). In general, social species are restricted to the lowland humid tropics up to about 2000 m above sea level, whereas subsocial species are common in the temperate zone, at higher elevations in the humid tropics (down to about 1000 m), and in more arid habitats in the tropics (Avilés et al. 2007). At intermediate elevation and latitude social and subsocial species may overlap. The most speciose Anelosimus community assembled in this way occurs in Serra do Japi, Brazil at the southern edge of the tropical zone, where at least five Anelosimus species, displaying a wide range of social behaviours, overlap in one 354 km2 protected area (Gonzaga & Santos 1999). By studying the species in this area, we can investigate how species with similar social phenotypes interact with each other versus how they interact with species exhibiting different social phenotypes. The hypothesis of limiting similarity predicts that sympatric species will differ from one another along at least one ecological axis (MacArthur & Levins 1967). Here we investigate whether the five Anelosimus species coexisting in Serra do Japi segregate along 85  spatial and temporal axes. Along the spatial axis, we test the hypothesis that the species are segregated at (at least) one of three spatial scales (H1). If segregated regionally (i), each species would occupy a different area of the Serra do Japi reserve. If segregated at a more local scale (ii), different species could occupy different adjacent habitats, such as the forest edge vs. the interior, or, if occurring in the same habitat (iii), they could occupy different types of plants or different positions on the same plant. Species that do not differ along any of the three spatial scales could alternatively differ along the temporal axis (H2). Here, we investigate differentiation at all three spatial scales and one temporal scale, phenology, for each species in Serra do Japi. Non-random segregation within one or more of these dimensions would support the hypothesis of limiting similarity. We also hypothesize that the level of sociality displayed by these five species may correlate with the patterns of spatial segregation, a factor that has not been well explored in previous studies of species coexistence. Presuming that similarity of social system results in dietary similarity (Guevara & Avilés 2007; Powers & Avilés 2007; Yip et al. 2008) and competition for resources takes priority over a species’ habitat or nest architectural needs, we expect that species with similar social systems will be segregated in different regions or habitats within the park (H3). Alternatively, if species with similar social systems share habitat requirements and habitat constraints or competition for nest sites take priority over competition for prey, then species with similar social systems may occupy the same habitats (H4). In the latter case, the hypothesis of limiting similarity would be supported if similar species sharing common habitats are separated either temporally or at finer spatial scales. 4.2 METHODS 4.2.1 Species descriptions At least five Anelosimus species coexist in Serra do Japi, Brazil (Gonzaga & Santos 1999). Anelosimus species exhibit social behaviours ranging from highly social to solitary, with the majority of species being subsocial. In this site, one species, Anelosimus dubiosus Keyserling, is considered highly social (permanent, non-territorial social; Marques et al. 1998). Social (or quasisocial) species occur in nests shared with multiple reproductive adult females and their offspring. Colony members cooperate with each other in brood care, nest maintenance, and prey capture (Avilés 1997). Dispersal does not usually occur every 86  generation, and new nests are always initiated by gravid females alone or in small groups. In contrast, subsocial Anelosimus species generally occur in nests shared by a single female and her offspring. Siblings cooperate for a portion of their life cycle in the natal nest, but always disperse prior to sexual maturity (Avilés 1997). A second species in Japi, Anelosimus jabaquara Levi, displays a phenotype that appears to be intermediate between social and subsocial, since some females fail to disperse from their natal nest (Gonzaga & Vasconcellos-Neto 2001). However, although A. jabaquara’s nests may reach sizes as large or larger than those of A. dubiosus, most individuals do disperse as subadults prior to maturation and the sex ratio of this species is only slightly, if at all, biased (Marques et al. 1998), whereas social species usually show a highly female biased sex ratio (Avilés 1993). Two of the species that we measured, Anelosimus studiosus Hentz and Anelosimus baeza Agnarsson are typical subsocial species, with dispersal occurring at subadult instars each generation. Some subsocial species disperse at an earlier age than others, and these are considered to be ‘less social’ than those that disperse closer to sexual maturity (Lubin & Bilde 2008). The fifth species, Anelosimus nigrescens Keyserling, is considered a nearlysolitary subsocial species due to the early dispersal of immature individuals and the reduced maternal care phase (Agnarsson et al. 2006; M. O. Gonzaga, unpublished data). Lab studies have also shown A. nigrescens young to be more aggressive toward their siblings when dispersal does not occur (J. Vasconcellos-Neto, unpublished data). Anelosimus jabaquara and A. dubiosus occur in the same clade and are closely related. Anelosimus studiosus and A. baeza occur in different clades, but are also fairly close relatives (Agnarsson 2006). These four species are moderately interrelated, whereas A. nigrescens is a much more distant relative of the other four species (Agnarsson 2006). Both A. dubiosus and A. jabaquara only occur in Brazil. A. dubiosus has been documented between 900 and 1600 m above sea level, whereas A. jabaquara has been found from 600-1100 m (Agnarsson 2006; M. O. Gonzaga, unpublished data). The remaining 3 species have a wider geographical and altitudinal distribution. Anelosimus studiosus is found throughout North and South America from 0-2500 m above sea level (Agnarsson 2006; Avilés et al. 2007). Anelosimus baeza is widely distributed in South America from Panama to Peru and east to Brazil (Agnarsson 2006; Avilés et al. 2007; M. O. Gonzaga, unpublished data). However both A. baeza and A. studiosus are absent from lowland tropical rainforests. 87  Anelosimus nigrescens has been found in Brazil from sea level in São Paulo state to about 1300m in Minas Gerais state (M. O. Gonzaga, unpublished data), as well as Guyana and possibly Venezuela (Agnarsson 2005). 4.2.2 Habitat Description Serra do Japi is located between the latitudes 23°12'-23°22'S and longitudes 46°57'47°05'W, comprising an area of about 354 Km2. The vegetation is composed mainly of semi-deciduous forest, markedly seasonal, with leaf fall occurring especially during the dry and relatively cool season (from April to September). Between 700-900 m, the taller trees reach about 20-25 m with a predominance of Myrtaceae, Lauraceae, Meliaceae, Caesalpinaceae, Mimosaceae, Euphorbiaceae and Fabaceae. Areas located above 1000 m are usually covered by semi-deciduous altitudinal forest, with a lower canopy (8-10 m high) and scattered plants in the shrub and herbaceous strata. Our study area was located from about 900 m-1100 m. Rains are concentrated in the first months of the summer (from October to January) and annual precipitation is about 1350 mm in the region of Jundiaí, Brazil, the town nearest Serra do Japi (Morellato et al. 1989; Morellato & Leitão Filho 1992; Pinto 1992). 4.2.3 Sampling We surveyed the nests of five Anelosimus species along six randomly placed transects in Serra do Japi in November, 2005. Transects were positioned using the T-squared sampling method (Krebs 1999). Random points were selected along the network of roads and trails inside the park, and the spider colony nearest that point was located. We then initiated the transect at the first colony’s nearest neighbour, and followed a random bearing from there. Each transect was 200 m long by 5 m wide. Transects initiated near the road always diverged into denser forest, so all transects included a sample from the forest edge and inside the forest. In most transects, the forest edge consisted of shrubby habitat along human created edges (roads, trails, and overgrown pastures). In one case, we also found forest edge habitat along a natural swamp. We acknowledge that some of the forest interior habitat that we surveyed may have been influenced by the edge effect, but the plants available for web construction differed considerably between these two habitat types. Anelosimus spider nests consist of three-dimensional baskets built with dense webbing. The cross section area of the basket is proportional to the colony size (Purcell & Avilés 2007). Above the basket, the ‘prey capture’ area of the nest consists of a mesh of 88  looser threads where most prey items are intercepted. In each transect, we located Anelosimus spider colonies, measured the height above the ground, cross section area, and prey capture height of the nests (Purcell & Avilés 2007), characterized the vegetation supporting the nests, and classified the habitat. Our ‘prey capture height’ variable measures the height of this structure from the top of the basket to the supporting branches above. The ‘height above ground’ variable measures the distance from the lowest part of the basket to the ground below it. We identified plant substrates to family (when possible), estimated the tree or shrub trunk diameter at breast height (DBH), classified into categories for 0 (i.e., plant < 1.4 m tall), small (up to 10 cm), medium (up to 40 cm) or large (greater than 40 cm). We then determined whether the spider nest was located on a branch tip, in the middle of a branch, or in the core of the plant. We also visually estimated the percent canopy cover directly above the nest. Because we did not have a convex mirror, this was a coarse estimate, but the observations were made sequentially during the same field trip by the same observer. We measured the distance from the nearest forest edge, up to 10 meters, and estimated the average height of the canopy above each nest. We collected voucher specimens from nests for which species identification was not possible in the field. These individuals were then reared until adulthood in the lab and identified. Because the protected area prohibited destructive sampling, we were not able to collect vouchers from some nests that were positioned high above the ground, and these colonies were not included in this analysis. This inability to identify the highest nests may have skewed our height above ground comparisons, since the highest nests were not included. For each nest, we visually assessed the instar of the inhabitants (juvenile, subadult, adult, egg sac present). These stages are relatively straightforward to distinguish based on the size and sexual development of the spiders. Anelosimus species tend to have fairly discrete generations, and most nests contained individuals of a single instar. If nests were at a stage when mother and offspring overlapped, or when two instars were represented, we assigned them to the category representing the most common instar. To assess the characteristics of available habitat, we collected the same habitat and vegetation data that we assessed for each spider nest at 20 randomly selected points along each transect. In these surveys, we identified the four plants nearest the random point. We then estimated the trunk diameter and the approximate number of possible nest locations at 89  the branch tips, mid branch, and core of the plant nearest the point. The latter measurement did not always result in an equal number of locations, since many branches diverged from the same point on the trunk (core), and small branches did not have room for a nest in the “mid branch” area. We also repeated the measurements of canopy cover, distance from forest edge, and forest height for each random point. Finally, starting at the plant nearest the random point, we measured the amount of continuous horizontal vegetation on which a spider nest could expand as the maximum length of leafy vegetation with gaps of less than 30 cm (Purcell & Avilés 2007). 4.2.4 Analysis In total, we analyzed the characteristics of 34 A. nigrescens nests, 58 A. baeza nests, 7 A. studiosus nests, 52 A. jabaquara nests, and 31 A. dubiosus nests. We performed a principal components analysis (PCA) on all of the nest size, habitat, and temporal data except substrate identity (9 variables). We coded categorical data as ordinal variables along hierarchical axes, so instars were ordered from 0 (egg sac) through 6 (adult), and position on plant was ordered from the core (0) to the edge (2). We repeated the PCA without the nest size variables. For the first three principal components axes, we compared the degree of separation between the five Anelosimus species using an ANOVA. We also calculated the correlation coefficients of all of the variables used in the principal components analysis. To control for multiple comparisons across all five species, we used the Tukey-Kramer test to compare nest area. The prey capture height and habitat data remained non-normal even after transformation, so we used non-parametric tests for the remaining comparisons. We compared prey capture height, canopy cover, distance from the forest edge, forest height and height above ground with the Kruskal-Wallis test across all five species and the Wilcoxon rank-sum test for pairwise comparisons. We also compared each species distribution to our random sample and the continuous vegetation measures between forest edge and forest interior habitats with the Wilcoxon rank-sum test. For substrate diameter and nest position on plant substrate comparisons, we used Pearson Chi-squared tests. For nest size and habitat variables, we compared each species (10 comparisons). Because the five species segregated into two distinct groups at the habitat level, we only compared forest edge species with each other (3 comparisons) and forest interior species with each other (1 comparison) for the phenology and fine scale spatial variables. We used the Bonferroni correction to account for 90  these multiple comparisons. For each species, we used independent contrasts to compare the vegetation used for nesting and position of the nests on the vegetation substrate versus the vegetation and positions available to it. We used the binomial test to determined if plant substrates were used more often than expected based on their abundances and to compare the habitat of social and subsocial nests to the neutral expectation. We compared the proportion of nests represented by instar categories (juvenile, subadult, adult female, females with egg sacs) between species sharing the same habitat with a Pearson Chi-squared test. We used JMP 5.1 (SAS Institute, Cary, NC) for all statistical tests. 4.3 RESULTS The five species separated from one another along the first three principal component axes (Fig. 4.1), which described about 64% of the observed variation (Table 4.1). We found a similar result when we removed nest size and prey capture height from the analysis (Appendix 3). The first principal component correlated with nest size and the variables that differentiate forest edge and forest interior habitat (% canopy cover, distance from forest edge, and forest height), accounting for 34% of the observed variation. The second and third principal components correlated with local nest position variables and phenology, accounting for an additional 30% of the observed variation (Table 4.1, Appendix 3). All three PC axes revealed significant differences between the five species (ANOVA PC1: F(4, 172) = 35.25, p < 0.0001; PC2: F(4, 172) = 13.66, p < 0.0001; PC3: F(4, 172) = 7.36, p < 0.0001). Nest size and prey capture height were positively correlated with % canopy cover, distance from forest edge, and forest height variables (Table 4.2). Two variables influencing principal component 2, plant diameter and nest height above ground, were also positively correlated (Table 4.2, Appendix 3). Looking at individual variables separately, we found a significant difference in nest cross-section area and prey capture height across the five species found in Serra do Japi (Fig. 4.2). Specifically, the nearly-solitary A. nigrescens and subsocial A. baeza had significantly smaller nests and shorter prey capture webs than the social A. dubiosus and the intermediate social A. jabaquara. Anelosimus studiosus was statistically intermediate in both traits, but our sample size was very small (Fig. 4.2). For each figure, we align the species along the xaxis from least to most social. We found significant differences between the five species for 91  our measures of canopy cover (χ2 = 33.28, DF = 4, p < 0.0001), distance from the forest edge (χ2 = 57.69, DF = 4, p < 0.0001) and forest height (χ2 = 51.44, DF = 4, p < 0.0001). In particular, the three species with smaller nests (less social) usually occurred on the forest edge (Fig. 4.2c), while the two species with larger nests (more social) occurred inside the forest (Fig 4.2c; Table 4.2; Appendix 4). One of the two forest interior species, A. jabaquara, was often found in areas with a taller forest canopy and greater cover (Fig. 4.2d, e; Appendix 4). We also compare the habitat position of the nests of each species to the random measures of each habitat variable. We found that A. jabaquara occurred in habitat with a significantly greater canopy cover, higher forest, and was significantly farther from the forest edge than would be expected if the species were distributed randomly. Similarly, A. baeza and A. nigrescens were both found significantly closer to the forest edge in habitat with shorter plants and less canopy cover than would be expected based on the distribution of habitat types in the transect. The habitat position of A. dubiosus did not differ from the random expectation. The nests of A. studiosus were found on the forest edge more often than expected by chance, but the other comparisons were not significant (Table 4.2; Appendix 4). Overall, about 50% of the nests that we found were located on the forest edge, while the other 50% were located inside the forest. A binomial test demonstrates that the distribution of nest positions of more social and less social nests, respectively, was non-random (binomial test p<0.0001 for both comparisons). We did not find differences in the vegetation substrate among the three forest edge species or among the two forest interior species, although most species were found on specific plants more often than expected based on our measures of plant availability. On the forest edge, A. nigrescens and A. baeza nests were constructed on plants in the family Asteraceae more often than expected by chance, and A. studiosus showed a similar, nonsignificant trend (binomial test: A. nigrescens p = 0.015; A. baeza p = 0.00090; A. studiosus p = 0.30). Anelosimus jabaquara nests were built on Myrtaceae trees more often than expected by chance and A. dubiosus showed a similar trend (binomial test: A. jabaquara p = 0.00015; A. dubiosus p = 0.071). We did, however, find that locally coexisting species tended to build their nests at different heights above the ground (Fig. 4.3a, 4.3b). Anelosimus nigrescens built nests significantly lower than A. baeza (χ2 = 13.2, p = 0.00030) and A. studiosus (χ2 = 6.76, p = 0.0093), and A. jabaquara built nests significantly higher than A. dubiosus (χ2 = 9.65, p = 92  0.0019). All three forest edge species tended to build their nests on plants with greater diameter than expected given the available substrate (Fig. 4.3c, Appendix 4). Anelosimus jabaquara nests were found on plants that were both larger than expected and larger than plants with A. dubiosus nests (Fig. 4.3d, Table 4.2). The position of nests on the substrate also varied for each species living at the forest edge (Fig. 4.3e) and inside the forest (Fig. 4.3f, Appendix 4). We found a non-significant trend suggesting that in Serra do Japi, the forest interior habitat has more room for nest expansion (greater continuous vegetation length, see methods) than the forest edge habitat (Wilcoxon test χ2 = 3.31, p = 0.069). The three species on the forest edge showed substantial differences in age structure, although each contained overlapping instars (Fig. 4.4). During our study, A. baeza was at the earliest life cycle stage of the three species, with most nests (57%) containing subadult individuals. A. nigrescens was more advanced, with the majority (62%) of nests containing adult individuals (only 18% contained subadults). A few A. nigrescens nests (15%), however, contained young juveniles (alone or with mother), suggesting that this species might have more than one generation a year or be less seasonal than the other four species (Fig. 4.4). This impression is also supported by the discovery of adult males and females of this species in the dry season (July-September, M. O. Gonzaga, unpublished data). Anelosimus studiosus appeared to show the most advanced phenology with 50% of nests containing adult females with egg sacs or young clutches, although our sample size was small (Fig. 4.4). In this study, we did not find a significant difference in the age structure of the forest interior species, A. jabaquara and A. dubiosus (Fig. 4.4). Most transects contained nests of at least 4 of the 5 species investigated in this study. The density of nests of all Anelosimus species in the six random transects that we sampled was about 0.055/m2. Of the nests that we were able to positively identify, we found the highest densities of A. jabaquara and A. baeza nests, with 0.013/ m2 and 0.011/ m2 respectively. A. nigrescens and A. dubiosus were about half as abundant, with about 0.006 nests/m2 each. A. studiosus was much less common, and was, in fact, absent from three of the six transects, with only 0.001 nest/m2. Nests that were not positively identified, either because we were unable to view or collect the occupants, or because the specimens we collected died prior to reaching sexual maturity, made up the difference in density (0.018 nests/m2.) This nest density is markedly greater than density recorded for Anelosimus species 93  living in Ecuador (J. Purcell, unpublished data) or Arizona (L. Avilés, unpublished data), where this genus has been extensively studied. In both Brazil and Ecuador, the density of individual Anelosimus spiders (estimated at 0.5 spiders/m2) is very high relative to the density of solitary spider species in many tropical habitats (I. Agnarsson, unpublished data). 4.4 DISCUSSION Serra do Japi, Brazil is the only location in the Americas where at least five Anelosimus spider species are known to coexist. These five species showed marked differences in habitat use and phenology along three principal component axes (Fig. 4.1, Appendix 3). We found that all five species were present in many of our random transects, which suggests that they are not segregated at the regional level. However, we did find segregation at the habitat and local spatial scales, thus supporting our first hypothesis. The three species building smaller nests tended to occur at the forest edge in both human induced and natural shrubby habitat, while the two species with larger nests were usually found inside the forest (Fig. 4.2). Within the species clustering at the forest edge and inside the forest, we found differences in vegetation substrate size, height of nests above the ground, and nest position (Fig. 4.3). In our comparisons of nest size and habitat, A. studiosus was often statistically intermediate between the forest edge and forest interior species, but this may be due to the low density of A. studiosus in our study area, and the conservative tests used in these analyses. Because the habitat use of this species is more similar to A. baeza and A. nigrescens, we lump A. studiosus into the forest edge group in this discussion. We also point out that our measurement of nest height above ground may be slightly skewed by our inability to collect voucher specimens from the highest nests (see results). Nevertheless, our results suggest that these spiders do segregate into different local habitats and occupy different positions on plants within those habitats (H1). Although we found fine-scale spatial segregation between the species (Fig. 4.3), we were surprised that the nests of species on the forest edge and species inside the forest were found on plants of particular families (Asteraceae and Myrtaceae, respectively) more often than expected based on the plants’ abundance in our transects (see results). This type of pattern could emerge in one of several ways. Either the spiders dispersed and settled on all available vegetation substrate, but experience higher mortality on some types of plants than 94  others, or they preferentially settled on certain types of plants. This pattern may reflect the importance of an underlying microhabitat factor that is beneficial to both spiders and the plants being used. We cannot yet differentiate between these mechanisms, but they do suggest that some plants or their locations provide more suitable substrate for these spiders than others. We also detected a difference in phenology among the three forest edge species, supporting our second hypothesis (Fig. 4.4). Although we did not find a significant difference in the age structure of the two forest interior species, a previous study demonstrated that these species were offset by about one month in their annual phenologies (Marques et al. 1998). Spider species may also vary on a finer temporal scale, if they forage or perform nest maintenance at different times of day. This possibility is addressed in another study of the spiders at Serra do Japi (J. Guevara, M. O. Gonzaga J. VasconcellosNeto, and L. Avilés, unpublished manuscript). Taken together, our results demonstrate that these five spider species occupy different positions in two spatial dimensions (H1) and at least one temporal dimension (H2) in Serra do Japi, thus supporting the hypothesis of limiting similarity in this community (MacArthur & Levin 1967). In this system, we suggest that differences in social systems may facilitate the coexistence of so many related species in this habitat. These species vary from highly social (A. dubiosus) to nearly-solitary (A.nigrescens), and the nest basket size and prey capture height differ accordingly by an order of magnitude (Fig. 4.2). In its effect on the interaction between the organism and its environment, this nest size difference may be analogous to body size differences noted in other studies of species coexistence (e.g. Schluter 1996; Murray & Baird 2008). In the context of this study, we hypothesized that species with similar social systems would either occur in different habitats, in order to reduce direct competition for prey (H3), or they could occur in the same habitat, if habitat requirements take priority over competition (H4). Our results support the latter hypothesis (H4). The social species (A. dubiosus) and the intermediate social species (A. jabaquara), both of which had the largest nests in the system, tended to occur inside the forest, while the two subsocial species (A. baeza and A. studiosus), as well as the nearly-solitary species (A. nigrescens) tended to occur at the forest edge (Fig. 4.2). We also found significant positive correlations between all three variables used to distinguish between forest edge and forest interior habitats (canopy cover, 95  distance from forest edge, and forest height) and our measures of nest size and prey capture height (Appendix 3). We suggest that more social species would require sturdier branches and more continuous horizontal vegetation to support their larger nest baskets, as well as branches above the nest to allow the construction of enough prey capture webbing to support the greater number of individuals in the colonies (Yip et al. 2008, Fig. 4.5). We found that the plants were larger and generally had more continuous horizontal vegetation (non-significant) inside the forest compared to the forest edge in Serra do Japi. In Ecuador, Purcell and Avilés (2007; 2008) similarly found that only social- and not subsocial- Anelosimus species occurred inside the forest. They also found that measures of vegetation continuity were correlated with the maximum nest size in different habitats (Purcell & Avilés 2007). Smaller subsocial nests may not require as much room to expand horizontally or vertically. Nest size differences may also facilitate differential capture of prey resources, a possibility that is being investigated in the same system (J. Guevara et al., unpublished manuscript). However, these results suggest that the habitat requirements of social versus subsocial species are more important than competition for prey in driving the spatial distribution pattern observed in this community. This result is not surprising in light of previous studies of species coexistence in spiders, which suggest that competition for prey is rarely the primary force structuring spider communities (Wise 1993). The level of sociality also influences the timing and means of dispersal in Anelosimus spiders (Avilés 1997). Interestingly, we found the greatest age structure difference in the two species that share the most similar social phenotype (A. baeza and A. studiosus; Fig. 4.4). Colonies of both species occur at the forest edge (Fig. 2) and are expected yield subadult dispersers. The species with the intermediate phenology, A. nigrescens, is the least social species, with dispersal usually occurring during the juvenile instars (M. O. Gonzaga, unpublished data). This species is also thought to produce two generations per year and be less seasonal (see results) compared to the two remaining subsocial species. Thus, the three forest edge species would tend to disperse and to raise young offspring at different times of the year from each other. The forest interior species (A. dubiosus and A. jabaquara) tend to employ different dispersal mechanisms (Marques et al. 1998). When each group yields dispersers (not necessarily each generation, see methods), A. jabaquara individuals tend to 96  disperse prior to sexual maturity, while A. dubiosus dispersers are solitary or small groups of gravid females (Marques et al. 1998). We speculate that differential dispersal times and mechanisms may facilitate coexistence in species with similar social phenotypes, potentially by reducing competition for nest sites or for prey (see below). Phylogenetic relationships could also contribute to similarities in habitat use. The closely related species (A. dubsiosus and A. jabaquara; A. baeza and A. studiosus) tended to use similar habitat types (forest interior vs. forest edge, respectively), and we cannot yet separate the role of species differentiation from the role of sociality for these species. However, we did find that A. nigrescens, which is only distantly related to the other four species, showed ecological convergence with the two species sharing a more similar social phenotype (A. baeza and A. studiosus). This result suggests that social phenotype likely contributes to at least some of the observed spatial differentiation. In comparing our results with the habitat data collected at random points along each transect, we found that the distribution of nests of more social and less social species differed significantly from the null expectation (see results; Table 4.2; Appendix 4). Interestingly, the relationship between the spatial scales that we document here was the reverse of the pattern described by Gilbert et al. (2008) in their study of sympatric mosquito larvae living in bromeliad wells. They showed that more functionally similar species segregated at a coarser spatial scale, while functionally different species were more likely to overlap at finer scales. This pattern was used to demonstrate that the mosquito community was not neutrally assembled. We hypothesized that a similar pattern might emerge in our system (H3). However, we found that functionally similar species shared the same habitat, but differed from one another in their use of space within the habitat, thus supporting our alternative hypothesis (H4). We suggest that both patterns demonstrate a non-neutral community assembly, but probably result from different stabilizing mechanisms- possibly to avoid negative interspecific competition in the former case, and as a result of habitat constraints in the latter case. The primary goal of this study was to take a preliminary look at the coexistence of several species where each shows a different degree of sociality. So far, our results demonstrate that these five spider species differ from one another and from random expectation in two spatial and at least one temporal dimensions in Serra do Japi. The degree 97  of sociality appears to facilitate differentiation between the more social and less social species, at least in their habitat use. Although we cannot yet determine the mechanism causing these patterns, we speculate about the factors that may be driving the observed spatial and temporal differentiation, and suggest further studies that would test these ideas. Some possible mechanisms include: (1) Competition for nest sites. While there may seem to be an unlimited number of potential nest sites, several factors including Anelosimus nest architectural requirements, use of specific plant types, and very high nest densities (for this genus) could drastically reduce the number of effective sites and drive nest site competition. Under this scenario, temporal variation in dispersal times for the subsocial spiders may then drive the local nest position differences that we observed in the field (e.g. Herberstein 1998). (2) Prey size and availability. The food resources available to the spiders may interact with the spatial and temporal differences described in this study. Yip et al. (2008) demonstrated that larger Anelosimus colonies tend to catch larger prey items. Thus, the prey size captured by subsocial species should vary over the course of the year: newly founded nests tend to be small, while nests nearing the dispersal stage are large. The differences in phenology (Fig. 4.4) could reduce competition for prey during critical stages of the spiders’ life cycle by staggering the feeding intensive portions of their life cycles and by causing each species to specialize on different prey sizes at any given time. A concurrent study investigates differential prey capture in this system for a part of the year (J. Guevara et al., unpublished manuscript), but a long-term study will be needed to verify the links between spatial, temporal, and dietary axes. (3) Microclimate optimization. Previous studies have shown that arthropods often seek nesting sites with optimal temperature or moisture conditions (e.g. Wise 1993; Harwood et al. 2003; Fischer & Vasconcellos-Neto 2005). In this system, spatial variables such nest height above the ground may allow spiders to adjust the nest’s microclimate, as humidity and temperature may vary with proximity to the ground. Microclimate conditions may also vary seasonally, especially in a wet/dry season habitat like Japi. This may place constraints on the phenology of a species. 4.4.1 Conclusion In this study, we present evidence that social and subsocial Anelosimus spiders coexisting in Serra do Japi, Brazil employ different functional strategies along both spatial and temporal axes. Many studies have demonstrated a similar phenomenon in other web98  building spiders (reviewed by Wise 1993), but few studies have focused on so many closely related species. We offer two key insights in this paper, both of which will be interesting to explore in future studies. First, we suggest that studying habitats in which closely related social and non-social species overlap will facilitate a deeper understanding of the role of ecological factors in shaping the distribution of sociality in social organisms (J. Purcell, unpublished manuscript). In this case, we have found that the different habitat requirements of social versus subsocial species may contribute to their spatial separation. Second, we propose that social behaviour in general may contribute to the ability of otherwise similar species to coexist because species with different social systems should differ from one another in the manner in which they interact with their environment, thereby differentiating in their diet, their habitat requirements, and/or the timing of key life cycle events.  99  Table 4.1: Principal components analysis results. PC Axis Variable  1  2  3  4  5  Nest Size  0.437  -0.236  -0.178  0.26  0.128  Prey Capture Height  0.31  -0.171  -0.226  0.721  -0.135  Percent Canopy Cover  0.408  0.147  -0.163  -0.407  0.0819  Distance from Forest Edge  0.453  0.0264  0.00645  -0.28  -0.22  Forest Height  0.476  0.186  -0.0209  -0.199  -0.129  Height Above Ground  -0.18  0.615  -0.0959  0.154  0.00011  DBH (Ord)  0.151  0.54  -0.154  0.19  0.597  Location Plant (Ord)  0.147  0.399  0.545  0.255  -0.526  Instar (Ord)  0.196  -0.171  0.749  0.0612  0.511  Eigenvalue  3.073  1.649  1.012  0.918  0.814  % Variance  34.145  18.323  11.244  10.197  9.048  100  Table 4.2: Summary of two-way statistical analyses for spatial comparisons. Species are either compared with each other or with available habitat measures. NS: Non-significant; *: 0.05>p>0.005; **: 0.005>p>0.0005; ***: p<0.0005. Distance from Canopy Cover Species  Comparison  Forest Edge (Wilcoxon)  A. baeza  A. studiosus  Nest Position  Above Ground  (Edge, Mid,  (Wilcoxon)  Core) (Pearson)  DBH Category  (Wilcoxon) (Wilcoxon)  A. nigrescens  Nest Height Forest Height  (Pearson)  A. baeza  NS  NS  NS  ***  NS1  NS1  A. studiosus  NS  NS  NS  **  ***  NS  A. jabaquara  ***  ***  ***  -  -  -  A. dubiosus  NS1  **  NS  -  -  -  Available  ***  ***  **  -  ***  **  A. studiosus  NS  NS  NS  NS  NS  NS  A. jabaquara  ***  ***  ***  -  -  -  A. dubiosus  NS1  ***  **  -  -  -  Available  ***  ***  ***  -  *  ***  A. jabaquara  NS1  **  NS1  -  -  -  A. dubiosus  NS  NS  NS  -  -  -  Available  NS  *  NS  -  NS  **  101  Distance From Canopy Cover Species  Comparison  Forest Edge (Wilcoxon)  A. dubiosus  1  Nest Position  Above Ground  (Edge, Mid,  (Wilcoxon)  Core) (Pearson)  DBH Category  (Wilcoxon) (Wilcoxon)  A. jabaquara  Nest Height Forest Height  (Pearson)  A. dubiosus  NS1  **  ***  **  ***  ***  Available  **  *  ***  -  ***  ***  Available  NS  NS  NS  -  NS  NS  In these cases, p-value was less than 0.05, but was considered non-significant due to multiple comparisons.  102  Figure 4.1: Mean and 95% confidence intervals of each species along principal components axes 1-3.  103  104  Figure 4.2: Comparison of nest sheet area (a), nest prey capture height (b) and broad-scale habitat measurements (c-e) across all five sympatric species. Letters show statistically significant differences between species (a: Tukey Kramer test; b-e: Wilcoxon rank-sum test, Appendix 4). The species with the largest nests (a; A. jabaquara and A. dubiosus) tend to be found furthest from the forest edge (d). Between these two species, A. jabaquara is found in areas with the highest percent canopy cover (c) and the tallest forest height (e). The species with smaller nests (a; A. nigrescens and A. baeza) usually occur on the forest edge and in areas with less canopy cover (c, d). A. studiosus is statistically intermediate in nest size and each of the habitat variables, but the comparison is based on a very small sample size for this species. Because our observations suggest that A. studiosus tends to build smaller nests in forest edge habitat, we group this species with A. nigrescens and A. baeza.  105  Figure 4.3: Local scale spatial comparisons of forest edge species (left panel) and forest interior species (right panel) show that species in both habitats show some differences in nest site preference. Significant differences between comparisons are shown with letters (Table 4.2, Appendix 4). The proportion of nests on substrate of each DBH category and locations on the substrate were also compared with the null expectation based on available vegetation and possible nest locations.  106  107  Figure 4.4: Number of nests observed for each species and the most common instar present in each. The three subsocial species (all of which live on the forest edge) show significant differences with respect to the distribution of age classes. A. baeza had the greatest frequency of subadults in nests, which differed significantly from A. nigrescens (χ2=290.25, DF=6, p<0.0001) and A. studiosus (χ2=290.25, DF=6, p<0.0001). A. nigrescens nests frequently contained adults (both male and female), while A. studiosus nests often contained adult females with egg sacs (χ2=23.6, DF=6, p=0.0006). A. jabaquara and A. dubiosus, both of which prefer forested habitat, did not show a significant difference in instar frequency in our study (χ2=6.16, DF=3, p=0.1039), but a more detailed study of their phenology showed that major events, such as dispersal and reproduction, were offset by a month (Marques et al. 1998).  108  Figure 4.5: Webs of A. jabaquara (a, scale bar 10 cm), A. baeza (b), A. studiosus (c), A. nigrescens (d), and A. dubiosus (e, b-e scale bar 5 cm). Also shown are male and female of A. nigrescens (f), female of A. baeza (g), and female of A. jabaquara. 109  4.5 REFERENCES Agnarsson, I. 2005. Revision and phylogenetic analysis of American ethicus and rupununi groups of Anelosimus (Araneae, Theridiidae) Zool. Scr. 34, 389-413. Agnarsson, I. 2006. A revision of the New World eximius lineage of Anelosimus (Araneae, Theridiidae) and a phylogenetic analysis using worldwide exemplars. Zool. J. Linn. Soc.-Lond. 146, 453-493. Agnarsson, I., Barrantes, G., & May-Collado, L. 2006. Notes of the biology of Anelosimus pacificus Levi, 1963 (Theridiidae, Araneae)- evidence for an evolutionary reversal to a less social state. J. Nat. Hist. 40, 2681-2687. Albrecht, M. & Gotelli, N.J. 2001. Spatial and temporal niche partitioning in grassland ants. Oecologia 126, 134-141. Amarasekare, P. 2007. Trade-offs, temporal variation, and species coexistence in communities with intraguild predation. Ecology 88, 2720-2728. Avilés, L. 1993. Interdemic selection and the sex ratio: a social spider perspective. Am. Nat. 142, 320-345. Avilés, L. 1997. Causes and consequences of cooperation and permanent sociality in spiders. In The evolution of social behaviour in insects and arachnids. (eds. J. C. Choe & B. Crespi), pp. 476-498. Cambridge, UK: Cambridge University Press. Avilés, L., Agnarsson, I., Salazar, P., Purcell, L., Iturralde, G., Yip, E., Powers, K.S. & Bukowski, T. 2007. Altitudinal patterns of spider sociality and the biology of a new mid-elevation social Anelosimus species in Ecuador. Am. Nat. 170, 783-792. Behmer, S.T. & Joern, A. 2008. Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc. Natl. Acad. Sci. U.S.A. 105, 1977-1982. Brown, K. M. 1981. Foraging ecology and niche partitioning in orb-weaving spiders. Oecologia 50, 380-385. Fischer, M.L., & Vasconcellos-Neto, J. 2005. Microhabitats occupied by Loxoceles intermedia and Loxoceles laeta (Araneae: Sicariidae) in Curitiba, Paraná, Brazil. J. Med. Entomol. 42, 756-765. Gause , G. F.1934. The Struggle for Existence. Williams and Wilkins, Baltimore, MD. Gilbert, B., Srivastava, D. S., & Kirby, K. R. 2008. Niche partitioning at multiple scales facilitates coexistence among mosquito larvae. Oikos 117, 944-950. 110  Gillespie, R. 2004. Community assembly through adaptive radiation in Hawaiian spiders. Science 303, 356-359. Gonzaga, M. O. & Santos, A. J. 1999. The females of Anelosimus dubiosus and Anelosimus jabaquara (Araneae, Theridiidae). J. Arachnol. 27, 432-434. Gonzaga, M. O. & Vasconcellos-Neto, J. 2001. Female body size, fecundity parameters and foundation of new colonies in Anelosimus jabaquara (Araneae, Theridiidae). Insectes Soc. 48, 94-100. Greenstone, M. H. 1980. Contiguous allotropy of Pardosa ramulosa and Pardosa tuoba (Araneae: Lycosidae) in the San Francisco Bay Region, and its implications for patterns of resource partitioning in the genus. Am. Midl. Nat. 104, 305-311. Harmon, L.J., Harmon, L.L., & Jones, C.G. 2007. Competition and community structure in diurnal arboreal geckos (genus Phelsuma) in the Indian Ocean. Oikos 116, 1863-1878. Harwood, J. D., Sunderland, K. D., & Symondson, W. O. C. 2003. Web-location by linyphiid spiders: prey-specific aggregation and foraging strategies. J. Anim. Ecol. 72, 745-756. Herberstein, M. E. 1998. Web placement in sympatric linyphiid spiders (Arachnida, Araneae): individual foraging decisions reveal inter-specific competition. Acta Oecologia 19, 67-71. Hurd, L.E. 1988. Consequences of divergent egg phenology to predation and coexistence in two sympatric, congeneric mantids (Orthoptera: Mantidae). Oecologia 76, 549-552. Hutchinson, G.E. 1959 Homage to Santa Rosalia, or why are there so many kinds of animals? Am. Nat. 93, 145-159. Kaplan, I., & Denno, R. F. 2007. Interspecific interactions in phytophagous insects revisited: a quantitative assessment of competition theory. Ecol. Lett. 10, 977-994. Krebs, C. J. 1999. Ecological Methodology, 2nd edn. Benjamin/Cummings, Menlo Park, CA. Kronfeld-Schor, N. & Dayan, T. 1999. The dietary basis for temporal partitioning: food habits of coexisting Acomys species. Oecologia 121, 123-128. Leibold, M. A. 1995. The niche concept revisited: mechanistic models and community context. Ecology 76, 1371-1382. Lubin, Y., & Bilde, T. 2008. The evolution of sociality in spiders. Adv. Stud Behav. 37, 83145. 111  MacArthur, R.H. 1958. Population ecology of some warblers of northeastern coniferous forests. Ecology 39, 599-619. MacArthur, R.H. & Levins, R. 1967. The limiting similarity, convergence and divergence of co-existing species. Am. Nat. 101, 377-385. Marques, E. S. A., Vasconcellos-Neto, J., & Mello, M. B. 1998. Life history and social behavior of Anelosimus jabaquara and Anelosimus dubiosus (Araneae, Theridiidae). J. Arachnol. 26, 227-237. Morellato, L. P. C., Rodrigues, R. R., Leitão Filho, H. F., & Joly, C. A. 1989. Estudo fenológico comparativo de espécies arbóreas de floresta de altitude e floresta mesófila semidecídua na Serra do Japi, Jundiaí, SP. Revista Brasilera Botânica 12, 85-98. Morellato, L. P. C. & Leitão Filho, H. F. 1992. Padrões de frutificação e dispersão na Serra do Japi. In História natural da Serra do Japi: ecologia e preservação de uma área florestal no sudeste do Brasil. (ed. L. P. C. Morellato) pp. 112-140. Editora da Unicamp/Fapesp, Campinas, Brazil. Murray, M.G. & Baird, D.R. 2008. Resource-ratio theory applied to large herbivores. Ecology 89, 1445-1456. Pianka, E. R. 1969. Sympatry of desert lizards (Ctenotus) in Western Australia. Ecology 50, 1012-1030. Pinto, H. S. 1992. O clima da Serra do Japi. In História natural da Serra do Japi: ecologia e preservação de uma área florestal no sudeste do Brasil. (ed. L. P. C. Morellato) pp. 3038. Editora da Unicamp/Fapesp, Campinas, Brazil. Purcell, J. & Avilés, L. 2007. Smaller colonies and more solitary living mark higher elevation populations of a social spider. J. Anim. Ecol. 76, 590-597. Purcell, J. & Avilés, L. 2008. Gradients of precipitation and ant abundance may contribute to the altitudinal range limit of subsocial spiders: insights from a transplant experiment. Proc. R. Soc. B 275, 2617-2625. Schaefer, M. 1978. Some experiments on regulation of population density in the spider Floronia bucculenta (Araneidae: Linyphiidae). Symp. Zool. Soc. Lond. 42, 203-210. Schluter, D. 1996. Ecological speciation in postglacial fishes. Philos. T. Roy. Soc. B 351, 807-814. 112  Schoener, T.W. 1968. The Anolis lizards of Bimini: Resource partitioning in a complex fauna. Ecology 49, 704-726. Silvertown, J. 2004. Plant coexistence and the niche. Trends Ecol. Evol. 19, 605-611. Spiller, D. A. 1984. Competition between two spider species: experimental field study. Ecology 18, 1-35. Stone, G., Willmer, P. & Nee, S. 1996. Daily partitioning of pollinators in an African Acacia community. Proc. R. Soc. B 263, 1389-1393. Vasconcellos-Neto, J. 1991. Interactions between Ithomiinae butterflies and Solanaceae: Feeding and reproductive strategies. In Plant-animal interactions: Evolutionary ecology in tropical and temperate regions. (eds. P. W. Price, T. M. Lewinsohn, G. W. Fernandes, and W. W. Benson) pp. 291-313. John Wiley & Sons, New York. Wise, D. H. 1981 Inter- and intraspecific effects of density manipulations upon females of two orb-weaving spider species (Araneae: Araneidae). Oecologia 48, 252-256. Wise, D. H. 1993. Spiders in ecological webs. Cambridge University Press, Cambridge, UK. Yip, E. C., Powers, K. C, & Avilés, L. 2008. Cooperative capture of large prey solves the problem of a declining surface area to volume ratio of large social spider colonies. Proc. Natl. Acad. Sci. U.S.A. 105, 11818-11822  113  CHAPTER 5: ARE ALL SOCIAL TRAITS CREATED EQUAL? COEVOLUTIONARY DYNAMICS BETWEEN THREE SOCIAL TRAITS AND DISPERSAL5 5.1 INTRODUCTION Intuitively, selection on social traits should be influenced by the costs and benefits of dispersal behaviours. As dispersal rates decrease, we might expect the emergence of some type of social behaviour to improve conditions for individuals remaining in their natal habitat. Reciprocally, an increase in social traits may lead to decreased propensity to disperse. Classic models investigated the costs and benefits of dispersal behaviours versus philopatry. Dispersal may be physically demanding and dangerous. For instance, dispersers may experience increased predation or starvation risk, or they may be unable to locate new suitable habitat (reviewed by Johnson & Gaines 1990; Dieckmann et al. 1999). However, philopatry also carries the heavy cost of subsequent competition between kin (Hamilton & May 1977) and inbreeding (Bengtsson 1978). Hamilton and May (1977) found that under simple conditions, an evolutionarily stable strategy emerges in which about half of a parent’s offspring remain in the natal patch, while the other half disperse. Thus evolution strikes a balance between both strategies, thereby mitigating the intrinsic costs of dispersal and philopatry. The evolution of different types of social behaviours may influence the evolutionarily stable strategy that balances dispersal and philopatry. Many theoretical studies have explored the conditions under which different classes of helping behaviours can evolve. In their review, Lehmann and Keller (2006) divided these studies of social evolution into two distinct categories, those investigating strong altruism (net cost>0) and those investigating cooperation (net cost<0). Hamilton’s (1964) simple analytical solution to the problem of how altruism evolved ushered in a long period of interest in the role of relatedness in social evolution. In the model, Hamilton demonstrated that strong altruism could evolve when the benefits (b) received by relatives of the actor (r) outweigh the cost (c) of performing an 5  A version of this chapter will be submitted for publication. Purcell, J., Brelsford, A., & Avilés, L. Are all social traits created equal? Co-evolutionary dynamics between three social traits and dispersal. 114  altruistic action (rb>c). Inherent in the construction of this simple rule are two distinct fitness accounting mechanisms: direct fitness gained or lost by the focal individual, and indirect fitness, gained through increases in the fitness of relatives of the focal individual (in proportion to their degree of relatedness). Subsequent studies have looked at the local kinship conditions that would predispose populations to the evolution of strong altruism both from a theoretical (reviewed by West Eberhard 1975; Michod 1982; West et al. 2007) and from an empirical perspective (e.g. Queller & Strassman 1998; Clutton-Brock 2002). In particular, strong altruism is considered to be a useful representation of the evolution of castes and sex allocation in the Hymenoptera (Chapuisat & Keller 1999). Trivers (1971) developed the alternative idea that altruistic behaviours can evolve if the actor is likely to receive reciprocal benefits from another altruist. This form of helping behaviour has later come to be viewed as a form of weak altruism (Wilson 1979, 1980, 1990). Weak altruism differs from strong altruism in that the cost of an altruistic act may be regained through subsequent beneficial interactions with other altruists. Studies of reciprocal altruism are classified as ‘cooperative’ under Lehmann and Keller’s (2006) classification. Subsequent studies have explored the evolutionary dynamics of social behaviours that do not necessarily carry a direct fitness cost to the actor (e.g. Vehrencamp 1983; Pulliam & Caraco 1984; Slobodchikoff 1984; Emlen 1991). For instance, Avilés (1999) suggested that cooperation may allow group members to experience a greater mean individual fitness in certain habitats, either through increasing access to resources, or through reducing the probability of predation. In reality there appear to be many examples of organisms that exhibit helping behaviours without paying extremely high fitness costs (Avilés et al. 2002). The consideration of less costly social behaviours and social systems that do not rely on kin structuring has been the focus of many recent empirical studies investigating the role of ecological pressures in shaping the evolution of sociality (reviewed by Purcell in review; see also Jarvis et al. 1994; Pollock & Cabrales 2008; Verdolin & Slobodchikoff 2008). Whether they are altruists or cooperators, group members may have less incentive to risk dispersal if the benefits of social behaviours outweigh the costs of local competition. Early models considered the role of population viscosity in the evolution of social behaviours (Taylor 1992; Wilson et al. 1992). They proposed that limited mobility would increase the number of close relatives in local areas, thus predisposing the organisms to the evolution of 115  altruism. However, in the simplest models, they found that the benefits of cooperating with relatives in a local neighbourhood were exactly cancelled out by the costs of competing with the same relatives (Taylor 1992; Wilson et al. 1992). Subsequent investigations of dispersal and sociality have explored the range of behavioural and ecological conditions under which altruism can evolve, including both the inclusive fitness benefits of helping kin (e.g. Perrin & Lehmann 2001; Lehmann & Perrin 2002), and the ecological benefits of altruism under particular spatial (Le Galliard et al. 2003) or demographic conditions (e.g. Lehmann et al. 2006; Alizon & Taylor 2008). Each of these studies has identified a subset of specialized conditions under which altruism can evolve, and they have shown that the evolution of altruism is not necessarily linked to a decrease in dispersal behaviours. However, these studies almost universally focus on the case of strong altruism (but see Vainstein et al. 2007). Here, we compare the evolutionary trajectories of dispersal with three distinct social traits, tolerance (allows grouping, but no benefits from group living), cooperation (benefits to grouping, but no cost to cooperation) and weak altruism (benefits to grouping, moderate cost) that have not been carefully considered in previous co-evolutionary models with dispersal. Our model does not include any mechanism to enhance helping behaviours among kin, although kin recognition has been an important component of previous models of altruism and dispersal (Perrin & Lehmann 2001; Lehmann & Perrin 2002; Lehmann et al. 2006). We use an individual based simulation model (Lion & van Baalen 2008) to investigate the following hypotheses: For each of the three social traits, we expect selection against dispersal behaviour to result in directional selection increasing each social trait, even if sociality is costly (H1), and selection against each social trait to result in directional selection increasing dispersal propensity, even if dispersal is costly (H2). These classic hypotheses were not supported in the simplest models (e.g. Taylor 1992; Wilson et al. 1992), but have been supported under some specific conditions in subsequent models of the coevolution between strong altruism and dispersal (e.g. Lehmann & Perrin 2002). In the tolerance model, we expect tolerance behaviour to relax the local competition for sites that was cited as the primary force maintaining moderate dispersal rates in Hamilton and May’s (1977) classical model of dispersal in stable habitats. As a result, we predict that the equilibrium tolerance level will increase linearly with increasing selection against dispersal (H3). In the cooperation and weak altruism models, we expect cooperation values to reach 116  their highest levels in cooperation simulations with a high dispersal cost, because cooperation in these models brings synergistic benefits to group members without an individual relative fitness cost and would be most beneficial when there is strong selection against dispersal (H4). Conversely, cooperation values should reach their lowest levels in altruism simulations with a low dispersal cost, because altruism would always carry a cost when it is rare, and this cost would only be counteracted if dispersal were costly (H5). In each model, we also explore two distinct ecological scenarios. We either impose a global individual carrying capacity, which limits the total number of individuals in the population, or we impose random patch extinctions (patch carrying capacity), which limits the number of patches that can be used in each generation. Previous models using similar simple scenarios have demonstrated that stochastic mortality at the individual and the patch levels have different effects on the evolution of dispersal (e.g. Lehmann et al. 2006) and cooperation (e.g. Avilés et al. 2002). As a result, we expect these two ecological scenarios to influence the evolutionary outcomes of the models differently. Under the global individual carrying capacity scenario, selection should maximize individual fitness. In this case, a large number of individuals are randomly removed from the population each generation, so producing more offspring would increase the probability that at least some of a given individual’s offspring survive. This, in turn, should result in very high equilibrium sociality values, because members of groups with higher cooperation levels can gain the highest possible fecundity (H6). Under the patch carrying capacity scenario, all of the individuals living in entire patches are eliminated each generation, so maintaining intermediate dispersal levels should be most important, even when dispersal is costly. Because dispersal rates will remain high and groups will experience constant turnover, we expect lower equilibrium sociality values to evolve in this scenario compared to the individual carrying capacity scenario (H7). We suggest how the predictions generated by this model may be used to inform future empirical studies of sociality and dispersal. 5.2 THE MODEL 5.2.1 Model Structure We construct an individual based simulation model to explore the co-evolutionary trajectories of dispersal combined with one of three distinct social traits: tolerance, 117  cooperation, or weak altruism. Individuals in the model disperse and interact with their patch mates according to their discrete phenotypic trait values, but with no consideration for the level of kinship with their neighbours. Reproduction is asexual, and the phenotypic values of parents are passed down to offspring, unless random mutation occurs. Mutation occurs with a discrete probability (Table 5.1) for each locus and for each offspring during reproduction. The magnitude and direction of the mutation is randomly sampled from a truncated normal distribution with the mean equal to the phenotypic value of the parent and a fixed standard deviation (Table 5.1). Thus all phenotypic variation in the population is generated through mutation in this model. The spatial structure is a finite island model (Wright 1931), so there are a fixed number of discrete habitat patches (Table 5.1), and there is no spatial structure in the distribution of the patches. Individuals dispersing from a given patch have an equal chance of landing in any patch, including their natal patch. Dispersal, then, is modelled as the probability of changing (or attempting to change) patches, and ranges from 0 (never disperse) to 1 (always disperse). A cost of dispersal (Table 5.1) may be imposed in the form of a probability of mortality applied to each individual that attempts dispersal. 5.2.2 Global Population Control Scenarios For each social trait of interest, we simulate two distinct forms of global population control. In the fixed individual carrying capacity scenario, we impose a ceiling population size. If the population is greater than the carrying capacity, the population is randomly reduced to the carrying capacity. In our patch carrying capacity scenario, if the population is greater than the carrying capacity, the inhabitants of a fixed number of randomly selected patches go extinct. Both of these population controls are applied each generation, but are expected to influence the evolution of dispersal in different ways (Lehmann et al. 2006; Alizon & Taylor 2008). 5.2.3 Tolerance Model The model exploring interactions between tolerance and dispersal is a simple modification of Hamilton and May’s (1977) classic model. In this case, tolerant mutants are able to share single habitat patches with one another. However, individuals sharing a habitat patch gain no additional benefit from living in a group, and may pay a cost manifested in decreased individual fitness (tolerant versus solitary fecundity; Table 5.1). In this model, 118  there are only two phenotypes: the mutant ‘tolerant’ individuals and the baseline ‘solitary’ ones. The mutation rate (Table 5.1) controls the probability of switching from one phenotype to the other. Dispersal mutates as described above. When mutation of the sociality locus in this simulation is not permitted (Appendix 5), the dynamics resemble those found in the original model (Hamilton & May 1977). In this model, the number of ‘tolerant’ individuals in a given site is capped at a discrete maximum group size (Table 5.1). 5.2.4 Cooperation and Altruism Models In the model of cooperation, we employ the equation described by Avilés (1999), in which cooperation brings synergistic benefits that counterbalance to a certain extent the effects of local competition for resources, so that fitness reaches its maximum at intermediate group sizes (Appendix 5). This model contains three parameters: an intrinsic rate of growth, which gives the baseline fitness for solitary individuals (r), a cooperation parameter (γ), and a crowding parameter (c). The group size at which fitness is maximum is a function of the latter two parameters, according to the equation nopt = γ/c (Avilés et al. 2004). In our model, the cooperation (γ) levels evolve. Cooperation is initiated at a value of 0 (no cooperation) for each individual in the simulation. Trait values vary continuously between 0 and 1, with higher values representing more cooperative individuals. Mutation occurs as stated above. Mean fitness values are calculated based on the average cooperation level of each group, and are sampled from the unimodal fitness distribution described by equation 1 (Appendix 5). In the models of weak altruism, we add a cost to cooperation that will be greatest when an altruist enters a patch with non-altruists and will decrease in patches with a higher mean cooperation level (Trivers 1971; Pfeiffer et al. 2005). We employ an equation developed by Avilés et al. (2002) that imposes a cost with a fixed magnitude (β, appendix 5). In all other respects, this model works in the same manner as the cooperation model described in the previous paragraph. 5.2.5 Simulations and Baseline Conditions We ran the simulation for each parameter combination at least 100 times for 10,000 generations in Matlab (Mathworks, Inc.). All populations were initiated with one individual present in each patch, and with cooperation (or tolerance) phenotypes of 0 (solitary). Each individual had a dispersal propensity phenotype of 0.7. We carried out sensitivity analysis on  119  each of our variables, both to identify a practical baseline value and to evaluate the influence of each parameter on the model (Appendix 5). 5.3 RESULTS The equilibrium values resulting from co-evolution between dispersal and each of three different social traits are quite distinct (Fig. 5.1). Moreover, the evolution of these traits is strongly influenced by the simple ecological scenarios applied in this model (Fig. 5.1). If tolerance is not costly or is moderately costly, tolerance will evolve to relatively high, stable values. However, if tolerance imposes a large fecundity cost (tolerant individuals attain less than half of the individual fecundity of solitary individuals in the population), selection will maintain the solitary strategy in the population (Fig. 5.1a, blue). Interestingly, the equilibrium tolerance values do not correlate strongly with the dispersal equilibrium levels, which vary depending on dispersal risk (Figs. 5.1, 5.2). The individual carrying capacity scenario seems to allow a slightly higher proportion of tolerant individuals to be maintained in the population, and allows dispersal to evolve to lower values than the patch carrying capacity scenario (Fig. 5.1a). Individuals in the weak altruism model (Fig. 5.1b, blue) attain lower cooperation values than individuals in the cooperation model (Fig. 5.1b, red) under all dispersal and ecological conditions. In these simulations, the ecological scenarios have a large impact on the evolutionary trajectories of both dispersal and cooperation. In the individual carrying capacity simulations, dispersal stabilizes at very low levels in both the weak altruism and the cooperation scenarios. In these cases, increasing dispersal cost has almost no impact on the dispersal equilibrium, and has a weak effect of reducing the cooperation level with increasing dispersal cost (Fig. 5.1b, circles). Under the patch carrying capacity scenario, dispersal is maintained at relatively high levels even with increasing dispersal costs (Fig. 5.1b, pluses). In the cooperation model (red), populations evolve higher cooperation levels in the patch carrying capacity scenario, but altruists have lower cooperation equilibria under the patch versus the individual carrying capacity scenario (Fig. 5.1b). Dispersal costs impact the rate of evolution and the equilibrium level of the dispersal propensity phenotypes in most simulations (Fig. 5.2; Appendix 6). In particular, the altruism individual carrying capacity model (Fig. 5.2c) shows a much more gradual change in the 120  dispersal propensity and cooperation phenotypes when no dispersal cost is imposed versus when a moderate dispersal cost is imposed (Fig. 5.2f). Other simulations show a similar, though less extreme, pattern (Fig. 5.2). In the tolerance model, dispersal costs have clear effects on the dispersal propensity and more subtle effects on the tolerance phenotype in both the individual and the patch carrying capacity scenarios (Fig. 5.2a, d, g, j). In the cooperation and altruism models, dispersal cost is more influential under the patch carrying capacity scenario (Fig. 5.2k, l). Selective pressure on dispersal influences the evolutionarily stable state of cooperation more in the altruism models than in the cooperation models. The two ecological scenarios also influence the group dynamics of each population in the cooperation and altruism models in very different ways (Fig. 5.3). In the individual carrying capacity scenario, selection works to maximize fecundity, as manifested in mean population sizes at intermediate levels, and relatively high fecundity rates. In the patch carrying capacity scenario, on the other hand, group size oscillates around the stable group size identified by Avilés (1999), where groups are comparatively large, and where the mean per capita rate of growth is equal to one (Fig. 5.3f). Raising the dispersal costs in these scenarios decreases the equilibrium group size (thereby increasing fecundity slightly), particularly in the patch carrying capacity scenario. 5.4 DISCUSSION This model explores how dispersal and three different social traits interact over evolutionary time under two simple ecological scenarios. In general, the influence of dispersal cost on sociality was unexpectedly subtle in all of the simulations. In many instances, the social traits were not inversely proportional to dispersal propensity as we expected. This suggests that dispersal and colony establishment risks may not be enough by themselves to drive the evolution of sociality. Some of our results did conform to our first hypothesis, that sociality should increase when dispersal is more costly. For instance, tolerance levels were marginally higher when dispersal was most costly compared to when dispersal was less costly (Figs. 5.1a, 5.2). In this model, mean tolerance level represents a balance between tolerant and solitary individuals, so values closer to zero indicate increasingly strong selection against tolerant individuals, whereas values closer to one indicate increasingly strong selection against 121  solitary individuals. Interestingly, for altruism and cooperation, our two continuously varying social traits, sociality only increased when dispersal was more costly in the patch carrying capacity scenario. The exceptions to this pattern are also interesting to consider. Under the individual carrying capacity scenario, the altruism and cooperation models actually had lower mean cooperation levels in simulations with increasingly high dispersal risks (opposite of expected in H1; Fig. 5.1b, circles). In this case, the cooperation level is initially highest in simulations with a large dispersal cost. However, dispersal propensity decreases rapidly and approaches zero earlier in these simulations than in simulations in which dispersal carries no extra cost (Fig. 2). When selection against dispersal is reduced as dispersal approaches zero, cooperation values level off. Thus, although cooperation increases most rapidly in the early generations of simulations with costly dispersal, it levels off sooner and at slightly lower levels. In most cases, increasing the cost of each social trait resulted in an increase in dispersal propensity at all dispersal cost values, thus supporting our second hypothesis (Fig. 1). However, this was not the case in the comparison between cooperation and altruism under the patch carrying capacity scenario (contrary to H2; Fig 5.1b, blue pluses). This result can be attributed to the difference in the fecundity cost experienced by cooperators entering a patch with non-cooperators in the altruism versus the cooperation models. In the altruism model, dispersers may suffer both the cost of dispersal and an added relative fecundity cost if they move to a patch in which the mean cooperation level is less than their own cooperation phenotype. The additional relative fecundity cost would be particularly likely and severe early in the simulation when altruists would be rare mutants. This added selection against dispersers prevents dispersal from reaching the highest level in these simulations, even when dispersal does not carry an additional mortality cost. Dispersers in the cooperation model, on the other hand, will always produce as many offspring as other members of the groups they join, so there is no additional selection constraining dispersal. On the contrary, individuals in the patch carrying capacity cooperation model should be under selection for greater dispersal propensities to prevent groups from becoming over-crowded, to increase the chances that some offspring will land in a patch that survives random patch extinction, and to have the opportunity to recolonize empty patches (Fig. 5.3f). Both Le Galliard et al. (2005) and Vainstein et al. (2007) also reported some unexpectedly complex co-evolutionary dynamics 122  between mobility and altruism levels, where helping could increase with higher dispersal rates under specific ecological conditions. The results did not support our third hypothesis that tolerance levels, in particular, would show a strong linear increase with increasing selection against dispersers. Instead, we found altruism in the patch carrying capacity scenario to be the most sensitive to different levels of selection against dispersal (Fig. 5.1b, blue pluses). Tolerance and cooperation equilibrium values were not influenced as strongly by differences in dispersal propensity. Neither of these behaviours results in a large fitness cost to individuals that disperse to patches with a less favourable social context. On the other hand, the weak altruists are more dependent on finding themselves in a patch with other, similar individuals. In this respect, only our weak altruism model resembles previous models of the co-evolution between strong altruism and dispersal, in that a dispersal risk may be required to drive the evolution of this costly helping behaviour (Perrin & Lehmann 2001; Lehmann & Perrin 2002; Lehmann et al. 2006). However, the weak altruism model does not require the kin-recognition or phenotype matching abilities that are cornerstones of the previous strong altruism models. As expected (H4), we found that the highest cooperation levels were reached in the cooperation model with the patch carrying capacity scenario and a high dispersal cost (Fig. 5.1). Interestingly, the altruism model with the patch carrying capacity scenario and small or moderate dispersal costs reached the lowest cooperation values, thus supporting our fifth hypothesis. Both the cooperation and altruism models with individual carrying capacities showed intermediate cooperation values (Fig. 5.1b). Because the patch dynamics in the patch carrying capacity scenario were unstable, altruistic dispersers were unlikely to find other cooperators. This imposed strong selection against cooperation that was entirely absent from the cooperation model, where cooperators receive the same fitness as non-cooperators within each group. In the individual carrying capacity scenario, groups were much more stable and long-lived. This allowed dispersal propensity to evolve to very low levels. In both the altruism and cooperation models under this scenario, cooperation trait values levelled off when dispersal propensity approached zero. This, in turn, prevented cooperation levels from reaching the extremely high values attained in the patch carrying capacity cooperation model, where the dispersal rate was maintained at relatively high levels. In the latter case, a constant influx of mutation was the primary factor preventing cooperation from 123  fixing in the population, particularly when dispersal was costly. In the individual carrying capacity altruism model, the presence of stable groups allowed cooperation to evolve in a favourable social context and mitigated the risk of an altruist landing in a group with lower cooperation values. The results also supported our hypotheses regarding the two ecological scenarios investigated. We hypothesized that imposing a global individual carrying capacity would cause the population to maximize fecundity (H6). In this case, having more offspring would increase a parent’s chance of producing young that survived the random mortality each generation. We suggested that individuals in the cooperation and altruism models would evolve to have relatively high levels of cooperation, thus enabling them to attain higher individual fitness rates on average (Fig. 5.3). In fact, we did find that cooperation evolved to relatively high levels in both the altruism and cooperation models (Figs. 5.1b, 5.2). In the cooperation models, individuals in both ecological scenarios attained very high cooperation levels, but the mechanism maintaining the cooperation levels was different. In the individual carrying capacity model, selection for increasing cooperation levels decreased as the dispersal propensity levelled off at very low levels. In the patch carrying capacity scenario, cooperation levels were maintained by a combination of factors, including the tension between selection against dispersal (imposed) versus selection for dispersal (a product of the patch carrying capacity scenario) and mutations preventing cooperation from reaching fixation. We also found that selection maintained higher dispersal levels under the patch carrying capacity scenario than under the individual carrying capacity scenario (H7; Fig. 5.2). In both the cooperation and the altruism models, dispersal propensity was maintained at relatively high levels in the population even when dispersal was very costly (Fig. 5.1b, pluses), because under this scenario dispersers had the opportunity to spread into empty habitat patches. As predicted, we observed very low equilibrium cooperation values in the patch mortality scenario for the altruism model. This result is similar to the findings of Lehmann et al. (2006) in their environmental stochasticity model. They suggested that, without propagule dispersal (see also Alizon & Taylor 2008), the evolution of cooperation depends primarily on the direct cost of the helping behaviour. Accordingly, we did not observe a clear effect of dispersal propensity in the cooperation model. The patch carrying 124  capacity scenario also resulted in the formation of larger groups, which caused the average individual fitness to be lower in these models (Fig. 5.3). In the tolerance model, on the other hand, we found that the evolutionary trajectories were not strongly influenced by the ecological scenario. This model highlights some key differences that emerge when different helping traits are permitted to co-evolve with dispersal. First, we showed that social tolerance evolved to near fixation, even when tolerance was moderately costly, regardless of the degree of selection against dispersal (Fig. 5.1a). When tolerance was very costly, neither the cost of dispersal behaviour nor the ecological scenario seem to influence the strong selection against tolerance. Second, helping that was not costly (cooperation) evolved to extremely high levels under different ecological scenarios, even when the dispersal rate was high (Fig. 2b). This result mirrors the findings of Avilés et al. (2002) in their model of cooperation and grouping. However, the addition of dispersal behaviour in the current model caused the evolutionary trajectories of cooperation and altruism to be more influenced by the two ecological scenarios (Fig. 5.3), in contrast with previous findings (Avilés et al. 2002). Third, we found that weak altruism attained higher levels when dispersal was costly and/or when a global individual carrying capacity was imposed than were ever attained under any conditions in the cooperation and grouping co-evolutionary model described by Avilés et al. (2002). These findings may be relevant to the future study of many empirical social systems. As Avilés et al. (2002) pointed out, many social systems, ranging from social bacteria (Velicer et al. 2000) to tree-killing bark beetles (Raffa & Berryman 1987) to colonial nesting birds (e.g. Brown & Brown 1996) resemble the sociality model developed here. The novel contribution of our model is to combine previous investigations of these less costly social behaviours with the body of theory exploring the co-evolutionary dynamics between dispersal and altruism. We find that some social behaviours that do not necessarily carry a direct fitness cost to the actor, including tolerance and cooperation, can invade a population very rapidly even when dispersal rates are very high. This is not a surprising result, but is probably extremely important in the formation of large groups of non-kin, which share a common ecological benefit from grouping, as in the tree-killing bark beetles in their joint attacks on trees (Raffa & Berryman 1987). In keeping with previous studies of strong 125  altruism and mobility (Perrin & Lehmann 2001; Lehmann & Perrin 2002; Le Galliard et al. 2005), we find that the evolution of weak altruism is more sensitive to both dispersal risk and other environmental factors that may limit the population size. Overall, the findings of this model open up several future empirical and theoretical avenues of study. First, the evolutionary dynamics described in this and previous models have not yet been explored in natural systems, and have rarely been investigated in the lab (Kummerli et al. 2009). Our results suggest that comparing intrinsic dispersal propensities and rates in different social systems would be interesting and worthwhile. Based on this model, social traits are not necessarily inversely proportional to dispersal propensity, suggesting that risky dispersal and risky colony foundation may not be sufficient in themselves to select for sociality. We would, however, expect the variance in dispersal ability and/or behaviour to increase as the cost of social behaviours decreases. Second, we find that different population level demographic constraints can impose very different selection regimes on social systems. This difference may be readily detectable in natural systems. For example, Avilés and Tufiño (1998) found that the social spider Anelosimus eximius forms extremely large groups that far exceed the intermediate group size at which they would maximize reproduction. However, Salazar and Avilés (in prep) found that another social spider, Anelosimus guacamayos, seems to disperse when groups reach intermediate sizes at which fecundity is maximized. We do not yet know what causes this difference in group size, but the results of this model suggest that differential mortality at the colony versus the individual level could contribute to such a difference. This hypothesis could be tested empirically in this and other natural systems. We suggest that several extensions of this model would also be interesting to explore in the future. First, we would like to explore more context-dependent dispersal behaviours. Previous studies have demonstrated that dispersal via propagule formation can expand the conditions under which strong altruism can evolve (Lehmann et al. 2006; Alizon & Taylor 2008). It follows that other dispersal strategies observed in natural social systems may also expand the ecological conditions under which altruism can evolve. We also suggest that modelling ecological conditions that do not interact with the characteristics of social groups is naïve. Thus, we secondarily plan to explore the influence of several ecologically relevant  126  population control mechanisms that would feedback with social groups in different ways (e.g. predation, parasitism, etc.). 5.4.1 Conclusions The results of this model clearly demonstrate that not all social behaviours are equal in their interactions with dispersal. Strong altruism, which has been explored extensively in previous models (Perrin & Lehmann 2001; Lehmann & Perrin 2002; Le Galliard et al. 2005), can only evolve when kinship and ecological conditions are just right. We see the effect of these constraints in our model of weak altruism, but we find that kin recognition is no longer required to support the spread of this behaviour. The constraints on the evolution of cooperation (no cost) and tolerance are further reduced, such that both behaviours can invade a population under a broad range on conditions. Interestingly, dispersal seems to have less impact on these behaviours, but does impact the rate of evolution for all three social traits.  127  Table 5.1: Description of parameters and list of baseline values. Notation  Baseline Value  Definition  gen num_sites  10,000 200  number of generations run in each simulation number of habitat patches available  disp_risk  0, 0.25, 0.5, 0.75  probability of mortality for dispersing individuals, simulated from no risk (0) to very risky (0.75)  mutation_rate  0.01  probability that a mutation will occur in a single locus (of an offspring during reproduction)  mut_stdev Individual Carrying Capacity Scenario  0.1  magnitude of mutation (sample standard deviation)  Carrying_capacity  num_sites (200)  number of individuals that can survive in the habitat  Patch_mort  num_sites/2 (100)  number of patches that can be occupied in each generation  Tolerance Specific Parameters max_group_size sol_fecundity tol_fecundity  5 5 2,3,5  maximum number of individuals that can survive in a single patch average fecundity experienced by non-tolerant individuals average fecundity experienced by tolerant individuals  r  1  models the intrinsic rate of growth for individuals living alone, and is the maximum mean fitness that can be attained by non cooperative individuals (appendix 5)  c  0.05  crowding parameter, determines the group size at which the costs of local competition begin to outweigh the benefits of grouping (appendix 5)  Altruism Specific Parameters beta  0.8  imposes relative fitness cost on cooperators (appendix 5)  General Parameters  Patch Carrying Capacity Scenario  Cooperation Specific Parameters  128  129  Figure 5.1: Equilibrium values for the last 5000 generations, averaged over at least 100 replicates, are shown for each parameter combination. (A) In the tolerance model, tolerance cannot evolve when subjected to a high fecundity cost (blue), but is maintained at very high levels under moderate cost (green) or no cost (red) models. The ecological scenarios (individual carrying capacity, pluses; patch carrying capacity, circles) have relatively little effect on the mean tolerance level or dispersal propensity. Increasing the dispersal cost from no cost (bold) to high cost (fine, dispersal cost=0.75) strongly influences the equilibrium dispersal propensity in all of the tolerance simulations, but has a relatively small effect on the tolerance level. (B) In the cooperation models (red), the equilibrium cooperation level is always significantly higher than the equilibrium levels reached in the altruism models (blue). The patch carrying capacity scenario (pluses) results in substantially higher dispersal propensities than the individual carrying capacity scenario (circles) in both the cooperation and the altruism models. Increasing dispersal costs (bold-fine) have different effects on both dispersal propensity and cooperation level in each distinct scenario.  130  131  Figure 5.2: A time series of 1000 generations, averaged over at least 100 iterations for each model type (top row: tolerance; middle row: cooperation; bottom row: altruism), under each ecological scenario (individual carrying capacity, left; patch carrying capacity, right), and for either no dispersal costs (left) or moderate dispersal costs (right). In the tolerance model (A, D, G, J; simulations of intermediate tolerance cost), the transition from no dispersal costs to moderate dispersal costs results in decreased dispersal propensities (red) and slightly increased mean tolerance levels (blue). Dispersal is maintained at slightly higher levels in the patch carrying capacity scenario. In the cooperation model (B, E, H, K), dispersal propensity evolves to very low levels in the individual carrying capacity scenarios, but is maintained at relatively high levels in the patch mortality scenario. The initial dip in dispersal propensity (K) can be attributed to the cost of dispersal. This selection pressure is counteracted by the invasion of cooperation behaviours after about 500 generations. In the altruism model (C, F, I, L), the rate of evolution of both cooperation and the decrease in dispersal are both substantially slower in the no dispersal cost simulations relative to the moderate dispersal cost runs. Imposing an individual carrying capacity results in a moderate level of cooperation, whereas cooperation is maintained at low levels in the patch carrying capacity scenario.  132  133  Figure 5.3: Group and fecundity frequencies are compared with the per capita growth rate by group size model generated by equation 1 (r and c=baseline conditions, γ=0.9; Avilés 1999; Appendix 5). Each scenario shows the dynamics of the cooperation model with moderate dispersal cost (0.5), and includes the group size and fecundity value for each occupied patch for the last 100 generations of a simulation repeated 20 times. In the individual carrying capacity scenario, the mean group size oscillates around the high point of the unimodal curve, maximizing the fecundity rate. Because this is not a stable equilibrium point and because mortality is random, groups often over or undershoot the optimum group size (C), so the mean fecundity rate is usually somewhat lower than the theoretical maximum (B). Because dispersal is low in these simulations, individuals that do disperse are also likely to land in an empty patch, which explains the relatively large numbers of solitary individuals or small groups (A). The patch carrying capacity dynamics, on the other hand, result in group sizes oscillating around the stable equilibrium point where the curve crosses the point at which the per capita rate of growth equals 1 (F; Avilés 1999).  134  5.5 REFERENCES Alizon, S. & Taylor, P. 2008. Empty sites can promote altruistic behavior. Evolution 62, 1335–1344. Avilés, L. 1999. Cooperation and non-linear dynamics: An ecological perspective on the evolution of sociality. Evol. Ecol. Res. 1, 459-477. Avilés, L. & Tufiño, P. 1998. Colony size and individual fitness in the social spider Anelosimus eximius. Am. Nat. 152, 403-418. Avilés, L. Abbot, P. & Cutter, A.D. 2002. Population ecology, nonlinear dynamics, and social evolution. I. Associations among nonrelatives. Am. Nat. 159, 115-127. Avilés, L., Fletcher, J.A., & Cutter, A.D. 2004. The kin composition of social groups: trading group size for degree of altruism. Am. Nat. 164, 132-144. Bengtsson, B.O. 1978. Avoiding inbreeding- at what cost. J. Theor. Biol. 73, 439-444. Brown, C.R. & Brown, M.B. 1996. Coloniality in the cliff swallow: the effect of group size on social behavior. Chicago, IL: University of Chicago Press. Chapuisat, M. & Keller, L. 1999. Testing kin selection with sex allocation data in eusocial Hymenoptera. Heredity 82, 473-478. Clutton-Brock, T. 2002. Breeding together: Kin selection and mutualism in cooperative vertebrates. Science 296, 69-72. Dieckmann, U., O’Hara, B., & Weisser, W. 1999. The evolutionary ecology of dispersal. Trends Ecol. Evol. 14, 88-90. Emlen, S.T. 1991. Cooperative breeding in birds and mammals. In Behavioral Ecology: An Evolutionary Approach (eds. J. R. Krebs & N. B. Davis), pp. 305-339. Oxford, UK: Blackwell Scientific. Hamilton, W.D. 1964. The genetical evolution of social behaviour. I & II. J. Theor. Biol. 7, 1-52. Hamilton, W.D. & May, R.M. 1977. Dispersal in stable habitats. Nature 269, 578-581. Jarvis, J.U.M., O’Riain, M.J., Bennett, N.C., & Sherman, P.W. 1994. Mammalian eusociality: A family affair. Trends Ecol. Evol. 9, 47-51. Johnson, M.L. & Gaines, M.S. 1990. Evolution of dispersal- theoretical models and empirical tests using birds and mammals. Ann. Rev. Ecol. Syst. 21, 449-480.  135  Kummerli, R. Gardner, A., West, S.A., & Griffin, A.S. 2009. Limited dispersal, budding dispersal, and cooperation: An experimental study. Evolution in press. Le Galliard, J., Ferrière, R. & Dieckmann, U. 2005. Adaptive evolution of social traits: Origin, trajectories, and correlations of altruism and mobility. Am. Nat. 165, 206-224. Lehmann, L. & Keller, L. 2006. The evolution of cooperation and altruism- a general framework and a classification of models. J. Evol. Biol. 19, 1365-1376. Lehmann, L. & Perrin, N. 2002. Altruism, dispersal, and phenotype-matching kin recognition. Am. Nat. 159, 451-468. Lehmann, L., Perrin, N., & Rousset, F. 2006. Population demography and the evolution of helping behaviors. Evolution 60, 1137-1151. Lion, S. & van Baalen, M. 2008. Self-structuring in spatial evolutionary ecology. Ecol. Lett. 11, 277-295. Michod, R.E. 1982. The theory of kin selection. Ann. Rev. Ecol. Syst. 13, 23-55. Perrin, N. & Lehmann, L. 2001. Is sociality driven by the costs of dispersal or the benefits of philopatry? A role for kin-discrimination mechanisms. Am. Nat. 158, 471-483. Pfeiffer, T., Rutte, C., Killingback, T., Taborsky, M., & Bonhoeffer, S. 2005. Evolution of cooperation by generalized reciprocity. Proc. R. Soc. B 272, 1115-1120. Pollock, G.B. & Cabrales, A. 2008. Suicidal altruism under random assortment. Evol. Ecol. Res. 10, 1077-1086. Pulliam, H.R. & Caraco, T. 1984. Living in groups: Is there an optimal group size? In Behavioral Ecology: An Evolutionary Approach (eds. J.R. Krebs & N.B. Davis), pp. 305-339. Oxford, UK: Blackwell Scientific. Queller, D.C. & Strassman, J.E. 1998. Kin selection and social insects. Bioscience 48, 165175. Raffa, K.F. & Berryman, A.A. 1987. Interacting selective pressures in conifer-bark beetle systems: A basis for reciprocal adaptations? Am. Nat. 129, 234-262. Slobodchikoff, C.N. 1984. Resources and the evolution of social behavior. In A New Ecology: Novel Approaches to Interactive Systems (eds. P.W. Price, C.N. Slobodchikoff, & W.S. Gaud), pp. 227-251. New York, NY: Wiley. Taylor, P.D. 1992. Inclusive fitness in a homogeneous environment. Proc. R. Soc. B 249, 299-302. 136  Trivers, R.L. 1971. The evolution of reciprocal altruism. Q. Rev. Biol. 46, 1-35. Vainstein, M.H., Silva, A.T.C., & Arenzon, J.J. 2007. Does mobility decrease cooperation? J. Theor. Biol. 244, 722-728. Vehrencamp, S.L. 1983. A model for the evolution of despotic versus egalitarian societies. Anim. Behav. 21, 667-682. Velicer, G.J., Kroos, L., & Lenski, R.E. 2000. Developmental cheating in the social bacterium Myxococcus xanthus. Nature 404, 598-601. Verdolin, J.L. & Slobodchikoff, C.N. 2008. Resources, no kinship, determine social patterning in the territorial Gunnison’s prairie dog (Cynomys gunnisoni). Ethology 115, 59-69. West Eberhard, M.J. 1975. The evolution of social behavior by kin selection. Q. Rev. Biol. 50, 1-33. West, S.A., Griffin, A.S., & Gardner, A. 2007. Evolutionary explanations for cooperation. Curr. Biol. 17, R661-R672. Wilson, D.S. 1979. Structured demes and trait-group variation. Am. Nat. 113, 606-610. Wilson, D.S. 1980 The Natural Selection of Populations and Communities. Menlo Park, CA: Benjamin/Cummings. Wilson, D.S. 1990. Weak altruism, strong group selection. Oikos 59, 135-140. Wilson, D.S., Pollock, G.B. & Dugatkin, L.A. 1992. Can altruism evolve in purely viscous populations? Evol. Ecol. 6, 331-341. Wright, S. 1931. Evolution in Mendelian populations. Genetics 16, 97-159.  137  CHAPTER 6: GENERAL CONCLUSIONS 6.1 GENERAL OVERVIEW OF THE THESIS The manuscripts included in this thesis are a part of a recent resurgence of studies that focus on the ecological side of social evolution (Choe & Crespi 1997; Hunt & Amdam 2005; Costa 2006; Korb & Heinze 2008; Yip et al. 2008; Zammit et al. 2008). For the past 40 years, the majority of research on social organisms was focused on investigating the role of relatedness in the evolution of sociality (Hamilton 1964; Anderson 1984; Crozier & Pamilo 1996; West et al. 2007). With both elements combined, I believe future research will lead to a more comprehensive and general answer to the question of how and why social behaviours have evolved in such a diverse range of organisms, but are totally absent from others. In chapter 2 (Purcell & Avilés 2007), I document a clear pattern of decreasing sociality with increasing elevation in social spider populations along the eastern slope of the Andes in Ecuador. I developed this idea in the context of similar research performed on halictine and allodapine bees (Packer 1990; Eickwort et al. 1996; Cronin & Schwarz 1999; Cronin & Hirata 2003). These bees are considered to be an ideal system for investigating questions of how, why, and where sociality has evolved, because they contain so much variation in social behaviour and because sociality has evolved repeatedly in both lineages (Schwarz et al. 2007). In the spider genus Anelosimus, Furey (1998) first documented among-population differences in the number of adult spiders sharing nests in two temperate populations of the subsocial spider Anelosimus studiosus. This chapter, along recent studies documenting an interspecific gradient of sociality (Avilés et al. 2007) and demonstrating that sociality has arisen repeatedly in the genus (Agnarsson 2006; Agnarsson et al. 2006), suggests that Anelosimus may be another group (along with halictine and allodapine bees) in which to explore the conditions that facilitate the evolutionary origin of sociality. In my review manuscript (chapter 1), I synthesize other studies that have reported biogeographic patterns in the distribution of sociality. I believe that comparative studies across social systems should help us to form a clearer idea of what factors commonly contribute to the evolution of social behaviours. Some of the mechanisms driving the patterns of social distribution documented in chapter 2 are explored experimentally in chapter 3. I transplant small groups of subsocial 138  spiders within their own habitat, to their lower range boundary, and to the lowland tropical rainforest where they do not naturally occur. I find that subsocial spiders transplanted to the lowland rainforest ultimately cannot survive, and that groups transplanted to the lower range boundary show decreased dispersal tendencies and increased mortality relative to the groups transplanted within their own habitat. I suggest that small groups and solitary spiders face greater risks and, conversely, that large groups are increasingly beneficial as one approaches the lowland tropical rainforest. I correlate these results with two specific environmental factors, rainfall intensity and predator abundance, both of which increase with decreasing elevation. By sheltering small groups of social and subsocial spiders from the rain in the lowland rainforest, I demonstrate that rainfall causes both increased individual mortality and decreased web construction. Related studies of social spiders have recently demonstrated that a third factor, prey size, decreases with increasing elevation and latitude (Guevara & Avilés 2007; Powers & Avilés 2007). Yip et al. (2008) showed that large social groups must capture relatively large insect prey to provide adequate nutrition to their members. Thus it appears that a lack of large prey may prevent large groups from colonizing higher elevation habitats in the tropics, while other factors including intense rainfall and abundant predators may select against small groups (both subsocial and social) in the lowland rainforest (Avilés & Tufiño 1998; Avilés et al. 2007; Purcell & Avilés 2007, 2008). Both rainfall and predation have been suggested to influence sociality in a similar manner in other social organisms. Intense rainfall is expected to be particularly detrimental to groups forming relatively fragile, but long lasting nest structures, including other social spiders (Riechert et al. 1986), ants (Murphy & Breed 2007), termites (Picker et al. 2006) and wasps (Barlow et al. 2002). As far as I know, however, none of these earlier studies have demonstrated this experimentally, as was done here. Defence against predation has also been suggested to be one of the major benefits of sociality in many organisms (Lin & Michener 1972; Jeanne 1991; Henschel 1998; Pike & Foster 2008). Ants are known to be one of the most important predators of other social insects (Hölldobler & Wilson 1990). My study is one of several papers that documents increasing ant abundance with decreasing elevation (Janzen et al. 1976; Samson et al. 1997; Bruhl et al. 1999; Purcell & Avilés 2008; Guevara & Avilés in press) and decreasing latitude (Jeanne 1975, 1979; Hölldobler & Wilson 1990). So far, much  139  of the evidence supporting this hypothesis is correlational, and the idea has been explored experimentally in relatively few natural systems (e.g. Zammit et al. 2008). As we discovered in chapters 2 and 3, social Anelosimus species are generally restricted to the lowland tropical rainforest, while subsocial species are abundant at higher latitudes and at higher altitudes in the tropics (Avilés et al. 2007; Purcell & Avilés 2007, 2008). We find that these species tend to overlap at intermediate latitudes and altitudes. The fourth chapter presents the novel idea that social behaviour may facilitate species coexistence in these zones of overlap. Specifically, I investigate the idea that social and subsocial species coexist in one particularly speciose area in Brazil by differing in their spatial distributions and their phenologies. A companion paper by Guevara et al. (in prep) explores the dietary dimension in the same habitat. I find that the two species with higher degrees of sociality tend to occur inside the forest, while the three subsocial species tend to occur at the forest edge. This finding supports our hypothesis that the position of social and subsocial nests is dictated by the specific habitat requirements of the nests built by each behavioural type. As in chapter 2, the size and continuity of the plant substrate where the nest is constructed may be particularly important in driving the habitat requirement of species with large versus small nests. Within these habitats, I document further segregation on a local spatial scale and in phenology, thus supporting the hypothesis of limiting similarity. Although many studies have investigated how similar sympatric species partition their habitat (e.g. MacArthur 1958; Schoener 1968; Pianka 1969; Schluter 1996; Kronfeld-Schor & Dayan 1999; Harmon et al. 2007; Behmer & Joern 2008), I do not know of any other studies that have yet explored how sociality may shape community dynamics. Each of the three empirical chapters provides novel insight into the interplay between ecology and sociality in Anelosimus spiders. Modelling complements these empirical studies by permitting the exploration of the role of different behaviours and ecological factors over evolutionary time, thereby facilitating a more general understanding of the selective pressures that act on each behaviour under a wide variety of simple conditions. In the last several years, there has been a great deal of theoretical interest in the influence of dispersal on the evolution of strong altruism (Perrin & Lehmann 2001; Lehmann & Perrin 2002; Le Galliard et al. 2005; Lehmann et al. 2006; Vainstein et al. 2007; Alizon & Taylor 2008; Kummerli et al. 2009). At the same time, over the last few decades, many theoretical (e.g. 140  Vehrencamp 1983; Pulliam & Caraco 1984; Slobodchikoff 1984; Emlen 1991; Avilés 1999; Avilés et al. 2002) and empirical studies (e.g. Jarvis et al. 1994; Pollock & Cabrales 2008; Verdolin & Slobodchikoff 2008) have suggested that many real social systems may not experience the extreme costs of cooperation that are implicit in these models of strong altruism. Thus, chapter 5 seeks to marry these two bodies of literature by exploring the coevolution of three less costly social behaviours (weak altruism, cooperation, and tolerance) with dispersal behaviours under two different ecological scenarios. We find several key results that should inform future empirical work on social spiders and other cooperative societies. First, we show that the evolutionary equilibria of tolerance and cooperation are not strongly influenced by selection against dispersal. Second, we demonstrate that many of the specific conditions required for the evolution of strong altruism are relaxed in our models of less costly social behaviours, including the need for kin recognition or phenotype matching (Perrin & Lehmann 2001; Lehmann & Perrin 2002). Finally, we show that the group dynamics that emerge in the cooperation and altruism models are strongly influenced by whether random mortality is imposed at the individual or at the patch level. In chapter 3, I suggest that strong selection against dispersers in the lowland tropical rainforest may be one key factor driving the distribution of sociality documented in chapter 2 and by Avilés et al. (2007). However, the results of chapter 5 and previous models (Le Galliard et al. 2005; Vainstein et al. 2007) suggest that selection against dispersal alone probably cannot drive the evolution of social behaviours, unless sociality provides additional benefits. This suggests that both synergistic benefits reaped by individuals in groups, in this case through access to large prey (Avilés 1999; Yip et al. 2008) and selection against small groups or solitary dispersers (chapter 3; Purcell & Avilés 2008) are probably both necessary to drive the evolution of spider sociality in the tropics. The model also provides some possible insight into a puzzling pattern that was documented both in chapter 2 (Purcell & Avilés 2007) and in previous studies of social spiders, showing that some populations and species form much larger groups than others (Avilés & Tufiño 1998; Salazar & Avilés unpublished manuscript). Based on the model results, I speculate that some habitats have relatively high individual mortality rates, so selection should optimize group sizes at intermediate levels, thereby maximizing the per capita rate of growth (Avilés 1999). In habitats where individual mortality is relatively low, but whole groups may be subject to a 141  higher probability of extinction, the model suggests that extremely large groups should form, particularly if larger groups are buffered against extinction, as I suggested in chapter 3. 6.2 FUTURE DIRECTIONS One of the major strengths of this thesis is to put forward new and interesting ideas that will hopefully ignite future research. Because several of the studies I pursued herein addressed questions that have not yet been explored, much of this work has been somewhat preliminary. I therefore look forward to following up on some of the possibilities raised in these chapters with a more experimental approach. The patterns of sociality documented in chapters one and two, for instance, present many exciting possibilities for future research, as identifying geographic patterns in the distribution of social organisms, both invertebrate and vertebrate, should provide key insight into the ecological drivers of social evolution. The fact that a variety of social organisms show latitudinal and altitudinal patterns in the distribution of their social systems (chapter 1) suggests that there may be common thread shaping the evolution of such diverse social systems. Along these lines, I hope to pursue future studies on the ecology of sociality in a systematic way. To follow up on one of the unanswered questions in this thesis, for example, I would like to experimentally test the effect of the increased abundance of predators that I observed in chapter 3 on dispersers, small groups, and large groups. A study of African social spiders from the genus Stegodyphus found that in populations where predators (ants) were common, spiders formed large groups, whereas in areas where predators were uncommon, spiders were more likely to form small groups or establish nests alone (Henschel 1998). I hope to follow up on these studies with a manipulative experiment. Unfortunately, I was not able to design a means of preventing ants from accessing colonies, but I could instead place ant baits near spider colonies of different sizes. I would then observe the behaviours of spiders in the presence of ants, and the behaviours of different ant species approaching or entering spider nests under more controlled conditions. I predict that ants would enter spider colonies of all sizes, but that the spiders would be more likely to attack ants or retreat to nest areas that ants could not access in larger colonies. I would like to then repeat this experiment or, when possible, to apply the predator exclusion experimental design of Zammit et al. (2008), to other social arthropods that show similar social distribution 142  patterns. Likewise, I hope to repeat my rain exclosure experiment (chapter 3) both on larger social spider groups in the lowland tropical rainforest and on social spiders living in other habitats with less intense rainfall, as well as in other social organisms that may be influenced by rainfall. This thesis also usefully combines three different approaches to research: a literature review, three empirical studies, and a simulation model. Theory should be used as a tool to explore problems that cannot be studied empirically, but should always be relevant to real systems and real questions. Thus I also plan to investigate the evolution of cooperation and altruism under more realistic ecological scenarios, including predation and parasitism, which will respond to the level of sociality in each group. Specifically, I would like to determine how each of these environmental factors would influence the evolutionary trajectory of different social traits in isolation and in combination. I hope that this approach of coordinating theoretical and empirical research programs will provide further insights into how social dynamics are influenced by specific factors in real, complex systems. The fourth chapter of the thesis suggests that sociality may facilitate species coexistence by separating social and subsocial species into distinct habitats. While this is an exciting result, I cannot yet determine whether other similar communities are structured in the same way, or whether this structure is specific to the study area in which the research was conducted. Future research should repeat this study in other Anelosimus communities and in communities of other organisms, so that we can determine whether social traits facilitate species coexistence more generally. Another extension to this preliminary work would be to investigate whether character displacement of social traits may occur. This investigation would require the detailed study of each species both in shared habitat and alone. While this may not be plausible for all of the species described in chapter 3, there are many areas in Ecuador where individual species could be studied in isolation and compared to areas where they overlap with other species. If sociality is truly important in facilitating species coexistence in these intermediate habitats, we would expect social species to be more social and subsocial species to be less social in areas where they overlap with one another. The results of my second chapter documenting a gradient of sociality in the social spider Anelosimus eximius suggest that this is not always the case, since this species becomes more “subsocial-like” in populations where it overlaps with subsocial species. 143  6.3 SIGNIFICANCE Ultimately, the chapters in this thesis suggest that the interactions between ecological factors and social behaviours are very influential, both over evolutionary time in the emergence of new social traits, and over ecological time in the formation of communities. I hope that this contribution, along with recent studies exploring similar questions in other social organisms, facilitates the return of the study of social biology to a more balanced position that considers both ecology and relatedness as necessary ingredients in the emergence and persistence of social behaviours. 6.4 REFERENCES Agnarsson, I. 2006. A revision of the New World eximius lineage of Anelosimus (Araneae, Theridiidae) and a phylogenetic analysis using worldwide exemplars. Zool. J. Linn. Soc.-Lond. 146, 453-493. Agnarsson, I., Avilés, L., Coddington, J. A., & Maddison, W. P. 2006. Sociality in theridiid spiders: repeated origins of an evolutionary dead end. Evolution 60, 2342-2351. Alizon, S. & Taylor, P. 2008. Empty sites can promote altruistic behavior. Evolution 62, 1335–1344. Anderson. M. 1984. The evolution of eusociality. Ann. Rev. Ecol. Syst. 15, 165-189. Avilés, L. 1999. Cooperation and non-linear dynamics: An ecological perspective on the evolution of sociality. Evol. Ecol. Res. 1, 459-477. Avilés, L. & Tufiño, P. 1998. Colony size and individual fitness in the social spider Anelosimus eximius. Am. Nat. 152, 403-418. Avilés, L. Abbot, P. & Cutter, A.D. 2002. Population ecology, nonlinear dynamics, and social evolution. I. Associations among nonrelatives. Am. Nat. 159, 115-127. Avilés, L., Agnarsson, I., Salazar, P., Purcell, L., Iturralde, G., Yip, E., Powers, K.S. & Bukowski, T. 2007. Altitudinal patterns of spider sociality and the biology of a new mid-elevation social Anelosimus species in Ecuador. Am. Nat. 170, 783-792. Barlow, N.D., Beggs, J.R., & Barron, M.C. 2002. Dynamics of common wasps in New Zealand beech forests: a model of density dependence and weather. J. Anim. Ecol. 71, 663-671. 144  Behmer, S.T. & Joern, A. 2008. Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc. Natl. Acad. Sci. U.S.A. 105, 1977-1982. Bruhl, C.A., Mohamed, M. & Linsenmair, K.E. 1999. Altitudinal distribution of leaf litter ants along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. J. Trop. Ecol. 15, 265-277. Choe, J.C. & Crespi, B. 1997. The evolution of social behaviour in insects and arachnids. Cambridge, UK: Cambridge University Press. Costa, J.T. 2006. The Other Insect Societies. Cambridge, MA, USA: Belknap Press. Cronin, A.L. & Hirata, M. 2003. Social polymorphism in the sweat bee Lasioglossum (Evylaeus) baleicum (Cockerell) (Hymenoptera, Halictidae) in Hokkaido, northern Japan. Insectes Soc. 50, 379-386. Cronin, A.L. & Schwarz, M.P. 1999. Latitudinal variation in the life cycle of allodapine bees (Hymenoptera; Apidae). Can. J. Zool. 77, 857-864. Crozier, R.H. & Pamilo, P. 1996. Evolution of Social Insect Colonies: Sex Allocation and Kin Selection. Oxford, UK: Oxford University Press. Eickwort, G.C., Eickwort, J.M., Gordon, J. & Eickwort, M.A. 1996. Solitary behavior in a high altitude population of the social sweat bee Halictus rubicundus (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 38, 227-233. Emlen, S.T. 1991. Cooperative breeding in birds and mammals. In Behavioral Ecology: An Evolutionary Approach (eds. J. R. Krebs & N. B. Davis), pp. 305-339. Oxford, UK: Blackwell Scientific. Furey, R.E. 1998. Two cooperatively social populations of the theridiid spider Anelosimus studiosus in a temperate region. Anim. Behav. 55, 727-735. Guevara, J. & Avilés, L. 2007. Multiple sampling techniques confirm differences in insect size between low and high elevations that may influence levels of spider sociality. Ecology 88, 2015-2023. Guevara, J. & Avilés, L. In press. Elevational changes in insect and other arthropod composition at tropical latitudes: a comparison of multiple sampling methods and social spider diets. Insect Conserv. Biodiv. Hamilton, W.D. 1964. The genetical evolution of social behaviour. J. Theor. Biol. 7, 1-52. 145  Harmon, L.J., Harmon, L.L., & Jones, C.G. 2007. Competition and community structure in diurnal arboreal geckos (genus Phelsuma) in the Indian Ocean. Oikos 116, 1863-1878. Henschel, J.R. 1998. Predation on social and solitary individuals of the spider Stegodyphus dumicola (Araneae, Eresidae) J. Arachnol. 26, 61-69. Hölldobler, B. & Wilson, E.O. 1990. The Ants. Cambridge, MA: Belknap Press. Hunt, J.H. & Amdam, G.V. 2005. Bivoltinism as an antecendent to eusociality in the paper wasp genus Polistes. Science 308, 264-267. Janzen, D.H., Ataroff, M., Farinas, M., Reyes, S., Rincon, N., Soler, A., Soriano, P. & Vera, M. 1976. Changes in the arthropod community along an elevational transect in the Venezuelan Andes. Biotropica 8, 193-203. Jarvis, J.U.M., Oriain, M.J., Bennett, N.C. & Sherman, P.W. 1994. Mammalian eusociality – a family affair. Trends Ecol. Evol. 9, 47–51. Jeanne, RL. 1975. The adaptiveness of social wasp nest architecture. Q. Rev. Biol. 50, 267287. Jeanne, RL. 1979. A latitudinal gradient in rates of ant predation. Ecology 60, 1211-1224. Jeanne, RL. 1991. The swarm founding Polistinae. In The Social Biology of Wasps. (eds. K.G. Ross & R.W.), pp. 191-231. Ithaca, NY: Cornell University Press. Korb, J. & Heinze, J. 2008. Ecology of Social Evolution. Berlin, Germany: Springer-Verlag. Kronfeld-Schor, N. & Dayan, T. 1999. The dietary basis for temporal partitioning: food habits of coexisting Acomys species. Oecologia 121, 123-128. Kummerli, R. Gardner, A., West, S.A., & Griffin, A.S. 2009. Limited dispersal, budding dispersal, and cooperation: An experimental study. Evolution in press. Le Galliard, J., Ferrière, R., & Dieckmann, U. 2005. Adaptive evolution of social traits: Origin, trajectories, and correlations of altruism and mobility. Am. Nat. 165, 206-224. Lehmann, L. & Perrin, N. 2002. Altruism, dispersal, and phenotype-matching kin recognition. Am. Nat. 159, 451-468. Lehmann, L., Perrin, N., & Rousset, F. 2006. Population demography and the evolution of helping behaviors. Evolution 60, 1137-1151. Lin, N. & Michener, C.D. 1972. Evolution of sociality in insects. Q. Rev. Bio. 47, 131-159. MacArthur, R.H. 1958. Population ecology of some warblers of northeastern coniferous forests. Ecology 39, 599-619. 146  Murphy, C.M & Breed, M.D. 2007. A predictive distribution map for the giant tropical ant, Paraponera clavata. J. Insect Sci. 7, insectscience.org/7.08 Packer, L. 1990. Solitary and eusocial nests in a population of Augochlorella striata (Provancher) (Hymenoptera; Halictidae) at the northern edge of its range. Behav. Ecol. Sociobiol. 27, 339-344. Perrin, N. & Lehmann, L. 2001. Is sociality driven by the costs of dispersal or the benefits of philopatry? A role for kin-discrimination mechanisms. Am. Nat. 158, 471-483. Pianka, E. R. 1969. Sympatry of desert lizards (Ctenotus) in Western Australia. Ecology 50, 1012-1030. Picker, M.D., Hoffman, M.T., & Leverton, B. 2006. Density of Microhodotermes viator (Hodotermitidae) mounds in southern Africa in relation to rainfall and vegetative productivity gradients. J. Zool. 271, 37-44. Pike, N. & Foster, W.A. 2008. The ecology of altruism in a clonal insect. In Ecology of Social Evolution. (eds. J. Korb & J. Heinze), pp. 37-56. Berlin, Germany: SpringerVerlag. Pollock, G.B. & Cabrales, A. 2008. Suicidal altruism under random assortment. Evol. Ecol. Res. 10, 1077-1086. Powers, K.S. & Avilés, L. 2007. The role of prey size and abundance in the geographic distribution of spider sociality. J. Anim. Ecol. 76, 995-1003. Pulliam, H.R. & Caraco, T. 1984. Living in groups: Is there an optimal group size? In Behavioral Ecology: An Evolutionary Approach (eds. J.R. Krebs & N.B. Davis), pp. 305-339. Oxford, UK: Blackwell Scientific. Purcell, J. & Avilés, L. 2007. Smaller colonies and more solitary living mark higher elevation populations of a social spider. J. Anim. Ecol. 76, 590-597. Purcell, J. & Avilés, L. 2008. Gradients of precipitation and ant abundance may contribute to the altitudinal range limit of subsocial spiders: insights from a transplant experiment. Proc. R. Soc. B 275, 2617-2625. Riechert, S.E., Roeloffs, R. & Echternacht, A.C. 1986. The ecology of the cooperative spider Agelena consociata in equatorial Africa (Araneae, Agelenidae). J. Arachnol. 14, 175191. 147  Samson, D.A., Rickart, E.A. & Gonzales, P.C. 1997. Ant diversity and abundance along an elevational gradient in the Philippines. Biotropica 29, 349-363. Schluter, D. 1996. Ecological speciation in postglacial fishes. Philos. T. Roy. Soc. B 351, 807-814. Schoener, T.W. 1968. The Anolis lizards of Bimini: Resource partitioning in a complex fauna. Ecology 49, 704-726. Schwarz, M.P., Richards, M.H., & Danforth, B.N. 2007. Changing paradigms in insect social evolution: new insights from halictine and allodapine bees. Ann. Rev. Entomol. 52, 127150. Slobodchikoff, C.N. 1984. Resources and the evolution of social behavior. In A New Ecology: Novel Approaches to Interactive Systems (eds. P.W. Price, C.N. Slobodchikoff, & W.S. Gaud), pp. 227-251. New York, NY: Wiley. Vainstein, M.H., Silva, A.T.C., & Arenzon, J.J. 2007. Does mobility decrease cooperation? J. Theor. Biol. 244, 722-728. Vehrencamp, S.L. 1983. A model for the evolution of despotic versus egalitarian societies. Anim. Behav. 21, 667-682. Verdolin, J.L. & Slobodchikoff, C.N. 2008. Resources, no kinship, determine social patterning in the territorial Gunnison’s prairie dog (Cynomys gunnisoni). Ethology 115, 59-69. West, S.A., Griffin, A.S., & Gardner, A. 2007. Evolutionary explanations for cooperation. Curr. Biol. 17, R661-R672. Yip, E. C., Powers, K. C, & Avilés, L. 2008. Cooperative capture of large prey solves the problem of a declining surface area to volume ratio of large social spider colonies. Proc. Natl. Acad. Sci. U.S.A. 105, 11818-11822. Zammit, J., Hogendoorn, K. & Schwarz, M. P. 2008. Strong constraints to independent nesting in a facultatively social bee: quantifying the effects of enemies-at-the-nest. Insect. Soc. 55, 74-78.  148  Appendix 1: Regressions of nest cross-section area to number of females per colony for Cuyabeno River (a), Cuyabeno Forest (b), Jatun Sacha (c), and Via a Loreto (d).  149  Appendix 2: Summary of methods, experimental treatments, and transplant duration for all nests (chapter 3). In the lower montane rainforest habitat, we transplanted 9 nests that had been recollected from Jatun Sacha. We transplanted these to see if they would be able to recover in a more favourable habitat. These nests performed as well as the 10 nests that were transplanted directly from the source habitat. Natural  Habitat  Elevation Species  # of  Occurrence  Source  in Habitat  Habitat  # of Nests  Number of Transplant  Transplanted Spiders/Nest  Sites  Additional  Transplant  Measurements  Duration  Notes  46 days  Control  Rainfall and Yanayacu: 2100 m  A. baeza  Common  Yanayacu  201  12 to 100  5  Predator  Cloud Forest Observations Transplant nests consisted of 9 groups 'recycled' from Jatun Hollin: Lower Montane Rainforest  Sacha, 10 groups transplanted  Rainfall and 1000 m  A. baeza  Rare  Yanayacu  192  12 to 65  4  Predator Observations  30-35 days  directly from native cloud forest; no performance difference  150  Natural  Habitat  Elevation Species  # of  Occurrence  Source  in Habitat  Habitat  # of Nests  Number of Transplant  Transplanted Spiders/Nest  Sites  Additional  Transplant  Measurements  Duration  Notes  All nests  Jatun Sacha: Lowland  400 m  A. baeza  Absent  Yanayacu  50  12 to 100  5  Rainforest  Rainfall and  completing  30 nests failed before or soon  Predator  transplant (n=15)  after net removal; 9 of the  Observations; extinct within 4  remaining nests were the  Rain Exclosure days; remaining source for groups 'recycled' to Experiment  nests removed  Hollin.  after 2 weeks. Test of rain exclosure effect on Jatun Sacha: Lowland  small groups of the native 400 m  A. eximius Common Jatun Sacha  223  Rain Exclosure ~100  3  25-30 days  species; 11 nests in each  Experiment Rainforest  treatment (sheltered versus exposed)  1  Six nests at two sites were removed by people during the experiment (2 very early, and 4 late). One nest was destroyed by passers by late in the experiment. 3 One exposed nest was destroyed by passers by late in the study period. 2  151  Appendix 3: Additional figures and tables documenting the results of the principal components analyses.  Appendix 3, Figure 1: Position of individual colonies of each species along principal components axes 1 and 2. The colours match those shown in figure 1 of the text. The influence of each variable along each principal components axis is shown with a labelled vector. We contend that variables that differentiate forest edge and forest interior habitat, forest height, % canopy cover, and distance from forest edge, as well as nest size and prey capture height variables, are the most influential for principal component axis 1. Principal component axis 2 appears to be most influenced the local spatial scale variables, nest height above ground and plant diameter. The third principal component axis is strongly influenced by instar and location on plant (not shown).  152  Appendix 3, Figure 2: Position of individual colonies of each species along principal components axes 1 and 2. We suggest that removing the nest size and prey capture height variables has a minimal impact on the quantitative results of the principal component analysis. The species are still significantly different from each other along the first three principal components axes (ANOVA, PC 1 F=25.2, p<0.0001; PC2 F=17.3, p<0.0001; PC3 F=4.69, p=0.0013). The influence of each variable along each principal components axis is shown with a labelled vector. As above, the three variables that differentiate forest edge and forest interior habitat, forest height, % canopy cover, and distance from forest edge are the most influential for principal component axis 1. Principal component axis 2 and principal component 3 (not shown) are influenced by local spatial variables and instar, respectively, as above.  153  Appendix 3, Table 1: Results of the principal components analysis that did not include nest size or prey capture height variables (Appendix 3 Fig. 2). When compared with Table 4.1 in the text, it is clear that excluding these two variables does not greatly affect the weighting of the spatial and temporal variables, the eigenvalues, or the percent of the variance explained by each principal component axis. Principal Component Axis Variable 1 2 3 4 5 Percent Canopy Cover 0.499 -0.0323 -0.316 0.128 -0.375 Distance From Forest Edge 0.515 -0.156 -0.125 -0.208 0.526 Forest Height 0.563 0.00985 -0.114 -0.113 0.118 Height Above Ground -0.0853 0.696 0.0116 -0.0258 0.562 Plant Diameter (Ord) 0.239 0.547 -0.034 0.591 -0.244 Position on Plant (Ord) 0.247 0.324 0.595 -0.571 -0.372 Instar (Ord) 0.207 -0.293 0.719 0.503 0.233 Eigenvalue 2.472 1.526 0.994 0.799 0.527 % Variance 35.313 21.797 14.194 11.416 7.531  154  Appendix 3, Table 2: Correlation coefficients for the variables used in the principal components analysis. Variable  Prey Capture Height  Nest Size (Cross Section Area)  Rho=0.728, p<0.0001 ***  Prey Capture Height  XXX  % Canopy Cover Distance from Forest Edge Forest Height Estimate Nest Height Above Ground Plant Diameter (ord) Location on Plant (ord)  Distance from Forest Edge Rho=0.362, Rho=0.524, p<0.0001 p<0.0001 *** *** Rho=0.220, Rho=0.454, p=0.0029 p<0.0001 ** *** Rho=0.599, XXX p<0.0001 *** % Canopy Cover  XXX  Nest Height Plant Above Diameter Ground (ord) Rho=0.489, Rho=-0.417, Rho=0.119, p<0.0001 p<0.0001 p=0.110 *** *** Rho=0.356, Rho=-0.307, Rho=0.069, p<0.0001 p<0.0001 p=0.358 *** *** Rho=0.658, Rho=0.282, Rho=-0.105, p<0.0001 p=0.0001 p=0.159 *** *** Rho=0.744, Rho=-0.191, Rho=0.113, p<0.0001 p=0.0096 p=0.130 *** * Rho=0.296, Rho=-0.122, XXX p<0.0001 p=0.101 *** Rho=0.365, XXX p<0.0001 ***  Forest Height Estimate  XXX  Location on Plant (ord) Rho=-0.033, p=0.655 Rho=-0.025, p=0.737  Instar (ord) Rho=0.229, p=0.0022 ** Rho=0.174, p=0.0207 *  Rho=0.052, p=0.489  Rho=-0.009, p=0.905  Rho=0.156, p=0.036 * Rho=0.271, p=0.0002 *** Rho=0.177, p=0.017 * Rho=0.157, p=0.035 *  Rho=0.182, p=0.015 *  XXX  Rho=0.106, p=0.158 Rho=-.278, p=0.0002 *** Rho=-0.013, p=0.869 Rho=0.033, p=0.662  155  Appendix 4: Full statistical results for nest size and spatial variables. Statistics are summarized in Table 2 and in Figures 2, 3, and 4 in the text. Appendix 4, Table 1: For each 2-species comparison, prey capture height is shown above the line and distance from the forest edge. For these comparisons, alpha=0.005 and one degree of freedom. A. nigrescens A. nigrescens  XXX 2  A. baeza A. studiosus A. jabaquara A. dubiosus  χ =0.078, p=0.78 χ2=0.026, p=0.87 χ2=27.9, p<0.0001 χ2=8.24, p=0.0041  A. baeza χ2=5.85, p=0.016 XXX 2  χ =0.0042, p=0.95 χ2=47.7, p<0.0001 χ2=14.8, p=0.0001  A. studiosus χ2=12.2, p=0.0005 χ2=2.87, p=0.09 XXX 2  χ =10.8, p=0.001 χ2=3.09, p=0.079  A. jabaquara A. dubiosus χ2=32.2, χ2=18.7, p<0.0001 p<0.0001 χ2=24.1, χ2=8.56, p<0.0001 p=0.0034 χ2=2.56, χ2=0.41, p=0.11 p=0.52 χ2=1.64, XXX p=0.20 2 χ =7.44, XXX p=0.0064  Appendix 4, Table 2: Canopy cover statistical comparisons are displayed above the line and forest height below the line. Alpha=0.005 and one degree of freedom for both comparisons.  A. nigrescens A. baeza A. studiosus A. jabaquara A. dubiosus  A. nigrescens  A. baeza  A. studiosus  A. jabaquara  XXX  χ2=0.17, p=0.68  χ2=0.061, p=0.81 χ2=0.11, p=0.73  χ2=19.6, p<0.0001 χ2=24.9, p<0.0001 χ2=4.79, p=0.029  χ2=0.33, p=0.57 χ2=0.044, p=0.83 χ2=22.3, p<0.0001 χ2=1.96, p=0.16  XXX χ2=0.0018, p=0.97 χ2=47.4, p<0.0001 χ2=9.17, p=0.0025  XXX χ2=4.15, p=0.042 χ2=1.05, p=0.31  XXX χ2=13.0, p=0.0003  A. dubiosus χ2=6.17, p=0.013 χ2=6.93, p=0.0085 χ2=2.1, p=0.15 χ2=2.97, p=0.085 XXX  156  Appendix 4, Table 3: Independent contrasts comparing the habitat variables measured for each species to the habitat variables measured for each random point. We use the Wilcoxon rank-sum test for all comparisons. Species  Canopy Cover  Distance from Forest Edge  Forest Height  A. nigrescens  less than random expectation: χ2=12.2, p=0.0005  less than random expectation: χ2=18.5, p<0.0001  less than random expectation: χ2=11.0, p=0.0009  A. baeza  less than random expectation: χ2=15.0, p=0.0001  less than random expectation: χ2=39.0, p<0.0001 less than random expectation: χ2=5.31, p=0.021  less than random expectation: χ2=33.4, p<0.0001  A. jabaquara  NS: χ2=1.80, p=0.18 greater than random expectation: χ2=9.35, p=0.0022  greater than random expectation: χ2=5.91, p=0.015  NS: χ2=2.37, p=0.12 greater than random expectation: χ2=17.7, p<0.0001  A. dubiosus  NS: χ2=0.27, p=0.60  NS: χ2=0.95, p=0.33  NS: χ2=1.65, p=0.20  A. studiosus  Appendix 4, Table 4: Local spatial comparisons within habitats (forest edge and forest interior), as well as comparisons between each species and the available nest locations calculated at each random vegetation sample along each transect. The latter comparisons are independent contrasts. Results for the comparison of nest position on the plant substrate are shown above the line (two degrees of freedom), and substrate size (DBH) is shown below the line (three degrees of freedom). For the comparison of the three edge species, alpha=0.0167. A. nigrescens A. nigrescens  XXX 2  A. baeza χ2=6.07, p=0.048  A. studiosus χ2=13.2, p=0.0003 χ2=1.35, p=0.51  χ =9.38, p=0.025 χ2=3.53, p=0.32  χ =3.06, p=0.38  XXX  A. jabaquara  NA  NA  NA  A. dubiosus  NA  NA  NA  A. baeza A. studiosus  2  Available  χ =16.1, p=0.0011  XXX 2  2  χ =44.0, p<0.0001  A. jabaquara A. dubiosus NA  NA  NA  NA  NA  NA 2  2  χ =12.6, p=0.0057  XXX χ2=62.5, p<0.0001 χ2=71.0, p<0.0001  χ =228.0, p<0.0001 XXX 2  χ =6.73, p=0.081  Available χ2=20.18, p<0.0001 χ2=8.48, p=0.041 χ2=0.031, p=0.98 χ2=25.6, p<0.0001 χ2=1.64, p=0.44 XXX  In addition, we found no significant difference in the height of nests above the ground between A. baeza and A. studiosus (χ2=1.54, DF=1, p=0.22).  157  Appendix 5: Detailed description of model and parameter sensitivity testing results. Accounting Individuals are represented in the simulation as a row in a series of four vectors, which contain information about their location, cooperation or tolerance phenotype, dispersal propensity phenotype, and status (alive or dead). The vectors are updated after each phase in the life cycle when the individuals that died are removed. Migration During each generation in the simulation, individuals go through a series of discrete stages or behaviours. First, individuals either disperse or remain in their natal patch. Each individual is subjected to a random chance of dispersing, with the probability increasing with the dispersal propensity phenotype. When an individual migrates, it is equally likely to occupy any patch in the habitat. There is no discrimination based on whether the patch is empty or occupied. Each individual that disperses is then subjected to the possibility of random mortality. The probability of mortality during dispersal depends on the dispersal_risk variable (Table 5.1). Competition for Habitat Patches (Tolerance Model Only) In the tolerance model, habitat patches can either contain solitary or tolerant individuals (modified from Hamilton & May 1977). If patches contain both solitary and tolerant individuals after the migration phase, one phenotype is randomly selected to inhabit the patch, and all individuals with the other phenotype are removed. In the model presented here, there is always an equal chance of each phenotype retaining each patch. Reduction to Local Carrying Capacity (Tolerance Model Only) For both solitary and tolerant phenotypes in the tolerance model, patches have a limited carrying capacity: they can contain either one solitary individual or a group of up to max_group tolerant individuals (Table 5.1). If a patch exceeds these dimensions, individuals are randomly removed until the patch population is at the local carrying capacity. Global Mortality Scenarios During each generation, the population experiences some form of global mortality: either the population is reduced to a fixed individual carrying capacity, or the population is reduced to the individuals living in a fixed number of groups (patch 158  carrying capacity). For the individual carrying capacity scenario, living individuals are counted. If the population exceeds the carrying capacity, individuals are randomly sorted and the appropriate number are removed from the population. For the patch carrying capacity scenario, if the population exceeds a fixed individual carrying capacity, all of the individuals living in a fixed number of random patches (rm_sites, Table 5.1) are removed from the population. Calculating Mean Individual Fitness (Cooperation and Altruism Models Only) In the cooperation and altruism models, the individual fitness (or fecundity) depends on the group size and mean cooperation level at each patch. In the cooperation model, each individual will have the same mean fitness, following the model developed by Avilés (1999): f(N) = Nγere-cN  (1)  This equation represents the average reproductive rate of individuals living in a group of size N with the cooperation phenotype averaged over all group members equal to γ. The actual individual fitness in this model is then sampled from a normal distribution with a mean equal to the function in equation 1.  This function is unimodal when γ>0, but is a  decreasing function of group size when γ=0. In the latter case, individuals living alone achieve the highest possible fecundity, equal to er, or the intrinsic rate of growth. The remaining term, e-cN, imposes negative density dependence effects on the group (Avilés 1999). This factor simulates local competition (Hamilton & May 1977). In the altruism model, we incorporate a relative fitness function that causes individuals with a high cooperation level (γ) to pay a fecundity cost when they are grouped with less cooperative individuals. We apply the equation developed by Avilés et al. (2002): Relative_fitnessi = α - βγi  (2)  where 159  α = 1-βγ  (3)  In this case, γi and γ represent the individual and group cooperative tendencies, respectively. Thus, an individual with average cooperation tendencies in a group (γi = γ) will have a relative fitness of one. The other parameter, β represents the slope of this function, and is a fixed parameter in our altruism model (Table 5.1). In our cooperation model, β=0, so there are no relative fitness costs applied to any group members. Reproduction Each individual in this stage of the model produces a randomized number of offspring that is centred around a mean value determined by the tolerance phenotype (tolerance model) or by the average reproductive rate (cooperation model) and relative fitness (altruism model). Reproduction is asexual. The offspring then replace their parents in the natal patch. Offspring have the same phenotypes as their parent unless mutation occurs. Mutation During the production of each offspring, mutations can occur in both the tolerance/cooperation phenotype and the dispersal propensity phenotype with a fixed mutation probability of 10-2 for each trait (Table 5.1). The mechanism is described in the methods section. Parameter sensitivity testing In order to establish the baseline parameters for the simulations described in this paper, we tested the influences of each of the non-focal variables. We then selected nonfocal parameter values that both made sense from a biological perspective and allowed the model to run efficiently. The general parameters (Table 5.1) were included in each of the three models. We ran each simulation for 10,000 generations to ensure an accurate measure of the stable equilibria that were reached. In at least one case (Appendix 6c), the traits evolving in the simulation took almost 10,000 generations to asymptote. In several other cases (not shown) chaotic dynamics made short-term simulations unreliable in revealing long-term equilibrium values. The equilibrium values presented in Fig. 5.1 in the text represent the average trait values over the last 5000 generations of each simulation. Simulations were then replicated at least 100 times to control for random 160  variation in each run, and to detect any alternate stable states. Each habitat contained 200 sites. Using very few sites influenced trait evolution, but this effect subsided when the habitat contained more than about 50 patches. We initiated each model with one individual in each patch. All individuals were solitary (tolerance model) or had a cooperation level of zero (cooperation and altruism models), and each had a moderately high dispersal propensity. We initiated the parameters thus because we are interested in the transition from solitary living, where individuals are expected to maintain moderate dispersal propensities under simple conditions (Hamilton & May 1977), to group living based on each of the three social traits investigated here. The mutation rate and the mutation standard deviation were also fixed for all simulations. Not surprisingly, increasing or decreasing the mutation rate increased or decreased the rate of evolution of both dispersal propensity and cooperation or tolerance phenotypes. Mutation rate had both the direct effect of slowing the occurrence of mutations, and the indirect effect of making mutants with a social trait less likely to find similar individuals in the population. We selected a mutation rate that has been previously used in similar models (e.g. Avilés et al. 2002; Table 5.1). Increasing the mutation standard deviation allowed for more mutations of large effects and both increased the rate of evolution and created more noise at the equilibrium level. When the mutation standard deviation was decreased, mutation did not provide enough phenotypic variance to the population to allow natural selection to act, so populations did not evolve (or at least did not evolve within our time frame). The mutation standard deviation that we used in these simulation allowed evolution to occur over a reasonable time frame, and did not lead to large fluctuations in the equilibrium values of each phenotype (Table 5.1). In each of our population control scenarios, we set the threshold population size, below which mortality would not occur and above which mortality would occur, to the number of sites (200). In the individual carrying capacity scenario, we found that the results were not sensitive to higher carrying capacity values. However, at moderately small values, the rate of evolution was decreased. When the individual carrying capacity was very small relative to the number of patches available, such that individuals were unlikely to land in patches with other individuals, the phenotypes remained stable near 161  the initial values (cooperation = 0; dispersal_propensity = 0.7). In the patch carrying capacity model, half of the patches went extinct each generation that the individual carrying capacity was exceeded. Like the individual carrying capacity scenario, the more patches that went extinct, the slower cooperation evolved to its equilibrium level. Unlike the individual carrying capacity scenario, cooperation evolved even when very few patches contained inhabitants each generation. In the tolerance model, the relative fecundity of the tolerant individuals compared to the solitary ones was the most important variable. If tolerant individuals produced fewer than half the number of offspring produced by solitary individuals, tolerance could not invade the population. Increasing the maximum group size only allowed tolerant individuals to invade under a slightly more detrimental relative fecundity rate. The actual fecundity values also had a relatively small effect on the ultimate equilibrium tolerance values. The behaviour of the novel parameters in the cooperation and altruism models has already been explored extensively (Avilés 1999; Avilés et al. 2002), and our results were similar. We found that the population went extinct before cooperation could spread when r < 0. Like Avilés et al. (2002), we found that changing the intrinsic rate of growth (r) had relatively little effect on the rate of evolution or on the evolutionary equilibrium of cooperation. Increasing ‘r’ just meant that more individuals were produced each generation. We used r = 1 to keep the model running at a reasonably speed. We found that adjusting the crowding parameter (c) influenced the rate of evolution of both cooperation (γ) and dispersal propensity. When local competition was imposed at smaller colony sizes, the traits reached their equilibrium values more slowly. Decreasing the crowding level led to increasingly unstable dynamics, particularly in the population mean group size and fecundity (Avilés 1999). In their model, Avilés et al. (2002) found that the  β value explained most of the variance observed in their comparisons of cooperation evolution. In our altruism model, we employed the highest β value investigated by Avilés et al. (2002), to maximize the differences in behaviours between the two models. We also explored the evolutionary dynamics of our focal traits, dispersal propensity and tolerance, cooperation, or altruism, at fixed values of each (Appendix 5 Fig. 1 below). 162  163  Appendix 5, Figure 1: Simulations of one trait evolving with fixed values of the other for each model with no dispersal cost. Colours represent fixed dispersal or social trait values of zero (magenta), 0.25 (blue), 0.5 (green), 0.75 (yellow), and 1 (red). When tolerance is fixed and dispersal evolves (D, J), individuals are either tolerant (red) or solitary (blue).  164  Appendix 6  165  Appendix 6, Figure 1: Full time series (10000 generations) averaged over at least 100 replications for each model type (top row: tolerance; middle row: cooperation; bottom row: altruism), under each ecological scenario (individual carrying capacity, left; group carrying capacity, right), and for either no dispersal costs (left) or moderate dispersal costs (right). In all of the simulations except altruism under the individual carrying capacity scenario (C), the phenotypes become relatively stable within the first 1000 generations or so.  166  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0067207/manifest

Comment

Related Items