UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fitness consequences of group living and the loss of dispersal in spiders Salazar, Patricio Alejandro 2006

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2006-0641.pdf [ 2.21MB ]
Metadata
JSON: 831-1.0092746.json
JSON-LD: 831-1.0092746-ld.json
RDF/XML (Pretty): 831-1.0092746-rdf.xml
RDF/JSON: 831-1.0092746-rdf.json
Turtle: 831-1.0092746-turtle.txt
N-Triples: 831-1.0092746-rdf-ntriples.txt
Original Record: 831-1.0092746-source.json
Full Text
831-1.0092746-fulltext.txt
Citation
831-1.0092746.ris

Full Text

FITNESS CONSEQUENCES OF GROUP LIVING A N D THE LOSS OF DISPERSAL IN SPIDERS by PATRICIO A L E J A N D R O S A L A Z A R L i e , Pontificia Universidad Catolica del Ecuador, 2001 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE -in THE F A C U L T Y OF G R A D U A T E STUDIES (Zoology) THE UNIVERSITY OF BRITISH C O L U M B I A August 2006 © Patricio Alejandro Salazar, 2006 A B S T R A C T The evolutionary transition from solitary to group living in spiders implies the progressive loss of natal dispersal behaviour. This process is likely mediated by the fitness consequences of two interrelated factors: the "choice" that individuals make to stay or disperse from their natal group and, as a result of this, the group size they experience at different life stages. In this study I investigated the consequences of these factors in three components of female lifetime reproductive success in the social spider Anelosimus guacamayos (Araneae: Theridiidae). B y regularly recording changes in spider counts and nest proliferation events in a total sample of 105 naturally occurring colonies, my collaborators and I found: (1) an overall higher survival probability for philopatric females relative to emigrants, (2) some survival improvement related to group size for philopatric females living in colonies smaller than -20 individuals, and (3) an increase in the offspring survival probability, mainly associated with group size rather than being merely related with philopatry or dispersal. We found no effect of either dispersal or group size on the probability of female reproduction. Despite the fitness costs of dispersal, its occurrence in group living spiders, such as A. guacamayos, is best explained by the negative density-dependent effects of group living. On the other hand, the rapid improvement in offspring survival is a straight forward explanation for the evolution of group living because it does not require active cooperation between the spiders staying in their natal nest. Because mother spiders sometimes die before their offspring are self-sufficient, the first benefit of group living is most likely a geometric decrease for each adult that is added to the group in the chances of ending up in a nest with dependent offspring but no adults. i i T A B L E O F C O N T E N T S ABSTRACT : i i T A B L E OF CONTENTS iii LIST OF T A B L E S iv LIST OF FIGURES v A C K N O W L E D G M E N T S vi CO-AUTHORSHIP STATEMENT vii CHAPTER 1: 1 G E N E R A L INTRODUCTION: T H E EVOLUTION OF SOCIAL BEHAVIOUR IN SPIDERS 1 BIBLIOGRAPHY '.' 4 CHAPTER 2: 5 FITNESS CONSEQUENCES OF GROUP LIVING AND DISPERSAL IN T H E SOCIAL SPIDER ANELOSIMUS GUACAMAYOS : 5 INTRODUCTION 5 METHODS 8 The study species: Anelosimus guacamayos 8 Fitness decomposition 9 Field data collection 9 Estimation and analyses of demographic parameters 11 Other factors affecting fitness 15 RESULTS 15 Colony dispersal as a function of group size 15 Consequences of dispersal on the survival probability of mature females to reproduction ..16 The probability of female reproduction 19 The probability of offspring survival to maturity : 20 DISCUSSION .. 21 BIBLIOGRAPHY 25 CHAPTER 3: : 28 G E N E R A L DISCUSSION 28 BIBLIOGRAPHY 32 iii LIST OF TABLES CHAPTER 1 (No tables) CHAPTER 2 T A B L E 2.1 : Equations used to calibrate field censuses error 11 T A B L E 2.2 : Females surviving in and disappearing from their natal nest during the transition from maturity to reproduction 19 T A B L E 2.3 : Females surviving from maturity to reproduction inside the four 150x5m sampled plots 19 CHAPTER 3 (No tables) i v LIST OF FIGURES CHAPTER 1 (No figures) CHAPTER 2 FIGURE 2.1: Group size distribution at the start of the egg sac laying period 12 FIGURE 2.2 : Probability of colony dispersal as a function of group size 16 FIGURE 2.3 : Survival probability of philopatric (psp) and emigrant (pse) females as a function of the fraction of disappearances assigned to emigration (e) 17 FIGURE 2.4 : Proportion of maturing females surviving to reproduction in their natal nests 18 FIGURE 2.5 : A. Proportion of females reproducing in their natal versus newly founded nests B. Proportion of females reproducing in relation to their group size 20 FIGURE 2.6: A. Proportion of offspring surviving to maturity in relation to their nest of birth B. Proportion of offspring surviving to maturity in relation to group size 21 CHAPTER 3 (No figures) v ACKNOWLEDGMENTS Thanks to all the people that made it possible my graduate studies at U B C and that, directly or indirectly, contributed to the completion of this thesis work. To my supervisor and co-author of the second chapter, Leticia Aviles, for her work throughout the project. To my second co-author, Gabriel Iturralde, for his contributions during the fieldwork, data processing and analyses. Neuza Gallardo made financially possible my trip to Canada. My mother, Patricia Salazar, has constantly contributed emotionally and logistically. Esperanza Garcia has supported and encouraged me in all aspects of my studies. Yadira Mera, Raul Barrera, and especially Patricia Salazar, Cecilia Puertas, Pepa Duran-Ballen and Jorge Luna helped and provided invaluable support during the field work phase. Patricio Taco, from the Ecuadorian Ministry of Environment, allowed the use of the Antisana Ecological Reserve facilities, at Sector Cocodrilo, at no cost. The park guards: Antonio Quilumba, Franklin Tanguila, Javier Haro, Luis Yupa and Saul Alvarado, were always helpful and made field work very enjoyable. My gratitude to all the staff of the Zoology Department, particularly the graduate secretaries Allison Barnes and Marie Wallden, for their support and friendship. The following people provided suggestions and insightful discussion at different stages of the data analyses and manuscript writing: my committee members Michael Whitlock and Judith Myers, my lab mates Ingi Agnarsson, Jessica Purcell, and Jennifer Guevara, the Zoology Department professors Sarah Otto and Diane Srivastava, the attendants to SOWD at UBC and Les Ecologistes seminar series at Simon Fraser.University. Esperanza Garcia and Jeff Fletcher deserve special thanks for long hours of fruitful discussion. Important financial support came from grants NSF DEB-9815938 and NSERC RGPIN 261354-03 to L. Aviles, and a Graduate Entrance Scholarship from the Zoology Department and a U B C University Graduate Fellowship to P. A . Salazar. v i CO-AUTHORSHIP STATEMENT Patricio A. Salazar performed a leading role in the following phases of this research: design and establishment of the field study, data collection and analyses, and manuscript preparation. Leticia Aviles shared the leading role in the design of the study and contributed to the data analyses and manuscript preparation of chapter two. Gabriel Iturralde, from the Pontificia Universidad Catolica del Ecuador (Quito, Ecuador), shared a leading role during the field data collection and contributed to the data processing and analyses. vii CHAPTER 1: General introduction: the evolution of social behaviour in spiders Social behaviour is not common in spiders. From approximately 39,500 described species (e.g. Platnick 2006), fewer than a hundred are known to exhibit some form of social behaviour (for reviews see: Aviles 1997; Whitehouse and Lubin 2005). Hence, it is intriguing why some spider species are social. Furthermore, the questions of why and how social behaviour evolves are central in evolutionary biology. Three chapters exploring some aspects of the evolutionary transition from solitary to group living in spiders compose this thesis work. This first chapter is a general introduction to the problem of the evolution of sociality in spiders. The second chapter is the complete report of a field study, co-authored with Leticia Aviles and Gabriel Iturralde, on the fitness consequences of group living and dispersal in the social spider Anelosimus guacamayos Agnarsson 2006. A n d finally, the third chapter is a broader discussion of the evolutionary implications of the results reported in chapter two, and some ideas about the causes, mechanisms and consequences of the evolution of social complexity in spiders and other organisms. The non-territorial social spiders can be classified into two fairly discrete categories: periodic- . social and permanent-social species (Aviles 1997). The first category concerns species in which members of a social group disperse invariably every generation, usually just before reaching reproductive maturity. The colonies of this type of spider typically consist of an adult female with her offspring, or the brood without the presence of the mother. In permanent-social species, members of a colony remain together their entire life cycle, so that successive generations continue to occupy the same communal web, until the colony eventually either proliferates or becomes extinct. Colony proliferation normally occurs at relatively large colony sizes when some or all individuals disperse to start new colonies, and probably reflects the point at which the costs of overcrowding exceed the risks of emigration. The analyses of the evolution of sociality in spiders involve three aspects: (1) the possible ancestral stages that allowed and/or favoured group living, (2) the geographic distribution and ecological correlates of the social species, and (3) the possible adaptive value, i.e. the balance 1 between costs and benefits of evolving sociality. In terms of the first aspect, certain authors (e.g. Burskirk 1981; Aviles 1997) have proposed that the existence of a three-dimensional irregular web constitutes a "preadaptation" for the evolution of non-territorial aggregations in spiders . This idea is based on two arguments: first, by providing a physical connection between individuals, a web facilitates group cohesion and, second, because of the energetic cost of building and maintenance, the web provides an incentive for cooperation. With only one exception, all non-territorial permanent-social species belong to lineages that build three-dimensional irregular webs. Still, a web and its architecture do not seem to be the only requirements for evolving permanent-sociality. By looking at the phylogenetic distribution of the non-territorial permanent-social spiders, it becomes evident that they only occur in clades with species showing extended maternal care. This suggests a transition from periodic to permanent sociality, because if the period of tolerance and cooperation between siblings of a given periodic-social species extends through adulthood the permanent-social condition would arise (Aviles 1997). How this transition occurs, and what its benefits and costs are, constitute aspects of current active research (e.g. Bilde et al. 2005; Aviles and Bukowski 2006; L i and Kuan 2006). The geographic distribution of the social spiders gives us insight into the ecological conditions that may play a role in the transition from periodic to permanent sociality. Most of the social species, as well as the ones with more complex sociality, inhabit tropical lowland regions (Riechert 1985; Aviles 1997). Moreover, periodic-social species tend to live in either temperate zones or relatively high/mid elevation sites in the tropics; in contrast, permanent-social species are normally restricted to the lowlands. Some ecological reasons to explain the geographic distribution of periodic vs. permanent sociality are: (1) the more or less constant year-round food supply in the lowland-tropics that would promote a delay of the brood's dispersal phase, so that the social groups persist for several generations, (2) the high rain frequency, which tends to damage and destroy the webs, would promote communal building and maintenance, and (3) an abundance of relatively large-body prey that would not be accessible to solitary species. A different possibility, that does not exclude the previous arguments is simply that, given the overall higher species diversification of spiders in the tropical lowlands, the opportunities for evolving sociality are also greater (Aviles 1997). 2 The adaptive value of group living has been studied by either examining the dispersal phase in periodic-social species (e.g. Aviles and Gelsey 1998; Jones and Parker 2002) or the relationship between individual fitness and group size in permanent-social species (e.g. Uetz and Hieber 1997; Aviles and Tufifio 1998). Studies in periodic-social species have found that delaying natal dispersal is beneficial whereas the actual dispersal process is costly. Studies in permanent-social species have shown that lifetime reproductive success (or at least some of its components) is maximum at intermediate colony sizes. This is the first study that combines those two approaches, by taking advantage of a species that represents an intermediate stage in the transition from periodic to permanent sociality. As I describe in more detail in chapter two, the ecological and evolutionary context of A. guacamayos makes it interesting because, even though this species can already be defined as permanently social, its natural history and behaviour shares many characteristics with periodic-social species. Principally, colony dispersal occurs relatively frequently and its geographic distribution (mid elevation sites in tropical Ecuador) is intermediate in relation to the habitats of more strict periodic and permanent social species in the same genus (P. Salazar, L. Aviles and G. Iturralde, unpubl. results) The main goal of this thesis work is to contribute to the understanding of the processes occurring during the transition from solitary to group living in spiders, and by doing so to explore some general principles involved in the evolution of sociality. In particular I would like to make a case for the interrelated nature of the evolution of group living and the evolution of dispersal. In many contexts they are two sides of the same coin and, as I will try to show in the following chapters, considering them as interactive processes is productive.. 3 BIBLIOGRAPHY Aviles, L. 1997. Causes and consequences of cooperation and permanent sociality in spiders, Pages 476-498 in J. C. Choe, and B. J. Crespi, eds. The evolution of social behaviour in insects and arachnids. 476-498. Cambridge University Press. Cambridge. Aviles, L., and T. C. Bukowski. 2006. Group living and inbreeding depression in a subsocial spider. Proceedings of the Royal Society of London - Series B: Biological Sciences 253:157-163. Aviles, L., and G, Gelsey. 1998. Natal dispersal and demography of a subsocial Anelosimus species and its implications for the evolution of sociality in spiders. Canadian Journal of Zoology 76:2137-2147. Aviles, L., and P. Tufino. 1998. Colony size and individual fitness in the social spider Anelosimus eximius. American Naturalist 152:403-418. Bilde, T., Y . Lubin, D. Smith, J. M . Schneider, and A. A. Maklakov. 2005. The transition to social inbred mating systems in spiders: role of inbreeding tolerance in a subsocial predecessor. Evolution 59:160-174. Burskirk, R. 1981. Sociality in the Arachnida, Pages 282-397 in H . R.. Hermann, ed. Social Insects. 282-397. Academic Press. New York. Jones, T. C , and P. G. Parker. 2002. Delayed juvenile dispersal benefits both mother and offspring in the cooperative spider Anelosimus studiosus (Araneae: Theridiidae). Behavioral Ecology 13:142-148. L i , D., and J. Y . X . Kuan. 2006. Natal dispersal and breeding dispersal of a subsocial spitting spider (Scytodes pallida) (Araneae: Scytodidae), from Singapore. Journal of Zoology 268:121-126. Platnick, N . I. 2006. The world spider catalogue, version 7.0, American Museum of Natural History.http://research.amnh.org/entomology/spiders/catalog/index.html. (Accessed: 1-Aug-04). Riechert, S. E. 1985. Why do some spiders cooperate? Agelena consociata, a case study. Florida Entomologist 68:105-116. Uetz, G. W., and C. S. Hieber. 1997. Colonial web-building spiders: balancing the cost and benefits of group-living, Pages 458-474 in J. C. Choe, and B. J. Crespi, eds. The evolution of social behaviour in insects and arachnids. 458-474. Cambridge University Press. Cambridge. Whitehouse, M . E. A. , and Y. Lubin. 2005. The functions of societies and the evolution of group living: spider societies as a test case. Biological Review 80:347-361. C H A P T E R 2: Fitness consequences of group living and dispersal in the social spider Anelosimus guacamayos1 1 2 1 Patricio A . Salazar , Gabriel Iturralde and Leticia Aviles 1 Department of Zoology, University of British Columbia, Vancouver, Canada 2 Escuela de Biologia, Pontificia Universidad Catolica del Ecuador, Quito, Ecuador INTRODUCTION Why and how social behaviour evolves are major questions in evolutionary biology. Behavioural and ecological comparisons between phylogenetically related social and non-social organisms (e.g. Choe and Crespi 1997; Hunt 1999; Thorne and Tranielo 2003) suggest that the evolutionary transition from solitary to group living often implies two interrelated behavioural changes: (1) a reduction of the dispersal tendency of individuals from their natal place and (2) an increase of their tolerance to conspecifics. These evolutionary changes are logical requirements for the evolution of higher levels of cooperation, such as division of labour and reproduction, and it is reasonable to expect that they w i l l be associated with changes in the patterns of survival and reproduction of individuals. Although it is unlikely that we wi l l observe and record these evolutionary transitions in real time (at least for animal taxa) it is certainly possible to infer what factors are involved, how they operate, and what their consequences are by looking at organisms that represent intermediate stages in the transition (Thorne and Tranielo 2003). In this paper we analyse the fitness consequences of group living and dispersal in one of such organism, the social spider Anelosimus guacamayos Agnarsson 2006 (Araneae: Theridiidae). Non-territorial permanently social behaviour has evolved independently at least 21 times in spiders (Agnarsson et al. Submitted). Lineages where sociality has evolved generally include species with two social systems: a periodic-social (also called "subsocial") and permanent-' A version of this chapter is going to be submitted for publication: Salazar, P. A. , G. Iturralde and L. Aviles. Fitness consequences of group living and dispersal in the social spider Anelosimus guacamayos. 5 social (often called just "social"). Social groups (colonies) of periodic-social species typically consist of one adult female with her offspring, and eventually the brood without the presence of the mother. In these species all colony members disperse solitarily before reaching reproductive maturity. On the other hand, colonies of permanent-social species may contain many adult individuals - up to tens of thousands in some species - that tend to stay together the entire life cycle so that successive generations continue to occupy the same communal web. In this case, dispersal of some or all individuals may take place every few generations at relatively large colony sizes (for reviews see: Aviles 1997; Whitehouse and Lubin 2005). Permanent-social spiders are thought to evolve from periodic-social ancestors. This idea is mainly supported by the observation that permanent-social species occur only in lineages characterized by extended maternal care, and that permanent sociality is generally derived from periodic sociality in phylogenetic reconstructions (Aviles 1997; Agnarsson 2004; 2006; Agnarsson et al. Submitted). This evolutionary route is the most parsimonious because it only requires that the period of tolerance and cooperation between siblings of a given periodic-social species extends through adulthood, losing their characteristic fixed dispersal phase. Several adaptive benefits can result from the evolution of permanent-sociality in spiders, including a reduction in the per-capita silk production for nest building and maintenance (Riechert 1985); access to relatively large prey items, that solitary spiders of equivalent conditions are unable to capture (Nentwig 1985; Ward 1986; Powers 2004); reduction in the mortality due to fixed dispersal every generation (Aviles and Gelsey 1998); and increase in offspring survival as group size increases (Aviles and Tufino 1998). On the other hand, the evolution of permanent-sociality also implies costs. For instance, a shift from an outbred to a highly inbred mating system (Bilde et al. 2005; Aviles and Bukowski 2006), and reduced per-capita prey availability as colony size increases (Ward 1986, E.Yip, K. Powers and L. Aviles, unpubl. results). Hence it is reasonable to expect that the transition from periodic to permanent sociality involves a compromise between the fitness benefits and costs of staying (philopatry) versus dispersing from the natal group. A. guacamayos is interesting in this context because, even though it can be already defined as permanently social, its natural history and behaviour share many characteristics with periodic social species. For instance, colony dispersal occurs relatively frequently, compared to what has been observed in other permanent-social spiders, and can even occur (although rarely) in 6 the first generation after colony foundation (P. Salazar, L. Aviles and G. Iturralde, unpubl. results). As a consequence, colony size appears not to exceed a few hundred individuals. Also, it has been observed that adult females caring for egg sacs or young spiderlings have greater degree of physical separation within colonies compared to more heavily populated nests of other permanent-social species in the same genus (E.Yip, K. Powers and L. Aviles unpubl. results). This is reminiscent of the territorial behaviour sometimes observed in adult periodic-social Anelosimus, when occasionally two or more nests of solitary females, originally located in close proximity, fuse to become a single shared nest (pers. obs.). Finally, A. guacamayos inhabits sites where environmental conditions are intermediate in relation to the geographic distribution of sociality in spiders. Permanent-social species, particularly those with more complex sociality, tend to occur in tropical lowland regions - characterized by heavy rainfall and availability of relatively large potential prey (Riechert 1985; Aviles 1997), whereas periodic-social species are mainly distributed in either temperate zones or high/mid elevations in the tropics. A. guacamayos has only been recorded at mid-elevations (1200-1900m) in the eastern Ecuadorian Andes (Agnarsson 2006; P. Salazar, L. Aviles and G. Iturralde, unpubl. results), and interestingly is often sympatric (therefore share the same ecological conditions) with A elegans, its closest known periodic-social relative, from which appears to have diverged relatively recently (I. Agnarsson pers. comm.). The fact that two closely related species with those two alternative social systems can coexist in the same environment implies not only that such environment can sustain the existence of group living, but that some of these ecological conditions are likely to influence the transition from periodic to permanent sociality. In the present article we first evaluated how the occurrence of colony dispersal in A. guacamayos increases with group size, and what are the consequences of dispersal for the group living environment (specifically group size) experienced by individuals during reproduction and brood care. We then studied the short-term fitness consequences of these two interrelated attributes of group living: the "choice" to stay or disperse from the natal group, and group size. We analysed three components of females' lifetime reproductive success (LRS): (1) the survival probability from recently acquired maturity to the reproductive period, (2) the probability of reproducing, and (3) the probability of offspring survival to maturity. We end with a discussion of the possible causes and evolutionary implications of the demographic patterns found in this study. 7 METHODS The study species: Anelosimus guacamayos Anelosimus guacamayos inhabits open-disturbed areas, such as landslides or roadsides. Before reproducing (laying of egg sac), adult females can be found living either solitarily or in groups (colonies) of up to -250 adult individuals (>1000 including juveniles). Solitary females and relatively small groups are normally the product of recent dispersal, whereas medium and relatively large groups have usually been established either in previous generations or by fission of established nests (P. Salazar, L. Aviles and G. Iturralde, unpubl. results). Individuals living in groups engage in cooperative nest building and maintenance, prey capture and communal feeding. Communal brood care does not seem to be as developed as has been described for other permanent-social species (e.g. A. eximius, Chistenson 1984; Stegodyphus mimosarum, Ward and Enders 1985), but newly born juveniles certainly benefit from being free of web building and maintenance, and prey capture (P. Salazar, L. Aviles and G. Iturralde, unpubl. results). Individual females lay their clutch in a silk egg sac. Eggs take approximately 48 days to hatch (estimated range: 32-64days). Average clutch size is 34 ± 6 eggs per sac (mean ± SD). There is no evidence that females lay more than one egg sac during their lifetime. Newly born spiders grow in their natal nests; females undergo seven moults to reach sexual maturity (adults) whereas males only six. Adult females mate in their natal nest and often stay there to lay their egg sac. However, depending on colony size (see Results below), females can disperse and establish a new nest, either solitarily or by joining a few other dispersing females that usually come from the same source colony. Colony dispersal, when it occurs, is a massive event that involves a large fraction (60-100%; see Results) of the colony members. Males die before females start to lay egg sacs, and generally before dispersal, in colonies where it happens. Reproduction and individual growth is highly synchronized within colonies, so that adults of two successive generations very rarely overlap in the same nest (P. Salazar, L. Aviles and G. Iturralde, unpubl. results). This fact greatly facilitates the estimation of the demographic parameters analysed here, because spiders of the same cohort (colony-generation) can be followed through their lifetime without marking them individually, and the number of individuals of approximately the same age (one or two 8 consecutive instars), inhabiting a nest at a given point in time, corresponds to the total number of individuals of the monitored cohort that have survived until that age. Fitness decomposition Building on the approach by Aviles and Tufmo (1998), we estimated individual fitness as female lifetime reproductive success, i.e. the expected number of sexually mature offspring females produced per mature female in the maternal generation, and decomposed it as follows: LRS=psxprxfxpsFI (1) where ps is the probability of female survival from recently acquired maturity to reproduction,^ is the probability of female reproduction,/is female fecundity and pSFi is the probability of offspring survival to maturity. We were able to analyse the consequences of philopatry, dispersal and group size for three of these parameters: ps,pr and pSFi. We did not analyse the fecundity component because we preferred not to collect egg sacs from the colony sample we used to track individual survival and reproduction. Field data collection We estimated ps, pr and pspi for philopatric and emigrant females from census data collected between September 2002 and November 2003, in a total sample of 105 naturally occurring nests, corresponding to the natural range and distribution of colony sizes (when necessary, we use subscripts p and e to differentiate parameters for philopatric and emigrant females respectively). Monitored colonies were located on the sides of the road Baeza-Tena,. at the Cordillera de los Guacamayos (Sector Cocodrilo: 0°38'44,9" S; 77°47'26,8" W; 1700-1900m elevation), Napo province, Ecuador. Censuses were carried out once every two weeks, and consisted in counting the total number of individuals in a colony, classifying them by sex and instar (juveniles, subadults and adults), and recording the number of egg sacs. Our censuses were greatly facilitated by the natural construction of A. guacamayos nests, that consist of a very translucent silk web that tends to accumulate fewer dry leaves compared to other permanent-social species (e.g. A. eximius, A. domingo; P. Salazar and L. Aviles pers. obs.). In addition, differentiating between sexes and instars, based on morphological features such as body size, coloration and development of reproductive organs (male palpi in particular), 9 requires only brief training and is highly consistent between observers. Censuses were carried out at night using a flashlight, in order to avoid over-lighting caused by sunlight and because spiders are less likely to hide at night in the most enclosed areas of the nest than during the day. In order to detect dispersal events, every time a census was performed, we searched during daytime for newly established nests 5-5.5 m around the monitored colonies. We chose this surveying ratio based on previous studies of other Anelosimus species (Aviles and Gelsey 1998; Powers and Aviles 2003), and our own preliminary observations of the clumped spatial distribution of A. guacamayos nests, which suggested that most dispersal occurs in the first few meters adjacent to the colony source. We avoided touching the surrounding vegetation to minimize any possible impact on the natural movement and settling patterns of dispersing individuals. We recorded and marked all new nests, and included them in the set of monitored colonies. We considered a colony dispersal event only when a decrease in the number of spiders in the monitored colonies coincided simultaneously with the appearance of newly founded nests on the surroundings, and the age of individuals in these newly founded nests was the same as individuals in the putative source colony. Because some individuals may disperse beyond the monitored area, to investigate the survival probability of emigrant females we recorded colony dispersal and proliferation in larger areas (four 150x5m plots). This way more emigrants from monitored colonies would be recorded, and the movement of individuals in and out the surveyed area is expected to balance to some extend. A l l Anelosimus nests inside our sampled plots were marked, mapped and monitored between January and November 2003 (10~11 months, which is approximately equal to one A. guacamayos generation). We used this data subset, relative to a standard area and time frame, to evaluate possible values and to obtain confident estimates of pse. A l l remaining parameters were estimated from data obtained by targeting individual colonies and their surroundings. Despite the ease that A. guacamayos offers for collecting demographic data, from previous experiences (Aviles and Tufino 1998; Salazar 2001), we did not expect our censuses to be completely accurate. Thus, near the end of the study, we collected 10 colonies covering the range of sizes studied and examined their contents in the laboratory. As in previous studies (Aviles and Tufino 1998; Salazar 2001), we found a general tendency to underestimate the number of individuals as colony size increased. Such a tendency, however, was highly 10 predictable (see table 2.1 for R2). We calibrated our field census data using the regression equations that predict the number of individuals in a colony as a function of the field count. We used separate calibration equations for individuals of different instars because field census error varied with age/size. Nevertheless, it is important to mention that we found no census error in adult female counts of colonies up to 20-25 individuals. Thus, to keep calibrated data as accurate as possible, estimates obtained from the regression equations were rounded down to the nearest integer. This rule restored small estimates to the original counts without having much of an impact on calibrated estimates of larger colonies. Table 2.1. Regression equations used to predict the number of individuals in a colony as a function of the field census (x). N o census error calibration was performed for egg sacs because field counts were highly accurate (linear regression slope « 1). Instar Equation R2 Adul t females v = 1.128* 0.984 Subadult+adult females (last two instars) y= 1.2158* 0.993 Egg sacs N o calibration Estimation and analyses of demographic parameters Dispersal and group size-— Even though colony dispersal in permanent-social spiders is known to occur at relatively large colony sizes (Aviles 1997; Bodasin et al. 2001), to our knowledge, this density-dependent pattern has neither been quantitatively described nor tested in the field, for any species yet. We quantified this increase in the probability of colony dispersal in relation to group size by performing logistic regression analyses on the colony dispersal events as a function of the number of maturing females inhabiting a nest before dispersal. We estimated the number of females reaching maturity in a colony-cohort as the largest number of subadult + adult females (last two instars) recorded, once spiders of the two instars were present and, in cases where colony dispersal occurred, before detecting the first newly founded nest in the surroundings. Because emigrant females are more likely to end up living either solitarily or in relatively small groups (figure 2.1), the process of colony dispersal may not only affect fitness directly via mortality, but indirectly by changing the group living environment that individual females experience for reproduction and brood care. As a result, fitness differences between philopatric and emigrant females may not only be influenced by the dispersal process, but also by factors associated with group size. For that reason, besides direct comparisons of 11 demographic parameters between philopatric and emigrant females, we also explored the possible effects of group size. A l l estimates are reported ± their 95% confidence limits. When comparing two proportions . against each other we applied z tests accounting for differences in sample size. When analysing the bivariate relationship between a certain parameter and group size we fitted linear models to the arcsine transformed proportions (weighting each observation by the denominator to ensure that residuals are uniform) as a function of the In group size. When possible we treated dispersal occurrence and group size as covariates in an analyses of covariance framework. We used a 0.05 significance level. Emigrant Philopatric Figure 2.1. Group size at the start of the egg sac laying period for emigrant and philopatric females. The density of data points and box plots reflect the natural distribution of colony sizes. After dispersing and establishing a new nest, emigrant females are more likely to end up taking care of their offspring either solitarily or in relatively small groups, while most philopatric females will tend to experience larger groups. Survival probability ofphilopatric vs. emigrant females— Estimating and comparing survival of philopatric and dispersing individuals is one of the hardest aspects in demographic studies. First, it is hard to differentiate between mortality and emigration when only census data are available; in particular, most individual deaths and dispersal movements are rarely witnessed, but inferred from individual disappearances and appearances between censuses at different localities (Waser et al. 1994). Second, the ability to detect successful dispersal is reduced (as in our case) or null beyond the study area. In order to deal with these limitations, we first 12 analysed how possible values of the survival probability of philopatric (psp) and emigrant (pse) females vary depending on the fraction of individual disappearances assigned to emigration. We then obtained estimates of these parameters by combining data from our sampling of individual colonies - and their surroundings - with data from the four 150x5m plots. For all these analyses, we estimated the number of females surviving until the reproductive period as the largest number of adult females recorded in a colony-cohort (both in established and newly founded nests) once the first egg sac was laid and, in cases of colony dispersal, after no more newly founded nests were observed in the surroundings. The range of values that psp and pse can take, given the observed number of individuals surviving in both their natal and newly founded nests, depends on the fraction of individual disappearances assigned to death of philopatrics and to death of emigrants. If that is the case, as Waser et al. (1994) pointed ouX,psp andpse are related as follows: P.=-E (2) eN where P is the sum of all females surviving until reproduction in their natal nests, / the number of emigrants that succeeded in immigrating into a new nest, N the total number of maturing females before dispersal, and e the proportion of individual disappearances assigned to emigration (complementarily, 1 -e corresponds to the proportion of individual disappearances assigned to philopatry). Plotting equations (1) and (2), using estimates of P, I and N for each of the plots, enables us to evaluate all biologically possible values that psp and pse can take as a function of e (see Results figure 2.3). We calculated an average psp estimate as the proportion of females surviving from maturity to reproduction in all colonies where no dispersal was detected; i.e. a subset of the studied population for which individual disappearances can confidently be assigned to deaths (Waser et al. 1994). We then used this estimate to calculate a corresponding pse value for each of the sampled plots, applying, the relationship between those two parameters implicit in equations 13 (1) and (2). We finally estimated the average pse as the mean of the plot values, being aware that each plot represents statistically only one response unit and the confidence interval for our estimate is broad. We recognize some limitations of our sampling approach in plots. First, the number of individuals observed in newly founded nests inside a plot is only an indirect and minimum measurement of /, whose accuracy depends on the extent to which the number of individuals dispersing and surviving outside the monitored area is balanced with individuals immigrating successfully inside the plots (Waser et al. 1994). Second, the topography of the study area limited our ability to equally perform standard censuses to all nests. Specifically, one established nest in plot 4 was inaccessible during the first three months of monitoring, and a second nest in plot 2 contained too many dry leaves where the spiders hid out. Consequently, and despite the census error calibration, we recorded in these nests less subadult/adult females than adult females at the start of the reproductive period. Similarly, six newly founded nests in plot 3 and five in plot 4 were established in tall branches that made it impossible to examine their contents. We were forced then, in the first case, to compensate the number of subadult/adult females with the number expected assuming that spiders survived at the same rate than females surviving the same life history transition in all the other available nests, where the accessibility to perform reliable censuses was satisfactory. In the case of the inaccessible newly founded nests, we estimated their contents as equal to the number of females founding nests of similar dimensions, since nest size is a good predictor of the number of adult individuals that it contains (Aviles 1992). The probability of female reproduction— We calculated pr for philopatric and emigrant females as the proportion of females laying one egg sac in either their natal or in newly founded nests respectively. We estimated the total number of egg sacs produced by a colony-cohort, assuming that once an egg sac was recorded it most likely remained viable until either the appearance of newly born juveniles, the documentation of its hatching or destruction, or no longer than four successive censuses since the average development time for one egg sac is 42-56 days (P. Salazar, L. Aviles and G. Iturralde, unpubl. results). The number of maternal generation females that could potentially lay an egg sac was estimated as the largest number of adult females recorded at the time the first egg sac was laid. In nests where no egg sacs were laid, we used the largest number of females recorded three censuses after they 14 reached maturity. This is the median time between recently acquired maturity and the start of the reproductive period in colonies where egg sacs were laid. Offspring survival to maturity— We estimated pspi for philopatric and emigrant females as the proportion of female offspring born in either natal or newly founded nests that survived to maturity. We estimated the expected number of female offspring produced by a colony-cohort, by multiplying the total number of egg sacs the cohort produced by the expected number of females per egg sac in this species, assuming a mean proportion of females per colony of 0.92 (Iturralde 2004) and a mean number of 34 eggs per sac (P. Salazar, L. Aviles and G. Iturralde, unpubl. results). We excluded those offspring cohorts for which we did not observe whether the maternal generation was born in its nest of reproduction or it established the nest (7 out of 49 cases). We did however include data from nests of uncertain foundation in the analyses across group sizes. Other factors affecting fitness Besides the main effects of staying or dispersing from the natal nest, and group size, there are certainly other factors that affect the survival and reproduction of individual spiders. The most important reflected in our data are: parasitism by Ichneumonic wasps, nest destruction by environmental factors such as rain or plant substrate decaying, and predation by ants. Although such factors were associated with female survival, reproduction and even dispersal in some of the studied generations, their only effect was to increase the noise on the patterns that we analysed. When cohorts affected by these factors were excluded from the analyses there was no qualitative change in the relationships between fitness, dispersal and group size, but rather a clearer effect despite the sample size reduction: For that reason, in order to obtain the estimates and analyses reported here, we did not exclude or treat differently the data affected by those additional factors. RESULTS Colony dispersal as a function of group size According to our logistic regression analyses, the probability of colony dispersal increases • 2 gradually as the number of maturing females produced in a generation increases (R = 0.24; n = 27 colonies - 14 did disperse, 13 dispersed; likelihood ratio %2 = 8.93;p = 0.003; figure 15 2.2). In our sample, none of the six colonies containing <10 subadult/adult females dispersed whereas all five colonies containing >86 subadult/adult females (range: 86-240) did. Interestingly, approximately half of the colonies in the range 10-76 subadult/adult females did not disperse but the other half did, generating the gradual pattern of increase predicted. These results clearly show that dispersal is not a fixed life history strategy in this population of Anelosimus guacamayos, but depends on the number of females reaching maturity in a colony-generation. • • • •«•• • • • • Dispersal O No dispersal O O QQOCSD QD O O O ITT I'M | "VTTTT'ITT 1 t I M| I T I T I I 11 ] I I 1 I II 111 I I I I I 1 I I [ I I I I I I I 11 * I I "I ITI'lTf* 1 2 4 8 16 32 64 128 256 512 . G r o u p s i z e {log2 subadult adult females) Figure 2.2. Predicted increase in the probability of colony dispersal (dashed line) as the number of females reaching maturity in a generation increases; y = 1 / (1+ Exp(1.29-0.03x)). Dispersal occurrence data are plotted according to their corresponding group size; empty dots: colonies where no dispersal was detected, black filled dots: dispersal was detected. Consequences of dispersal on the survival probability of mature females to reproduction Figure 2.3 shows the range of values that psp and pse can take, given our data, as a function of the fraction of individual disappearances assigned to emigration (e) for the three plots where colony dispersal and proliferation was observed (plots 1, 3 and 4). In these plots we were also able to associate the establishment of 43 newly founded nests (84.31%) to the simultaneous decrease in numbers of a putative source colony; yet, we also recorded 8 newly founded nests (8 individuals in total) whose founder females most likely came from elsewhere. In plot 2, neither one of the three monitored colonies dispersed nor were any new nests recorded; \ 0 Q. (/> o o o CO -Q O 1.0 0.8 0.6 0.4 0.2 0.0 16 hence, no pse value could be estimated because all recorded disappearances in this plot (four spiders) were more confidently assigned to philopatric mortality. Plot 1 Plot 3 Plot 4 0) Q . CO x> 0.8 A ,0> M— Q. C CL o ' ••c o a. o 0.2 "SP (1-8)N -- philopatric - emigrant 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 Proportion females assigned to emigration (e) Figure 2.3. Range of possible values for the survival probability of philopatric (pxp) and emigrant (p,t.) females, for each of the three plots where colony dispersal and proliferation was observed. The values that psp and pse can take, given the observed number df females surviving in their natal (P) and newly founded (7) nests inside each plot, are functions of the proportion of individual disappearances assigned to emigration (e). P, I and N (the total number of maturing females before dispersal) were estimated from census data collected inside each plot. The grey bands indicate our average psp estimate ± 95% Cl (psp = 0.80 ± 0.04), from colonies where no dispersal was detected, and the corresponding pse for each plot. Two observations are worth highlighting from figure 2 . 3 . First, equations (2 ) and (3) define the lowest possible values for both psp and pse. At one extreme, the lowest psp possible, found when all individual disappearing from their natal nests are assumed to have stayed and died (consequently pse = 1 0 0 % ) , ranged from 1 2 . 4 4 % in plot 4 to 5 6 . 6 % in plot 1 (figure 2 . 3 ) . At the other extreme, the lowest pse possible, found when all disappearances are assumed to be emigrants that died in the dispersal process (consequentlypsp = 1 0 0 % ) , ranged from 3 . 3 7 % (plot 4 ) to 26.80%o (plot 3 ) (figure 2 . 3 ) . Second, althoughpsp could take both larger and smaller values than pse, all cases where psp < pse would imply the death of a major fraction of philopatric individuals - from at least 4 0 . 1 2 % in plot 1 up to at least 8 4 . 9 7 % in plot 4 . These fractions are much larger than any confident value observed in single colonies where no dispersal was detected (figure 2 . 4 ) . This suggests that the true averagepse value is more likely to be smaller than the average psp. This is consistent with the observation we made, while defining our sampling plots, that had we increased the area surveyed to record more dispersing individuals, it is unlikely that we would have found in newly founded nests a number of individuals close to the large numbers disappearing from dispersing nests. On the contrary, had we increased the sampling area it is possible we would have included additional 17 medium or large nests that would have increased the estimation of the proportion of philopatric survivors, in case of no dispersal, or reduced the proportion of emigrant survivors, in case of dispersal. ro c E .2 ^ o •§ 2 3 Q . ro £ E o c •> 9r w c o t o 1.0 0.8 J 0.6 4 0.4 0.2 0.0 • Dispersal O O No dispersal O rj C P O o o o o o o I ^ I i r i i i ] r 1111 i i i i i 1 2 4 8 16 32 64 128 256 512 G r o u p s i z e (log2 subadult/adult females) Figure 2.4. Proportion o f maturing females surviving to reproduction in their natal nests. Empty dots: colonies where no dispersal was detected, black f i l led dots: dispersal was detected. Figure 2.4 and table 2.2 summarize our analyses of the proportion of maturing females surviving to reproduction in their natal nests. This proportion was significantly higher in colonies where no dispersal was detected than in dispersing colonies (z = -23.61, one tailedp « 0.001; table 2.2). This pattern is particularly clear all through the range of group sizes where -50% of colonies dispersed (figure 2.4). Even though the proportion of females surviving in non-dispersing colonies seems to either be an asymptotic or a curvilinear function of group size, we found only significant support for its increase in the group-size range 1-20 subadult/adult females (R2 = 0.69; n = 10 colonies; y = 0.32 + 0.29 In (x); F = 13.05,p = 0.011; figure 2.4), and no support for its decrease in larger groups (R = 0.76; n = 4 colonies; y = 2.04 - 0.22 In (x); F=6.37,p = 0.128; figure 2.4). We found no relation with group size in colonies where dispersal occurred. Finally, the overall proportion of females surviving to reproduction in non-dispersing colonies, i.e. our average estimate for psp across the whole population, was 80.48 ± 4.55% (n = 292 individuals in 14 nests). 18 Table 2.2. Numbers and proportions of females surviving in and disappearing from their natal nest, during the transition from recently acquired maturity to the reproductive period. Both the absolute number and the proportion disappearing from colonies where no dispersal was detected were significantly much smaller than from dispersing colonies. Highlighted in bold is our estimate of the average survival probability for philopatric females (psp). Results shown in this table were obtained by sampling individual nests across the natural range of colony sizes. Dispersal # nests Colony size range Maturing females Females reproducing % surviving ± 95%CI Females disappearing % disappearing ± 95% CI No 14 3-77 292 235 80.48 ± 4.55% 57 19.52 ±4 .55% Yes 13 10-240 1188 155 13.05 ± 1.92% 1033 86.95 ± 1.92% Total 27 1480 390 26.35 ± 2.24% 1090 73.65 ±2.24% Table 2.3 summarizes the numbers and proportions of females estimated to have survived from recently acquired maturity to the reproductive period, both in their natal and in newly founded nests, inside the four 150x5m plots sampled. Using our averagepsp estimate we inferredpse values that ranged from 3.48% in plot 4 to 31.08% in plot 3. A l l these calculations and analyses support the idea that the average survival probability of dispersing A. guacamayos females is relatively low, and that it is lower than the survival of individuals that stay in their natal nest. Table 2.3. Numbers and proportions of females surviving the transition from recently acquired maturity to the reproductive period inside our four 150x5m sampled plots. Estimates of the survival probability for emigrants (/?.,„) were calculated using the estimated average survival probability for philopatrics (psp = 80.48%), and the relationship between those two parameters described by equations (2) and (3) (see text for details). Plot N # nests (N) Colony size range # nests that dispersed P # nests (P) %P / # nests % / Pse 1 172 10 4-35 4 90 10 52.33% 13 13 7.56% 21.61% 2 25 3 4-12 0 21 3 84.00% 0 3 240 1 240 1 87 1 36.25% 41 25 17.08% 31.08% 4 439 3 52-223 2 53 3 12.07% 13 13 2.96% 3.48% Mean psi±95% p = 18.72±15.87% Note: A ' = total number of females maturing inside each plot; P = adult females found surviving in their natal nests; 7 = adult females surviving in newly founded nests; # nests () = total number of nests recorded inside each plot before dispersal ( A O , after dispersal in natal nests (P) and in newly founded nests (/). The probability of female reproduction Over the entire population, we estimated that only 49.16 ± 3.17% (average ± 95% CI; n = 952 individuals in 81 nests) of the adult females alive at the start of the reproductive period lay one egg sac. When compared against each other, there is a statistically significant difference between the proportion of females reproducing in their natal (47.81 ± 3.32%, n = 868 individuals in 29 nests) versus newly established nests (63.1 ± 10.12%, n = 84 19 individuals in 52 nests) (z = -2.67, two tailed /J>=0.007; figure 2.5A). Although at first such a difference could be attributed to group size, it seems that the apparent negative relation between female reproduction and group size shown in figure 2.5B is in fact driven by the effect of colony dispersal in the data. Specifically, most of the relatively low pr values seen in large colonies reflect the fact that many of these colonies produced dispersers during the reproductive period, so that relatively large fractions of females disappeared from their natal nests before laying egg sacs. If this effect is accounted in the analyses or the data from dispersing nests are excluded, the difference in the probability of female reproduction in natal versus newly founded nests disappears (ANCOVA R = 0.35; F = 0.95 andp = 0.34 for nest of reproduction; F = 0.86 and p = 0.363 for group size; F = 5.63 and p = 0.026 for dispersal occurrence; figure 2.5B). Thus, we found no convincing evidence that the probability of females' reproducing in their natal nests differs from females that have dispersed and established new nests. c o Z3 T3 O i_ CL 0 c/> _0 CD c o '•c o CL 2 1.0-f 1.0 -f 0.8 0.6 0.4 0.2 0.0 B <S at least one dispersed colony ° ° °* ° • O S ® C L o o S ® Newly founded Natal Nes t of reproduct ion 4 8 16 32 64 G r o u p s i z e (log2 subadult/adult females) " U > > i 128 256 Figure 2.5. (A) Proportion of females laying egg sacs in their natal versus newly founded nests; error bars show ,95% CI. (B) Same data but plotted in relation to group size. Diagonally marked data points in (B) include at least one dispersed colony of that size. The probability of offspring survival to maturity Over the entire population, we estimated that only 15.02 ± 0.85% (n = 6719 expected offspring, from 215 egg sacs produced in 41 nests) of females, expected to come out from recorded egg sacs, survive to sexual maturity. Comparing pspi for those females whose nest of birth was known (213 out of 275 females in 42 nests), there is a statistically significant difference between philopatric (13 ± 1.07%, n = 3784 expected offspring, from 121 egg sacs 20 in 10 nests) and emigrant females (2.6 ± 0.88%, n = 1309 expected offspring, from 42 egg sacs in 32 nests) (z = -10.66, two tailedp < 0.001; figure 2.6A). In this case the detected difference is better attributed to group size than merely maternal philopatry, when both variables are incorporated into the analyses (ANCOVA R2 = 0.77; F= 0.59 and p = 0.459 for nest of birth; F= 24.39 and p < 0.001 for group size; figure 2.6B). D5 > </) c CL o c o "t: o Q. O i_ QL 0.4- A • Newly founded O Established 0.3-0.2-0.1 - s 0.0-s Newly founded Established 0.4- B c c C Uncertain 0.3- c o o c 0.2-0.1 - o o o 0.0-• • o © • o o 16 32 64 128 256 N e s t of birth G r o u p s i z e (log2 subadult/adult females) Figure 2.6. (A) Proportion of expected female offspring that survive to sexual maturity in nests established at least in the previous generation versus newly founded nests; error bars show 95% CI. (B) Same data but plotted in relation to group size. D I S C U S S I O N We found three main fitness consequences of the "choice" that adult Anelosimus guacamayos females make to either stay or disperse from their natal group. First, philopatric females have an overall higher survival probability than emigrants as a result of avoiding mortality due to dispersal (figures 2.3, 2.4; tables 2.2, 2.3). Second, there is some improvement in philopatrics' survival associated with increasing group size, but only in colonies smaller than -20 females (figure 2.4). The detection of potential survival costs attributable to living in even larger groups was prevented by the increasing occurrence of colony dispersal as groups grow larger (figure 2.2). Third, there is an increased offspring survival more strongly associated with group size than merely philopatry (figure 2.6). For emigrants the patterns are the opposite; reduced survival due to dispersal and lower offspring survival due to solitary living or small group size (figures 2.3, 2.4, 2.6; tables 2.2, 2.3). In this case, our data and methodology did not allow us to test for group size effects on emigrants' survival. In addition to these results, we found no clear evidence of either positive or negative consequences of 21 dispersal or group size on the probability of female reproduction (figure 2.5). These fitness patterns bring up several questions: why, despite the negative fitness consequences of dispersal, do some females still disperse rather than increase group size even more? Why does group living increase offspring survival? And finally, what do these fitness patterns contribute to our understanding of the evolution of sociality in spiders? To understand why A. guacamayos females still disperse, despite the fitness costs implied, it is necessary to evaluate both possible proximate and ultimate reasons for this behaviour. On the one hand, the density-dependent dispersal pattern observed (figure 2.2) suggest that the immediate triggers of the "decision" to leave nests must be negative consequences of group living, not explicit in our demographic approach. For instance, besides recording survival and reproduction, we did not observe systematically the physical condition of individuals in relation to their group size. Nevertheless, observations done in other studies make it reasonable to hypothesize that the trigger of dispersal is intraspecific competition for food resources due to the reduction of the prey availability per capita as group size increases. In a partly simultaneous study, E.Yip, K. Powers and L. Aviles (unpubl. results) found that colony size significantly affected per capita prey capture in A. guacamayos. Ward (1986) also found, in the permanent-social spider Stegodyphus mimosarum, that prey availability per capita decreases with nest size (indicator of the prey capture capacity of a group) and as the number of spiders in a colony increases. It has also been shown that when periodic-social spiders are artificially supplemented with food the natal dispersal phase is delayed or at least prolonged (e.g. Krafft et al. 1986; Ruttan 1990; Kim 2000), and conversely when individuals are starved dispersal occurs more quickly (Bodasin et al. 2002). Thus, it is most likely that the environment, specifically food availability, triggers dispersal in A. guacamayos. On the other hand, the evolution of dispersal tendencies and abilities is traditionally explained as a balance between the costs of dispersal and the separated or combined effects of kin interactions, inbreeding and spatio-temporal variability in habitat quality (For reviews see: Johnson and Gaines 1990; Clobert et al. 2001; Bowler and Benton 2005). A frequent result of theoretical models is that the factors favouring dispersal can make it to evolve even if it implies very high fitness costs. However, the fitness costs of dispersal have to always be < 100%, because they can not exceed the benefits of moving into a new patch (Johnson and Gaines 1990; Sylvain and Michalakis 2001; Bowler and Benton 2005). In our case, the most 22 intuitive evolutionary explanation for A. guacamayos dispersal would be that the fitness of spiders that take the risk of dispersing is higher than what would be expected had they stayed in their crowded natal group. Although our approach does not allow us to test this hypothesis, since it would be necessary to experimentally force individuals to stay in relatively large groups, our observations allow us to assess how plausible this idea is. The first thing to realize is that the fitness of A. guacamayos dispersers is low but it is not zero (figure 2.3; table 2.3). Therefore, any genetic component driving the behavioural tendency to disperse under stressful group living conditions would be favoured by the successful dispersal of some individuals. Conversely increasing philopatric tendencies would be enhanced by successful generations of philopatrics. A second aspect is that the dispersal phenomena in a permanently social spider, such as A. guacamayos, is not equivalent to the analogous process in their periodic social relatives. Natal dispersal in periodic-social spiders is a fixed life history strategy that all individuals experience during their lifetime, whereas dispersal in permanent-social spiders is a plastic response to environmental triggers (i.e. a "condition dependent strategy" sensu Bowler and Benton 2005). In the case of periodic-social species the evolutionary explanation for dispersal is likely to be a trade-off between factors enhancing maternal care, the costs of dispersal and the ever present action of factors favouring the evolution of increased dispersal (e.g. kin competition, inbreeding, habitat heterogeneity). In our case, as well as in other permanent-social species, the reduced frequency of dispersal is better understood as a partial loss of an ancestral behaviour, but its evolutionary maintenance is could be a response to natural selection favouring dispersal over, a perhaps more costly strategy, of remaining in increasingly larger groups. i • The main fitness benefit of group living found in our study was the increase in offspring survival for females living in larger groups. This pattern was also found by Aviles and Tufmo (1998) in another permanent-social Anelosimus species (A. eximius). Interestingly, group living benefits in this fitness component do not need to be a consequence of active cooperation among nest inhabitants. It might simply be that the chances of offspring survival are improved by the presence and activity of other adult spiders that indirectly help spiderlings, particularly in the absence of their mother, by maintaining the nest and capturing prey. It is common to observe, both in periodic and permanent social species, that when solitary nesting females die or are removed from their nest their offspring do not survive by themselves unless they are old enough to take care of the web and capture their own food. In 23 -fact, early death of solitary nesting females is perhaps the main failure of brood raising for periodic social spiders (Aviles and Bukowski 2006, pers. obs.). Under group living conditions the chances of ending up with no adults in the nest decrease geometrically for each adult that is added to the group. This is perhaps the first adaptive effect of group living, during the transition from periodic to permanent sociality, i f the environment allows the formation of groups. Evolutionary transitions involving the tension between staying and dispersing from the natal group have been claimed as a component of the evolutionary route towards more complex sociality in the majority of the groups where social behaviour has evolved (e.g. Thorne 1997; Hunt 1999; Hatchwell and Komdeur 2000). Although the factors involved in the evolution of sociality and the evolution of dispersal have often been associated with each other, they have seldom being explicitly considered as interactive processes (exceptions include Perrin and Lehmann 2001; Le Galliard et al. 2005). Linking these two processes will be useful in understanding even more complex social phenomena than just the nascent transition from solitary to group living. As Crespi (2004) pointed out, what we probably see when looking at lineages where social behaviour has evolved is the result of a positive feedback cycle through which more complex and tight social interactions evolve coupled with new dispersal strategies. 24 BIBLIOGRAPHY Agnarsson, I. 2004. Morphological phylogeny of cobweb spiders and their relatives (Araneae, Ananeoidea, Theridiidae). Zoological Journal of the Linnean Society 141:447-626. —. 2006. A revision of the New World eximius lineage of Anelosimus (Araneae, Theridiidae) anda phylogenetic analysis using worldwide exemplars. Zoological Journal of the , Linnean Society 146:453-593. Agnarsson, I., L. Aviles, J. A . Coddington, and W. P. Maddison. Submitted. Sociality, inbreeding and sex ratio bias in spiders - repeated evolution of an evolutionary dead-end strategy. Evolution. Aviles, L. 1992. Metapopulation biology, levels of selection and sex ratio evolution in social spiders. Ph.D. thesis, Harvard University, Cambridge, M A , USA. —. 1997. Causes and consequences of cooperation and permanent sociality in spiders, Pages 476-498 in J. C. Choe, and B. J. Crespi, eds. The evolution of social behaviour in insects and arachnids. 476-498. Cambridge University Press. Cambridge. Aviles, L., and T. C. Bukowski. 2006. Group living and inbreeding depression in a subsocial spider. Proceedings of the Royal Society of London - Series B: Biological Sciences 253:157-163. Aviles, L., and G. Gelsey. 1998. Natal dispersal and demography of a subsocial Anelosimus species and its implications for the evolution of sociality in spiders. Canadian Journal of Zoology 76:2137-2147. Aviles, L., and P. Tufino. 1998. Colony size and individual fitness in the social spider Anelosimus eximius. American Naturalist 152:403-418. Bilde, T., Y. Lubin, D. Smith, J. M . Schneider, and A. A . Maklakov. 2005. The transition to social inbred mating systems in spiders: role of inbreeding tolerance in a subsocial predecessor. Evolution 59:160-174. Bodasin, M . , T. Crouch, and R. Slotow. 2002. The influence of starvation on dispersal in the social spider, Stegodyphus mimosarum (Araneae: Eresidae). Journal of Arachnology 30:373-382. Bodasin, M . , R. Slotow, and T. Crouch. 2001. The influence of group size on dispersal in the social spider Stegodyphus mimosarum (Araneae, Eresidae). Journal of Arachnology 29:56-63. Bowler, D. E., and T. G. Benton. 2005. Causes and consequences of animal dispersal strategies: relating individual behaviour to spatial dynamics. Biological Reviews 80:205-225. Chistenson, T. R. 1984. Behaviour of colonial and solitary spiders of the Theridiid species Anelosimus eximius. Animal Behaviour 32:725-734. Choe, J. C , and B. J. Crespi, eds. 1997, The evolution of social behavior in insects and arachnids. Cambridge University Press. Cambridge Clobert, J., E. Danchin, A. A. Dhondt, and J. D. Nichols, eds. 2001, Dispersal. Oxford University Press. Oxford Crespi, B. J. 2004. Vicious circles: positive feedback in major evolutionary transitions. Trends in Ecology and Evolution 19:627-633. 25 Hatchwell, B. J., and J. Komdeur. 2000. Ecological constraints, life history traits and the evolution of cooperative breeding. Animal Behaviour 59:1079-1086. Hunt, J. H. 1999. Trait mapping and salience in the evolution of eusocial vespid wasps. Evolution 53:225-237. Iturralde, G. 2004. Aspectos evolutivos de la proportion de sexos en aranas sociales de la Amazonia ecuatoriana (Evolutionary aspects of sex ratio in social spiders from Amazonian Ecuador). Licenciature thesis, Ponticia Universidad Catolica del Ecuador, Quito, Ecuador. Johnson, M . L., and M . S. Gaines. 1990. Evolution of dispersal: theoretical models and empirical tests using birds and mammals. Annual Review of Ecology and Systematics 21:449-480. Kim, K. W. 2000. Dispersal behaviour in a subsocial spider: group conflict and the effect of food availability. Behavioral Ecology and Sociobiology 48:182-187. Krafft, B., A . Horel, and J. Julita. 1986. Influence of food supply on the duration of the gregarious phase of a maternal-social spider, Coleotes terrestris (Araneae, Agelenidae). Journal of Arachnology 14:219-216. Le Galliard, J.-F., R. Ferriere, and U . Dieckmann. 2005. Adaptive evolution of social traits, origin, trajectories and correlations of altruism and mobility. The American Naturalist 165:206-224. Nentwig, W. 1985. Social spiders catch larger prey: a study of Anelosimus eximius (Araneae: Theridiidae). Behavioral Ecology and Sociobiology 17:79-85. Perrin, N . , and L. Lehmann. 2001. Is sociality driven by the cost of dispersal or the benefits of philopatry? A role for kin-discrimination mechanisms. The American Naturalist 158:471-483. Powers, K. S. 2004. Prey abundance and the evolution of sociality in the social spider genus Anelosimus. Ph.D. thesis, University of Arizona, Tucson, A Z , USA. Powers, K. S., and L. Aviles. 2003. Natal dispersal patterns of a subsocial spider Anelosimus cf. jucundus (Theridiidae). Ethology 109:725-737. Riechert, S. E. 1985. Why do some spiders cooperate? Agelena consociata, a case study. Florida Entomologist 68:105-116. Ruttan, L. M . 1990. Experimental manipulations of dispersal in the subsocial spider, Theridion pic turn. Behavioral Ecology and Sociobiology 27:169-174. Salazar, P. A. 2001. Dinamica de las poblaciones de la arana social Anelosimus domingo en la Amazonia ecuatoriana (Population dynamics of the social spider Anelosimus domingo in the Ecuadorian Amazon). Licenciatura thesis, Pontificia Universidad Catolica del Ecuador, Quito, Ecuador. Sylvain, G., and Y. Michalakis. 2001. Multiple causes of the evolution of dispersal, Pages 155-167 in J. Clobert, E. Danchin, A. A . Dhondt, and J. D. Nichols, eds. Dispersal. 155-167. Oxford University Press. Oxford. Thorne, B. L. 1997. Evolution of eusociality in termites. Annual Review of Ecology and Systematics 28:27-54. Thorne, B. L., and J. F. A . Tranielo. 2003. Comparative social biology of basal taxa of ants and termites. Annual Review of Entomology 48:283-306. 26 Ward, P. I. 1986. Prey availability increases less quickly than nest size in the social spider Stegodyphus mimosarum. Behaviour 97:213-225. Ward, P. I., and M . M . Enders. 1985. Conflict and cooperation in the group feeding of the social spider Stegodyphus mimosarum. Behaviour 94:167-182. Waser, P, M . , S. R. Creel, and J. R. Lucas. 1994. Death and disappearance: estimating mortality risk associated with philopatry and dispersal. Behavioral Ecology 5:135-141. Whitehouse, M . E. A. , and Y. Lubin. 2005. The functions of societies and the evolution of group living: spider societies as a test case. Biological Review 80:347-361. 27 C H A P T E R 3: General discussion As I pointed out in the previous chapters, the transition from solitary to group living in spiders, as well as in several other animal taxa, involves the progressive loss of the natal dispersal behaviour. In this context, two of the fitness patterns observed in Anelosimus guacamayos are enlightening about why and how the frequency of individuals staying in their natal group increases in a population initially composed of only solitary living individuals. The first of these patterns is the increased survival probability of philopatric females relative to emigrants (chapter 2: figure 2.3, table 2.3). This is important because, if the environment allows the formation of groups, it is probably why philopatric tendencies are initially favoured over dispersal, and thus increase in frequency. The second interesting pattern is the increased offspring survival, more strongly associated with group size than merely philopatry (figure 2.6). This pattern constitutes an additional fitness benefit of philopatry on top of avoiding the dispersal cost. These observations suggest an intriguing question about A. guacamayos behaviour that was not explicitly discussed in the previous chapter: why, despite its potential benefits, do spiders not disperse collectively instead of as individuals? This question concerns a broader problem in social evolution; that is, what conditions are necessary for the evolution of increasingly complex social behaviours - such as cooperative group dispersal. In the following paragraphs I will discuss some ideas about how dispersal and the evolution of social complexity are related. Cooperative group dispersal has been reported for several permanent social spiders (e.g. Lubin and Robinson 1982; Vollrath 1982; Aviles 2000) and is a behaviour observed in other social organisms as well (e.g. Peeters and Ito 2001; Anderson et al. 2002; Janson et al. 2005). Still, dispersing^, guacamayos females leave their nest alone, and often build a new nest on their own. The cases of multifemale nest foundation are in fact the result of a sequence of solitary dispersing females colonizing a nest initially established by a single pioneer female. Cooperative group dispersal is a more complex social behaviour that could enhance successful dispersal but has not evolved in A. guacamayos. Species 28 where group dispersal is frequent also show other complex cooperative behaviours, such as active brood caring or division of labour (Aviles 1993; Peeters and Ito 2001; Anderson et al. 2002). So, the question of why A. guacamayos does not disperse in groups can be discussed in the broader context of the evolution of increasing social complexity. I argue that the progressive evolution of social complexity depends on the duration of social groups and the costs of dispersal, both ultimately dependent on attributes of the environment. In the spiders' case, environmental features, specifically food availability per capita, not only set a limit to colony growth but also determine how many generations groups last. This has two implications: on the one hand, the more often groups dissolve, the more frequently characteristics useful for solitary living would be favoured. On the other hand, the more generations groups persist, the more cooperative tendencies can be accumulated. This second implication arises because any nascent cooperative behaviour is expected to increase in frequency across the whole population just because groups with more cooperative individuals are more productive. Fletcher and Zwick (2004) showed theoretically that altruistic characters can evolve more easily in groups that last for an intermediate number of generations. This is because groups dominated by altruists can compound the benefits of cooperation to grow much larger than other groups, even though the fraction of altruists declines within all groups. As a consequence, over the whole population, the number and frequency of altruists can increase. This effect of group duration may explain in part the lack of highly complex social behaviours, such as cooperative group dispersal, in A. guacamayos, in contrast to species with more complex social behaviours that inhabit environments that can afford larger and longer lasting groups. Dispersal costs have long been discussed in the evolution ofcomplex social behaviours (e.g. Emlen 1982; Vehrencamp 1983). In general, the idea is that low expected fitness for emigrants, in addition to favouring philopatry, also favours improved social interactions (Vehrencamp 1983; Perrin and Lehmann 2001). Perrin and Lehmann (2001) demonstrated that, in the absence of elaborated kin-discrimination mechanisms - which is 29 probably the case for spiders, dispersal costs are a necessary prerequisite for the evolution of altruistic traits. Likewise, Vehrencamp (1983) showed that the degree of bias in fitness or resource acquisition within social groups (which leads to extreme division of labour), is limited by the opportunities to survive and reproduce that subordinated individuals have outside groups. Consistent with these ideas, many spider species that show more complex social behaviours, including cooperative group migration, inhabit lowland tropical environments that are likely to be harsher for dispersal (due to rainfall or predation; pers. obs.; L. Aviles pers. comm.) than the mid-elevations cloud forests that A. guacamayos inhabits (e.g. Lubin and Robinson 1982; Vollrath 1982; Riechert 1985; Aviles 2000). Not until recent years have sociality and dispersal been explicitly considered as jointly evolving characters (Perrin and Lehmann 2001; Le Galliard et al. 2005), most likely because each of these attributes is complex enough to deserve its own research program. In this thesis work, I have tried to think in the evolution of group-living and dispersal as a combined process, by looking at their consequences under a single methodological framework - demography. In this case, my collaborators in chapter two and I took advantage of a species whose ecological and phylogenetic context allowed us to make reasonable evolutionary interpretations of our empirical results. These interpretations concerned both the original questions that motivated this study - why some spider species are social? and how does social behaviour evolve?, and non-traditional implications of those problems - how are sociality and dispersal evolution related? However, as it often occurs, I have proposed some answers but opened more questions and opportunities for new projects. For instance, a necessary next step is a systematic research about the extend to which the fitness patterns found in this study are a general consequence of the transition from solitary to group-living in other organisms. Likewise, now that more confident and complete phylogenetic reconstructions of the genus Anelosimus - the most diverse in periodic and permanent social species in the neotropics - are available (Agnarsson 2004; 2006), new venues for comparative studies are possible. Especially taking advantage of the variety of social systems and environments of closely related species. Further ecological studies looking at both the factors allowing the 30 formation of groups (e.g. prey abundance and size) and the costs of dispersal (e.g. rainfall and predation) will be useful in developing a formal conceptual framework to understand the joint evolution of sociality and dispersal. Experimental manipulations of group size will be valuable to quantify the costs of group living. A l l these new research opportunities are possible because of the repeated evolution of group living and/or the loss of dispersal in spiders as well as in several other organisms. These phenomena has provided the natural experiment to test our ideas about the two major and interrelated behavioural and life history attributes here studied: sociality and dispersal. 31 BIBLIOGRAPHY Agnarsson, I. 2004. Morphological phylogeny of cobweb spiders and their relatives (Araneae, Ananeoidea, Theridiidae). Zoological Journal of the Linnean Society 141:447-626. —. 2006. A revision of the New World eximius lineage of Anelosimus (Araneae, Theridiidae) and a phylogenetic analysis using worldwide exemplars. Zoological Journal of the Linnean Society 146:453-593. Anderson, C., G. Theraulaz, and J. L. Deneubourg. 2002. Self-assemblages in insect societies. Insectes Sociaux 49:99-110. Aviles, L. 1993. Newly-discovered sociality in the neotropical spider Aebutina binotata Simon (Dictynidae?). Journal of Arachnology 21:184-193. —. 2000. Nomadic behaviour and colony fission in a cooperative spider: life history evolution at the level of the colony? Biological Journal of the Linnean Society 70:325-339. Emlen, S. T. 1982. The evolution of helping. I. An ecological constraints model. The American Naturalist 119:29-39. Fletcher, J. A. , and M . Zwick. 2004. Strong altruism can evolve in randomly formed groups. Journal of Theoretical Biology 228:303-313. Janson, S., M . Middendorf, and M . Beekman. 2005. Honeybee swarms: how do scouts guide a swarm of uninformed bees? Animal Behaviour 70:349-358. Le Galliard, J.-F., R. Ferriere, and U . Dieckmann. 2005. Adaptive evolution of social traits, origin, trajectories and correlations of altruism and mobility. The American Naturalist 165:206-224. Lubin, Y. , and M . H. Robinson. 1982. Dispersal by swarming in a social spider. Science 216:319-321. Peeters, C., and F. Ito. 2001. Colony dispersal and the evolution of queen morphology in social hymenoptera. Annual Review of Entomology 46:601-630. Perrin, N . , and L. Lehmann. 2001. Is sociality driven by the cost of dispersal or the benefits of philopatry? A role for kin-discrimination mechanisms. The American Naturalist 158:471-483. Riechert, S. E. 1985. Why do some spiders cooperate? Agelena consociata, a case study. Florida Entomologist 68:105-116. Vehrencamp, S. L. 1983. A model for the evolution of despotic versus egalitarian societies. Animal Behaviour 31:667-682. Vollrath, F. 1982. Colony foundation in a social spider. Zeitschrift fur Tierpsychologie 60:313-324. 32 

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}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0092746/manifest

Comment

Related Items