Effects of developmental plasticity and antagonistic

selection on phenotypic variation in spiders

by

Michael Matthew Kasumovic

A thesis submitted in conformity with the requirements

for the degree of Doctor of Philosophy

Graduate Department of Ecology and Evolutionary Biology

in the University of Toronto

© Copyright by Michael Matthew Kasumovic 2008 Library and Bibliotheque et 1*1 Archives Canada Archives Canada Published Heritage Direction du Branch Patrimoine de I'edition

395 Wellington Street 395, rue Wellington Ottawa ON K1A0N4 Ottawa ON K1A0N4 Canada Canada

Your file Votre reference ISBN: 978-0-494-39827-2 Our file Notre reference ISBN: 978-0-494-39827-2

NOTICE: AVIS: The author has granted a non­ L'auteur a accorde une licence non exclusive exclusive license allowing Library permettant a la Bibliotheque et Archives and Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par telecommunication ou par Plntemet, prefer, telecommunication or on the Internet, distribuer et vendre des theses partout dans loan, distribute and sell theses le monde, a des fins commerciales ou autres, worldwide, for commercial or non­ sur support microforme, papier, electronique commercial purposes, in microform, et/ou autres formats. paper, electronic and/or any other formats.

The author retains copyright L'auteur conserve la propriete du droit d'auteur ownership and moral rights in et des droits moraux qui protege cette these. this thesis. Neither the thesis Ni la these ni des extraits substantiels de nor substantial extracts from it celle-ci ne doivent etre imprimes ou autrement may be printed or otherwise reproduits sans son autorisation. reproduced without the author's permission.

In compliance with the Canadian Conformement a la loi canadienne Privacy Act some supporting sur la protection de la vie privee, forms may have been removed quelques formulaires secondaires from this thesis. ont ete enleves de cette these.

While these forms may be included Bien que ces formulaires in the document page count, aient inclus dans la pagination, their removal does not represent il n'y aura aucun contenu manquant. any loss of content from the thesis. Canada 11 The influence of antagonistic selection pressures on the maintenance of phenotypic variation

Doctor of Philosophy, 2008

Michael Matthew Kasumovic, Department of Ecology, Evolution and Behaviour, University of

Toronto

ABSTRACT

In this dissertation, I ask whether developmental plasticity can explain the variation of phenotypic distributions as an adaptive response to variation in population demography and changes in the strength and direction of selection. I use three spider species to examine this question. Using a combination of field attraction and mark-recapture experiments, I first ask how male spiders locate females and the choices they make while mate searching. I demonstrate that males can distinguish between females of different species, populations and ages using long-distance pheromones, and that males increase mate searching risks because they are searching for specific females. As these long-distance pheromones can provide cues of female density, I next examine whether juvenile males alter their development when reared in the presence or absence of female's pheromones (cues of high and low female density, respectively). My results demonstrate that males alter their ontogeny to mature the phenotype that is most beneficial in the competitive context they are likely to experience at maturity. Males develop significantly faster when females are present, and are significantly larger and in better condition when females are absent. Since small size is apparently the result of a decision that is independent of resource availability, I next examine whether male fitness is phenotype specific.

By testing small and large males in the competitive environments in which they mature, I show that although larger males are superior in direct competitions, smaller males have higher fitness when tested in the context that leads to their more rapid development. These results challenge the concept of male quality as a fixed trait value, demonstrating the necessity of taking life- history traits into consideration. 1 next use field populations to demonstrate that demographic iii variables fluctuate within a season, and depend on the scale of the examination. The strength of selection pressures also varied significantly throughout the breeding season. As a result, males are likely to experience different competitive challenges and selection pressures at different spatial and temporal scales, making a single phenotypic optimum unlikely. I end with a discussion regarding the adaptive nature of developmental plasticity, and when it is likely to evolve. iv

ACKNOWLEDGEMENTS

There are many people that I would like to thank for helping me during these last five years. However, out of everyone, there's no person that Id like to thank more than my supervisor, Maydianne Andrade. Maydianne has been a supportive and patient supervisor throughout my entire degree. She has provided guidance, offering excellent advice, and has allowed me to continue research in directions that interested me. Overall, this has helped me to grow as a researcher. I know that I could not have been as successful without Maydianne. I'd also like to thank all the others I've worked with in the Integrative Behaviour and Neuroscience group, especially DamianElias and Andrew Mason. Both have been great collaborators and friends. There were also a large number of undergrads that have helped me rear the necessary study organisms and helped with experiments, most notably Kuhan Perampaladas, Nafis Tin, and Shivani Bhalla. I'd also like to thank Mariella Herberstein for all her support and for giving me a home while I was in Australia. Mariella was like a second supervisor while I was in

Australia and I've been lucky to have a collaborator like her. I'd also like to thank my committee Darryl Gwynne, Helen Rodd, and Locke Rowe for their insightful comments and discussions regarding my thesis that have helped make it into what it is, and especially for offering advice regarding research and postdoctoral decisions. Thanks to all the graduate students in my cohort for their support and interaction, especially Kevin Judge, Mark

Fitzpatrick, Kevin Wilson, and of course, David Punzalan. They all provided many a great discussion on science that helped me to shape some of my ideas and also provided many opportunities to relax over a cold beer or a snowboard strapped to my feet. I also wouldn't be able to get as much done as I did if it weren't for the administrative staff that helped me and made my life easier, especially Brenda Brown, Lucy Pickering, Nella Semoff, and Josie Valotta.

Academic life ran more smoothly with all of their help. I would also like to take this opportunity to thank NSERC and OGS for their funding during the duration of my doctoral thesis and V Canadian Journal of Zoology, Current Biology, and Behavioral Ecology for providing copyright permission to allow reprinting of my already published papers in my thesis. And most importantly, I would also like to thank my wife Beth and my family for their support during my

Ph.D. and for being so patient during all the times where things needed to be done. My whole family's continued support made it easier to finish my thesis. VI

TABLE OF CONTENTS

ABSTRACT 11

ACKNOWLEDGEMENTS iv

TABLE OF CONTENTS vi

LIST OF TABLES ix

LIST OF FIGURES x

CHAPTER 1: GENERAL INTRODUCTION 1

Literature Cited 13

CHAPTER 2: DISCRIMINATION OF AIRBORNE PHEROMONES BY MATE-SEARCHING

WESTERN BLACK WIDOW MALES (Latrodectus hesperus): SPECIES-

AND POPULATION-SPECIFIC RESPONSES 25

Abstract 26

Introduction 27

Methods 29

Results 33

Discussion 35

Acknowledgements 39

Literature Cited 41

CHAPTER 3: RISKY MATE SEARCH AND MATE PREFERENCE IN THE GOLDEN

ORB-WEB SPIDER {Nephila plumipes) 50

Abstract 51

Introduction 52

Methods 53

Results 57

Discussion 60 Acknowledgements

Literature Cited 66

CHAPTER 4: MALE DEVELOPMENT TRACKS RAPIDLY SHIFTING SEXUAL

VERSUS NATURAL SELECTION PRESSURES 78

Acknowledgements 81

Literature Cited 82

Supplementary Data 88

Supplemental Literature Cited 92

CHAPTER 5: A CHANGE IN THE COMPETITIVE CONTEXT REVERSES SEXUAL

SELECTION ON MALE SIZE 95

Abstract 96

Introduction 97

Methods 100

Results 106

Discussion 108

Acknowledgements 113

Literature Cited 114

CHAPTER 6: SPATIAL AND TEMPORAL DEMOGRAPHIC VARIATION DRIVES

WITHIN-SEASON FLUCTUATIONS IN SEXUAL SELECTION 125

Abstract 126

Introduction 127

Methods 129

Results 132

Discussion 139

Acknowledgements 143 vm Literature Cited 144

CHAPTER 7: GENERAL DISCUSSION 162

Conclusion 170

Literature Cited 172 ix

LIST OF TABLES

Table 1.1: Examples of developmental plasticity in response to various 22

environmental factors

Table 1.2: Six criteria required for plasticity to be considered adaptive 24

Table 2.1: The number of males caught at different sampling cages 45

Table 2.2: The G-statistic values for effect of treatment on male

attraction to cages 46

Table 3.1: Nominal logistic analysis examining female attraction of males 71

Table 3.2: Results of two separate nominal logistic analyses examining

male choice of females 73

Table 4.1: Examination of the effect of female presence and diet on

male phenotypes 84

Supplemental Table 4.1: Distance of the nearest penultimate or adult

female in the field affects the adult phenotype 94

Table 6.1: Effects of population, body size and condition on male proximity

to the hub of female's webs when rival males are present. 151

Table 6.2: Non-linear univariate selection gradients (±S.E.) on size and

condition and non-linear bivariate selection gradients. 152 LIST OF FIGURES

Figure 2.1: Spatial distribution of webs of female L. hesperus across two

adjacent field sites 48

Figure 2.2: Box plot comparing male body size and embolus length 49

Figure 3.1: Survival analysis of rate of capture for all males that were

eventually recovered in females' webs during two release

experiments 74

Figure 3.2: The null distribution of male size of released individuals 75

Figure 3.3: The null distribution of male condition of released individuals 76

Figure 3.4: The number of males attracted to the webs of adult and

penultimate females in a field enclosure release experiment 77

Figure 4.1: The differences in phenotypes of males reared in the absence

or presence of virgin females across three diet treatments 86

Figure 5.1: A sample distribution of females in a field enclosure and the

relative location of male release 121

Figure 5.2: The number of small and large males that successfully located

a female in each treatment. 122

Figure 5.3: The number of small and large males that a) successfully

mated at least once versus b) those that successfully

mated first insuring increased paternity. 123

Figure 5.4: The mean predicted paternity of small and large males in

each treatment. 124

Figure 6.1: A frequency distribution of male a) size and b) body

condition in each population. 155 xi Figure 6.2: The distribution of females of different ages in each

population throughout the breeding season. 156

Figure 6.3: The proportion of penultimate instar and adult females at the

aggregations and local level in each population throughout

the season. 157

Figure 6.4: The operational sex ratio at the aggregation and local level

in each population throughout the season. 158

Figure 6.5: The number of multiple male webs found in each

population throughout the season. 159

Figure 6.6: Frequency distribution of the number of webs with

multiple males in each population. 160

Figure 6.7: The strength of selection on male size and body condition

in each population throughout the season. 161 1

CHAPTER ONE

General Introduction 2

Understanding the origin and maintenance of phenotypic variation has always interested evolutionary biologists. As phenotypes are a product of genes and the environment, phenotypic distributions have been considered a consequence of the expression of a variety of genotypes in a heterogeneous environment. However, such a conclusion does not distinguish between whether phenotypic distributions are a passive or an adaptive developmental response to the environment. The fundamental difference between the two is that the former implies that a proportion of individuals have developed a suboptimal phenotype and thus have decreased fitness (as evidenced by behavioural ecological studies performed in a single context; reviewed in Andersson 1994; Blanckenhorn 2005). In contrast, the latter implies the various phenotypes within a distribution are adaptive as each is a response to a specific competitive challenge found within a variable environment (i.e., developmental plasticity; Doughty & Reznick 2004). Understanding the relative contribution of each of these processes in maintaining phenotypic variation is important as it helps us to understand phenotype-fitness associations and the evolution of phenotypes.

In my thesis, I am particularly interested in understanding the conditions that lead to the evolution of adaptive developmental plasticity, and how well it can explain the observed distribution of animal phenotypes in nature. Examining the prevalence of adaptive developmental plasticity and the context in which it occurs can help expand our comprehension of factors underlying phenotypic distributions. Developmental plasticity is a subset of phenotypic plasticity that deals with an irreversible shift in ontogeny that can lead to the development of phenotypes closely matched to the competitive challenges encountered during later life stages (Dufty et al. 2002; Piersma & Drent 2003; Pigliucci 2001; West-

Eberhard 2003; West-Eberhard 2005). Developmental plasticity has been demonstrated in many different animal taxa (reviewed in Adler & Harvell 1990; Agrawal 2001; Bateson et al. 3

2004; Harvell 1990; Nylin & Gotthard 1998) in response to a number of biotic and abiotic variables (See Table 1.1 for examples). The ontogenetic shifts in response to environmental variation can result in the development of either alternative phenotypes (polyphenisms; qualitative variation), or continuous phenotypic variation within a unimodal distribution

(quantitative variation). In my thesis, I focus on quantitative variation as it this type of variation that is more likely to contribute to the maintenance of continuous phenotypic variation within animal populations.

Although developmental plasticity is common, the developed phenotype must result in an increase in fitness for the plasticity to be considered adaptive. However, few animal studies actually demonstrate adaptive plasticity mainly because studies rarely examine the fitness associations of the different phenotypes in various relevant environments (Doughty &

Reznick 2004; Pigliucci 2001). Without empirical evidence of increased fitness of the plastic phenotypes, the developmental plasticity observed in natural populations may simply reflect the inherent lability of developmental system under examination (i.e., the final phenotype developed may be a passive response to the environment, Doughty & Reznick 2004;

Gotthard & Nylin 1995; Meyers & Bull 2002; Pigliucci 2001). Doughty and Reznick (2004) suggest six criteria that must be met to label demonstrated plasticity as adaptive (Table 1.2).

Although Doughty and Reznick themselves state that some of the criteria are difficult to satisfy (criteria five and six), this list provides a good overview of the critical features of systems in which developmental plasticity has evolved as an adaptive response to a variable environment.

Developmental plasticity and environmental variation

Although developmental plasticity can occur in response to a variety of different environmental cues, the majority of examples of developmental plasticity in animal taxa are 4 in response to factors such as photoperiod, temperature, and resource availability (Table 1.1).

This is likely due to the relative ease of manipulating such variables. As these factors vary systematically throughout a breeding season, theory predicts that this variation is likely to lead to the evolution of a developmental response (Harvell 1990; Van Tienderen 1991), particularly in response to temporal variation (Moran 1992). However, even if environmental factors that affect plasticity do not vary systematically, a correlated cue that can reliably predict the approaching environmental change can trigger plasticity. This is the case in studies of predator presence, another factor that has been well studied in plasticity studies.

Although variation in predator presence and abundance can be inconsistent, both visual and chemical cues of predator presence have been known to trigger adaptive developmental responses in life-history and metric traits in a number of taxa (for reviews see Benard 2004;

Harvell 1990; Relyea 2003).

In contrast, relatively fewer studies have demonstrated developmental plasticity in response to variation in population demographics (e.g., Gage 1995; Rodd et al. 1997;

Stockley & Seal 2001; Tan et al. 2004). This is surprising as there is abundant evidence that population density and operational sex ratio can fluctuate throughout a breeding season. For example, population demographics can vary as a function of birth, death, and maturity rates

(Foellmer & Fairbairn 2004; Foellmer & Fairbairn 2005a; Maxwell 1998; Velez &

Brockmann 2006; Zimmermann & Spence 1992), resource availability (Gwynne et al. 1998;

Matter & Roland 2002; Wauters et al. 2001), immigration/emigration patterns (Clutton-

Brock et al. 1997; Matter 1997; Matter & Roland 2002; Maxwell 1998; Wauters & Dhondt

1993), timing of arrival to breeding grounds (Grant et al. 1995; Wiklund & Fagerstrom

1997), predation rates (Gwynne 1987; Gwynne & Bussiere 2002; Norrdahl & Korpimaki

1998; Su & Li 2006), and differences in mating optima between the sexes (Amqvist & Rowe 5

2006; Parker 2006). Variation in these above factors is likely to have a strong effect on the strength and direction of selection as it can affect which sex is more choosy (Clutton-Brock et al. 1997; Kokko & Monaghan 2001; Simmons & Kvarnemo 2006), the availability of mates, and the number of competitors males encounter (Kokko & Rankin 2006; Matter

1997). Although such variation may not be systematic, developmental plasticity should still evolve as long as there are reliable cues of population fluctuations within breeding seasons.

However, examining effects of demographic variation (i.e., population density and sex ratio) on developmental plasticity is unlike examining effects of variation in other environmental factors, as plasticity may feedback to alter demography, as well as being triggered by demography (ecogenetic feedback; reviewed in Kokko & Lopez-Sepulcre 2007).

For example, individual ontogeny may shift in response to variation in population demographics that change the strength and direction of selection pressures within a breeding season. However, these ontogenetic shifts can cause changes in the age distribution or sex ratio of the population, thus altering population demographics, and therefore, the environmental context that individuals will encounter at maturity. Therefore, to truly understand how developmental plasticity functions in nature and whether it can help maintain broad phenotypic distributions requires an understanding of the complex dynamics between demography, ecology, and evolution (Benton et al. 2006; Doughty & Reznick 2004; Kokko

& Lopez-Sepulcre 2007; Metcalf & Pavard 2006). Additional empirical studies must therefore examine the prevalence of developmental plasticity, the cues and environmental conditions that trigger plasticity, and the fitness effects of plasticity in the various environments that are known to trigger it (Table 1.2).

In my thesis, I ask whether the phenotypic variation observed in nature is a result of adaptive developmental plasticity in response to population demography and selection in nature, rather than a passive response to environmental variation. I present data in support of four conditions (adapted from Doughty & Reznick 2004) that are necessary if developmental plasticity is adaptive: (I) Individuals must have reliable cues of the challenges they will experience at maturity, and these must be available prior to maturation (Scheiner 1993).

This is critical as it would allow the evolution of physiological mechanisms that reliably produce phenotypes that will perform well in the future (Dufty et al. 2002; Zera & Harshman

2001). (2) Individuals must alter their development in response to these reliable cues (i.e., developmental plasticity must be apparent). If individuals alter their development in response to variance in the environment, phenotype-environment matching may result, and this can increase individual fitness (DeWitt et al. 1998; Pigliucci 2001; Thompson 1991). (3) This developmental plasticity must be adaptive. If this is the case, then phenotypes developed within the test environment should outperform phenotypes reared in a alternate environment

(e.g.j reverse transplant experiment) (DeWitt et al. 1998; DeWitt 1998). (4) The environment must vary either spatially or temporally. A varying environment would result in a range of potential competitive challenges, and therefore, a range of potential phenotypes that could be optimal for each competitive challenge. If the environment, and therefore the competitive challenges vary continuously, then phenotypic distributions may arise from individual adaptive responses to the continuous challenges found in any given environment.

Plasticity in size, condition & development time in spiders

As individuals generally face limitations in resource abundance throughout their lifetime, it is impossible for individuals to optimize all their phenotypic traits, resulting in trade-offs in resource allocation between competing traits (Zera & Harshman 2001).

Understanding developmental trade-offs between size and time to maturity is critical as both of these traits can be essential to fitness (Berner & Blanckenhorn 2007). I focus on body size 7 and condition (metric traits) and time to maturity (age; life-history trait) in this thesis for several reasons: (1) because these traits are likely critical to male mating success but the importance of each may vary depending on the most common competitive context in nature,

(2) because of the observed variation in these traits, and (3) because trade-offs in development of these metric traits against development time is likely (Roff 1992; Stearns

1992). I also examined whether pheromonal cues can trigger plasticity in these traits, and the relationships of these traits to population demographics in three species of spiders.

I worked on the Western black widow {Latrodectus hesperus), native to the Western

United States and Canada, and two species native to Australia, the Australian redback spider

{Latrodectus hasselti) and the golden orb-web spider (Nephila plumipes). These three species have similarities in their mating systems and life-history traits that allowed the use of a combination of controlled laboratory and field approaches in the work described in this thesis. For example, all three species show extreme reversed size dimorphism (N. plumipes:

Elgar & Fahey 1996; L. hasselti: Forster & Kingsford 1983; L. hesperus: Kaston 1970), with earlier development of small males suggesting some past selection for protandry (e.g.,

Maklakov et aL 2004). However, reported clustering of males on webs of females suggests there might also be size-mediated direct competition among males (L. hasselti: Andrade

1996; N. plumipes: Elgar & Fahey 1996; L. hesperus: Kasumovic and Andrade pers. obs.).

In addition, while females are sedentary in each species, males must wander to search for their mates (Foelix 1982). Wandering may allow males to sample the biotic and abiotic environment, but may also impose strong selection on males due to increased risk of predation (Andrade 2003; Foellmer & Fairbairn 2005b; Prenter et al. 1998; Voilrath 1980;

Vollrath 1998). Like most spiders, males of these species are likely able to detect females and perhaps other conspecifics via airborne pheromones (Gaskett 2007), and I conjectured that 8 these pheromones would provide a mechanism by which developing males might detect population demographics such as density and sex ratio. Finally, there is considerable variation in male size and body condition at maturity in each of these species (L. hasselti:

Andrade 2003; N. plumipes: Kasumovic pers. obs; L. hesperus: Kasumovic and Andrade pers. obs.), along with variation in development time (N. plumipes: Herberstein, pers comm.,

L. hesperus and L. hasselti: Andrade, pers. comm.).

1. Availability of reliable cues of female abundance.

In spiders, males may be able to assess female reproductive status and density using airborne pheromones (Gaskett 2007). Successful pheromone-mediated mate searching may be a critical determinant of male reproductive success in many spiders as high mortality rates have been associated with this phase of their life (Andrade 2003; Vollrath 1980). However, pheromone-mediated mate searching has not been well-studied in spiders. Currently, there are very few field studies of male attraction to airborne pheromones (see Gaskett 2007 for a review) and only three published studies of mortality during mate searching by male spiders

(Andrade 2003; Foellmer & Fairbairn 2005b; Vollrath 1980). Moreover, although mortality during mate searching can have significant effects on male reproductive strategies on a population level (Kokko & Wong 2007), it is poorly studied in general across taxa, perhaps because of the difficulty in examining survival during mate searching.

In Chapters Two and Three of my thesis, I examine male mate searching in two different species of spiders (L. hesperus andiV. plumipes, respectively) to determine whether male spiders use long distance pheromones when searching for mates (most species of spiders have poor vision, Foelix 1982). I examine whether searching males can distinguish conspecific over heterospecific females (L. hesperus Chapter 2), if they can distinguish mature from juvenile females, female mating status, and if they base their choice on their 9 own phenotype (N. plumipes, Chapter 3). Finally, I document mate-searching mortality rates and examine whether male size or condition are associated with mate searching success (TV. plumipes, Chapter 3). This work makes a significant contribution to the small but developing literature on the mortality rates associated with mate searching. It also suggests that juvenile male spiders could use long-distance pheromones as reliable cues of the density of available females in the environment.

2. Developmental plasticity occurs in response to these reliable cues.

The availability of receptive females is likely to vary throughout the breeding season as new females mature, and older females become mated, which can alter population density and operational sex ratio. These fluctuating demographics can change the phenotypic traits associated with male success. Given that developing males likely have reliable cues of the availability of virgin females present in the environment, I hypothesized that they might alter their ontogeny to develop the phenotype that best matches the competitive challenges they will face at maturity. In Chapter Four, I examined whether penultimate instar (one moult from maturity) male redback spiders (L. hasselti) altered their development in response to cues of female density. I reared penultimate males in two different competitive environments

(female presence: mimicking a high density female environment; female absence: mimicking a low density female environment) and three different diet treatments. Males were also reared in the presence of a variable number of rival males, which might also be detected by pheromones (Gaskett 2007). I examined male development time, body size, and body condition at maturity as these traits can affect male mating success. This laboratory study was designed to determine whether male development is plastic and whether any developmental trade-offs are resource dependent. There are several studies that examine the effect of density on plasticity of quantitative traits (in size and sperm production: e.g., Gage 1995; Stockley & 10

Seal 2001; Tan et al. 2004). Examinations of the affect of sex ratio on plasticity of

quantitative traits are fewer as studies have only indirectly examined the effects of male and

female conspecifics on the same traits (e.g., Walling et al. 2007). However, there is evidence

in other taxa that the adult sex ratio can affect developmental pathways as female fish are

known to alter their sex in response to the adult sex ratio and social conditions (Benton &

Berlinsky 2006; Manabe et al. 2007). Being able to distinguish between simultaneous effects

of male and female conspecifics may be critical if plasticity is triggered by the intensity of

sexual selection (rather than sex-independent resource limitation).

3. Developmental plasticity must be adaptive.

As redback males matured rapidly and at significantly smaller sizes when females

were present than when females were absent (Chapter Four), I proposed that small male size

might offer a competitive advantage when female density is high. In Chapter Five, I

examined the fitness benefits associated with small male size in two different competitive

contexts to determine whether the demonstrated developmental plasticity is adaptive. I

established a grid of females on their webs in field enclosures to mimic the natural

distribution of potential mates in the field. I then released small and large males in these

enclosures and tracked their success at finding and copulating with females. In the

simultaneous release treatment, small and large males competed directly for access to virgin females. In the staggered release treatment, small males were released and allowed to compete for females one day before large males to mimic the competitive advantage they would gain due to their faster development (see Chapter Four). Both of these competitive contexts are likely to occur in nature as a result of the continuous variation in female availability and density of males on female's webs (Andrade 1996; Andrade 2003) that is likely to trigger developmental shifts (Chapter Four). 11

4. The environment must vary in competitive challenges.

Examining how the environment may change within a single breeding season will provide information on the range of competitive challenges an individual can encounter during its reproductive life span. If competitive challenges vary either spatially or temporally in a continuous manner due to variation in population demography, then there is likely variance in selection and several phenotypic optima may exist. Few studies have examined variation in selection pressures within a single breeding season in animals (see Jann et al.

2000 for an example). Understanding how competitive challenges vary within a single season would help explain the evolution and maintenance of variation in metric and life-history traits.

In Chapter Six, I sample two different populations of N. plumipes that differ in population density to examine how population density and the operational sex ratio vary throughout the breeding season at different spatial scales. These demographic factors are likely to have strong effects on the intensity of sexual and natural selection on mate- searching males as they alter population densities and sex ratios (Kokko & Monaghan 2001;

Kokko & Rankin 2006). I also ask whether the strength and direction of selection fluctuates between sampling periods using selection gradients. This study adds to the handful of studies that have examined variation in selection within a breeding season in animals (Jann et al.

2000; Punzalan 2007).

SUMMARY

In my thesis, I examined whether populations fluctuate in density and sex ratio within a season as these factors can potentially alter the competitive context, and therefore, selection for traits critical to mating success. I asked whether individuals demonstrate developmental 12 plasticity in response to demographic variation, and then asked if that plasticity was adaptive.

Understanding how population demographies and selection fluctuate will provide a better understanding of the types of challenges males will face throughout a single breeding season, and how quickly these challenges could potentially change, affecting fitness. This will provide critical information on factors that could lead to the evolution of developmental plasticity and help link together demography, mating system evolution, behavioural ecology, and life-history theory (Benton et al. 2006; Kokko & Lopez-Sepulcre 2007; Metcalf &

Pavard 2006). 13

LITERATURE CITED

Adler, F. R. & Harvell, C. D. 1990 Inducible Defenses, Phenotypic Variability and Biotic

Environments. Trends in Ecology & Evolution 5, 407-410.

Agrawal, A. A. 2001 Ecology - Phenotypic plasticity in the interactions and evolution of

species. Science 294, 321-326.

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

Andrade, M. C. B. 1996 Sexual selection for male sacrifice in the Australian redback spider.

Science 111, 70-72.

Andrade, M. C. B. 2003 Risky mate search and male self-sacrifice in redback spiders.

Behavioral Ecology 14, 531-538.

Arnqvist, G. & Rowe, L. 2006 Sexual conflict. Princeton: Princeton University Press.

Bateson, P., Barker, D., Clutton-Brock, T., Debal, D., D'Uldine, B., Foley, R., Gluckman, P.,

Godfrey, K., Kirkwood, T., Lahr, M., McNamara, J., Metcalfe, N., Monaghan, P.,

Spencer, H. & Sultan, S. 2004 Developmental plasticity and human health. Nature

430,419-421.

Benard, M. F. 2004 Predator-induced phenotypic plasticity in organisms with complex life

histories. Annual Review of Ecology Evolution and Systematics 35, 651-673.

Benton, C. B. & Berlinsky, D. L. 2006 Induced sex change in black sea bass. Journal of Fish

Biology 69, 1491-1503.

Benton, T. G. & Beckerman, A. P. 2005 Population dynamics in a noisy world: lessons from

a mite experimental system. Advances in Ecological Research 37, 143-181.

Benton, T. G., Plaistow, S. & Coulson, T. M. 2006 Complex population dynamics and

complex causation: devils, details and demography. Proc. R. Soc. Lond. B 273,1173-

1181. 14

Bemer, D. & Blanckenhorn, W. U. 2007 An ontogenetic perspective on the relationship

between age and size at maturity. Functional Ecology 21, 505-512.

Blanckenhorn, W. U. 1998 Adaptive phenotypic plasticity in growth, development, and body

size in the yellow dung fly. Evolution 52, 1394-1407.

Blanckenhorn, W. U. 2005 Behavioral causes and consequences of sexual size dimorphism.

Ethology 111, 977-1016.

Clutton-Brock, T. H., Rose, K. E. & Guinness, F. E. 1997 Density-related changes in sexual

selection in red deer. Proceedings of the Royal Society B-Biological Sciences 264,

1509-1516.

Dahl, J. & Peckarsky, B. L. 2003 Developmental responses to predation risk in

morphologically defended mayflies. Oecologia 137, 188-194.

David, J. R., Legout, H. & Moreteau, B. 2006 Phenotypic plasticity of body size in a

temperate population of Drosophila melanogaster: when the temperature-size rule

does not apply. Journal of Genetics 85, 9-23.

De Neve, L., Soler, J. J., Perez-Contreras, T. & Soler, M. 2004 Genetic, environmental and

maternal effects on magpie nestling-fitness traits under different nutritional

conditions: a new experimental approach. Evolutionary Ecology Research 6, 415-431.

DeWitt, T., Sih, A. & Wilson, D. 1998 Costs and limits of phenotypic plasticity. Trends Ecol.

Evol. 13,77-81.

DeWitt, T. J. 1998 Costs and limits of phenotypic plasticity: Tests with predator-induced

morphology and life history in a freshwater snail. Journal of Evolutionary Biology 11,

465-480.

Dixon, A. F. G. & Kundu, R. 1998 Resource tracking in aphids: programmed reproductive

strategies anticipate seasonal trends in habitat quality. Oecologia 114, 73-78. 15

Doughty, P. & Reznick, D. N. 2004 Patterns and analysis of adaptive plasticity in animals. In

Phenotypic Plasticity: Functional and Conceptual Approaches (ed. T. J. DeWitt & S.

M. Scheiner), pp. 126-150. New York: Oxford University Press.

Dufty, A. M., Clobert, J. & Moller, A. P. 2002 Hormones, developmental plasticity and

adaptation. Trends in Ecology & Evolution 17, 190-196.

Elgar, M. A. & Fahey, B. F. 1996 Sexual cannibalism, competition, and size dimorphism in

the orb-weaving spider Nephilaplumipes Latreille (Araneae: Araneoidea). Behavioral

Ecology7,195-198.

Ernande, B., Boudry, P., Clobert, J. & Haure, J. 2004 Plasticity in resource allocation based

life history traits in the Pacific oyster, Crassostrea gigas. I. Spatial variation in food

abundance. Journal of Evolutionary Biology 17, 342-356.

Foelix, R. 1982 Biology of spiders. Cambridge: Harvard University Press.

Foellmer, M. W. & Fairbairn, D. J. 2004 Males under attack: sexual cannibalism and its

consequences for male morphology and behaviour in an orb-weaving spider.

Evolutionary Ecology Research 6, 163-181.

Foellmer, M. W. & Fairbairn, D. J. 2005a Competing dwarf males: sexual selection in an

orb-weaving spider. Journal of Evolutionary Biology 18, 629-641.

Foellmer, M. W. & Fairbairn, D. J. 2005b Selection on male size, leg length and condition

during mate search in a sexually highly dimorphic orb-weaving spider. Oecologia

142, 653-662.

Forster, L. M. & Kingsford, S. 1983 A preliminary study of development in two Latrodectus

species (Aranea: Theridiidae). N. Z. Entomol. 7,431-438. 16

Gage, M. J. G. 1995 Continuous variation in reproductive strategy as an adaptive response to

population density in the moth Plodia interpimctella. Proc. R. Soc. Lond. B 261, 25-

30.

Gaskett, A. C. 2007 Spider sex pheromones: emission, reception, structures, and functions.

Biological Reviews 82, 26-48.

Gotthard, K. & Nylin, S. 1995 Adaptive plasticity and plasticity as an adaptation - a selective

review of plasticiy in animal morphology and life-history. Oikos 74, 3-17.

Grant, J. W. A., Bryant, M. J. & Soos, C. E. 1995 Operational sex ratio, mediated by

snychrony of female arrival, alters variance of male mating success in Japanese

medaka. Artim. Behav. 49, 367-375.

Gwynne, D. T. 1987 Sex-biased predation and the risky mate-locating behaviour of male

tick-tock cicadas (Homoptera: Cicadidae). Animal Behaviour 35, 571-576.

Gwynne, D. T., Bailey, W. J. & Annells, A. 1998 The sex in short supply for matings varies

over small spatial scales in a katydid (Kawanaphila nartee, Orthoptera :

Tettigoniidae). Behavioral Ecology and Sociobiology 42, 157-162.

Gwynne, D. T. & Bussiere, L. 2002 Female mating swarms increase predation risk in a 'role-

reversed' dance fly (Diptera : Empididae : Rhamphomyia longicauda Loew).

Behaviour 139, 1425-1430.

Harvell, C. D. 1990 The Ecology and Evolution of Inducible Defenses. Quarterly Review of

Biology 65,323-340.

Jann, P., Blanckenhorn, W. U. & Ward, P. I. 2000 Temporal and microspatial variation in the

intensities of natural and sexual selection in the yellow dung fly Scathophaga

stercoraria. Journal of Evolutionary Biology 13, 927-938. 17

Johansson, F., Stoks, R., Rowe, L. & De Block, M. 2001 Life history plasticity in a

damselfty: Effects of combined time and biotic constraints. Ecology 82,1857-1869.

Johnson, J. B. & Belk, M. C. 2001 Predation environment predicts divergent life-history

phenotypes among populations of the livebearing fish Brachyrhaphis rhabdophora.

Oecologia 126, 142-149.

Kaston, B. J. 1970 Comparative biology of american black widow spiders. Transactions of

the San Diego Society of Natural History 16, 33-82.

Kause, A. & Morin, J. P. 2001 Seasonality and genetic architecture of development time and

body size of the birch feeding sawfly Priophorus pallipes. Genetical Research 78, 31-

40.

Kingsolver, J. G. 2007 Variation in growth and instar number in field and laboratory

Manduca sexta. Proceedings of the Royal Society B-Biological Sciences 214,977-

981.

Kokko, H. & Lopez-Sepulcre, A. 2007 The ecogenetic link between demography and

evolution: can we bridge the gap between theory and data? Ecology Letters In press.

Kokko, H. & Monaghan, P. 2001 Predicting the direction of sexual selection. Ecology Letters

4, 159-165.

Kokko, H. 8c Rankin, D. J. 2006 Lonely hearts or sex in the city? Density dependent effects

in mating systems. Phil Trans. R. Soc. B 361, 319-334.

Kokko, H. & Wong, B. B. M. 2007 What determines sex roles in mate searching ? Evolution

61,1162-1175.

Maklakov, A. A., Bilde, T. & Lubin, Y. 2004 Sexual selection for increased male body size

and protandry in a spider. Animal Behaviour 68,1041-1048. 18

Manabe, H., Ishimura, M., Shinomiya, A. & Sunobe, T. 2007 Inter-group movement of

females of the polygynous gobiid fish Trimma okinawae in relation to timing of

protogynous sex change. Journal of Ethology 25, 133-137.

Matter, S. F. 1997 Population density and area: The role of between- and within-patch

processes. Oecologia 110, 533-538.

Matter, S. F. & Roland, J. 2002 An experimental examination of the effects of habitat quality

on the dispersal and local abundance of the butterfly Parnassius smintheus.

Ecological Entomology 27, 308-316.

Maxwell, M. R. 1998 Seasonal adult sex ratio shift in the praying mantid Iris oratoria

(Mantodea : Mantidae). Environmental Entomology 27, 318-323.

Metcalf, C. J. E. & Pavard, S. 2006 Why evolutionary biologists should be demographers.

Trends Ecol Evol 22, 205-212.

Meyers, L. A. & Bull, J. J. 2002 Fighting change with change: adaptive variation in an

uncertain world. Trends in Ecology & Evolution 17, 551-557.

Moran, N. A. 1992 The Evolutionary Maintenance of Alternative Phenotypes. American

Naturalist 139, 971-989.

Norrdahl, K. & Korpimaki, E. 1998 Does mobility or sex of voles affect risk of predation by

mammalian predators? Ecology 79, 226-232.

Nylin, S. & Gotthard, K. 1998 Plasticity in life-history traits. 43, 63-83.

Parker, G. A. 2006 Sexual conflict over mating and fertilization: an overview. Phil. Trans. R.

Soc. B 361, 235-259.

Piersma, T. & Drent, J. 2003 Phenotypic flexibility and the evolution of organismal design.

Trends in Ecology & Evolution 18,228-233. 19

Pigliucci, M. 2001 Phenotypic plasticity: beyond nature and nurture John Hopkins

University Press.

Prenter, J., Elwood, R. W. & Montgomery, W. 1.1998 No association between sexual size

dimorphism and life histories in spiders. Proceedings of the Royal Society Biological

Sciences Series B 265, 57-62.

Punzalan, D. 2007 Evolution of sexual colour pattern dimorphism in the ambush bug

Phymata americana. Ecology Evolution and Behaviour, University of Toronto.

Relyea, R. A. 2003 How prey respond to combined predators: A review and an empirical test.

Ecology 84, 1827-1839.

Rodd, F. H., Reznick, D. N. & Sokolowski, M. B. 1997 Phenotypic plasticity in the life

history traits of guppies: Responses to social environment. Ecology 78, 419-433.

Roff, D. A. 1992 The evolution of life histories: Theory and analysis. New York: Chapman &

Hall.

Scheiner, S. M. 1993 Genetics and evolution of phenotypic plasticity. Annual Review of

Ecology andSystematics 24, 35-68.

Simmons, L. W. & Kvarnemo, C. 2006 Costs of breeding and their effects on the direction of

sexual selection. Proceedings of the Royal Society of London Series B Biological

Sciences 273, 465-470.

Stearns, S. C. 1992 The evolution of life histories. Oxford: Oxford University Press.

Stockley, P. & Seal, N. J. 2001 Plasticity in reproductive effort of male dung flies

(Scatophaga stercoraria) as a response to larval density. Functional Ecology 15, 96-

102.

Su, K. & Li, D. Q. 2006 Female-biased predation risk and its differential effect on the male

and female courtship behaviour of jumping spiders. Anim. Behav. 71, 531-537. 20

Tan, G. N., Govedich, F. R. & Burd, M. 2004 Social group size, potential sperm competition

and reproductive investment in a hermaphroditic leech, Helobdella papillornata

(Euhirudinea: Glossiphoniidae). Journal of Evolutionary Biology 17, 574-580.

Thompson, J. D. 1991 Phenotypic plasticity as a component of evolutionary change. Trends

in Ecology & Evolution 6, 246-249.

Van Buskirk, J. & Relyea, R. A. 1998 Selection for phenotypic plasticity in Rana sylvatica

tadpoles. BiologicalJournal of the Linnean Society 65, 301-328.

Van Tienderen, P. H. 1991 Evolution of generalists and specialists in spatially heterogeneous

environments. Evolution 45, 1317-1331.

Velez, M. J. & Brockmann, H. J. 2006 Seasonal variation in selection on male calling song in

the field cricket, Gryllus rubens. Animal Behaviour 72, 439-448.

Vences, M., Puente, M., Nieto, S. & Vieites, D. R. 2002 Phenotypic plasticity of anuran

larvae: environmental variables influence body shape and oral morphology in Rana

temporaria tadpoles. Journal of Zoology 257,155-162.

Vollrath, F. 1980 Male body size and fitness in the web-building spider Nephila clavipes.

Zeitschrift Tierpsychologie 53, 61-78.

Vollrath, F. 1998 Dwarf males. Trends in Ecology & Evolution 13,159-163.

Walling, C. A., Royle, N. J., Metcalfe, N. B. & Linstrom, J. 2007 Green swordtails alter their

age of maturity in response to the population level of male ornamentation. Biology

Letters 3, 144-146.

Wauters, L. & Dhondt, A. A. 1993 Immigration pattern and success in red squirrels.

Behavioral Ecology and Sociobiology 33,159-167. 21

Wauters, L. A., Gurnell, J., Preatoni, D. & Tosi, G. 2001 Effects of spatial variation in food

availability on spacing behaviour and demography of Eurasian red squirrels.

Ecography 24, 525-538.

West-Eberhard, M. J. 2003 Developmental plasticity and evolution. New York: Oxford

University Press.

West-Eberhard, M. J. 2005 Developmental plasticity and the origin of species differences.

Proceedings of the National Academy of Sciences 102, 6543-6549.

Wiklund, C. & Fagerstrom, T. 1997 Why do males emerge before females? A hypothesis to

explain the incidence of protandry in butterflies. Oecologia 31, 153-158.

Zera, A. J. & Harshman, L. G. 2001 The physiology of life history trade-offs in animals.

Annual Review of Ecology and Systematics 32, 95-126.

Zimmermann, M. & Spence, J. R. 1992 Adult-population dynamics and reproductive effort

of the fishing spider Dolomedes triton (Araneae, Pisuridae) in Central Alberta. Can.

J. Zool. 70,2224-2233. 22

Table 1.1: Examples of phenotypic plasticity in response to different environmental variables.

Factor Species Trait(s) Reference

Temperature* Tobacco hornworm Larval instar (Kingsolver 2007)

(Manduca sexto) number

Fruit fly (Drosophila Thorax length, (David et al.

melanogaster) wing size 2006)

Frog {Rana temporarid) Growth rate, tail (Vences et al.

depth, body width 2002)

Time of Season Yellow dung flies Size at maturity, (Blanckenhorn

(i.e., photoperiod) (Scatophaga time to maturity 1998)

stercoraria)

Damselfly (Lestes Development time (Johansson et al.

sponsa) and size at maturity 2001)

Willow-carrot aphid Body size, gonad (Dixon & Kundu

(Cavariella aegopodii) size, lipid reserves 1998)

Food availability* Oyster (Crassostrea Growth, survival (Ernande et al.

gigas) 2004)

Mocking birds (Pica Tarsus length, body (De Neve et al.

pica) mass, immune 2004)

response

Birch feeding sawfly Larval mass, (Kause & Morin

(Priophorus pallipes) development time 2001) 23

Predator presence Mayfly (Ephemerella Development Time (Dahl &

invaria) and size at maturity Peckarsky 2003)

Wood frog {Rana Body size and tail (Van Buskirk &

sylvatica) depth Relyea 1998)

Fish {Brachyrhaphis Body Size (Johnson & Belk

rhabdophord) 2001)

Density Moth (Plodia Sperm number, (Gage 1995)

interpunctelld) size and

development time

Yellow dung flies Size and (Stockley & Seal

(Scatophaga development time 2001)

stercoraria)

Leech (Helobdella Testis volume (Tan et al. 2004)

papillornata)

Population Soil mites (Sancassania Development time (Benton &

structure berlesei) Beckerman 2005)

Green swordtails development time (Walling et al.

(Xiphophorus helleri) and ornament size 2007)

* Plasticity in responis e to both temperature anc food availability are rnor e likely passive responses to variation rather than examples of adaptive plasticity. 24 Table 1.2: Six criteria that should be met for plasticity to be considered adaptive (adapted from

Doughty and Reznick 2004).

1. Different phenotypes must be produced in various environments.

2. Understanding the nature of the environmental heterogeneity that has the potential to

trigger developmental plasticity.

3. The various phenotypes must exhibit differences in fitness in different environments.

4. A reliable cue that predicts the future selective environment must be identified.

5. There must be genetic variation for plasticity and plasticity must evolve in response to

selection.

There must be comparative evidence that plasticity is correlated with environmental

heterogeneity. 25

CHAPTER TWO

Discrimination of airborne pheromones by mate-searching western black widow males

(Latrodectus hesperus): species- and population-specific responses

Michael M. Kasumovic and Maydianne C. B. Andrade

Integrative Behaviour and Neuroscience Group, University of Toronto at Scarborough, Toronto,

ON, MIC 1A4, Canada

Published in Canadian Journal of Zoology, 2004, 82: 1027-1034, reprinted by permission from

NRC Press 26

ABSTRACT

Males of many web-building spiders abandon their webs at maturity to search for a potential mate. Since wandering can be very risky, and females are often widely distributed, males should use any cues that might ensure rapid and accurate location of conspecific females.

Although it has long been assumed that mate-searching male spiders locate females using species-specific airborne pheromones released from webs, few studies have experimentally examined this phenomenon in the field. We show that western black widow spider males

{Latrodectus hesperus Chamberlin and Ivie, 1935) are attracted to females' webs by an airborne cue released from the web, can distinguish between conspecific and heterospecific females, and can discriminate between webs produced by conspecific females from different geographical populations. The latter result demonstrates a partial pre-mating block to fertilization between populations at the edges of the species range. Complementary inter-population laboratory matings suggest there may also be a post-mating block to fertilization as these copulations did not result in viable offspring. This study provides experimental field evidence of male attraction by airborne pheromones released from females' webs, shows the potential importance of these pheromones in species discrimination in black widow spiders, and suggests that northern and southern populations of L. hesperus may be incipient biological species. INTRODUCTION

Accurate species recognition is an important component of mate identification since it prevents hybridization with genetically incompatible heterospecifics. Recognition involves the use of species-specific cues during mate attraction or courtship—these cues may be visual, acoustic, tactile or olfactory in different groups. Species identification in spiders is particularly interesting because they utilize several different modalities to convey this information, such as complex visual displays, acoustic displays, and/or vibrational signalling (for a review see Krafft

1982). Spiders also use chemical signals—presumed to be one of the least evolutionarily derived forms of communication—in mating. The chemical composition of pheromones bound to female webs has recently been identified (Schulz and Toft 1993; Papke et al. 2000), and behavioural studies have revealed the importance of these contact pheromones in the recognition of conspecifics and the initiation of male courtship behaviours (Trabalon et al. 1997; Papke et al. 2001; Tichy et al. 2001). Contact pheromones can also provide male spiders with information about the sex, sexual maturity, and the mating status of females (Riechert and Singer 1995;

Searcy et al. 1999; Papke et al. 2001). In some of these groups, however, contact pheromones trigger courtship behaviour in heterospecific as well as conspecific males, suggesting these alone are not likely to be useful in preventing costly mating mistakes (e.g. Ross and Smith

1979). These studies may omit a necessary step in species recognition—discrimination by mate- searching males prior to web contact.

Contact pheromones require close proximity between individuals before identification can occur and thus cannot be used to attract males over long distances. This may be problematic for male web-building spiders which must often travel through dangerous terrain to reach the web of a potential mate (e.g. Henschel 2002) and experience high mortality during this time

(e.g. Andrade 2003; Vollrath and Parker 1992). The risk of mortality during mate searching should impose strong selection on males to use any available cues that might facilitate rapid and 28 accurate identification of potential mates at a distance. Selection may also favour the production of airborne pheromones by sedentary females, particularly if webs are widely distributed and encounter rates with males are low. Thus, airborne pheromones may be an important, but little- studied, aspect of pre-copulatory communication for arachnids (Pollard et al. 1987).

A few studies have examined the importance of airborne pheromones in spiders—these have demonstrated that airborne chemicals released from females' bodies attract males and can trigger courtship (Tietjen and Rovner 1982; Searcy et al. 1999; Papke et al. 2001). Although this may be sufficient to attract males, in web-building spiders the web itself (with its relatively large surface area) could provide a more efficient means for releasing attractive airborne chemicals.

Studies of black widow spiders {Latrodectus spp.) suggest that females incorporate pheromones into their webs to facilitate web location by males (Ross and Smith 1979; Anava and Lubin

1993), but these studies included direct male contact with web-bound chemicals. To date, the majority of studies examining the efficacy of chemical signals in spiders have experimentally placed males directly on webs, examined the influence of chemicals released by the female's body, or allowed direct contact between males and females. Thus, there is little experimental evidence that airborne pheromones released from webs attract males.

In this field study, we demonstrate that western black widow males {Latrodectus hesperus Chamberlin and Ivie, 1935) are attracted to females' webs by airborne pheromones, and discriminate conspecific female webs from those of heterospecific females (L. hasselti) using these airborne cues. We placed empty webs constructed by virgin L. hesperus and L. hasselti Thorell, 1870 females at a location where L. hesperus spiders were common and monitored the arrival of mate-searching males at these webs. We included webs built by females from two geographically separated populations of L. hesperus (British Columbia, Canada and

Arizona, USA) and show that males not only discriminate conspecific from heterospecific females, but are also disproportionately attracted to webs built by females from their own 29 population. This partial pre-mating block to fertilization suggests these populations may be

incipient species—a hypothesis supported by complementary laboratory pairings showing

matings across these populations are less likely to result in viable offspring than are matings

within a population.

MATERIALS AND METHODS

Study Site and Study Species.- This study was completed between May and June 2003 just outside Lac Du Bois Provincial Park in Kamloops, British Columbia (50°43 '20.47" N,

120°24'15.27 W). Female L. hesperus commonly inhabit desert habitat and, at our field site,

were most often found on the side of rocky outcrops and hillsides inside small cavities in the

rocks or soil. Females build refugia within these gaps and cob webs on the rock face outside of

their refugia for prey capture. Our study site consisted of two large rock outcrops, one

approximately 50 m below the first. The first outcrop was 145 m long, covered an area of 860

m2 , and contained 74 female's webs, while the second outcrop was 170 m, long covered an area

of 1030 m2, and contained 139 female's webs (Figure 2.1 A ,B). Webs were initially identified

by their distinct structure (e.g., Szlep 1965), we later confirmed these were L. hesperus webs

because all contained females and/or egg sacs. Webs were mapped by measuring the distance

from each web to two points of known location using a laser distance-meter (Leica Disto Lite,

Leica Geosystems). X and Y co-ordinates for webs were input to Arc View 3.2a (Environmental

Systems Research Institute Inc., 2000) where we calculated nearest-neighbour distances using

the Nearest Features v. 3.7a extension (Jenness Enterprises, 2004), and then plotted a histogram

depicting the distribution of nearest-neighbour distances using Systat 10.2 (SPSS, 2002). The

distribution of webs varied throughout both outcrops, but it was common to find refugia next to

one another with the edges of cob webs overlapping. Also, we had an instance of two females

retreating into the same refuge when disturbed—suggesting that some females may share webs. 30 Although the mating status of wild females was unknown, some webs containing egg sacs were adjacent to webs without egg sacs—suggesting the distribution of female mating status also varied spatially. The mean nearest-neighbour distance for webs was 0.50m ± 0.36 SD for the first site, and 0.72m ± 0.50 SD for the second site (Figure 2.1 C, D).

Latrodectus hesperus females are most active between dusk and dawn, when they leave the refuge to settle on the outer web. During this time, females clean, repair, and expand the web—most likely refreshing any existing web-borne chemicals. While females inhabit a web throughout their lives, males abandon their juvenile webs at sexual maturity and wander in search of a potential mate. This wandering phase is likely to result in high mortality for males

(e.g. Gwynne 1987; Andrade 2003).

Female genitalia consist of paired, independent sperm storage organs (spermathecae), each of which is inseminated via a separate opening and long, coiled duct. Males have corresponding coiled structures (emboli), located on paired anterior appendages (the pedipalps).

Each embolus is inserted in one spermatheca via the coiled duct at copulation. Recent work on congeners (L. hasselti, L. revivensis) suggests that the male's embolus tip must reach inside the spermatheca for insemination to be successful (Berendonck and Graven 2000; Snow 2003).

Field Experiment.- We examined whether or not males of the western black widow spider (L. hesperus) are differentially attracted to webs of conspecific females by placing empty webs in the field in a randomized block design that included 4 treatments: a control, webs from northern and southern populations of L. hesperus, and webs from heterospecific L. hasselti.

We used webs built by L. hesperus females descended from spiders collected from

British Columbia, Canada (n = 15) and from Arizona, USA (n =15) to determine whether males can distinguish between different populations of L. hesperus, and used webs built by L. hasselti females descended from spiders collected in Perth, Western Australia (n = 15) to determine whether males can distinguish between webs of different species. All females were from outbred 31 lines that had been reared in the laboratory for at least two generations. Females were used approximately six months after maturity, were virgins at the time of web construction, and had no prior direct contact with males. Each female was placed in a screen cage (10 x 10 x 7.5 cm) in the laboratory in Toronto and was given 17 days to build a web (females began web construction within 24 hours). Cages were shipped (overnight) to Kamloops, British Columbia where we removed the females and used the cages immediately. Before being placed in the field, each cage was surrounded on all sides by a 2.5 cm wide sticky strip cut from an insect glue trap.

We placed cages at 15 marked trap sites ('blocks'). Each trap site had four treatments: one web from each spider group (L hesperus British Columbia, L hesperus Arizona, L. hasselti) and one control. Trap locations were spread over both outcrops (eight covering a distance of 70 m in the first outcrop and seven covering a distance of 60 m in the second outcrop) and were in a row approximately in the centre of each outcrop, parallel to its long axis (deviation from ideal locations were due to variation in terrain and vegetation that made it impossible to place traps in some areas, Figure 1 A,B). At each trap site we marked a one metre square with a randomly determined orientation. A single cage (with glue trap frame) from each spider group was then randomly placed at each of three vertices. At the fourth vertex we placed an empty glue frame as a control. Our traps were placed near naturally-occurring female webs so inter-web distances were comparable to that found naturally on the outcrops (Figure 2.1 A, B), and the minimum distance between experimental cages (lm) was within one or two standard deviations of the mean nearest-neighbour distances found in natural webs in the second and first sites respectively

(Figure 2.1 C, D). Two of the traps had to be moved away from the centre of the outcrop during the study since they were damaged by yellow bellied marmots {Marmota flaviventris) on the first night. Males were attracted to both of these moved traps, so data from these traps are included in our analyses. 32 Each trap was checked daily at approximately 1100 hr PST for three days. Any males found on the glue strips or on the cages were collected. On the third day, all the cages and glue strips were collected at 1400 hr, and any other males found at this time were also collected. We then placed six empty cages (no webs) with sticky strips at six randomly chosen trap sites as an additional control and checked them over the following 3 days.

Laboratory Matings.- We conducted controlled laboratory matings to determine whether males could successfully mate with conspecifics from different populations.

Laboratory-reared L. hesperus males from lines established with spiders collected in British

Columbia were paired with lab-reared conspecific females from a population established with spiders collected in Arizona. Lab-reared Arizona males were also paired with lab-reared British

Columbia females (interpopulation matings). We also carried out control matings within each population (intrapopulation matings). Five matings were carried out in each group for a total of

20 matings.

Females were placed in clear Rubbermaid© containers (35 x 30 x 15 cm) on wood doweling and allowed to build a web for at least five days. Males and females were fed prior to trials to ensure variation in hunger level did not affect mating success. Trials began when a male was placed on a female's web, and were terminated after a maximum of eight hours, when a male had inserted both pedipalps (paired copulatory organs), or when a female knocked the courting male off the web. All mating trials were video recorded using Panasonic BP-330 cameras with Navitar Zoom 7000 zoom lenses.

For each trial we examined 1) male courtship behaviour as the presence or absence of six stereotyped courtship behaviours described by Ross and Smith (1979), 2) female receptivity as whether the female vibrated her abdomen (Ross and Smith 1979) and 3) male mating success 33 by examining the number of insertions. If the male mated successfully, we fed females twice a week for two months to determine whether females produced any viable egg sacs.

Finally, we examined male pedipalps to determine whether there was a difference between the two populations of L. hesperus in embolus length or coil number. Since male and female genitalic morphology are closely matched in Latrodectus (e.g. Kaston 1970; Berendonck and Graven 2000), we thought this might give insight into observed differences in fertilization success of males from the two populations (see results). We stretched the embolus away from the palp using fine forceps and took digital photographs of the embolus under a dissecting microscope. We then measured embolus length on digital photographs using Image Tools 3.0

(UTHSCSA, 2002).

Statistics.- We used a replicated goodness-of-fit test (G-statistic) to test whether there were differences between the number of males attracted to different cages since this test 1) pools data across days into a single value and 2) also treats each day as a replicate allowing the probabilities in each day to be combined into a single statistical test (Sokal and Rohlf 1995). We then used a chi-square test for post-hoc analyses between populations. For lab matings, we compared mating success and offspring production for inter vs. intrapopulation matings. We tested all data for normality and used non-parametric statistics where applicable. All values are expressed in mean ± SE.

RESULTS

Field Experiment.- A total of 47 males were caught during the three day study (Table

2.1). No L. hesperus males were attracted to empty cages or to the glue traps alone (controls), although the glue traps did catch a variety of other ground-walking invertebrates (e.g., crickets, ants, beetles, Lepidoptera larvae). We removed control cages (zeros) from the analysis to avoid biasing the goodness of fit test towards significance. More than 50% of L. hesperus males captured were on cages containing the web of females from the British Columbia L. hesperus 34 population (Table 2.1). Significantly more males were attracted to the cages of conspecific females from the same population than to cages with the web of Arizona females orL. hasselti females. This is true whether data is pooled across days or whether each day is used as a replicate (Table 2.2). In addition, twice as many males were collected from cages filled with web of heterospecific females (L. hasselti, 30% of males) than the conspecific females from the

Arizona population (15% of males) although this difference was not significant (%2 = 2.33, p =

0.127). Twenty-nine of the forty-seven captured males (62%) were found in, or on the cages, rather than on the sticky strips. The distribution of males attracted to the different spider groups was similar for males caught on sticky strips and those caught inside cages (Table 2.1). The relative attractiveness of the different treatment webs was similar each day (heterogeneity test, p

> 0.05). The total number of males caught each day decreased with time, although not significantly (%2 = 1.269, p = 0.53).

Lab matings.- The stereotyped behaviours outlined by Ross and Smith (1979) were performed by all males except for four (three of which are from interpopulation matings)—all of which led to unsuccessful matings (two of these males were knocked off the web and a third was killed). Eight often males from intrapopulation pairings and five often males from interpopulation pairings copulated. This difference was not statistically significant (Fisher's

Exact = 2.03, p = 0.16). In general, we found no behavioural differences between inter and intra- populational matings, but we had low power for our analyses because there were relatively few successful matings. For example, the number of copulations did not differ significantly between inter and intrapopulation matings (Fisher's Exact = 5.49, p = 0.06). We observed female vibrations in 8 of 20 trials although this behaviour did not occur more often in successful matings (Fisher's Exact = 0.60, p = 0.44) or in intrapopulation matings (Fisher's Exact = 0.84, p

= 0.36). However, intrapopulation matings were significantly more likely to result in viable egg 35 sacs than were interpopulation matings—fifty percent of females (4/8) mated to males from the same population produced egg sacs, while none of the females mated to males from a different population produced egg sacs (0/5) (one-tailed Fisher's exact test: p = 0.05).

Pedipalp structure.- Laboratory reared males from the Arizona population had significantly longer emboli than laboratory reared males from the British Columbia population

(4.65 ± 0.11 mm and 3.89 ±0.11 mm, respectively; pooled t = 4.79, df = 22, p < 0.001, Figure

2.2). It is unclear whether these longer emboli had more coils as small sample size results in low power for this test (Arizona: 3.75 ± 0.13 coils; British Columbia: 3.42 ± 0.12 coils; pooled t =

1.876, df = 22, p = 0.07). Arizona males were significantly larger than males from British

Columbia (Figure 2.1; pooled t = 2.428, df = 19, p = 0.03, Figure 2.2), which suggests that differences in embolus length could arise through allometry. However, we were unable to detect a relationship between male embolus length and body size within each population, perhaps because of our relatively small sample size.

DISCUSSION

Our results strongly suggest that L. hesperus males use airborne pheromones released from webs to locate conspecific female webs in nature. No black widow males were found on control cages although other invertebrates were captured on these glue traps. There were no visual or tactile differences among web-filled cages, and females were removed so males could not get cues from them directly. Nevertheless, large numbers of males were attracted to web- filled cages (Table 2.1). Moreover, males were found disproportionately on webs built by conspecific females from the same geographic population, suggesting airborne pheromones provide males with information they can use to discriminate among webs of even closely related females. Since all females used in this study had been laboratory reared on a common diet, this effect cannot be due to environmental or diet-based effects on the chemical content of the web. 36 Rather, this is likely to reflect heritable variation in the composition of pheromone produced by female L. hesperus as a function of species and population of origin.

Since about half of the males were caught on or in the cages (Table 2.1), it is possible that some males entered cages without getting stuck on sticky traps, were able to sample female webs directly, then left the traps again. Some of these males would likely get stuck on exit, but some might exit successfully. If this were the case, variation in attraction might be partly based on assessment of web-bound rather than airborne pheromones. However, several lines of evidence support the conclusion that the distribution of males across female webs is due to attraction by airborne pheromones. First, all males found on sticky traps were oriented toward the cages, suggesting they were trapped when first approaching the cage—no males were found in the opposite orientation, as would be expected if they entered cages initially then got stuck on exiting. Second, the distribution of males attracted across the spider groups was similar whether we pooled all data, examined only males caught on the sticky strips, or examined only those found inside cages (Table 2.1). Third, males of the closely related L. hasselti remain on the webs of virgin females for days, even if there is no female present, so it is likely that males would arrive on a female's web and remain there for the few hours before we collected them (Andrade, unpublished).

Our study adds to the few studies demonstrating the importance of airborne pheromones in mate attraction in spiders (see Searcy et al. 1999; Papke et al. 2001). First, males of some species cut and cover portions of the web of a potential mate as soon as they arrive on the web, and this apparently limits the release of attractive volatiles that might attract rivals (Watson

1986; Schuitz and Toft 1993). Second, some field studies report differential attraction of male spiders to the webs of virgin compared to previously mated or immature females, suggesting the importance of cues that are effective at a distance (Anava and Lubin 1993; Robinson and

Robinson 1980; Pollard et al. 1987). However, in the latter observational studies, females were 37 present in webs or males were in direct contact with webs so it was unclear whether these effects were due to differential male attraction to the web or variation in male departure from webs after exposure to contact pheromones or females.

Previous studies of Latrodectus spiders examined the effect of female contact pheromones on male courtship behaviour and showed that males initiate courtship on the webs of conspecific and heterospecific females (e.g. L. rivivensis, L. hesperus, L. mactans; Ross and

Smith 1979, Anava and Lubin 1993). Additionally, Kaston (1970) was able to successfully mate

L. hesperus with L. mactans, although only 3 of 27 attempts were successful, and no egg sacs were produced. In Kaston's study, the intensity of courtship was highest with conspecifics suggesting that contact pheromones may also contribute to genetic isolation of these species.

However, these studies also led to speculation that heterospecific copulations might be common in nature. Since matings with heterospecifics may be costly to males (aggression from heterospecific females [Kavale 1986] or no offspring produced [Kaston 1970] after risky mate searching) this led to questions about how males avoid these potentially costly interactions. Our study suggests that, even in areas where two Latrodectus species overlap (e.g., L. hesperus and

L. mactans in Texas, Levi 1959), most males could accurately locate the webs of conspecifics using airborne pheromones. Although we did not directly compare sympatric species, the discrimination of conspecific females from different populations suggests males are able to make fine-grained decisions using pheromones. If this is true, there may be relatively weak selection for males to discriminate females based on contact pheromones that elicit courtship behaviours—which would explain the observation of heterospecific courtship in these spiders

(Kaston 1970). However contact pheromones could serve as late protection from wasted courtship efforts since males do sometimes incorrectly identify females using long distance pheromones (e.g., Table 2.1). This function may be reflected in the variation in courtship intensity elicited by heterospecific versus conspecific contact pheromone (Kaston 1970). Further 38 field studies using different species whose ranges overlap (e.g., L. hesperus and L. mactans) would help determine whether this level of discrimination by males is common.

Perhaps the most interesting result of our study is the equal preference by males for webs built by heterospecific L. hasselti females and webs built by conspecific females from a different population. Pheromones are typically complex blends of chemicals (Prouvost et al.

1999) and female spiders are known to produce multiple pheromones that overlap between species (Trabalon et al. 1997). Thus male L. hesperus may be attracted to L. hasselti webs because of common chemical elements in the pheromones of the two species. Since there is no possibility of natural mis-matings between L. hasselti (Western Australia) and L. hesperus

(Western North America) there will not have been selection on males to discriminate cues produced by these females.

In comparison, although spiders from northern and southern populations of L. hesperus are also unlikely to interact directly, the species is distributed across western North America

(from British Columbia to southern Texas, Levi 1959), and because juveniles disperse by ballooning (Foelix 1982), it is possible that individuals from adjacent populations may interact throughout the north-south axis. Although results from our interpopulation matings should be treated with caution due to the small sample size, they suggest that, at least at the northern and southern extremes of this range, populations of L. hesperus may be completely or partially reproductively isolated and may be incipient biological species. Since matings across populations result in a higher rate of reproductive failure than matings within populations, selection would favour males that discriminate against mates from other geographical regions.

Our work suggests such discrimination could be mediated by differential response to female pheromones detected prior to arrival at a web, which would minimize the cost of mate searching for males. The evolution of differences in female pheromones and male response in these two populations may not require long time periods—in moths, where the genetic basis of pheromone 39 production and reception are well known, a single gene mutation can drastically alter the female pheromone, and male responses to pheromones are heritable and can be altered through selection in the laboratory (Evenden. et al. 2001; Roelofs et al. 1987). Further research comparing fitness effects of inter- and intra-population matings, relative male attraction, and heritability of differences in attractiveness are necessary to determine if divergence in airborne pheromones is coincident with incipient speciation in L. hesperus populations.

Divergence in morphological traits could be responsible for the observed failure of inter- population matings in L. hesperus due to physical incompatibility between populations. In addition, the observed population differences in embolus length (Figure 2.2) could be important since this affects the location of ejaculation. Longer emboli can reach into the female's sperm storage organ during mating. In the case of British Columbia males mating with Arizona females, the male's relatively short embolus might result in ejaculation in the insemination tubules, rather than in the organ itself—which reduces fertilization success in congeners

(Berendonck and Graven 2000; Snow 2003).

We also found a difference in body size between populations—males from the southern population were significantly larger than those from the northern population (Figure 2.1), despite a common laboratory rearing environment. This suggests the intriguing possibility that differential selection on male body size in northern and southern populations has resulted in allometric changes in male copulatory organs, which could directly affect inter-population mating success (Panhuis et al 2001; Schluter 2001). Future work will examine variation in ecological and social factors affecting selection on male body size and embolus length across the range of L. hesperus, and test whether this may be leading to parapatric speciation.

ACKNOWLEDGEMENTS

We thank S. Bhalla for her help in the lab, A. Sithamparanathan for his help in the field, and A. Leverette for acquiring L. hesperus from Arizona. We would also like to thank staff at 40 the BC Parks Service in Kamloops for expediting our permit application (Park Use Permit

TR0310300). This research was supported by grants from the Natural Sciences and Engineering

Research Council of Canada, Canadian Foundation for Innovation, and Ontario Innovation Trust to MCBA. 41

LITERATURE CITED

Anava, A., and Lubin, Y. 1993. Presence of gender cues in the web of a widow spider,

Latrodectus revivensis, and a description of courtship behaviour. The Bulletin of the

British Arachnological Society 9: 119-122.

Andrade, M.C.B. 2003. Risky mate search and male self-sacrifice in redback spiders. - Behav.

Ecol. 14: 531-538

Berendonck, B., and Greven, H. 2000. Morphology of female and male genitalia of Latrodectus

revivensis Shulov, 1948 (Araneae, Theridiidae) with regard to sperm priority patterns. In

European Arachnology 2000. Edited by S. Toft, and N. Scharff, Aarhus Univ. Press, pp.

157-167.

Evenden, M.L., Spohn, B.G., Moore, A.J., Preziosi, R.F., and Haynes K.F. 2002. Inheritance

and evolution of male response to sex pheromone in Trichoplusia ni (Lepidoptera :

Noctuidae). Chemoecology, 12: 53-59.

Foelix, R. 1982. Biology of Spiders. Harvard University Press, Cambridge, MA, USA.

Gwynne, D.T. 1987. Sex-biased predation and the risky mate-locating behaviour of male tick-

tock cicadas (Homoptera: Cicadidae). Anim. Behav. 35: 571-576.

Henschel, J. R. 2002. Long-distance wandering and mating by the dancing white lady spider

(Leucorchestris arenicold) (Araneae, Sparassidae) across Namib dunes. J. Arachnol. 30:

321-330.

Kaston, B.J. 1970. Comparative Biology of American black widow spiders. San Diego Soc. Nat.

Hist. 16: 33-82.

Kavale, J. 1986. The comparative biology of two Latrodectus species. MSc. Thesis. Zoology

Department, University of Otago, Dunedin, NZ. 42 Krafft, B. 1982. The significance and complexity of communication in spiders. In Spider

communication: mechanisms and ecological significance. Edited by

P. N. Witt, and J. S. Rovner, Princeton Univ. Press, pp. 15-66.

Levi, H.W. 1959. The spider genus Latrodectus (Araneae, Theridiidae). Trans. Am. Microsc.

Soc. 78: 7-43.

Panhuis, T.M., Butlin, R., Zuk, M., and Tregenza, T. 2001. Sexual selection and speciation.

Trends Ecol. Evol. 16: 364-371.

Papke, M., Schulz, S., Tichy, H., Gingl, E., and Ehn, R. 2000. Identification of a new sex

pheromone from the silk dragline of the tropical wandering spider Cupiennius salei.

Angewandte Chemie-International Edition 39: 4339-4341.

Papke, M. D., Riechert, S. E., and Schulz, S. 2001. An airborne female pheromone associated

with male attraction and courtship in a desert spider. Anim. Behav. 61: 877-886.

Pollard, S.D., Macnab, A.M., and Jackson, R.R. 1987. Communication with Chemicals:

Pheromones and Spiders. In Ecophysiology of Spiders. Edited by W. Nentwig, Springer-

Verlag. pp. 133-141.

Prouvost, O., Trabalon, M., Papke, M., and Schulz, S. 1999. Contact sex signals on web and

cuticle of Tegenaria atrica (Araneae, Agelenidae). Arch, of Insect Biochem. Physiol. 40:

194-202.

Riechert, S. E., and Singer, F. D. 1995. Investigation of potential male mate choice in a

monogmamous spider. Anim. Behav. 49: 715-723.

Robinson, M.H,. and Robinson, B. 1980. Comparative studies of the courtship and mating

behavior of tropical Araneid spiders. Pacific Insects Monog. 36: 1-218.

Roelofs, W., Glover, T., Tang, X.H., Sreng, I, Robbins, P., Eckenrode, C, Loftstedt, C,

Hansson, B.S., and Bengtsson, B.0.1987. Sex-pheromone production and perception in 43 European corn-borer mothers is determined by both autosomal and sex-linked genes.

Proc. Natl. Acad. Sci. USA 84: 7585-7589.

Ross, K., and Smith, R. L. 1979. Aspects of the courtship behavior of the black widow spider,

Latrodectus hesperus (Araneae: Theridiidae), with evidence for the existence of a

contact pheromone. J. of Arachnol. 7:69-77.

Schluter, D. 2001. Ecology and the origin of species. Trends Ecol. Evol. 16: 372-380.

Schulz, S., and Toft, S. 1993. Identification of a sex pheromone from a spider. Science, 260:

1635-1637.

Searcy, L. E., Rypstra, A. L., and Persons, M.H. 1999. Airborne chemical communication in the

wolf spider Pardosa milvina. J. Chem. Ecol. 25: 2527-2533.

Snow, L.S.E. 2003. Postcopulatory sexual selection in redback spiders {Latrodectus hasselti).

MSc. Thesis, University of Toronto at Scarborough. Scarborough, ON.

Sokal, R. R., and Rohlf, FJ. 1995. Biometry, 3rd Edition. W.H. Freeman and Company.

Szlep, R. 1965. The web-spinning process and web-structure of Latrodectus tredecimguttatus, L.

pallidus andL. revivensis. Proc. Zool. Soc. Lond. 145: 75-89.

Tichy, H., Gingl, E., Ehn, R., Papke, M. and Schulz, S. 2001. Female sex pheromone of a

wandering spider (Cupiennius salei): identification and sensory reception. J. Comp.

Physiol. A 187: 75-78.

Tietjen, W. J., and Rovner. J. S. 1982. Chemical communication in Lycosids and other spider. In

Spider communication: mechanisms and ecological significance. Edited by P.N. Witt,

and J.S. Rovner. Princeton Univ. Press, pp. 249-280

Trabalon, M., Bagneres, A.G., and Roland, C. 1997. Contact sex signals in two sympatric spider

species, Tegenaria domestica and Tegenariapagana. J. of Chem. Ecol. 23: 747-758.

Vollrath, F., and Parker, G.A. 1992. Sexual dimorphism and distorted sex ratios in spiders. Nature (Lond.), 360: 156-159. 44 Watson, P. J. 1986. Transmission of a female sex pheromone thwarted by males in the spider

Linyphia litigiosa (Linyphiidae). Science, 233: 219-221. I

45 Table 2.1 - The total number of males caught at cages containing web produced by females

from each of three groups or an empty control over the three day period of testing.

Population Day 1 Day 2 Day 3 Sticky In/on Total

strips only cages only (%)

(%) (%)

BC1^ L. hesperus 11 7 8 10 (63) 16 (52) 26 (55)

AZ* L. hesperus 4 3 0 1(6) 6(19) 7(15)

L. hasselti 6 3 5 5(31) 9(29) 14 (30)

Control 0 0 0 0(0) 0(0) 0(0)

Total 21 13 13 16 31 47

'Webs produced by lab-reared females descended from spiders collected in British Columbia.

* Webs produced by lab-reared females descended from spiders collected in Arizona. 46 Table 2.2 - The G-statistic values for effect of treatment on male attraction to cages for each of

three experimental days and with all days pooled.

Day df G~ ~~y~~~~

T~_ 2 3J14 015

2 2 2.462 0.29

3 2 7.538 0.02

Pooled1" 2 11.787 0.003

Total1 6 13.714 0.03

' Pooled G-statistic tests for significance over the whole experiment

*Total G-statistic tests for significance when each day is as a repeated measure. 47 Figure 2.1 - A, B. Spatial distribution of webs of female L. hesperus across two adjacent field sites (open squares) with the location of experimentally placed traps (closed circles). The Y-axis represents a direction with a central bearing of 146° and the X-axis represents a central bearing of 236°. Panels C and D are histograms showing the distribution of nearest-neighbour distances for webs in the first and second site respectively (Systat 10.2, 2002).

Figure 2.2 - Box plot comparing male body size (mean patella-tibia length of first pair of legs) and embolus length for laboratory-reared males from populations originally collected in British

Columbia (white boxes) and Arizona (grey boxes). 48

A B

• • Q 7- 7-; f \ 0d» 0 : nn 1 a rff • a ft C 4,; •** __ D s •to D '"rP (A • #* % a 2 - (D3 0 •ZI 1 • a I ^ 3 So 0 a -1—1—1—1—1—1—1- 20 40 I 80 100 120 20 40 60 80 100 120 140 160 Distance (m) Distance (m)

D

80 -0.5 70 k

"0 E 60 -0.4 ^ -i 'S 0.4 Q. 3

od e J2 50 E 3io n p .

Ba r * s 30 0.2 CO z1 20- 0.1 10- tL i r~i 0.0 12 3 4 5 0 12 3 4 5 6 7 Nearest neighbour distance (m) Nearest neighbour distance (m) 49

6.0

5.5-

5.0- 5K 4.5

4.0-1 I I XL X 3.5 T * 3.0 A T

2.5

2.0

Embolus i_eg

Structure 50

CHAPTER THREE

Risky mate search and mate preference in the golden orb-web spider (Nephilaplumipes)

Michael M. Kasumovic1, Matthew J. Bruce2, Marie E. Herberstein2, & Maydianne C. B.

Andrade1

'integrative Behaviour and Neuroscience Group, University of Toronto at Scarborough,

Toronto, ON, MIC 1A4, Canada

2Department of Biological Sciences, Macquarie University, New South Wales, 2109, Australia

Published in Behavioral Ecology, 2007, 18: 189-195, reprinted by permission of Oxford

University Press. License Number: 1717660922694 51

ABSTRACT

Mate searching is a risky behaviour that decreases survival by increasing predation risk and the risk of energy depletion. However, few studies have quantified actual mortality during mate search, making it difficult to predict mate searching and mating strategies. Using a mark and recapture study, we examined mate searching success in a highly sexually dimorphic species, the golden orb-web spider {Nephila plumipes). We show that despite the high density aggregations of this species, male survival during mate searching is extremely low (36%) and is phenotype independent. Surprisingly, males that survived mate search were in better condition after recapture than prior to release, most likely due to kleptoparasitism on females' webs. In a complementary release experiment in a field enclosure, we show that males are choosy, and adjust their choice of female depending on their own condition and weight. Thus, the high mortality rate of searching males in the field may be a cost of choosiness, since released males travelled further than necessary to settle on females. Although males were choosy about female phenotypes, they did not avoid webs with rival males already present. This suggests the cost of continued searching outweighs the cost of competition, but not the cost of mating with certain females. Further examinations of mate searching risk in other species in reference to their mating system and environmental conditions are necessary to determine the occurrence and effects of high mortality rates during searching. 52

INTRODUCTION

Before courtship and mating can begin, individuals must locate a mate. Although mate searching costs can be significant, they are rarely quantified in field studies. Mate searching is an energetically costly endeavour, and only those individuals in good condition can afford to begin searching (Proctor, 1992). This cost can be increased if individuals do not replenish their resources during searching, like in web-building spiders (Foelix, 1982). In addition to energetic costs (Byers et aL, 2005), searching individuals must also deal with an increased risk of predation. Studies that have examined mate searching costs show that predation rates can be high, and that mortality is often biased towards the sex that does the searching (Gwynne, 1987;

Gwynne and Bussiere, 2002; Norrdahl and Korpimaki, 1998; Sakaluk and Belwood, 1984; but see Su and Li, 2006). This is because searching individuals are more conspicuous, and because their greater activity increases the probability of encountering predators (Gwynne, 1987;

Gwynne and Bussiere, 2002; Sakaluk and Belwood, 1984). It is therefore no surprise that increased predation risk leads to individuals decreasing search times and activity (DeRivera et aL, 2003), or changing mate searching patterns and habitat use altogether (Sih, 1988).

Despite evidence demonstrating the costs of searching (Byers et aL, 2005; Gwynne,

1987; Gwynne and Bussiere, 2002; Sakaluk and Belwood, 1984), there are only a few studies that estimate survival rates during mate search. This limits the ability to associate a direct risk with mate searching, making it difficult to determine individual mate searching costs, and thus, to predict searching and mating strategies. The few studies that have quantified mate searching risk and the traits associated with searching success demonstrate that male mortality can be extremely high during searching (approximately 80% in two species of spiders: Andrade, 2003;

Vollrath, 1980), and that successful mate searching males tend to be larger and in better condition than average (Andrade, 2003; Foellmer and Fairbairn, 2005b; Vollrath, 1980). As mate search costs could impose a considerable constraint on multiple mating by males, it is 53 essential that mate search mortality be measured under natural conditions in a range of systems to determine whether such high costs are common.

Here we examine mate searching survival in the golden orb-web spider (Nephila plumipes). There were three goals of this study. First, we used a mark-recapture experiment to determine whether male mate search mortality in this aggregative species (Elgar, 1989) is similar to other spiders with more sparse web distributions (Latrodectus hasselti: 80-93% mortality, Andrade, 2003; N. clavipes: 88% mortality, Voilrath, 1980). Mate search mortality in spiders may be mainly due to predation and exhaustion (e.g. Andrade, 2003), so the cumulative risk should increase with time spent searching. Thus, it might be expected that mortality would be lower in a species with aggregated webs. Second, we examined whether male phenotype affected searching mortality. Third, we used a field enclosure study to determine whether males are choosy during mate searching, as this could elevate mortality rates. Male N. plumipes are monogynous due to a high frequency of injury and sexual cannibalism following attacks by their first mate (Elgar and Fahey, 1996; Schneider and Elgar, 2001). Recent work by Elgar et al.

(2003) shows that males did not choose females based on mating status or weight, however, males that chose virgin females were significantly heavier than those that chose mated females.

Thus males may be choosy about female phenotype or mating status to ensure their single mating has the highest payoff possible (Andrade and Kasumovic 2005; Bonduriansky 2001).

MATERIALS AND METHODS

Female N. plumipes build large orb-webs, which can be part of aggregations or can occur solitarily (Elgar, 1989), and are located at varying heights above the ground (Herberstein and

Elgar, 1994). Males mature either on or near the webs of females within these aggregations, or on their own orb-web separate from any aggregations (Kasumovic MK personal observation).

Once mature, males leave their web in search of females. Although male orb-web spiders are not known to feed while searching (Foelix, 1982), K plumipes males can kleptoparasitize prey 54 from a female's web once cohabitation begins (Kasumovic MK personal observation). While cohabiting with females, multiple males can settle on a single female's web, remaining there until an opportunity to mate arises (when the female is occupied with a prey item: Elgar et ah,

2003; Elgar and Fahey, 1996). Males assort according to size with larger males closer to the hub, allowing them mating priority (Elgar and Fahey, 1996). Although large size seems to play a role in mating success, direct competitions are rarely observed (Elgar and Fahey, 1996).

We collected male N. plumipes from our field site in Bicentennial Park, Pymble in

Sydney, NSW, Australia between February and March, 2005. Although it is common for male

K plumipes to autotomize legs while escaping cannibalism (only 31% of 327 males collected had all legs intact) (Elgar and Fahey, 1996), we only used males that had all eight legs to ensure no handicap in mate searching. Males can also break off the tip of the sclerotized portion of their intromittant organ (pedipalp) during copulation (Schneider et al., 2001). Since it is unclear if this affects male behaviour, we selected only males with intact pedipalps for our experiments.

All males and females used for the following experiments were released in nature once experiments were completed.

Mark and recapture

We completed two separate release experiments. In experiment 1, we released individual males (N = 52) throughout the field site to estimate survival. These males were marked using non-toxic fluorescent paint (Luminous paint, BioQuip Products) on the abdomen and the tibia of both of the first pair of legs. Since males commonly autotomize legs and would thus be likely to lose part of this marking, we marked all males identically. One week later in experiment 2, we used a separate group of males (N = 48) to make a second estimate of survival and to determine the distance travelled to a female's web and how mate searching affected male condition and survival. We weighed and individually marked males by gluing (Tarzan's Grip, Selleys) numbered labels (2 pt font printed on white paper approximately 2x3 mm in size) on their 55 abdomen. Both marking techniques have been successfully used for tracking bees (e.g. Fewell and Bertram, 2002). Size was measured as the average length of the patella-tibia of the first pair of legs (using callipers). In both experiments, males were anaesthetized using CO2 to facilitate marking. We then monitored males for an hour after marking and for five minutes after release to insure that neither marking method had an affect on locomotion.

Males were released in a field site (approximately 845 m2) composed of three separate islands of vegetation (approximately 45, 200, and 600 m2 in size) located on the Kuringai campus of the University of Technology campus in Sydney, NSW. The three islands consisted of predominantly Eucalypt forest and shrub and were less than 5m apart, surrounded by parking lots in all directions. Females found at this site settled both solitarily and in aggregations of up to nine spiders of various instars. Settlement of aggregations were similar in spacing to other described aggregations (Elgar, 1989) and other aggregations in the area (Kasumovic MM personal observation). Vehicular traffic was not observed during surveys, and was minimal throughout the study (completed during university holidays, February-March 2005).

We initially searched the entire site and located and mapped all females' webs. Although females were not marked, webs were found at the same locations during every survey. Males were released in the morning (08:00) on the ground in groups of 8-10 individuals within one metre of a conspecific female's web. Release locations were spread evenly throughout the field site. In experiment 1, 36 males were released in the largest island, and 16 males were released in the medium-sized island, while in experiment 2, 48 males were released in the largest island only. No males were released in the smallest island, as only two juvenile females' webs were found there. After release, we searched all three islands for marked males, surveying females' webs between 08:00 and 10:00 hours on each of the first five days, and then every two days for another 20 days. Any males found at this time were collected and returned to the laboratory.

Males from experiment 1 were measured (tibia patella length). Males from experiment 2 were 56 measured and weighed to determine individual differences in condition (see below) before and after release.

Male mate choice experiment

If cannibalism reduces male mating opportunities, males might be predicted to be choosy about potential mates (Andrade and Kasumovic, 2005; Bonduriansky, 2001). To examine whether males in the field may be choosing specific females and thus increasing their search times, we performed a mate choice experiment in a 3x3x2.5 m screened outdoor enclosure. For this experiment, we collected 60 adult males, as well as 30 penultimate and 30 adult females from the field. Female instar was distinguished by examining the epigyne. Adult females have a protruding epigyne that has two clear openings, while penultimate females have the same protrusion but the openings are covered. Although we could not identify the mating status

(virgin or mated) of adult females collected, they were representative of the females available in the field at the time of the mark-recapture experiment and thus provided a representative sample of choices available to those males.

We ran three replicates of a male mate choice experiment where we randomly placed 10 juvenile and 10 adult females along the walls of the enclosure. We released females at 10:00 and allowed them to build their webs over 24 hours. The following morning at 10:00, we counted all the adult and penultimate females within the enclosure. All the females built their webs in the upper part of the enclosure using the walls and ceiling for support at similar heights to those found in the field. We then released 20 males evenly spaced along the base of each of the four walls of the enclosure (five males per wall). Once again, we used only males that had all their legs and pedipalps intact. We returned to the enclosure six hours later to collect all the females, and any males found on each female's web. We also searched all the walls of the enclosure, and each of the females' webs for dead males. Males and females were weighed and measured as above and we also calculated male condition after recapture. 57 Condition estimate

To estimate an index of body condition, we used a residual index (a regression of body weight on linear size) that has recently been proven to perform well to be biologically relevant

(Schulte-Hostedde et al., 2001; Schulte-Hostedde et al., 2005). To determine the relationship between log (weight) and log (size) for N. plumipes males, we performed an Reduced Major

Axis regression (Green, 2001) on an independent group of males from a field captured population (N = 320, unpublished data). We regressed male log (weight) on log (size) and found the slope of the relationship to equal 4.22. We then used 4.22 as the exponent for the root of log

(weight) regressed on male log (size) (for greater details see Kasumovic and Andrade, 2006).

We did not calculate a condition index for females since the relationship between size and weight did not correlate as strongly. This is most likely due to natural variance in female weight and size due to variance in the instar and mating status of the females used. Thus, for any analyses involving female traits, we only used weight and instar as the descriptive variables, since females of different instars had significantly different sizes (t=7.961, df=54, p<0.001).

RESULTS

Mark and recapture experiment

A total of 34 (34%) males were recaptured throughout both replicates. There was no difference in the proportion of males recaptured in the two release experiments (Experiment 1:

19 males, 36.5%, Experiment 2: 15 males, 31.3%; Fisher's Exact two-tailed tesr=0.67). There was also no difference in the proportion of males caught in each island in experiment 1 (large island: 14 males, 38.9%, medium island: 5 males 31.3%; Fisher's Exact two-tailed test=0.76). In experiment 2, males were only recovered in the large island where they were released, suggesting that males either limited their search for females within each island, or that no males survived if they left the island. No males were found on the smallest island in either experiment.

Of the 34 males recaptured, 21 (62%) were found together with rivals on a female's web, thus, 58 competition over access to females may occur despite high male mortality rates. Up to five males (range 1-5, mean=1.3) were found cohabiting with the same female.

In the second experiment where we could follow individuals, the 15 recaptured males travelled an average of 9.91±1.24 m to successfully find a female (range 2.4-17.8 m), even though they were released within 1 m of a conspecific female. Recaptured males from both experiments took an average of 3.38±0.61 days to successfully find a female (range 1-15 days).

We assessed whether male success at finding webs depended on the time elapsed following initial release by using a survival analysis (Sokal and Rohlf, 1995). This analysis compares the observed pattern of male recapture over time to the expected pattern if there was an equal probability of discovering a male on each day following release (i.e. one or two males recovered each day until 34 males are recovered, Figure 3.1). This analysis suggests a much higher success rate for males that find a web quickly (Survival analysis, %2=56.91, d.f.=2, P<0.0001).

Since only 34 males were recaptured the power of parametric statistics to detect whether size is related to mate searching success was limited. In a more powerful analysis, we used a subsampling technique to create a null distribution to determine whether the males recaptured were on average larger than the males that were released (Manly, 1991). To create this null model, we randomly subsampled 34 males from the distribution of 100 released males and calculated the mean of this distribution. We repeated this subsampling a total of 10,000 times, and used the mean calculated each time to create a null distribution of 34 randomly selected males. This allowed us to compare the actual mean of recaptured males to a null distribution to determine whether our observed sample differed from random (Figure 3.2). We also compared variances in male size before and after recapture to examine whether stabilizing selection operates on male size during mate searching. The mean leg size of the recaptured males was

0.50±0.01 mm, which was not significantly different from the null distribution (0.49 mm,

P=0.19, Figure 3.2). There was also no difference in the variance in male size between released 59 and recaptured males (Bartlett's test; F=1.08, d.f.-l, P=0.30). Therefore, size was not a predictor of mate searching success and there is no evidence for directional or stabilizing selection.

We performed the same subsampling technique to determine whether male condition had an influence on mate searching success. In this case, we randomly subsampled 15 males from the 48 released in the second experiment 10,000 times to create a null distribution. In this analysis, males in better condition before release were not more likely to be recaptured with a female (recaptured mean=0.0028±0.0067; null distribution mean=0.0003, P=0.54, Figure 3.3).

Moreover, 14 of the 15 males recaptured in experiment 2 were heavier after recapture than prior to release (release= 17.41±0.26 mg, recapture=18.68±0.26 mg; paired t-test=4.83, d.f.=14,

PO.0003).

Male mate choice experiment

At the end of our field-enclosure trials, we collected a total of 35 adult females and 21 penultimate females from the enclosures. The deviation from the original number of adult and penultimate females was due to four of the original penultimate females either dying or being killed by neighbouring females, while five other penultimate females moulted to maturity overnight, before males were released. No inter-female aggression was observed during the trials. A total of 51 of the 60 (85%) males released were collected from female's webs. One other male was located on the wall of the enclosure. It is unlikely that males mated with females in the six hours allotted for the mate choice experiment since males cohabit with females for long periods of time and do not attempt to mate with females until females are distracted with a prey item (Elgar et al., 2003; Elgar and Fahey, 1996; Schneider and Elgar, 2001), and no prey items were found in females' webs. Furthermore, it was unlikely that males were cannibalized as females store all food caught (Griffiths et al., 2003) and we did not find any wrapped bodies of males in females' webs. Thus, males that were not recovered were likely those that failed to 60 find a female. This suggests that, even in the enclosure there was a moderate mate searching failure rate (15%).

Of the females that attracted males, the mean number of males per female's web was 1.7

(range 1-8, Figure 3.4). Female instar or weight were not predictors of whether a female attracted at least one male (Table 3.1). Since previous work found that a male's weight influenced the female males chose (Elgar et al, 2003), we used two separate nominal logistic analyses (JMPin 4.02, SAS Institute 2000) to examine whether successful males made their choice (1) based on their own size and/or weight (with condition calculated as a covariate,

Garcia-Berthou, 2001) and also (2) based on size and/or condition calculated by residuals (see methods). Our results show that heavier better condition males preferred penultimate females over adults (condition of males choosing penultimate: 0.0010±0.0007, and adult: -

0.0004±0.0005 females; Table 3.2). However, unlike in the mark-recapture experiment, males could not have increased their condition through kleptoparasitism since there were no prey items found in any of the females' webs.

DISCUSSION

We found no phenotypic correlates of male mate searching success in N. plumipes since neither condition nor leg size predicted success at finding a potential mate. Furthermore, there was no evidence for directional or stabilizing selection for male size. There are three other studies that have examined phenotypic correlates of male mate searching success in highly dimorphic spiders. The first two (Andrade, 2003; Vollrath, 1980) also used mark and recapture techniques to follow individual males and found no phenotypic predictors of mate searching success. The third study (Foellmer and Fairbairn, 2005b) compared the phenotypes of males found on female's webs (successful mate searchers) to recently-matured males found on their own webs (before mate searching began). Although they found that successful males were larger, this result was limited to a single population in one of two years. These results, together 61 with our study, suggest only limited evidence that male size may be under selection during mate searching, and this may be true only under certain environmental conditions. In this species, size is more likely determined by stronger selection pressures such as smaller size for scramble competition and larger size for direct competition (e.g. Kasumovic and Andrade, 2006).

We also provide the first evidence that successful males may be in better condition after recapture than before release. Since web-building spiders are thought to refrain from feeding while searching (Foelix, 1982), this difference is most likely due to kleptoparasitism on the female's web or perhaps nectar-feeding during mate search (Pollard et al., 1995). Although condition will undoubtedly be important during mate search, especially in species that do not feed as they search, it may be sufficient for males to achieve some threshold condition that makes it physiologically possible to survive the necessary period of mate searching (Proctor,

1992). Males may then be able to supplement their energy reserves that have been used during searching with resources found on a female's web. This may explain why examinations of condition in other mate searching web building spiders have never showed significant effects

(e.g. Andrade, 2003; Foellmer and Fairbairn, 2005b).

The most striking result of our experiment is the low male survival rate despite the fact that males are searching in dense web aggregations (Elgar, 1989). Such aggregations are thought to decrease mortality rates since the costs of sampling should be lower. Despite these aggregations, a total of 76% of males perished during mate search. We believe our recapture rates accurately represent male survival rates in the field. Once males mature, they apparently focus on mate searching. In our mate choice experiment, no males built their own capture webs instead of seeking females. This means that males that were not found on a female's web at our site were either dead (as we conclude), were still searching, or had emigrated from the site. First, it is unlikely that males were still searching within the site at the end of our experiment as recapture rates of males were high initially, but we found no marked males in the last 14 days of 62 our sampling period (Figure 3.1). Second, emigration rates are apparently low since no males were located on the smallest island in the first release study, and none were found outside the island of their release in the second experiment. Our survival estimate is in agreement with previous studies on other species with extreme sexual-size dimorphism {Latrodectus hasselti:

80-93% mortality, Andrade, 2003; N. clavipes: 88% mortality, Vollrath, 1980), and suggests there are several size and condition-independent factors influencing mortality.

Since survivorship decreases as the time spent searching increases (this study; Kotiaho et al., 1999), one of the most important features influencing male survival is mate searching time.

This could be for two reasons. First, actively searching males are more conspicuous (Sakaluk and Belwood, 1984) and increase their probability of encountering predators as time progresses

(Gwynne, 1987; Magnhagen, 1991) than individuals stationary on their webs. Second, this could also be due to senescence or due to depleted energy reserves in species with short-lived males

(e.g. Andrade, 2003; Bonduriansky and Brassil, 2005). Therefore, any variables that increase the time required to successfully find a mate are likely to decrease male survival during mate searching.

One strategy that is likely to increase search times, and thus mortality rate, is male choosiness. N. plumipes males typically only get a single opportunity to mate (Schneider and

Elgar, 2001) and are therefore predicted to be choosy (Bonduriansky, 2001). By increasing choosiness, males decrease the probability of choosing inappropriate females, but potentially increase the risk of predation and depletion of energy reserves since they may be searching for longer periods. Other studies examining attractive pheromones in spiders show that males can successfully find penultimate females (Andrade and Kasumovic, 2005). We show here that although males do not prefer females of a particular instar, a male's choice of female was based on male weight and condition (this study, Elgar et al., 2003), preferring penultimate instar females over adults when they are heavier. This pattern might also arise if males that cohabit 63 with juveniles are more likely to be successful at kleptoparasitism or if males in poor condition fight more vigorously to exclude rivals from webs of adults. These explanations are unlikely because (1) there were very few prey species in the enclosure, and none were found wrapped in or below females' webs. There was similarly no evidence of cannibalism. (2) males of this species engage in very little combat (Elgar and Fahey, 1996), and no inter-male aggressive interactions were observed in this study, even in webs with multiple cohabitants. As a result, males are often forgoing females on nearby webs, and travelling further than necessary (8 m in this study) to reach preferred females.

Males are predicted to be choosy in this species if the benefit gained by being choosy is greater than the increased mortality risk. Males that settle with penultimate females will be required to cohabit for longer periods of time and may require greater resources to do so.

However, by settling with penultimate females, males can ensure they mate first which may lead to a greater proportion of paternity (Elgar et al., 2003). Males in poorer condition may benefit by settling with adult females since this would allow them an opportunity to mate more quickly

(e.g. Elgar et al., 2003), decreasing their risk of starvation. Further studies of male mate choice and female attributes are required to help determine the factors influencing male choice. Despite being choosy with respect to female phenotype and instar, males are not leaving females' webs regardless of the presence of conspecifics both in the field and the enclosure experiment.

Other factors that may influence male searching times are likely to be associated with female density and environmental conditions. Decreases in female density should result in increased search times as should environmental conditions that decrease detectability of females. Since many male spiders use pheromones to locate potential mates (Andrade and

Kasumovic, 2005; Gaskett et al., 2004; Kasumovic and Andrade, 2004; Papke et al., 2001;

Searcy et al., 1999; but see Anderson and Morse, 2001), conditions that decrease the spread of pheromones will force males to search for longer than if pheromone gradients are clearly 64 detectable (Bell and Carde, 1984). These conditions could include unfavourable wind speeds or low temperatures since this not only leads to poorer airborne dispersion of pheromones (Bell and Carde, 1984), but also results in slower movement by males (Foelix, 1982).

Our study, along with others (Andrade, 2003; Vollrath, 1980) suggests that male survival during mate search is very low in some web-building spiders, and that phenotypic correlates have a weaker effect on success than do extrinsic factors. High mortality rates during mate searching have been previously proposed to (1) lead to decreased pre-copulatory competition between conspecifics leading to the evolution of the extreme sexual-size dimorphism (Prenter et al., 1998; Prenter et al., 1999; Vollrath and Parker, 1992) and (2) cause a female-biased sex ratio and thus decrease the risk of sperm competition and the likelihood of investment in paternity- protection mechanisms (e.g. Fromhage et al., 2005). However, here we found that, despite high mortality rates, the presence of rivals on a female's web did not seem to discourage settlement by newly arriving males, and multiple males commonly cohabit with females. A male-biased sex ratio on individual female's webs can occur if there is an overall male-biased sex ratio after to mate searching, or if a proportion of females do not attract males due to their location and therefore, remain unmated (as in our field enclosure). Thus, high mortality rates will not necessarily relax selection on male traits for combat or sperm competition since the local operational sex ratio can still be male-biased (e.g. Foellmer and Fairbairn, 2005a). Our study shows how male choosiness can have significant effects on the dynamics of systems with risky mate searching (e.g. Bonduriansky, 2001; Foellmer and Fairbairn, 2005a; Fromhage et al., 2005) and highlights the difficulty in establishing the direction of causality when examining links between mate search mortality and male mating strategies.

ACKNOWLEDGMENTS 65 We would like to thank MW Foellmer and two anonymous referees for comments that helped to improve the manuscript. This research was supported by an NSERC PGS B and OGS to MMK, Macquarie University to MJB, Macquarie University and ARC to MEH, and grants from NSERC, CFI and OIT to MCBA. 66

LITERATURE CITED

Anderson JT, Morse DH. 2001. Pick-up lines: cues used by male crab spiders to find

reproductive females. Behav Ecol 12:360-366.

Andrade MCB. 2003. Risky mate search and male self-sacrifice in redback spiders. Behav. Ecol.

14:531-538.

Andrade MCB, Kasumovic MM. 2005. Terminal investment strategies and male mate choice:

Extreme tests of Bateman. Int. Comp. Biol. 45:838-847.

Bell WJ, Carde RT, 1984. Chemical ecology of insects. Sunderland, MA: Sinauer Press.

Bonduriansky R. 2001. The evolution of male mate choice in insects: a synthesis of ideas and

evidence. Biological Reviews 76:305-339.

Bonduriansky R, Brassil CE. 2005. Reproductive ageing and sexual selection on male body size

in a wild population of antler flies {Protopiophila litigata). Journal of Evolutionary

Biology 18:1332-1340.

Byers J, Wiseman P, Jones L, Roffe T. 2005. A large cost of female mate sampling in

pronghorn. American Naturalist 166:661-668.

DeRivera CE, Backwell PRY, Christy JH, Vehrencamp SL. 2003. Density affects female and

male mate searching in the fiddler crab, Uca beebei. Behavioral Ecology and

Sociobiology 53:72-83.

Elgar MA. 1989. Kleptoparasitism - a cost of aggregating for an orb-weaving spider. Animal

Behaviour 37:1052-1055.

Elgar MA, Bruce MJ, de Crespigny FEC, Cutler AR, Cutler CL, Gaskett AC, Herberstein ME,

Ramamurthy S, Schneider JM. 2003. Male mate choice and patterns of paternity in the

polyandrous, sexually cannibalistic orb-web spider Nephila plumipes. Australian Journal

of Zoology 51:357-365. 67 Elgar MA, Fahey BF. 1996. Sexual cannibalism, competition, and size dimorphism in the orb- weaving spider Nephilaplumipes Latreille (Araneae: Araneoidea). Behavioral Ecology

7:195-198.

Fewell JH, Bertram SM. 2002. Evidence for genetic variation in worker task performance by

African and European honey bees. Behavioral Ecology and Sociobiology 52:318-325.

Foelix R, 1982. Biology of spiders. Cambridge: Harvard University Press.

Foellmer MW, Fairbairn DJ. 2005a. Competing dwarf males: sexual selection in an orb-weaving

spider. Journal of Evolutionary Biology 18:629-641.

Foellmer MW, Fairbairn DJ. 2005b. Selection on male size, leg length and condition during

mate search in a sexually highly dimorphic orb-weaving spider. Oecologia 142:653-662.

Fromhage L, Elgar MA, Schneider JM. 2005. Faithful without care: The evolution of

monogyny. Evolution 59:1400-1405.

Garcia-Berthou E. 2001. On the misuse of residuals in ecology: testing regression residuals vs.

the analysis of covariance. Journal of Animal Ecology 70:708-711.

Gaskett AC, Herberstein ME, Downes BJ, Elgar MA. 2004. Life-time male mating preferences

in a sexually cannibalistic orb-web spider (Araneae: Araneidae). Behaviour 141:1197-

1210.

Green AJ. 2001. Mass/length residuals: Measures of body condition or generators of spurious

results? Ecology 82:1473-1483.

Griffiths BV, Holwell GI, Herberstein ME, Elgar MA. 2003. Frequency, composition and

variation in external food stores constructed by orb-web spiders: Nephila edulis and

Nephila plumipes (Araneae : Araneoidea). Australian Journal of Zoology 51:119-128.

Gwynne DT. 1987. Sex-biased predation and the risky mate-locating behaviour of male tick-

tock cicadas (Homoptera: Cicadidae). Animal Behaviour 35:571-576. 68 Gwynne DT, Bussiere L. 2002. Female mating swarms increase predation risk in a 'role- reversed' dance fly (Diptera : Empididae : Rhamphomyia longicauda Loew). Behaviour

139:1425-1430.

Herberstein ME, Elgar MA. 1994. Foraging strategies of Eriophora transmarina and Nephila

plumipes (Aranea: Araneoida): Nocturnal and diurnal orb-weaving spiders. Australian

Journal of Ecology 19:451-457.

Kasumovic MM, Andrade MCB. 2004. Discrimination of airborne pheromones by mate-

searching male western black widow spiders {Latrodectus hesperus): species- and

population-specific responses. Can. J. Zool. 82:1027-1034.

Kasumovic MM, Andrade MCB. 2006. Male development tracks rapidly shifting sexual versus

natural selection pressures. Current Biology 16:R242-243.

Kotiaho JS, Alatalo RV, Mappes J, Parri S. 1999. Sexual signalling and viability in a wolf

spider (Hygolycosa rubrofasciata): measurements under laboratory and field conditions.

Behavioral Ecology and Sociobiology 46:123-128.

Magnhagen C. 1991. Predation Risk as a Cost of Reproduction. Trends in Ecology & Evolution

6:183-185.

Manly B, 1991. Randomization and Monte-Carlo methods in biology. New York: Chapman and

Hall.

Norrdahl K, Korpimaki E. 1998. Does mobility or sex of voles affect risk of predation by

mammalian predators? Ecology 79:226-232.

Papke MD, Riechert SE, Schulz S. 2001. An airborne female pheromone associated with male

attraction and courtship in a desert spider. Animal Behaviour 61:877-886.

Pollard SD, Beck MW, Dodson GN. 1995. Why do male crab spiders drink nectar? Animal

Behaviour 49:1443-1448. 69 Prenter J, Elwood RW, Montgomery WI. 1998. No association between sexual size dimorphism

and life histories in spiders. Proceedings of the Royal Society Biological Sciences Series

B 265:57-62.

Prenter J, Elwood RW, Montgomery WI. 1999. Sexual size dimorphism and reproductive

investment by female spiders: A comparative analysis. Evolution 53:1987-1994.

Proctor HC. 1992. Effect of food-deprivation on mate searching and spermatophore production

in male water mites (Acari, Unionicolidae). Functional Ecology 6:661-665.

Sakaluk SK, Belwood JJ. 1984. Gecko phonotaxis to cricket calling song: a case of satellite

predation. Animal Behaviour 32:659-662.

Schneider JM, Elgar MA. 2001. Sexual cannibalism and sperm competition in the golden orb-

web spider Nephila plumipes (Araneoidea): Female and male perspectives. Behavioral

Ecology 12:547-552.

Schneider JM, Thomas ML, Elgar MA. 2001. Ectomised conductors in the golden orb-web

spider, Nephila plumipes (Araneoidea): a male adaptation to sexual conflict? Behavioral

Ecology and Sociobiology 49:410-415.

Schulte-Hostedde AI, Millar IS, Hickling GJ. 2001. Evaluating body condition in small

mammals. Canadian Journal of Zoology 79:1021-1029.

Schulte-Hostedde AI, Zinner B, Millar JS, Hickling GJ. 2005. Restitution of mass-size

residuals: Validating body condition indices. Ecology 86:155-163.

Searcy LE, Rypstra AL, Persons MH. 1999. Airborne chemical communication in the wolf

spider Pardosa milvina. Journal of Chemical Ecology 25:2527-2533.

Sih A. 1988. The Effects of Predators on Habitat Use, Activity and Mating-Behavior of a Semi-

Aquatic Bug. Animal Behaviour 36:1846-1848.

Sokal RR, Rohlf FJ, 1995. Biometry, 3rd ed. New York: Freeman. 70

Su KT Li DQ. 2006. Female-biased predation risk and its differential effect on the male and

female courtship behaviour of jumping spiders. Anim. Behav. 71:531-537.

VollrathF. 1980. Male body size and fitness in the web-building spider Nephila clavipes.

Zeitschrift Tierpsychologie 53:61-78.

Vollrath F, Parker GA. 1992. Sexual dimorphism and distorted sex ratios in spiders. Nature

360:156-159. 71 Table 3.1 - Nominal logistic analysis demonstrating that whether a female attracted at least one male did not depended on female instar and/or weight.

Source DF Likelihood Ratio P

ChiSquare

Female instar 1 0.432 051

Female weight 1 0.067 0.80

Instar x Weight 1 0.727 0.39 72 Table 3.2 - Results of two separate nominal logistic analyses examining whether A) male size, weight and condition (calculated as a covariate) and B) male size, and condition (calculated by residuals) influence choice of available females. In both analyses, heavier and better condition males preferred penultimate females.

Source df Likelihood Ratio

ChiSquare

A Male size 1 2.66 0.10

Male weight 1 6.09 0.01

Size x Weight 1 5.38 0.02

(condition)

B Male size 1 2.90 0.09

Male condition 1 5.47 0.019

Size x Condition 1 2.65 0.10 73 Figure 3.1- Survival analysis of rate of capture for all males that were eventually recovered in females' webs during two release experiments (n = 34). Actual recapture rate (solid line) is compared to null distributions that assume a constant recapture rate per day (dotted line = single male/day; dashed line = two males/day). Most males that successfully reached a female's web do so within a few days of release (by day 3, only 40% of males that will eventually be recaptured are still searching). Relative to null models, very few males successfully reach a web after searching for more than 5 days.

Figure 3.2 - The null distribution of male size generated by randomly subsampling 34 individuals from the original population (N=100) 10,000 times. The arrow is the observed mean

(0.50 mm) of the recaptured individuals, which is not significantly different from that expected by chance (P=0.19).

Figure 3.3 - The null distribution of male condition generated by randomly subsampling 15 individuals from the second experimental population (N=48) 10,000 times. The arrow is the observed mean (-0.34) of the recaptured individuals, which is not significantly different from that expected by chance (P=0.74).

Figure 3.4 - The number of males attracted to the webs of adult (white) and penultimate (black) females in a field enclosure release experiment. 74

LOT • wm**• 0.9- I ••• , ! *•* D) 0.8- Hi I *«. 4 • taw "• *"i. 1: ^ ••**• % x: 0.7- 1 *. o " '<—» *•*. <5 0.6- 1 0) 1 CcO 0.5- "... o • 1 **l | "*l rt i 0.4- X o • t 1 -. 0.3- '- X, o i :.. « %, ::.t £L 0.2- | L ":» . 1 • -: 0.1- »-| fc • "L. 1- '. %.. n n U.U """*T r™ -r 1 1 i • i • i • i • i 5 10 15 20 25 30 3J Days Frequency

O O o o o

9L (D V) N'

3 76

3000

2500

2000 o § 1500 cr

LL 1000-

500-

-0.02 -0.01 0.00 0.01 0.02 Male Condition 77

m E .0)

CD -Q E 13 , • • ^^i 2 3 4 Number of males per web 78

CHAPTER FOUR

Male development tracks rapidly shifting sexual versus natural selection pressures

Michael M. Kasumovic and Maydianne C.B. Andrade

Integrative Behaviour and Neuroscience Group, University of Toronto at Scarborough, Toronto,

ON, MIC 1A4, Canada

Published in Current Biology, 2006,16: R242-R243, reprinted by permission from Elsevier 79 Sexual selection theory predicts that interplay between sexual and natural selection shapes phenotypic distributions over evolutionary time [1-3]. We show this tension also significantly affects individual development. Developmental plasticity, where individuals vary ontogeny in response to variation in the selection to be encountered upon maturity [4-6], has previously been demonstrated in response to single selective pressures (e.g. predation [7], sperm competition [8]). We show developmental shifts are also triggered by the relative magnitude of opposing natural and sexual selection—rapidly changing distributions of phenotypic traits critical to reproductive success. Males of the sexually cannibalistic redback spider {Latrodectus hasselti) show a tactical, condition-dependant shift between conflicting developmental strategies favoured by scramble competition or surviving mate search. Male body condition, development rate, and size changed with the relative importance of these selective pressures, which naturally fluctuate throughout a breeding season [9-11]. This has important implications for studies comparing fitness values of fixed traits without regard for plasticity.

We exposed penultimate-instar males to non-contact, pheromonal cues [12, 13] simulating dense (females present) or sparse (females absent) populations, and crossed this treatment with three diet levels (high, mid or low). When females are sparse, natural selection for provisioning to survive mate searching (80% mortality) [14] will be more intense than sexual selection for scramble competition for virgin females (90% median paternity for first male) [15].

Since multiple males commonly settle on females' webs [16], male density will also affect competitive success. We therefore examined male growth, development time, and body condition as a function of diet, female treatment and male density. We also assessed male phenotypes as a function of female proximity in the field (see supplemental materials).

Our laboratory and field results collectively show that males use pheromonal cues of female and male density to moderate development. When females are absent, redback males trade-off rapid development for increased size and body condition (Table 4.1, Figure 4.1) since 80 larger, better condition phenotypes are more likely to survive mate searching and direct competition when females are sparse. In contrast, males trade-off size and body condition for rapid development when females are present (Table 4.1, Figure 4.1), ensuring males reach virgin females first [15]. Similarly, in the field, as female proximity increases, male size (Two- way ANCOVA, F=8.96, P=0.004) and body condition (Two-way ANCOVA, F=5.14, P=0.028) decrease (Supplementary Table 4.1). Sexual selection for scramble competition has been proposed to contribute to sexual size dimorphism as a side-effect of rapid development of males

[17]. These results show the degree of dimorphism could change within seasons as selection fluctuates.

Although increased female density resulted in smaller, poorer condition males that developed rapidly, increased male density led to the development of smaller males in better condition (Table 4.1). This independent effect of neighbouring males suggests a trade-off between size and body condition. Thus, the relative importance of these traits to male fitness apparently changes as a function of both female and male density.

Although conflicting selection usually results in non-optimal phenotypes with moderate fitness [e.g. 18,19], in redbacks, conflicting pressures act on developmental plasticity which yields phenotypes optimized for specific competitive challenges [6]. This is critical because redback males are under strong selection to succeed in their single mating opportunity [16, 20].

Similar plasticity may be important where (1) males have few mating opportunities that occur over small spatial or temporal scales, (2) juveniles can detect cues reliably predicting adult challenges, and/or (3) allocation of limited resources yields disproportionate increases in selected fitness components. These conditions may occur in other mate-searching species with variable population density and relatively short male lifespan [17].

We demonstrate short-term shifts in conflicting sexual and natural selection can be a major source of variation in male phenotypic traits important to sexual competitiveness [1,17]. 81 Although decreases in resource availability can decrease male size and body condition [Table

4.1, Figure 4.1, 8], small size may also arise because the net fitness benefit of decreased development time outweighs potential costs of small size in the local environment. Thus the maintenance of classes of individuals thought to be competitively inferior [1] could be explained by plasticity and fine-scale environmental heterogeneity, rather than variation in condition [7].

This highlights the critical need to expand definitions of male quality beyond fixed adult traits

(e.g., size) to include an individual's ability to respond to changes in varying selection pressures

[9]. Most importantly, for the countless studies of sexual selection on male phenotypes, assessments of fitness effects measured without consideration of local conditions and development could seriously overstate the importance of fixed, heritable traits to lifetime reproductive success.

ACKNOWLEDGMENTS

We thank M.D. Biaggio, J.M. Brandt, R. Brooks, D.O. Elias, D.T. Gwynne, J.C. Johnson, A.P.

Moczek, M. Polak, L. Rowe and two anonymous reviewers for comments on this paper. This project was funded by an NSERC PGS B to MMK and grants from NSERC, CF1 and OIT to

MCBA. 82

LITERATURE CITED

1. Andersson, M. (1994). Sexual Selection (Princeton: Princeton University Press).

2. Darwin, C. (1871). The Descent of Man, and Selection In Relation to Sex (London:

Murray).

3. Fisher, R. A. (1930). The Genetical Theory of Natural Selection (Oxford: Clarendon

Press).

4. Roff, D.A. (1992). The Evolution of Life Histories: Theory and Analysis (New York:

Chapman & Hall).

5. Stearns, S.C. (1992). The Evolution of Life Histories (Oxford: Oxford University Press).

6. West-Eberhard, M.J. (2003). Developmental Plasticity and Evolution (New York:

Oxford University Press).

7. Benard, M.F. (2005). Predator-induced phenotypic plasticity in organisms with complex

life histories. Annual Review of Ecology and Systematics 35, 651-673.

8. Gage, M.J.G. (1995). Continuous variation in reproductive strategy as an adaptive

response to population density in the moth Plodia interpunctella. Proc. R. Soc. Lond. B

261,25-30.

9. Meyers, L.A., and Bull, J.J. (2002). Fighting change with change: adaptive variation in

an uncertain world. Trends in Ecology & Evolution 17, 551-557.

10. Jann, P., Blanckenhorn, W.U., and Ward, P.I. (2000). Temporal and microspatial

variation in the intensities of natural and sexual selection in the yellow dung fly

Scathophaga stercoraria. J. Evol. Biol. 13, 927-938.

11. Kassen, R. (2002). The experimental evolution of specialists, generalists, and the

maintenance of diversity. J. Evol. Biol. 15,173-190.

12. Andrade, M.C.B., and Kasumovic, M.M. (2005). Terminal investment strategies and

male mate choice: Extreme tests of Bateman. Int. Comp. Biol. 45, 838-847. 83 13. Kasumovic, M.M., and Andrade, M.C.B. (2004). Discrimination of airborne pheromones

by mate-searching male western black widow spiders {Latrodectus hesperus): species-

and population-specific responses. Can. J. Zool. 82,1027-1034.

14. Andrade, M.C.B. (2003). Risky mate search and male self-sacrifice in redback spiders.

Behav. Ecol. 14, 531-538.

15. Snow, L.S.E., and Andrade, M.C.B. (2005). Multiple sperm storage organs facilitate

female control of paternity. Proc. R. Soc. Lond. B 272,1139-1144.

16. Andrade, M.C.B. (1996). Sexual selection for male sacrifice in the Australian redback

spider. Science 271, 70-72.

17. Thornhill, R., and Alcock, J. (1983). The evolution of insect mating systems

(Cambridge, Massachusetts; Harvard University Press).

18. Candolin, U. (2004). Opposing selection on a sexually dimorphic trait through female

choice and male competition in a water boatman. Evolution 58,1861-1864.

19. Bonduriansky, R., and Rowe, L. (2003). Interactions among mechanisms of sexual

selection on male body size and head shape in a sexually dimorphic fly. Evolution 57,

2046-2053.

20. Forster, L.M. (1992). The stereotyped behaviour of sexual cannibalsim in Latrodectus

hasselti Thorell (Aranea: Theridiidae), the Australian redback spider. Australian Journal

of Zoology 40,1-11. 84 Table 4.1: Female presence and diet affect growth, body condition and development time of males. Values in bold are significant (table-wise p < 0.05).

MANOVA

Pillai's Trace F d.f. P

Female 0.086 6.01 3,198 0.0006

Diet 0.584 27.27 6,398 <0.0001

Female x Diet 0.059 1.83 6,398 0.09

Males 0.084 5.73 3,198 0.0009

Female x Males 0.009 0.48 3,198 0.70

Diet x Males 0.016 0.97 6,398 0.44

Female x Diet x Males 0.073 2.32 6,398 0.032

Univariate Analyses

Source F P

Development Time Female 4.80 0.03

Diet 80.70 < 0.0001

Female x Diet 0.25 0.78

Males 0.34 0.56

?emale x Males 0.01 0.90

Diet x Males 0.058 0.95

x Diet x Males 0.08 0.92

Body Condition Female 4.62 0.033

Diet 23.15 < 0.0001

Female x Diet 1.84 0.16

Males 15.22 0.0001 85 Female x Males 1.25 0.26

Diet x Males 2.68 0.07

Female x Diet x Males 2.67 0.07

Experimental Growth Female 5.23 0.023

Diet 23.73 < 0.0001

Female x Diet 2.58 0.078

Males 4.41 0.037

Female x Males 0.06 0.81

Diet x Males 0.07 0.94

Female x Diet x Males 3.21 0.042 86 Figure 4.1: Development time (A), body condition (B) and growth (C) of males reared in the absence (white) or presence (black) of virgin females across three diet treatments (X-axis). Error bars are one standard error. Asterisks and different letters represent significant results of Tukey-

Kramer HSD post-hoc tests. In (A and B), data for post-hoc tests were pooled across female presence treatments. Post-hoc tests in (C) were completed separately for the different female treatments due to a significant interaction between female presence and diet when controlled for male density (Table 1). In (B and C), asterisks represent significant differences between female presence treatments within the low-diet. Growth (mm) O Condition Development Time (Days) ^* o o o o o o o -a- -»• N> NJ CO W J^ b b b b b b b ocnocnocnocno O) 5 N> O N> •£• 0> I I I i i I II I I II I

03

O 88

SUPPLEMENTARY DATA

Supplemental Methods

Laboratory experiment

Spiders were outbred offspring of individuals collected in Perth (2000) and Sydney,

Australia (2001, rearing protocols in [21]). Spiderlings were held in separate 2x2x3 cm plastic cages and fed Drosophila sp. twice weekly. For our development experiment, 212 penultimate

(4th) instar males (identified by developing copulatory organs [22]) were removed from the population and reared in the presence or absence of females and their webs, on one of three diets

(high, mid, or low). High and mid-diet males were fed 3 times per week (6 and 3 Drosophila, respectively); low-diet males received 1 Drosophila each week. In the female-present treatment, each male's cage was surrounded by 4 web-building virgin females (each in their own cage).

There was no direct or visual contact between males and females or their webs, but cages were porous to allow the passage of airborne pheromones. In the female-absent treatment, males were kept in a similar room, adjacent to cages of other males, but with no female spiders present. In both treatments, the number of neighbouring males varied throughout the experiment due to the constant addition of newly moulted penultimate males and the removal of mature males from treatments. Temperature and light cycle was the same for all males. Males were monitored daily and date of adult moult noted. We measured the length of the patella-tibia of each male's two front legs at the penultimate instar and the adult stage using digital images and measurement software (Simple PCI, Compix Inc. Imaging systems, 2002). Adult males were also weighed

(Ohaus explorer balance accurate to 0.1 mg), then returned to their cages.

Male size was the average of the two leg measurements, and growth was the change in size after males were placed into a treatment, adjusted for variation in penultimate size ([adult size - penultimate size] / penultimate size). We ensured that variation in our measure of growth was 89 not biased by pre-experimental differences in male size (i.e., size at the penultimate instar) (Fi,

210= 2.44, P = 0.12).

Male body condition was estimated using residuals of log (cubed root of male weight) regressed on log (size) (see below). We analyzed the data using a MANOVA and a three-way

ANCOVA with (i) growth, (ii) adult body condition, and (iii) development time as the dependent variables, and (a) feeding treatment, (b) female presence/absence as the independent variables, and (c) average number of neighbouring males as a covariate for each male. We examined differences between diet treatments using a Tukey-Kramer HSD post-hoc test [23].

Field data

We tested our prediction that size and body condition of adult males should decrease as the distance to the nearest potential mate decreased in the field [field site described in 14]. We monitored solitary penultimate instar males (found on the webs on which they developed) daily in the field and weighed and measured them (as above) on the day they matured. We recorded distance to the nearest adult or penultimate female's web. We only used males observed during their penultimate instar or found during or immediately following moulting to insure accurate assessment of proximity to females during development and body condition at final moult.

Estimating male body condition

There are a variety of ways to estimate body condition, but there is considerable debate about which is the most statistically appropriate or biologically relevant [24-32]. Although there is no current consensus on any one method, it is desirable in many studies to estimate the relative size of energy stores available for use by individuals [27,31-33]. Although body condition indices may not directly correlate to fat reserves as measured by lipid content, they nevertheless measure a biologically relevant trait related to fitness in many studies [27,32,33].

One recommended method is to include body size as a covariate in a general linear model analysis of variables of interest [24,30]. While this method has some advantages [24,29,30] it 90 constrains the types of analyses available, and does not yield individual indices of body condition, as is required in many studies.

One commonly used body condition index, recently found to perform well statistically and to be biologically relevant [32, 33], is the residual index, which estimates body condition as the residual from a regression of body weight on linear size [27, 32]. Although the use of residuals has been questioned [24, 29, 30], new research demonstrates that using Ordinary Least

Squares (OLS) regression satisfies all critical statistical assumptions and performs better than

Reduced Major Axis (RMA) regression [32]. The relationship between weight and size has been argued to be log linear [27], so residual indices are often based on a regression of

log^J weight on log 4 size . However, the exponent of the log relationship of weight may vary across species, and depends on how weight scales with size [25].

To determine the relationship between log (weight) and log (size) for redback spiders, we used two independent groups of males. The first group was from a lab reared population (N

= 60, Andrade, unpublished data), and the second group was from a field captured population

(N = 400, see [14]). For each group we performed a separate Model II regression (reduced major axis regression, RMA) because of the error associated with measuring both the dependant and independent variables [23, 30]. We regressed male weight on size using RMA [34] and found the average slope of the relationship across the two analyses was 3 (Group 1: 2.615 ± 0.1826,

95% CI: 2.249 - 2.989; Group 2: 3.667 ± 0.1141 ; 95% CI: 3.443 - 3.890). We ran our statistical analyses using each calculated slope value independently and using the average slope

(3), but found no qualitative differences in our results. Thus we report our analyses using the average value (3) as our best estimate of the true exponent (Table 4.1).

Statistical Analyses 91 There was no difference in the initial size of penultimate-instar males placed in

treatments (Two-way ANOVA, all P > 0.31). Variation in adult male size at the end of the

experiment was within the range of variation seen in wild-caught males (mean ± SD; this study:

2.993 ± 0.250mm, data from [14]: 2.913 ± 0.399mm) suggesting experimental diets were

reasonable simulations of natural diets. Adult size and body condition were normally

distributed.

We used a MANOVA to test for effects of diet treatment, female density, male density

(and all possible interactions) on development time, experimental growth, and body condition.

Because of the significant multivariate interaction between diet, female density, and male density

(Table 4.1), we also report separate univariate ANOVA's for each treatment to determine how

each factor influences the dependant variables.

We tested for post-hoc differences between and within diet treatments using a Tukey-

Kramer HSD post-hoc test [23]. Development time decreased and body condition at maturity

increased significantly with each increase in food availability (low to mid to high-diet, Figure

4.1 A, B). When data were examined within each diet treatment, only in the low-diet treatment

did female presence lead to a significant decrease in male body condition and increase in growth

(Figure 4.1B, C). Due to a significant interaction between diet and female presence (controlled

for male density, Table 4.1), we examined effects of diet on growth separately within each

female treatment [23]. In the absence of females, high-diet males grew more than low-diet males

(Figure 4.1C). In the presence of females, high and mid-diet males grew more than low-diet

males (Figure 4.1C). 92

SUPPLEMENTAL LITERATURE CITED

SI. Andrade, M.C.B., and Banta, E.M. (2002). Value of remating and functional sterility in

redback spiders. Anim. Behav. 63, 857-870.

S 2. Forster, L.M., and Kingsford, S. (1983). A preliminary study of development in two

Latrodectus species (Aranea: Theridiidae). N. Z. Entomol. 7, 431-438.

S 3. Sokal, R.R., and Rohlf, F.J. (1995). Biometry, 3rd Edition (New York: Freeman).

S 4. Andrade, M.C.B. (2003). Risky mate search and male self-sacrifice in redback spiders.

Behav. Ecol. 74,531-538.

S 5. Garcia-Berthou, E. (2001). On the misuse of residuals in ecology: testing regression

residuals vs. the analysis of covariance. Journal of Animal Ecology 70, 708-711.

S 6. Kotiaho, J.S. (1999). Estimating fitness: comparison of body condition indices revisited.

Oikos 87, 399-400.

S 7. Darlington, R.B., and Smuiders, T.V. (2001). Problems with residual analysis. Animal

Behaviour 62, 599-602.

S 8. Jakob, E.M., Marshall, S.D., and Uetz, G.W. (1996). Estimating fitness: a comparison of

body condition indices. Oikos 77, 61-67.

S 9. Rolff, J., and Joop, G. (2002). Estimating condition: pitfalls of using weight as a fitness

correlate. Evolutionary Ecology Research 4, 931-935.

S 10. Freckleton, R.P. (2002). On the misuse of residuals in ecology: regression of residuals

vs. multiple regression. Journal of Animal Ecology 71, 542-545.

S 11. Green, A.J. (2001). Mass/length residuals: Measures of body condition or generators of

spurious results? Ecology 82,1473-1483.

S 12. Marshall, S.D., Barrow, J.H., Jakob, E.M., and Uetz, G.W. (1999). Re-estimating fitness:

can scaling issues confound condition indices. Oikos 87,401-402. 93 S 13. Schulte-Hostedde, AX, Zinner, B., Millar, J.S., and Hickling, G.J. (2005). Restitution of

mass-size residuals: Validating body condition indices. Ecology 86,155-163.

S 14. Schulte-Hostedde, AX, Millar, J.S., and Hickling, G.J. (2001). Evaluating body

condition in small mammals. Canadian Journal of Zoology 79, 1021-1029.

S 15. Bohonak, A.J., and van der Linde, K. (2004). RMA: Software for Reduced Major Axis

regression, Java version. Website: http://www.kimvdlinde.com/professional/rma.html. 94 Supplemental Table 4.1: Distance of the nearest penultimate or adult female affects adult male size and condition in the field. Values in bold are significant.

Source

Adult Size Collection date 3.74 0.059

Distance 8.96 0.0004

date x Distance 1.17 0.28

Condition Collection date 0.49 0.48

Distance 5.14 0.028

date x Distance 0.008 0.937 95

CHAPTER FIVE

A change in competitive context reverses sexual selection on male size

Michael M. Kasumovic & Maydianne C. B. Andrade

Integrative Behaviour and Neuroscience Group, University of Toronto at Scarborough, Toronto,

ON, MIC 1A4, Canada ABSTRACT

There is a common view in behavioural ecology that large size increases male fitness, with numerous studies demonstrating fitness benefits of large size within single competitive contexts. However, in the Australian redback spider (Latrodectus hasselti), despite ample resources, males develop as small adults when reared in the presence of females—trading-off size for a faster development time. In this study, we demonstrate that this developmental plasticity in response to female availability is adaptive under some competitive contexts. As male redbacks can experience a broad spectrum of competitive challenges, phenotype-fitness associations may change in different environments. When females are abundant, rapid development and mate searching can ensure a male is the first to reach a virgin female

(protandric competition). In contrast, when females are sparse, males may have to compete with up to 5 other rivals after arriving at a female's web (direct competition). We simulated these contexts within field enclosures and demonstrated that relatively smaller males lose when competing directly for females, but have ten times higher fitness than larger males when given the temporal advantage in competition that is derived from their more rapid development. We calculated linear selection gradients for each type of competition and confirm sexual selection in favour of small male size under protandric competition only. Male fitness is therefore not determined only by body size or body condition, but by compatibility with the competitive environment, suggesting that the life-history traits associated with the development of the phenotypic traits under examination are also important. Our results demonstrate the adaptive benefit of plasticity and the importance of contextual examinations of phenotypes. 97

INTRODUCTION

Studies of sexual selection across taxa demonstrate that male fitness can depend on the

ability to outcompete rivals (Andersson 1994). Excluding systems with discrete alternative phenotypes (Moran 1992), the traits most often considered to maximize fitness under competition are large size and weaponry (Blanckenhorn 2005). Demonstrations of positive directional selection on these traits in species with unimodal phenotypic distributions are common (Blanckenhorn 2005; Kingsolver et al. 2001) with only a few studies demonstrating a benefit of being small (e.g., a locomotory advantage in three-dimensional courtship displays;

Crompton et al. 2003). Even in species where protandrous males are smaller than females, relatively larger males often have higher net fitness (Candolin & Voigt 2003; Foellmer &

Fairbairn 2005). Examples of a large size benefit are so overwhelming that it has been argued that an understanding of "these mechanisms.. .no longer need special attention" (Blanckenhorn

2005, pg. 1004).

Although a large size benefit is widespread, it is important to understand the effects of life-history traits associated with the development of large size because of the potential trade-off between adult size and timing of maturity (Roff 1992; Stearns 1992). Maturing with the appropriate phenotype at an inappropriate time may lead to a decrease in fitness since the phenotype may not match the environmental challenges at the time of maturity. It may therefore be critical that development time also be considered, as the ontogenetic pathway required to attain a particular phenotype under a given set of conditions may also affect male performance relative to rivals. For example, if selection on phenotypes varies temporally and spatially within a season (Caruso et al. 2003; Jann et al. 2000; Punzalan 2007; Chapter 6), individuals may shift their ontogeny to ensure the development of a phenotype that is optimal to the most relevant competitive challenges that will be faced at maturity (i.e., developmental plasticity, Pigliucci

2001; West-Eberhard2003). 98 There are many examples of developmental plasticity in life history and metric traits

(reviewed in Agrawal 2001; Nylin & Gotthard 1998; Pigliucci 2001; West-Eberhard 2003).

However, examples of adaptive developmental plasticity are fewer as this requires demonstrating phenotype-fitness associations in the various environments that lead to developmental shift (see Doughty & Reznick 2004 for a review of the necessary criteria).

Without such examinations, these developmental shifts may simply be passive biophysical responses to environmental variation (Doughty & Reznick 2004; Meyers & Bull 2002; Pigliucci

2001). Furthermore, the majority of examples of adaptive developmental plasticity are in response to environmental variation (e.g., temperature, food availability; reviewed in Doughty

& Reznick 2004; Pigliucci 2001; West-Eberhard 2003). In contrast, relatively fewer studies have demonstrated adaptive developmental plasticity in response to population density or sex ratio that may reflect variation in the strength or direction of selection (e.g., Gage 1995;

Kasumovic & Andrade 2006; Stockley & Seal 2001; Tan et al. 2004; Walling et al. 2007). Most of these studies are in response to tactile cues of population density (Gage 1995; Stockley &

Seal 2001; Tan et al. 2004), while examples outside of tactile cues are even more rare (visual cues: Walling et al. 2007). However, demonstrations of the adaptive nature of developmental plasticity in response to population demographics are necessary as these forms of plasticity may provide a link between the complex dynamics associated with demography, ecology, and evolution to that will result in a better understand of the evolution of phenotypes and phenotype- fitness associations (Benton et al. 2006; Kokko & Lopez-Sepulcre 2007).

Here, we use the Australian redback spider {Latrodectus hasselti) to demonstrate that the developmental plasticity males show in response to variation in the presence of females

(Kasumovic & Andrade 2006) is adaptive when tested within the environment that led to the development of the phenotype. Redback males alter their development time in response to pheromones that provide cues of the local density of virgin females and rival males. These cues 99 indicate the type of competitive challenges males are likely to face upon maturity (Kasumovic &

Andrade 2006), and thus whether rapid development or larger body size are likely to be

favoured by selection. For example, at high population densities, where males are likely to

detect virgin females nearby, rapid development may be favoured to ensure mating priority

given the high first-male sperm precedence in this species (Snow et al. 2006; Snow & Andrade

2005). Consistent with this scenario, redback males mature more rapidly (on average two days

earlier) at a smaller body size and lower body condition at high female densities (Kasumovic &

Andrade 2006).

Redback females overwinter at several developmental stages, and develop much more

slowly than males (Forster & Kingsford 1983). As a result, each maturing cohort of males in a

given area exists within a different array of potential mating opportunities and risks of direct

competition. Local density and mate availability can change rapidly throughout the season as

individuals mature and mate (Andrade & Banta 2002). At a given time in a given microhabitat, time to maturity for a focal male is affected by temperature (Forster & Kingsford 1983) and resource availability, as well as the local density of virgin females and adult males (Kasumovic

& Andrade 2006). The competitive environments that males can experience thus range from

intense multi-male competition for access to females (up to 6 males have been recorded on a

female's web, Andrade 1996) where larger males are competitively superior (Stoltz et al.

2007a), to intense selection for rapid location and mating with virgin females due to sperm precedence patterns in this species (Snow et al. 2006; Snow & Andrade 2004; Snow & Andrade

2005).

There are currently no data on the relative success of smaller redback males as a function of their faster development in mate searching. Such data are necessary to determine whether the developmental plasticity in response to conspecific pheromones (Kasumovic & Andrade 2006) is adaptive. Here we tested the prediction that rapidly developing (small) males have increased 100 fitness when the sex ratio is equal or female-biased, but that larger males have increased fitness

when the sex ratio is male-biased. More specifically, we predicted that earlier maturity would

allow smaller males to reach virgin females and mate first, increasing their fitness relative to

larger (slower developing) males (Kasumovic & Andrade 2006). In contrast, when the sex ratio

is male-biased and direct competition is likely to be common, male redbacks develop more

slowly but are larger and better provisioned (i.e. increased body condition) as adults

(Kasumovic & Andrade 2006). As in other systems, we predicted that larger size would increase

male success in direct inter-male competition (Andersson 1994). Thus, if the developmental

plasticity shown by male redbacks (Kasumovic & Andrade 2006) is adaptive, we predict a

switch in the competitive context (direct competition over females versus a protandry

advantage) should lead to a switch in the phenotype that maximizes fitness, and therefore, a

reversal in the direction of selection on body size. To test this, we examined mate searching and

copulatory success of large and small males released in field enclosures. We examined the

outcome of direct competition over mates (simultaneous release treatment) and protandric

competition (with smaller males initiating mate search sooner; staggered release treatment). We

also calculated linear selection gradients on size and body condition based on known

relationships between paternity and the pattern and frequency of mating (Snow & Andrade

2005).

MATERIALS AND METHODS

Natural History

In nature, redback males mature at a broad distribution of size and body conditions throughout the breeding season (Andrade 2003). As juveniles, individual males inhabit their own webs, which are surrounded by conspecifics of a range of ages (Kasumovic and Andrade pers. obs.). In redbacks, males are limited to mating with a single cannibalistic female (Andrade

1996; Forster 1992). Females on the other hand, are known to mate with a maximum of three males, although most apparently mate with one or two (Andrade 1996). As first-mating males

achieve sperm precedence by depositing a sperm plug (Snow & Andrade 2005), mated females

have low reproductive value for males. Mated females no longer produce attractive pheromones

one day after mating (Stoltz et al. 2007b) and virgin females are preferred by mate-searching

males, although males do sometimes settle with previously-mated females (Andrade &

Kasumovic 2005),

The timing of male maturity as well as adult body size may be important to male fitness.

Maturing after nearby females have already mated reduces a male's chance of successfully

finding virgin females in an already costly search (80% of males perish while mate searching;

Andrade 2003). However, maturing rapidly at small size in the presence of direct competition

over few females may also be costly as relatively larger males are competitively superior (Stoltz

et al. 2007a). Redback males can therefore experience a natural gradient of selection pressures

resulting from competitive environments that range from single males on webs to contests with

up to five rivals (Andrade 1996). Males apparently use pheromonal cues of the local density of

males and females as a trigger for developmental plasticity, and this yields phenotypes predicted

to be optimal under each competitive condition (Kasumovic & Andrade 2006).

Collecting and housing spiders

Redback spiders were collected from the campus of Macquarie University in Sydney,

NSW, Australia. Males were collected as adults or in the penultimate instar (one moult from maturity, stage determined by copulatory organ development; Forster & Kingsford 1983) and reared to maturity in the lab in 2x2x3 cm clear plastic cages. We ensured that all males

collected as adults had intact copulatory organs and thus were virgins (copulatory organ morphology changes with mating, see Snow et al. 2006; Snow & Andrade 2005). We did not use males until five days after capture or final eclosion to insure they had induced sperm into their palps (Foelix 1982). We collected females as juveniles or adults. To distinguish between 102 mated and virgin adult females, we housed females in the laboratory for a month prior to the experiment, fed them ad libitum and checked them daily for egg sac production (females survive up to 2 years in the laboratory, Andrade & Banta 2002). Mated females usually produce an egg sac within two weeks under such a feeding regime (Kasumovic pers. obs.). All spiders were held in individual plastic cages and fed Drosophila melanogaster (male spiders), or appropriate sized

Acheta domesticus (female spiders). We weighed all spiders before experiments began

(±0.1mg), and measured size as the average length of the patella-tibia of the two front legs. We estimated condition as the residuals of the log cubed root of weight on log size after estimating the allometric coefficient (3) using another population of spiders (see Jakob et al. 1996;

Kasumovic & Andrade 2006; Schulte-Hostedde et al. 2005 for greater details).

For all experimental trials, females were housed in aluminum cylinders 15 cm in length and 10 cm in diameter (with open ends covered in plastic wrap) for a minimum of one week to allow the construction of large webs. Males were separated into two size classes to mimic early and late developing males. As redback males show a large distribution of continuous variation in weight and size (CV weight=35.5%, CV size=13.7%; calculated from Andrade 2003), we ensured that all small males (mean=2.67mm, S.D.=0.24) were a minimum of one standard deviation smaller in size than all large males (mean=3.03mm, S.D.=0.25). Males from each size class were randomly assigned to two treatments (see below). There was no difference in the proportion of field-collected compared to lab-reared adult males placed in either size class

(Fisher's Exact two-tailed test, P=0.77) or treatment group (Fisher's Exact two-tailed test,

P=0.77).

We painted adult males with individual colour markings using non-toxic fluorescent paint (Luminous paint, BioQuip Products) on the tibia of each front leg, and on the abdomen to allow individual identification during the experiment. Males were marked the day before a trial 103 by placing them between two pieces of fibreglass screening under a dissecting microscope to

minimize movement.

Experimental procedure

All trials were completed in 3x3x2 m screened outdoor enclosures in the Fauna Park at

Macquarie University. This field enclosure size is reasonable given typical web distributions in

the field (nearest neighbour distances between conspecifics is 2.40 - 3.40m, Andrade 2003) and

average male mate-search distances in the field (~3m, Andrade 2003, Figure 5.3). Each trial

consisted of a simultaneous release where all males competed simultaneously with one another

(simultaneous release treatment) to simulate an environment where a large male phenotype

would be developed, and a treatment where we delayed the release of larger males by one day

(staggered release treatment) to simulate an environment where a small male phenotype would

be developed (Kasumovic & Andrade 2006). As smaller males develop on average two days

earlier than larger males (Kasumovic & Andrade 2006), releasing smaller males only one day

earlier in the staggered release treatment provided a conservative experimental treatment where

we could examine the effect of the temporal advantage smaller males would gain due to their

faster development. Both treatments were run simultaneously in individual enclosures separated

by 20 m. There were six adult females (four virgin and two mated), six juvenile females, and six

males (three large and three small) in each treatment. Juvenile and mated females were included

to simulate distributions found in nature (Andrade 2003). We completed four replicates of paired trials (staggered release and simultaneous release treatments).

To simulate these natural distributions, the cylinders of adult and juvenile females were

randomly (but equally) distributed through four columns and three rows set up in one half of the

enclosure (Figure 5.1). The openings to the cylinders were parallel to the wind direction to allow

effective transmission of pheromones. Males were placed in the centre of the opposite end of the

enclosure, downwind of females. Both males and females were placed in their individual containers within the enclosure a day before the trial began to allow individuals to acclimate to

environmental conditions. Each spider was used only once, except for juvenile females, which

were used again as virgin adults once they matured (total females, M=74).

On the first day of the trial, we removed the plastic wrap from each female's cylinder

one hour before dusk (redbacks are nocturnal). We released males from their cages once all

females became active, typically with web construction/repair. In the staggered treatment, we

released only the small males on the first day. After releasing the small males, we checked all

female's webs in the first hour then every half hour after release using a black light to illuminate

males' markings. We noted male arrival, number of matings, and mating order. When males and

females became inactive at dawn, we covered all the cylinders, enclosing all individuals within.

On the second day, we once again removed the plastic wrap from the ends of the cylinders one

hour before dusk. We then released the three larger males simultaneously after females became

active and again observed male arrival, the number of matings, and mating order until dawn.

We followed the same procedure in the simultaneous release treatment except that both

small and large males were released simultaneously, and no additional males were released on

the second day. All males began searching for females immediately after release; we never

observed inter-male interactions during searching. Mated females were monitored after the experiment to ensure they produced viable eggs.

As copulation duration in redback spiders ranges from 6 to 31 minutes (Andrade 1996),

copulations may have occurred between our scans. However, males typically do not abandon

female's webs once they arrive (Andrade 2003; Andrade & Kasumovic 2005; Kasumovic &

Andrade 2004), and they are usually injured or killed during copulation (Andrade 1996;

Andrade 1998) so we would have found dead males in or under females' webs even if we missed a copulation. Moreover, courtship lasts a median of 4.6hr (Andrade 1996), which reduces the likelihood we missed any copulations. Finally, the number of copulations achieved 105 by males (0,1 or 2) was later confirmed by examining male copulatory organs under a dissecting microscope. Males lose the terminal sclerite from their intromittent organ at copulation; this sclerite acts as a sperm plug inside the female genitalia (Snow et al. 2006).

Males have paired copulatory organs that are functionally sterile after they have copulated once with each organ (Andrade & Banta 2002), so exannning sclerite loss provides a good estimate of male copulatory success. We assessed male reproductive success with a single female as male redbacks are limited to mating with only one female in nature (Andrade 2003; Andrade & Banta

2002).

Paternity and calculation of selection gradients

We predicted each male's paternity based on the number and order (determined by observations) of copulations with a given female and known sperm use patterns in redbacks

(Snow et al. 2006; Snow & Andrade 2005). Females have paired, independent sperm storage organs (spermathecae), each of which is inseminated by one of the male's two pedipalps. Sperm are apparently released equally from the two spermathecae and mix at fertilization (Snow et al.

2006; Snow & Andrade 2005), and the first male to inseminate both spermathecae deposits a sperm plug which ensures 89% of the fertilizations if a rival male mates second (Snow et al.

2006; Snow & Andrade 2005). We used these empirical results to predict paternity in the following way. First-mating males were assigned 89% paternity if they copulated with both organs and a rival also mated with both organs. If the rival mated with only one spermatheca, the first male's paternity was assigned 94.5% (1/2 x [100% from one organ + 89% from the second organ]). If the first male mated with only one spermatheca, paternity would be shared equally (50%) with a later-mating rival that mated once (second-mating males always copulate with the empty spermatheca, Snow & Andrade 2005). If the first male copulated once but the second male copulated twice, paternity of the first male would be 44.5% (1/2 x 89% from one organ -+- 0% from the second organ). We assigned 0% paternity to males that were the third to 106 copulate with a given female, or that did not copulate. We assigned 100% paternity if a female mated with only one male. For each scenario above, the paternity of the second male would be

100% - (paternity of the first male).

We calculated selection gradients to determine whether the strength and direction of selection on male phenotype differed between treatments (Lande & Arnold 1983). We used estimated paternity as our measure of relative reproductive success and used mean-standardized values for both male body size and body condition. In each treatment, we calculated selection gradients using a multiple linear regression of relative fitness on size and condition (Lande &

Arnold 1983).

RESULTS

Male success

A total of 68.8% (33/48) of males successfully located females across all trials. All males that located females during the 2-day experiment did so within the first hour following their release. There was no difference in body condition (residual mass, see methods) between large and small males (t=0.25, d.f.=46, P=0.81), males used in either treatment (t=0.38, d.f.=46,

P=0.71), or males that did and did not successfully find females (t=-l.l 1, d.f.=46, P=0.27).

More males found females in the direct competition treatment (20/24 males) than in the scramble competition treatment (13/24 males; %2=3.72, P=0.05). This difference was due to the substantially reduced searching success of larger males in the scramble treatment, where only two large males located a female's web (Fisher's Exact one-tailed test<0.001; Figure 5.2). In comparison, searching success of small and large individuals was equal in the direct competition treatment (Fisher's Exact two-tailed test=0.70; Figure 5.2).

Males had a significant preference for settling on the webs of virgin females (%2=36.88,

PO.0001, Figure 5.1). However, while searching for virgin females, three males passed through the cylinders containing juvenile females, with a single male settling on a web of a juvenile 107 female. No males settled on the webs of mated adult females. Although there was no difference in the size (t= -0.784, d.f.=32, P=0.44), weight (t= -1.113, d.f.=32, P=0.27), or condition (t = -

0.907, d.f.=32, P=0.37) between virgin females that did and did not attract males, males settled on the closest virgin females (Front=l 1/13, Middle=3/8, Back=5/12; x2=5.97, P=0.05). This resulted in 16 of the 33 virgin females remaining unmated due to their location, despite the fact that the effective sex-ratio was male biased. All females observed to mate during trials later produced viable eggs.

Of the 33 males that successfully found females in both treatments (Figure 5.2), 24 mated, and 23 of these mated within 6 hours of reaching a female. We examined whether mating success of males was related to size within each treatment. In the simultaneous release treatment, large and small males were equally likely to copulate at least once (Fisher's Exact one-tailed test, P=0.33; Figure 5.3a), but larger males were more likely to mate first (Fisher's

Exact one-tailed test, P=0.05; Figure 5.3b). In contrast, in the staggered release treatment, smaller males were more likely than larger males to copulate at least once (Fisher's Exact one- tailed test, P=0.01; Figure 5.3a) and first (Fisher's Exact one-tailed test=0.002; Figure 5.3b). We used a two-way analysis of variance (ANOVA) to examine whether estimated paternity differed between small and large males in each treatment. There was no difference in average paternity of small and large males (Fj, 44=1.33, P=0.26). The simultaneous release treatment had a significant effect on male paternity (F3;44=4.98, P=0.031) likely due to the decrease in average paternity of large males in the staggered release treatment. There was a significant treatmentxsize interaction (Two-way ANOVA: F3j44=6.59, P=0.014) with larger males having a significantly higher predicted paternity in the simultaneous release treatment, while smaller males had a significantly higher predicted paternity in the staggered release treatment (Figure

5.4).

Selection gradients We tested for differences between standardized linear (P±S.E.) selection gradients on male size and body condition (Lande & Arnold 1983) by using an analysis of covariance

(ANCOVA) to examine whether slopes were equal (as in Caruso et al. 2003; Conner 1989). To detect differences in selection among size and body condition measurements between release treatments, we entered both body size and body condition in the model, along with treatment as the categorical variable. A significant treatmentxtrait interaction indicated that the strength of selection varied significantly in each treatment There was an overall effect of treatment on reproductive success (Fi(42=4.73, P=0.035). When males were under direct competition

(simultaneous release treatment), there was no significant selection on body size

(P=0.142±0.096, P=0.15) or body condition (H).052±0.096, P=0.59). In contrast, in the staggered release treatment, there was significant negative selection on body size (P= -

0.171±0.07, P=0.02); a significant difference from the simultaneous release treatment (body sizextreatment: Fi,42=7.04, P=0.011). Selection on body condition in the staggered release treatment was not significant (P= 0.010±0.069, P=0.88), and there was no significant difference from the simultaneous release treatment (body conditionxfreatment: Fi>42=0.12, P=0.73).

DISCUSSION

In our study, we assessed the effects of male size and body condition on mating success when we varied release times of males and thus the relative timing of male entry into the competitive arena. We have demonstrated a reversal in the relationship between male size and reproductive success as a function of the context and form of inter-male competition. When large and small males were released simultaneously, there was no difference in mate searching success, replicating the field results of Andrade (2003). However, upon arrival at female's webs, when all males competed directly, larger males had superior mating success, which resulted in higher predicted paternity. When smaller males were released one day earlier in the staggered release treatment, these smaller males had greater mating success, resulting in ten times higher predicted paternity than larger males (Figure 5.4). Thus, the plasticity in response to female presence demonstrated in a previous study (Kasumovic & Andrade 2006) was shown to be adaptive when males were tested within a developmentally realistic context (Doughty &

Reznick 2004). This result is predicted when life-history traits and developmental plasticity are both taken into consideration, yielding phenotypes that are optimum for a particular set of adult challenges (Doughty & Reznick 2004; Pigliucci 2001; Scheiner 1993; Via & Lande 1985; West-

Eberhard2003).

Variation in the type of competition (e.g., female choice vs. intrasexual competition) males encounter at maturity is not uncommon in redbacks where there is variance in development time as a result of resource availability, female availability, and male density

(Kasumovic & Andrade 2006), and males mature at different times throughout the season in the field (Andrade 2003). The competitive environments that males can experience range from intense multi-male competition for access to females (up to 6 males have been recorded on a female's web, Andrade 1996) to intense selection for rapid location and mating with virgin females before other males arrive. Although we examined two extremes of the intrasexual competition males are likely to face in nature, there is likely a continuum of variation in the relative importance of rapid development compared to large size and body condition. For example, while up to 6 males have been found per web (high likelihood of direct competition), a median of 2 males per adult female are found in nature and some virgin females are found without cohabiting males (Andrade 1996). Thus, an ontogenetic strategy that allows a trade-off between size and development time as a function of the variable competitive context is likely to maximize fitness in such a variable environment.

In the staggered release treatment, we demonstrated that rapid-developing males, despite the potential competitive handicap of their small size, reach females first, copulate first and more frequently, and can thus insert a sperm plug that ensures paternity through sperm priority 110 (Snow et aL 2006; Snow & Andrade 2005). Earlier arrival of small, rapidly-developing males also allows males to decrease post-copulatory competition in two ways. First, copulation results in a change in the composition or production of pheromones by females (Prouvost et al. 1999;

Trabalon et al. 1997), so that mated females are less attractive to other males (Andrade &

Kasumovic 2005; Gaskett et al. 2004). Second, males can decrease the release of attractive pheromones emitted from the web through web reduction (Schulz & Toft 1993; Watson 1986), and thus reduce the chance that a female will attract a second mate. This is important, as pheromones can remain on the web for several days after the female ceases production of these chemicals (Andrade & Kasumovic 2005). Our results suggest the drop in attractiveness of mated females occurs rapidly after undisturbed matings (less than 9 hours), and this may explain the failure of larger males to find potential mates in the staggered release treatment. In our trials, males were never attracted to previously mated females on the first day, and on the second day, only one of the newly-mated females in the staggered release treatment attracted a male. This one exception was also the only case in which the female remained unmated until the second day of the experiment (the first-arriving male did not copulate immediately).

Although small males outperformed large males in our staggered release treatment, there are contexts in which early development is not advantageous because, in redbacks, it is coupled with smaller body size (Kasumovic & Andrade 2006). The larger male phenotype is thus maintained in redbacks because it offers a benefit when the probability of rival males clustering on females' webs is high. In our trials, larger males excluded smaller males, and had increased mating success in direct competitions. So, as in other systems, increased size is favoured when the likelihood of direct competition increases (for a review see Andersson 1994; Kingsolver &

Pfennig 2004). This is likely in situations when the sex ratio is male-biased. However, even if the population-wide sex ratio is equal, local fluctuations in the density of receptive mates can lead to males clustering with available females (Foellmer & Fairbairn 2005; Gwynne et al. Ill

1998; Kasumovic et al. 2007). Thus, local variation in sex ratio and density will likely cause spatial and temporal variation in the benefit of large size. Larger males might also have higher survivorship during mate searching over the long distances expected when females are sparsely distributed, although there is currently limited empirical support for this idea (Andrade 2003;

Foellmer & Fairbairn 2005; Kasumovic et al. 2007; Vollrath 1980).

Our estimates of linear selection gradients also varied depending on the competitive context. Although our sample size was relatively small, resulting in limited statistical power to detect selection of 'typical' strength (see Kingsolver et al. 2001), the direction and magnitude of selection on the focal traits differed significantly as predicted. In the simultaneous release treatment, selection on body size was positive (but not significant). In contrast, when protandry was taken into consideration, there was significant negative selection on body size, which can be interpreted as selection for rapid development. Furthermore, the change between treatments was significant. Since microspatial variation in competitive challenges can significantly alter the sign and magnitude of selection on size, single-context or whole-season and population-wide estimates of the strength of selection without taking into consideration the multiple competitive contexts that can potentially occur in nature can provide a misleading impression of how sexual selection is affecting phenotypic evolution. For example, biases in the sex ratio on a very local scale, similar to the scale of our field enclosure (roughly 3m2) may be sufficient to switch selection, as this area is comparable to the distance searched by the average successful redback male (approximately 2.5 to 3.5m, Andrade 2003).

In many species, the presence of multiple competitive challenges has led to the evolution of different morphs (i.e., polyphenisms) that allow individuals to specialize in the different strategies suited to each context (reviewed in Moran 1992). In redback spiders however, rather than discrete morphs, males show plasticity resulting in continuous (quantitative) variation in adult size and body condition as a function of the density of neighbouring males and females 112 (Kasumovic & Andrade 2006), which indicates the most relevant competitive challenges. The rapid response of males to short-term variation in competitive context is most likely because males are limited to mating with a single female (Andrade 1996; Andrade 2003; Forster 1992), so optimize their phenotype to match the single context they are likely to encounter at maturity.

Such a developmental strategy acts via variation in time to maturity, which yields significant variation in male size and body condition (mirroring variation seen in nature). Thus, despite size-mediated competitive interactions (simultaneous release results), no single size optimum exists for redback males. We provide further evidence that sexual selection can favour context- specific changes in development rather than specialization in environments where temporally and spatially varying competitive challenges within a breeding season are predictable (van

Tienderen 1991; Via & Lande 1985).

This study suggests that developmental plasticity can lead to phenotype-environment matching that ensures phenotypes are closely coordinated to local competitive challenges

(Kasumovic & Andrade 2006). This explanation is consistent with evidence from other studies of developmental plasticity in size at maturity in response to population density (e.g. Gage 1995;

Johansson et ai. 2001; Stockley & Seal 2001; Tan et al. 2004). Developmental plasticity is likely to be especially important in species where environments are temporally and spatially heterogeneous between generations, as offspring would rarely experience the same competitive environment as their parents (e.g., Danielson-Francois et al. 2005; Greenfield & Rodriguez

2004; Kokko & Heubel 2007). In such a situation, a highly heritable phenotype would lead to lower fitness in the next generation when the environment changes. Thus, when genotype by environment interactions play a significant role in adult condition, lower heritability of quantitative traits and higher heritability of developmental plasticity to respond to a heterogeneous environment may be adaptive. 113 Larger size may signal male quality to females in some systems (Andersson 1994;

Blanckenhorn 2005), however females have been found to mate with smaller (less dominant) males in a range of species (Qvamstrom & Forsgren 1998). This behaviour is puzzling if such females are copulating with poorer quality males. However, in redbacks, by mating with the first male to arrive (irrespective of size), females are in fact selecting the males most able to tactically respond to the inherent heterogeneity of the environment. In situations where multiple males also compete for access to females, size may play an important role since it allows success in direct competitive interactions. Definitions of male quality should therefore include life-history decisions associated with the development of the traits under assessment.

Competitive contexts are likely to vary throughout a breeding season as a consequence of variation in the availability of mature females and the number of potential competitors (Kokko

& Monaghan 2001; Kokko & Rankin 2006). To assess the effects of selection on male traits then, requires knowledge of the range of competitive challenges that males may face (i.e., how demography affects selection), the phenotypes that will maximize fitness in each situation, and how quickly and over what spatial scale competitive contexts are likely to fluctuate (Benton et al. 2006; Kokko & Ldpez-Sepulcre 2007; Metcalf & Pavard 2006).

ACKNOWLEDGEMENTS

We would like to thank M. Herberstein for graciously providing a place to call home while in

Australia and for logistical support, P.W. Taylor and C.S. Evans for access to outdoor enclosures, and W. Blanckenhorn, J.M. Brandt, R. Brooks, D.O. Elias, D. Gwynne, K.A. Judge,

H. Kokko, F.H. Rodd, D. Punzalan, L. Rowe and the Integrative Behaviour & Neuroscience

Group (UTSC) for comments on the manuscript. This project was funded by an NSERC PGS B and an OGS to MMK and grants from NSERC, CFI and OIT to MCBA. 114

LITERATURE CITED

Agrawal, A. A. 2001 Ecology - Phenotypic plasticity in the interactions and evolution of

species. Science 294, 321-326.

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

Andrade, M. C. B. 1996 Sexual selection for male sacrifice in the Australian redback spider.

Science 271, 70-72.

Andrade, M. C. B. 1998 Female hunger can explain variation in cannibalistic behavior despite

male sacrifice in redback spiders. Behavioral Ecology 9, 33-42.

Andrade, M. C. B. 2003 Risky mate search and male self-sacrifice in redback spiders.

Behavioral Ecology 14, 531-538.

Andrade, M. C. B. Sc Banta, E. M. 2002 Value of remating and functional sterility in redback

spiders. Anim. Behav. 63, 857-870.

Andrade, M. C. B. & Kasumovic, M. M. 2005 Terminal investment strategies and male mate

choice: Extreme tests of Bateman. Int. Comp. Biol. 45, 838-847.

Benton, T. G., Plaistow, S. & Coulson, T. M. 2006 Complex population dynamics and complex

causation: devils, details and demography. Proc. R. Soc. Lond. B 273, 1173-1181.

Blanckenhorn, W. U. 2005 Behavioral causes and consequences of sexual size dimorphism.

Ethology 111, 977-1016.

Candolin, U. & Voigt, H.-R. 2003 Size-dependent selection on arrival times in sticklebacks:

why small males arrive first. Evolution 57, 862-871.

Caruso, C. M., Peterson, S. B. & Ridley, C. E. 2003 Natural selection on floral traits of Lobelia

(Lobeliaceae): Spatial and temporal variation. American Journal of Botany 90, 1333-

1340.

Conner, J. K. 1989 Density-dependent sexual selection in the fungus beetle, Bolitotherus

cornutus. Evolution 43,1378-1386. 115 Crompton, B., Thomason, J. C. & McLachlan, A. 2003 Mating in a viscous universe: the race is

to the agile, not to the swift. Proceedings of the Royal Society B-Biological Sciences

270, 1991-1995.

Danielson-Francois, A. M., Kelly, J. K. & Greenfield, M. D. 2005 Genotype x environment

interaction for male attractiveness in an acoustic moth: evidence for plasticity and

canalization. Journal of Evolutionary Biology 19, 532-542.

Doughty, P. SL Reznick, D. N. 2004 Patterns and analysis of adaptive plasticity in animals. In

Phenotypic Plasticity: Functional and Conceptual Approaches (ed. T. J. DeWitt & S. M.

Schemer), pp. 126-150. New York: Oxford University Press.

Foelix, R. 1982 Biology of spiders. Cambridge: Harvard University Press.

Foellmer, M. W. & Fairbairn, D. J. 2005 Competing dwarf males: sexual selection in an orb-

weaving spider. Journal of Evolutionary Biology 18, 629-641.

Forster, L. M. 1992 The stereotyped behaviour of sexual canmbalsim in Latrodectus hasselti

Thorell (Aranea: Theridiidae), the Australian redback spider. Australian Journal of

Zoology 40,1-11.

Forster, L. M. & Kingsford, S. 1983 A preliminary study of development in two Latrodectus

species (Aranea: Theridiidae). N. Z. Entomol. 7,431-438.

Gage, M. J. G. 1995 Continuous variation in reproductive strategy as an adaptive response to

population density in the moth Plodia interpunctella. Proc. R. Soc. Land. B 261,25-30.

Gaskett, A. C, Herberstein, M. E., Downes, B. J. & Elgar, M. A. 2004 Life-time male mating

preferences in a sexually cannibalistic orb-web spider (Araneae: Araneidae). Behaviour

141,1197-1210.

Greenfield, M. D. & Rodriguez, R. L. 2004 Genotype-environment interaction and the reliability

of mating signals. Animal Behaviour 68,1461-1468. 116 Gwynne, D. T., Bailey, W. J. & Annells, A. 1998 The sex in short supply for matings varies

over small spatial scales in a katydid (Kawanaphila nartee, Orthoptera : Tettigoniidae).

Behavioral Ecology and Sociobiology 42, 157-162.

Jakob, E. M., Marshall, S. D. & Uetz, G. W. 1996 Estimating fitness: a comparison of body

condition indices. Oikos 77, 61-67.

Jann, P., Blanckenhorn, W. U. & Ward, P. I. 2000 Temporal and microspatial variation in the

intensities of natural and sexual selection in the yellow dung fly Scathophaga

stercoraria. Journal of Evolutionary Biology 13, 927-938.

Johansson, F., Stoks, R., Rowe, L. & De Block, M. 2001 Life history plasticity in a damselfly:

Effects of combined time and biotic constraints. Ecology 82, 1857-1869.

Kasumovic, M. M. & Andrade, M. C. B. 2004 Discrimination of airborne pheromones by mate-

searching male western black widow spiders (Latrodectus hesperus): species- and

population-specific responses. Can. J. Zool. 82,1027-1034.

Kasumovic, M. M. & Andrade, M. C. B. 2006 Male development tracks rapidly shifting sexual

versus natural selection pressures. Current Biology 16, R242-243.

Kasumovic, M. M., Bruce, M. J., Herberstein, M. E. & Andrade, M. C. B. 2007 Male mate

choice may increase the cost of mate searching in the golden orb-web spider (Nephila

plumipes). Behavioral Ecology 18, 189-195.

Kingsolver, J., Hoekstra, H., Hoekstra, J., Berrigan, D., Vignieri, S., Hill, C, Hoang, A., Gibert,

P. & Beerli, P. 2001 The strengths of phenotypic selection in natural populations.

American Naturalist 157, 245-261.

Kingsolver, J. & Pfennig, D. 2004 Individual-level selection as a cause of Cope's rule of

phyletic size increase. Evolution 58,1608-1612.

Kokko, H. & Heubel, K. 2007 Condition-dependence, genotype-by- environment interactions

and the lek paradox. Genetica In Press. 117 Kokko, H. & Lopez-Sepulcre, A. 2007 The ecogenetic link between demography and evolution:

can we bridge the gap between theory and data? Ecology Letters In press.

Kokko, EL & Monaghan, P. 2001 Predicting the direction of sexual selection. Ecology Letters 4,

159-165.

Kokko, H. & Rankin, D. J. 2006 Lonely hearts or sex in the city? Density dependent effects in

mating systems. Phil. Trans. R. Soc. B 361, 319-334.

Lande, R. & Arnold, S. J. 1983 The measurement of selection on correlated characters Evolution

37, 1210-1226.

Metcalf, C. J. E. & Pavard, S. 2006 Why evolutionary biologists should be demographers.

Trends Ecol Evol 22, 205-212.

Meyers, L. A. & Bull, J. J. 2002 Fighting change with change: adaptive variation in an uncertain

world. Trends in Ecology & Evolution 17, 551-557.

Moran, N. A. 1992 The Evolutionary Maintenance of Alternative Phenotypes. American

Naturalist 139,971-989.

Nylin, S. & Gotthard, K. 1998 Plasticity in life-history traits. 43,63-83.

Pigliucci, M. 2001 Phenotypic plasticity: beyond nature and nurture John Hopkins University

Press.

Prouvost, O., Trabalon, M., Papke, M. & Schulz, S. 1999 Contact sex signals on web and cuticle

of Tegenaria atrica (Araneae, Agelenidae). Archives of Insect Biochemistry and

Physiology 40, 194-202.

Punzalan, D. 2007 Evolution of sexual colour pattern dimorphism in the ambush bug Phymata

americana. Ecology Evolution and Behaviour, University of Toronto.

Qvarnstrom, A. <& Forsgren, E. 1998 Should females prefer dominant males. Trends in Ecology

& Evolution 13, 498-501. 118 Roff, D. A. 1992 The evolution of life histories: Theory and analysis. New York: Chapman &

Hall.

Schemer, S. M. 1993 Genetics and evolution of phenotypic plasticity. Annual Review of Ecology

andSystematics 24, 35-68.

Schulte-Hostedde, A. I., Zinner, B., Millar, J. S. & Hickling, G. J. 2005 Restitution of mass-size

residuals: Validating body condition indices. Ecology 86, 155-163.

Schulz, S. & Toft, S. 1993 Identification of a sexpheromone from a spider. Science 260,1635-

1637.

Snow, L. S. E., Abdel-Mesih, A. & Andrade, M. C. B. 2006 Broken copulatory organs are low-

cost adaptations to sperm competition in redback spiders. Ethology 112, 379-389.

Snow, L. S. E. & Andrade, M. C. B. 2004 Pattern of sperm transfer in redback spiders:

implications for sperm competition and male sacrifice. Behavioral Ecology 15, 785-792.

Snow, L. S. E. & Andrade, M. C. B. 2005 Multiple sperm storage organs facilitate female

control of paternity. Proc. R. Soc. Lond. B 272, 1139-1144.

Stearns, S. C. 1992 The evolution of life histories. Oxford: Oxford University Press.

Stockley, P. & Seal, N. J. 2001 Plasticity in reproductive effort of male dung flies (Scatophaga

stercorarid) as a response to larval density. Functional Ecology 15, 96-102.

Stoltz, J. A., Elias, D. O. & Andrade, M. C. B. 2007a Females penalize male competition and

reward courtship in a cannibalistic spider. Behavioral Ecology and Sociobiology In

Review.

Stoltz, J. A., McNeil, J. N. & Andrade, M. C. B. 2007b Males assess chemical signals to

discriminate just-mated females from virgins in redback spiders. Anim. Behav. In Press.

Tan, G. N., Govedich, F. R. & Burd, M. 2004 Social group size, potential sperm competition

and reproductive investment in a hermaphroditic leech, Helobdella papillornata

(Euhirudinea: Glossiphoniidae). Journal of Evolutionary Biology 17, 574-580. 119 Trabalon, M., Bagneres, A. & Roland, C. 1997 Contact sex signals in two sympatric spider

species, Tegenaria domestica and Tegenaria pagana. Journal of Chemical Ecology 23,

747-758. van Tienderen, P. H. 1991 Evolution of generalists and specialists in spatially heterogeneous

environments. Evolution 45, 1317-1331.

Via, S. & Lande, R. 1985 Genotype-environment interaction and the evolution of phenotypic

plasticity. Evolution 39, 505-522.

Vollrath, F. 1980 Male body size and fitness in the web-building spider Nephila clavipes.

Zeitschrift Tierpsychologie 53, 61-78.

Walling, C. A., Royle, N. J., Metcalfe, N. B. & Linstrom, J. 2007 Green swordtails alter their

age of maturity in response to the population level of male ornamentation. Biology

Letters 3, 144-146.

Watson, P. J. 1986 Transmission of a female sex pheromone thwarted by males in the spider

Linyphia litigiosa (Linyphiidae). Science 233, 219-221.

West-Eberhard, M. J. 2003 Developmental plasticity and evolution. New York: Oxford

University Press. Figure 5.1: A sample distribution of females m a field enclosure and the relative location of male release. Each female is inside an open cylinder in which they have built a web. Larger females are adults (black=virgins, n=4 and white=mated, n=2) while smaller females are juveniles. Female location was assigned randomly with respect to age and mating status during replicated trials. Values on the right margin are the percent (and sample size) of virgin females in each row that attracted at least one male across both treatments.

Figure 5.2: The number of small (black bars) and large (white bars) males that successfully located a female when all males were released simultaneously (direct competition treatment) or when smaller males were released first (temporal competition treatment) in a field enclosure.

Figure 5.3: The number of small (black bars) and large (white bars) males that a) successfully mated at least once versus b) those that successfully mated first insuring increased paternity.

Figure 5.4: The mean predicted paternity of small (black bars) and large (white bars) males in the direct and temporal competition treatments based on the number and order of copulations achieved. Error bars are standard error. 121

42% \y*/ ~w w \>m/ (12)

38% \**^ w \w \w (8)

85% w w w \)W (13)

/fof\ A**, 122

Simultaneous Staggered Release Treatment 123

10i

T3 .+-< iS 3 Q. O O CD

CO

o

Simultaneous Staggered

10i CO

^2 8H jo 3 64 O o J"C=D -t_» J CO E O

Simultaneous Staggered Release Treatment 124

Simultaneous Staggered Release Treatment 125

CHAPTER SIX

Spatial and temporal demographic variation drives within-season fluctuations in sexual

selection

Michael M. Kasumovic1, Matthew J. Brace2'3, Maydianne C. B. Andrade1, & Marie E.

Herberstein

'integrative Behaviour and Neuroscience Group, University of Toronto at Scarborough,

Toronto, ON, MIC 1A4, Canada

2Department of Biological Sciences, Macquarie University, New South Wales, 2109, Australia

3Current address: Behavioural Biology, Utrecht University, P.O. Box 80.086, 3508 TB, Utrecht,

The Netherlands. ABSTRACT

Our understanding of phenotypic selection in nature stems mainly from whole-season and cross-sectional estimates of selection gradients within single breeding seasons. These studies suggest that selection is relatively constant within a season, but can fluctuate rapidly between seasons. In addition, it is knowathat population demographics can fluctuate within seasons, but there is a gap in our understanding regarding the extent to which these fluctuations may cause variation in selection. Here we use two different populations of the golden orb-web spider {Nephila plumipes) that differ in population density to examine how population demographics change within a breeding season and the effect this has on the competitive challenges and strength of selection that males encounter. We calculated both the operational sex ratio and population density at three spatial scales and show that the estimation of both demographic factors depends on the scale of the examination. Sex differences at maturity and mortality likely caused the demographic fluctuations, which in turn led to changes in the competitive challenges that males encountered at different times of the season. This was demonstrated by within-season variation in the number of webs with multiple rival males. We also demonstrate significant variation in selection on male size and body condition in each population. Overall, our results demonstrate that multiple competitive challenges can exist within a single breeding season, that these can vary at small temporal and spatial scales, and that this may favour different phenotypes throughout the breeding season. We argue that studies may underestimate the true variation in selection by averaging values throughout the season, leading to misinterpretation of the influence of selection on phenotypic evolution. INTRODUCTION

The strength and direction of selection pressures acting on heritable traits can predict the evolution of phenotypic distributions (Fisher 1930; Kingsolver et al. 2001; Lande & Arnold

1983). The advent of statistical methods to quantify phenotypic selection (Arnold & Wade

1984a; Arnold & Wade 1984b; Brodie et al. 1995; Lande & Arnold 1983) has led to a better understanding of how selection shapes phenotypes through time. Using estimates of selection gradients derived from cross-sectional sampling or average fitness (longitudinal estimates) across a breeding season, a number of studies have provided information on variation in the strength of selection on a variety of traits associated with fitness (see Kingsolver et al. 2001 for a review). Comparisons of spatially separated populations (e.g., Carroll & Salamon 1995) or the same population over time (e.g., Grant & Grant 2002; Wilson et al. 2006) have led to a better understanding of how demographic factors such as the operational sex ratio (OSR) and/or population density (e.g. Blanckenhorn 1998; Jann et al. 2000) correlate to changes in the strength and direction of selection. However, the time frame over which selection is measured can have significant effects on estimates of selection, and values averaged over long periods can mask fine-scale variation that will affect selection on phenotypic distributions (Hoekstra et al.

2001; Losos et al. 2006).

The strength and direction of selection is likely to vary as a result of variation in population demographics (Kokko & Monaghan 2001; Kokko & Rankin 2006). Variation in population density can shift the strength of selection on male traits by altering conspecific encounter rates, leading to changes in female preference functions, the importance of direct competition, and thus, the degree of mating skew (Kokko & Rankin 2006; Matter 1997).

Population density can change throughout a breeding season as a function of birth, death, and maturity rates (Foellmer & Fairbairn 2004; Foellmer & Fairbairn 2005a; Kasumovic et al. 2007;

Maxwell 1998; Velez & Brockmann 2006), resource availability (Gwynne et al. 1998; Matter & Roland 2002; Wauters et aL 2001) and immigration/emigration patterns (Matter 1997).

Similarly, changes in the OSR can lead to changes in which sex is more choosy, and therefore, the direction and strength of selection on each sex (Clutton-Brock et aL 1997; Kokko &

Monaghan 2001; Simmons & Kvarnemo 2006). The OSR can also vary due to sex differences in: (1) the rate at which individuals become available for mating either through differences in maturity rates (Foellmer & Fairbaim 2005a) or arrival at breeding grounds (Grant et al. 1995;

Wiklund & Fagerstrom 1997), (2) mating optima leading to differences in the number of mating opportunities (Arnqvist & Rowe 2006; Parker 2006), and (3) immigration/emigration rates (e.g.,

Clutton-Brock et al. 1997; Wauters & Dhondt 1993). As population demographics can shift at small spatial and temporal scales, it is essential to measure within-season fluctuations in density and OSR and determine how these affect the strength and direction of selection. Such an approach will reveal the relevance of fluctuating demographics in estimations of selection, and how selection and population demographics interact (Kokko & L6pez-Sepulcre 2007).

In this study, we used populations of golden orb-web spiders (Nephila plumipes) to examine (1) whether conspecific density and OSR vary at different spatial scales within populations and throughout the breeding season, (2) whether there is spatial or temporal variation in the intensity of inter-male competition, and (3) whether selection (Brodie et al.

1995; Lande & Arnold 1983) on male body size and condition varies through the season along with demographic variation.

N. plumipes is an excellent species to address these questions because of the ease with which male phenotypic traits, mating success, and demographic variables can be assessed in the field. Females build webs alone or as part of aggregations with females of different ages (Elgar

1989; Herberstein & Elgar 1994; Kasumovic & Andrade 2007), and remain on these webs throughout the breeding season (Herberstein & Elgar 1994; Higgins 1990). Adult males are found on or near the webs of females and male competition is mediated by size (Elgar & Fahey 129 1996; Vollrath 1980). Furthermore, distance to the hub (centre) of the web is a predictor of male mating priority with larger males found closer to the hub (Elgar & Fahey 1996; Vollrath 1980) as they displace smaller males from the hub position (Vollrath 1980).

To examine how selection pressures varied through the breeding season, we surveyed two separate populations of K plumipes in New South Wales, Australia every two weeks throughout one breeding season (three surveys: early, mid, and late-season). We collected every male, and noted the number and instar of all males and females within each aggregation, and the number of females within five metres of each aggregation. We used these data to estimate density and OSR at three different levels: 1) within the aggregation, 2) within 5m of the focal aggregation (local), and 3) within each population. We estimated male reproductive success based on known relationships between the position on a female's web and paternity (see methods), and calculated selection gradients (Brodie et al. 1995; Lande & Arnold 1983) on male size and body condition at each time frame within each population to determine whether selection fluctuated with changes in demographic variables.

MATERIALS AND METHODS

Natural History

N. plumipes is a univoltine species that breeds over two months of the Australian summer (January-February). Females build webs either solitarily or as part of aggregations which are defined as groups of webs that share support strands (Elgar 1989). Males mature on their own web in or near the aggregated webs of females (Higgins 1990), leaving their web upon maturity in search of available females. Males preferentially take residence on the webs of penultimate instar or adult females (Higgins 1990; Kasumovic et al. 2007; Vollrath 1980), and wait for an opportunity to mate. As adults, male N. plumipes face two main selection pressures.

First, males are often clustered on female's webs (Kasumovic et al. 2007; Vollrath 1980) and competition with up to five rivals selects for larger size (Elgar & Fahey 1996; Vollrath 1980). 130 Second, mating first ensures the highest paternity (Elgar et al. 2003; Schneider & Elgar 2001),

likely resulting in selection for faster development in the presence of virgin females (e.g.,

Kasumovic & Andrade 2006). Also, males are limited to a single mating due to a high frequency

of fatal sexual cannibalism (-60%) and a high injury rate while mating (Elgar & Fahey 1996;

Schneider & Elgar 2001; Schneider et al. 2001). Even if males survive their first mating, a 76%

mortality rate during mate searching (Kasumovic et al. 2007) makes further matings unlikely.

Thus, to maximize their fitness, males must maximize success in a single mating, and this is

likely to be most strongly affected by their success in direct or scramble competition

(Kasumovic & Andrade 2007) rather than in avoiding cannibalism (Elgar 1992; Schneider &

Elgar 2002).

Male Collection

We surveyed two field populations of N. plumipes in Bicentennial Park, Pymble in

Sydney, NSW, Australia. The two populations were approximately 5,500m2 (North Pymble; low

density population) and 1,700m2 (South Pymble; high density population) in size and were

separated by a mowed, grassy field (170 m wide). As previous research demonstrated that males

do not cross gaps (Kasumovic et al. 2007), we considered the two populations independent of

one another. Both habitats consisted mainly of shrubs and eucalypts. Each population was

surveyed every two weeks throughout the breeding season for a total of three surveys (early, mid, late-season). Surveys were completed between 10:00 and 14:00h. For each survey, we located all N. plumipes webs at each site. We found female's webs in the same location each survey as females are mostly stationary, continually adding to their web (Herberstein & Elgar

1994; Higgins 1990). Webs of females can either occur solitarily or as part of aggregations

(Elgar 1989; Herberstein & Elgar 1994; Kasumovic et al. 2007). Web elevation varied from low to the ground in shrubs and saplings, to the upper canopy of trees (Herberstein & Elgar 1994). 131 We surveyed all webs that were below two metres in elevation. Only two aggregations above two metres were observed.

We collected all males found during these surveys, noting their distance to the hub of the web (in cm), whether there were any other adult males found on the web or within the aggregation, and the age of the female on whose web they were found (juvenile, penultimate, and adult). We aged females using the colouration and shape of the epigyne. Adult females have a protruding epigyne that has two clear openings, penultimate females have the same protrusion, but the openings are covered, and juvenile females lack a protrusion (Higgins 1992; Kasumovic et al. 2007). We also examined the age structure of females in all aggregations and within a 5m radius of each focal aggregation. We used these data to calculate three separate estimates of the operational sex ratio (OSR) and female density at the: (1) aggregation, (2) local (5m radius), and

(3) population level.

All collected males were brought into the laboratory where they were immediately weighed and measured using the average length of the patella-tibia of the two front legs as a measure of male size. We used a regression of the log (^J weight ) on log (leg length) to estimate body condition (Jakob et al. 1996; Kasumovic & Andrade 2006; Kasumovic et al. 2007;

Schulte-Hostedde et al. 2005), where x=4.22. The allometric exponent of 4.22 was estimated from the slope of an ordinary least squares regression of weight on leg size (see Kasumovic &

Andrade 2006, for details and justification of this method of estimating body condition; see

Kasumovic et al. 2007). None of the males collected were released back into the monitored field populations as male lifespan is short (5 days on average, Vollrath 1980), and it is unlikely that males collected at that time would remain on the web for the next collection.

Predicted Paternity 132 The majority of males (94%, Schneider et al. 2001) only mate with one female due to cannibalism and high injury rates while mating (Elgar & Fahey 1996; Schneider & Elgar 2001;

Schneider et al. 2001). We predicted paternity farKplumipes males found on female's webs using known sperm use patterns from two (Schneider & Elgar 2001) and three male trials (Elgar et al. 2003). Males closest to the hub mate first (Elgar & Fahey 1996), and in double mating trials, gain an average of 54% paternity (Schneider & Elgar 2001). However, in mating trials involving three males, the third male gains an average of 23% paternity, diluting the paternity of the first and second males (45%, 32%, respectively; Elgar et al. 2003). Thus, when we found two males together, we assigned 54% paternity to the male closest to the hub, and 46% paternity to the second male. If a third male was present, we assigned paternity as 45%, 32%, and 23% to the first, second and third male respectively. We assigned 0% paternity to males that were the fourth furthest from the hub and 100% paternity to a male if he was the only male on the web.

We used the predicted paternity as our estimate of fitness.

RESULTS

We collected a total of 327 males throughout the breeding season at both sites. Eight of these males were found on their own webs outside of aggregations, and seven were collected on webs from which the resident female was missing and had likely been depredated (webs were partially destroyed with large central holes). All our analyses used only the males collected from webs on which females could be identified (N=312) as males on their own webs or otherwise empty webs would have no current opportunity to mate. We collected 182 males from population 1 (North Pymble, early=28, mid=79, late=75) and located 454 females' webs

(early=88, mid=195, late=171). We collected fewer males (N=130) from population 2 (South

Pymble, early=43, mid=53, late=34), and also found fewer females' webs (N=334; early=91, mid=133,late=110). 133 We used separate two-way analysis of variances (ANOVAs) to test whether there were differences in male size and condition between populations and at each sampling time. There were significant effects of population (F5, ^6=6.84, P=0.009; Figure 6.1a) and collection time

(F5,306=16.35, P<0.0001) on male size with larger males occurring earlier in the season and in the lower density population (population 2). However, male size followed the same pattern of change in both populations (populationxtime: F5,306=0.006, P=0.99). There was also no difference in the variance in male size between populations (Bartlett's Test: F=0.996, P=0.32).

Male body condition increased through time (Fs, 305=15.49, PO.0001). There were no differences between populations (Fs, 305=2.39, P=0.12, Figure 6. lb) or between populations through time (populationxtime: F5,305=0.82, P=0.44) in male body condition.

Across the two sites, the majority of males were collected from webs of penultimate- instar and adult females (males found on webs of: juveniles=40, penultimates=142, adults=130;

X2=34.2 df=2, P<0.0001). As males prefer settling with adult and penultimate-instar females

(Higgins 1990; Kasumovic et al. 2007; Schneider et al. 2001), and only these classes of females are likely to be available for mating within a male's lifetime (Higgins 1993), we used these age classes to estimate female density and OSR throughout the season. The age structure of females in each population changed throughout the breeding season (low: x2=153.0, df=2, PO.0001, high: %2=73.6, df=2, P<0.0001; Figure 6.2), with an initial increase in the number of penultimate females (one instar prior to sexual maturity) followed by an increase in adult females.

Density Estimates

Population 1 had a lower density of penultimate and adult females (per 100m2) throughout the season (= low density population; early=0.35, mid=1.18, late=1.60) than did population 2 (= high density population; early=l .06, mid=4.19, late=3.01). As the density of available females is likely to alter selection on males, we used separate two-way ANOVAs to examine whether the population and survey date affected the density of females experienced by males at each measurement scale (aggregation or local, Figure 63). At the level of the aggregation, males in each population experienced similar average female densities overall (F5,

304=0.47, P=0.50, Figure 6.3). Densities within the aggregations changed as the season progressed (F5,304= 16.03, PO.0001), but there was no difference between populations in the pattern of change over time (F5,304=0.54, P=0.58). In contrast, at the local level, there was a significant difference in average female density between populations (F5; 303=54.33, P<0.0001,

Figure 6.3), between sampling times (F5> 303= 12.12, P<0.0001), and also in the pattern of change in density over time between populations (F5,303=14.34, PO.0001).

Variance in female availability (density) could affect optimal searching or mating strategies for males (Carroll & Corneli 1995), or may affect the adaptive value of plasticity in time to maturity (Kasumovic & Andrade 2006). Variance in female density was affected by the scale of measurement both between and within populations. There was no difference in the variance in female density between populations at the level of the aggregation (high density population: o2=2.34, low density population: a2=1.84; Bartlett's test: F=2.20, P=0.14). At the local level however, variance in female density was more than four times higher in the high density population (a2=36.87) than in the low density population (a2=9.12; Bartlett's test:

F=72.80, P<0.0001). The measurement scale (aggregation or local) affected variance in density in both the high density population (Bartlett's test: F=l92.48, P<0.0001) and the low density population (Bartlett's test: F=104.93, PO.0001).

Sex Ratio Estimates

The population-wide male:female OSR became more female-biased later in the season in both the low (early=1.68, mid=1.23, late=0.91) and the high density population (early=2.44, mid=0.75, late=0.78). Once again, spatial scale had a strong effect on variation in OSR (Figure

6.4). We used a Two-way ANOVA to test whether there were differences in OSR between populations and sampling times at each spatial scale. We used a box-cox transformation to 135 normalize OSR estimates. At the level of the aggregation, OSR was similar between populations

(F5,3oi=1.69, P=0.20), although it did change throughout the breeding season (F5,30i=30.16,

P<0.0001). There was a trend towards a difference in the pattern of change in OSR between populations (F5> 3oi=2.53, P=0.082) due to the sharper decrease in OSR in the high density population. At the local level, there was a significant difference between populations (F5;

3oi=19.30, PO.0001), between sampling times (F5,3oi=7.15, P=0.0009), and also between populations through time (F5,3oi=4.17, P=0.009).

We examined variance in sex ratio as a function of spatial scale, time of season and population using a Barlett's test, as above. There was a difference in the variance in female OSR between populations at the level of the aggregation (high: c2=0.83, low: o2=0.45; F=14.28,

P<0.0001) and also a difference at the local level (high: o2=0.018, low: a2=0.085; F=76.74,

PO.0001). Variance in OSR was an order of magnitude higher at the level of the aggregation compared to the local scale in both populations (high density: F=320.50, P<0.0001; low density:

F=l 12.01, PO.0001).

Many sexually cannibalistic species show a shift towards a female-biased OSR as the breeding season progresses as a result of cannibalism (Hurd et al. 1994; Maxwell 1998; Vollrath

1980; Zimmermann & Spence 1992), and we predicted a similar pattern in N. plumipes since females cannibalize or injure most males that mate. For our study, it was necessary to collect males to ensure accurate measurement of size and body condition. It could be argued that this could cause any observed shift in OSR. However, our collections are unlikely to have had an effect on the overall OSR because males found on female's webs at any given survey would have been very unlikely to survive until the next survey if left in nature. There are several reasons for this: first, the high rate of cannibalism and injury with mating (Elgar & Fahey 1996;

Schneider & Elgar 2001; Schneider et al. 2001); second, male tenure on the webs of females is short relative to female lifespan (only approximately 10% of males remained on webs after five days, Vollrath 1980; and the interval between our surveys was 14 days) and mortality rates for males that search for new females is high (Kasumovic et al. 2007). Thus, our sampling regime allowed for an accurate assessment of density, OSR, and male traits through the season, while still allowing male and female development to progress naturally between collections.

Furthermore, if our collections had depleted the population of males, we would expect the number of males collected to decrease rather than increase and remain approximately constant as was seen.

Male Competition

There were a total of 45 webs that had multiple males (N=l 13 males) throughout the season (Figure 6.5) with the majority of webs having a single male (Figure 6.6). To examine whether the average number of males per web changed through the season in each population, we performed a two-way ANOVA using aggregations as the level of analysis with the total number of males on a web as the dependent variable, and time of season and population as the independent variables. The average number of males per web was similar at the early and mid sampling period and decreased as the season progressed (early: 2.03±0.12, mid: 1.99±0.09, late:

1.01±0.10; F5,241=29.21, P<0.0001). There was no difference between populations in the

= average number of males per web (F5; 24i 1.06, P=0.30) nor was there a timexpopulation

= interaction (F5>24i 1.15, P=0.32).

In a similar analysis, we examined the number of clustered penultimate-instar and adult females within aggregations. The average number of females per aggregation increased as the breeding season progressed (F5i24i=13.86, P<0.0001). There was no difference in the average

= number of females per aggregation between populations (F5,24i 1.21, P=0.27) nor was there a timexpopulation interaction (F5> 24i=1.17, P=0.31). We used a multiple logistic regression to examine whether males closest to the hub of the web were larger or in better body condition than males further from the hub, and whether these traits covaried with survey time or 137 population. Males closest to the hub were significantly larger {y£=6J2, df=2, P=0.0095, Table

6.1) and there was a non-significant trend towards males closest to the hub having greater body condition (%2=2.79, df=2, P=0.09) than other males. There was no effect of population or survey time on the size or condition of males at the hub, nor were there any interactions among model variables. Despite the benefit of large size (Elgar & Fahey 1996; Vollrath 1980), there was considerable variation in male size through the season with mean male size across all sampling periods being 51.24±0.40mm, and male size ranging between 34.5 - 72.5mm (CV=13.82%,

Figure 6.1).

Fitness and Selection Estimates

The variation in female density and the operational sex ratio has the potential to alter the

selective pressures that males experience at different times of the season (Kokko & Monaghan

2001; Kokko & Rankin 2006). To determine whether the strength and direction of selection on male size and body condition changed within and between populations through the season, we predicted paternity based on the published relationship between paternity and proximity to the female at the time of each survey (Elgar et al. 2003; Elgar & Fahey 1996; Schneider & Elgar

2001). For these values, we used only males that were found with females. In doing so, we assumed that mate searching did not impose significant selection on male size or condition in N. plumipes. This assumption is consistent with the literature in this and similar species (Andrade

2003; Foellmer & Fairbairn 2005b; Kasumovic et al. 2007; Vollrath 1980).

The strength of selection on size and body condition fluctuated in both populations as the breeding season progressed and varied from the seasonal average (Figure 6.7). However, there was only significant positive selection on male size at mid-season in the low density population

(F2/75-6.11, P=0.016) and on body condition in the early-season sample in the high density population (F2,39=6.69, P=0.014). We tested for differences between standardized linear (^±S.E.) selection gradients on male size and body condition (Lande & Arnold 1983) by using an analysis of covariance

(ANCOVA) to examine whether slopes were equal (as in Caruso et al. 2003; Conner 1989). To detect differences in selection among size and body condition measurements between sampling periods within populations, we performed separate ANCOVAs for each population. We entered fitness as the response variable, both phenotypic traits (body size and body condition) as the

continuous terms and sampling time as the random variable in the model, along with both sampling timextrait terms. A significant sampling timextrait interaction indicates that the

strength of selection varied significantly over time for that trait. The strength of selection on male size differed between sampling periods in the low density population (F8,i73=3.05,

= P=0.050, Figure 6.7), but not the high density population (Fgji20 0-04, P=0.96). In contrast, the

strength of selection on body condition differed significantly throughout the season in the high density population (F8,i20=3.73 , P=0.027), but not the low density population (F8,i73=0.27,

P=0.77, Figure 7).

To test whether there was a difference in selection between populations within time periods, we entered both traits as the continuous terms and population as the categorical term in the model, along with each populationxtrait term. A significant populationxtrait term indicates that the strength of selection varied between populations. The strength of selection on male size did not differ between populations within sampling times for either body size (Early: Fs^s^.l 8,

P=0.68; Mid: F5,i25=1.16, P=0.28; Late: F5,io3=1.18, P=0.28), or body condition (Early:

F5>65=1.29, P=0.26; Mid: F5,i25=0.09, P=0.77; Late: F5,i03=0.97, P=0.33).

Finally, to simultaneously test whether selection varied between populations through time, we entered both traits as continuous terms, both population and sampling time as the categorical terms in the model, along with each categorical variable xtrait term. Again, a significant sampling timextrait term indicates a significant difference in selection between sampling periods, while a significant populationxtrait term indicates a significant difference m selection between populations. There was no significant effect of population on either trait

: (body size: Fn)299= 0.32, P==0.57; body condition: Fii,299=0.47, P=0.49) or sampling time on either trait (body size: Fi 1,299=1.72, P=0.18; body condition: FU;299=0.59, P=0.21).

Although the nonlinear selection gradients also fluctuated for both size and body condition, there was only significant nonlinear selection on body size in the early sampling period in the low density site. In general, there was no strong evidence for stabilizing or disruptive selection in our study populations either in each survey period, or as a cross-season average (Table 6.2).

DISCUSSION

Our study examined how OSR and density changed in two different populations within a single breeding season to determine whether demographic fluctuations were related to shifts in the strength and direction of selection. We demonstrate that both demographic variables differed between populations and fluctuated within populations throughout the breeding season (Figures

6.3 & 6.4). Furthermore, we measured high variance in OSR and female density over the season, and show that this depends on both the population surveyed and the spatial scale of the examination. As a result of these demographic fluctuations, males encountered very different levels of competitive challenges, as was evidenced by changes in the intensity of intrasexual competition over the season (Figures 6.5 & 6.6). This correlated with some significant changes in the strength of selection through time within two closely situated populations (Figure 6.7).

Despite the fact that the number of males and females changed similarly throughout the breeding season, the OSR became progressively female biased (Figure 6.4). This progressive shift towards a female biased OSR can either be a result of: (1) our male collections, or (2) differential changes in population demographics between the sexes. First, it is unlikely that our collection influenced male density since the absolute number of males we collected peaked in 140 mid-season, matching the pattern of female density and clustering. If our collection influenced the absolute number of males, we would expect male density to decrease over time, and to differ from female patterns as females were not collected. A male's tenure on a web is short under natural circumstances and it is likely the majority of males perish or are cannibalized within five days of arriving on a female's web (Vollrath 1980), or die while searching for other females if they abandon the web (Kasumovic et al. 2007). Thus, the males that we collected would not have been present at the next collection period two weeks later.

Sex differences in movement (Matter & Roland 2002; Wauters & Dhondt 1993), development (Foellmer & Fairbairn 2005a; Forster & Kingsford 1983) and/or mortality

(Gwynne 1987) rates are all potential factors that could lead to observed biases in the OSR.

Changes in immigration and emigration rates are unlikely to explain the variation in OSR in this species as a previous study has shown that males do not move between isolated populations

(Kasumovic et al. 2007). This is also likely the case with females as they are sedentary once they begin web building (Foelix 1982). The progressive female bias seen in N. plumipes is likely a result of differences in maturity rates between the sexes and higher mortality rates in males

(e.g., Kasumovic et al. 2007). Females take much longer to mature than males (Higgins 1992) and are present for the majority of the breeding season. For example, females that are counted as juveniles and penultimates earlier in the season are the same females that are counted as adults later in the season. Males however, mature much more quickly and then leave their web in search of females. During this period of mate search, approximately 80% of males perish due to their vulnerability to predators (Kasumovic et al. 2007). Furthermore, as N. plumipes is known to sexually cannibalize males before the end of the copulation (Elgar & Fahey 1996; Schneider

& Elgar 2001), multiply-mating cannibalistic females could skew the OSR as seen in other cannibalistic species (Hurd et al. 1994; Maxwell 1998; Vollrath 1980; Zimmermann & Spence

1992). Our results demonstrate the importance of identifying the relevant scale for examining population demographics. Both populations had similar estimates of OSR and female density at the level of the aggregation (Figure 6.3 & 6.4). Thus, despite the overall difference in density between populations, aggregations were of similar size. However, at the local level, variation in female density was significantly different between populations with the high density population having a significantly greater local OSR and female density throughout the season. This arises because of the higher density of aggregations in the high density population. The scale selected for comparisons could therefore have a large impact on the assumptions made regarding the competitive challenges individuals are likely to encounter in nature. In N. plumipes, the relevant scale of female density and the OSR is most likely the aggregation as males mature either within, or nearby aggregations of females (only 8 of 320 males collected were found on their own web outside of aggregations). Thus, assessing selection on male phenotypes as a function of larger scale estimates could lead to incorrect interpretations regarding the importance of population demographics in this species.

The natural shift in population demographics reported here will alter the competitive challenges that males experience and potentially the opportunity for selection on males within a season, as it alters the number of available females and competing males at different temporal and spatial scales. However, despite the female-biased sex ratio, there were still males clustering on webs of individual females in our study, similar to results mArgiope aurantia, another orb- web spider (Foellmer & Fairbairn 2005a). These studies suggest that males may be limited in their ability to find suitable females, and that a female biased OSR may not necessarily relax selection on male traits if clustering leads to intense competition (Fromhage et al. 2005). For example, selection on male size was weakest in magnitude in the beginning of the season when male clustering on female's webs was least intense (Figure 6.5) even though there was a male- biased OSR (Figure 6.4). Selection on male size was strongest in the low density population when male clustering on female's webs was most intense (mid-season, Figure 6.5), while selection on male body condition was strongest in the high density population when the OSR was male-biased (Figure 6.4 & 6.7). Although estimates of OSR and density at the population level could aid the understanding of how selection varies (Kokko & Monaghan 2001; Kokko &

Rankin 2006), local variance in either factor due to clustering effects within a population may weaken correlations between population level demographics and estimates of the strength and direction of selection.

Microspatial variation at levels below the population may thus cause biologically relevant variation in selection pressures. This means that some studies may underestimate the true variation in the strength and/or direction of linear and correlative selection by averaging values throughout the season, multiple years, and/or populations (Jann et al. 2000; Kingsolver et al. 2001). Furthermore, fluctuations in the strength and direction of selection, such as those measured here, have the potential to dilute the net influence of selection, reducing the response to selection on heritable traits at the level of the population (e.g. Wilson et al. 2006). Although the available data will not always allow examinations of spatial variation in selection pressures within seasons, it is nonetheless critical to consider the existence of such fluctuations, especially for species likely to show plastic development or behaviour in response to such variation (e.g.,

Kasumovic & Andrade 2006) since this plasticity can skew estimates of fitness associated with estimating the strength of selection. Determining the salient spatial and temporal scale for examining demographics and selection will be especially important when: (1) individuals move on a smaller spatial scale than the entire population on which the estimates are based (e.g.,

Andrade 2003; Foellmer & Fairbairn 2005 a), (2) demographic variables can shift rapidly due to the reproductive ecology of the species (e.g., Blanckenhorn et al. 1999; Jann et al. 2000), and (3) an individual's lifespan is shorter than the entire breeding season and thus individuals would not experience a seasonal average (Bradshaw 1965) (e.g., sexually cannibalistic species: Andrade

1996; Maxwell 1998).

We demonstrate that multiple competitive contexts can exist within a single population when examined at different temporal and/or spatial scales as a result of natural fluctuations in population demographics (e.g. male clustering in different areas of a population, differences in maturity rates between the sexes). Coupled with the idea that heritability and selection vary depending on environmental quality, potentially limiting evolution (Wilson et al. 2006), spatial and temporal variation in selection within a population may help explain why estimates of the strength of selection examined over an entire breeding season may seem low. Further studies examining how demographic variation at different scales can affect estimates of selection at different times of the breeding season are necessary to better understand the relationship between demography, the strength and direction of selection, and the evolution of phenotypes

(Benton et al. 2006; Kokko & Lopez-Sepulcre 2007; Metcalf & Pavard 2006).

ACKNOWLEDGEMENTS

We would like to thank D.O. Elias, D. Gwynne, K.A. Judge, H. Kokko, F.H. Rodd, D.

Punzalan, L. Rowe and the Integrative Behaviour & Neuroscience Group (UTSC) for comments on the manuscript. This research was supported by an NSERC PGS B and OGS to MMK,

Macquarie University to MJB, Macquarie University and ARC to MEH, and grants from

NSERC, CFI and OIT to MCBA. LITERATURE CITED

Andrade, M. C. B. 1996 Sexual selection for male sacrifice in the Australian redback spider.

Science 271, 70-72.

Andrade, M. C. B. 2003 Risky mate search and male self-sacrifice in redback spiders.

Behavioral Ecology 14, 531-538.

Arnold, S. J. & Wade, M. J. 1984a On the measurement of natural and sexual selection:

applications. Evolution 38, 720-734.

Arnold, S. J. & Wade, M. J. 1984b On the measurement of natural and sexual selection: theory.

Evolution 38, 709-719.

Arnqvist, G. & Rowe, L. 2006 Sexual conflict. Princeton: Princeton University Press.

Benton, T. G., Plaistow, S. & Coulson, T. M. 2006 Complex population dynamics and complex

causation: devils, details and demography. Proc. R. Soc. Lond. B 273, 1173-1181.

Blanckenhorn, W. U. 1998 Adaptive phenotypic plasticity in growth, development, and body

size in the yellow dung fly. Evolution 52, 1394-1407.

Blanckenhorn, W. 11, Morf, C, MiihlMuser, C. & Reusch, T. 1999 Spatiotemporal variation in

selection on body size in the dung fly Sepsis cynipsea. Journal of Evolutionary Biology

12, 563-576.

Bradshaw, A. D. 1965 Evolutionary significance of phenotypic plasticity in plants. Advances in

Genetics 13, 115-155.

Brodie, E. D., Moore, A. J. & Janzen, F. J. 1995 Visualizing and quantifying natural selection.

Trends in Ecology & Evolution 10, 313-318.

Carroll, S. P. & Corneli, P. S. 1995 Divergence in Male Mating Tactics between 2 Populations

of the Soapberry Bug .2. Genetic Change and the Evolution of a Plastic Reaction Norm

in a Variable Social-Environment. Behavioral Ecology 6,46-56. Carroll, S. P. & Salamon, M. H. 1995 Variation in sexual selection on male body size within

and between populations of the soapberry bug. Animal Behaviour 50, 1463-1474.

Caruso, C. M., Peterson, S. B. & Ridley, C. E. 2003 Natural selection on floral traits of Lobelia

(Lobeliaceae): Spatial and temporal variation. American Journal of Botany 90, 1333-

1340.

Clutton-Brock, T. H., Rose, K. E. & Guinness, F. E. 1997 Density-related changes in sexual

selection in red deer. Proceedings of the Royal Society B-Biological Sciences 264,1509-

1516.

Conner, J. K. 1989 Density-dependent sexual selection in the fungus beetle, Bolitotherus

cornutus. Evolution 43, 1378-1386.

Elgar, M. A. 1989 Kleptoparasitism - a cost of aggregating for an orb-weaving spider. Animal

Behaviour 37, 1052-1055.

Elgar, M. A. 1992 Sexual cannibalism in spiders and other invertebrates. In Cannibalism:

Ecology and Evolution among diverse taxa (ed. M. A. Elgar & B. J. Crespi), pp. 128-

155. Oxford: Oxford University Press.

Elgar, M. A., Bruce, M. J., de Crespigny, F. E. C, Cutler, A. R., Cutler, C. L., Gaskett, A. C,

Herberstein, M. E., Ramamurthy, S. & Schneider, J. M. 2003 Male mate choice and

patterns of paternity in the polyandrous, sexually cannibalistic orb-web spider Nephila

plumipes. Australian Journal of Zoology 51,357-365.

Elgar, M. A. & Fahey, B. F. 1996 Sexual cannibalism, competition, and size dimorphism in the

orb-weaving spider Nephila plumipes Latreille (Araneae: Araneoidea). Behavioral

Ecology 7, 195-198.

Fisher, R. A. 1930 The genetical theory of natural selection. Oxford: Clarendon Press.

Foelix, R. 1982 Biology of spiders. Cambridge: Harvard University Press. Foellmer, M. W. & Fairbaim, D. J. 2004 Males under attack: sexual cannibalism and its

consequences for male morphology and behaviour in an orb-weaving spider.

Evolutionary Ecology Research 6,163-181.

Foellmer, M. W. & Fairbairn, D. J. 2005a Competing dwarf males: sexual selection in an orb-

weaving spider. Journal of Evolutionary Biology 18, 629-641.

Foellmer, M. W. & Fairbairn, D. J. 2005b Selection on male size, leg length and condition

during mate search in a sexually highly dimorphic orb-weaving spider. Oecologia 142,

653-662.

Forster, L. M. & Kingsford, S. 1983 A preliminary study of development in two Latrodectus

species (Aranea: Theridiidae). N. Z. Entomol. 7,431-438.

Fromhage, L., Elgar, M..A..& Schneider, J. M. 2005 Faithful without care: The evolution of

monogyny. Evolution 59, 1400-1405.

Grant, J. W. A., Bryant, M. J. & Soos, C. E. 1995 Operational sex ratio, mediated by snychrony

of female arrival, alters variance of male mating success in Japanese medaka. Anim.

Behav. 49, 367-375.

Grant, P. R. & Grant, B. R. 2002 Unpredictable evolution in a 30-year study of Darwin's

finches. Science 296, 707-711.

Gwynne, D. T. 1987 Sex-biased predation and the risky mate-locating behaviour of male tick-

tock cicadas (Homoptera: Cicadidae). Animal Behaviour 35, 571-576.

Gwynne, D. T., Bailey, W. J. & Annells, A. 1998 The sex in short supply for matings varies

over small spatial scales in a katydid (Kawanaphila nartee, Orthoptera : Tettigoniidae).

Behavioral Ecology and Sociobiology 42, 157-162.

Herberstein, M. E. & Elgar, M. A. 1994 Foraging strategies of Eriophora transmarina and

Nephilaplumipes (Aranea: Araneoida): Nocturnal and diurnal orb-weaving spiders.

Australian Journal of Ecology 19,451-457. Higgins, L. 1993 Constraints and plasticity in the development of juvenile Nephila clavipes in

mexico. Journal of Arachnology 21, 107-119.

Higgins, L. E. 1990 Variation in foraging investment during the intermolt interval and before

egg-laying in the spider Nephila clavipes (Araneae, Araneidae). Journal of Insect

Behavior 3,773-783.

Higgins, L. E. 1992 Developmental plasticity and fecundity in the orb-weaving spider Nephila

clavipes. Journal of Arachnology 20, 94-106.

Hoekstra, H., Hoekstra, J., Berrigan, D., Vignieri, S., Hoang, A., Hill, C, Beerli, P. &

Kingsolver, J. 2001 Strength and tempo of directional selection in the wild. Proceedings

of the National Academy of Sciences 98, 9157-9160.

Hurd, L. E., Eisenberg, R. M., Fagan, W. F., Tilmon, K. J., Snyder, W. E., Vandersall, K. S.,

Datz, S. G. & Welch, J. D. 1994 Cannibalism reverses male-biased sex-ratio in adult

mantids - female strategy against food limitation. Oikos 69,193-198.

Jakob, E. M., Marshall, S. D. & Uetz, G. W. 1996 Estimating fitness: a comparison of body

condition indices. Oikos 77,61-67.

Jann, P., Blanckenhorn, W. U. & Ward, P. I. 2000 Temporal and microspatial variation in the

intensities of natural and sexual selection in the yellow dung fly Scathophaga

stercoraria. Journal of Evolutionary Biology 13, 927-938.

Kasumovic, M. M. & Andrade, M. C. B. 2006 Male development tracks rapidly shifting sexual

versus natural selection pressures. Current Biology 16, R242-243.

Kasumovic, M. M. & Andrade, M. C. B. 2007 Changes in competitive context reverses sexual

selection on male size. American Naturalist Submitted.

Kasumovic, M. M., Bruce, M. J., Herberstein, M. E. & Andrade, M. C. B. 2007 Male mate

choice may increase the cost of mate searching in the golden orb-web spider {Nephila

plumipes). Behavioral Ecology 18, 189-195. 148 Kingsolver, J., Hoekstra, H., Hoekstra, J., Berrigan, D., Vignieri, S., Hill, C, Hoang, A., Gibert,

P. & Beerli, P. 2001 The strengths of phenotypic selection in natural populations.

American Naturalist 157, 245-261.

Kokko, H. & Lopez-Sepulcre, A. 2007 The ecogenetic link between demography and evolution:

can we bridge the gap between theory and data? Ecology Letters In press.

Kokko, H. & Monaghan, P. 2001 Predicting the direction of sexual selection. Ecology Letters 4,

159-165.

Kokko, H. & Rankin, D. J. 2006 Lonely hearts or sex in the city? Density dependent effects in

mating systems. Phil. Trans. R. Soc. B 361, 319-334.

Lande, R. & Arnold, S. J. 1983 The measurement of selection on correlated characters Evolution

37,1210-1226.

Losos, J. B., Schoener, T. W., Langerhans, R. B. & Spiller, D. A. 2006 Rapid temporal reversal

in predator-driven natural selection. Science 314,1111.

Matter, S. F. 1997 Population density and area: The role of between- and within-patch

processes. Oecologia 110,533-538.

Matter, S. F. & Roland, J. 2002 An experimental examination of the effects of habitat quality on

the dispersal and local abundance of the butterfly Pamassius smintheus. Ecological

Entomology 27, 308-316.

Maxwell, M. R. 1998 Seasonal adult sex ratio shift in the praying mantid Iris oratoria (Mantodea

: Mantidae). Environmental Entomology 27, 318-323.

Metcalf, C. J. E. & Pavard, S. 2006 Why evolutionary biologists should be demographers.

Trends Ecol Evol 22, 205-212.

Parker, G. A. 2006 Sexual conflict over mating and fertilization: an overview. Phil. Trans. R.

Soc. B 361, 235-259. Schneider, J. M. & Elgar, M. A. 2001 Sexual cannibalism and sperm competition in the golden

orb-web spider Nephila plumipes (Araneoidea): Female and male perspectives.

Behavioral Ecology 12, 547-552.

Schneider, J. M. & Elgar, M. A. 2002 Sexual cannibalism in Nephila plumipes as a consequence

of female life history strategies. Journal of Evolutionary Biology 15, 84-91.

Schneider, J. M., Thomas, M. L. & Elgar, M. A. 2001 Ectomised conductors in the golden orb-

web spider, Nephila plumipes (Araneoidea): a male adaptation to sexual conflict?

Behavioral Ecology andSociobiology 49, 410-415.

Schulte-Hostedde, A. I., Zinner, B., Millar, J. S. & Hickling, G. J. 2005 Restitution of mass-size

residuals: Validating body condition indices. Ecology 86, 155-163.

Simmons, L. W. & Kvarnemo, C. 2006 Costs of breeding and their effects on the direction of

sexual selection. Proceedings of the Royal Society of London Series B Biological

Sciences 273, 465-470.

Velez, M. J. & Brockmann, H. J. 2006 Seasonal variation in selection on male calling song in

the field cricket, Gryllus rubens. Animal Behaviour 72, 439-448.

Vollrath, F. 1980 Male body size and fitness in the web-building spider Nephila clavipes.

Zeitschrift Tierpsychologie 53,61-78.

Wauters, L. & Dhondt, A. A. 1993 Immigration pattern and success in red squirrels. Behavioral

Ecology and Sociobiology 33,159-167.

Wauters, L. A., Gurnell, J., Preatoni, D. & Tosi, G. 2001 Effects of spatial variation in food

availability on spacing behaviour and demography of Eurasian red squirrels. Ecography

24,525-538.

Wiklund, C. & Fagerstrom, T. 1997 Why do males emerge before females? A hypothesis to

explain the incidence of protandry in butterflies, Oecologia 31,153-158. Wilson, A. J., Pemberton, J. M., Pilkmgton, J. G., Coltman, D. W., Mifsud, D. V., Clutton-

Brock, T. H. & Kruuk, L. E. B. 2006 Environmental coupling of selection and

heritability limits evolution. PloSBiology 4, 1270-1275.

Zimmermann, M. & Spence, J. R. 1992 Adult-population dynamics and reproductive effort of

the fishing spider Dolomedes triton (Araneae, Pisuridae) in Central Alberta. Can. J.

Zaol. 70, 2224-2233. Table 6.1: Effects of population, body size and condition on male proximity to the hub of female's webs when rival males are present (n=45 webs).

1 Source df Wald x P

Population 1 0.011 0.91

Time 2 2.08 0.35

Size 1 6.72 0.0095

Body condition 1 2.79 0.095

Size x Time 2 0.027 0.99

Condition x Time 2 2.96 0.23

Size x Population 1 0.076 0.78

Condition x Population 1 0.18 0.67 152 Table 6.2: Non-linear univariate selection gradients (±S.E.) on size and condition and non-linear bivariate selection gradients.

High Density Low Density Both

Selection y y Populations

Early 0.025±0.041 0.125±0.060* -

Quadratic size Mid 0.009±0.027 -0.020±0.022 -

Late -0.037±0.045 0.003±0.029 -

Overall 0.010±0.018 -0.005±0.016 0.002±0.012

Early 0.016±0.026 -0.084±0.059

Quadratic condition Mid 0.017±0.022 0.031±0.031

Late -0.022±0.052 -0.034±0.025

Overall 0.001±0.014 -0.021±0.016 -0.011±0.018

Early -0.153±0.078f -0.177±0.144 -

Correlational Mid -0.003±0.040 O.O29±O.040 -

Late 0.045±0.068 0.026±0.054 -

Overall -0.005±0.029 0.027±0.025 0.013±0.019 f PO.IQ, * P<0.05 Figure 6.1: A frequency distribution of male a) size and b) body condition within the high

(black bars) and low (white bars) density population.

Figure 6.2: The distribution of adult (white bars), penultimate instar (one moult prior to adulthood, grey bars), and juvenile (more than one moult prior to adulthood, black bars) females in the low and high density populations throughout the breeding season.

Figure 6.3: The density of penultimate instar and adult females at the aggregations (# of females/aggregation, top) and local level (# females/5m, bottom) at three survey times in two populations (high density: squares with a solid line; low density: circles with a dashed line).

Figure 6.4: The operational sex ratio (male: female) at the aggregation (top) and local (bottom) level at three survey times in two populations (high density: squares with a solid line; low density: circles with a dashed line).

Figure 6.5: The number of webs containing multiple males found in the high (black bars) and low (white bars) density site at each sampling time.

Figure 6.6: Frequency distribution of the numbers of webs with multiple males in the high

(black bars) and low (white bars) density population.

Figure 6.7: The strength of selection on male size (top) and body condition (bottom) during three surveys in two populations (high density: squares with a solid line; low density: circles with a dashed line). Selection gradients calculated from data summed across the three survey 154 periods are also shown (average: solid red line; standard error: dotted red line). Asterisks denote statistically significant selection gradients at the P

-». N> W 4^ oi O) O O O O O O O o o o o o o o

o

o bo' 156

1.0-1 0.9- 0.8- 0.7- 5 0.6- o tor 0.5- 8- °-4" i: o.3- 0.2- 0.1- O.OJ Low High Low High Low High Early Mid Late Sampling Period 2.5.

2.04

1.5J

1.04

0.5

121

10

8

6 4 t Early Mid Late Sampling Time Operational Sex Ratio

00 159

14 0) -Q 0 12 5 j> 8 Q. +-* 3 s 6 «*— o i_ 4 0 JQ E 2 z3

Early Mid Late Sampling Time Number of Webs ho w oooooooo o o o i I i i i I i > i I i I i I i VL- J I I

mmm C 3 U" CD

Q3_ CD C/> "O CD Oi-T CD

OJ-L

ON o 0.15.

-0.1 (N1 -0.15 0.20,

-0.1 O^U Early Mid Late Sampling Time 162

CHAPTER SEVEN

General Discussion The goal of this thesis is to determine whether developmental plasticity in response to a demographically variable environment is an adaptive explanation for the maintenance of variation in size and body condition. I postulated that four conditions are required for phenotypic distributions to be adaptive developmental responses to environmental variation, rather than the result of constraints on development imposed by environmental variation: (1)

Individuals must have reliable cues of the challenges they will experience at maturity, and these must be available prior to maturation, (2) Individuals must alter their development in response to these reliable cues (i.e., developmental plasticity must be apparent), (3) This developmental plasticity must be adaptive, and (4) The environment must vary either spatially or temporally within the lifespan of the focal animal. My work focused on plasticity in body size, body condition and time to maturity in males, as these traits are critical to success under sexual competition, but the importance of each may change with local conditions.

In my work, I used three different species of spiders, two species native to Australia

(Latrodectus hasselti and Nephila plumipes) and a species native to North America (L. hesperus). Spiders provide an excellent system in which to examine the effect of demographic variation on developmental plasticity of males for several reasons. First, males use long-distance pheromones to locate sedentary females (Gaskett 2007; Chapters Two and Three), and pheromones may provide cues of receptive female density, and potentially, the adult sex ratio.

Second, the demographic environment of a male spider fluctuates throughout a single breeding season as the sexes differ in their rates of maturity and survival (Forster & Kingsford 1983;

Higgins 1992), altering the competitive landscape within a single breeding season (Chapter Six).

And third, males vary in their development time (Chapter Four) and in their phenotypic distributions of body size and condition (Chapters Three, Four, and Five), which make them primary candidates for examinations of life-history trade-offs. 1. Availability of reliable cues of future competitive challenges.

In Chapters Two and Three, I demonstrated that males used long-distance pheromones to distinguish between females of different species (Chapter Two: L. hesperus), and also between females of different mating status, basing their choice of female on their own phenotype

(Chapter Three: N. plumipes). Most importantly, these studies suggest that pheromones could provide reliable cues of how population demographics are fluctuating in adult populations of spiders, and therefore, the competitive challenges males are likely to encounter at maturity. The use of chemical cues such as pheromones is not uncommon as many taxa use pheromones to locate individuals of the opposite sex (for a general review see Shorey 1976) (spiders: Gaskett

2007; Pollard et al. 1987; Schulz 2004) (moths: Phelan 1997) and recognize the presence of predators (reviewed in Benard 2004). However, I also demonstrated that male redback spiders respond to chemical cues of neighbouring conspecific males (Chapter Four). This is the first evidence that male spiders can recognize pheromonal cues of potential competitors for mates

(for reviews: Gaskett 2007; Schulz & Toft 1993). In this species, and potentially other animals that use pheromones as a means of communication, pheromonal cues can thus potentially provide reliable cues of the relative strength of sexual selection in the adult's environment.

2. Developmental plasticity occurs in response to these reliable cues.

In Chapter Four, I demonstrated that juvenile male redback spiders responded to pheromonal cues from females and males by shifting their development. Despite a common diet, males reared in the absence of females matured significantly larger and in better body condition than males reared in the presence of females, but required significantly longer to do so. Male developmental trajectories were further mediated by male density, with males developing significantly larger when in the presence of a greater number of neighbouring males. These results are some of the first to demonstrate that males are developmentally plastic in response to pheromonal cues of both male and female density, and suggest that males simultaneously use information on the density of both sexes to determine the phenotype that can maximize fitness.

The phenotypic trade-offs that I demonstrate in Chapter Four also suggest there are constraints on male development. Although males in each treatment were fed the same amount of food, there were still trade-offs between body size, body condition, and development time.

There are a large number of studies that demonstrate trade-offs between size and time at maturity (reviewed in Berner & Blanckenhorn 2007; Roff 1992; Stearns 1992). Because of limited resource acquisition during an individual's lifetime, decisions must be made regarding where resources must be allocated, resulting in certain traits having priority (Zera & Harshman

2001). However, rapid growth may not only be constrained by resource availability, but also because of the risk associated with growing large too quickly (e.g., Higgins & Rankin 2001).

Further examinations of this phenotypic trade-off dynamic are necessary to fully understand the trade-offs between traits associated with fitness in various environments.

3. Developmental plasticity must be adaptive.

Studies of developmental plasticity are common (reviewed in Adler & Harvell 1990;

Agrawal 2001; Bateson et al. 2004; Harvell 1990; Nylin & Gotthard 1998), however, developmental plasticity may simply result from an internally labile system, rather than an evolutionary solution to environmental variation (Doughty & Reznick 2004; Meyers & Bull

2002; Pigliucci 2001), further studies demonstrating the adaptive nature of developmental plasticity are therefore necessary to distinguish between the two factors above. In Chapter Five, using a similar protocol to a reciprocal transplant experiment seen in plant literature (Gotthard &

Nylin 1995), I showed that developmental plasticity in response to conspecific density (Chapter

Four) was adaptive. In this experiment, I found that the smaller males (which develop in the presence of females, Chapter Four) outperformed larger individuals when given the developmental advantage associated with their faster development. That is, the temporal developmental advantage of small males can yield significant advantages in mate searching and mating, particularly when females are in relatively close proximity (high density). However, larger males were superior in direct intraspecific competition for females. Thus, developmental plasticity is maintained in this species as larger males have increased fitness when the local sex ratio is male biased, and smaller males have increased fitness when the local sex ratio is female biased.

My results provide a view of male quality that contradicts a mainstream view in behavioural ecology. Behavioural ecological studies of intraspecific competition suggest that larger phenotypes are optimal for males of most taxa under both forms of sexual selection

(Andersson 1994; Blanckenhorn 2005). With this view, the development of a smaller male phenotype, despite sufficient resources to develop a larger size (e.g., Chapter Four), is puzzling.

However, because the majority of competition studies test male phenotypes in a single competitive environment (e.g., lab studies using field-collected individuals for studies of female choice or intrasexual competition; Elgar & Fahey 1996; Forslund 2000; Schaefer & Uhl 2003) they do not include the importance of the interplay between local competitive challenges and male developmental or behavioural strategies. This can be problematic because the fitness associated with a phenotype is partly a consequence of the biotic and abiotic environment in which it developed (ecogenetic feedback, Kokko & Lopez-Sepulcre 2007). My results bring together life-history and behavioural ecology theory, demonstrating the importance of testing male phenotypes within the context in which they developed.

4. The environment must vary in competitive challenges.

The environment has been shown to vary between breeding seasons, and this can have an affect on the evolution of phenotypes (Losos et al. 2006; Wilson et al. 2006). Similar within- season estimates are necessary to clarify whether small scale temporal variation may also have an effect on the process of evolution. However, examinations of within-season variation of selection pressures in animals are few (e.g., Jann et al. 2000; Punzalan 2007). In Chapter Six, I demonstrated that operational sex ratio and population density in N. plumipes fluctuate within a single breeding season, and estimates of these variables depend on the spatial scale of the examination. There is considerable evidence that population demographics are not static in natural environments within a single habitat and breeding season. For example, population demographics can vary with sex differences in birth, death, and maturity rates (Foellmer &

Fairbairn 2004; Foellmer & Fairbairn 2005a; Zimmermann & Spence 1992), immigration/emigration rates (Clutton-Brock et al. 1997; e.g., Matter 1997; Matter & Roland

2002; Maxwell 1998; Wauters & Dhondt 1993), time of arrival to breeding grounds (Grant et al.

1995; Wiklund & FagerstrSm 1997), predation rates (Gwynne 1987; Gwynne & Bussiere 2002;

Norrdahl & Korpimaki 1998; Su & Li 2006), and differences in mating optima between the sexes (Arnqvist & Rowe 2006; Parker 2006). In N. plumipes, I suggest that the fluctuation in demographics is likely due to variation in lifespan between the sexes, and due to higher male mortality as a result of mate searching and sexual cannibalism.

Chapter Six also demonstrates the importance of understanding the ecology of the organism under study to determine the relevant temporal and spatial scale for measurement of demographic variation and selection gradients. In this study, there were very different estimates of population demographics as a function of temporal and spatial scale and this would affect correlations between selection and demography. Although many studies measure average selection pressures across seasons, my thesis demonstrates that such averaging may conceal significant patterns of fluctuation in selection, and potentially mask factors that could underlie plasticity-generated phenotypic distributions. I demonstrate that the strength of sexual selection can change rapidly within a season (within two weeks) resulting in a phenotypic optimum that is likely to shift over space and time. 168 Small scale variation in the environment is likely to be important whenever changes in population demographics occur on a shorter temporal scale than an individual's lifespan. For example, in many invertebrates, individuals live for only a short period of time (often less than one complete breeding season for the species), and during that lifespan, rapid maturation and/or mortality of conspecifics is likely to quickly alter selection pressures. Only a few studies have demonstrated within-season variation in selection pressures in response to variation in demographics in animal taxa (Dung flies, Scathophaga stercoraria: Jann et al. 2000; Ambush bugs, Phymata americana: Punzalan 2007). Further studies examining variation in population demographics and selection pressures within seasons are necessary to determine the relevance of demographic variation to phenotypic evolution and plasticity in phenotypic development.

Future directions of study

Overall, my thesis suggests that developmental plasticity can occur in response to demographic variation that is likely to occur naturally within a season, and that different phenotypic optima may exist for the different competitive challenges resulting from those shifting demographics. It is critical that developmental plasticity in response to population density and sex ratio be examined in a wider range of taxa. Such work is necessary for understanding male developmental tactics and the evolution of phenotypic distributions

(Pigliucci 2001; Scheiner 2002). It is currently unclear how widespread developmental plasticity in response to short-term demographic variation among taxa. However, given our understanding of the cues necessary to determine developmental trajectories and how populations can vary in nature, it is possible to make qualitative predictions regarding the variables that are necessary for developmental plasticity to be favoured and contribute to the observed diversity of life- history and metric traits.

(1) Plasticity is more likely if individuals have reliable cues of future competitive challenges. The predictability of the environment is an important determinant of the evolution 169 of plasticity, with plasticity more likely to evolve in accurately predictable environments (Getty

1996; Scheiner 1993). Temperature and photoperiod are easily predictable by individuals as they follow a natural progression within a season. However, predicting demographic variation may be more difficult as it is less likely to follow a clear temporal pattern. In such instances, reliable cues that accurately predict future variation are necessary if developmental plasticity is to be a viable adaptive strategy. These cues are likely to vary among taxa depending on the life- history and ecology of the species under examination. For example, individuals may have pheromonal cues of available females and surrounding competitors (e.g., Kasumovic & Andrade

2006), or may have visual cues of competitors if generations are overlapping (e.g., Walling et al.

2007). Individuals may also have tactile cues if the interaction rate during the larval stage correlates with the density of adults that will mature (e.g., Gage 1995; Simpson et al. 1999;

Stockley & Seal 2001). Species with a substantial wandering phase during the juvenile stage

(e.g., Vollrath & Parker 1992) that would allow individuals to estimate population demographics may also be more likely to show plasticity of traits associated with fitness.

(2) Plasticity should be more likely if selection pressures vary and there is a strong phenotype-fitness association. If the competitive context that individuals experience upon maturity varies either temporally or spatially (Scheiner 1993; Scheiner 2002), a single phenotype will be unlikely to perform well in all contexts, and this favours the evolution of plasticity. This is especially important if opposing traits determine male fitness at different times of the season. Furthermore, if environments fluctuate significantly between seasons (e.g., Losos et al. 2006; Wilson et al. 2006) or within seasons (Chapter Six; Jann et al. 2000; Punzalan

2007), offspring would rarely experience the same competitive environment as their parents. If phenotypes have high heritability, a parent with a successful phenotype in one context could potentially have offspring with much lower fitness in the next generation when the environment changes. Thus, when competitive challenges fluctuate, higher developmental flexibility of 170 quantitative traits (and thus lower heritability) may be adaptive. In other words, phenotypic plasticity itself may be a heritable trait that can be critical to fitness (Agrawal 2001; Pigliucci

2005; Pigliucci et al. 2006; Via et al. 1995; West-Eberhard 2005).

(3) Plasticity is more likely if individuals have limited mating opportunities. If individuals are likely to mate a few times, then the phenotypic traits that are likely to increase fitness for the particular challenges they are likely to face for those limited reproductive opportunities may be more strongly favoured by selection. This is in contrast to multiply-mating or long lived species, for which competitive challenges may change substantially between mating opportunities, and an average phenotype may be favoured (van Tienderen 1991). Mating opportunities may be limited for one sex because mortality risk during mate searching is high

(Andrade 2003; Foellmer & Fairbairn 2005b; Gwynne 1987; Kasumovic et al 2007), the probability of locating the opposite sex is low (e.g., mantids: Hurd et al. 1994; Maxwell 1998), a short mating season or longer interaction time (e.g., mate guarding, biparental care) between the sexes reduces mate searching opportunities (Carroll & Corneli 1995; Herberstein et al. 2005), and/or the sexes differ in either their mating optima (e.g., Gwynne et al. 1998), or in their lifespan (e.g., Higgins 1990; Maxwell 1998).

CONCLUSION

Population distributions in nature can be seen as a consequence or response to environmental variation. The difference is that when viewed as a response to environmental variation, broad phenotypic distributions can be a result of adaptive developmental plasticity. In my thesis, I demonstrate that in JV. plumipes and L. hasselti, the environment varies in competitive challenges within a breeding season, and that the strength and direction of selection can fluctuate. As a result, there may be a range of competitive contexts that arise within a breeding season, likely resulting in a continuous distribution of phenotypes. I show that juvenile male redbacks demonstrate adaptive developmental plasticity in response to variation in 171 conspecific density—factors that signal the degree of intrasexual competition at maturity, allowing males to develop the phenotype that will potentially maximize their reproductive success. Overall, I demonstrate that broad phenotypic distributions can be considered a result of adaptive developmental plasticity as they are a response to the range of competitive challenges that individuals are likely to encounter at maturity. 172

LITERATURE CITED

Adler, F. R. & Harvell, C. D. 1990 Inducible Defenses, Phenotypic Variability and Biotic

Environments. Trends in Ecology & Evolution 5, 407-410.

Agrawal, A. A. 2001 Ecology - Phenotypic plasticity in the interactions and evolution of

species. Science 294, 321-326.

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

Andrade, M. C. B. 2003 Risky mate search and male self-sacrifice in redback spiders.

Behavioral Ecology 14, 531-538.

Arnqvist, G. & Rowe, L. 2006 Sexual conflict. Princeton: Princeton University Press.

Bateson, P., Barker, D., Clutton-Brock, T., Debal, D., D'Uldine, B., Foley, R., Gluckman, P.,

Godfrey, K., Kirkwood, T., Lahr, M., McNamara, J., Metcalfe, N., Monaghan, P.,

Spencer, H. & Sultan, S. 2004 Developmental plasticity and human health. Nature 430,

419-421.

Benard, M. F. 2004 Predator-induced phenotypic plasticity in organisms with complex life

histories. Annual Review of Ecology Evolution and Systematics 35, 651 -673.

Berner, D. & Blanckenhorn, W. U. 2007 An ontogenetic perspective on the relationship between

age and size at maturity. Functional Ecology 21, 505-512.

Blanckenhorn, W. U. 2005 Behavioral causes and consequences of sexual size dimorphism.

Ethology 111, 977-1016.

Carroll, S. P. & Corneli, P. S. 1995 Divergence in Male Mating Tactics between 2 Populations

of the Soapberry Bug .2. Genetic Change and the Evolution of a Plastic Reaction Norm

in a Variable Social-Environment. Behavioral Ecology 6, 46-56.

Clutton-Brock, T. H., Rose, K. E. & Guinness, F. E. 1997 Density-related changes in sexual

selection in red deer. Proceedings of the Royal Society B-Biological Sciences 264, 1509-

1516. Doughty, P. & Reznick, D. N. 2004 Patterns and analysis of adaptive plasticity in animals. In

Phenotypic Plasticity: Functional and Conceptual Approaches (ed. T. J. DeWitt & S. M.

Scheiner), pp. 126-150. New York: Oxford University Press.

Elgar, M. A. & Fahey, B. F. 1996 Sexual cannibalism, competition, and size dimorphism in the

orb-weaving spider Nephila plumipes Latreille (Araneae: Araneoidea). Behavioral

Ecology!, 195-19S.

Foellmer, M. W. & Fairbaim, D. J. 2004 Males under attack: sexual cannibalism and its

consequences for male morphology and behaviour in an orb-weaving spider.

Evolutionary Ecology Research 6,163-181.

Foellmer, M. W. & Fairbaim, D. J. 2005a Competing dwarf males: sexual selection in an orb-

weaving spider. Journal of Evolutionary Biology 18, 629-641.

Foellmer, M. W. & Fairbaim, D. J. 2005b Selection on male size, leg length and condition

during mate search in a sexually highly dimorphic orb-weaving spider. Oecologia 142,

653-662.

Forslund, P. 2000 Male-male competition and large size mating advatage in European earwigs,

Forficula auricularia. Animal Behaviour 59, 753-762.

Forster, L. M. & Kingsford, S. 1983 A preliminary study of development in two Latrodectus

species (Aranea: Theridiidae). N. Z. Entomol. 7, 431-438.

Gage, M. J. G. 1995 Continuous variation in reproductive strategy as an adaptive response to

population density in the moth Plodia interpunctella. Proc. R. Soc. Lond. B 261, 25-30.

Gaskett, A. C. 2007 Spider sex pheromones: emission, reception, structures, and functions.

Biological Reviews 82, 26-48.

Getty, T. 1996 The maintenance of phenotypic plasticity as a signal detection problem.

American Naturalist 148, 378-385. Gotthard, K. & Nylin, S. 1995 Adaptive plasticity and plasticity as an adaptation - a selective

review of plasticiy in animal morphology and life-history. Oikos 74, 3-17.

Grant, J. W. A., Bryant, M. J. & Soos, C. E. 1995 Operational sex ratio, mediated by snychrony

of female arrival, alters variance of male mating success in Japanese medaka. Anim.

Behav. 49, 367-375.

Gwynne, D. T. 1987 Sex-biased predation and the risky mate-locating behaviour of male tick-

tock cicadas (Homoptera: Cicadidae). Animal Behaviour 35, 571-576.

Gwynne, D. T., Bailey, W. J. & Annells, A. 1998 The sex in short supply for matings varies

over small spatial scales in a katydid (Kawanaphila nartee, Orthoptera : Tettigoniidae).

Behavioral Ecology and Sociobiology 42, 157-162.

Gwynne, D. T. & Bussiere, L. 2002 Female mating swarms increase predation risk in a 'role-

reversed' dance fly (Diptera : Empididae : Rhamphornyia longicauda Loew). Behaviour

139,1425-1430.

Harvell, C. D. 1990 The Ecology and Evolution of Inducible Defenses. Quarterly Review of

Biology 65,323-340.

Herberstein, M. E., Barry, K. L., Turoczy, M. A., Wills, E., Youssef, C. & Elgar, M. A. 2005

Post-copulation mate guarding in the sexually cannibalistic St Andrew's Cross spider

(Araneae Araneidae). Ethology Ecology & Evolution 17, 17-26.

Higgins, L. E. 1990 Variation in foraging investment during the intermolt interval and before

egg-laying in the spider Nephila clavipes (Araneae, Araneidae). Journal of Insect

Behavior 3, 773-783.

Higgins, L. E. 1992 Developmental plasticity and fecundity in the orb-weaving spider Nephila

clavipes. Journal of Arachnology 20, 94-106.

Higgins, L. E. & Rankin, M. A. 2001 Mortality risk of rapid growth in the spider Nephila

clavipes. Functional Ecology 15, 24-28. 175 Hurd, L. E., Eisenberg, R. M., Fagan, W. F., Tilmon, K. J., Snyder, W. E., Vandersall, K. S.,

Datz, S. G. & Welch, J. D. 1994 Cannibalism reverses male-biased sex-ratio in adult

mantids - female strategy against food limitation. Oikos 69,193-198.

Jann, P., Blanckenhorn, W. U. & Ward, P. I. 2000 Temporal and microspatial variation in the

intensities of natural and sexual selection in the yellow dung fly Scathophaga

stercoraria. Journal of Evolutionary Biology 13, 927-938.

Kasumovic, M. M. & Andrade, M. C. B. 2006 Male development tracks rapidly shifting sexual

versus natural selection pressures. Current Biology 16, R242-243.

Kasumovic, M. M., Bruce, M. J., Andrade, M. C. B. & Herberstein, M. E. 2007 Spatial and

temporal demographic variation drives within-season fluctuations in sexual selection.

PloS Submitted.

Kokko, H. & Lopez-Sepulcre, A. 2007 The ecogenetic link between demography and evolution:

can we bridge the gap between theory and data? Ecology Letters In press.

Losos, J. B., Schoener, T. W., Langerhans, R. B. & Spiller, D. A. 2006 Rapid temporal reversal

in predator-driven natural selection. Science 314,1111.

Matter, S. F. 1997 Population density and area: The role of between- and within-patch

processes. Oecologia 110, 533-538.

Matter, S. F. & Roland, J. 2002 An experimental examination of the effects of habitat quality on

the dispersal and local abundance of the butterfly Parnassius smintheus. Ecological

Entomology 21, 308-316.

Maxwell, M. R. 1998 Seasonal adult sex ratio shift in the praying mantid Iris oratoria (Mantodea

: Mantidae). Environmental Entomology 27, 318-323.

Meyers, L. A. & Bull, J. J. 2002 Fighting change with change: adaptive variation in an uncertain

world. Trends in Ecology & Evolution 17,551-557. Norrdahl, K. & Korpimaki, E. 1998 Does mobility or sex of voles affect risk of predation by

mammalian predators? Ecology 79, 226-232.

Nylin, S. & Gotthard, K. 1998 Plasticity in life-history traits. 43, 63-83.

Parker, G. A. 2006 Sexual conflict over mating and fertilization: an overview. Phil. Trans. R.

Soc. B 361, 235-259.

Phelan, P. L. (ed.) 1997 Evolution of mate-signaling in moths: phylogenetic considerations and

predictions from the asymmetric tracking hypothesis. The Evolution of Mating Systems

in Insects and Arachnids. Cambridge: Cambridge University Press.

Pigliucci, M. 2001 Phenotypic plasticity: beyond nature and nurture John Hopkins University

Press.

Pigliucci, M. 2005 Evolution of phenotypic plasticity: where are we going now? Trends in

Ecology & Evolution 20, 481-486.

Pigliucci, M., Murren, C. J. & Schlichting, C. D. 2006 Phenotypic plasticity and evolution by

genetic assimilation. Journal of Experimental Biology 209, 2362-2367.

Pollard, S. D., Macnab, A. M. & Jackson, R. R. (ed.) 1987 Communication with chemicals:

pheromones and spiders. Ecophysiology of Spiders. Berlin: Springer-Verlag.

Punzalan, D. 2007 Evolution of sexual colour pattern dimorphism in the ambush bug Phymata

americana. Ecology Evolution and Behaviour, University of Toronto.

Roff, D. A. 1992 The evolution of life histories: Theory and analysis. New York: Chapman &

Hall.

Schaefer, D. & Uhl, G. 2003 Male competition over access to females in a spider with last-male

sperm precedence. Ethology 109, 385-400.

Schemer, S. M. 1993 Genetics and evolution of phenotypic plasticity. Annual Review of Ecology

and Systematics 24, 35-68. Scheiner, S. M. 2002 Selection experiments and the study of phenotypic plasticity. Journal of

Evolutionary Biology 15, 889-898.

Schulz, S. (ed.) 2004 Semiochemistry of spiders. Advances in Chemical Ecology. Cambridge:

Cambridge University Press.

Schulz, S. & Toft, S. 1993 Identification of a sex pheromone from a spider. Science 260,1635-

1637.

Shorey, H. H. 1976 Animal communication by pheromones: Academic Press.

Simpson, S. J., McCaffery, A. R. & Hagele, B. F. 1999 A behavioural analysis of phase change

in the desert locust. Biological Reviews of the Cambridge Philosophical Society 74,461-

480.

Stearns, S. C. 1992 The evolution of life histories. Oxford: Oxford University Press.

Stockley, P. & Seal, N. J. 2001 Plasticity in reproductive effort of male dung flies (Scatophaga

stercorarid) as a response to larval density. Functional Ecology 15,96-102.

Su, K. & Li, D. Q. 2006 Female-biased predation risk and its differential effect on the male and

female courtship behaviour of jumping spiders. Anim. Behav. 71, 531-537. van Tienderen, P. H. 1991 Evolution of generalists and specialists in spatially heterogeneous

environments. Evolution 45, 1317-1331.

Via, S., Gomulkiewicz, R., Dejong, G., Scheiner, S. M., Schlichting, C. D. & Vantienderen, P.

H. 1995 Adaptive phenotypic plasticity - consensus and controversy. Trends in Ecology

& Evolution 10, 212-217.

Vollrath, F. & Parker, G. A. 1992 Sexual dimorphism and distorted sex ratios in spiders. Nature

360, 156-159.

Walling, C. A., Royle, N. J., Metcalfe, N. B. & Linstrom, J. 2007 Green swordtails alter their

age of maturity in response to the population level of male ornamentation. Biology

Letters3, 144-146. 178 Wauters, L. & Dhondt, A. A. 1993 Immigration pattern and success in red squirrels. Behavioral

Ecology andSociobiology 33, 159-167.

West-Eberhard, M. J. 2005 Developmental plasticity and the origin of species differences.

Proceedings of the National Academy of Sciences 102, 6543-6549.

Wiklund, C. & Fagerstrom, T. 1997 Why do males emerge before females? A hypothesis to

explain the incidence of protandry in butterflies. Oecologia 31,153-158.

Wilson, A. J., Pemberton, J. M., Pilkington, J. G., Coltman, D. W., Mifsud, D. V., Clutton-

Brock, T. H. & Kruuk, L. E. B. 2006 Environmental coupling of selection and

heritability limits evolution. PloS Biology 4,1270-1275.

Zera, A. J. & Harshman, L. G. 2001 The physiology of life history trade-offs in animals. Annual

Review of Ecology and Systematics 32, 95-126.

Zimmermann, M. & Spence, J. R. 1992 Adult-population dynamics and reproductive effort of

the fishing spider Dolomedes triton (Araneae, Pisuridae) in Central Alberta. Can. J.

Zool. 70, 2224-2233.