|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 ENVIRONMENTAL HETEROGENEITY AND PHENOTYPIC VARIATION: THE EVOLUTION OF MALE BODY SIZE IN A GOLDEN ORB WEB SPIDER By CLARE C. RITTSCHOF A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTI AL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011
2 2011 Clare C. Rittschof
3 To my family
4 ACKNOWLEDGMENTS I thank my mentor Jane Brockmann for her guidance and support, and I am grateful t o the faculty and students in the Department of Biology for providing a positive academic environment. I thank my parents for providing me a free education. I acknowledge the help of my undergraduate assistants, including Kristie Vetter, Jason Ziegler, Mar cos Mills, Nerine Constant, and Diego Valbuena for assistance in molecular work. In terms of faculty, I thank Rebecca Kimball for use of her lab space and molecular expertise, as well as David Reed, Jamie Gillooly, Ginger Clark, and Marta Wayne for shared equipment and lab space. Ben Bolker and Colette St. Mary were invaluable for statistical advice. For my work on sperm depletion, I thank my collaborator Peter Michalik and the University of Greifswald, Germany for hosting me. In addition, I thank Marion Sa ndhop (Zoological Institute and Museum, University of Greifswald, Germany) for her technical support especially with the ultramicrotomy, and Oliver Vcking and Christine Putzar (Zoological Institute and Museum, University of Greifswald, Germany) for rearin g the animals. I am especially grateful to Gabriele Uhl (Zoological Institute and Museum, University of Greifswald, Germany) for her invaluable support and the helpful discussions. I thank Matja Kuntner and Simona Kralj ( Scientific Research Centre, Slovenian Academy of Sciences and Arts, Ljubljana, Slovenia ) for the Nephilengys specimen and Klaas Welke (University of Hamburg, Germany) for the specimen of Argiope lobata Additionally, I thank the Ordway Swisher Biological Station, Glenn Hall and the B ee Research Unit, and Mark Brenner and Susan Milbraith for use of their property as field sites. I thank my funding support, which included Sigma Xi Grants in Aid of Research, the American Arachnology Society, the Animal Behavior Society Student Research G rant, the H.H.M.I. G.A.T.O.R. program,
5 Department of Zoology travel grants, Graduate Student Council travel grants, and a National Science Foundation Doctoral Dissertation Improvement Grant ( DDIG IOS 0909367).
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .............. 4 LIST OF TABLES ................................ ................................ ................................ ......................... 9 LIST OF FIGURES ................................ ................................ ................................ ..................... 10 ABSTRACT ................................ ................................ ................................ ................................ 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ................. 15 2 A COMPARATIVE ANALYSIS OF THE MORPHOLOGY AND EVOLUTION OF PERMANENT SPERM DEPLETI ON IN SPIDERS ................................ ....................... 19 Background ................................ ................................ ................................ .......................... 19 Material and Methods ................................ ................................ ................................ ......... 22 Collectio n and Rearing ................................ ................................ ................................ 22 Sample Dissection and Preparation ................................ ................................ ......... 23 Measurements and Calculations ................................ ................................ ............... 24 Male Sperm Induction ................................ ................................ ................................ 25 Statistics ................................ ................................ ................................ ........................ 25 Character Optimization and Phylogenetic Comparative Analyses ...................... 26 Results ................................ ................................ ................................ ................................ .. 27 Mechanistic Basis of Permanent Sperm Depletion ................................ ................ 27 Evolut ionary History of Permanent Sperm Depletion ................................ ............ 29 Discussion ................................ ................................ ................................ ............................ 29 Final Remarks ................................ ................................ ................................ ...................... 34 3 MALE DENSITY AFFECTS LARGE MALE ADVANTAGE IN THE GOLDEN SILK SPIDER, NEPHILA CLAVIPES ................................ ................................ ............... 43 Background ................................ ................................ ................................ .......................... 43 Materials and Methods ................................ ................................ ................................ ....... 45 Study Sites ................................ ................................ ................................ .................... 45 Male Group Size Survey ................................ ................................ ............................. 46 Mating Experiment: Procedures ................................ ................................ ................ 46 Mating Experiment: Behavioral Observations ................................ ......................... 49 Statistical Analyses ................................ ................................ ................................ ...... 51 Results ................................ ................................ ................................ ................................ .. 52 Male Group Size ................................ ................................ ................................ .......... 52 Experiment: Behavioral Analysis ................................ ................................ ............... 52 Discussion ................................ ................................ ................................ ............................ 55
7 4 MORTALITY RISK AFFECTS MATING DECISIONS IN THE SPIDER NEPHILA CLAVIPES ................................ ................................ ................................ ............................ 67 Backg round ................................ ................................ ................................ .......................... 67 Study System ................................ ................................ ................................ ....................... 69 Methods ................................ ................................ ................................ ................................ 71 Male Removal Field Experimen t ................................ ................................ ............... 71 Mating Experiment ................................ ................................ ................................ ....... 72 Animal collection and housing ................................ ................................ ............ 72 Treat ments ................................ ................................ ................................ ............. 73 Male sperm counts ................................ ................................ ............................... 76 Parental genotypes ................................ ................................ ............................... 76 Offspring an alysis ................................ ................................ ................................ 77 Condition Index and Egg Development in Wild Caught Females ........................ 78 Results ................................ ................................ ................................ ................................ .. 78 Male Removal Experiment ................................ ................................ ......................... 78 Mating Experiment ................................ ................................ ................................ ....... 79 Discussion ................................ ................................ ................................ ............................ 81 5 MALE MULTIPLE MATING AND ALTERNATIVE TACTICS IN THE GOLDEN ORB WEB SPIDER NEPHILA CLAVIPES ................................ ................................ ..... 93 Background ................................ ................................ ................................ .......................... 93 S tudy System ................................ ................................ ................................ ....................... 96 Methods ................................ ................................ ................................ ................................ 98 Overview ................................ ................................ ................................ ....................... 98 The Basic Model ................................ ................................ ................................ .......... 98 Model structure ................................ ................................ ................................ ...... 98 Decision 1: search ................................ ................................ .............................. 100 Decision 2: stay ................................ ................................ ................................ ... 102 Size and season dependent strategies ................................ ........................... 103 Model Manipulations ................................ ................................ ................................ 104 Data An alysis ................................ ................................ ................................ .............. 104 Results ................................ ................................ ................................ ................................ 106 Male Mating Strategies ................................ ................................ ............................. 106 Basic model ................................ ................................ ................................ ......... 106 Restricted sperm model ................................ ................................ ..................... 107 Unrestricted sperm model ................................ ................................ ................. 107 Reprod uctive Outcomes ................................ ................................ ........................... 108 Basic model ................................ ................................ ................................ ......... 108 Model comparisons ................................ ................................ ............................ 109 Discus sion ................................ ................................ ................................ .......................... 111 Male Mating Rate ................................ ................................ ................................ ....... 111 Constraints on Mating Rate ................................ ................................ ...................... 112 Mat e guarding ................................ ................................ ................................ ..... 112 Male sperm limitation and choosiness ................................ ............................ 114 Male Body Size Variation ................................ ................................ ......................... 117
8 Final Remarks ................................ ................................ ................................ .................... 118 6 CONCLUSIONS ................................ ................................ ................................ ................ 127 Summary ................................ ................................ ................................ ............................ 127 Environmental Heterogeneity and Plasticity ................................ ................................ 128 Plasticity and Size Variation ................................ ................................ ..................... 130 The Link between Size and Behavior ................................ ................................ ..... 135 Evolutionary Consequences of Environmental Variation ................................ .... 137 Assumptions and Biases: Models Versus Data ................................ ........................... 138 Novel Insights for Behavioral Ecology ................................ ................................ ........... 142 Sperm Depletion ................................ ................................ ................................ ........ 142 Partner Mortality Risk ................................ ................................ ................................ 144 Final Remarks ................................ ................................ ................................ .................... 146 APPENDIX A SUPPLEMENTARY MATERIAL FOR CHAPTER 2 ................................ .................... 147 B SUPPLEMENTARY MATERIAL FOR CHAPTER 3 ................................ .................... 148 C SUPPLEMENTARY MATERIAL FOR CHAPTER 5 ................................ .................... 151 LIST OF REFERENCES ................................ ................................ ................................ ......... 154 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ..... 176
9 LIST OF TABLES Table page 2 1 Overview of sperm depletion in spiders as it pertains to sexual cannibalism and genital mutilation. ................................ ................................ ................................ .... 36 2 2 Hypotheses of the evolution of PSD. Hypotheses were tested using the concentrated change test for one of the phylogenies used by Miller (2007; ................................ ................................ .......... 3 7 3 1 Sample sizes across data types ................................ ................................ .................. 63 5 1 Parameter terms for the dynam ic state model. ................................ ...................... 120 5 2 Summary of male strategies by degree of male sperm limitation, size, and season. ................................ ................................ ................................ ........................... 121 A 1 Real age (numbe r of days after penultimate or final molt when sacrificed) and standardized age (age relative to the youngest sub adult male) for all males in the study ................................ ................................ ................................ .................... 147 B 1 Microsatellite loci used in pat ernity analysis ................................ ............................ 150 C 1 Parameter values that change with female type ................................ ..................... 151 C 2 Parameter value that change with season and female type ................................ 152 C 3 Of males that mated, the proportion of males within each size class that mated with each of three types of females. ................................ ............................. 153
10 LIST OF FIGURES Figure page 2 1 Testis volume vs. Deferent duct volume in Nephila clavipes ................................ 38 2 2 Trajectory of testis volume during male de velopment in Nephila clavipes .......... 39 2 3 Development of the male genital system in Nephila clavipes ................................ 40 2 4 Ultrastructural aspects of spermatogenesis in Nephila clavipes .. .......................... 41 2 5 Evolution of PSD in araneoid spiders. PSD is optimized on the phylogeny of Araneoidea according to Miller (2007; topolo ....... 42 3 1 Frequency distributions for male group sizes from field surveys of juvenile (top panel) and adult (bottom panel) female webs. ................................ .................. 64 3 2 Distribution of body size (cephalothorax width, mm) for male N. clavipes i n the mating experiment (N=139) ................................ ................................ ................... 65 3 3 Size ranks for fathers versus males that were not fathers in the 6 male treatment group Lower ranked males are larger ................................ ...................... 66 4 1 Average males per day versus males present at t 0 ................................ ................. 88 4 2 Body condition index versus days until oviposition. ................................ .................. 89 4 3 The proportion of P1 and P2 males observed mating for the non gravid (early) and gravid (late) treatment groups ................................ ................................ .. 90 4 4 P2 p aternity share versus treatment ................................ ................................ ........... 91 4 5 P2 pater nity share versus clutch number ................................ ................................ ... 92 5 1 Mean reproductive success and 95% confidence limits are shown for polygynous (Poly) versus monogynous (Mono) mating in the basic (A) and unrestricted (B) sperm models ................................ ................................ ................... 122 5 2 Mean repro ductive success and 95% confidence limits as a function of mate number for the basic (A) an d unrestricted sperm (B) models ............................... 123 5 3 Mean reproductive success with 95% confidence limits for mono gynous (Mono) and polygynous (Poly) mating outcomes divided by female type: virgin (V), mated (M), and gravid (G) ................................ ................................ ........ 124 5 4 The frequency of mating events where males chose to stay and guard the female for at least one day after copulation ................................ ............................. 125
11 5 5 Mean reproductive success with 95% confidence limits for small (S) and large (L) males in each of the three models, restricted sperm (A), basic (B), and un restricted sperm (C) ................................ ................................ ......................... 126
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philos ophy ENVIRONMENTAL HETEROGENEITY AND PHENOTYPIC VARIATION: THE EVOLUTION OF MALE BODY SIZE IN A GOLDEN ORB WEB SPIDER By Clare C. Rittschof May 2011 Chair: H. Jane Brockmann Major: Zoology Phenotypic variation is a guiding principle that underlies most major processes in evolutionary biology, including speciation, extinction, and adaptation. Here I address how environmental heterogeneity in selection pressure can act on a small spatial scale to maintain male body size variation in the golden orb web spi der Nephila clavipes Spider webs partition the environment into discrete patches, and for males who travel among female webs during adulthood in order to mate, female webs form a heterogeneous social environment. In the genus Nephila females build unusua lly strong and long standing webs, and males have a remarkable degree of variation in body size. Because large males have an advantage during male male contests, the maintenance of this broad distribution of sizes over evolutionary time is a puzzle. I addr ess how web to web variation in social environment could affect selection on male body size. To do this, I use field studies and mating experiments to investigate how female webs differ in two main characteristics that could affect the strength and directi on of selection on male size: 1) the number of male competitors that cohabit on the web with the female, and 2) the quality of the female mate herself. In addition, using a modeling approach, I address how males could employ size dependent strategies in
13 or der to optimize their reproductive success in a heterogeneous environment. In this model I incorporate a variety of factors that could constrain male reproductive behavior, including male sperm limitation. Female webs vary greatly in the number of male co mpetitors present on the web (i.e. male density). In addition, the strength of selection on male size differs with density. Large males have a competitive advantage only in intermediate to large sized male groups. When the number of competitors is low, sma ll males are as likely to achieve copulations as large males. Secondly, the quality of the female web owner varies from off between sperm competition risk and female mortality risk. Males inv est the most sperm when mating with a virgin female or a non virgin female who is close to oviposition. As a result of this increased sperm investment, males sire more offspring when mating with either of these female types compared to a non virgin female who is one month from oviposition. Furthermore, field studies show that the level of male competition at the web is a function of the quality of the female, which suggests that male density and female quality interact to affect male reproductive success in a size dependent way. Behavior and behavioral plasticity could play an important role in enabling small males to achieve similar fitness to large males if males can navigate the reproductive environment in an optimal way. However, for males of all sizes, male reproductive potential and male mating strategies could be constrained by male sperm limitation in this species. Using a histological and comparative phylogenetic approach, I assess the degree to which male mating rate could be limited by sperm deple tion in this species. Findings show that spermatogenesis is synchronous, and testes are non functional in
14 adult males. As a result, males have a finite quantity of available sperm, and once this sperm is depleted during one or a few copulations, males lose their ability to fertilize eggs. I incorporate male sperm limitations and other constraints to male reproductive success as well as environmental heterogeneity into a dynamic state model in order to determine how male strategies and reproductive outcomes change as a function of body size. The model shows that male strategies differ particularly in terms of post copulatory behavior, and large males are more likely than smaller males to stay behind and guard a female after copulation. Reproductive outcomes also differed between large and small males. Large males were slightly more likely to be monogynous compared to small males because they mate with virgin females at a higher rate and as a result deplete their sperm stores in a single mating. Thus, large ma le reproductive success is partially constrained by sperm limitation. Sperm limitation, however, did not promote choosiness in males unless it was very extreme (i.e. males were limited to a single mate). In this system, a variety of factors contribute to m ale reproductive success and male mating strategies, and as a result fitness differs more within a size class than between size classes. Males of all sizes achieve similar levels of reproductive success on average, which suggests that environmental heterog eneity may weaken selection on male body size by adding substantial within size class variation in fitness. A second interpretation is that males are able to use environmental heterogeneity to their advantage, navigating the reproductive environment in an optimal way for their size. Both of these scenarios could result in the maintenance of a broad distribution of male sizes in this species.
15 CHAPTER 1 INTRODUCTION Orb web spiders show some of the most extreme examples of sexual size dimorphism in animals. Many studies have addressed the evolution of this intersexual variation in size. Females can develop to large sizes in part because they are the sedentary sex (Fairbairn 2007) and lar ge size corresponds to increased fecundity (Prenter et al. 1999) Males are smaller, and maximum size may be constrained by the demands of moving among female webs as adults in order to mate (Fairbairn 2007) However, in addition to extreme sexual size dimorphism, members of the genus Nephila the largest web building spiders in the world, show very broad within sex size variation, but only in males (Vollrath 1980) In some species adult male size varies by an order of magnitude (Schneider & Elgar 2005) Because large males have a competitive advantage in almost every species in this genus, the maintenance of a broad distribution of male size s remains a puzzle. For arthropods in particular, body size is a complex trait that has a strong environmental component (Mousseau & Roff 1989) While there may be many proximate causes for size variat ion in Nephila size and size variation must have a heritable component in order for the trait to be subject to evolutionary change. Some studies in spiders have addressed the degree to which environmental and genetic components contribute to size variatio n. For example, in the Mediterranean tarantulas (Lycosidae), low variance in f emale size relative to variance in male size has been suggested as evidence for stabilizing selection on female size and weak directional selection on male size (Fernandez Montraveta & Moya Larano 2007) The most comprehensive treatment of heritability in growth and size parameters in a spider is a
16 study of a cellar spider (Pholcida e) by Uhl et al. (2004) They found a narrow sense heritability of 0.44 for body size (father son regression). Furthermore, this study found significant family effects on the response to diet supplementation in growth and adult size. In spiders generally, s ize as well as gene by environment effects on size appear to have significant heritable components, and thus size and size variation may be subject to selection in these species (Uhl et al. 2004; Fernandez Montravet & Moya Larano 2007). Nephila species als o show evidence of heritable variation in size. Environmental variation in food availability affects molt timing, but not necessarily adult male size, at least during the last three instars (Higgins 2000) Maximum size appears to be constrained by genetics and maternal effects (i.e. starting size; Higgins 1992; Higgins 1993, 2000) Furthermore, Higgins (1992, 1993, 2000) reports differences among populations within N. clavipes in average male size, which are presumably a function of genetic and environmental variation, or possibly an interaction between these two fo rces. Michalik and Rittschof (in press) found that a small sample of N. clavipes males collected in the wild and reared in a common garden lab setting had unusually long developmental times, and still showed a 1.5 fold range in size (cephalothorax width; M in = 1.43 mm, Max = 2.08 mm, S.D. = 0.14 mm), which suggests that maturation timing is plastic and broad, and at least some portion of the substantial variation in male size common to N. clavipes is due to genetic or gene by environment effects. In Nephila mal es, some component of size, and potentially the degree of plasticity in size, are heritable characteristics, with the potential to change over time in response to selection.
17 Previous studies that have investigated how natural selection could maintain body size variation in Nephila have attempted to sum and compare the costs and benefits for males that are large versus small (Vollrath 1980; Cohn 1990; Uhl & Vollrath 1998; Vollrath 1998; Schneider et al. 2000; Uhl & Vol lrath 2000; Schneider & Elgar 2001; Foellmer & Fairbairn 2005a; Schneider & Elgar 2005; Kasumovic et al. 2007) Here however, I take a broader approach by asking how, generally, natural selection acts to maintain phenotypic variation in any trait. Underst anding the selection processes that maintain or constrain phenotypic variation is a central goal for many fields within evolutionary biology (Charlesworth & Charlesworth 1987; Badyaev & Foresman 2000; Bijlsma et al. 2000; Robertson & Rosenblum 2009) In this study I focus on one feature that affects phenotypic variation, environmental heterogeneity. Environmental heterogeneity is an important force that can lead to speciation on large geographic scales. However here I specifically address whether heterogeneity on a smaller scale, web to web variation in selection on male size, may facilitate male body size variation in Nephila Environmental heterogeneity promotes adaptive phenotypic variation because it can shift the optimal level of a phenotype spatially or temporally, and in some cases completely reverses the optimal level of a phenotype (Candolin et al. 2007; Engstrom Ost & Candolin 2007; Wong et al. 2007; Gosden & Svensson 2 008) Environmental heterogeneity can occur at a variety of scales in space and time, and relevant environmental differences can be both abiotic (Grant & Grant 2002; Juenger & Bergelson 2002) and biotic (Svensson & Sinervo 2004) Variation in social environment in particular is expected to affect the strength and direction of selection on a fine
18 grained spatial scale (Levins 1968; Gosden & Svensson 2008) Web building spiders are a special case of spatial environmental heterogeneity because spider webs partition the environment into discrete patches (Agnarsson 2003; Kasumovic et al. 2008; Rittschof & Ruggles 2010) with different social characteristics. Female webs can vary in the number and size of male competitors, which in turn is a function of the female herself (i.e. the female males move from web to web searching for a female mate (Christenson & Goist 1979) and have the ability to modify their behavioral strategy in response to environmental variation. Here I use field and lab experiments as well as a modeling approach to investigate how spatial variation in female q uality affects male mating strategies, and may ultimately affect male body size evolution, in Nephila clavipes First, I demonstrate that N. clavipes males show an unusual degree of sperm limitation, a reproductive constraint that affects male mating strat egies. Second, I demonstrate that male competitive advantage is density dependent in this species, and that female webs are highly variable in terms of male density, creating spatial heterogeneity in the strength of sexual selection on male body size. Thir d, I show that male reproductive investment changes as a function of female age, resulting in variation in female quality associated with age and mated status. Last, I integrate the factors relevant to male mating decisions, including male sperm limitation variation in male male competition, and variation in female value, into a dynamic state model, and using this model, I assess the strength of selection on male body size, and describe the strategies and tactics males use to negotiate a spatially variable reproductive environment.
19 CHAPTER 2 A COMPARATIVE ANALYSIS OF THE MORPHOLOGY AND EVOLUTION OF PERMANENT SPERM DEPLETION IN SPIDERS Background Sperm are small and numerous compared to eggs, which has lead to the assumption that sperm are cheap to produ ce. However, recent studies have shown that sperm production is costly, and mal es can become sperm depleted (i.e. functionally sterile ) at least for some period of time after mating (Nakatsuru & Kramer 1982; Dewsbury 1983; Wedell et al. 2002) The amount of sperm transferred during copulation (Parker 1970) (Smith et al. 2009) As a result, sperm cost and male sperm depletion have broad im plications for the evolution of male mating strategies in sexually reproductive animals (Mller 1991; Simmons et al. 1993; Bateman et al. 2001; Engqvist & Reinhold 2006) Sperm depletion, the decrease in sperm numb er over successive ejaculates, is widespread in animals (Preston et al. 2001) but there is variation across species in the degree to which sperm depletion limits male mating opportunities (Radhakrishnan et al. 2009) Sperm depletion can be temporary, where males must undergo a reproductive latency period after mating in order to replenish their ejaculate (e.g. Lemaitre et al. 2009) In other cases, sperm depletion is permanent, where males are unable to replenish their sperm once it is used (Boivin et al. 2005) Because permanent sperm depletion (hereafter PSD) strongly constrains male mating ability the mechanistic, ecological, and evolutionary bases of this phenomenon are of special importance to a variety of research areas, including the evolution of mating systems, male mate choice, sperm competition, and female sperm limitation.
20 Web building sp iders are popular for male mating strategy studies because in some species males show a variety of behavioral and morphological features that limit mating rate, including male sacrifice behavior (i.e. sexual cannibal ism, e.g. Elgar 1991; Foellmer & Fairbairn 2003, 2004) and complete or partial genital breakage during copulation (genital mutilation, e.g. Schneider et al. 2001; Nessler et al. 2007b; Kuntner et al. 2009c) Mating mate are collectively called terminal investment strategies (Andrade & Kasumovic 2005) Because these behaviors occur across a variety of distantly related species (Miller 2007; Uhl et al. 2010b) many studies have examined the selection pressures that maintain these extreme behaviors, and whether these selection pr essures can be generalized across spider taxa (Schneider & Elgar 2001; Herberstein et al. 2002; Andrade 2003; Fromhage et al. 2003; Fromhage & Schneider 2006; Snow et al. 2006) However, some species of spiders ex hibiting terminal investment strategies also show evidence of PSD In order to interpret the evolutionary history, causes, and consequences of terminal investment strategies in spiders, it is necessary to verify the amount of sperm available to a male acro ss successive copulations (Herberstein et al. 2005; Michalik et al. 2010) Despite these broad implications, and although PSD was first suggested in spiders over 20 years ago (Christenson 1989) little is known about this phenomenon. Verifying whe ther PSD occurs, understanding its mechanistic basis, and assessing its evolutionary history are essential to interpreting the selection pressures driving terminal investment behaviors in this arthropod group. The male reproductive system in spiders is u nusual because males transfer sperm to the female using paired prosomal appendages called pedipalps that are
21 separate from the male genital opening. After the maturation molt, males ejaculate sperm through their genital pore onto a sperm web and draw the s perm into their pedipalps, a process called sperm induction (Foelix 1996) The sperm remains in the pedipalps until copulation. Studies that propose sperm depletion in spiders ha ve examined the pedipalps for the presence or absence of sperm (Christenson 1989) which would indicate temporary sperm depletion because males can re induct sperm before, d uring, or after copulation (Table 2 1) However, examination of the testes is required in order to demonstrate that sperm depletion is permanent. The male testes are simple, paired cylindrical organs that are connected to the genital pore by paired deferent ducts (Michalik 2009) Sperm is produced in the testes and then temporarily stored in the deferent duct s until ejaculation. Cursory examination of the adult male testes suggests that sperm depletion may result b ecause male testes atrophy or do not produce sperm during adulthood (Table 2 1; Herberstein et al. 2005; Michalik et al. 2010) In the current study, I verify that sperm depletion in spiders is permanent using the golden orb web spider Nephila clavipes. Spiders in the genus Nephila are known for genita l mutilation and male sacrifice behavior, which suggests permanent sperm depletion may be present in this group. More importantly, e xtensive mating experiments in N. clavipes provide the best evidence of any species studied that sperm depletion may be perm anent (Christenson 1989) Using morphological measurements from specimens sacrificed at different life stages, I examine how the size of the testes changes as males mature, induc t sperm, and age. I use light and transmission electron microscopy to confirm that decreased testes size corresponds to a decrease in the
22 amount of tissue devoted to sperm at ogenesis. In addition, using a comparative phylogenetic framework and phylogeny bas ed statistics we test the evolutionary relationships between PSD and other characters that limit male mating rate, including genital mutilation, male sacrifice behavior, and sexual size dimorphism. Material and Methods Collection and Rearing Third and f ourth instar juvenile Nephila clavipes were collected from mixed oak habitats within the Ordway Swisher Biological Station in Melrose, Florida (Putnam County). Juveniles were reared at the Lab of General and Systematic Zoology of the University of Greifswa ld in cylindrical plastic containers (5x10 cm). The tops of the containers were covered with cheesecloth, and the bottoms were open. The open bottoms sat on top of moistened towels to provide humidity, and males were housed together on a shelf under natura l light cycle at room temperature In the containers males constructed prey capture webs and were fed daily with two Drosophila flies. Because I was interested in comparing changes in the testes as males approach and pass their maturation molt, I sacrifi ced sub adult males across a range of times that spans the period between the sub adult and adult molt ( see Table A 1: Real age and standardized age for all males in the study ). Furthermore, I sacrificed adult males across a range of times that spans the pe riod from the maturation molt to death (N sub adult = 18, N adult = 19; see additional file 1) My experimental set up simplifies the true reproductive experience for adult males because we did not allow males to mate, nor did we provide females as cues for sperm production. However, in a series of behavioral experiments in this species, Christenson (1989) gave males the opportunity to copulate with multiple rec eptive females in succession, and males failed to transfer
23 any sperm in this scenario. This suggests that any processes occurring in the male genital system that limit re mating ability are independent of male mating history and female cues. Sample Dissect ion and Preparation Virgin male specimens were dissected in phosphate buffer (0.1M, pH 7.2) with 1.8% sucrose added (PB). The isolated genital systems were fixed in 2.5% glutaraldehyde ( Merck Chemicals Ltd., Nottingham UK) in PB and pictures for the analy ses of the gross morphology were taken using an Olympus DP10 digital camera mounted on an Olympus ZX 7 stereomicroscope. For the histological and transmission electron microscope (TEM) analyses samples were post fixed in PB buffered 2% OsO 4 (SERVA Electrop horesis GmbH, Heidelberg, Germany). After being washed in PB, the genital system was dehydrated in graded ethanol and embedded in Spurr's resin (Spurr 1969) For the light microscope (LM) a nalyses semi thin sections (700 nm) were made with a Diatome HistoJumbo diamond knife at a Leica ultramicrotome UCT and stained according to Richardson et al. (Richardson et al. 1960) Sections were documented using a Zeiss MCr digital camera mounted on an Olympus BX60 compound microscope. For the TEM analyses ultra thin sections (50 nm) were made with a Diatome Ultra 35 diamo nd knife at a Leica ultramicrotome UCT and stained with uranyl acetate and lead citrate according to Reynolds (Reynolds 1963) Examination was performed with a JEOL JEM 1011 electron microscope at 80 KV. Images were taken with a side mounted Olympus MegaView III digital camera usin g the iTEM software (Olympus Soft Imaging Solutions GmbH, Mnster, Germany).
24 Measurements and Calculations In order to compare changes in the genital system before and after the maturation molt, it was necessary to evaluate both male age groups along a si ngle is the age at which they were sacrificed relative to the sub adult males in the study ( Table A 1: Real age and standardized age for all males in the study ). For exam ple, the oldest sub adult male was sacrificed on day 72 after his sub adult molt, so the adult sacrificed one day after his own maturation molt received an age of 73. In order to determine how the male reproductive system changes with age, measurements o f body size, testis and deferent duct width and length, and the ratio of generative tissue to total testis tissue were taken for all males in the study (N = 37). All body measurements were taken from digital photographs (Zeiss Discovery V20 with Zeiss MCr camera) using the IntMess module in the program Zeiss AxioVision 4.8 ( Carl Zeiss MicroImaging GmbH, Gttingen, Germany ). A linear measurement of male body size was taken as the width of the prosoma at its widest point. In addition, after dissection, the le ngth and width of one testis and one deferent duct were measured per individual. For each individual, the most intact of each organ was selected for measurement. Because of their irregular shapes, total testis and deferent duct lengths were measured using the curve tool of the IntMess module, which traces along the length of non linear objects. Widths of the testis and deferent duct were taken as linear measurements. Because testis and deferent duct widths vary along the length of both organs, the width of each organ was measured at three locations per individual, the 25%, 50%, and 75% points along the length. These widths were averaged to estimate the true organ width. Using the assumption that the testis and deferent duct are
25 approximately cylindrical, the length and average width measurements were used to calculate testis and deferent duct volumes. To assess changes in generative tissue over male lifetime, I calculated the ratio of generative tissue to total testis area from a single stained testis cross section per individual. Cross sections were magnified using an Olympus BX60 light microscope and photographed with a Zeiss MCr camera. In spiders, spermatogenesis occurs in cysts, which are bordered by thin extensions of the somatic cells located at the p eriphery of the testis (e.g. Michalik et a l. 2006) At the end of sperm ato genesis, sperm cells accumulate in the lumen of the testis. Thus, I defined generative tissue as the total area of the testis filled with either spermatogenic cysts or lumen. The borders of the testis, the testis lumen, and the spermatogenic cysts were traced using the IntMess module to approximate area. The resulting testis cyst s area and lumen area were summed, and this value, divided by the total testis area, gives the ratio of generative tissue to total testis tissue. Ma le Sperm Induction In order to assess whether the process of sperm induction corresponds to changes in the testes, we assessed male pedipalps for the presence of sperm. To do this, for each adult male in the study (N = 19), I removed the left pedipalp and soaked it in clove oil (Sigma Aldrich Chemie GmbH, Munich, Germany) for 3 5 hrs and then examined the pedipalp under a light microscope. Using this method, we could visualize sperm through the pedipalp cuticle in order to determine sperm presence or absen ce. Statistics All statistical analyses were performed using the SAS program JMP 7.0 (SAS Institute Inc., Cary, NC, USA). For comparisons of testis and deferent duct volume for
26 sub adult versus adult males, the data were natural log transformed in order to normalize the data distributions, and analyzed using two tailed t tests. All other data were analyzed without transformation. Male generative tissue changes were analyzed using a Kruskal Wallis Test. Character Optimization and Phylogenetic Comparative Analyses In order to assess the evolutionary history of PSD and other male mate limiting (Miller 2007) phylogenetic hypotheses of araneoid spiders, which assessed the evolutionary history of genital mutilation, male sacrifice behavior, male accumulatio n on female webs (i.e. male biased operational sex ratio), and monogamy in this group. All analyses were based on the trees and character matrix of Miller (2007; see for detailed description of the phylogenetic hypotheses and characters:  genitalia mutilated during copulation (G),  sa crifice behavior (S),  sexual size dimorphism (D),  male accumulation (A),  male monogamy (M)) Additionally, I coded the character based on the present data and published and unpublished results (Michalik 2006; Michalik et al. 2006; Michalik & Hormiga 2010; Michalik et al. 2010; Michalik unpublished observations) as follows: 0, spermatogenesis ongoing in adulthood ; 1, spermatogenesis only in subadult stage The Miller (2007) phylogeny is based on morphological data, but given the available genetic data, at the time of the study, this tree was the most robust phylogeny, particularly for resolving relationships among the Nephilids. The character optimization was carried out wi th the Maximum Parsimony method (MP) implemented in the software Mesquite 2.7.2 (Maddison & Maddison 2010) To
27 (CCT; Maddison 1990) The CCT was run in MacClade 4.0 (Maddison & Maddison 2000) using 100 000 replicate simulations with the ancestral state unspecified and actual changes considered. We ana lyzed the correlation between PSD and genital mutilation, male sacrifice behavior, sexual size dimorphism, monogamy, and male accumulation based on the four phylogenies used by Miller (2007) Since the tests in all four phylogenies resulted in nearly ident ical p values we restricted Tab. 2 to the results based on the phylogenetic hypothesis shown in Fig. 2 5 (topology Kuntner et al. ((TA)(EN)) ; Miller 2007) Results In order to account for the effect of body size on testis size, I analyzed the variation in male body size and the relationship between body size and testis size. For males used in our analyses, body size showed limited variation; prosoma width ranged from 1.43 to 2.08 mm (Mean = 1.85 mm, SE + 0.023 mm), and regression analyses showed that prosoma width was not significantly correlated with either testis or deferent duct volume (F 1,35 = 0.022, R 2 = 0.0006, P = 0.88; F 1,35 = 2.55, R 2 = 0.068, P = 0.12). For these reasons, I did not account for body size variation in any analyses o f the genital system. M echanistic Basis of Permanent Sperm Depletion Testis volume was significantly greater for sub adult males compared to adult males (two tailed t test, t 35 = 5.89; P<0.0001; Fig. 2 1A). In contrast to testis volume, male deferent duct v olume did not differ between sub adult and adult males (two tailed t test, t 35 = 0.34; P = 0.74, Fig. 2 1B). Thus, for my analyses, I focused on processes in the testis.
28 Male testis volume changed dramatically with male standardized age (Fig. 2 2 A ). Within th e adult male group, testis volume showed a steep decline with male age that can be described as a negative logarithmic function (Fig. 2 2; Slope = 0.018, R 2 = 0.72). The further decrease in volume for adult males is related to the induction of the seminal f luid into the male pedipalps (Fig. 2 2 B ). The changes in testis volume corresponded to changes in the proportion of testis tissue devoted to sperm production. Combining all data, these changes over time can be described as three distinct phases (Fig. 2 3; Kr uskal Wallis Test, X 2 = 17.26, P<0.0002). In phase one (Fig. 2 3, left), testis volume increases, and the majority of the testis tissue is devoted to sperm production. The peak testis volume corresponds to early spermatogenesis during which germ cells divide and spermatids start to differentiate nearly synchronously The sperm cells in this phase are large and spherical (Fig. 2 4A). Over the course of spermatogenesis, the spermatids become more compact, indicated by the condensed chromatin in the nucleus and th e relatively small amount of cytoplasm in the cell (Fig. 2 4B). At the end of spermatogenesis, the main cellular components coil within the sperm cell (Fig. 2 4C), a process that occurs in all other spiders studied (Alberti 1990) The loss of volume of the spermatids due to the condensing and coiling processes induces phase two (Fig. 2 3, middle), where the spermatogenic cysts around the sperm cells contract, resulting in shrinking of the testes and a decrease in generative tissue. By the end of phase two, sperm cells have moved from the contracted cysts into the lumen of the testes (Fig. 2 4D) and they subsequently move into the lumen of the deferent duct. Sperm induction initiates phase three (Fig. 2 3, right); seminal fluid is discha rged out of the testes and deferent duct through the
29 genital opening. The absence of seminal fluid in the lumen results in further shrinking of the testes and near absence of generative tissue. The seminal fluid consists of spermatozoa and secretion (Fig. 2 4D), which is produced by the somatic cells of the testis. Evolutionary History of Permanent Sperm Depletion The optimization of PSD suggests at least three independent origins of this trait in araneoid spiders i.e. within Nephilidae Araneidae and T heri diidae ( Fig. 2 5 ). Based on the lack of behavioral data (scored as missing data, see also Miller (2007)), the presence of PSD in Clitaetra and Herennia (Nephilidae) as well as in the closest related taxa Argiope (Araneidae) is ambiguously optimized and shou ld be addressed in future studies. As indicated by the CCT, the evolution of PSD is significantly correlated to and dependent on genital mutilation, male sacrifice, male monogamy, male accumulation, and sexual size dimorphism (Table 2 2 ). Discussion Nephi la clavipes sperm maturation occurs only during the sub adult instar, and spermatogenesis is completely absent in adult males. In addition, after the maturation molt, all sperm is inducted into the pedipalps at one time (Myers & Ch ristenson 1988) As a consequence, the amount of sperm available to males for mating is limited to the sperm contained in the male pedipalps, and once it is used, males lose their ability to fertilize eggs. These data strongly suggest that when males depl ete their sperm, it is permanent. PSD, characterized by nearly synchronous spermatogenesis and termination of spermatogenesis in the sub adult instar, is unusual in spiders. In most spiders, spermatogenesis is ongoing throughout adulthood (Michalik & Uhl 2005) and all stages
30 of spermatogenesis are observed in the testes at the same time (Michalik & Huber 2006) Moreover, in most species, males generally recharge their pedipalps before, after or even during courtship and copulation (Huber 1998; Knoflach 2004) indicating active sperm production in adult males PSD occurs in distantly related spider groups (Fig. 2 5). It has been documented in two s pider families, Theridiidae (Michalik et al. 2010) and Nephilidae ( N ephila clavipes and N. senegalensis; Nephilengys malabarensis and N. borbonica ; this study and Michalik and Kuntner unpublished data). In addition, it has been suggested to occur in a third family, Araneidae (Herberstein et a l. 2005) Because these three families of spiders are distantly related, and because PSD occurs in some but not all members of these families, (Fig 2 5; Michalik, unpublished data), it is likely that PSD has evolved independently at least three times (Fig 2 5 ) One plausible hypothesis for this distinct distribution of PSD is that the trait evolved in correlation with other traits that limit male reproductive rate (Miller 2007) Terminal investment behaviors, which eliminate male re mating ability (e.g. male sacrifice behavior and genital mutilation; An drade & Kasumovic 2005) o ften co occur (Miller 200 7) and PSD appears to have evolved in lineages following the evolution of traits asso ciated with male monogamy (Fig. 2 5, Table 2 2). Data suggest that in lineages with high levels of male competition (i.e. male accumulation; Fig. 2 5; Miller 2007), which is o ften a consequence of extreme sexual size dimorphism (Miller 2007) males are limited to monogamy, and genital mutilation and male sacrifice behaviors have evolved as mechanisms of paternity assurance (Fig. 2 5; Miller 2007; Uhl et al. 2010b) As a consequence, unlimited sperm production became unnecessary in these
31 groups because males typically mate once (Knoflach & van Harten 2001; Foellmer & Fairbairn 2003) The evolution of PSD may be an energy saving mechanism that has evolved in some groups with terminal investment strategies and monogamy because it ving copulations with a female. Sperm production and maintaining the function of the testes is energetically costly (Van Voorhies 1992) particularly if males do not feed as adults. In the moth Plodia interpunctella starved males show decreased sperm numbers (Gage & Cook 1994) Similarly, in the adder Vipera berus males lose as much mass during periods of sperm production (when males are immobile) as they do during periods of active mate search (Olsson et al. 1997) In spiders, extreme sexual size dimorphism, which is a consequence of fecundity selection on females (Coddington et al. 1997; Hormiga et al. 2000; but see also Corcobado et al. 2010) results in large sedentary females and mobile searching males. In species that require webs for prey capture, males that search for females cannot build webs and so typically do not eat during adulthood (Christenson et al. 1985) which makes them vulnerable to starvation. The hypothesis that PSD is an energy sa ving adaptation has some support in the genus Tidarren (Theridiidae) Males in this genus have unusually large pedipalps for their body size and so they castrate one of their pedipalps prior to sperm induction (Knofl ach & van Harten 2000, 2001) Pedipalp removal increases male locomotor performance, giving males more stamina during mate search, and allowing males to find femal es more quickly (Ramos et al. 2004) The high locomotor performance that is characteristic of Tidarren is unusual among s pider species (Ramos et al. 2004 ) Thus in this genus one result of increased energetic demands could be that males divert oxygen
32 and other resources away from t he testes and into the muscles leading to a loss of testes function (PSD; Michalik et al. 2010). The need for physical endura nce is one characteristic that Tidarren appears to share with another sperm depleting group, Nephilidae. M ost of the Nephilid spiders have high post copulatory energy requirements. Male sacrifice behavior is rare but genital mutilation is common, limiting male re mating ability (Fig. 2 5; Miller 2007; Kuntner et al. 2009b) However, males guard females after mating, which decreases the probability that the female w ill re mate (Christenson & Goist 1979; Christenson et al. 1985; Linn et al. 2007) The termination of spermatogenesis might be an energy saving measure that allows these males, who no longer have functional pedipalp s, to spend extended periods of time fighting after copulation (up to two weeks in N. clavipes ; (Christenson et al. 1985). In contrast, in the genus Latrodectus (Theridiidae), the widows male sacrifice behavior is common, but preliminary results suggest that PSD does not occur in this genus (Michalik et al. unpublished data for L. hasselti L. hesperus and L. geometricus ; Table 2 1 ). However, in Latrodectus even when males are not cannibalized during copulation, they die soon after mating (Forster 1992) Thus Latrodectus species do not appear to share the level of physical stamina and survivorship found in Tidarren and the Nephilids, possibly because Latrodectus species do not have the same pre and post copulatory energetic demands. This preliminary hypothesis might explain the conspicuous absence of PSD in Latrode ctus even though these species are well known for high rates of genital mutilation, male sacrifice behavior, and monogamy (Andrade 2003; Stoltz & Andrade 2010) Future studies should address this problem.
33 In the third family of spiders that show PSD, the Araneids, PSD is only found in Argiope (Herberstein et al. 2005) one of the few genera in this family with high rates of genital mutilation (Ness ler et al. 2007a, 2008) and male sacrifice behavior (Foellmer & Fairbairn 2003, 2004) Although extreme sexual size dimorphism is common in the Araneids, it does not typically lead to monogamy in this group (Miller 2007) which may explain the rarity of PSD in the family. Thus although sexual size dimorphism can lead to monogamy in some cases (Fromhage et al. 2005; Miller 2007) the patterns in the Araneids suggest t hat either genital mutilation or male sacrifice behavior, which impose the strongest constraints on male re mating opportunity, must be present for PSD to evolve. Further analyses of the A raneids should test for PSD more broadly to determine whether it occurs only in groups with genital mutilation or sacrifice behavior, or if extreme sexual size dimorphism alone (which only somewhat constrains mal e mating rate) is sufficient to favor its evolution. It is important to note that there is variation in the occurrence of genital mutilation within the family Nephilidae, even though PSD may be present in all species ( see Fig 2 5 ). In Nephila fenestrata males commonly break off the distal tip of the embolus during mating rendering the pedipalp useless (Fromhage & Schneider 2006) while in other species (e.g. N. e dulis ), embolus breakage is rare or absent (Kuntner et al. 2009b; Uhl et al. 2010b) In N. clavipes specifically, the evidence for embolus breakage is equivocal (Kuntner et al. 2009b; Uhl et al. 2010b) and recent work in Nephilengys borbonica has shown that genital mutilation is a labile trait that occurs only in certain mating contexts (Kuntner et al. 2009a)
34 In the Nephilids, character optimization and CCT suggests that genital mutilation preceded the evolution of PSD, and mutilation was secondaril y lost in certain species (Kuntner et al. 2009b; Fig 2 5 Table 2 2 ). The loss of genital mutilation could have from mating multiply, or because females escaped male plugging beh avior over evolutionary time (i.e. antagonistic co evolution of genitalia; Kuntner et al. 2009b). However, while males in these species may have regained the ability to maintain intact pedipalps during copulation, they may not be able to regain functional testes once PSD has evolved. Instead, in some species (e.g. N. clavipes ) males may prudently allocate sperm, particularly when mating with non virgin females (Christenson & Cohn 1988) Thus it remains unclear whether PSD results in monogamy in all cases. Future studies should confirm the occurrence of PSD in Nephilids, determine the relationship between PSD and male mating rate, and examine the evolutionary lability of PSD relative to other terminal investment behaviors. The Nephilid spiders, which show a broad range of mating systems and terminal investment behaviors, make an ideal group for this compar ative study. Final Remarks Here we verify that sperm depletion in Nephila clavipes is permanent, and describe its mechanistic basis. Cursory studies in other species suggest that spiders exhibiting PSD share this common mechanism (Michalik et al. 2010) Although PSD is an unusual phenomenon, it appears to have evolved multiple times in association with genital mutilation, male sacrifice behavior, and other traits associated with monogam y (e.g. sexual size dimorphism). In general, PSD could be an energy saving adaptation, although the factors favoring it (e.g. pre copulatory mate search or post copulatory mate
35 guarding) may be species specific. Future work will explore the costs of sperm production, the energetic benefits of PSD, and employ a broad comparative phylogenetic approach to address the relationships between PSD, male terminal investment behaviors, and environmental factors that constrain male mating rate in spiders.
36 Table 2 1. Overview of sperm depletion in spiders as it pertains to sexual cannibalism and genital mutilation. Family Species PSD TSD Sexual cannibalism Genital damage Source Nephilidae Nephila clavipes yes yes F no no present study; (Christenson 1989) Nephila plumipes ? no yes yes, partial genital breakage (Schneider & Elgar 2001; Schneider et al. 2001; Schneider et al. 2008) Nephilengys malaba riensis yes ? yes yes, emasculation (Danielson Franois 2006) Tetragnathidae Glenognatha emertoni ? no no no (Danielson Francois & Bukowski 2005) Tetragnatha versicolor ? no no no (Herberstein et al. 2005; Uhl et al. 2010b) Araneidae Argiope keyserlingi yes yes y es yes, partial genital breakage (Schneider et al. 2005b; Schneider et al. 2006) Argiope bruennichi ? yes T yes yes (Bukowski & Christenson 1997b) Micrathena gracilis ? yes T no no (Bukowski et al. 2001) Gasteracantha cancriformis ? yes T 14.7 %** no (Andrade & Banta 2002; Snow & Andrade 2004) ; Michalik and Andrade (unpublished) Theridiidae Latrodectus hasselt i no yes yes yes (Knoflach 2004; Molina & Christenson 2008) Netiscodes rufipes ? yes T no no (Knoflach & van Harten 2001; Michalik et al. 2010) Tidarren argo yes ? yes yes, emasculation (Costa 1998) Lycosidae Schizocosa malitiosa no* no no no (Huber 1995; Eberhard 2004) Anyphaenidae Anyphae na accentuata ? no no no (Danielson Franois 2006) In order t o summarize direct evidence for sperm depletion in spiders, only studies that explicitly addressed sperm usage by sperm counts were considered. PSD (permanent sperm depletion; ye s/no): In cases where the testes were examined, was PSD confirmed? TSD (tempor ary sperm depletion; yes/no): In cases where the pedipalps were assessed for the presence of sperm after copulation, were the pedipalps absent of sperm? The superscript T denotes that the occurrence of TSD depends on the time spent copulating, and F denote s that the occurrence of TSD depends on the status of the female mate (i.e. virginity). Note that TSD does not necessarily suggest PSD since spider males usually re induct sperm. *males that were able to re induct fathered significantly more offspring than those that were prevented from re inducting, **male is killed by female but not consumed.
37 Table 2 2. Hypotheses of the evolution of PSD. Hypotheses were tested using the concentrated change test for one of the phylogenies used by Miller (2007; Independent Dependent p value g, l of dep. (g ind.) PSD genital mutilation 0.52 +10, 4 (+3) PSD male sacrifice 0.49 +6 (+2) PSD male monogamy 0.086 +7, 1 (+3) PSD male accumulation 0.166 +6, 1 (+3) PSD size dimorphism 0 .25 +7, 5 (+3) genital mutilation PSD 0.006 +3 (+3) male sacrifice PSD 0.0018 +3 (+2) male monogamy PSD 0.00069 +3 (+3) male accumulation PSD 0.005 +3 (+3) size dimorphism PSD 0.03 +3 (+3) When optimization was ambiguous, accelerated transformations were preferred to minimize the number of gains in the character of interest. The right column indicates the total number of gains (g) and Losses (l) of the dependent character (dep). The number of gainesv in the presnce of the independent character (ind) is given in parentheses.
38 Figure 2 1. Testis volume vs. Deferent duct volume in Nephila clavipes (A) Testis volume in sub adult versus adult males (N subadult =18, N adult = 19; F 1,35 = 34.7; P<0.0001). (B) Deferent duct volume in sub adult versus adult males (N subadult = 18, N adult = 19; F 1,35 = 0.11; P = 0.73).
39 Figure 2 2. Trajectory of testis volume during male development in Nephila clavipes The adult stage (box in upper graph) is enlarged to show the decrease in testis volume after sperm inducti on (Slope = 0.018, R 2 = 0.72).
40 Figure 2 3. Development of the male genital system in Nephila clavipes Three phases in the development of the male genital system (left to right) characterized by a dorsal view of the whole genital system (top, scale: 1m m), a stained cross section of the testis (middle), and the analysis of the proportion of the generative tissue in testis (bottom; Kruskal Wallis Test, one way; X 2 = 17.26, P<0.0002).
41 Figure 2 4. Ultrastructural aspects of spermatogenesis in Nephila cl avipes (A) Early spermatids. (B) Longitudinal section of mid stage spermatids with condensed and elongated nuclei. (C) Coiled spermatids in a spermatogenic cyst. (D) Coiled sperm cells and secretion in the lumen of the testis. Abbreviations: AV, acrosomal vacuole; AX, axoneme; CyL, lumen of spermatic cyst; N, nucleus; SC, somatic cell; Sec, secretion.
42 Figure 2 5. Evolution of PSD in araneoid spiders. PSD is optimized on the phylogeny of Note the ambiguous optimization in the Nephilidae node and within Araneidae. Boxes indicate states of the following characters: permanent sperm depletion (P), male sacrifice behavior (S), male genital mutilation (G), extreme sexual size dimorphi sm (D), male accumulation (A), and male monogamy (M). The pictures in the right column show examples from all genera where PSD is present indicated by the equal diameter of the deferent duct and the testis in the male reproductive system ( T. argo from Mich alik et al. 2010).
43 CHAPTER 3 MALE DENSITY AFFECTS LARGE MALE ADVANTAGE IN TH E GOLDEN SILK SPIDER, NEPHILA CLAVIPES Background Understanding the factors that either constrain or maintain phenotypic variation is a fundamental puzzle in biology (Kingsolver & Pfennig 2007) The evolution of body size in particular is of interest because body size affects many aspects of survival and reproduction (e.g. Badyaev 2002; Lafferty & Kuris 2002; Hone & Benton 2005) For males, large male advantage in male male contests is common and pla ys an important role in body size evolution (Andersson 1994) However male body size is a trade off between male male competition and other factors such as growth rate, mate searching efficiency, and predator avo idance. In many species, even though large males may have an advantage during contests, body size is under stabilizing selection, and variation in size is constrained because very large or very small sizes are selected against (Falconer & Mackay 1996; Blanckenhorn 2000) In contrast, in other species, male body size is highly variable, and a wide range of sizes coexist ove r evolutionary time (e.g. Schneider & Elgar 2005) In sexu ally dimorphic species, s election pressures differ between the two sexes to maintain divergent optimal male and female sizes (Fairbairn 2007) However, stabilizing selection that maint ains these size optima is predicted to decrease within sex size variation. In contrast, in some dimorphic species, a broad range of male body sizes is maintained in a single population over time. One factor that could maintain male size variation is densit y dependent selection, where large male competitive advantage changes with density (Knell 2009) Increased male density can have a positive or a negative effect on large male advantage. In some cases, when male density is high and
44 male male interactions are chaotic, females are difficult to monopoliz e, resulting in scramble competition that favors small males that find mates quickly (Conner 1989; McLain 1992) or efficiently (Fairbairn 2007) In other cases high male density excludes small males and favors large aggressive males (Zeh 1987; Bertin & Cezilly 2005) limiting small male reproductive opportunities to contexts where male interactions are inf requent. I f density varies spatially or temporally, and large male advantage is density dependent then selection will favor high variance in body size within a population. Size variation may be maintained because the optimal male size shifts with density or because extreme sizes are favored in different contexts (e.g. alternative tactics). Spiders of the genus Nephila are regularly featured in studies that address the evolution of body size due to their strong sexual size dimorphism (e.g. Higgins 1992; Head 1995; Hormiga et al. 2000; Schneider & Elgar 2001; Foellmer & Fairbairn 2005b) In addition, male body size in Nephila shows unusual ly high variation (Vollrath 1980) as much as an order of magnitude in N. edulis (Schneider et al. 2000) Because larger male Nephila have a compe titive advantage and typically win male male agonistic encounters (Christenson & Goist 1979; Vollrath 1980) the persistence of a high degree of male size variation ha s received special attention. The prevailing hyp othesis for the maintenance of variance in male body size in Nephila populations is that trade offs associated with different male sizes result in an intermediate optimal body size (e.g. Vollrath 1980) However, stabilizing selection on an intermediate body size should erode genetic variation over time (Falc oner & Mackay 1996) not maintain the high variance that is characteristic of Nephila An alternative hypothesis for high variance in size is that seasonal variation in population density and operational sex ratio creates
45 density dependent selection gradi ents on male size (e.g. Kasumovic et al. 2008) In addition to seasonal variation in population density, local male density (i.e. male group size season. Variation in male density could maintain the high degree of variation in male body size if increased density changes the strength or direction of sexual selection on male size (e.g. Danforth & Desjardins 1999; Emlen et al. 2005) Here I investigate whether variation in local male density could contribute to body size evolution in male Nephila clavipes In these web building spiders female webs are discrete competitive patches where males fight amongst themselves for opportunities to mate. I surveyed natural variation in male group size on female webs and tested how group size affect ed the reproductive success of males in a controll ed mating experiment. I address how male group size affect ed male male competition, the relative size s of successful father(s), and female re mating rate s Materials and Methods Study Sites The male group size survey and mating experiment took place on a r anch in Jonesville, Florida (latitude 29.654, longitude 82.523). This site contains pasture as well as oak forest with palmetto understory and has a high density of N. clavipes Survey data were collected along the perimeter of a hay field, which was boun ded on all sides by oak forest. Females and males used in the mating experiment were collected from other forested sites on the property. The mating experiment was conducted in and around a covered but open 3.5m x 3m x 4m shed in the oak forest, which prov ided shelter from the rain and maintained humidity and temperature similar to the surrounding forest.
46 Male Group Size Survey I surveyed male group sizes on 1 August 2007 between 1300 and 1400 hrs checking all webs along the perimeter of the hay field (N = 129). I chose this sampling method in order to get the best idea of male group size variation during the peak of the breeding season when adult males are in highest abundance (C. Rittschof, unpublished data). Because males do not change webs during the daytime, sampling at one time point during the day is an accurate measure of the group size on that day (Christenson and Goist 1979). To be included in my survey, webs had to: 1) contain either an adult female or a juvenile female within 1 2 instars of mat urity; 2) be within 3 m of the ground; and 3) have at least one male present. I determined female maturity using genital morphology (Higgins 20 00) and estimated the age of juvenile females by visually approximating their body size (about 5 6 mm cephalothorax width). I excluded smaller juvenile females because these webs typically do not contain males (Cohn et al., 1988). For each web, I recorde d the number of males present. Mating Experiment: Procedures I tested the effects of male group size on male reproductive success in a mating experiment conducted between 18 July 2007 and 23 September 2007. I conducted the mating experiment in cages to e nsure that each female was a virgin at the start of the terminal molt and left the males to mate for two days. I used two treatments, a 2 male group size and a 6 male group siz e. The 6 male group size was the maximum group size observed on the webs of virgin females (Farr 1977; T. Christenson, pers. comm.) I collected p enultimate instar females (N = 47) and housed them individually in 90 cm x 90 cm wood frame cages 15 cm deep. Cages were built from pressure treated
47 er Seal and covered with fiberglass screening (1 mm mesh size ). One corner of the screen on one side of each cage was lined with Velcro so that I could detach the screen to feed the spiders and add and remove males. Females built and maintained normal pr ey capture webs in these frame cages. I sprayed a water mist on webs daily and fed each female 1 large mealworm (approximately 2 cm in body length) per day for the duration of the experiment (Christenson & Cohn 1988) Through most of the season, I collected mature males for the mating experiment from juvenile female webs. After their terminal maturation molt, males have dark body coloration and sclerotized pedipalps (Myers & Christenson 1988) Because males were current fem ale web owner, but I could not ensure virginity because males travel between webs ( Cohn et al. 1988) and a male could have mated with a different female prior to collection. Towards the end of the season, juvenile females and males became scarce, and so two trials from each treatment used males collected from adult female webs Althoug h mating experience may have affected male behavior, due to my method of collection, any effects of mating experience are randomized across the two treatments. Each day male collection was required, I collected a minimum of 8 males at a time ( enough mal es for 1 trial from the 2 male treatment and 1 from the 6 male treatment) in order to standardize any effects of collection date across the two treatment groups. I collected a range of male body sizes, and males were assigned to treatments in the order of collection Because previous studies have shown that male dominance
48 on female webs is a function of relative and not absolute body size (Christenson & Goist 1979) I did not control for absolute body size in my treatments. Male body size was determined by measuring cephalothorax width, and each male within a treatment trial was assigned a size rank starting with r ank 1 for the largest male. The length of the first tibia patellar (TP) leg segment can influence the outcome of male contests in spiders (Foellmer & Fairbairn 2005b) However, I validated that TP length for all legs scales linearly with cephalothorax width for males of this species (C. Rittschof, unpublished data). This suggests that cephalothorax width alone is adequate to describe the size differences among males, regardless of which element of body size gives large males a enamel paint. I housed males individually in 9 46 mL Tupperware containers with a portion of the lid cutout and replaced with fiberglass screening. I sprayed the males with water daily until the initiation of the mating experiment (mean waiting time = 3.1 d, S.E. = + 0.06). Because mature males do not build prey capture webs and eat only occasionally as adults on female webs (Cohn & Christenson 1987) I did not feed the males between ca pture and the initiation of the mating experiment. No males died while awaiting the beginning of the mating experiment. Every morning, I checked female cages to determine if any females had molted. If a female molted, I verified she was mature by genital morphology (Higgins 2000) and initiated the experiment between 0800 and 1200 hours. As females matured, I assigned them to a treatment group, alternating between the 2 male and 6 male treatment. When more than one female molted on a single day, I started one trial at a time in order to
49 observe and record male behaviors for the first hour after males were introduced to the uced males sequentially in order from smallest to largest in less than 1 min. Males were left in the cage for 48 h to mate with the female. I did not feed behavior (Christenson et al. 1985) If a male died before the experiment was completed, I removed him and froze his body immediately. After 48 h, I removed all remaining males and killed them by placing them in a 20 C fr eezer for tissue preservation. After mating, females were housed in their cages and fed and misted as they were prior to the mating experiment until they laid their first clutch of eggs. Immediately after clutch production each female was collected, killed measured, and preserved in the same way as the males. I left egg c lutches where they were laid and collected them after the offspring hatched. Most clutches were collected and preserved in 95% ethanol after the offspring molted to their second instar. Ho wever 6 clutches were preserved before their second instar because they were at risk of predation by ants. Paternity analysis was performed using 3 polymorphic microsatellite markers developed for this species. The loci had 2, 3, and 18 alleles (see Appe ndix B). I used the maximum likelihood calculation from the program GERUD 2.0 (Jones 2005) to determine paternal genotype(s) using the known maternal genotype and the genotypes of 24 progeny per clutch (se e Appendix B). Mating Experiment: Behavioral Observations I recorded male male aggression and mating behavior for the first hour of each experimental trial. A behavior was considered an aggressive act only when the male performing the behavior was orien ted in the direction of another male. Using th is
50 criterion, any behaviors or interactions directed towards the female were excluded. When males encounter one another, one or both males perform a series of behaviors that escalate from long range threats (wh en males are as much as 20 cm apart) to physical contact (Christenson & Goist 1979) There are three basic sta ges of escalation (adapted from Christenson & Goist 1979) : (1) strand plucking, where one male turns to face the other male, shakes his body from side to side and jerks the web with his front pair of legs (Christenson et al. 1985) ; (2) chasing, where the aggressor moves forward with his front legs raised in response to another male who either flees from or turns toward the aggressor in preparation for a fight; (3) grappling (Cohn et al. 1988) which is where males come into contact and hit one another with the front pairs of legs, attempt to bite, and push with the front legs to knock the opponent off of the web (Christenson & Goist 1979) Other male behaviors, such as avoidance or running away from other males were not tal lied as aggressive behaviors. I recorded any male male interaction escalating threatening and fighting behaviors. This general definition of male fighting behavior was necessary because the transition from strand plucking through grappling can occur in a matter of seconds. I recorded the male who initiated each challenge, the identity of the male targeted in the challenge, and the winner and loser of the challenge. The male who initiated the challenge was the first male to behave aggressively in web, or chased another male as he approached the female (after Christenson & Goist 1979) The loser was the male that first turned away or fell off of the web during a grapple. I the male who initiated the most
51 challenges within his group. In order to test if males in the 6 male treatment behaved more aggressively than those in the 2 male treatment, I calculated the number of male male chal lenges per male. I measured mating behavior for the first hour of each trial. During mating, a male her epigynum (Ch ristenson et al. 1985) Because pedipalp insertions were difficult to see through the cage screen, I used mounting as a measure of mating attempts. A male was considered mounted when he climbed onto the ventral surface of the female's abdomen. I recorded the latency to any male mounting the female, the identity of the first male to mount, and the cumulative time a male spent in the mounted position. A dismount was scored if the male lost contact with the female. Statistical Analyses Both behavioral and pa ternity analyses comparing treatment groups are unpaired analyses. For standard parametric and non parametric statistical tests (Wilcoxon Signed Rank test, T tests and Chi Squared test) I used the program JMP 7.0 (SAS Institute, Cary NC). However, in order to determine if body size affects either male dominance or paternity success in a 6 male group, I needed to test whether large males were dominant males (or fathers) more often than would be expected if size had no effect on dominance and paternity. Due t o non independence, I could not directly compare male size ranks for dominant and non dominant males (as well as fathers versus non fathers) within each treatment group using a standard non parametric test. Instead, I tested whether, across all 6 male tria ls, the sum of the body size ranks for dominant males (or fathers) was different from a sum of size ranks sampled randomly.
52 The mean (m) of six ranks is 3.5 with a variance (v) of 3, and t he distribution of the sum of N number ranks is approximately norm al (confirmed by simulation). I tested the sum of the ranks for behaviorally dominant males (or fathers) in the 6 male treatment against a normal distribution of sums with mean N*m and standard deviation the square root of N*v. I computed a one tailed p va lue for the hypothesis that dominant males (or fathers) fathers have a larger rank than the random expectation. Results Male Group Size On juvenile female webs (N = 75), male group size ranged from 0 6 males (Mean = 1.48, SE = + 0 .16; Fig 3 1 top), 65.3% o f webs had 0 1 males, 17.3% had two males, and 17.3% had 3 or more males. On adult female webs (N=54), male group size ranged from 0 16 males (Mean = 2.4, SE = + 0 .36; Fig 3 1 bottom) 46% of webs had 0 1 males, 24% had 2 males, and 30% had three or more ma les. Overall there were significantly more males on adult female webs (Wilcoxon Signed Rank Test, Z = 2.29, P = 0.022). Experiment: Behavioral Analysis Due to observation there are behavioral da ta for 35 total mating trials across the 2 male and 6 male treatment groups, and there are paternity data for 29 trials in which f emales laid viable clutches (Table 3 1). Males used in the mating experiment spent 0 11 days in captivity before they entered the experiment, and there was no difference between the two treatments in captivity time prior to use in the experiment (Mean 2 male = 3.2 d, S.E. = 1.0 d, Mean 6 male = 2.9 d, S.E. = 0.81 d, two tailed T test, t = 0.18, P = 0.86). Time in captivity did not
53 affect male behavior. T here was no relationship between the number of days males spent in captivity and the total number of male male challenges ( F 1,33 = 0.012, P = 0.91). In addition, t he time before first mount was independent of time in captivity (F 1,3 3 = 1.43, P = 0.08), and the total time spent mounted was independent of time in captivity ( F 1,33 = 1.15, P = 0.21). Within the first hour of observation, c hallenges occurred in 83 % of 6 male trials (N = 15 of 18) and 35% of 2 male trials (N=6 of 17) a significant difference (Chi Squared Test, P<0.0021). The number of challenges in the 2 male group ranged from 0 to 6 (Mean = 1.1, S.E. = 0.46), and in the 6 male group from 0 to 12 (Mean = 4.9, S.E. = 0.9). Across all trials (N = 35), t he total challenges calculated per male did not differ between the treatment groups ( Mean 2 male = 0.56, S.E.= 0.23, Mean 6 male = 0.82, S.E. = 0.15, two tailed T test; t 3 = 0.97, P = 0.34 ). Male body size had a unimodal distribution and ranged from 1.3 mm to 3.45 mm (Mean = 2.25 SE = + 0.04, N = 139; Fig 3 2). The range of male body sizes represented within a single trial did not differ between treatments (Mean 2 male = 0.71, S.E. = 0.09, Mean 6 male = 0.94, S.E. = 0.09, two tailed T test, t 33 = 1.9, P = 0.07). In addition, abso lute male body sizes within each 6 male treatment trial were evenly distributed across the size range represented in the trial; the average difference in absolute body size between each size rank within a trial was 0.20 mm (S.E. = 0.02 mm; range = 0.08 mm to 0.3 mm). Across all challenges observed in this study, I observed 39 challenges in which there were clear winners. Of these contests, l arger ranked males won contests more often than smaller ranked males ( Chi Squared Test X 2 = 11.3 P = 0.0008). Howe ver,
54 b ecause males did not always respond to challenges, a winner was determined for only 39 of 76 total challenges observed across all trials (51%). In the 6 male treatment group, behaviorally dominant males (N = 11) were larger ranked males than predicte d under the hypothesis that male size has no effect on dominance (P<0.001). The mean rank for dominant males was 1.9 (S.E. = 0.47) and for non dominant males the mean rank was 3.8 (S.E.=0.21). In the 2 male treatment, the dominant male was the larger male in all cases in which male interactions occurred (N = 4 trials where a dominant male was identified). Across both treatments (N = 35), males mounted within the first hour in 16 trials ( N 2 male =7, N 6 male =9) There was no difference among treatment groups in latency to mount (Mean 2 male = 32.1 min S.E. = 5.9 min Mean 6 male = 26.9 min S.E. = 5.2 min ; two tailed T test, t 14 = 0.67, P = 0.51) or in total time spent mounted (Mean 2 male = 16.0, S.E. = 4.6, Mean 6 male = 10.3, S.E. = 3.9, two tailed T test, t 1 4 = 0.95, P = 0.36). For the trials where males mounted within the first hour and a behaviorally dominant male could be determined (N 2 male = 4, N 6 male = 8), the behaviorally dominant male mounted first in 100% (2 male) and 75% (6 male) of the cases. In 85% (N = 13) of all trials where males mounted and paternity could be determined, the first male to mount became the father. This included 75% of 6 male trials (N = 8) and 100% of the 2 male trials ( N = 5 ) The overall rate of multiple paternity was low; one male fathered all offspring in 83% of clutches (N = 29). There was no difference in the rate of multiple paternity in 2 male (N= 2) and 6 male (N = 3) groups (Chi Square Test, X 2 = 0.33, P = 0.56). I n 4 trials from the 2 male treatment and 2 trials fro m the 6 male treatment, the father died before
55 the end of the experiment, but even among these trials there was only 1 case of multiple paternity (in a 6 male trial ) Across both treatments, behavioral dominance corresponded to paternity in 80% of all tri als (N = 15). Behaviorally dominant males were fathers 73% of the time (N = 11) in 6 male trials and 100% of the time in 2 male trials (N = 4). However, across all 2 male trials where paternity was determined, there was no difference in the probability tha t the smaller or larger male was the father (N = 15, Chi Square Test, X 2 = 0.529, P = 0.47), while fathers in the 6 male treatment were larger males than predicted under the hypothesis that male size has no effect on paternity ( Mean rank for fathers = 2.6 (S.E. = 0.4); Mean rank for non fathers = 3.7 (S.E. = 0.2); N =14, P<0.006, Fig 3 3). The number of offspring produced per female was not correlated with the size of the father (F 1,17 = 0.70, P = 0.41) it did not differ among trials with one (Mean = 560.7, S.E = 25.2 offspring) versus two fathers (Mean = 606.6 S.E. = 55.2 offspring ) and it was not significantly different between the 2 male (Mean = 552.6, S.E. = 31.9 offspring) and 6 male (Mean = 585.5 S.E. = 33.0 offspring) treatment groups ( F 1,27 = 0.5 2, P = 0.48). The number of offspring produced was positively correlated with female size (R 2 = 0.19, P = 0.007) and female size was negatively correlated with maturation date (R 2 = 0.71, P<0.0001). Discussion Nephila clavipes males gather in groups of compete to mate (Fig 3 1). In contrast to some other density dependent systems where smaller, non competitive males find females and mate faster when at high densities (Knell 2009) larger males mated first in N. clavipes only when group size was large (Fig 3 3). Ther e was no large male advantage when group size was small and male
56 interactions were infrequent. Thus, on many webs (Fig 3 1 ) small males face no competitive disadvantage. This may allow small body sizes to persist in the population. In the mating experiment, the frequency of male male challenges increased with increased male group size, and across all trials, larger males typically won challenges. Male group size did not affect the rate of multiple paternity, which occurred in 17% of all trials Because there was typically only a single successful father per male group, male reproductive success was determined by his ability to reach the female first; in 85% of all trials the first male to mount was the father. In the wild, males compete on juvenile, virgin, and non virgin adult female webs (Christenson et al. 1985) However, in the mating experiment, I staged male male competitions on virgin female webs to enable accurate paternity analysis. In addition, I chose to use virgin females because they are the most likely to be sexually receptive (Christenson et al. 1985; Cohn et al. 1988) Because neither the intensity nor the outcome of male contests changes with female t ype (Cohn et al. 1988) the positive relationship between male density and large m ale advantage found on virgin female webs in this experiment may also apply to contests on juvenile and non virgin adult female webs. Nephila clavipes males arrive and compete to establish dominance on older juvenile webs in order to mate with the female once she has matured (Christenson et al. 1985) but males are also attracted to attempt to mate with adult females (Fi g 3 1, bottom; Christenson et al. 1985) The survey of juvenile female webs (Fig 3 1, top) shows the degree of male male competition over females that are close to their maturation molt. Because only 17.3% of juvenile webs had three or more males,
57 competition is low on most webs, and small and large males may be equally likely to mate with a large percentage of virgin females in the population. This range of male group size is comparable to group sizes reported in other studies in this species (Farr 1977; Vincent & Lailvaux 2006) In contrast, the range of male group sizes and the frequency of groups with three or more males are higher on adult female webs (Fig 1). Group sizes of three or more occur on about 30% of these webs, which suggests that large male advantage affects male reproductive success to a higher degree on adult webs compared to juvenile webs. Although the frequency of highly competitive webs is below 50% for adult females, males are not uniformly attrac ted to these females, and group size is positively correlated with female condition (Vincent & Lailvaux 2006; C. Rittschof unpublished data) Thus females with large groups of males may be higher quality mates compa red to females with small groups. While male body size and density may be important determinants of male mating success, the quality of the female mate ultimately affects total male reproductive potential. Future work should address male dynamics on non vi rgin webs, including the causes for variation in group size, and the relative fitness pay offs for males that mate with females that attract large versus small male groups. Male encounters and contests were infrequent in the 2 male treatment group. This o utcome may have occurred because on the large webs characteristic of N. clavipes males are unable to detect one another when group sizes are small. An alternative explanation is that males behave less aggressively when group size is small. Male aggression can vary as a function of operational sex ratio or the number of
58 at higher rates as the operational sex ratio becomes increasingly male biased (Jirotkul 1999) Similarly in the mosquitofish, decreased female density results in increased male male competition (Smith 2007) Thus a minimum number of rival males could be required to stimulate aggressive interactions between males, and when these interactions occur, large males, who typically win contests, have a reproductive advant age. Two other factors that could affect male reproductive success in a size dependent manner are the order of male arrival to the web and male competitive endurance over the course of his tenure on the web (e.g. Blanckenhorn 2000) B ecause female web sim ultaneously when I initiated the mating experiment. This method does not account for the order of male arrival to the web or male endurance. However, because in many cases in the mating experiment males required more than an hour to reach and mount the fem ale, and because multiple males arrive to a web within hours of each other in the field (Rittschof & Ruggles 2010) my method approximates co mpetitive scenarios that occur on female webs in the wild. Females rarely re mated within the time frame of this experimen t even in 5 trials where the single male mate died before the end of the experiment. Prior work suggests this result could be becaus e females are initially unreceptive to mating after their first copulation (Christenson & Cohn 1988; Cohn et al. 1988) In all other experiments that report immediate second male mating, the experimenters present fo od to the female in order to facilitate mating (Christenson & Cohn 1988) In my experiment, females were not fed during the 48 hrs males were present, which may have prevented second male
59 mates from gaining access to the female. In this 2 day time frame, the factor most critical to male reproductive success is his ability to mate first, which is a function of male body size and male density. It is possible that the 6 male group size used for my high density treatment was not adequately large to overcome large male advantage (Knell 2009) However, certain features of N. clavipes could facilitate a large male competitive advantage even at very high densities. First, f emale N. clavipes spen d most of their time in the center of the web. She moves only to collect prey that is caught in the web (Christenson & Goist 1979; Cohn et al. 1988) M ales that lose male male encounters move to the periphery of the web while contest winners move closer to the center and defend only the small area around the female. Second, males threaten one another at a distance by shaking the web. These long range contests allow males to defeat multiple intruders in succession wit hout physical contact. Third, the spider web is a two dimensional surface, and unlike animals that fly or burrow underground, male intruders can only approach the female on the plane of the web. The small defendable central area, long range threatening beh aviors, and two dimensional structure of the web could allow males to monopolize females even at high densities. Body size in N. clavipes has a genetic component and a degree of plasticity in response to environmental conditions and food availability (Vollrath 1983; Higgins 1992) However, stabilizing selection on a single body size optimum would select against plasticity that could result in a body size that deviates from the optimum. Conversely, if changes in male density facilitate variation in reproductive success for male s of different sizes, body size variation could be maintained by spatial variation in
60 the strength of sexual selection. Local variation in the strength or direction of sexual selection can be co nsidered a heterogeneous mosaic of selection intensities. In the case of the damselfly Ishnura elegans spatial variation in the density of females and the frequency of female morphotypes changes the strength and direction of selection on male body size (Gosden & Svensson 2008) Spider webs are discrete habitat patches (Agnarsson 2003) with different male densities. As a result, there is spatial variation in the strength of sexual selection on male body size. My results are similar to f indings in Nephila plumipes where seasonal changes in population density and operational sex ratio result in temporal variation in selection gradients on male size (Kas umovic et al. 2008) V ariation in male body size in web building spiders has been explained in several different ways. For example, Elgar and Fahey (1996) emphasized an advantage to small size due to reduced sexual cannibalism. Moya Larano et al. (2009) showed an intermediate optimum attributable to climbing speed. However, these studies suggest that stabilizing selection is acting to optimize male body size, which predicts decreased, not increased size vari ation (Falconer & Mackay 1996) In N. edulis Schneider et al. (2000) showed that females prefer small males during the physical act of copulation, and small males typically copulate l onger and father more offspring than large males. The authors speculate that some other variable is required to maintain large male sizes in this species, for example high or variable male densities. An additional alternative hypothesis for the maintenanc e of small male sizes and size variation in N. clavipes is that large male competitive advantage is balanced by a small male advantage in mate search efficiency, particularly if female density is low
61 (Blanckenhorn 2000; Fairbairn 2007) This hypothesis predicts that small males arrive to a web and mate befo re large males are able to reach the web. However, if small males arrive to a web first, this does not exclude the possibility that these males are avoiding competitive environments and choosing webs with few competitors. Furthermore, many studies in spide rs have tested for an association between male body size and mate search efficiency and have failed to find a relationship (e.g. Prenter et al. 1998; Foellmer & Fairbairn 2005b; Kasumovic et al. 2007; De Mas et al. 2 009) In N. clavipes there is evidence that rather than small males, intermediate sized males have the highest survival rate during mate search (Vollrath 1980) Thus stabilizing selection on male size would result from the trade off between male competition that favors larger size and mate search efficiency that favors intermediate size. Such selection should limit variation in male size (Falconer & Mackay 1996) Because females occur at such high densities, mate search efficiency may not limit male mating opportunity in N. clavipes (Far r 1977; Hodge & Uetz 1992) Furthermore, adult webs in close proximity contain variable male group sizes. Males appear to assess some aspect of female webs through a period of residence time on the web. They arrive at a web and stay for variable periods o f time depending on their success at competing with other males to mate (Christenson & Goist 1979) The fact t hat males are able to assess webs after they arrive and in some cases leave suggests movement between different webs may not be a strong constraint on male reproductive behavior. on about the level of competition on the web and allow males to decide to stay or leave
62 depending on their own relative body size. Small males might avoid highly competitive webs while large males might target those webs because they have a competitive adv antage in large groups. Such divergent mating tactics often contribute to variation in sexually selected traits such as body size (Brockmann & Taborsky 2008) Thus density dependent selection on body size could maintain male size variation and facilitate the evolution of size dependent mating tactics. In addition, if the level of competition on a web is co rrelated with female value, these tactics carry different risks and pay offs. Future studies could explore the contexts for male alternative mating tactics, specifically the relationship between female value, male competition intensity, and male reproducti ve success with respect to body size
63 Table 3 1. Sample sizes across data types Number of Trials 2 male 6 male Total B ehavioral 17 18 35 Paternity 15 14 29 Both 13 13 26
64 Figure 3 1. Frequency distributions for male group sizes from field survey s of juvenile (top panel) and adult (bottom panel) female webs.
65 Figure 3 2. Distribution of body size (cephalothorax width, mm) for male N. clavipes in the mating experiment (N=139).
66 Figure 3 3. Size ranks for fathers versus males that were not fath ers in the 6 male treatment group. Lower ranked males are larger.
67 CHAPTER 4 MORTALITY RISK AFFECTS MATING DECISIONS IN THE SPIDER NEPHILA CLAVIPES Background For animals, decision making involves risk assessment (S eger & Brockmann 1987; Austad 1989; Creel 1993; McNamara et al. 2006; Beaumont et al. 2009; Wiklund & Friberg 2009; Beauchamp 2010; Nevoux et al. 2010) For example, many studies have addressed how males and females navigate the risks associated with sear ching for, competing for, and winning mates, including predation risk (e.g. Hedrick & Dill 1993) and sperm competition risk (Parker et al. 1997) For males, one element of risk that is rarely considered (Dunn et al. 2001) is that a female mate may not survive to offspring independence, resulting in total reproductive failure. This risk fac tor would be particularly important (a) when males, due to limited resources, mate with only one or a few females in their lifetime; and (b) when there is high variance in breeding success among reproductive females because of high and variable adult morta lity. These traits are common in species with short life spans and few reproductive opportunities because there is strong selection on individuals of low quality to reproduce despite their low chances of success (Parker 1983; Roff 1992) In contrast, in species with longer reproductive life spans, low quality individuals typically forego a reproductive cycle if they can improve their quality as non breeding individuals, and thus improve their lifetime fitness by passing up a risky reproductiv e opportunity (e.g. Angerbjorn et al. 1991; Vanheezik et al. 1994) If female mortality risk is a significant source of variation in female reproductive out put, males who choose low risk females will have a fitne ss advantage. Although sperm competition and sperm precedence patterns are common sources of variation in
68 male reproductive success (Parker et al. 1997) female fecundity differences are often the greatest source of variation, and as a result, choosy males pay attention to indicators of fecundity (Bonduriansky 2001) fecundity, a female must survive to offspring independence for males to gain any fitness at all. Therefore, i f males can as sess female mortality risk, they might benefit by choosing low risk females even at a trade off to other features that affect fecundity or number of potential offspring sired, e.g. female body size or sperm competition risk. Few studies have considered whether males use female survivorship as a mate choice criterion, even though there is evidence that males can choose females using traits that predict survivorship. For example, in the seaweed fly, males choose females on the basis of vigor and vigorous f emales are more likely to survive the long period between mating and oviposition in this species (Dunn et al. 2001) In a number of taxa including fish and insects, there is evidence of male choice for gravid females (Bonduriansky & Brooks 1998; Katvala & Kaitala 2001; Benson 2007) although m ost of these studies argue that male preference for gravid females is a consequence o f last male sperm precedence patterns (Bonduriansky & Brooks 1998; Katvala & Kaitala 2001; Benson 2007) Gravid f emales, who are close to oviposition, are most likely to survive to offspring independence, and thus are a lower risk mate choice for males than females earlier in the egg development process. Particularly in cases where females store sperm over their adult lifetime survival probability and proximity to oviposition may be stronger predictors of offspring number (and male success) than sperm precedence patterns. In this study I evaluate the role of mortality risk in male mate choice decisions
69 using a study system that allows me to separate risk from other likely mate choice criteria, including female body size and sperm precedence patterns. In the golden orb web spider, Nephila clavipes males are attracted to and attempt to mate with both virgin and non virgin females even though there is first male sperm precedence (Christenson & Cohn 1988) Furthermore, prior studies have suggested that male attraction to non virgin females i ncreases as females approach oviposition, but that attraction is not related to female size, which predicts clutch size (e.g. Vincent & Lailvaux 2006; Rittschof 2010) Females that are close to oviposition may have a better chance of surviving to produce a clutch of offspring compared to females who are earlier in the process of egg development. If males prefer to mate with females close to oviposition regardless of female body size and reproductive costs due to sper m competition, then this result would support the hypothesis that males choose female mates on the basis of survival probability. Study System Female Nephila clavipes mature sometime between July and late September and live approximately 3 4 months as a dults (Christenson & Cohn 1988; Rittschof 2010) They require 30 40 days between matur ation and first oviposition and lay at most 5 clutches of eggs in a laboratory setting (C hristenson et al. 1985; Higgins 2000) before dying at the onset of winter No females over winter in temperate populations of N. clavipes (Hig gins 2000) Mature males spend adulthood (estimated to be 3 weeks) searching for mates (Brown 1985) and can visit multiple female webs in their lifetime Males are sperm limited in this species. In spiders, males induct and store sperm in paired prosomal appendages called pedipalps (Foelix 1996) Nephila clavipes males induct sperm into
70 their pedipalps only once (Michalik and Rittschof, in press). Males adjust the amount of sperm they transfer during a copulation event, and so they may be able to mate multiply (Christenson & Cohn 1988) but once males deplete the sperm in their pedipalps, they cannot replenish it. Males are attracted to and attempt to mate with virgin females, and males use all of their lifetime sperm stores when mating with virgins (Christenson & Cohn 1988) However, males also mate with non virgin females, and when they do so, they use only a portion of their sperm stores (Christenson & Cohn 1988) There is some evidence in this species that males are attracted to females who are close to oviposition. Despite first male sperm precedence patterns (Christenson & Cohn 1988) the number of males found on non with increased female abdomen size (Vincent & Lailvaux 2006; Rittschof 2010) Abdomen size is cons idered a measure of gravidity or time to oviposition in spiders (Ortlepp & Gosline 2008) In contrast, female cephalothorax width, which determines clutch size in this species (Higgins 2000; R ittschof 2010) does not correlate with an increase in male attraction (Vincent & Lailvaux 2006) In this study, I test the hypothesis that males are attracted to adult females who are close to oviposition, irrespe ctive of female body size. Even though the largest females have the highest potential fecundity, males that choose females close to oviposition may maximize offspring number because these females are more likely to survive to lay a clutch of eggs. I n a m ale removal field experiment, I verify that males are consistently attracted to some females over others i.e. that males make mate choice decisions In a second experiment I determine how male mating behavior, sperm usage, and subsequent reproductive succ ess change as a function of female time to oviposition. Finally, with
71 these same females, I determine that female abdomen size is strongly correlated with time to oviposition, and that the closer a female is to oviposition, the less likely she is to die be fore laying her first clutch of eggs. Methods Male Removal Field Experiment In this experiment I evaluated whether the occurrence of large groups of males on certain female webs could indicate male mate choice. Because female N. clavipes have long web t enure (Rittschof & Ruggles 2010) and males usually do not follow females when they change web sites, even over short distances (Cohn et al. 1988) the he length of time a female has been at a web site. To demonstrate that males in the field are consistently attracted to certain females and not others, I performed a male removal experiment. On August 23 2008, 71 mature female webs were identified and fem ales were marked with enamel paint (Testors ) and measured following Rittschof and Ruggles (2010) On this day (day 0), the n umber of males on each web was recorded (hereafter the number of initial males). Males were then removed and released at least 5 m from the nearest treatment web. Each subsequent day for 14 days (or until the female abandoned her web site), I returned to e ach web, counted the number of males present, and removed males as I had on day 0. Due to time constraints, I did not mark males, and so some males may have returned to their original web or appeared on more than one web. While the short release distance c ould bias the removed males to return to their original web by chance, female density is very high during this peak time of the breeding season (Christenson et al. 1985) and because females are often fou nd within 1 m of one another (Moore 1977) males choosing webs at random are unlikely to
72 hus any patterns of male accumulation on webs would suggest a male web preference. For each female web, I totaled the number of males observed across all days of the experiment and divided this value by the number of days the female was observed on her web This calculation is the average number of males present on the web per day. Mating Experiment Animal collection and housing The mating experiment was conducted from 6 July 2009 to 21 August 2009. Females and males were collected about twice a week throu ghout this time period from three locations, the Ordway Swisher Biological Station in Melrose, Florida (Putnam (Alachua County, latitude longitude 82 ), and San Felasco Hammock Preserve State Park (Alachua County, latitude longitude ) The mating experiment was conducted at the University of Florida Bee Biology Unit (Alachua County, latitude ude which is a fenced and protected research area. In order to ensure virginity, females were collected during their sub adult instar. Female maturity was determined by inspecting the epigynum, which is black with two distinct openings i n mature females, and smooth and maroon colored in sub adult females (Higgins 2000) Females were housed outside in wooden cages (90 cm x 90 c m x 15 cm) and fed two large mealworms and sprayed with a mist of water daily. Females built and maintained normal prey capture webs in these frame cages (after Rittschof 2010) The cages were arranged along a chain linked fence, and shielded from the sun and rain with a tarp awning. I collected mature males for the mating
73 experiment from juvenile female webs. After their terminal maturation molt, males have dark body coloration and sclerotized pedipalps (Myers & Christenson 1988) Because males were with the current female web owner, but I could not ensure virginity because a male could have mated with a different female prior to collection. However, because I could not identify which males may have mated previously, male experience was evenly distributed across treatments. Treatments The goal of the mating experiment was to determine whether males that mate with non virgin females close to oviposition copulate more frequently, t ransfer more sperm, and father more offspring than males that mate with younger non virgin females. To do this, I designed an experiment with two treatments that differed in the timing of the second male (hereafter P2) mating. For both treatments, females were allowed to mate with a first male mate (hereafter P1). Then, for females in treatment 1 (early treatment), a second male was introduced and allowed to mate with the female immediately following the removal of the first male. For treatment 2 (late trea tment), I waited until the female was close to laying her eggs (see below) before introducing the P2 in to the female cage. I initiated each mating experiment trial on the day that the female molted to maturity (day 1), which was usually within one week o f female collection. Females were checked daily from 0700 hrs to 0900 hrs. If a female was mature, I introduced the P1 three full days and was removed on the morning of day 4. Because females typically do
74 not eat immediately after their maturation molt (Christenson & Cohn 1988) I did not feed females on day 1. However, I resumed the normal feeding on day 2. I assigned each female to a treatment once she matured, alternating between treatment 1 (early) and treatment 2 (late). In the early treatment I introduced P2 to the after P1 was removed). P2 was left for three full days to mate with the female, and females were fed normally throughout this time period. In the late treatment, after P1 w as removed, females were left alone until they reached a body condition index of 1.5 (see below). The morning of the day the female reached this body condition index value, P2 was introduced and allowed three full days to mate with the female. Because male body size affects sperm number and could influence male sperm competition (Christenson & Cohn 1988; Cohn 1990) for each replicate, P1 and P2 males were matched for body size. Male body size was determined by measu ring the tibia patellar length of the 4 th leg after it was removed for DNA extraction (see below). I fed the females normally during the time that the P2 males were with the female. For both treatments, after P2 was removed, females were kept and fed as th ey were prior to the start of the experiment and allowed to lay 2 clutches of eggs as in Rittschof (2010) Female body condition index was defined as the abdomen height (the dorsal ventral height of the abdomen measured just posterior to the epigynal slit; Vincent & Lailvaux 2006; Rittschof & Ruggles 2010) divided by the cephalothorax width (measured at its widest point). Cephalothorax width was measured using dial calipers. I measured abdomen height from digital photographs taken daily usin g a PowerShot A400 Digital Camera (Canon, Lake Success, New York). Photographs enabled me to
75 obtain accurate measurements of abdomen height without removing the females from their cages. I held a ruler in the plane of the abdomen directly above the female to provide a scale for each photograph. I viewed each picture using Windows Picture and Fax Viewer (Microsoft, Redmond, WA) and measured the abdomen height and the scale by holding a ruler flush with the computer screen. From these measurements I calculat ed female abdomen height. Each day a male was present, I checked the cage to observe whether or not the male was copulating with the female. Although males copulate throughout the day, they are most likely to copulate with the female while she is eating (Christenson et al. 1985) and so I checked for copulation 30 min after food was introduced to the female web in order to estimate copulation effort. When a male is copulating with a female, he mounts the ventral side of her abdomen and inserts the distal part of his pedipalp (the conductor) into her epigynal opening (Christenson et al. 1985) I checked for copulation after feeding each day that the male w as present in the cage (3 days). I report whether or not the male was observed copulating on any days of the mating trial. Females laid egg sacs on the walls and screening of their cages. Initially, clutches were left to hatch where they were laid within the cages. However, due to high rates of parasitism by mantisflies (Neuroptera: M antispid ae) all un hatched clutches were removed on 19 September 2009 and placed in 946 mL screened boxes in a screened shed. Any clutch laid after this date was also moved i mmediately and treated in the same way. The shed kept the eggs at ambient temperatures and humidity, but provided protection from parasites. Eleven females were selected at random and their clutch sizes were determined by counting the total number of offsp ring and eggs laid.
76 Male sperm counts In N. clavipes and Rittschof, in press). It is not possible to determine the number of sperm in the male pedipalps without sacrificing the male. How ever, a count of the number of sperm remaining in the pedipalps after copulation provides a basis to estimate the number of sperm used during copulation (Christenson & Cohn 1988) which is useful for comparing sperm usage across treatment groups. Once males were removed from female cages following the mating experiment, male pedip alps were assessed for sperm number. Two assistants who were blind to male mate order and treatment group scored male sperm counts. Males were placed in a 20 C freezer for five minutes in order to anesthetize them, and then both pedipalps were removed by pinching the femur with forceps. The pedipalps were placed in a 1.5 mL microcentrifuge tube with 50l of tap water and pulverized with a tissue grinder for 3 min. The samples were then vortexed for one minute, and centrifuged for one minute. Because centr ifugation pulls the contents of the tube into a pellet at the bottom of the tube, the pedipalp solution was homogenized before counting by drawing the solution in and out of the pipette tip until the pellet was no longer visible. A 10 L sample of the sper m solution was dispensed onto a hemacytometer (Fisher Scientific) and the sperm were counted using a light microscope ( Leica, Wetzlar, Germany) Two 10 L samples were counted for each male, and total sperm density per mL was calculated. Parental genotypes I assessed paternity using microsatellite marke rs developed for this species (Rittschof 2010) which s how low variability but reliable allelic peaks. In order to obtain full paternity exclusion and maximize the level of precision for paternity analysis, I pre
77 screened male and female genotypes in order to match the P1 and P2 males and the female for full exclusion at one locus. To do this, I removed one leg per individual for DNA extraction and genotype analysis prior to the mating experiment. I chose one 4 th leg from each male and one 1 st leg from each female. To remove a leg, I pinched the femur between a pair of forceps until the individual autotomized it s leg which I transferred immediately to cell lysis buffer (Gentra Puregene Tissue Kit, Qiagen, Valencia, CA). No males or females died as a re sult of the leg removal treatment. I extracted parental DNA and amplified and scored three polymorphic microsatellite loci for each individual (Rittschof 2010) For each mating trial, I chose males whose genotypes allowed for full paternity exclusion at one locus. Offspring analysis For most of t he clutches included in paternity analyses, offspring were preserved in 95% ethanol once they had molted to their second instar. However, due to parasitism and poor egg sac construction, 5 clutches began to desiccate before all offspring had hatched or mol ted to their second instar. In these clutches, offspring were preserved in developmental stages including embryos and first and second instar juveniles. Twenty offspring from each clutch were selected for paternity analysis. Offspring were chosen at rand was extracted following Rittschof (2010) In the 5 cases where offspring were preserved at different developmental stages, both embryos and hatched offspring were included in the paternity analysis. To improve DNA yiel d, embryo DNA was extracted using the QIAamp DNA Microkit (Qiagen, Valencia, CA). Within this small sample group however, first and second male paternity share did not differ among the different offspring developmental stages.
78 Condition Index and Egg Deve lopment in Wild Caught Females Because female abdomen size is a consequence of both egg development and the amount of food a female has consumed, I collected wild caught females across a range of abdomen sizes in order to test for a relationship between a bdomen size and ovary mass (a measure of the degree or state of egg development). A positive relationship would suggest that females with larger abdomens are closer to oviposition time. Sixteen females were collected between 27 July 2010 and 12 September 2 010. Females were anesthetized and the ovary was removed within 1 day of collection. To anesthetize the females, each female was placed for 7 minutes in a 20 C freezer. After this period, female abdomen height was photographed with a measurement scale, a nd abdomen height was determined by measuring these photographs (see above). I measured cephalothorax width using dial calipers. The ovary was dissected out of the abdomen in 10mM PBS and preserved in 95% ethanol. Prior to assessing mass, ovaries were remo ved from ethanol and dried in plastic weigh boats in a drying oven at 37 C. Results Male Removal Experiment There was a weak positive correlation between the number of initial males found 2 = 0.0 8, P<0.01). There was no correlation between cephalothorax width and the number of initial males found on the web (N = 71; R 2 = 0.003, P = 0.65). When removed, the rate at which males returned to female webs was positively correlated with the number of ini tial males present on day 0 (Fig 4 1; N = 71, R 2 = 0.48, P<0.0001).
79 Mating Experiment Female body condition index was negatively correlated with the number of days remaining before oviposition (Fig 4 2; R 2 = 0.73; P<0.0001). Females in the gravid treatment (N = 19) reached a body condition index of 1.5 (the day on which the P2 male was introduced) an average of 24.4 days after their maturation molt (S.E. = 0.75 days), and 7.7 days before first oviposition (S.E. = 0.64 days). For wild caught females, abdomen s ize ranged from 4.4 to 11.9 mm (Mean = 9.2 mm; S.E. = 0.55 mm), and condition index ranged from 0.68 to 1.95 (Mean = 1.47; S.E. = 0.09). Ovary mass ranged from 0.0006 g to 0.2987 g (Mean = 0.095 g; S.E. = 0.03 g). Because of this broad range of masses rela tive to the range of abdomen sizes and condition indices, ovary mass was log transformed for analysis. There was a strong positive relationship between ovary mass and abdomen size (R 2 = 0.80, P<0.0001) as well as ovary mass and condition index (R 2 = 0.76, P<0.0001). Twenty two females in the early treatment successfully mated with both the P1 and P2 males, and in the late treatment 19 females mated with two males. The percentage of trials in which the P1 male was observed mating did not differ between the two treatment groups (Fig 4 3; Early = 82%, Late = 74%; X 2 = 0.39, P = 0.53). However, a significantly higher frequency of P2 males mated in the late treatment group compared to the early treatment group (Fig 4 3; Early = 34%, Late = 84%; X 2 = 9.6, P<0.002). A cross both treatment groups where sperm counts were obtained for both the P1 and P2 males (N Early = 20, N Late = 18), 100% of P1 males were completely sperm depleted after copulation regardless of treatment. However, P2 males from the early treatment group had significantly more sperm remaining compared to P2 males from the late treatment group (Two tailed T test, t 36 = 3.4; P<0.002).
80 Twenty three females across both treatment groups laid viable first clutches of eggs (N Early = 12, N Late = 11). In the firs t clutch, a higher percentage of P2 males fathered offspring in the late treatment compared to the early treatment (Fig 4 4; Early = 17%; Late = 73%; X 2 = 7.3, P<0.007). When P2 males fathered offspring (N Early = 2, N Late = 8), paternity ranged from 44 100% (Mean = 79%; S.E. = 7%). There were 19 females that laid two viable clutches. For 10 of these females (all in the early treatment), clutch 1 had a single father, and this father was the P1 male. In these 10 cases, clutch 2 also had a single father, which w as the P1 male. For the remaining 9 females that laid 2 viable clutches (1 female from the early treatment, 8 females from the late treatment), there was mixed paternity in clutch 1. Among these females, P2 paternity ranged from 50 100% (Mean = 78%; S.E. = 6%) in clutch 1. However, P2 paternity in clutch 2 was significantly different than clutch 1 (Fig 4 5; Two tailed paired T test, t = 4.3, p<0.003). In all but one case, P2 paternity share was lower in clutch 2 than in clutch 1, and the mean difference was 39% (S.E. = 9%). Of 47 females that matured during the course of the experiment, 11 females died before oviposition (23% mortality). Six females that died mated only once (i.e. they did not make it to the body condition index required before the P2 male c ould be introduced), 4 females that died successfully mated with both the P1 and P2 males in the early treatment, and 1 female that died successfully mated with both the P1 and P2 roductive failure for mating with a virgin female is approximately 23% (11 of 47 mature females) ; the risk for mating with a non virgin female early in adulthood is about 18% (4 of 22 early treatment females) ; and the risk for mating with a gravid female i s about 5% (1 of
81 19 late each female progresses from virgin to mated to gravid in her lifetime, these categories are not mutually exclusive in terms of the calculations of risk. Fo r example, the gravid female that died was also tallied for the total calculation of risk for mating with a virgin female (this female just happened to die at the gravid stage). For this reason, it is not possible to compare mortality risk statistically ac ross female categories. F emale egg number ranged from 357 to 908 (averaged across clutch 1 and 2 for each female, N = 11, Mean = 673, S.E. = 28) and was positively correlated with female body size (cephalothorax width; R 2 =0.49, P<0.025). Clutch 2 trended towards having fewer offspring compared to clutch 1 (Mean clutch one = 711, S.E. = 26.5; Mean clutch two = 634, S.E. = 48; Two tailed paired T test, t = 1.86; P = 0.09). The mean difference was 76 offspring (S.E. = 41 offspring), which corresponds to a mean difference of 11% (S.E. = 6%). Discussion The age and reproductive state of non virgin females significantly affects male mating behavior. Field data show that males are consistently attracted to some mature females over others (Fig 4 1) and that high lev els of attraction are correlated with female abdomen size, which is an indicator of time until oviposition (Fig 4 2). In contrast, males are not differentially attracted to females with large body sizes (cephalothorax width), even though body size predicts a verage clutch size (see results). In the mating experiment, when paired with non virgin females close to oviposition, males mated at higher frequencies (Fig 4 3), transferred more sperm, and fathered more offspring in the first clutch (Fig 4 4) compared to mal es paired with younger non virgin females. These findings support the hypothesis that males invest more
82 reproductive effort in gravid females who are close to oviposition compared to younger non virgin females. Furthermore, at least within the controlled conditions of this mating experiment, mating with a female close to oviposition (i.e. a gravid female) decreased a mortality rate of 5%, a value much lower than the rat e for both virgin females (23%) and young non virgin females (18%). More work is needed to evaluate differences in survivorship among females in the wild, although mortality in general is probably even greater because food availability is variable and fema les are subject to predation. M ale preference for females that are close to oviposition suggests that female survival probability could be a more important predictor of offspring number than other fecundity parameters like body size in this species. Partn er survivorship could be an important mate choice criterion in species where there is delayed reproductive pay off after copulation, i.e. species where females store sperm. However, almost no studies have suggested a role for partner mortality in mate choi ce. In long lived species, poor condition individuals can forego reproduction for a season or breeding cycle in order to optimize their lifetime reproductive success (Roff 1992) The result is that yearly adult survival rates are high because individuals at risk do not enter the mating pool (e.g. Vanheezik et al. 1994) However, in short lived species with few reproductive opportunities, there i s high variation in reproductive output (Roff 1992) and there is a high cost associated with reproductive failure for both sexes because in some cases individu als have as little as one opportunity to reproduce. In cases where male reproductive investment ends at sperm deposition, female mortality risk directly affects male reproductive out put. In these species (e.g.
83 Nephila clavipes ), males could incorporate female mortality risk into their choice criteria by choosing females on the basis of probable survivorship (Dunn et al. 2001) Similar to the current s tudy, other studies in insects and fish demonstrate male choice for gravid females who are close to egg deposition (Bonduriansky & Brooks 1998; Katvala & Kaitala 2001; Benson 2007) In these cases, as opposed to inc reased probability of female survival, the benefit of mating with gravid females is attributed to last male sperm precedence patterns. Sperm precedence studies assess whether ejaculates stored by a female fertilize eggs disproportionately due to mate order As is evidenced by the current study, sperm precedence studies should be interpreted with caution because in most cases, it is difficult to assess differences in the number of sperm transferred by each male. Because N. clavipes males have a limited numbe r of sperm, and all sperm are contained in the pedipalps (Christenson 1989) the sperm remaining in the pedipalps after copulation provides an estimate of sperm usage by mal es (Christenson & Cohn 1988) P2 males in the late mating treatment used mor e sperm and were more likely to father offspring (Fig 4 4) than P2 males in the early mating treatment, a finding that emphasizes the importance of sperm numbers in paternity outcomes. Previous findings in N. clavipes have suggested that there is first mal e sperm precedence, presumably a consequence of the conduit shape of the female reproductive tract (Christenson & Cohn 1988) However, my findings suggest that paternity patterns may reflect a complex relationship between the number of sperm transferred and mate order. Regardless of treatment, P1 males always transferred more sper m than P2 males, and across multiple clutches (e.g. clutch number 2), the male that
84 transferred more sperm (P1 male) showed an increase in paternity (Fig 4 5). This f inding is similar to the predictions of a fair raffle (e.g. Ball & Parker 2000) However, within the first clutch, sperm number and mate order affect paternity share, and the P2 male has an advantage over the first male even though he transfers fewer sperm. This disparity could be due to post copulatory processes within the reproductive tract of the f emale, although the mechanisms are unknown. Other studies in spiders have suggested that sperm precedence patterns are not simply a function of female genital morphology (reviewed in Huber 2005) In Nephila species specifically, multiple factors affect the n umber of sperm transferred by males, as well as the sperm stored and used for fertilization by females. Second male paternity share is affected by copulation duration and frequency in N. edulis and N. plumipes (Schne ider et al. 2000; Elgar et al. 2003; Schneider & Elgar 2005) In addition, in N. plumipes the number of sperm transferred and stored by the female does not show a simple positive correlation with paternity share, which suggests a role for sperm manipulat ion by the female prior to fertilization (Schneider & Elgar 2001) In spiders generally, it is common for post copulatory processes, influenced by both males and females, to affect sperm storage and usage (Parker 1990; Bukowski & Christenson 1997a; Eberhard 2004; Huber 2005; Aisenberg 2009; Schneider & Lesmono 2009; Welke & Schneider 2009; Burger 2010; Peretti & Eberhard 2010) In some cases, these processes are associated with copulation interval and th (Schneider & Elgar 2001; but see Jones & Elgar 2008) The current study shows that in N. clavipes both the amount of sperm transferred to the female as well as mate order play an important r ole in determining the number of offspring sired by a male.
85 M ating with multiple non virgin females or a single virgin female could be mutually exclusive mating strategies in this species. Male sperm use and mating effort indicate that males prefer to ma te with either virgin females or gravid females close to oviposition (Fig 4 3). Males may avoid mating with younger non virgin females because these females have only slightly lower mortality risk compared to virgin females, and are costly to males in terms of sperm competition. Although mating with a virgin female is a high risk strategy in terms of female mortality, this risk is off set by other potential advantages. First, i f the female does not re mate, the male could father 100% of her offspring. In a ce nsus survey in my population, females were visited 0 17 males over the course of their lifetime, which suggests female mating rate is variable. Second, because first male paternity share increases in the second clutch, the total reproductive pay off for ma ting with a virgin female, even if she re mates, may be higher than mating with a non virgin female as long as the female lays more than one clutch (particularly if the clutches are of similar size; Higgins 2000) Female clutch number in the wild is highly variable, and can depend on factors including the time of the season at w hich the female matures (Higgins 1992, 2000) However, in my studies, most females easily lay 2 clutches in captivity, although three clutches is rare. Thus depending on the time of the season, local density, and op erational sex ratio, the optimal male strategy may shift. In addition to sperm competition risks and female mortality risks, male N. clavipes face other pre copulatory challenges that could affect whether they successfully mate with a virgin or non virgin female. In other spiders, male mortality risk during mate search appears to constrain male mating rates and presumably male choosiness (Andrade 2003; Andrade & Kasumovic 2005; Fromhage et al. 2007;
86 Kasumovic et al. 2007) Mate search mortality could be a function of female density and mating probability (Fromhage et al. 2008) In addition, f emale webs differ in the number of male competitors present, (Elgar & Fahey 1996; Schneider & Elgar 2005; Rittschof 2010; Rittschof & Ruggles 2010) Variation in mating succe ss may be highest on gravid female webs where male group sizes are largest (Vincent & Lailvaux 2006; Rittschof 2010) The probability of successfully mating, the ability to re mate, and variation in female survival and clutch number could result in an evolutionar il y stable state maintaining males in the population who adopt mating strategies that result in copulations with virgin and gravid non virgin females (ESS; Brockmann & Taborsky 2008) Other work in Nephila suggests that these strategies could depend on male body size (Schneider et al. 2000; Schneider & Elgar 2001, 2005; Elgar & Jones 2008; Jones & Elgar 2008; Rittschof 2010) Because there is conflict between the sexes over mating rate (Chapman et al. 2003) one alternative explanation for the results of this study is that females are in control of mating interactions, and that the increased frequency of copulation with gravid females occurred because females show increased sexual receptivity just before ovi position. Prior work in this species has suggested that, because non virgin females are less receptive compared to virgins, male mating opportunity with non virgin females is limited to times when the female is eating (Christenson et al. 1985; Christenson & Cohn 1988) Males presumably wait until the female is eating in order to avoid cannibalism (e.g. Fromhage & Schneider 2005) which suggests sexual conflict. Because cannibalism rates are uniformly low in N. clavipes (Christenson et al. 1985) it
87 is difficult to infer how cannibalism impacts male copulation behavior. Regardles s however, the current study shows that, even when food is present, and as a result males have access to all non virgin mates, males copulate with non virgin females at different rates depending on female age, which demonstrates some role for male control during mating events.
88 Figure 4 1. Average males per day versus males present at t 0 The average number of number of males present prior to the beginning of the removal experiment (R 2 =0.48, P<0.0001). Each data point represents one female web (N=71) on which males were removed daily.
89 Figure 4 2. Body condition index versus days until oviposition. Female body condition index increases as the female approaches oviposition (R 2 =0.73 P<0.0001). Each data point represents a single female on a given day of her reproductive cycle (N=35 females; N=938 total female days).
90 Figure 4 3. The proportion of P1 and P2 males observed mating for the non gravid (early) and gravid (late) treatmen t groups. Mating frequency was similar for P1 males across treatments (P=0.53), but P2 males mated with significantly higher frequency in the gravid (late) treatment group (P<0.002).
91 Figure 4 4. P2 paternity share versus treatment. P2 males fathered a g reater proportion of offspring in the first clutch in the late treatment versus early treatment group.
92 Figure 4 5. P2 paternity share versus clutch number. Pooled across both treatments, there were 9 cases where females laid 2 viable clutches and the P2 male fathered offspring in clutch 1 (N Early =1, N Late =8). Of these cases, P2 male paternity share was significantly lower in the second clutch compared to the first clutch (P<0.003).
93 CHAPTER 5 MALE MULTIPLE MATING AND ALTERNATIVE TACTICS IN THE GOLDEN ORB WEB SPIDER NEPHILA CLAVIPES Background Recent empirical work shows that males and females often mate with multiple partners during a single reproductive episode (Arnqvist & Nilsson 2000; Jennions & Petrie 2000; Zeh & Zeh 2003) T he discovery of high rates of multiple mating in female s, typically considered the choosy sex shows that rather than minimize their mate number, females optimize their mate number and mate choice depend ing on factors such as their condition their genotype, their genetic compatibility, and environmental conditions (Zeh 1997; Fox & Rauter 2003; Klemme & Ylonen 2010) In contrast, for males, it is wide ly accept ed that fitness should increase with increased mate number (Bateman 1948) As a result, s tudies of male mating strategies have focused on factors that constrain male mate number, for example the number of available mates (Emlen & Oring 1977; Fromhage et al. 2008) dominance status (Emlen & Oring 1977) costs of obtaining a mate (Segoli et al. 2006; Oliver & Cordero 2009) and the costs of sp erm production (Wedell et al. 2002; Smith 2009) However, empirical evidence shows that males do not always mate at the highest possible rate, which suggests that males, like females, follow decision rules to optimi ze their mate number (Wedell et al. 2002) Male reproductive effort with a single mate can come at the expense of re mating (Bonduriansky 2001; Matessi et al. 2009) and males use environmental cues as well as their own condition and competitive ability to optimi ze this trade off For example, when sperm competition risk is high, a male can increase his sperm allocation to one female, but potentially limit his future mating ability (Ball & Parker 2007; Teng & Zhang 2009)
94 S imilarly, the switch point between post copulatory mate guarding, which prevents female re mating, and searching for new mates can depend on male density, which affects the likelihood that a female will re mate once she is abandoned (Schubert et al. 2009) mate with an ad his chances of winning an additional mate (Burley 1988). In all of these cases, it is beneficial for a male to adjust his reproductive investment and optimize his mate number in order to maximize total reproductive success. In extreme cases, male mate number optimization could result in the evolution of male monogamy (hereafter monogyny; Fromhage et al 2008). Web building spiders, particularly those in the genera Argiope Nephila and L atrodectus have been the focus of empirical and theoretical studies addressing the evolution of monogyny. This is because members of these groups exhibit conspicuous behaviors like sexual cannibalism and genital mutilation during copulation, two traits th at prevent males from re mating (Andrade & Kasumovic 2005; Schneider et al. 2005a; Fromhage & Schneider 2006; Segoli et al. 2006; Uhl et al. 2010a; Michalik and Rittschof, unpublished manuscript) In contrast, the g olden orb web spider Nephila clavipes provides an opportunity to address male mate number optimization because, although males use all of their life time sperm in a single mating in certain contexts (Christenson 1989; Michalik and Rittschof, in preparation) they are not always constrained to monogyny because they are rarely cannibalized (Christenson et al. 1985) and they do not consistently mutilate their genitalia during copulation. Thus males are able to mate multiply if they retain some sperm after insemination, and I
95 predict that males follow decision rules to o ptimize their mate number depending on their own condition as well as environmental factors. The constraints to reproduction faced by male Nephila clavipes are also found in a variety of other animal taxa. For this reason, N. clavipes can serve as a model system to address the relative importance of factors like sperm depletion, operational sex ratio, mate search mortality risk, mate guarding, female value, and male competitive ability (e.g. body size), in determining male mating rate and mate number optimi zation. In this study, I parameterize a dynamic state variable model for the spider Nephila clavipes (Clark & Mangel 2000) to (1) assess optimal mate number for males, and (2) examine the conditions that have the largest effect on male mating rate. In addition, because body size is highly variable in this species and is known to affect male reproductive outcomes (Christenson et al. 1985; Christenson & Cohn 1988; Rittschof 2010) I (3) assess whether males show size dependent differences in strategies, tactics, and mating rates. The model generates a matrix of opti mal male decisions for every value of male state considered. This decision matrix defines the male reproductive strategy. By solving the model using size specific parameter values, I can compare strategies across different male sizes. After assessing diffe rences in strategies, I simulate a theoretical population of males whose behavior is constrained to the optimal decisions generated by the model. This simulation gives the reproductive outcomes and behavioral patterns (including male mating rate) associate d with a particular male strategy. I refer to reproductive outcomes and behavioral patterns generated by the simulation as tactics.
96 Study System Adult male N. clavipes travel between female webs at night and co habit with mature females for one day to sev eral weeks (Christenson et al. 1985) Females periodically change web sites and attract new males (Rittschof & Ruggles 2010) and field and empirical data suggest females in this species are likely polyandrous (C. Rittschof, in preparation). Females can store sperm for several months (Christen son et al. 1985) and lay their first clutch of eggs 30 40 days after their maturation molt (Christenson & Cohn 1988) Males can visit multiple webs (C. Rittschof, unpublished data) but m ate search is risky, and male mortality across a 20 day period can be as high as 88% (Vollrath & Parker 1992) On a web, males fight among themselves to mate with the f emale. Larger males typically win male male contests (Cohn et al. 1988; Rittschof 2010) Upon mating, males become permanently sperm depleted if they mate with a virgin female (Christenson 1989) However, if they mate with a non virgin female, males retain some sperm and are capable of re mating (Christenson & Cohn 1988) After copulation, m ales may continue to guard females, especially when mate d with a virgin (Christenson et al. 1985) Females vary in the degree of male competition on their webs as well as in the pay off associated with a successful mating. Large male group sizes (5 6 males) are typically observed on virgin female webs (Moore 1977) and webs with non virgin females (Rittschof 2010) that are close to oviposition (Vincent & Lailvaux 2006; C. Rittschof, unpublished manusc ript) A male that mates with a virgin female fathers the majority of offspring in her first clutch (Christenson et al. 1985; Christenson & Cohn 1988) and males that mate with a non virgin female fertilize approxi mately 20% of her clutch (Christenson & Cohn 1988) Thus, fertilization rate s do not explain why large
97 numbers of males are attracted to non virgin females close to oviposition (Vincent & Lailvaux 2006) Here I hypothesize that these females are attractive because they have a high probabili ty of surviving to lay eggs. Female death before oviposition (hereafter female mortality risk) could occur because of random processes like predation, starvation, or injury. Generally, this risk is lower for females close to oviposition, but the risk incre ases overall at the end of the reproductive season when the ambient temperature may not be adequate for a female to complete egg development before she dies (Higgins 2000) Considering both sperm competition and female mortality risk, in the model I define three types of females that correspond to three stages in adult female life: (1) newly matured virgin females (hereafter virgin females), ( 2) non virgin adult females that are more than one week from oviposition (hereafter mated females), and (3) non virgin gravid adult females that are within a week of oviposition (hereafter gravid females). Because of first male sperm precedence (Christenson & Cohn 1988) virgin females are the highest quality in terms of potential offspring sired, followed by mated and gravid females. Gravid females, who are closest to oviposition, have the lowest mortality risk, followed by mated and virgin females. In the model I address how spatial variation in female quality and male male com mating opportunity) or after copulation (i.e. mate guard or not). I also evaluate how strategies and tactics change as a result of seasonal variation in female quality, op erational sex ratio, and mate search mortality risk. Finally, because limited sperm
98 numbers are predicted to have a large effect on male mating rate and strategies, I manipulate male sperm availability across three versions of the model (see below). Metho ds Overview In this study, I build a dynamic state model to evaluate male mating rate and the accompanying strategies and tactics. In order to evaluate the effects of male sperm limitation, I design three separate models that differ in the amount of sperm males have available for mating. The Basic Model is parameterized most realistically to N. clavipes using empirical data. In the Restricted Sperm Model, males use all of their sperm in a single copulation, regardless of female type, and are thus limited t o monogyny. This situation is similar to other spider species where males mutilate their genitalia during copulation or are cannibalized and thus no longer able to mate after their first copulation (Andrade & Kasumov ic 2005; Schneider et al. 2005a; Fromhage & Schneider 2006; Segoli et al. 2006; Uhl et al. 2010a; Michalik and Rittschof, unpublished manuscript) In the third model, the Unrestricted Sperm Model, males use fewer sperm than is typical during copulations w ith non virgin females (see below), which increases their potential mating rate. This level of sperm availability is similar to species where males are able to re induct sperm during or after copulation (Huber 1998; Knoflach 2004) All models are described in detail below. The Basic Model Model structure Each time step in the model (t) corresponds to a single 2 4 hour day in which males can either (1) search for a female; or (2) stay on the current female web. In
99 each time step until he has mated the female or he leaves the web Because females mate immediately after molting to maturity (Christenson et al. 1985) if a male finds a virgin female web and does not mate within one time step, his web type changes from occupying a vi rgin web to occupying a mated web. However, within the time horizon of the model (21 days), females do not progress from mated to gravid (Cohn et al. 1988; C. Rittschof, unpublished data) The time horizon (T) is the lifespan of a male (21 days). Because all males die at the end of tim e (T), the expected future fitness (F) at T is zero. If a male does not have enough time to complete a mating, his fitness is discounted in proportion to the time shortage ( ). This is important when a male mates with a virgin female, because these matings require 48 hrs (Christenson et al. 1985) while all other matings require only 24 hrs. Four state variables affect male behavior and reproductive success: (1) web type (J; a male can occupy one of thre e types of webs, J = 1 3, or no web at all, J = 4); (2) energy level (E); (3) available sperm (X); and (4) mating history (I; i.e. whether the male has mated or guarded the current female for at least one time step). Following the natural patterns, differe nt types of females differ in the level of male male competition (m J ), the energetic costs of attempting to mate (c J ), the sperm requirements for successful mating ( J ), and the pay off to the male if he mates (f J ; Tables 5 1, C 1 C 2 ). Each male starts with 21 energy units (E). Males lose energy at a rate of s per unit time (Table 5 1). Males also expend energy fighting with each other and mating with females (c J ; Table C 2 ). Males rarely eat as adults (Christenson 1985), and so for simplicity, in the model the y never gain energy. If energy falls below 1 unit, the male
100 dies. If a behavior requires more energy than a male has, the fitness gained from that behavior is discounted in proportion to the shortage of energy ( ; Table 5 1). Males start with pedipalps ful l of sperm (100%) (X; Table 5 1). The minimum amount of sperm required to gain maximum fitness from mating with a given female type ( J ) is 100% for virgin females and estimated to be 40% for both mated and gravid females following empirical studies (Table C 1; Christenson & Cohn 1988; Christenson 1989) If a male does not have the required amount of sperm at the time of mating, his fitness from that mating is discounted in proportion to the shortage of sperm ( ; Table 5 1). The mating history variable (I; Table 5 1) tells whether the male has mated (if unmated, I = 1, if mated, I > 1) and also tells the number of days the male has guarded the female and successfully prevented other males from copulating with her (number of successful guarding days = I 2). For each time step a male successfully guards his mate, he accrues additional reproductive success. Decision 1: search If a male decides to search his expected fitness is: ( 5 1) where is male mortality risk for searching, r J is the encounter rate for each female type, m J is the probability of mating with the female encountered (Table C 2). C mate,J is the fitness gained from mating, where ( 5 2)
101 fitness, and depends on whether the male successfully mates ( F mate,J ) or not ( F notmated ): ( 5 3) ( 5 4) ( 5 5) Male mortality risk per web movement ( ; Table C 2) is uniform across male body sizes in order to allow for the possibility that size dependent mortality, as observed in census studies (Vollrath & Parker 1992) reflects a trade off between lifespan and reproductive success corresponding to size based alternative strategies. Overall, male mortality risk increases over the course of the season. Female encounter rate (r J ; Table C 2 ) is a function of the total density of females and the relative proportions of each type of female in the population. These proportions change over the season: total femal e density declines over the season ; the density of virgin females peaks mid season; and the density of gravid females peaks late in the season (C. Rittschof, unpublished data). J ; Table C 2) depends on the web type (J) bec ause male group size, and thus competition, changes across the three female web types (Christenson and Goist 1979; Rittschof 2010). Mating probability for each female type remains uniform because male group size on these webs is consistent throughout the s eason (C. Rittschof, personal observation). I assume that on virgin and mated webs, larger males have a higher mating probability (Rittschof 2010), but on gravid webs mating probability is equal with respect to male size because group sizes are very large which could eliminate large male advantage (C. Rittschof, personal observation).
102 first clutch a male sires (f J ). At the time of mating, a male gains only a portion of his p otential paternity share (y J *f J ; Table C 1 ) because successfully guarding the female is required to accrue the highest possible fitness from a mating (Linn et al. 2007; see below) Although female size (cephalotho rax width) predicts the number of eggs laid by a female to some degree (Higgins 2000, Rittschof 2010), there is a lot of residual variation in offspring number that cannot be explained by body size. Furthermore, male attraction is not correlated with female body size (C. Rittschof unpublished manuscript). Thus I do not consider female body size in the model as a component that affects male mating strategies, and I measure male fitness in terms of proportion, not total number, of offspring sired. Female surv ival probability (d J ; Table C 2 ) incorporates female mortality risk due to random processes and seasonal risk. Random risk is constant throughout the season. However, at the latest of the three time periods within the season, risk increases due to declining temperatures. During the third period, virgin females have no chance of oviposition, mated females have a 30% chance of surviving, while gravid female mortality risk is the same as in early and middle season periods (98.8% survival). Decision 2: stay A components of the structure of the model for this decision are very similar to the decision to search. If a male stays on a web and he has not mated with the female (I = 1) he wil l attempt to mate again, and his expected fitness is:
103 ( 5 6) If the male has already mated with the female (I>1) and stays, he will attempt to guard the female and prevent her from re mating: ( 5 7) A male increases his fitness incrementally (z J *f J ; Table C 1 ) each day he successfully guards a female he has mated: ( 5 8) He guards successfully with a probability k J (Table C 1) : ( 5 9) ( 5 10) However, there is a chance the female will abandon her web (w) during the time step, and if she does, the male is forced to search in the next time step (eqn 5 1). ( 5 11) Given the choice to search or stay, I can evaluate wh ether males decide to leave a web without mating (e.g. if they are choosy about their mates). As a limitation of the coding of the model (i.e. because a male always attempts to mate with a female if he is on her web), there is some probability a male will choose to search, find a web, and mate successfully during a single time step. In this case, the male does not have to opportunity to choose against the female whose web he finds. Size and season dependent strategies I designed nine sets of parameters (Tables C 1 and C 2) to evaluate differences in male strategy associated with three male body sizes (small, medium, and large) at
104 model correspond to males that range i n cephalothorax width from approximately 1.3 is approximately the month of August, and October in North Central Florida Model Manipulations In order to assess how male sperm limitation changes male reproductive strategies, I created two manipulations of the basic model where I increase and decrease rate, which is predicted to affect male mating strategies and tactics (Wedell et al. 2001). In the first manipulation the restricted sperm model, I increased the value of J so t hat males deplete their sperm stores after a single mating, regardless of female type. In the second manipulation the unrestricted sperm model, I lowered the value of J for mated and gravid females so that males only use half as much sperm ( in comparison to the basic model) to successfully inseminate these females The changes in sperm usage for both model manipulations affect potential mate number, but they do not change the proportion of the clutch the male sired, i.e., the pay off associated with each female Data Analysis The solution to the above models is a state and time dependent series of optimal decisions and associated expected fitness values. These decision matrices can be considered male mating strategies. In order to assess how these strate gies affect reproductive outcomes and behavioral patterns (including male mating rate), I
105 conducted Monte Carlo forward simulations of a population of males adopting the state dependent strategies identified in the Basic, Restricted, and Unrestricted model s (Clark & Mangel 2000) From these simulations, I collected data on the reproductive success, mating experience, and behavior of 60 males in each size class and season al period. Male reproductive success was calculated as proportion of clutch sired. Because most of the data distributions for reproductive success were non normal and in some cases bimoda l, I calculated the mean and 95% confidence limits using a bootstrapping method. To do this I re sampled the data set 1000 times with replacement to generate a distribution of sample means. I calculated the mean of this distribution as well as the 95% conf idence limits. Logistic regression and Chi squared Tests were used for categorical data analyses. Tests using logistic regression are noted in the results. In all other cases with a X 2 value listed, a Chi squared Test was used. Finally, I classified male s as having employed one of six behavioral tactics that incorporate mate number (monogyny versus polygyny) as well as female type (virgin, mated, or gravid). Because polygynous tactics can encompass mates of more than one female type, I established a hiera rchy for strategy assignment. If a polygynous male mated with a virgin female (regardless of the types of his other mates), he was assigned to the polygynous virgin category. If he did not mate with a virgin female and mated with at least one gravid female he was assigned to the polygynous gravid category. If he mated only with mated females, he was assigned to the polygynous mated category. All monogynous males were assigned to categories based on their single female mate monogynous virgin, monogynous ma ted, monogynous gravid.
106 Results Male Mating Strategies Basic model I defined 9 strategies in the basic model, 3 different male body sizes (small, medium, and large) each at 3 different season periods (early, middle, and late). In a pair wise comparison o f the strategies, I tallied differences in the optimal decision associated with each combination of state values and found that the 9 strategies ranged from 0.8% to 19% different from one another overall. Differences among male sizes in the strategies pred icted by the model were most pronounced in the early season period (11% different on average versus 8% different for the middle season and 5% different in the late season period), and large male strategies changed the most over the course of the season (12 % difference in early versus late season strategies, compared to 2% difference for small males and 9% different for medium males). Because medium male strategies and middle season strategies were intermediate between small and large males and early and lat e season respectively, here I will compare only small and large males and the early and late season in order to simplify the results. However, the data for medium sized males and the middle time period are included in calculations of male mating rate and m ale reproductive outcomes from the Monte Carlo simulations (see below). In general, across all strategies, if a male located a web, he remained on the web until mating successfully. There was evidence of gravid female abandonment towards the end of a male the web and then leave without mating in favor of attempting to mate on a virgin or
107 females at a higher rate than small males (about 23% of the time for large males, and 2% for small males). In the basic model, almost all differences among strategies corresponded to differences in male behavior after mating. If a male mated with a non virgin female (either mated or gravid), he had to deci de whether to leave the female after mating or stay behind to guard her. In the early season period, large males were very likely to stay behind to guard their non virgin female mates (Table 5 2; Fig 5 4). In contrast, in the late season period, large males be haved similarly to small males by becoming less likely to guard their mates. Restricted sperm model In the restricted sperm model, males were limited to a single mate. Because a single mating depleted males of their sperm, there was no variation in post copulatory guarding behavior among strategies (all males guarded; see methods). The major difference between male strategies in the basic and restricted sperm model was that in the restricted sperm model, for 63% of the state space when large males were u nmated but on a mated female web, they abandoned the web without mating. This was the optimal behavior up until the last 3 reverted to the strategy of staying on the web until mating. Small male strategies in the early season did not differ from the basic model. Unrestricted sperm model In the unrestricted sperm model, males used only half the sperm for each non virgin mating compared to the basic model (see methods). The general trends in mating strategies we re similar between the unrestricted sperm model and the basic model, although overall in the unrestricted sperm model, both small and large males spent a
108 lower proportion of the state space guarding non virgin females compared to the basic model (Table 5 2). Reproductive Outcomes Basic model Male mating rate. Across all body sizes and season periods 184 of 270 males successfully mated. Most males mated polygynously (N Polygyny = 117, N Monogyny = 67), and polygynous males mated with either two or three female s. Mean reproductive success was higher for polygyny relative to monogyny (Fig 5 1 ). Comparing males that had one, two, or three mates, mean reproductive success increased with mate number, although the difference between two and three mates was slight (F ig 5 2A). This is b ecause males used a minimum of 40% of their sperm at each mating, thus, males mating for a third time never achieved full pay off with their last mate. The effect of mate number on mean reproductive success changed with female type (Fig 5 3A ). Males that mated monogynously with mated or gravid females had lower mean reproductive success compared to the other four strategies. Monogynous males that mated with virgin females had mean reproductive success comparable to polygynous males, although the variance in reproductive success was much lower for polygynous males who mated with mated and gravid females compared to males who mated monogynously with virgin females Male size. Of the 90 males simulated for each size class (across all season peri ods) 61 small males, 63 medium males, and 60 large males mated successfully (X 2 2 = 0.24, P = 0.89). There was no difference in mean reproductive success across male size classes (Mean Small = 0.19, 95% CI: 0.16 0.24; Mean Medium = 0.2, 95% CI: 0.16 0.24; Me an Large = 0.18, 95% CI: 0.14 0.22). There was a trend towards a difference in
109 rates of monogyny and polygyny among male sizes where large males mated monogynously at a higher rate compared to other male sizes (X 2 2 = 3.6, P = 0.17). Size differences in ra tes of monogyny do not reflect differences in male mate guarding strategies with non virgin females (see above). Large males guard ed in 76% of mating events versus 12% for small males (X 2 1 = 80.1, P<0.0001; Fig 5 4 A B). However, mate guarding had no effect o n the number of subsequent matings achieved comparing large males (who typically guard) to small males (who typically do not guard ; X 2 2 = 0.89, P = 0.7). Difference in rates of polygyny among size classes could have occurred because large males mated with virgin females at slightly higher rates (X 2 2 = 5.2, P<0.075; Table C 3 ) and became sperm depleted after these matings. Seasonal effects. Male mean reproductive success declined across the time periods within the season for males of all body sizes (Fig 5 5B). In addition, as the season progressed, a greater proportion of males adopted a polygynous strategy (X 2 2 = 6.1, P<0.047). Guarding also decreased across time periods within the season (X 2 2 = 30.6, P<0.0001; Fig 5 4 ). Model comparisons Model similarities. In the restricted sperm model, males were limited to a single mate (see above). In the unrestricted sperm model, males had twice the number of available sperm compared to the basic model when mating with non virgin females. The basic model and the unrestri cted sperm model were similar in terms of male tactics. In both models most males were polygynous (Basic = 65% polygynous, Unrestricted = 64% polygynous). Polygyny was not possible in the restricted sperm model because males were only able to mate once. In the basic and unrestricted sperm models, the mean reproductive success for each of the six strategies showed similar patterns (Fig
110 5 3B). Combining the three time periods, equal proportions of small and large males successfully mated in all models (Restric ted model: X 2 1 = 0.17, P = 0.68, Unrestricted: X 2 1 = 0.85, P = 0.36), and t here were only small size based differences in mean reproductive success across models. Male mate guarding showed similar size based patterns in the basic and unrestricted sperm mod els. However, in general, guarding rates were much lower in the unrestricted sperm model for all male sizes. Model differences: male mating rate Under reduced sperm limitation, the advantage of being polygynous increased ( Fig 5 1B ), although mating monogyno usly had equal reproductive success to polygyny if the male mated with a virgin female (Fig 5 3B). In the unrestricted sperm model, polygynous males mated with between two and five females, compared to a maximum of three mates in the basic model (Fig 5 2A B). Unlike the basic model where males lost reproductive success due to sperm limitation at the third mating, in the unrestricted sperm model, the relationship between mate number and reproductive success continued to increase without bound because males coul d mate with up to five non virgin females in their lifetime with no sperm limitation (Fig 5 2B) In the basic model, there was a trend towards increased monogyny with increased male size. However, t here were no size based differences in the rate of monogyny in the unrestricted sperm model (X 2 1 = 0.001, P = 0.98). Model differences: seasonal effects. In contrast to the basic and unrestricted model where multiple mating was possible, there were seasonal differences in the number of small and large males that s uccessfully mated in the restricted sperm model In the early season, more small males successfully mated compared to large males (X 2 1 = 4.4, P<0.034), and in the late season this trend was reversed although the difference
111 was not significant (X 2 1 = 2.4, P = 0.12). Nonetheless, overall, across all models and season periods, large and small males had similar mean reproductive success (Fig 5 5). Only in the late season period of the restricted sperm model were there suggestions of large male advantage (Fig 5 5A) D iscussion I used a dynamic state modeling approach to (1) determine optimal male mating rate; (2) evaluate the factors that constrain male mating rate; and (3) determine whether males of different sizes adopt alternative mating strategies and tactics in the spider Nephila clavipes Male Mating Rate Model results suggest that there may be more than one mating rate optimum for male Nephila clavipes In the basic model, males achieved high reproductive success from mating at a maximum rate (i.e. polygynou sly with 2 or 3 females; Fig 5 2A), but they also achieved similar reproductive gain from mating monogynously with a female if she was a virgin (Fig 5 3A). Due to female mortality risk, the variance in reproductive success was much higher for mating monogynous ly with a virgin female compared to mating polygynously with either mated or gravid females (Fig 5 3A). These findings suggest that polygyny is a lower risk tactic compared to monogyny, even though the mean reproductive outcome between the two tactics is sim ilar. Males could mate polygynously as a way to spread female mortality risk and minimize the risk of total reproductive failure. There are patterns of reproduction in other species that suggest that males may benefit by mating multiply to distribute risk of brood failure. For example, in some species males exhibit a preference for novel females, which manifests as decreased
112 mating investment in a familiar female and increased investment in unfamiliar females (the Co olidge Effect; Wilson et al. 1963; Dewsbury 1981; Pierce et al. 1992; Pizzari et al. 2003; Koene & Ter Maat 2007) This preference for novelty is puzzling when it occurs despite intrinsic disparities in quality between the familiar and unfamiliar female. For example in the burying beetle Nicrophorus vespilloides, males decrease their latency to copulate when presented with a novel female, regardless of her virginity status and therefore the level of sperm competition the male may face. Males choose female s on the basis of novelty as opposed to reproductive pay off (in terms of number of offspring sired), which could result in males mating with multiple low pay off females instead of restricting their reproductive investment to a few high pay off females (Steiger et al. 2008) The Coolidge Effect is expected in c ases where there are diminishing returns from repeated copulations with a single female. These diminishing returns could occur because continuous mating results in physical exhaustion without increased sperm competition benefits (Jordan and Brooks 2010), o r because sperm are limited, and so it benefits males to allocate sperm prudently (Wedell et al. 2001). Here I suggest that a preference for novelty in general may simply promote a risk spreading mating tactic, beneficial in cases where mating at a higher rate with multiple females increases average reproductive success but also decreases variance in reproductive success that could result from female mortality or brood failure due to inadequate parental care. Constraints on Mating Rate Mate guarding One tr aditional factor thought to impose a constraint on male mating rate, mate search mortality risk (Vollrath and Parker 1992), did not constrain male mating rate in the model. However, similar to other species (Stockley et al. 1996; Zamudio & Sinervo
113 2000; Schubert et al. 2009) mate guarding affected male mating strategies, although it did not affect male mating rate. For males, mate guarding is typically considered to be a trade off between preventing female re mating and achieving matings with additional females (Alcock 199 4) However, in the model, male re mating frequency was comparable between large and small males even though large males spent more time guarding (Fig 5 4A). This could be because for large males there is little cost to mate guarding relative to small males because large males can guard and still have time to mate with the maximum number of females (3). However, small males, who are less competitive and have to wait longer once they reach a web to achieve successful copulation, do not have time to guard. Female density affected male guarding behavior in the opposite way than predicted based on empirical studies in other species, where decreased mate availability increases guarding rates (Latty 2006). At the end of the season, female density (i.e. mate avai lability) was lowest, but the frequency of guarding also decreased (Fig 5 4). Even with low female density and high mate search mortality risk (a 20% chance of failing to find a female coupled with 26% mortality risk per search), it was more beneficial for m ales to spend additional time searching for new mates in order to reach their maximum reproductive rate. Similarly, males with the greatest chances of re mating, large males, actually spent the most time guarding, because they were able to spend less time waiting to mate compared to small males once a new web was found. These results could in part be a result of the fact that guarding successfully carried little additional benefit for males. However, the substantial differences in guarding behavior
114 between the basic and unrestricted sperm models (where the benefit of guarding were the same; Fig 5 4) suggest that sperm availability may play an even greater role. The unrestricted sperm model revealed that guarding frequency was strongly affected by potential re productive rate. Changes in potential reproductive rate however, were not caused by decreased female density. Rather, because males had more sperm available for copulation, guarding frequency decreased in the unrestricted sperm model because males needed t o spend more time searching and copulating in order to reach an optimal mate number (5 mates; Fig 5 4B). Thus in general, it seems that males must have either an extremely low probability of re mating or very high benefits from guarding in order to off set t he potential benefits of mating with additional females. As a result, some other mating rate limitations (e.g. male sperm limitation, which is common across a number of taxa) may be required for guarding behavior to evolve. Sperm limitation could have a gr eater effect on mate guarding behavior than other traditional measures of male mating potential (e.g. female density and male competitive ability). Male sperm limitation and choosiness Sperm depletion constrained male mating rate in general (Fig 5 2A), bu t it also affected male reproductive tactics. For instance, there was a trend towards increased monogyny for large males compared to small males in the basic model. This difference was not a result of differences in male strategies, but rather a consequenc e of male sperm depletion: large males mated virgin females at a higher rate compared to small males (Table C 3), which depleted their sperm and eliminated re mating ability. Increased sperm limitation (restricted sperm model) also caused males to become c hoosy in some cases. Large males abandoned low quality mated female webs without mating in the early season period. This result is not surprising, because in
115 species in which females vary in quality, males are expected to be choosy about their reproductive investment, particularly when mating opportunities are limited (e.g. Hardling et al. 2008; Barry & Kokko 2010) However, female quality and number of mating opportunities are not the only important variables to consider when evaluating whethe r males should be choosy. In the basic model, even though males were sperm limited, males were never choosy, which is a consequence of the low cost associated with mating with a non virgin female, the reproductive benefits of mating multiply (Fig 5 1 5 3), and the mortality risk associated with searching for a new mate. It is only in the restricted sperm model, when males are most severely sperm limited (but all other parameters remain the same), do the benefits of choosiness outweigh the costs and risks associ ated with foregoing a mating opportunity. small males abandoned gravid female webs at a consistent rate across all three models during the early season period. Because males aban doned gravid females but not mated females in the basic and unrestricted models, it suggests that the reason for female abandonment was not that males were choosing against low quality females, but instead that males were attempting to re locate to a web w here their mating probability was higher. The mating probability on gravid webs is low for all males (due to large male group sizes on these webs). At the end of their lifetime, it is more beneficial for males to search for less competitive webs. This is e specially true for large males who have a good chance of out competing other males on other webs types. In contrast to abandonment behavior with gravid females, only large males chose against low quality mated females in favor of higher quality females in the restricted model.
116 Choosiness, as defined as discrimination against females based on quality, was For this reason, choosiness was exclusively a large male strategy. Larg e males abandoned mated female webs in order to pursue a mating with a higher quality female (in the early season of the forward simulation, a virgin female). In contrast, choosiness was not beneficial for small males because the alternative to mating with a low quality female was to move to a highly competitive web of a different type where small males have a low chance of successfully mating compared to large males. Thus there are size dependent costs and benefits of choosiness, a finding consistent with work in other species that suggests that male mate investment and mate choice strategies can change as a function of male condition, size, or ability to secure mates (Burley 1988; Hardling & Kokko 2005; Bel Venner et al. 2008; Candolin & Salesto 2009) In spiders in particular, male reproductive behavior is affected by experience during male male contests (Kasumovic et al. 2010) and so it is possible for males to assess their own competitive ability as well as the competitive environment on a web, and make decisions a ccordingly. In N. clavipes specifically, there is evidence that males arrive on webs and assess the competitive environment before deciding to move to a new web (Christenson & Goist 1979) Because small males were unable to be choosy, and thus mated indiscriminately, I might expect a large male reproductive advantage when large males are able to be choosy. Indis criminate mating with low quality mates compared to choosy mating with (Lee 2005) Surprisingly however, the ability to be choosy did not provide a reproductive advantage
117 for large males. Instead, when choosiness occurred (the early season period of the res tricted sperm model), this high risk strategy resulted in significantly fewer large males achieving copulations compared to small males. In addition, due to the variance associated with mating with a virgin female, there were no significant size based diff erences in male reproductive success during times where males were choosy (Fig 5 5A). These findings suggest that choosiness and indiscriminate mating may represent stable alternative strategies (Fawcett & Johnstone 2003; Hardling & Kokko 2005) Male Body Size Variation In my model, variation in female value coupled with constraints on male mating rate resulted i n weak selection on male body size despite competitive advantages of large size Past studies in other Nephila species have suggested that trade offs associated with male size result in weak stabilizing selection (Vollrath 1980) For example, while there is a large male advantage in contests ( Christenson & Goist 1979) small males have lower rates of sexual cannibalism (Elgar & Fahey 1996) an int ermediate size is favored in terms of climbing speed (Moya Larano et al. 2009) and there can be seasonal fluctuations in the strengt h of selection on size (Kasumovic et al. 2008) However, here I also demonstrate that spatial variation in the strength of sexual n part a result of variation in female attractiveness, provides opportunities for small males to achieve lifetime reproductive success similar to that of large males. The only scenario in which the model recovered body size differences in reproductive su ccess was extreme sperm limitation, where each male could mate only once (Fig 5 5A), and in this case, large male advantage only occurred in the late season period. This suggests that the ability to mate multiply coupled with variation in
118 female quality coul d lessen size based differences in reproductive success and weaken selection on male body size in species similar to N. clavipes Within the genus Nephila there are interspecific differences in the degree of male body size variation (Elgar & Fahey 1996; Schneider et al. 2000; Rittschof 2010) In some species males are constrained to a single mate because they are cannibalized during copulation (Schneider et al. 2000; which is somewhat analogous to our restricted sperm model; Kasumovic et al. 2007) If multiple mating weakens selection on body size, male body size variation should be more extensive in species where males potentially re mate. This hypothesis could be tested thr ough comparative studies of mating rates and body size variation among populations and species. For instance, two sympatric species of Nephila found in Australia show different rates of sexual cannibalism and different levels of male body size variation. I n N. edulis cannibalism is rare, giving males the potential to re mate, w hereas cannibalism is very common in N. plumipes (Schneider et al. 2000; Kasumovic et al. 2007) As I would expect from our model, the coeffi cient of variation for male body mass is 104% in N. edulis and only 31.7% in N. plumipes (Elgar & Fahey 1996; Schneider et al. 2000; Elgar et al. 2003) Final Remarks In this study, I use a modeling approach to ass ess male mating rate optima and alternative reproductive tactics in the spider N. clavipes Multiple mating rate optima are a function of female type and mate number. Mate guarding and male sperm depletion constrain male mating rate. Multiple factors contr ibute to whether a male will be choosy, including potential reproductive rate, female quality, benefits of multiple mating, and the costs of mate search. Different sized males adopt mating strategies and tactics that differ in terms of degree of mate guard ing and male choosiness. However, these
119 strategies and tactics result in similar reproductive pay offs, which suggests that alternative tactics may be present in this species.
120 Table 5 1 Parameter terms for the dynamic state model. Term Definition Stat e variables: T time horizon (21 days) t time step (1 day) J web status E energy level X proportion of sperm available for mating I mating history of the male at his current web Factors affecting state: cJ cost of male male competition tjJ tim e required to mate (days) J proportion of total sperm used during mating event s daily senescence (energy) cost Probabilistic events: male mortality risk (per search) rJ probability of encountering a female mJ probability of mating dJ female survival probability kJ probability of successfully guarding w probability that the female abandons her web Fitness: fJ total fitness (proportion of clutch sired) yJ proportion of fitness gained by mating fitness discount due to energy shortag e fitness discount due to time shortage fitness discount due to sperm shortage zJ proportion of fitness gained with each day guarding Subscript (J) denotes variables that change with female type.
121 Table 5 2. Summary of male strategies by degree of male sperm limitation, size, and season. Model Size Season Choosiness Guard Basic Small Early Yes (G) + Large Yes (G) +++++ Small Late No + Large No +++ Restricted Small Early Yes (G) Large Yes (G,M) Small Late No Large No Unres tricted Small Early Yes (G) + Large Yes (G) ++++ Small Late No + Large No ++ males guarded at least one day: +++++ = >50%, ++++ = 40 50%, +++ = 25 40%, ++ = 10 24%, + = <10% G denotes the gravid female type, M denotes the mated female type.
122 Figure 5 1. Mean reproductive success and 95% confidence limits are shown for polygyno us (Poly) versus monogynous (Mono) mating in the basic (A) and unrestricted (B) sperm models. Mating polygynously yields higher mean reproductive success compared to mating monogynously in both models. However, the degree of difference in reproductive succ ess between monogyny and polygyny is greater with weaker sperm limitation, i.e., in the unrestricted model (B versus A).
123 Figure 5 2. Mean reproductive success and 95% confidence limits as a function of mate number for the basic (A) and unrestricted sper m (B) models. In both models mean reproductive success increases with mate number. However, in the basic model, males are sperm limited at the third mating, and reproductive success asymptotes (A) while reproductive success increases without bound in the u nrestricted sperm model (B).
124 Figure 5 3. Mean reproductive success with 95% confidence limits for monogynous (Mono) and polygynous (Poly) mating outcomes divided by female type: virgin (V), mated (M), and gravid (G). Mean reproductive success for mati ng monogynously with a virgin female is similar to mating polygynously with lower quality (M or G) females.
125 Figure 5 4. The frequency of mating events where males chose to stay and guard the female for at least one day after copulation. Data only inclu de males that had sperm remaining after copulation (Basic model: N Small, Early = 33, N Large, Early = 32, N Small, Late = 27, N Large, Late = 28; Unrestricted sperm model: N Small, Early =81, N Large, Early =77, N Small, Late =68, N Large, Late =67). Results are divi ded by model and season: basic (A,B) and unrestricted sperm (C,D) models, early (A,C) and late (B,D) seasons. P values are the results of Chi squared Tests comparing small and large males for each model and season period. Males are more likely to guard if they are larger, if its early in the season, and if potential reproductive rate is low.
126 Figure 5 5. Mean reproductive success with 95% confidence limits for small (S) and large (L) males in each of the three models, restricted sperm (A), basic (B), and unrestricted sperm (C). Closed symbols correspond to values in the early season period, and open symbols correspond to the late season period. Across all models, reproductive success is lower in the late season period. However, size based differences were only evident in the restricted sperm model (A), and significant differences occurred only in the late season period.
127 CHAPTER 6 CONCLUSIONS Summary In the golden orb web spider, Nephila clavipes multiple factors interact to affect male body size vari ation. Large males have an advantage during contests over females, but this advantage depends on the number of other competitors present on the web. Large males have an advantage when male density is high (Fig 3 3), but large and small males copulate at eq ual frequencies when density is low (Chapter 3). Furthermore, male density on a web is highly variable (Fig 3 1) which is, at least in part, a result of variation in female attractiveness. Males in the field are attracted to adult females who are close to oviposition (Chapter 4), and in a mating experiment, males mated at a higher rate (Fig 4 3) and transferred more sperm when paired with non virgin females close to oviposition compared to younger non virgin females. Males that mated with gravid females clo se to oviposition or virgin females had the highest pay off in terms of total offspring sired (Fig 4 4, 4 5), although virgin females may be a high risk mating investment compared to gravid females because there is a long period of sperm storage between mati ng and oviposition, and female mortality is common during this time period (Chapter 4). Thus Chapters 3 and 4 show that male reproductive success is determined by a complex interaction between multiple measures of female quality and male competitor density These interactions have implications for the strength of selection on male body size. Using a dynamic state model (Chapter 5), I assessed how different sized males could optimize their reproductive success in a variable mating environment where females d iffer in quality and the rate of male attraction. In the model I account for factors
128 that could affect male mating strategies and reproductive success, including mortality risk while moving from web to web, and seasonal changes in male and female density. In addition, I consider male sperm depletion as a factor in the model because this unusual characteristic of this species has implications for male mating rate (Chapter 2). Male N. clavipes have a finite number of sperm, and once this sperm is depleted, th ey can no longer fertilize eggs (Chapter 2). I showed in Chapter 4 that males alter their sperm use depending on the quality of their mate, and the dynamic state model (Chapter 5) suggests that sperm limitation will only affect male reproductive strategies in very extreme cases (i.e. when males are limited to a single mate). Sperm depletion, however, could have implications for male body size evolution (Chapter 5; discussed below). The integrative dynamic state model showed that different sized males could employ subtly different strategies and tactics that optimize their reproductive success in response to a heterogeneous reproductive environment, resulting in similar levels of success across male sizes (Chapter 5). These findings suggest that environmental heterogeneity in selection pressure, coupled with male behavioral variation, could maintain male body size variation in N. clavipes Environmental Heterogeneity and Plasticity Environmental heterogeneity occurs across a variety of scales in space and tim e (Svensson & Sinervo 2004) On a large ecological scale, populations m ay be distinct, capable of local adaptation, and depending on the level of gene flow, speciation (Lande 1982; Arnqvist 1992; Johannesson et al. 1995; GarciaRamos & Kirkpatrick 1997; Kawecki 1997; Case & Taper 2000; H endry et al. 2002; Doebeli & Dieckmann 2003; Nosil & Crespi 2004) Such inter population dynamics can be complex, for example, environmental clines can result in stable hybrid zones between neighboring species or
129 populations (Hewitt 1988) However, environmental heterogeneity can also operate on very fine grained spatial scales, and as opposed to clinal variation, it may be more accurately described as a mosaic (Svensson & Sinervo 2004; Gosden & Svensson 2008) In this context, where there are high levels of genetic mixing and poorly defined sub populations and population boundaries, environmental variation remains a force that can spatially alter the stren gth or direction of selection and thus contribute to phenotypic variation (Levins 1968) The latter description of environmental heterogeneity applies to spider webs. Webs partition the environment into a mosaic of discrete patches, and in the case of N. clavipes the characteristics o f these patches vary (Chapters 3, 4; Agnarsson 2003; Rittschof 2010; Rittschof & Ruggles 2010) with implications for the evolution of male size. The resulting heterogeneous environment has continuity due to gene fl ow, but fine scale differences in the strength of selection on male body size could maintain variation in this trait, which in this species has a broad, unimodal distribution (Fig 3 2). For a single population in a heterogeneous environment, several proces ses can result in a broad unimodal trait distribution, including multiple phenotypic optima (Lande 1982; Brooks 2002) frequency or density dependent selection (Gross 1991; Pu nzalan et al. 2005) phenotypic plasticity (Gotthard & Nylin 1995) and behavioral plasticity (West Eberhard 2003) In N. clavipes I found evidence that all of these processes may contribute in part to the evolution of male size and the maintenance of a broad size distribution. In Chapter 3 I discuss how different webs could select for different optimal male sizes because large male advantage is density dependent. In this section I focus on 1) the rol e of phenotypic and behavioral plasticity in the maintenance of size
130 variation in N. clavipes and 2) whether size and behavior are co evolving or are independent phenotypes, with consequences for the evolution of size. Plasticity and Size Variation Phenot ypic plasticity is a common adaptive response to environmental variation ( Levins 1968 ). Many species show phenotypic plasticity for body size, maturing at the size best suited for the current environmental conditions (Indermaur et al. 2010) In N. clavipes large male size is advantageous in high male densities (Chapter 3), but because there are typically costs to growing large in terms of growth rate and emergence time (Roff 1992) males would benefit if they could emerge at a sm aller size when male density is low. There is evidence in other web building species that individuals can respond during development to pheromone cues that signal density (Kasumovic & Andrade 2006; Kasumovic et al. 2009a) in order to mature at an optimal size. For this reason it is important to examine the possibility that male size variation in N. clavipes is a consequence of density dependent phenotypic pl asticity in adult size. In the sister species Nephila plumipes (Kasumovic et al. 2009a) there is evidence that males assess local density cues during development (Andrade and Kasumovic 2006), and emerge at larger sizes when male density is high, and small sizes when male density is low. In N. plumipes females live in stable aggregations and as an adult, a male joins an aggregation close to where the male matured. On a small spatial scale, males might sense and respond to changes in density. On a larger sc ale (>5m) however, Kasumovic et al. (2009a) found no relationship between male size and female and male density, and so it appears that males are unable to assess larger scale differences in density. In contrast to N. plumipes the larger spatial scale is most relevant for N. clavipes because females are much more mobile in this species: they do
131 not have stable aggregations and they change web sites often (Rypstra ; Vollrath ; Vollrath & Houston ; Rittschof & Ruggles) In addition, due to frequent web movement and differential mate attraction, the number of males found on a female web (i.e. local of a male. Furthermore, males can travel large distances (C. Rittschof, unpublished data), and they do not always inhabit female webs found near their juvenile webs. Thus if male N. clavipes respond developmentally to changes in density, they should do so on a relatively large (>5 m) spati al scale. On a larger spatial scale, in N. clavipes there are population wide seasonal shifts in male density (from high to low), which means that large males could have a competitive advantage early in the season (Higgins) Similar shifts have been observed in the sister species N. plumipes and in addition, in this species, average male size decreases over the course of the se ason (Kasumovic et al 2009a). Such a decrease could be a plastic response to a decrease in male density and an increase in the number of available females (Kasumovic et al. 2009a). In N. clavipes too, there appears to be a slight decrease in average male b ody size over the course of the season (C. Rittschof, personal observation). However, in both species size variation within a particular time of the season (e.g. on a given day) is much greater than the degree of change in average size across the season. For instance, from the beginning to the end of the season, average male size in N. plumipes decreases by about 10%, however, within the population throughout the season the largest males are 110% the size of the smallest males. Kasumovic et al (2009a) conc lude that seasonal variation in size explains only a small component of the total degree of male size variation. Similarly, in
132 N. clavipes seasonal shifts in size do not explain the 2.5 fold variation in male size that is evident throughout the season. Fu rthermore, female N. clavipes have seasonal shifts in size that are similar to that of males, which suggests that both males and females may show decreased size as a result of a shared selection pressure. Perhaps at the end of the season all adults show de creased size as a result of accelerated development because of the short reproductive time period available to adults before the onset of winter (no individuals are capable of overwintering; Higgins 2000). Although there are seasonal changes in male densit y in N. clavipes the greatest source of variation occurs on a finer spatial and temporal scale. Neighboring webs differ greatly in the number of males present, and thus the intensity of male male competition (Fig 3 1). The degree of competition on a web c an change in a very short time period due to female web movement or egg development (Chapters 3,4; Rittschof & Ruggles 2010) and thus N. clavipes males may not benefit as much from phenotypic plasticity as they would from targeting webs best suited for their size. This response to environmental heterogeneity, behavioral plasticity, is fine tuned level and allows an organism to respond quickly to environmental change. Individuals can respond to changes in the environment over time by adopting behavioral strategies that account for systematic or cyclical variation in environmental conditions (Levins 1968) However, strategies vary greatly in t heir degree of plasticity and in the cues individuals use to assess the environmental conditions and alter their behavior. For instance, strategies may be responses to seasonal shifts in the environment (Gotthard and Nylin 1995), or they may be responses t o less predictable environmental changes, like shifts in prey availability, or conspecific or predator density
133 (Milinski and Heller 1978). At the simplest level, the behavioral repertoire of an organism allows the individual to respond to a complex environ ment in real time rather than wait for population level heritable changes at the DNA sequence level (Milinski & Heller 1978; West Eberhard 2003) Behavioral plastic ity is a form of phenotypic plasticity, and may be particularly important for an organism that is navigating a spatially variable reproductive environment. The dynamic state model (Chapter 5) assumes that N. clavipes males are capable of some level of ada ptive behavioral plasticity. In this case, males alter their strategy depending on the level of sperm they have available, the type of female on the web they occupy (which affects reproductive pay off), the time of the season (which is associated with chan ges in conspecific density and mate search mortality risk), their body size, and their energy level. In spiders generally, there is evidence that these assumptions about behavioral plasticity are valid. Spiders can change their behavior in response to past experiences (Whitehouse 1997; Kasumovic et al. 2009b; Kasumovic et al. 2010) which suggests that they may have knowledge about their own mating experience, and thus perhaps their sperm numbers or competitive abili ty. It is well known that male spiders can identify female reproductive state (and perhaps quality) using airborne and substrate born chemical cues (reviewed by Gaskett 2007) Males can also sense female body weight through the silk, which would be an indicator of gravidity (Chapter 4). In general, a number of arthropods alter their re productive behavior in response to their experience of conspecific density as a juvenile (e.g. Harrison 1980; Bouaichi & Simpson 2003; Higaki & Ando 2003; Poniatowski & Fartmann 2009) which means males may also hav e information about mate search
134 mortality risk or the level of male male competition in the population at a given time. Given this ability for adaptive plasticity, I used the dynamic state model to attempt to describe the optimal strategy different sized m ales could use to maximize their reproductive success. The dynamic state model revealed the types of strategies males should employ to maximize their fitness in a variable environment. Although there were some size based differences in male strategies in N clavipes males in the basic model (which was the similar basic strategy regardless of size or time of the season. Males, once a web was found, would remain on the we b until they had the opportunity to mate. Males did not target certain types of females or exhibit choosiness in most cases (although small males did leave highly competitive webs in some cases). The biggest difference across sizes was in guarding behavior Large males spent more time guarding females in order to prevent re mating compared to small males, and small males mated on average, at a slightly higher rate. Similar strategies across male sizes had different outcomes in terms of the types of females males successfully mated (Table C 3), but in terms of reproductive success, large and small males did equally well. Given that the assumptions of the model are true, male behavioral plasticity (whether size dependent and size independent) could be a poten t force that affects the evolution of male size in this species. The idea that males behave in size dependent ways in order to optimize their behavior could explain the results of the Kasumovic et al. (2009a) study that found that male body size variation results from density dependent phenotypic plasticity in N.
135 plumipes Upon surveying adult male body size in the field, the authors found that adult males in female aggregations with high male densities were typically larger compared to males found in low male density aggregations. The authors assume that this is because males matured in proximity to these aggregations and used density cues to optimize their body size in response to local conditions. As an alternative however, large males may target or spen d a longer time on webs or in aggregations with high male male competition because they have a density dependent competitive advantage (similar to N. clavipes ; Chapter 3) when in these areas In this case, when surveying only adult males, large males shoul d be more common on webs with higher male densities. Thus size dependent behavioral variation instead of phenotypic plasticity may also explain these results. A survey of both local sub adult and adult males may reveal whether adult males are responding be haviorally or morphologically to local conditions. The Link between Size and Behavior The connections between behavioral and morphological evolution can be difficult to ascertain, but it is clear that behavior influences and sometimes precedes the evolutio n of morphological traits (reviewed by West Eberhard 2003). For instance, visual courtship displays typically consist of behaviors that are enhanced by exaggerated morphological traits (Candolin 2003) Thus a behavi or may precede the evolution of morphology, but in some cases become evolutionarily linked to a particular morphological trait variant (Wcislo 1989) if the two traits are most beneficial when inherited together. However, behavioral plasticity can also be independent of any particular morphological trait (e.g. experience based changes in behavior; Kasumovic et al. 2009b) particularly if behavioral plasticity itself is the target of selection (West Eberhard 2003) If this is the case, a capacity for behavioral variation that allows
136 individuals to achieve similar levels of reproductive success regardless of morphology can weaken selection on that morpho logical trait. This latter concept can be considered a form of phenotypic accommodation (West Eberhard 2003). In addition, there could be consistent and repeatable inter individual differences in behavior that are not coupled to a particular phenotype, con sidered behavioral syndromes or personalities (Dall et al. 2004; Sih et al. 2004) Whether or not behavior and body size are linked in N. clavipes has important implications for the long term evolutionary consequences of behavioral plasticity in this species. In the case of N. clavipes male s, it is unclear whether body size and behavior are co evolving traits. The model (Chapter 5) assumes that body size is linked to behavior, because I developed the decision matrix (male strategies) using different parameterizations for different sized male s. Similarly, for example, in many alternative tactics, behavioral polymorphism and phenotypic polymorphism are often strongly coupled (reviewed by Taborsky et al. 2008) In N. clavipes it makes sense to couple behavior to size because there is evidence of size dependent differences in reproductive outcomes in this species (Christenson and Goist 1979; Cohn 1989 ; Rittschof 2010), and if males are behaving optimally, they are predicted to incorporate size based differences in reproductive outcomes into their strategies over evolutionary time. Body size and behavioral changes may also be related even if the two phe notypes are not linked in an evolutionary sense. For example, if males adjust their behavior in contests based on their own experience of winning, large males may be more likely to show increased levels of aggression compared to small males because they ar e more likely to win fights and learn that they have a competitive advantage (Kasumovic et al.
137 2010) It is difficult to evaluate the link between size and behavior empirically in N. clavipes because, as the model demonstrated, there may be only subtle differences in behavioral strategies among male sizes. Furthermore, male body size shows a continuous distribution. Are behavioral strategies distributed in a continuous manner? The dynamic state model is one approach to attempt to capture the true complexity of male strategies, and to provide a connection bet ween size and behavior. Evolutionary Consequences of Environmental Variation As discussed above, environmental variation corresponds to the degree of variation in a phenotypic trait under selection (Vanvalen 1965) This is because, over time, low environmental variation will increase the efficiency of selection, which decreases trait variation (Falconer & Mackay 1996) As a result, if environmental conditions change, historical processes, which have decreased trait variance, will limit the degree to which a species is capable of responding to their new environment (Lovette et al. 2002; Arbogast et al. 2006; Losos 2010; Mahler et al. 2010) Because the evolutionary history of a species affects its capacity to respond to a changing environment (Travisano et al. 1995) lineages vary in th eir potential evolvability, or ability to exploit a diverse or novel environment (Lovette et al. 2002; Arbogast et al. 2006; Losos 2010; Mahler et al. 2010) The concept of evolvability has theoretical implications but also practical implications for conservation initiatives. For example, why are some species able to exploit the (novel) urban environment while others fail? Are certain species more vulnerable to habitat fragmentation than others? What evolutionary les species? Interspecific differences in adaptability could be the result of evolutionary
138 the unpreceden ted and increasing levels of anthropogenic environmental change. These questions are particularly relevant for the field of behavior and the study of behavioral plasticity because behavior is often the first trait to respond to changed environmental condit ions, and as a result, behavioral plasticity can buffer selection on other morphological traits by providing a mechanism for individuals to adjust to changes in conditions (Wcislo 1989; West in real tim e, it may be useful to attempt to examine plasticity (i.e. either morphological or behavioral plasticity within a species) and its mechanisms across multiple time scales. Short term plasticity, a product of molecular and physiological mechanisms functionin g within a single individual, could have implications for large scale plasticity, linked to the capacity for ecological and evolutionary change in a species over time. Assumptions and Biases: Models Versus Data Even though the model allows for a high degr ee of male behavioral plasticity, male strategies are remarkably consistent. In almost all contexts, males remained on a be considered fairly conservative, especially be cause there is some evidence in this species that males are choosy and forego reproductive opportunities in certain contexts (Chapter 4). Disparities between the model strategies and behavioral data reflect the model parameterization that I used, for examp le, I estimated the benefit of mating with low quality females to be too high for males to overcome the costs of passing up a mating opportunity in the model, or I set the mortality risk of mate search too high to reveal male choosiness. Such disparities a re to be expected however, because I estimated parameters from empirical data, which were collected with certain assumptions and biases.
139 Although the model has parameter biases, inferences from the model also reveal interesting potential biases in empiric al assessment of male tactics. In the model, despite strong similarities among male strategies, these strategies still resulted in differences in male reproductive outcomes. For example, the model showed some differences in rates of multiple mating for mal es of different sizes, and large and small males mated with different types of females at different rates (Table C 3). According to the model, these outcomes do not reflect major differences in decision rules and strategies among male sizes (e.g. certain s ized males choose against certain kinds of females). However, a male who employs an optimal strategy is still subject to probabilistic events that affect the outcome and pay off associated with the decision. Size based differences in outcome probabilities in this case yielded very different reproductive outcomes and tactics even though males were using similar decision rules. In most systems, male mating tactics are defined from empirical and observational data (reviewed by Taborsky et al. 2008) and observed tactics are used to infer behavioral decision rules. For example, if I had observed large and small male s mating with different types of females in the wild, I may have assumed these males were following different decision rules. If I also observed large males mating at higher rates with virgin females, I might have assumed large males have higher reproducti ve measured empirically as tactics, may not relate directly to differences in behavioral decision rules (Brockmann 2002)
140 In some ways a modeling approach may more accurately capture t he fitness differences associated with alternative tactics. Brockmann (2008) reviews several reasons why it can be difficult to estimate and compare lifetime reproductive success across reproductive tactics accurately using empirical data. For example, the average success of a given tactic should include all animals that employ the tactic but it is easy to exclude animals that die without mating at all, because they were not observed (S.M. Shuster, as cited by Brockmann 2008) Similarly it can be difficult to assess accurately the outcome of certain lifetime fitness trade offs associated with a tactic, for example trade offs between l ifespan and reproductive success for a given mating event, especially while observing individuals within a short time frame (Banks & Thompson 1985) Both of these problems could be resolved by monitoring all individuals performing a given tactic (or at least an accurate subset of individuals that encompasses the range of variation) across their entire lifetimes. These data can be difficult to obtain even though they may change the interpretation of the costs and benefits associated with di fferent phenotypes or strategies. For instance, many studies have put forth arguments pertaining to the evolution of male body size and the maintenance of size variation in Nephilid spiders (e.g. Elgar & Fahey 1996; Schneider et al. 2000; Moya Larano et al. 2009) However, the broad and complex array of factors that could affect reproductive outcomes in these species, and the inability to observe individuals over their lifetimes, have made it difficult to interpret t he importance of body size in terms of male reproductive success and male strategies. However, a dynamic state modeling approach allows me to incorporate many of the standing hypotheses that attempt to explain male size, e.g. male mate search mortality ris k.
141 My model of N. clavipes mating behavior incorporates the parameter male mate search mortality risk, which has been hypothesized to be an important factor that affects male reproductive outcome and presumably male strategies and body size evolution (Vollrath & Parker 1992; Higgins & Rankin 2001; Andrade 2003; Foellmer & Fairbairn 2005b; Fromhage et al. 2007; but see Fromhage et al. 2008; De Mas et al. 2009) Mate search mortality risk is difficult to account for in the wild, because it is unclear whether males who disappear are dead or simply unobservable. In addition, it is difficult to say whether size based differences in mortality estimated from field survey data (Vollrath & Parker 1992) result from size based tactics that ultimately have equal pay offs, or if they truly reflect a cost associated with a particular size. Finally, if mate search mortality contributes substantial variation in reproductive success, this factor in the model more accurately captures the variance in reproduc tive success associated with different male sizes. When mate search mortality risk and the resulting variation in reproductive success is accounted for, it suggests that there are only weak size based differences in reproductive success. In addition to de scribing male strategies more accurately and associated reproductive success, factoring in mortality risk also provides insight about the evolution of permanent sperm depletion in this species. Males have a finite volume of sperm (Chapter 2). However, if m ost males fail to use all of their sperm in their lifetime, there may not be strong selection on males to regenerate sperm. The percentage of males calculation, because these m ales die with all of their sperm remaining. The model allows me to account for this group of males, which may be ignored in empirical studies
142 because males are extremely cryptic when they are not on a web. Model data suggests that most males in this specie s die with at least some sperm remaining, due in part to mortality during mate search (Chapter 2). Thus there may be very weak selection on males to regenerate sperm even if most males that mate do so with multiple partners (Chapter 5). A combination of mo deling and empirical approaches may lead to the most accurate inferences about behavioral decision rules, and the complex interplay among factors that influence male reproductive success in this species. Novel Insights for Behavioral Ecology Sperm Depleti on Sperm depletion is an unusual phenomenon, and its evolutionary causes and consequences in spiders are unclear (Chapter 2). Nephila clavipes is an unusual sperm depleting species because unlike other nephilids males are not cannibalized during copulati on, and they do not break off their pedipalps, and thus they presumably can re mate if they have sperm remaining (Chapter 2, Discussion). It appears that sperm depletion is the ancestral state in the Nephila and presumably evolved following the evolution of genital mutilation and male sacrifice behavior in this group (Chapter 2). One hypothesis to explain why male N. clavipes fail to regenerate sperm even though they retain functional pedipalps is that, unlike pedipalp breakage and cannibalism, which are c ontext dependent traits (see Chapter 2 for discussion), testes function may not be as evolutionarily labile, and once function is lost, it is hard to re gain. An alternative hypothesis is that male N. clavipes are not under strong selection to regenerate s perm because most males mate only once or not at all due to ecological constraints of locating mates (e.g. mate search mortality risk). Even with limited sperm, males can mate multiply because they can prudently allocate sperm to females
143 depending on the q uality of the female (Christenson & Cohn 1988; Chapter 4). Males may gain significantly from mating multiply in spite of their low potential mating rate (Chapter 5). Nephila clavipes males retain permanent sperm depletion even with the ability to mate mult iply, and this could either be a result of evolutionary constraints tradeoffs, or weak selection on sperm regeneration. Sperm depletion may have implications for body size evolution in N. clavipes because it limits the potential reproductive rate of high ly competitive males. In the model simulations, large males become sperm limited early in their lifetime because they have a higher success rate of copulating with a female once a web is found. As a result, large males may run out of sperm, potentially bef ore they encounter gravid females, which are rare but high pay off mates. Because large male mating rate is constrained, small males are able to match their mating rate and their lifetime reproductive success. In addition, because they often fail to copula te with a female before she abandons her web, small males retain sperm for a longer portion of their lifetime compared to large males and therefore have a greater chance of encountering and copulating with rarer gravid females. Male sperm depletion, which imposes a constraint on large male mating rate, may be one characteristic that results in equal reproductive success among different sized males. The suite of mate limiting traits found in spiders, including sperm depletion, genital mutilation, and sexual cannibalism, make spiders an interesting group to assess male costs of reproduction across multiple levels. One logical next step is to evaluate the physiological costs of sperm production in spiders. Although these costs are often assumed (Dewsbury 1982) or demonstrated to impact males behaviorally, few studies
144 have bridged the gap between the physiological costs of maintai ning testes function and sperm production and the implications for male sperm use and mating strategies. Comparing related species in which males cease sperm production altogether with ones that produce sperm throughout adulthood is a unique opportunity to directly address the energetic demands associated with sperm production. Partner Mortality Risk I address how males may respond to female mortality risk during mate choice, a mate choice criterion that has only been considered in one other species (Dunn et al. 2001) Male mate choice criteria typically include female size or weight (indicators of potential reproductive output), and female age or mated status (indicators of sperm competition risk; B onduriansky 2001) My findings in Nephila clavipes provide support for the hypothesis that males select females on the basis of mortality risk, in part because males appear to ignore female body size during mate choice, which is an indicator of fecundity in this species (Chapters 3, 4). Males instead select females on the basis of abdomen size, which can indicate condition and therefore fecundity (Higgins 2000) but is very strongly correlated with proximity to oviposition (Chapter 4). Model data (Chapter 5) and empirical data (Chapter 4) reveal two ways that males may account for female mortality risk in their mating strategies. Males can mate multiply and distribute mortality risk across multiple female mates (Chapter 5). Males can also be choosy, and allocate more sperm to low risk females (Chapter 4). Both of these tactics can increase male fitness and decrease fitness variance in comparison to mating monogynously, and in N. clavipes these two strategies are not mutually exclusive. If mortality risk accounts for more variation in male reproductive success than factors like sperm competition and female size, males should choose low risk femal es at
145 the expense of other indicators of potential reproductive success. In N. clavipes two factors suggest that males would benefit from selecting females on the basis of mortality risk. First, similar to the Dunn et al. (2001), in N. clavipes female su rvival is variable and there is a relatively long period of time between mating and oviposition. These factors contribute to variation in female reproductive output, and are important regardless of female size. Second, mortality risk may be a better predic tor of female fecundity than other factors like female size or sperm competition risk. Female body size, a typical indicator of fecundity, explains only a portion of the variation in offspring number among females (Higgins 2000; Chapters 3, 4) Furthermore, these data do not include females who die without reproducing at all. In addition, because females mate pol yandrously, there is high variation in paternity share as a function of mate order (Christenson & Cohn 1988; Chapter 4) For instance a male that mates with a virgin fertilizes 0 100% of her offspring in the first clutch if the female re mates, regardless of the number of sperm transferred. The range o f reproductive success for the second male is similarly variable (range 0 100%), and on average the first and second male split paternity almost equally in the first clutch (Chapter 4). Therefore males may not benefit from selecting against previously mate d females without also accounting for female mortality risk, which could have a more predictable impact on male offspring number compared to sperm competition risk. Mortality risk could influence male reproductive outcome in any species where females sto re sperm. However, the length of time between maturation and oviposition relative to female lifespan may be the key feature that predicts whether males respond to this risk factor. In N. clavipes
146 devote d to maturing and laying her first clutch of eggs. However in spiders generally, there is a long period of time between maturation and oviposition. For instance, in the bowl and doily spider, time between maturation and first oviposition can range from abo (Austad 1982) In comparison, in Drosophila melanogaster where adult females can live for roughly 45 days (Sun & Tower 1999) first oviposition can occur within 24 48 hrs only 2 lifespan, a relatively short period of time. Thus female mortality risk may be particularly important for male spiders to consider because females in this group have a characteristically long time interval between maturation and oviposition. Although other species may face a similar reproductive delay (e.g. the seaweed f ly; Dunn et al. 2001) such long periods before oviposition may not be common in other sperm storing species, e.g. insects. Final Remarks Here I have investigated how environmental heterogeneity may contribute to male body size variation in the spider Nep hila clavipes The high degree of variation in local density on female webs creates a variable reproductive environment for males. One phenotypic characteristic, behavioral plasticity, may be critical in allowing males to achieve similar levels of reproduc tive success regardless of size in this species. Future work should study the mechanistic coupling between body size and behavior, the lifetime, the link between short term behavioral plasticity and long term evolutionary ability to adapt to a novel environment, particularly in terms of behavior.
147 APPENDIX A SUPPLEMENTARY MATERIAL FOR C HAPTER 2 Table A 1. Real age (number of days after penultimate or final molt when sacrificed) and standardized age (age relative to the youngest sub adult male) for all males in the study
148 APPENDIX B SUPPLEMENTARY MATERIAL FOR CHAPTER 3 In order to det ermine paternity, I developed three polymorphic microsatellite markers (Table B 1) using protocols adapted from Fleischer & Lowe (1995) and Kandpal et al. (1994) Genotyping was done on a 3730 Automated Sequencer (Applied Biosys tems) using fluorescent labeled primers. GeneMarker ( Version 1.75 ) was used to determine allele sizes, and each score was verified visually. I screened for polymorphism, estimated allele frequencies, and tested for neutrality and linkage disequilibrium usi ng 190 parents from the mating experiment. Each locus was tested for neutrality using the Hardy Weinberg Exact Test and for linkage disequilibrium using (Raymond & Rousset 1995) The three m icrosatellite loci used in paternity analysis (NC_F, NC_R and NC_BB) did not differ significantly from Hardy Weinberg equilibrium (HW Exact Test, P=1.0, P=0.61, P=0.12 for NC_F, NC_R, and NC_BB, respectively), and were not in 2 = 0.99, P=0.61, X 2 =1.4, P=0.51, X 2 = 0.21 p=0.90). With the female parent known, the three polymorphic loci had a combined exclusion probability of 0.72 Mothers, candidate fathers, and 24 spiderlings for each experimental trial were genotyped at the three loci using fluorescent labeled primer s. DNA was extracted from adult and spiderling tissue using the Gentra Puregene Tissue Kit (Qiagen). I adjusted the amount of tissue used in DNA extraction with sex and age because of variation in extraction efficiencies. For adult females, one of the thi rd pair of legs was used, for adult males 2 3 full legs, and for all spiderlings the entire body.
149 Each microsatellite locus was amplified using polymerase chain reaction (PCR), and the products were pooled together in order to genotype all three loci sim ultaneously. Each 10uL PCR reaction mixture contained 4.96uL of H2O, 10X Taq polymerase buffer at 1X concentration, 1mM MgCl 2 200mM of each of four deoxynucleotide triphosphates (dNTPs), 0.3mM of each forward and reverse primer, 0.2 U of Taq DNA polymeras e (New England Biolabs), 2 uL of template DNA (approximately 10 ng). PCR conditions were similar across loci, although annealing temperatures varied (Table B 1): 94 for 5 min ; 94 for 30 s; variable (see Table B 1) 60 s; 72 for 60 s; repeat steps 2 4 for 39 cycles; 72 for 10 min. Each genotype was scored using GeneMarker and verified visually. The 24 offspring genotypes were sco red blind to the genotypes of the mother and candidate fathers. Once all genotypes were scored, the number of fathers respons ible for each clutch was determined using the maximum likelihood approach in the program GERUD 2.0 (Jones 2005) The likelihood calculation is based on Mendelian segregation probabilities and population le vel allele frequencies (Jones 2005) Given the progeny genotype array, GERUD 2.0 subtracts the maternal genotype and calculates the most likely genotype(s) for the father(s) of the progeny. If the likeliho od calculation determines there is one paternal genotype for the progeny array, GERUD 2.0 lists this genotype. In this case, the genotype listed is the sole father of the progeny. In cases of multiple paternity, more than one paternal genotype is required to explain the progeny array, and GERUD 2.0 lists the paternal genotypes as well as the likely number of offspring attributed to each father. In some cases of multiple paternity, more than one
150 combination of paternal genotypes is consistent with the progen y array. If this is the case, the program calculates a likelihood score for each paternal genotype combination. I compared the parental genotype outputs from GERUD 2.0 to the genotypes of the candidate fathers in each experimental group in order to deter mine the identity of the father(s). If GERUD 2.0 listed more than one paternal genotype solution for a progeny array, I chose paternal genotype solution with the highest likelihood score that was also consistent with genotypes for the candidate fathers wit hin the experimental group. Table B 1. Microsatellite loci used in paternity analysis Locus Motif Length (bp) N Temp ( C) Primer Sequence NC_F CA 247 2 62.1 ACCCATCTTGGGACCTTTTC AG AAAAAGCCAAGACCCAGA NC_R AT 162 3 54.0 AAAAATCTGTGATACCCACTGC TGTGTTGCGTTGTCCAAAAT NC_BB CT 289 18 55.4 5 GGAGAAATTACAGTTTAGATGCTTGA TCGTGTTAAGGAGCTTGGA TTT N is the number of alleles at each locus, and Temp refers to the annealing temperature for each locus.
151 APPENDIX C SUPPLEMENTARY MATERIAL FOR CHAPTER 5 Table C 1. Parameter values that change with female type Female Type Virgin Mated Gravid tj J time required to mate 2 0 0 J proportion of sperm used at mating 1 0.4 0.4 f J reproductive success 0.8 0.2 0.2 y J proportion of f J gained by mating 0.8 0.9 0.9 k J probability of successfully guarding S 0.5 0.5 0.25 M 0.61 0.6 1 0.37 L 0.87 0.01 0.87 0.005 0.76 0.005 z J proportion of f J gained by guarding per t *Value changes across model manipulations. Small, medium, and large denote male body sizes.
152 Table C 2. Parameter value that change with season and femal e type Early Middle Late Virgin Mated Gravid Virgin Mated Gravid Virgin Mated Gravid Male searching mortality risk 0.21 0.23 0.26 r J female encounter probability 0.25 0.35 0 0.2 0.4 0.2 0.2 0.4 0.2 m J mating probability S 0.17 0.2 5 0.17 0.17 0.25 0.17 0.17 0.25 0.17 M 0.25 0.5 0.17 0.25 0.5 0.17 0.25 0.5 0.17 L 0.5 0.75 0.17 0.5 0.75 0.17 0.5 0.75 0.17 d J female survival probability 0.828 0.885 0.988 0.864 0.909 0.988 0 0.3 0.988 c J cost o f male male competition (E) S 1 1 1 1 1 1 1 1 1 M 3 1 3 3 1 3 3 1 3 L 1 1 1 1 1 1 1 1 1 Early, middle, and late denote season periods, virgin, mated, and gravid denote female types, and S,M,L refer to three male si zes, small, medium, and large.
153 Table C 3. Of males that mated, the proportion of males within each size class that mated with each of three types of females. Size Virgin Mated Gravid Basic Model Small 0.07 0.87* 0.32 Medium 0.11 0.94* 0 .30 Large 0.22 0.77* 0.15 Restricted Sperm Model Small 0.05* 0.89* 0.06* Large 0.16* 0.61* 0.23* Unrestricted Sperm Model Small 0.16* 0.84 0.63 Large 0.43* 0.71 0.51 Significant differences among male size classes (P<0.05)
154 L IST OF REFERENCES Agnarsson, I. 2003. Spider webs as habitat patches The distribution of kleptoparasites (Argyrodes, theridiidae) among host webs (Nephila, tetragnathidae). Journal of Arachnology 31 344 349. Aisenberg, A. 2009. M ale performance and body size affect female re mating occurrence in the orb web spider Leucauge mariana (Araneae, Tetragnathidae). Ethology 115 1127 1136. Alberti, G. 1990. Comparative spermatology of Araneae. Acta Zoologica Fennica 190 17 34. Alc ock, J. 1994. Post insemination associations between males and females in insects the mate guarding hypothesis. Annual Review of Entomology 39 1 21. Andersson, M. B. 1994. Sexual Selection Princeton, N.J.: Princeton University Press. Andrade, M. C. B. 2003. Risky mate search and male self sacrifice in redback spiders. Behavioral Ecology 14 531 538. Andrade, M. C. B. & Banta, E. M. 2002. Value of male remating and functional sterility in redback spiders. Animal Behaviour 63 857 870. Andrade, M. C. B. & Kasumovic, M. M. 2005. Terminal investment strategies and male mate choice: Extreme tests of Bateman. Integrative and Comparative Biology 45 838 847. Angerbjorn, A., Arvidson, B., Noren, E. & Stromgren, L. 1991. The effect of winter food o n reproduction in the arctic fox, Alopex lagopus a field experiment. Journal of Animal Ecology 60 705 714. Arbogast, B. S., Drovetski, S. V., Curry, R. L., Boag, P. T., Seutin, G., Grant, P. R., Grant, B. R. & Anderson, D. J. 2006. The origin and div ersification of Galapagos mockingbirds. Evolution 60 370 382. Arnqvist, G. 1992. Spatial variation in selective regimes sexual selection in the water strider Gerris odontogaster Evolution 46 914 929. Arnqvist, G. & Nilsson, T. 2000. The evolutio n of polyandry: multiple mating and female fitness in insects. Animal Behaviour 60 145 164. Austad, S. N. 1982. 1st male sperm priority in the bowl and doily spider, Frontinella pyramitela (Walckenaer). Evolution 36 777 785. Austad, S. N. 1989. Gam e Theory and the evolution of animal contests. Trends In Ecology & Evolution 4 2 3.
155 Badyaev, A. V. 2002. Growing apart: an ontogenetic perspective on the evolution of sexual size dimorphism. Trends In Ecology & Evolution 17 369 378. Badyaev, A. V. & Foresman, K. R. 2000. Extreme environmental change and evolution: stress induced morphological variation is strongly concordant with patterns of evolutionary divergence in shrew mandibles. Proceedings of the Royal Society B Biological Sciences 267 3 71 377. Ball, M. A. & Parker, G. A. 2000. Sperm competition games: A comparison of loaded raffle models and their biological implications. Journal of Theoretical Biology 206 487 506. Ball, M. A. & Parker, G. A. 2007. Sperm competition games: the risk model can generate higher sperm allocation to virgin females. Journal of Evolutionary Biology 20 767 779. Banks, M. J. & Thompson, D. J. 1985. Lifetime mating success in the damselfly Coenagrion puella Animal Behaviour 33 1175 1183. Barry, K. L. & Kokko, H. 2010. Male mate choice: why sequential choice can make its evolution difficult. Animal Behaviour 80 163 169 Bateman, A. J. 1948. Intra sexual selection in Drosophila Heredity 2, 349 368. Bateman, P. W., Gilson, L. N. & Ferguson, J. W. H. 2001. Investment in mate guarding may compensate for constraints on ejaculate production in the cricket Gryllodes sigillatus. Ethology 107 1087 1098. Beauchamp, G. 2010. Relaxed predation risk reduces but does not eliminate sociality in birds. Bio logy Letters 6, 472 474. Beaumont, H. J. E., Gallie, J., Kost, C., Ferguson, G. C. & Rainey, P. B. 2009. Experimental evolution of bet hedging. Nature 462 90 U97. Bel Venner, M. C., Dray, S., Allaine, D., Menu, F. & Venner, S. 2008. Unexpected male choosiness for mates in a spider. Proceedings Of The Royal Society B Biological Sciences 275 77 82. Benson, K. E. 2007. Enhanced female brood patch size stimulates male courtship in Xiphophorus helleri 2007 Copeia 212 217. Bertin, A. & Cezilly, F. 2005. Density dependent influence of male characters on mate locating efficiency and pairing success in the waterlouse Asellus aquaticus : an experimental study. Journal of Zoology 265 333 338.
156 Bijlsma, R., Bundgaard, J. & Boerema, A. C. 2000. Does in breeding affect the extinction risk of small populations? predictions from Drosophila. Journal Of Evolutionary Biology 13 502 514. Blanckenhorn, W. U. 2000. The evolution of body size: What keeps organisms small? Quarterly Review of Biology 75 385 407. Boivin, G., Jacob, S. & Damiens, D. 2005. Spermatogeny as a life history index in parasitoid wasps. Oecologia 143 198 202. Bonduriansky, R. 2001. The evolution of male mate choice in insects: a synthesis of ideas and evidence. Biological Reviews 76 305 339. Bonduriansky, R. & Brooks, R. J. 1998. Male antler flies (Protopiophila litigata; Diptera : Piophilidae) are more selective than females in mate choice. Canadian Journal Of Zoology Revue Canadienne De Zoologie 76 1277 1285. Bouaichi, A & Simpson, S. J. 2003. Density dependent accumulation of phase characteristics in a natural population of the desert locust Schistocerca gregaria. Physiological Entomology 28 25 31. Brockmann, H. J. 2002. An experimental approach to altering mating tactics in male horseshoe crabs ( Limulus polyphemus ). Behavioral Ecology 13 232 238. Brockmann, H. J. 2008. Alternative reproductive tactics in insects. In: Alternative Reproductive Tactics: An Integrative Approach (Ed. by Oliveira, R. F., Taborsky, M., Brockmann, H.J.), pp. 177 223. Cambridge: Cambridge University Press. Brockmann, H. J. & Taborsky, M. 2008. Alternative reproductive tactics and the evolution of alternative allocation phenotypes. In: Alternative Reproductive Tactics: An Integrative Approach (Ed. by Oliveira, R. F., Taborsky, M., Brockmann, H.J.), pp. 25 43. Cambridge: Cambridge University Press. Brooks, R. 2002. Variation in female mate choice within guppy populations: population divergence, multiple ornaments and the maintenance of polymorphism. Genetica 116 343 358. Brown, S. G. 1985. Mating behavior of the golden orb weaving spider, Nephila clavipes .2. Sperm capacitation, sperm competition, and fecundity. Journal of Comparative Psychology 99 167 175. Bukowski, T. C. & Christenson, T. E. 1997a. Determinants of sperm release and storage in a spiny orbweaving spider. Animal Behaviour 53 381 395.
157 Bukowski, T. C. & Christenson, T. E. 1997b. Natural history and copulatory behavior of the spiny orbweaving spider Micrathe na gracilis (Araneae, Araneidae). Journal of Arachnology 25 307 320. Bukowski, T. C., Linn, C. D. & Christenson, T. E. 2001. Copulation and sperm release in Gasteracantha cancriformis (Araneae: Araneidae): differential male behaviour based on female mating history. Animal Behaviour 62 887 895. Burger, M. 2010. Complex female genitalia indicate sperm dumping in armored goblin spiders (Arachnids, Araneae, Oonopidae). Zoology 113 19 32. Burley, N. 1988. The differential allocation hypothesis An experimental test. American Naturalist 132 611 628. Candolin, U. 2003. The use of multiple cues in mate choice. Biological Reviews 78 575 595. Candolin, U. & Salesto, T. 2009. Does Competition Allow Male Mate Choosiness in Threespine Sticklebacks? American Naturalist 173 273 277. Candolin, U., Salesto, T. & Evers, M. 2007. Changed environmental conditions weaken sexual selection in sticklebacks. Journal of Evolutionary Biology 20 233 239. Case, T. J. & Taper, M. L. 2000. Interspecific compe tition, environmental gradients, gene flow, and the coevolution of species' borders. American Naturalist 155 583 605. Chapman, T., Arnqvist, G., Bangham, J. & Rowe, L. 2003. Sexual conflict. Trends In Ecology & Evolution 18 41 47. Charlesworth, D. & Charlesworth, B. 1987. Inbreeding depression and its evolutionary consequences. Annual Review of Ecology and Systematics 18 237 268. Christenson, T. E. 1989. Sperm depletion in the orb weaving spider Nephila clavipes (Araneae, Araneidae). Journal of Arachnology 17 115 118. Christenson, T. E., Brown, S. G., Wenzl, P. A., Hill, E. M. & Goist, K. C. 1985. Mating behavior of the golden orb weaving spider, Nephila clavipes .1. Female receptivity and male courtship. Journal Of Comparative Psychology 99 160 166. Christenson, T. E. & Cohn, J. 1988. Male advantage for egg fertilization in the golden orb weaving spider ( Nephila clavipes ). Journal Of Comparative Psychology 102 312 318.
158 Christenson, T. E. & Goist, K. C. 1979. Costs and benefits of m ale male competition in the orb weaving spider, Nephila clavipes Behavioral Ecology And Sociobiology 5 87 92. Clark, C. W. & Mangel, M. 2000. Dynamic State Variable Models in Ecology: Methods and Applications New York: Oxford University Press. Cod dington, J. A., Hormiga, G. & Scharff, N. 1997. Giant female of dwarf male spiders? Nature 385 687 688. Cohn, J. 1990. Is it the size that counts palp morphology, sperm storage, and egg hatching frequency in Nephila clavipes (Araneae, Araneidae). Jo urnal of Arachnology 18 59 71. Cohn, J., Balding, F. V. & Christenson, T. E. 1988. In defense of Nephila clavipes Postmate guarding by the male golden orb weaving spider. Journal of Comparative Psychology 102 319 325. Cohn, J. & Christenson, T. E. 1987. Utilization of resources by the male golden orb weaving spider Nephila clavipes (Araneae). Journal Of Arachnology 15 185 192. Conner, J. 1989. Density dependent sexual selection in the fungus beetle, Bolitotherus cornutus Evolution 43 137 8 1386. Corcobado, G., Rodriguez Girones, M. A., De Mas, E. & Moya Larano, J. 2010. Introducing the refined gravity hypothesis of extreme sexual size dimorphism. BMC Evolutionary Biology 10 236. Costa, F. G. 1998. Copulation pattern and fertilization success in male wolf spiders without pre or post copulatory sperm induction. Journal of Arachnology 26 106 112. Creel, S. 1993. Why cooperate Game theory and kin selection. Trends In Ecology & Evolution 8 71 72. Dall, S. R. X., Houston, A. I. & McNamara, J. M. 2004. The behavioural ecology of personality: consistent individual differences from an adaptive perspective. Ecology Letters 7 734 739. Danforth, B. N. & Desjardins, C. A. 1999. Male dimorphism in Perdita portalis (Hymenoptera, And renidae) has arisen from preexisting allometric patterns. Insectes Sociaux 46 18 28.
159 Danielson Franois, A. M. 2006. Natural history of Glenognatha emertoni (Araneae, Tetragnathidae): mating behavior and sperm release in a haplogyne Journal of Arach nology 34 387 398. Danielson Francois, A. M. & Bukowski, T. 2005. Female mating history influences copulation behavior but not sperm release in the orb weaving spider Tetragnatha versicolor (Araneae, Tetragnathidae). Journal of Insect Behavior 18 13 1 148. De Mas, E., Ribera, C. & Moya Larano, J. 2009. Resurrecting the differential mortality model of sexual size dimorphism. Journal Of Evolutionary Biology 22 1739 1749. Dewsbury, D. A. 1981. The Coolidge effect in northern grasshopper mice ( Onych omys leucogaster ). Southwestern Naturalist 26 193 197. Dewsbury, D. A. 1982. Ejaculate cost and male choice. American Naturalist 119 601 610. Dewsbury, D. A. 1983. Recovery from sexual satiety in deer mice ( Peromyscus maniculatus Bairdi) Journal o f Comparative Psychology 97 34 42. Doebeli, M. & Dieckmann, U. 2003. Speciation along environmental gradients. Nature 421 259 264. Dunn, D. W., Crean, C. S. & Gilburn, A. S. 2001. Male mating preference for female survivorship in the seaweed fly Gl uma musgravei (Diptera : Coelopidae). Proceedings of the Royal Society of London Series B Biological Sciences 268 1255 1258. Eberhard, W. G. 2004. Why study spider sex: Special traits of spiders facilitate studies of sperm competition and cryptic fem ale choice. Journal of Arachnology 32 545 556. Elgar, M. A. 1991. Sexual cannibalism, size dimorphism, and courtship behaviour in orb weaving spiders (Araneidae). Evolution 45 444 448. Elgar, M. A., Bruce, M. J. & de Crespigny, F. E. C. 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. Elgar, M. A. & Jones, T. M. 2008. Size dependent mating strategies and the risk of cannibalism. Biological Journal of the Linnean Society 94 355 363.
160 Emlen, D. J., Hunt, J. & Simmons, L. W. 2005. Evolution of sexual dimorphism and male dimorphism in the expression of beetle horns: Phylogenetic evidence for modularity, evolutionary lability, and constraint. American Naturalist 166 S42 S68. Emlen, S. T. & Oring, L. W. 1977. Ecology, sexual selection, and evolution of mating systems. Science 197 215 223. Engqvist, L. & Reinhold, K. 2006. Theoretical influence of female mating status and remating propensity on male sperm allocation patterns. Jour nal of Evolutionary Biology 19 1448 1458. Engstrom Ost, J. & Candolin, U. 2007. Human induced water turbidity alters selection on sexual displays in sticklebacks. Behavioral Ecology 18 393 398. Fairbairn, D. J. 2007. Introduction: the enigma of sex ual size dimorphism. In: Sex, Size, and Gender Roles: Evolutionary Studies of Sexual Size Dimorphism (Ed. by Fairbairn, D. J., Blanckenhorn, W.U., Szekely, T.), pp. 1 12. Oxford: Oxford University Press. Falconer, D. S. & Mackay, T. F. C. 1996. Metric characters under natural selection. In: Introduction to Quantitative Genetics (Ed. by Falconer, D. S.), pp. 345 348. London: Longman Group. Farr, J. A. 1977. Social behavior of the golden silk spider Nephila clavipes (Linnaeus) (Araneae, Araneidae). Jo urnal of Arachnology 4 137 144. Fawcett, T. W. & Johnstone, R. A. 2003. Mate choice in the face of costly competition. Behavioral Ecology 14, 771 779. Fernandez Montraveta, C. & Moya Larano, J. 2007. Sex specific plasticity of growth and maturation size in a spider: implications for sexual size dimorphism. Journal of Evolutionary Biology 20 1689 1699. Fleischer, R. C. & Lowe, S. 1995. Construction and screening of microsatellite enriched genomic libraries. In: Molecular Zoology: Advances, Strate gies and Protocols (Ed. by Ferraris, J. a. P., S.), pp. 459 468. New York: Wiley Liss. Foelix, R. F. 1996. Biology of Spiders Second edition. Oxford: Oxford University Press. Foellmer, M. W. & Fairbairn, D. J. 2003. Spontaneous male death during copula tion in an orb weaving spider. Proceedings of the Royal Society London B Biological Sciences 270 S183 S185.
161 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. 1992. The stereotyped behavior of sexual cannibalism in Latrodectus hasselti Thorell (Ar aneae: Theridiidae), the Australian redback spider. Australian Journal of Zoology 40 1 11. Fox, C. W. & Rauter, C. M. 2003. Bet hedging and the evolution of multiple mating. Evolutionary Ecology Research 5, 273 286. Fromhage, L., Elgar, M. A. & Schn eider, J. M. 2005. Faithful without care: The evolution of monogyny. Evolution 59 1400 1405. Fromhage, L., Jacobs, K. & Schneider, J. M. 2007. Monogynous mating behaviour and its ecological basis in the golden orb spider Nephila fenestrata Ethology 113 813 820. Fromhage, L., McNamara, J. M. & Houston, A. I. 2008. A model for the evolutionary maintenance of monogyny in spiders. Journal of Theoretical Biology 250 524 531. Fromhage, L. & Schneider, J. M. 2005. Safer sex with feeding females: sex ual conflict in a cannibalistic spider. Behavioral Ecology 16 377 382. Fromhage, L. & Schneider, J. M. 2006. Emasculation to plug up females: the significance of pedipalp damage in Nephila fenestrata Behavioral Ecology 17 353 357. Fromhage, L., U hl, G. & Schneider, J. M. 2003. Fitness consequences of sexual cannibalism in female Argiope bruennichi Behavioral Ecology and Sociobiology 55 60 64. Gage, M. J. G. & Cook, P. A. 1994. Sperm size or numbers effect of nutritional stress upon eupyre ne and apyrene sperm production strategies in the moth Plodia interpunctella (Lepidoptera, Pyralidae). Functional Ecology 8 594 599. Garcia Ramos, G. & Kirkpatrick, M. 1997. Genetic models of adaptation and gene flow in peripheral populations. Evoluti on 51 21 28.
162 Gaskett, A. C. 2007. Spider sex pheromones: emission, reception, structures, and functions. Biological Reviews 82 26 48. Gosden, T. P. & Svensson, E. I. 2008. Spatial and temporal dynamics in a sexual selection mosaic. Evolution 62, 8 45 856. Gotthard, K. & Nylin, S. 1995. Adaptive plasticity and plasticity as an adaptation a selective review of plasticity in animal morphology and life history. Oikos 74 3 17. Grant, P. R. & Grant, B. R. 2002. Unpredictable evolution in a 30 year study of Darwin's finches. Science 296 707 711. Gross, M. R. 1991. Evolution of alternative reproductive strategies frequency dependent sexual selection in male bluegill sunfish. Philosophical Transactions of the Royal Society of London Series B B iological Sciences 332 59 66. Hardling, R., Gosden, T. & Aguilee, R. 2008. Male mating constraints affect mutual mate choice: prudent male courting and sperm limited females. The American Naturalist 172 259 271. Hardling, R. & Kokko, H. 2005. The e volution of prudent choice. Evolutionary Ecology Research 7 697 715. Harrison, R. G. 1980. Dispersal polymorphisms in insects. Annual Review of Ecology and Systematics 11 95 118. Head, G. 1995. Selection on fecundity and variation in the degree of sexual size dimorphism among spider species (class Araneae). Evolution 49 776 781. Hedrick, A. V. & Dill, L. M. 1993. Mate choice by female crickets is influenced by predation risk. Animal Behaviour 46 193 196. Hendry, A. P., Taylor, E. B. & McPhai l, J. D. 2002. Adaptive divergence and the balance between selection and gene flow: Lake and stream stickleback in the misty system. Evolution 56 1199 1216. Herberstein, M. E., Gaskett, A. C., Schneider, J. M., Vella, N. G. F. & Elgar, M. A. 2005. Li mits to male copulation frequency: Sexual cannibalism and sterility in St Andrew 's Cross spiders (Araneae, Araneidae). Ethology 111 1050 1061. Herberstein, M. E., Schneider, J. M. & Elgar, M. A. 2002. Costs of courtship and mating in a sexually canni balistic orb web spider: female mating strategies and their consequences for males. Behavioral Ecology and Sociobiology 51 440 446.
163 Hewitt, G. M. 1988. Hybrid zones natural laboratories for evolutionary studies. Trends In Ecology & Evolution 3 15 8 167. Higaki, M. & Ando, Y. 2003. Effects of crowding and photoperiod on wing morph and egg production in Eobiana engelhardti subtropica (Orthoptera : Tettigoniidae). Applied Entomology and Zoology 38 321 325. Higgins, L. 1992. Developmental plastic ity and fecundity in the orb weaving spider Nephila clavipes Journal Of Arachnology 20 94 106. Higgins, L. 1993. Constraints and plasticity in the development of juvenile Nephila clavipes in Mexico. Journal Of Arachnology 21 107 119. Higgins, L. 2 000. The interaction of season length and development time alters size at maturity. Oecologia 122 51 59. Higgins, L. E. & Rankin, M. A. 2001. Mortality risk of rapid growth in the spider Nephila clavipes Functional Ecology 15 24 28. Hodge, M. A. & Uetz, G. W. 1992. Antipredator benefits of single species and mixed species grouping by Nephila clavipes (L) (Araneae, Tetragnathidae). Journal of Arachnology 20 212 216. Hone, D. W. E. & Benton, M. J. 2005. The evolution of large size: how does Cope 's Rule work? Trends In Ecology & Evolution 20 4 6. Hormiga, G., Scharff, N. & Coddington, J. A. 2000. The phylogenetic basis of sexual size dimorphism in orb weaving spiders (Araneae, Orbiculariae). Systematic Biology 49 435 462. Huber, B. A. 199 5. Genital morphology and copulatory mechanics in Anyphaena accentuata (Anyphaenidae) and Clubiona pallidula (Clubionidae: Araneae). Journal of Zoology 235 689 702. Huber, B. A. 1998. Spider reproductive behaviour: a review of Gerhardts work from 19 11 1933, with implications for sexual selection. Bulletin of the British Arachnological Society 11 81 91. Huber, B. A. 2005. Sexual selection research on spiders: progress and biases. Biological Reviews 80 363 385. Indermaur, L., Schmidt, B. R., To ckner, K. & Schaub, M. 2010. Spatial variation in abiotic and biotic factors in a floodplain determine anuran body size and growth rate at metamorphosis. Oecologia 163 637 649.
164 Jennions, M. D. & Petrie, M. 2000. Why do females mate multiply? A review of the genetic benefits. Biological Reviews 75 21 64. Jirotkul, M. 1999. Operational sex ratio influences female preference and male male competition in guppies. Animal Behaviour 58 287 294. Johannesson, K., RolanAlvarez, E. & Ekendahl, A. 1995. In cipient reproductive isolation between two sympatric morphs of the intertidal snail Littorina saxatilis Evolution 49 1180 1190. Jones, A. G. 2005. GERUD 2.0: a computer program for the reconstruction of parental genotypes from half sib progeny array s with known or unknown parents. Molecular Ecology Notes 5 708 711. Jones, T. M. & Elgar, M. A. 2008. Male insemination decisions and sperm quality influence paternity in the golden orb weaving spider. Behavioral Ecology 19 285 291. Juenger, T. & Bergelson, J. 2002. The spatial scale of genotype by environment interaction (GEI) for fitness in the loose flowered gilia, Ipomopsis laxiflora (Polemoniaceae). International Journal of Plant Sciences 163 613 618. Kandpal, R. P., Kandpal, G. & Weissma n, S. M. 1994. Construction of libraries enriched for sequence repeats and jumping clones, and hybridization selection for region specific markers. Proceedings of the National Academy of Sciences of the United States of America 91 88 92. Kasumovic, M M. & Andrade, M. C. B. 2006. Male development tracks rapidly shifting sexual versus natural selection pressures. Current Biology 16 R242 R243. Kasumovic, M. M., Bruce, M. J., Andrade, M. C. B. & Herberstein, M. E. 2008. Spatial and temporal demograp hic variation drives within season fluctuations in sexual selection. Evolution 62 2316 2325. Kasumovic, M. M., Bruce, M. J., Herberstein, M. E. & Andrade, M. C. B. 2007. Risky mate search and mate preference in the golden orb web spider ( Nephila plum ipes ). Behavioral Ecology 18 189 195. Kasumovic, M. M., Bruce, M. J., Herberstein, M. E. & Andrade, M. C. B. 2009a. Evidence for developmental plasticity in response to demographic variation in nature. Ecology 90 2287 2296. Kasumovic, M. M., Elias, D. O., Punzalan, D., Mason, A. C. & Andrade, M. C. B. 2009b. Experience affects the outcome of agonistic contests without affecting the selective advantage of size. Animal Behaviour 77 1533 1538.
165 Kasumovic, M. M., Elias, D. O., Sivalinghem, S., Mason A. C. & Andrade, M. C. B. 2010. Examination of prior contest experience and the retention of winner and loser effects. Behavioral Ecology 21 404 409. Katvala, M. & Kaitala, A. 2001. Male choice for current female fecundity in a polyandrous egg carr ying bug. Animal Behaviour 62 133 137. Kawecki, T. J. 1997. Sympatric speciation via habitat specialization driven by deleterious mutations. Evolution 51 1751 1763. Kingsolver, J. G. & Pfennig, D. W. 2007. Patterns and power of phenotypic selection in nature. Bioscience 57 561 572. Klemme, I. & Ylonen, H. 2010. Polyandry enhances offspring survival in an infanticidal species. Biology Letters 6 24 26. Knell, R. J. 2009. Population density and the evolution of male aggression. Journal of Zoolo gy 278 83 90. Knoflach, B. 2004. Diversity in the copulatory behaviour of comb footed spiders (Araneae, Theridiidae). In: Diversity and Biology of Spiders, Scorpions and other Arachnids (Ed. by Thaler, K.), pp. 161 256. Knoflach, B. & van Harten, A. 2000. Palpal loss, single palp copulation and obligatory mate cosumption in Tidarren cuneolatum (Tullgren, 1910) (Araneae, Theridiidae). Journal of Natural History 34 1639 1659. Knoflach, B. & van Harten, A. 2001. Tidarren argo sp. nov. (Araneae: The ridiidae) and its exceptional copulatory behaviour: emasculation, male palpal organ as a mating plug and sexual cannibalism. Journal of Zoology 254 449 459. Koene, J. M. & Ter Maat, A. 2007. Coolidge effect in pond snails: male motivation in a simult aneous hermaphrodite. Bmc Evolutionary Biology 7 212. Kuntner, M., Agnarsson, I. & Gregori 2009a. Spider eunuch phenomenon induced by female or rival male aggressiveness. Journal of Arachnology 37 266 271. Kuntner, M., Coddington, J. A. & Schneider, J. M. 2009b. Intersexual arms race? Genital coevolution in Nephilid spiders (Araneae, Nephilidae). Evolution 63 1451 1463. Kuntner, M., Kralj Fiser, S., Schneider, J. M. & Li, D. 2009c. Mate plugging via genital mutilation in nephilid spiders: an evolutionary hypothesis. Journal of Zoology 277 257 266.
166 Lafferty, K. D. & Kuris, A. M. 2002. Trophic strategies, animal diversity and body size. Trends In Ecology & Evolution 17 507 513. Lande, R. 1982. Rapid origin of sexual isolation and character divergence in a cline. Evolution 36 213 223. Lee, J. S. F. 2005. Alternative repro ductive tactics and status dependent selection. Behavioral Ecology 16 566 570. Lemaitre, J. F., Rigaud, T., Cornet, S. & Bollache, L. 2009. Sperm depletion, male mating behaviour and reproductive 'time out' in Gammarus pulex (Crustacea, Amphipoda). A nimal Behaviour 77 49 54. Levins, R. 1968. Evolution in Changing Environments Princeton: Princeton University Press. Linn, C. D., Molina, Y., Difatta, J. & Christenson, T. E. 2007. The adaptive advantage of prolonged mating: a test of alternative hy potheses. Animal Behaviour 74 481 485. Losos, J. B. 2010. Adaptive Radiation, Ecological Opportunity, and Evolutionary Determinism. American Naturalist 175 623 639. Lovette, I. J., Bermingham, E. & Ricklefs, R. E. 2002. Clade specific morphological diversification and adaptive radiation in Hawaiian songbirds. Proceedings of the Royal Society of London Series B Biological Sciences 269 37 42. Maddison, W. P. 1990. A method for testing the correlated evolution of two binary characters: are gains or losses concentrated on certain branches of a phylogenetic tree? Evolution 44 539 557. Maddison, W. P. & Maddison, D. R. 2000. MacClade version 4: Analysis of phylogeny and character evolution. Sunderland Massachusetts: Sinauer Associates. Maddison W. P. & Maddison, D. R. 2010. Mesquite. Mahler, D. L., Revell, L. J., Glor, R. E. & Losos, J. B. 2010. Ecological opportunity and the rate of morphological evolution in the diversification of Greater Antillean anoles. Evolution 64 2731 2745. Matess i, G., Carmagnani, C., Griggio, M. & Pilastro, A. 2009. Male rock sparrows differentially allocate nest defence but not food provisioning to offspring. Behaviour 146 209 223.
167 McLain, D. K. 1992. Population density and the intensity of sexual selection on body length in spatially or temporally restricted natural populations of a seed bug. Behavioral Ecology And Sociobiology 30 347 356. McNamara, J. M., Wilson, E. M. K. & Houston, A. I. 2006. Is it better to give information, receive it, or be igno rant in a two player game? Behavioral Ecology 17 441 451. Michalik, P. 2006. Zur Morphologie des mnnlichen Genitalsystems von Spinnen (Araneae, Arachnida) unter besonderer Bercksichtigung der Ultrastruktur der Spermien und deren Genese. PhD thesis, Ernst Moritz Arndt Universitt. Michalik, P. 2009. The male genital system of spiders (Arachnida, Araneae) with notes on the fine structure of seminal secretions. Contributions to Natural History 12 959 972. Michalik, P. & Hormiga, G. 2010. Ultrastr ucture of the spermatozoa in the spider genus Pimoa new evidence for the monophyly of Pimoidae plus Linyphiidae (Arachnida: Araneae). American Museum Novitates 3682 1 17. Michalik, P. & Huber, B. A. 2006. Spermiogenesis in Psilochorus simoni (Berlan d, 1911) (Pholcidae, Araneae): Evidence for considerable within family variation in sperm structure and development. Zoology 109 14 25. Michalik, P., Knoflach, B., Thaler, K. & Alberti, G. 2010. Live for the moment male adaptations in a sexually ca nnibalistic spider (Theriidae, Araneae). Tissue & Cell 42 32 36. Michalik, P., Sacher, P. & Alberti, G. 2006. Ultrastructural observations of spermatozoa of several tetragnathid spiders with phylogenetic implications (Araneae, Tetragnathidae). Journa l of Morphology 267 129 151. Michalik, P. & Uhl, G. 2005. The male genital system of the cellar spider Pholcus phalangioides (Fuesslin, 1775) (Pholcidae, Araneae): development of spermatozoa and seminal secretion. Frontiers in Zoology 2 12. Milinsk i, M. & Heller, R. 1978. Influence of a predator on optimal foraging behavior of sticklebacks ( Gasterosteus aculeatus L). Nature 275, 642 644 Miller, J. A. 2007. Repeated evolution of male sacrifice behavior in spiders correlated with genital mutilati on. Evolution 61 1301 1315. Molina, Y. & Christenson, T. E. 2008. Effects of mating experience on subsequent reproductive behaviour and sperm release in the spider Nesticodes rufipes Animal Behaviour 76 1327 1334.
168 Mller, A. P. 1991. Sperm competition, sperm depletio n, paternal care, and relative testis size in birds. American Naturalist 137 882 906. Moore, C. W. 1977. Life cycle, ha bitat and variation in selected web parameters in spider, Nephila clavipes Koch (Araneidae). American Midland Naturalist 98 95 108. Mousseau, T. A. & Roff, D. A. 1989. Adaptation to seasonality in a cricket patterns of phenotypic and genotypic vari ation in body size and diapause expression along a cline in season length. Evolution 43 1483 1496. Moya Larano, J., Vinkovic, D., Allard, C. M. & Foellmer, M. W. 2009. Optimal climbing speed explains the evolution of extreme sexual size dimorphism in spiders. Journal Of Evolutionary Biology 22 954 963. Myers, L. & Christenson, T. 1988. Transition from predat ory juvenile male to mate searching adult in the orb weaving spider Nephila clavipes (Araneae, Araneidae). Journal Of Arachnology 16 254 257. Nakatsuru, K. & Kramer, D. L. 1982. Is Sperm Cheap Limited Male Fertility and Female Choice in the Lemon T etra (Pisces, Characidae). Science 216 753 755. Nessler, S. H., Uhl, G. & Schneider, J. M. 2007a. A non sperm transferring genital trait under sexual selection: an experimental approach. Proceedings of the Royal Society of London Series B Biological S ciences 274 2337 2341. Nessler, S. H., Uhl, G. & Schneider, J. M. 2007b. Genital damage in the orb web spider Argiope bruennichi (Araneae : Araneidae) increases paternity success. Behavioral Ecology 18 174 181. Nessler, S. H., Uhl, G. & Schneider, J. M. 2008. Sexual cannibalism facilitates genital damage in an orb weaving spider, Argiope lobata Behavioral Ecology and Sociobiology 63 355 362. Nevoux, M., Forcada, J., Barbraud, C., Croxall, J. & Weimerskirch, H. 2010. Bet hedging response to en vironmental variability, an intraspecific comparison. Ecology 91 2416 2427. Nosil, P. & Crespi, B. J. 2004. Does gene flow constrain adaptive divergence or vice versa? A test using ecomorphology and sexual isolation in Timema cristinae walking sticks Evolution 58 102 112. Oliver, C. & Cordero, C. 2009. Multiple mating reduces male survivorship but not ejaculate size in the polygamous insect Stenomacra marginella (Heteroptera: Largidae). Evolutionary Ecology 23 417 424.
169 Olsson, M., Madsen, T. & Shine, R. 1997. Is sperm really so cheap? Costs of reproduction in male adders, Vipera berus Proceedings of the Royal Society London B Biological Sciences 264 455 459. Ortlepp, C. & Gosline, J. M. 2008. The scaling of safety factor in spider drag lines. Journal Of Experimental Biology 211 2832 2840. Parker, G. A. 1970. Sperm competition and its evolutionary consequences in the insects. Biological Reviews 45 525 567. Parker, G. A. 1983. Mate quality and mating decisions. In: Mate Choice (Ed. by Bateson, P.), pp. 141 162. Cambridge: Cambridge University Press. Parker, G. A. 1990. Sperm competition games sneaks and extra pair copulations. Proceedings of the Royal Society of London Series B Biological Sciences 242 127 133. Parker, G. A. Ball, M. A., Stockley, P. & Gage, M. J. G. 1997. Sperm competition games: a prospective analysis of risk assessment. Proceedings of the Royal Society of London Series B Biological Sciences 264 1793 1802. Peretti, A. V. & Eberhard, W. G. 2010. Crypti c female choice via sperm dumping favours male copulatory courtship in a spider. Journal of Evolutionary Biology 23 271 281. Pierce, J. D., Obrien, K. K. & Dewsbury, D. A. 1992. No effect of familiarity on the Coolidge effect in prairie voles ( Microt us ochrogaster ). Bulletin of the Psychonomic Society 30 325 328. Pizzari, T., Cornwallis, C. K., Lovlie, H., Jakobsson, S. & Birkhead, T. R. 2003. Sophisticated sperm allocation in male fowl. Nature 426 70 74. Poniatowski, D. & Fartmann, T. 2009. E xperimental evidence for density determined wing dimorphism in two bush crickets (Ensifera: Tettigoniidae). European Journal of Entomology 106 599 605. Prenter, J., Elwood, R. W. & Montgomery, W. I. 1998. No association between sexual size dimorphism and life histories in spiders. Proceedings of the Royal Society of London Series B Biological Sciences 265 57 62. Prenter, J., Elwood, R. W. & Montgomery, W. I. 1999. Sexual size dimorphism and reproductive investment by female spiders: A comparative analysis. Evolution 53 1987 1994.
170 Preston, B. T., Stevenson, I. R., Pemberton, J. M. & Wilson, K. 2001. Dominant rams lose out by sperm depletion A waning success in siring counters a ram's high score in competition for ewes. Nature 409 681 682. Punzalan, D., Rodd, F. H. & Hughes, K. A. 2005. Perceptual processes and the maintenance of polymorphism through frequency dependent predation. Evolutionary Ecology 19 303 320. Radhakrishnan, P., Perez Staples, D., Weldon, C. W. & Taylor, P. W. 2009 Multiple mating and sperm depletion in male Queensland fruit flies: effects on female remating behaviour. Animal Behaviour 78 839 846. Ramos, M., Irschick, D. J. & Christenson, T. E. 2004. Overcoming an evolutionary conflict: removal of a reproduct ive organ greatly increases locomotor performance. Proceedings of the National Academy of Science 101 4883 4887. Raymond, M. & Rousset, F. 1995. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity 86 248 249. Reynolds, E. S. 1963. The use of lead citrate at high pH as electron opaque stain in electron microscopy. Journal of Cell Biology 17 208 212. Richardson, H. H., Jarett, L. & Finke, E. H. 1960. Embedding in Epoxy Resins for ultrathin sect ioning in electron microscopy. Stain Technology 35 313 323. Rittschof, C. C. 2010. Male density affects large male advantage in the golden silk spider, Nephila clavipes Behavioral Ecology 21 979 985. Rittschof, C. C. & Ruggles, K. V. 2010. The comp lexity of site quality: multiple factors affect web tenure in an orb web spider. Animal Behaviour 79 1147 1155. Robertson, J. M. & Rosenblum, E. B. 2009. Rapid divergence of social signal coloration across the White Sands ecotone for three lizard spec ies under strong natural selection. Biological Journal of the Linnean Society 98 243 255. Roff, D. 1992. The Evolution of Life Histories: Theory and Analysis New York: Routledge, Chapman, & Hall, Inc. Rypstra, A. L. 1981. The effect of kleptoparasit ism on prey consumption and web relocation in a Peruvian population of the spider Nephila clavipes Oikos 37 179 182. Schneider, J., Herberstein, M. E., Bruce, M. J., Kasumovic, M. M., Thomas, M. L. & Elgar, M. A. 2008. Male copulation frequency, sper m competition and genital damage in the golden orb web spider ( Nephila plumipes ). Australian Journal of Zoology 56 233 238.
171 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. 2005. The combined effects of pre and post insemination sexual selection on extreme variation in male body size. Evolutionary Ecology 19 419 433. Schneider, J. M., Fromhage, L. & Uhl, G. 2005a. Copulation patterns in the golden orb web spider Nephila madagascariensis Journal Of Ethology 23 51 55. Schneider, J. M., Fromhage, L. & Uhl, G. 2005b. Extremely short copulations do not affect ha tching success in Argiope bruennichi (Araneae, Araneidae). Journal of Arachnology 33 663 669. Schneider, J. M., Gilberg, S., Fromhage, L. & Uhl, G. 2006. Sexual conflict over copulation duration in a cannibalistic spider. Animal Behaviour 71 781 788 Schneider, J. M., Herberstein, M. E., De Crespigny, F. C., Ramamurthy, S. & Elgar, M. A. 2000. Sperm competition and small size advantage for males of the golden orb web spider Nephila edulis Journal Of Evolutionary Biology 13 939 946. Schneider, J. M. & Lesmono, K. 2009. Courtship raises male fertilization success through post mating sexual selection in a spider. Proceedings Of The Royal Society B Biological Sciences 276 3105 3111. Schneider, J. M., Thomas, M. L. & Elgar, M. A. 2001. Ectomise d conductors in the golden orb web spider, Nephila plumipes (Araneoidea): a male adaption to sexual conflict? Behavioral Ecology and Sociobiology 49 410 415. Schubert, M., Schradin, C., Rodel, H. G., Pillay, N. & Ribble, D. O. 2009. Male mate guardin g in a socially monogamous mammal, the round eared sengi: on costs and trade offs. Behavioral Ecology And Sociobiology 64 257 264. Seger, J. & Brockmann, H. J. 1987. What is bet hedging? In: Oxford Surveys in Evolutionary Biology (Ed. by Harvey, P. H. & Partridge, L.), pp. 182 211. Oxford: Oxford University Press. Segoli, M., Harari, A. R. & Lubin, Y. 2006. Limited mating opportunities and male monogamy: a field study of white widow spiders, Latrodectus pallidus (Theridiidae). Animal Behaviour 72 635 642. Sih, A., Bell, A. M., Johnson, J. C. & Ziemba, R. E. 2004. Behavioral syndromes: An integrative overview. Quarterly Review of Biology 79 241 277.
172 Simmons, L. W., Craig, M., Llorens, T., Schinzig, M. & Hosken, D. 1993. Bush Cricket spermatop hores vary in accord with sperm competition and parental investment theory. Proceedings of the Royal Society of London Series B Biological Sciences 251 183 186. Smith, C., Pateman Jones, C., Zieba, G., Przybylski, M. & Reichard, M. 2009. Sperm deplet ion as a consequence of increased sperm competition risk in the European bitterling, Rhodeus amarus Animal Behaviour 77 1227 1233. Smith, C. C. 2007. Independent effects of male and female density on sexual harassment, female fitness, and male compet ition for mates in the western mosquitofish Gambusia affinis Behavioral Ecology And Sociobiology 61 1349 1358. Smith, H. M. 2009. The costs of moving for a diurnally cryptic araneid spider. Journal Of Arachnology 37 84 91. Snow, L. S. E., Abdel M esih, A. & Andrade, M. C. B. 2006. Broken copulatory organs are low cost adaptations to sperm competition in Redback spiders. Ethology 112 1 11. Snow, L. S. E. & Andrade, M. C. B. 2004. Pattern of sperm transfer in redback spiders: implications for s perm competition and male sacrifice. Behavioral Ecology 15 785 792. Spurr, A. R. 1969. A low viscosity epoxy resin embedding medium for electron microscopy. Journal of Ultrastructure Research 26 31 43. Steiger, S., Franz, R., Eggert, A. K. & Muller J. K. 2008. The Coolidge effect, individual recognition and selection for distinctive cuticular signatures in a burying beetle. Proceedings Of The Royal Society B Biological Sciences 275 1831 1838. Stockley, P., Searle, J. B., Macdonald, D. W. & Jo nes, C. S. 1996. Correlates of reproductive success within alternative mating tactics of the common shrew. Behavioral Ecology 7 334 340. Stoltz, J. A. & Andrade, M. C. B. males to mate with reduced effort. Proceedings of the Royal Society London B Biological Sciences 277 585 592. Sun, J. T. & Tower, J. 1999. FLP recombinase mediated induction of Cu/Zn superoxide dismutase transgene expression can extend the life span of adult Drosophila melan ogaster flies. Molecular and Cellular Biology 19 216 228.
173 Svensson, E. I. & Sinervo, B. 2004. Spatial scale and temporal component of selection in side blotched lizards. American Naturalist 163 726 734. Taborsky, M., Oliveira, R. F. & Brockmann, H. J. 2008. The evolution of alternative reproductive tactics: concepts and questions. In: Alternative Reproductive Tactics: An Integrative Approach (Ed. by Oliveira, R. F., Taborsky, M., Brockmann, H.J.), pp. 1 21. Cambridge: Cambridge University Press. Teng, Z. Q. & Zhang, Q. W. 2009. Determinants of male ejaculate investment in the cotton bollworm Helicoverpa armigera : mating history, female body size and male age. Physiological Entomology 34 338 344. Travisano, M., Mongold, J. A., Bennett, A. F. & Lenski, R. E. 1995. Experimental tests of the roles of adaptation, chance, and history in evolution. Science 267 87 90. Uhl, G., Nessler, S. & Schneider, J. 2010a. Securing paternity in spiders? A review on occurrence and effects of mating plugs and male genital mutilation. Genetica 138 75 104. Uhl, G., Nessler, S. H. & Schneider, J. M. 2010b. Mating plugs and genital mutilation in spiders (Araneae). Genetica 138 75 104. Uhl, G., Schmitt, S., Schafer, M. A. & Blanckenhorn, W. 2004. Food and s ex specific growth strategies in a spider. Evolutionary Ecology Research 6 523 540. Uhl, G. & Vollrath, F. 1998. Little evidence for size selective sexual cannibalism in two species of Nephila (Araneae). Zoology Analysis Of Complex Systems 101 101 106. Uhl, G. & Vollrath, F. 2000. Extreme body size variability in the golden silk spider ( Nephila edulis ) does not extend to genitalia. Journal Of Zoology 251 7 14. Van Voorhies, W. A. 1992. Production of sperm reduces nematode lifespan. Nature 360 456 458. Vanheezik, Y. M., Seddon, P. J., Cooper, J. & Plos, A. L. 1994. Interrelationships between breeding frequency, timing and outcome in king penguins Aptenodytes patagonicus are king penguins biennial breeders? Ibis 136 279 284. Vanvalen, L 1965. Selection in natural populations. 3. Measurement and estimation. Evolution 19 514 528. Vincent, S. E. & Lailvaux, S. P. 2006. Female morphology, web design, and the potential for multiple mating in Nephila clavipes : do fat bottomed girls make the spider world go round? Biological Journal Of The Linnean Society 87 95 102.
174 Vollrath, F. 1980. Male body size and fitness in the web building spider Nephila clavipes Zeitschrift Fur Tierpsychologie Journal Of Comparative Ethology 53 61 78. Vo llrath, F. 1983. Relative and absolute growth in Nephila clavipes (Arachnida, Araneae, Argiopidae). Verhandlungen des Naturwissenschaftlichen Vereins in Hamburg 26 277 290. Vollrath, F. 1985. Web spider dilemma a risky move or site dependent growth. Oecologia 68 69 72. Vollrath, F. 1998. Dwarf males. Trends In Ecology & Evolution 13 159 163. Vollrath, F. & Houston, A. 1986. Previous experience and site tenacity in the orb spider Nephila (Araneae, Araneidae). Oecologia 70 305 308. Vollrath, F. & Parker, G. A. 1992. Sexual dimorphism and distorted sex ratios in spiders. Nature 360 156 159. Wcislo, W. T. 1989. Behavioral environments and evolutionary change. Annual Review of Ecology and Systematics 20 137 169. Wedell, N., Gage, M. J. G & Parker, G. A. 2002. Sperm competition, male prudence and sperm limited females. Trends in Ecology & Evolution 17 313 320. Welke, K. & Schneider, J. M. 2009. Inbreeding avoidance through cryptic female choice in the cannibalistic orb web spider Arg iope lobata Behavioral Ecology 20 1056 1062. West Eberhard 2003. Developmental Plasticity and Evolution Oxford: Oxford University Press. Whitehouse, M. E. A. 1997. Experience influences male male contests in the spider Argyrodes antipodiana (Ther idiidae:Araneae). Animal Behaviour 53 913 923. Wiklund, C. & Friberg, M. 2009. The evolutionary ecology of generalization: among year variation in host plant use and offspring survival in a butterfly. Ecology 90 3406 3417. Wilson, J. R., Kuehn, R. E. & Beach, F. A. 1963. Modification in sexual behavior of male rats produced by changing stimulus female. Journal of Comparative and Physiological Psychology 56 636 &.
175 Wong, B. B. M., Candolin, U. & Lindstrom, K. 2007. Environmental deterioration co mpromises socially enforced signals of male quality in three spined sticklebacks. American Naturalist 170 184 189. Zamudio, K. R. & Sinervo, E. 2000. Polygyny, mate guarding, and posthumous fertilization as alternative male mating strategies. Proceedi ngs of the National Academy of Sciences of the United States of America 97 14427 14432. Zeh, D. W. 1987. Aggression, density, and sexual dimorphism in Chernetid pseudoscorpions (Arachnida, Pseudoscorpionida). Evolution 41 1072 1087. Zeh, J. A. 1997 Polyandry and enhanced reproductive success in the harlequin beetle riding pseudoscorpion. Behavioral Ecology and Sociobiology 40 111 118. Zeh, J. A. & Zeh, D. W. 2003. Toward a new sexual selection paradigm: Polyandry, conflict and incompatibility (Invited article). Ethology 109 929 950.
176 BIOGRAPHICAL SKETCH Clare was born and raised in Morehead City, a small town on the coast of North Carolina. She attended a local Catholic School before transitioning to public school for junior high and high s chool. Clare spent a lot of time assisting her father, who is a marine b iologist, and later his graduate students, in fieldwork. This mostly involved finding and collecting local marine animals including crabs, fish, and mollusks, and leading student tours of the local estuary at night. Clare entered Cornell University in the fall of 2002 as a biology major. As her first research experience during college, Clare spent a summer working on she was mentored by Durrell Kappan, a post doctoral fe llow, who encouraged her to consider research science as a career goal. Unsure of her field of interest, Clare returned to school that fall where she took two formative classes, Animal Behavior and Spider Biology. The following semester, Clare began workin g on social huntsman spiders with Linda Rayor at Cornell. That summer, she switched to working on Clare continued her research on bees through her senior year. Following thes e experiences, Clare decided to pursue a degree in z oology, applying to work in Jane work, she plans to explore the mechanisms of behavioral plasticity and experienced Illinois.