1 ON THE VARIATION IN BODY SIZE AND MALE REPRODUCTIVE TACTICS OF HORSESHOE CRABS, LIMULUS POLYPHEMUS By MATTHEW DENMAN SMITH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Matthew Denman Smith
3 To my grandparents, James and Lois Denman, who introduced me to the s ea
4 ACKNOWLEDGMENTS I am forever grateful to my advisor, Jane Brockmann, for all of her wisdom, patience, and mentorship throughout my dissertation. I deeply appreciate the guidance of my committee : Daniel Hahn, Doug Levey, Steve Phelps, and Colette St. Mary I thank my paren ts, Donald and Margaret Agnoli for their ever present support I am in debt to the Nexus Biology Group and all of my friends who made this all so wonderful. Adrian Stier conducted the data extraction for the estimates of food abundance and Franois Michonneau conducted the random mating simulation Tina Moore at the North Carolina Division of Marine Fisheries, and Mark Maddox at the South Carolina Office of Fisheries Management provided body size data I wish to thank Melissa Clark, Jessica Diller Hunter Schrank, Kim Barbeitos, and Lindsay Keegan who provided field assistance; Jason Curtis and the U F Mass Spectrometry Lab who conducted the isotope measurements; and Ben Olivar, Pete Ryschkewitsch, and Mike Gunter who prov ided support with lab maint enance. The research was carried out under a special use permit from the Cedar Keys National Wildlife Refuge. I thank the Univ. of GA Marine Institute at Sapelo Island, GA; the Georgia Dept. Skidaway Island State Park; and the Lower Suwannee National Wildlife Refuge for their support of this proj ect. I also thank the Univ. of Florida Marine Lab at Seahorse Key, its Director, Harvey Lillywhite, and station managers Al Dinsmore and Bronko Gukanovich. This research was supported by a Univ. of Florida Grinter Fellowship; the Sigma Xi Grants in aid of research Lerner Gray Grants for Marine Research; and the National Science Foundation (grant # IOB 0641750 to Jane Brockmann).
5 TABLE OF CON TENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ 4 LIST OF TABLES ................................ ................................ ................................ ........... 8 LIST OF FIGURES ................................ ................................ ................................ ........ 9 LIST OF ABBREVIATIONS ................................ ................................ .......................... 10 ABSTRACT ................................ ................................ ................................ .................. 11 CHAPTER 1 PHENOTYPIC VARIATION IN MORPHOLOGY AND BEHAVIOR ........................ 13 Introduction ................................ ................................ ................................ ............ 13 Common Patterns of Phenotypic Variation in Nature ................................ ............. 14 Phenotypic Variation in Horseshoe Cra bs ................................ ............................. 18 Chapter 2: Body Size ................................ ................................ ...................... 20 Chapter 3: Sexual Size Dimorphism ................................ ................................ 21 Chapter 4: Reproductive Tactics ................................ ................................ ..... 21 Implications ................................ ................................ ................................ ........... 22 2 AN UNUSUAL PATTERN OF BODY SIZE DISTRIBUTION AMONG POPULATIONS: A LIFE HISTORY PERSPECTIVE ................................ .............. 23 Introduction ................................ ................................ ................................ ............ 23 Predicted Optimal Life History Strategies for Ecogeographic Patterns ............ 26 Factors Affecting Growth Rate and Survivorship ................................ ............. 27 Study Species ................................ ................................ ................................ 29 Methods ................................ ................................ ................................ ................ 33 Pattern of Body Size Distribution ................................ ................................ ..... 33 Ecological and Environmental Variables ................................ ......................... 34 Analysis ................................ ................................ ................................ ........... 36 Results ................................ ................................ ................................ .................. 36 Pattern of Body Size Distribution ................................ ................................ ..... 36 Ecological and Environmental Variables ................................ ......................... 37 Discussion ................................ ................................ ................................ ............. 38 Proximate Hypotheses to Explain Ecogeographic Clines ................................ 43 .......................... 44 Hypotheses based on physical constraints ................................ ............... 45 Hypotheses Regarding Food Availability to Explain Ecogeographic Clines ..... 48 Dome Shaped Distributions ................................ ................................ ............ 49 Conclusions ................................ ................................ ................................ .... 51
6 3 AN EVALUATION OF THE ULTIMATE FACTORS UNDERLYING THE EVOLUTION AND MAINTENANCE OF SEXUAL SIZE DIMORPHISM IN HORSESHOE CRABS ................................ ................................ .......................... 65 Introduction ................................ ................................ ................................ ............ 65 Study Species ................................ ................................ ................................ 66 Adaptive Hypotheses That Explain SSD ................................ ......................... 68 Natural selection for divergence in size ................................ ..................... 68 Sexual selection for small male size ................................ ......................... 69 Sexual selection for large female size ................................ ....................... 72 Methods ................................ ................................ ................................ ................ 73 Competitive Displacement Hypothesis ................................ ............................ 73 Agility Advantage Hypothesis ................................ ................................ .......... 74 Loading Constraints Hypothesis ................................ ................................ ...... 76 Nesting Competition Hypothesis ................................ ................................ ..... 77 Fecundity Advantage Hypothesis ................................ ................................ .... 78 Results ................................ ................................ ................................ .................. 80 Competitive Displacement Hypothesis ................................ ............................ 80 Agility Advantage Hypotheses ................................ ................................ ......... 8 0 Loading Constraints Hypothesis ................................ ................................ ...... 81 Nesting Competition Hypothesis ................................ ................................ ..... 82 Fecundity Advantage Hypothesis ................................ ................................ .... 82 Discussion ................................ ................................ ................................ ............. 83 4 COSTS OF ALTERNATIVE MATING TACTICS TO FEEDING IN MALE HORSESHOE CRABS ................................ ................................ ........................ 103 Introduction ................................ ................................ ................................ .......... 103 Methods ................................ ................................ ................................ .............. 106 Measuring Waste Production ................................ ................................ ........ 107 Measuring Food Transit Time ................................ ................................ ........ 109 Stable Isotope Analysis of Feces ................................ ................................ .. 110 Effect of Starvation on Fecal Stable Isotope Value ................................ ........ 112 Gut Contents Analysis ................................ ................................ ................... 113 Results ................................ ................................ ................................ ................ 115 Waste Production ................................ ................................ .......................... 115 Transit Time ................................ ................................ ................................ .. 115 Stable Isotope Analysis of Feces ................................ ................................ .. 116 Experimental Starvation ................................ ................................ ................ 116 Gut Contents ................................ ................................ ................................ 116 Discussion ................................ ................................ ................................ ........... 116 5 SUMMARY AND CONCLUSIONS ON PHENOTYPIC VARIATION IN BODY SIZE AND REPRODUCTIVE BEHAVIOR ................................ ........................... 127 Introduction ................................ ................................ ................................ .......... 127 Chapter 2 ................................ ................................ ................................ ...... 127
7 Chapter 3 ................................ ................................ ................................ ...... 128 Chapter 4 ................................ ................................ ................................ ...... 128 Conclusions ................................ ................................ ................................ ......... 129 APPENDIX A REFERENCES FOR TABLE 2 1. ................................ ................................ ........ 131 B LIST OF TAXA USED IN ESTIMATING FOOD ABUNDANCE ............................ 132 C REFERENCES FOR TABLE 3 2 ................................ ................................ ......... 133 LIST OF REFERENCES ................................ ................................ ............................ 134 BIOGRAPHICAL SKETCH ................................ ................................ ......................... 156
8 LIST OF TABLES Table page 2 1 Locations, latitude, and body size for the size distribution pattern of horseshoe crabs in North America ................................ ................................ ... 54 2 2 Values for environmental and ecological variables hypothesized to affect the growth and survival of horseshoe crabs ................................ ........................... 56 2 3 Average values for clusters of body size of horseshoe crabs found across locations in Nort h America ................................ ................................ ................ 58 2 4 Results of stepwise multiple regression model for environmental and ecological variables influencing body size ................................ ........................ 59 3 1 Hypotheses and predictions for the evolution of sexual size dimorphism in horseshoe crabs. ................................ ................................ ............................... 90 3 2 Data for the comparison of male size and operational sex ratio, and female size and nesting density across North America ................................ ................. 91 3 3 The number of 1F satellite males that obtained either the under or over position, with respect to their body size and condition ................................ ...... 92 3 4 The number of satellite males that obtained the desired position near a female with respect to their body size and condition ................................ ......... 93 3 5 Comparison of the estimated effects of body size on fecundity ........................ 94 4 1 Summary of the predictions and assumptions of t he reduced feeding hypothesis ................................ ................................ ................................ ...... 121 4 2 Values of various measures for horseshoe crabs collected from Seahorse Key, FL ................................ ................................ ................................ ............ 122 A 1 References used for T able 2 1. ................................ ................................ ....... 131 A 2 List of taxa used in estimating food abundance ................................ .............. 132 A 3 References used for Table 3 2 ................................ ................................ ........ 133
9 LIST OF FIGURES Figure page 2 1 Hypothetical scenarios that could result in the dome shaped size distribution seen in horseshoe crabs ................................ ................................ ................... 60 2 2 Scaling relationship between carapace width and interocular distance ............. 61 2 3 Relationship between body size and latitude; and between sexual size dimorphism and latitude ................................ ................................ ................... 62 2 4 Plots of environmental variables and body size across North America .............. 63 2 5 Contour plots for body size across North America ................................ ............. 64 3 1 Illustrations of horseshoe crab mating groups from Seahorse, Key, FL ............. 95 3 2 Stable isotope values for claw chitin and feces of wild caught crabs ................. 96 3 3 Effects of operational sex ratio and density on body size ................................ .. 97 3 4 Relationship between female and male size for observed mating pairs ............. 98 3 5 Probability density for simulated female to male size ratios ............................... 99 3 6 Allometries of mass and carapace width ................................ ......................... 100 3 7 Relationship between female nesting density and size across tides ................ 101 3 8 Relationship between size and fecund ity ................................ ......................... 102 4 1 Fecal mass and food transit time of wild caught horseshoe crabs from Seahorse Key, FL ................................ ................................ ............................ 123 4 2 Value s of 15N stable isotopes of feces produced by wild caught horseshoe crabs ................................ ................................ ................................ ............... 124 4 3 Pre and post treatment values of 15N from satellite males that were either starved or fed ................................ ................................ ................................ 125 4 4 G ut fullness and mass of seagrass found in attached and satellite male s ....... 126
10 LIST OF ABBREVIATION S 13 C Ratio (R) of 13 C to 12 calculated by the standard formula: [(R sample / R standard ) 1] x 10 3 15 N Ratio (R) of 15 N to 14 N isotopes, measured in permil calculated by the standard formula: [(R sample / R standard ) 1] x 10 3 O SR Operational Sex ratio. The ratio of sexually competing males that are ready to mate to sexually competing females that are ready to mate. This value reflects the inte ns ity of sexual competition in a species. S SD Sexual Size Dimorphism. The difference in body size between males and females of a given species (F / M in horseshoe crabs) P PT Parts Per Thousand. A commonly used measure of salinity concentration. P PS Practical Salinity Scale. A commonly used, unit less measure of salinity concentration. Q 10 M ORTALITY Thermal S ensitivity of Mortality Describes how mortality rate is influenced by temperature. Q 10 PRODUCTION Thermal Sensitivity of P roduction. Describes how p roduction rate ( the difference between rates of energy ass imilation and metabolic rates) is influenced by temperature.
11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment o f the Requirements for the Degree of Doctor of Philosophy ON THE VARIATION IN BODY SIZE AND MALE REPRODUCTIVE TACTICS OF HORSESHOE CRABS, LIMULUS POLYPHEMUS By Matthew Denman Smith August 2012 Chair: H. Jane Brockmann Major: Zoology Understanding the ultimate and proximate factors underlying patterns of phenotypic variation is a key goal in the study of behavioral and evolutionary ecology. Organisms commonly show variation in body size and reproductive behavior. Such variation is in teresting because it can represent adaptations to the environment. In my dissertation I used horseshoe crabs, Limulus polyphemus to explore the ultimate and proximate factors underlying variation in: 1) body size across populations, 2) size between males and females within a population, and 3) reproductive behavior among male within a population. In Chapter 2, I examined size at maturity which is one of the most fu ndamental traits of an organism. My investigation into large scale patterns of body size d istribution discovered that horseshoe crabs are one of just a few species to show a dome shaped distribution in body size. This pattern appears to be due to how temperature season length, salinity, oxygen levels, and food abundance a ffects juvenile morta lity and growth rates that control body size What is the ultimate explanation for why females and males same species rarely have the same adult body size? Sexual size dimorphism (SSD) is thought to be a result
12 of disruptive selection on body size between the sexes, due to reproductive roles. SSD is likely the product of a suite of conflicting pressures on each sex. In Chapter 3, I evaluated fived hypotheses to explain the evolution of SSD, and conclude that fecundity selection favoring late maturity and large female size, along with protandry favoring earlier maturity and small male size, are the likely selection pressure s responsible for SSD in this species. Horseshoe crab males exhibit two condition dependent, alternative mating tactics: males in bett er condition arrive on spawning beaches attached to females, while males in poorer condition join spawning pairs as satellites In Chapter 4, I investigated a cost to the attached male tactic that has not been considered previously: a restricted ability to feed. My results indicate that a period of nutritional stress caused by reduced food consumption is a cost of the attached tactic; and may explain why the alternative reproductive tactics take the form they do in this system.
13 CHAPTER 1 PHENOTYPIC VARIATION IN MORPHOLOGY AND BE HAVIOR Introduction Understanding ultimate and proximate factors underlying patterns of phenotypic variation is a key goal in the study of behavioral and evolutionary ecology. In a break from the common typological thinking at the time, Darwin recognized the fundamental impo rtance of trait variation ( Darwin 1859 ) In formin g his hypothesis of evolution by natural selection, Darwin first inferred that because most populations were relatively stable, but capable of exponential growth ( Malthus 1798 ) there must exist a struggle for existence. Because t rait variation among individuals in a population is common a second inference was that some variants survive or reproduc e better than others. This differential survival or reproduction of certain phenotypes constitutes natural selection ( Mayr 1991 ) We now understand that evolution by natural selection is inevitable if there is variation in a trait, if it is heritable, and if this variation confers a selective advantage ( Maynard Smith 1958 ) Thus, phenotypic variation forms the raw material for natural selection, and is at the heart of adaptation. Phenotypic variation in nature is ubiquitous, and can be classified based on trait category (e.g., morphological, behavioral); the trait itself (e.g., body size, foraging behavior); the type of variation that occurs (e.g., continuous, discrete) ; and where that variation occurs (e.g., among populations, between sexes within a sex ). Variation in continuous traits such as body size or coloration can be described with a frequency distribution. The mean of a continuous trait can be shifted higher or lower by directional selection, or the variance can be altered due to stabilizing selection if there are fitness advantages associated with those shifts ( Brockmann 2001 ) In addition, discrete
14 variation can occur if the intermediate phenotype is selected against by disruptive selection. Understanding the patterns of phenotypic variation and the causes and consequences of that variation has been a major theme in evolutionary research, and comprises the foundation for my d issertation. Common Patterns of Phenotypic Variation in Nature D ifferences between populations are frequently seen for morphological traits such as body size and coloration. For example, variation in adult size among populations often follows patterns acr oss large scale gradients such as altitude, latitude, or temperature. One pattern, termed is that adult body size increases with increasing latitude ( Bergmann 1847 Park 1949 Atkinson and Sibly 1997 Mousseau 1997 Blackburn et al. 1999 ) and occurs in most species of birds and mammals ( Ashton et al. 2000 Meiri and Dayan 2003 ) Another pattern, termed is the opposite pattern ( Park 1949 Mousseau 1997 ) and occurs in squamates ( Ashton and Feldman 2003 ), Urodules ( Olalla Tarraga and Rodriguez 2007 ) and many invertebrates ( Mousseau 199 7 Blanckenhorn and Demont 2004 ) But why is there variation in size among populations? In addition to simply documenting phenotypic variation, evolutionary and behavioral ecologists focus on tr ying to understand the ultimate evolutionary explanations for such differences and the proximate mechanisms by which that variation arises For example, differences in body size can reflect optimal life history strategies based on season length ( e.g., the decreasing body size of water striders, Aquarius remigis, in colder latitudes; Blanckenhorn and Fairbairn 1995 ) ; or based on food availability, ( e.g., the increase in body size of many Cervi dae in locations with high net
15 primary productivity; Huston and Wolverton 2011 ) In Chapter 2, I document the pattern of size variation in horseshoe crabs, and evaluate explanations for that pattern. Variation in adult body size between the sexes, termed sexual size dimorphism, is a regular pattern among a wide range of species. Sexual size dimorphism in endotherms is often male biased (i.e., males are larger), whereas in ectotherms it tends to be fem ale biased (Fairbairn 2007). Coloration is a nother trait that commonly varies among populations and between sexes For example, in guppies males have brightly colored spots on their bodies, while females tend to be more uniformly grey in color ; and among populations, the color of males, and the intensity of coloration is variable ( Endler 1980 Endler 1991 ) Male biased sexual si ze dimorphism in endotherms often evolves due to differential reproductive success via male mating competition ( e.g., in fallow deer, Dama dama ; McElligott et al. 2001 ) whereas fecundity advantages associated with larger size are often selected for in fema le biased ectotherms ( e.g., spiders; Head 1995 ) In an ultimate sense, differences in the male coloration of gu ppies functions to reduce predation ( Endler 1980 ) Mechanistically, this occurs because guppy coloration is matched to the visual sensory system of common predators located in those areas Fo r example, the major predator in some populations is a cichlid that lack s a blue cone and can perceive only green through red wavelengths. In these populations the predominant color of guppies is blue, thus allowing them to be more camouflaged from predators ( Endler 1991 ) In Chapter 3, I examine five hypotheses to explain the evolution and maintenance of sexual size dimorphism in horseshoe crabs. Variation among populations also can be found in behavioral phenotypes. For example, the degree of mate guarding behavior varies between populations in
16 soapberry bugs, Jadera haematoloma In Florida, copulation lasts for 10 hours on average, whereas in Oklahoma the average copulation lasts 26 hours ( Carroll 1988 1991 ) Variation in duration of mate guarding in soapberry bugs is an adaptation to different levels of sperm competition in different populations, males mate guard for longer periods in populations where the operational sex ratio (e.g., the ratio of mating males to females) is greater ( Carroll 1988 1991 ) What is more difficult to explain is discrete variation (e.g., a bimodal distribution ) of a trait within one sex in a single population ( Brockmann 2001 ) In general, one optimal value for morphological traits or one tactic for behavioral traits is expected because if one form or tactic had even slight ly higher fitn ess, then that genotype should spread through the population ( Brockmann 2001 ) Despite this, discrete variation can evolve through disruptive selection, and can be found in morphological traits, such as horns in major and minor morphs of the Onthophagus beetles ( Emlen 1994 ) Discrete variation is also found in behavioral phenotypes, such as the age polyethism in roles of worker honeybees, Apis mellifera The role of newly emerged bees progresses from cleaning comb ce lls, to building wax cells and feeding larvae (nurse bees), and finally to foraging for nectar and pollen ( Lindauer 1961 ) M orphology and behavior are often intertwined. For example, in Onthophagus beetles d iscrete differences in horn length are accompanied by discrete differences in reproductive tactics. Males with large horns guard entrances to burrows that contain females, and compete for mates with other males for access to females, whereas short ( Hunt and Simmons 2002 ) The behavioral changes in the roles of worker bees are accompanied by shifts in
17 morphology needed for those roles. For example, younger nurse bees have wax glands, while older forager bees do not; and mushroom bodies of the b rain (that aid in memory and learning) are large in older foragers, but small in young nurse bees ( Withers et al. 1993 ) The age polyethism in honey bees functions t o maintain the number of nurse and forager bees at an optimal ratio. Proximately, this is regulated by social interactions between foragers and nurse bees. When foragers exchange food with nurse bees, they also transfer a compound called e thyl o elate, wh ich functions to suppress juvenile hormone levels in nurse bees ( Leoncini et al. 2004 ) When the ratio of foragers to nurse bees in a colony declines, the number of interactions declines and the levels of juvenile hormone in nurses rise. High levels of juvenile hormo ne cause changes in gene expression associated with the ontogenetic progression from nurse bee to forager bee ( Whitfield et al. 2003 ) Thus, the colony ma intains the ratio of nurse and forager bees at a level that confers the highest fitness to the queen and her workers. Such suites of correlated morphological and behavioral traits are common when phenotypes show discrete variation. However, bimodal variat ion in behavior is not always associated with morphological changes. This is particularly common in alternative tactics that are based on condition. For example, in some frog species, a subset of males acquire s females by actively calling, while others a dopt a satellite tactic by staking out a position near calling males and intercept ing the phonotactic females ( Wells 1977b ) In the alternative mating tactics of frogs, calling and maintaining a territory are costly. Thus, by adopting a satellite tactic, males in poor condition can gain weight by not expending energy on defending a territory ( Wells 1977b Robertson 1986a McCauley et al. 2000 ) while still
18 having some benefit of mating if they intercept a female. These alternative tactics do not confer the same fitness as the calling tactic, b ut the decision to switch between mating tactics is an evolved strategy based on condition that maximize s fitness of an individual ( Brockmann 2001 ) In Chapter 4, I examine a potential cost of the alternative mating tactics in male horseshoe crabs. Specifically, I test the hypothesis that the attached tactic comes at a cost of a reduction in feeding. Phenotypic Variation in Horseshoe Crabs North American horseshoe crabs, Limulus polyphemus, are a unique and ancient Mesozoic era ( Fisher 1984 ) But of course they have changed, and they display phenotypic variation in a myriad of traits. Horseshoe crabs are divided into genetically distinct subpopulations in many parts of their range ( King et al. 2005 ) along the eastern and southern coasts of the US, and along the nor thern and western coasts of the Yucatan Peninsula, Mexico. Horseshoe crabs are external fertilizers, and usually nest on sandy, low energy beaches of bays and barrier islands ( Shuster 1982 ) On the Gulf of Mexico coast of Florida, the breeding season occurs from February to May, and again from late August to October ( Rudloe 1980 Brockmann and Johnson 2011 ) ; more northern populations have a single breeding season in spring ( Smith et al. 2010 ) Within a breeding season, spawning occurs during high tides around the new and full moons which are the high est tides of the month ( Rudloe 1980 Cohen and Brockmann 1983 Barlow et al. 1986 ) The mating system of horseshoe crabs is characterized as in which receptive females are abundant for a brief time during which mating is frequent ( Thornhill and Alcock 1983 Brockmann 1990 ) Spawning is highly synchronized both temporally ( Rudloe 1980 ) and spatially
19 ( Penn and Brockmann 1994 ) and this often results in high nesting densities. Ind ividual females generally nest for only one week in a breeding season, whereas males continue to mate with different nesting females for many weeks ( Brockmann and Penn 1992 Brockmann and Johnson 2011 ) For many populations this results in a highly male biased operational sex ratio (OSR) and intense competition for females on s pawning beaches ( Brockmann 1990 Botton and Loveland 1992 ) The locations where horseshoe crabs are found vary greatly in a number of environmental conditions (e.g., water temperature, salinity, oxygen levels), and ecological traits ( e.g., population size, population density, operational sex ratio; Brockmann and Smith 2009 ) T hese differences among locations are reflected in consistent phenotypic variation of horseshoe crabs from these locations. For example, the extent to which temperature affects their activity cycles differs among populations. In the spring, animals from n orthern populations begin to exhibit circatidal rhythms associated with spawning at colder temperatures whereas conspecifics in southern populations do not ( Sekiguchi 1988 Wenner and Thompson 2000 Watson et al. 2009 Chabot and Watson 2010 James Pirri 2010 Schaller et al. 2010 Smith et al. 2010 ) Morphologically, horseshoe crabs are highly variable in average body size among populations ( Riska 1981 Botton and Loveland 1992 Brockmann and Smith 2009 Faurby et al. 2011 ) ; individuals in the northern and southern portions of their range are smaller than animals in the center of their range ( Shuster Jr. 1957 Reynolds and Casterlin 1979 Shuster Jr. 1979 Riska 1981 Botton and Loveland 1992 Thompson 1998 Anderson and Shuster Jr. 2003 Brockmann and Smith 2009 Graham et al. 2009 ) Horseshoe crabs also exhibit sexual size dimorphi sm : within a pop ulation
20 females are larger than males, and this size difference is due to an additional molt and year of growth by females compared to males ( Smith et al. 2009a ) Furthermore, m ales in each population display discrete variation in their reproductive behavior. Male horseshoe crabs exhibit two condition dependent, alternative mating tactics ( Brockmann and Penn 1992 Brockmann et al. 2000 Brockmann 2002 ) Males that are younger and in better condition, attach to females at sea, and arrive on spawning beaches in amplexus with a female. M a les that are older and in poorer condition, roam the shoreline not attached to a female. These unpaired males join spawning pairs as satellites and engage in sperm competition with the attached males and other satellite males ( Brockmann and Penn 1992 Brockmann et al. 1994 Brockmann et al. 2000 ) In my dissertation, I seek to understand the proximate and ultimate factors underlying the variation in these morphological and behavioral traits in three chapters. Chapter 2: Body Size I will describe the pattern of body size distribution of horseshoe crabs a mong populations Why do they show the unusual pattern of a dome shaped size distribution rather than follow ing s as the vast majority of othe r organisms do? Why do they vary in their average body size? Mechanistically, how are these differences in size achieved? Many studies on ecogeographic patterns of body size are conducted using data across partial species ranges, or with species that ha ve relatively small ranges. In this chapter, I synthesize literature and gather new data to explore the body size distribution across the entire range of horseshoe crabs, which are a widely ranging species and thus may reveal patterns not seen in other s tudies on organisms with smaller ranges In addition, this research is framed with an infrequently used life history perspective that allows for an understanding of both the
21 ultimate and proximate reasons for the rare pattern of body size distribution see n in horseshoe crabs. I n particular, I examine how variation in environmental and ecological variables affect s growth and survival leading to optimal life history strategies reflected by the ecological and environmental conditions in a given location. Cha pter 3: Sexual Size Dimorphism At a proximate level, males and females are dimorphic in size because females delay sexual maturity and undergo one more year of growth and one additional molt prior to maturation compared to males (Smith 2009) But at an ul timate level, why is sexual size dimorphism maintained in this species? What are the benefits for large size in females and small size in males? Many studies on sexual size dimorphism investigate only one or two hypotheses about selection for larger and smaller individuals, or are focused on one sex only. In this chapter I evaluate five alternative hypotheses for selective pressures acting on both males and females This allows for a better understanding of the suite of selective pressure that can resul t in sexual size dimorphism Chapter 4: Reproductive Tactics What factors underlie the evolution and maintenance of alternative male reproductive tactics? What are the trade offs between tactics? Attached males benefit in terms of overall paternity and l ower probability of stranding on a beach, but are the re also costs of the attached tactic? In this chapter, I use field and lab methods to explore a previously untested hypothesis that attached males suffer a period of reduced feeding ability and nutritio nal stress to which satellites males are not subject. This is the first study to use stable isotopes to seek nutritional stress in a population of wild animals. Better understanding the trade off between tactics will help augment our understanding
22 of the evolution and maintenance of alternative reproductive tactics in this species. Furthermore, this begins to illuminate why the alternative tactics in this species take the form they do. Implications Overall, this research uses horseshoe crabs as a detaile d system to better understand the factors that underlie variation in body size and mating strategies across animal species. Because horseshoe crabs show an unusual dome shaped distribution, they offer a unique opportunity for future research to uncover a unifying mechanism underlying ecogeographic clines in body size. The stable isotope techniques used in Chapter 4 are novel, and will provide other researchers with a means to test for nutritional stress in other systems. This research also adds to our lim ited understanding of horseshoe crab biology in locations at the edge of their range ( Anderson and Shuster 2003 ) Such information is essential to the conservation of horseshoe crabs because it reveals details on the large scale constraints to the abundance and distribution of the species ( Shuster and Sekiguchi 2009 ) This research may also be useful to the management of horsesho e crabs. Fisherman use horseshoe crabs as bait, and they have a preference for taking the largest animals, generally females, but also large males. Understanding why males and females differ in size, and how costs of alternative tactics might influence r eproductive success may also be important to management decisions ( Berkson 2009 Smith et al. 2009b )
23 CHAPTER 2 AN UNUSUAL PATTERN O F BODY SIZE DISTRIBU TION AMONG POPULATIO NS: A LIFE HISTORY PERSPEC TIVE Introduction Adult b ody size is a key factor in history strategy ( Peters 1983 Calder 1984 Chown and Gaston 2010 ) Intraspecific variation in adult body size across populations is common, and researchers have described a number of patterns of variation in size across large scale environmental gradients such as latitude and temperature The two most common of these ecogeographic clines are with increasing latitude ( Bergmann 1847 Park 1949 Atkinson and Sibly 1997 Mousseau 1997 Blackburn et al. 1999 ) and c describes the opposite pattern ( Park 1949 Mousseau 1997 ) The taxonomic patterns associated with clines that follow rules are complex and both patterns can be found across and within taxa to differing degrees ( Park 1949 Gilligan 1991 Mousseau 1997 Ashton et al. 2000 Ashton 2002 Belk and D. 2002 Ashton and Fe ldman 2003 Meiri and Dayan 2003 Morrison and Hero 2003 Angilletta et al. 2004a Blanckenhorn and Demont 2004 Heibo et al. 2005 Blanckenhorn et al. 2006 Measey and Van Dongen 2006 Olalla Tarraga and Rodriguez 2007 Adams and Church 2008 Fisher et al. 2010 ) In contrast, some species show no relationship between size and latitude ( e.g., many species of Lepidoptera; Nylin and Svrd 1991 ) while others show more complex non linear patterns For example, a s aw tooth distribution (in animals that can adjust the number of generations to the growing season; Masaki 1978 Mousseau and Roff 1989 ) a dome shaped (mid latitudinal peak) distribution ( Geist 1987 Elmes et al. 1999 Blanckenhorn
24 and Demont 2004 Laugen et al. 2005 Virgs et al. 2011 ) or a bowl shaped (mid latitudinal trough) distribution ( Kaustuv and Martien 2001 Johansson 2003 ) Clearly no one pattern characterizes how body size changes with latitude Moreover, arguably more interesting questions focus on the processes and mechanisms that shape these patterns Whether intraspecific patterns of ecogeographic clines and their causes are generally applicable across taxa has been doubted because of simil ar taxa showing opposite patterns, differences in overall biology and ecology among organisms (such as growth strategy), and because the majority of species have relatively small geographical ranges across relatively narrow ranges of environmental conditio ns. Hence, they are unlikely to experience selection that would result in intraspecific patterns ( Gaston et al. 2008 ) F urthermore, f ew explanations for size clines attempt to couple ultimate evolutionary pressures with the proximate physiological factors affecting size variation ( Chown and Gaston 2010 ) Understanding these questions is especially important for organisms that show rare patterns (e.g., d ome shaped distribution ), and may help to provide or confirm a unifying explanation for ecogeographic clines. H orseshoe crabs are one of only a few organisms that appea r to show a dome shaped distribution in size. S maller body sizes have been reported to occur to the north and south of some locations ( Shuster 1957 Reynolds and Casterlin 1979 Shuster 1979 Riska 1981 Botton and Loveland 1992 Thompson 1998 Ander son and Shuster 2003 Brockmann and Smith 2009 Graham et al. 2009 ) but the pattern is not entirely clear, nor has it been rigorously c Here I survey the body size distribution of horseshoe crabs across a large latitudinal
25 range that covers a broad spectrum of environmental conditions. I then use a life history perspective to evaluate the potential proximate and ultimate explanations for that pattern. There has been a great deal of debate surrounding ec ogeographic clines. Are these patterns adaptive or are they simply the result of physiological constraints ( Conover and Present 1990 Atkinson 1994 Van Voorhies 1996 Atkinson and Sibly 1997 Mousseau 1997 Partridge and Coyne 1997 Van Voorhies 1997 Angilletta and Dunham 2003 Blanckenhorn and Demont 2004 Stillwell 2010 ) ? P hysiologists have questioned the heritability of body sizes in animals that follow e cogeographic clines and have explained them as plastic responses to constraints a t the cellular level that are imposed by environmental variables such as temperature and oxygen ( von Bertalanffy 1960 Bradford 1990 van der Have and de Jong 1996 Van Voorhies 1996 1997 Wood s 1999 van der Have 2002 ) In contrast others have argued that adaptive explanations are warranted because ecogeographic clin al patterns evolve repeatedly and predictably ( Huey et al. 2000 ) ; because common garden experiments have shown that genetic variation in body size exists within and among populations ( Partridge and Coyne 1997 Huey et al. 2000 Blanckenhorn and Demont 2004 Schutze and Clarke 2008 Angilletta 2009 ) ; and because some studies show that fitness optima are related to body size and temperature ( McCabe and Partridge 1997 Reeve et al. 2000 ) However like all phenotypic traits, these patterns are surely the result of both genetic and environmental influences ( Alcock 20 09 Angilletta 2009 ) Using a life history framework to explore clinal variation in body size allows the investigation of both proximate and ultimate factors ( Angilletta et al. 2004a Chown and Gaston 2010 )
26 When viewed in the context of life history theory, size clines reflect the evolution and maintenance of optimal body size at maturity. Optimal size at maturity is influenced by trade offs between survival, growth, and fecundity ( Stearns 1992 Teriokhin 1999 Roff 2002 Angilletta et al. 2004a ) The benefits of earlier maturation at a smaller size are a higher probability of survival b ecause organisms spend less time as juveniles, and have higher fitness because their offspring are born earlier and start reproducing sooner ( Stearns 1992 ) However, in many organisms fecundity is positively related to body size ( Darwin 1874 Williams 1966 Thornhill and Alcock 1983 Calder 1984 Shine 1988 Stearns 1992 Andersson 1994 Shine 2005 ) ; thus, higher fecundity is one benefit of longer growth and later maturation at a larger size. In addition, delayed maturation may allo w for an increase in offspring quality, and as a result, reduced instantaneous juvenile death rates ( Stearns 1992 ) He nce, patterns of adult size are driven by the relationships between ecological processes (e.g., physiological tolerance, food limitation) and two key life history traits: growth rate and juvenile survivorship ( Angilletta et al. 2004a Chown and Gaston 2010 ) Predicted Optimal Life History Strategies for Ecogeographic Patterns In each of the three following ecogeographic patterns, size at maturity may be an adaptive response to either temperature or season length or to both (respectively) if maturation at a given size is the optimal outcome of the trade off between growth, survival, and fecundity. Optimal life history strategies can be consistent with when there is lower survivorship of juveniles at lower latitudes (e.g., viability selection for earlier maturation at a sma ller size in the south); and when maturity is delayed in higher latitudes via a longer duration of growth or an accelerated growth rate ( e.g., fecundity selection for delayed maturation at a larger size in the north;
27 Angilletta et al. 2004a ) O ptimal life history strategies can be consistent with c onverse high latitudes (e.g., viability selection for earlier maturation at a smaller size in the north); when there is not a prolonged period of growth at high latitudes (e.g., benefits of increased fecundity by growing larger do not outweigh the costs of loss i n future fitness or the decreased probability of survival during delayed growth in the north); and when animals grow faster or longer in lower latitudes resulting in larger sizes in the south ( Angilletta et al. 2004a ) o be driven by different mechanisms (temperature and season length respectively), they may not be mutually exclusive ( Blanckenhorn and Demont 2004 Chown and Gaston 2010 ) Temperature and season length can operate in conjunction and may cancel each other to varying degrees if they interact additively; if they interact multiplicatively, dome shaped c lines could also occur ( Blanckenhorn and Demont 2004 ) Optimal life history strategies theoretica lly can be consistent with dome shaped distributions when: a) survivorship is high across most latitudes, but growth rates decline in high and low latitudes; or b) when juvenile survivorship is low in high and low latitudes, regardless of the effects of gr owth rate (Figure 2 1) Additionally in each of these three, food availability may play a role in influencing growth and survival (positively with high food abundance and vise versa; Wolverton et al 2009 Ho et al. 2010 ) F actors A ffect ing G rowth R ate and S urvivorship Understanding the mechanisms underlying latitudinal clines in body size is essential to studies of geographic variation in size ( Watt et al. 2010 ) Angilletta et al. ( 2004a ) stated that the problem of understanding the evolution and maintenance of size clines is one of understanding the evolution of thermal reaction norms for age and size
28 at maturity. However, in aqua tic and marine environments, the influence of other environmental factors such as salinity and oxygen on growth and survivorship must also be considered ( Gaston et al. 2008 Angilletta 2009 ) T emperature has a dramatic effect on growth In ectothermic animals growth rate typically increases with increasing developmental temperature, while body size increases with decreasing temperature (i.e., the temperature size rule ; Atkinson 1994 Kingsolver and Huey 2008 ) ; but there In ectotherms, temperature itself has no general physiological effect on survivorship ( Angilletta et al. 2004a ) h owever, the interactive effect of multiple environmental stressors (e.g., temperature x oxygen, temperature x salinity) can negatively affect growth and survivorship ( Cooney et al. 1983 Hanazato and Dodson 1995 Folt et al. 1999 Gaston et al. 2008 Angilletta 2009 ) Additionally v ariation in length of the growing season can affect growth ( Chown and Gaston 1999 Blanckenhorn and Demont 2004 Stillwell 2010 ) These environmental factors change system atically with latitude, and thus can be responsible for ecogeographic clines in body size ( Stillwell 2010 ) Additionally, e cological factors such as competition ( Warren et al. 2006 ) and predation ( Gaston et al. 1997 Stoks et al. 2006 ) and food abundance vary with latitude and can influence size (e.g., high qual ity and quantity of food positively affect growth rate, body size, and survival; Rosenzweig 1968 McNab 1971 Blackburn et al. 1999 Davidowitz et al. 2004 Davidowitz and Nijhout 2004 Yom Tov and Geffen 2006 Wolverton et al. 2009 Ho et al. 2010 ) Exactly how these factors constrain growth and survival is significant area of debate and n umerous hypotheses have been proposed ( reviewed in Chown and Gaston 2010; see discussion )
29 Study Species Horseshoe crabs are unique and ancient animals that live in shallow water s along the east coast of the North America from Maine, USA to the Yucatan, Mexico. There is little gene flow between populations, and so this species is divided into genetically distinct subpopulations in many parts of its range ( King et al. 2005 ) Horseshoe crabs are determinant growers (i.e., they stop growing after a terminal molt into adulthood, at which point they are sexually mature) and due to their molting cycles they display a staircase growth pattern over time During their first year, horseshoe crab juveniles m olt five or six times; during their second year they molt three times, and during their third year they molt twice. After the third year, they molt once per year until their terminal molt at 8 10 years old ( Shuster and Sekiguchi 2003 ) They also have a relatively long life cycle: sexual maturity occurs after their terminal molt and life span is approximately 20 years ( Shuster and Sekiguchi 2003 ) Lab experiments and observational studies in the field have shown that horseshoe crabs from different populations vary in their thermal tolerance and in how temperature affects their activity cycles. The thermal death point of individuals from Woods Hole, MA was 41 C, while for animals from Marquesas Keys in southern Florida it was 46.25 C ( Mayer 1914 ) Maximum activ ity occurred at 41 C in animals from Florida and at 38 C. in animals from Massachusetts ( Mayer 1914 ) In laboratory setting s animals from Great Bay, NH express circatidal rhythms from 12 17 C, rarely from 5 11 C, and never below 4 C a t which point activity is greatly attenuated ( Watson et al. 2009 Chabot and Watson 2010 ) Corresponding field data from telemetry studies on the same population s how that animals move very little below 8 C ( Schaller et al. 2010 )
30 When temperature dr op s in the fall, horseshoe crabs move to deeper water, bury in the substrate ( Sekiguchi 1988 Wenner and Thompson 2000 ) and become inactive ( Sekiguchi 19 88 Shuster and Sekiguchi 2009 Schaller et al. 2010 ) When temperature s rise in the spring, horseshoe crabs move from their wintering grounds to their spawning areas. When and at what temperature this activity occurs varies among population s For example, horseshoe crabs remain active in the fall until temperatures drop below 12 14 C in Maine and New Hampshi re ( Schaller 2002 Watson et al. 2009 Schaller et al. 2010 ) ; in Cape Cod, MA they appear to remain active until 15 16 C ( James Pirri 2010 ) ; and in mid Atlantic and more southern locations they are active until 16 20 C ( Sekiguchi 1988 Wenner and Thompson 2000 ) In spring, horseshoe crabs move up estuary out of their deeper water wintering grounds to their shallower spawning areas when water rises to 11 13 C in ME and NH ( Schaller 2002 Schaller et al. 2010 ) ; 12 15 in Cape Cod, MA ( James Pirri 2010 ) ; and at least 15 C i n Dela ware Bay where the first day of spawning is fo u r days earlier for every 1 C rise in mean daily temperatures in May ( Smith et al. 2010 ) Horseshoe crabs have substantial variation in morphology. First, t hey are sexually size dimorphic: within a population females are larger than males because females undergo an additional molt and year of growth compared to males ( Smith et al. 2009a ) T he average body size is highly variable among populations ( Riska 1981 Botton and Loveland 1992 Brockmann and Smith 2009 Faurby et al. 2011 ) and previous authors be larger in the center of their species range ( Shuster 1979 ) However, where the peak in body size occurs is n ot entirely clear. The largest animals have been stated to occur:
31 (a) in the Atlantic states ( Shuster 1957 1979 ) (b) from Maryland to New York ( Shuster 1982 ) (c) in the South Atlantic Bight which occurs from West Palm Beach, FL to Cape Hatteras, NC ( Thompson 1998 ) (d) from Georgia to Long Island, NY ( Anderson and Shuster 2003 ) (e) in South Carolina and North Carolina ( Graham et al. 2009 ) and (f) from Georgia to the southern shores of Cape Cod ( Botton et al. 2010 ) Effects of environmental variables on growth and survival Horseshoe crab embryos larvae and juveniles are robust and able to withstand fluctuating and harsh environ mental conditions ( Jegla and Costlow 1982 Palumbi and Johnson 1982 Laughlin 1983 Botton 1988 Sugita 1988 Ehlinger and Tankersley 2004 Bademan 2009 ) Studies examin ing the effects of temperature and salinity have found, survival is generally higher, and development time and inter molt growth intervals are generally shorter ( Laughlin 1983 Ehlinger and Tankersley 20 04 ) at higher temperature s and salinit ies (within the normal range). Extremes in temperature and salinity are detrimental to survival, and delay development and growth ( Jegla and Costlow 1982 Laughlin 1983 Sugita 1988 Ehlinger and Tankersley 2004 ) result ing in smaller animals ( Laughlin 1983 ) For example, u sing eggs from the Indian River Lagoon on the Atlantic coast of Florida, Ehlinger and Tankersley ( 2004 ) found that at high rearing temperatures (40 C) eggs failed to develop (regardless of salinity), and embryos fail ed to hatch into trilobite larvae (35 C). Additionally, using eggs from the northern Florida G ulf C oast Laughlin ( 1983 ) showed a reduced survival at low temperatures especially when coupled with low salinity (e.g., 2% survival at 20 C and 10 ppt ) Furthermore, t he least growth occurred in the highest (35 C) and lowest (20 C) temperature s and the longest development
32 was in the lowest ( 20 C and 25 C ) temperatures ( Laughlin 1983 ) Ehlinger & Tankersley ( 2004 ) found that e xtremes in salinity (i.e., 5, 80, 90 ppt ) substantially reduced survivorship and that hypersaline conditions delayed hatching but did not affect the duration of the larval (trilobite) stage Similarly, Laughlin ( 1983 ) found that extremes in salinity delayed development in general. In low salinities (below 18 ppt) Shuster ( 1979 ) found significantly smaller adults that matured one or two molts earlier. T emperature and salinity also affect rates of oxygen consumption ( Laughlin 1983 ) T emperature is thought to be the more important fact or: as temperature increased, so did the rates of O 2 consumption; and within each temperature treatment, O 2 consumption decreased as salinity increased ( Laughlin 1983 ) In general though, the oxygen content of water decreases with increasing temperature so it is hard to separate these factors. S ensitivity to these condit ions is age dependent, with adults and juveniles better able to withstand extreme temperatures, while embryos and larvae are better able to withstand extreme salinities and hypoxic or anoxic conditions ( Mangum et al. 1976 Reynolds and Casterlin 1979 Laughlin 1981 1983 Ehlinger and Tankersley 2004 ) Unfortunately, there are no corresponding data on the effects of temperature, salinity (or how they affect oxygen consumption) for cold water populati ons. However, these patterns are still useful in interpreting the effects of environmental variables on optimal adult body sizes. N o studies have examined how the environmental factors that affect growth and survival are correlated with average body size i n horseshoe crabs. Therefore, I examine the pattern of body size variation in horseshoe crabs along a 3,600 km
33 latitudinal gradient (18 44N) across North America to address how environmental factors might produce the presumed dome shaped size distributi on and evaluate the proximate hypotheses to explain size clines for this species. Methods Pattern of Body Size Distribution I searched journal articles, book s and federal and state Fish and Wildlife research reports for published information on the body size of adult horseshoe crabs in North America (see Appendix A for sources) I also located an unpublished manuscript that added information from Delaware Bay, and the Smithsonian Ft. Pierce website that provided sizes for animals from the Indian River La goon in FL ( see Appendix A). I supplemented these with field data collected by H. J. Brockmann on Seahorse Key, FL from 1995 1997, and with data that I collected from Stony Brook, NY (2007); Skidaway Island, GA (2007); Sapelo Island, G A (2009); and Seahor se Key, FL ( 2007 2009). In total, I compiled measurements from over 35,000 individuals from 29 different locations in North America (in 14 US states and 3 Mexican states), represent ing nearly the entire geographic range of the species (Table 2 1). Measurements of body size predominantly come from animals that were collected while spawning on the beach, but some were collected offshore during trawl surveys (North Carolina, South Carolina, and some an imals from Ochlockonee Bay, FL). Body size was mea sured as maximum carapace width (CW) for all locations with the exception of four f rom Mexico where size was measured as inter ocular distance (IO). For these locations, I estimated CW (cm) from IO (cm) with data from 396 individuals (from Delaware, Flo rida, Georgia, and Yucatn) that had both measurements. These
34 two traits are highly correlated (r 2 = 0.97, F 1, 395 = 11051, P > 0.0001; Figure 2 2 ), allowing me to accurately calculate CW wi th the following equation: ( 2 1) Sexual size dimorphism (SSD) is the result of females growing for an additional year (one additional molt) compared to males ( Smith et al. 2009a ) Therefore, the degree of dimorphism may be a reasonable proxy for growth rates from the penultimate t o ultimate stages in different populations, and potentially reflects overall growth rates. Ecological and Environmental Variables For each of the 29 locations I gathered data on the climatological mean for: 1) annual sea temperature ( C ) at 10 m, 2) salin ity (PSS) at 10 m, 3) dissolved oxygen (mg/L) at 10 m, 4) season length (days) and 5) food availability (#/m 3 ; Table 2 2). Data for annual temperature, salinity, and dissolved oxygen were obtained from the National s 2009 World Ocean Atlas; data on food availability were obtained from the National Oceanographic Data Center 2009 World Ocean database; and data for season length were obtained from the National Oceanic s database I chose to use values for temperature, salinity, and oxygen at 10 m rather than surface measurements because horseshoe crabs are often found on the ocean floor. Season length was defined as the number of days per year above the minimum temperature at which horseshoe crabs are active. Previous studies have shown that threshold temperatures for activity vary among populations ( Wenner and Thompson 2000 Schaller 2002 Watson et al. 2009 James Pirri 2010 Schaller et al. 2010 Smith et al. 2010 ) so I used
35 these sources to select the minimum temperatures when calculating season length: ME to NH = > 14.5 MA to GA = >15.5 and FL to YUC = > 16.5 This calculation of food abundance makes the ass umption that trawling surveys are a good representation of food availability. The dataset on trawling surveys from 1953 1999 that I used to calculate f ood availability was edited in the following way. Each line of the dataset contained a cast ID (each in dividual cast of the trawling equipment) and a catch value (number of a given taxonomic group per m 3 ) Therefore, each cast ID had several lines associated with it; each with a catch value for each different taxonomic group. The data set initially contained 45,092 unique entries of c ast ID x catch value; from this I first extracted all lines containing tax a that are known to be in the diet of horseshoe crabs (Appendix B), leaving 25,771 unique entries of cast ID x catch value (for horseshoe crab prey only) I then summed the catch across all taxonomic groups within a cast ID with total catch values. I then needed to calculate catch values for the latitudes of my locations. At this point, the datas et contained multiple entries for a given latitude Some of those entries were from surveys far out at sea, or for some reason were errors showing the survey location on land. I removed these entries, and then f or any given latitude I then averaged all o f the catch values leaving 546 unique latitudes with average number of prey items per cubic meter. For each exact latitude of my sample locations I took catch data for 1 latitude, and averaged that swath of data, leaving an average number of food ite ms / m 3 for 19 of my sample locations. There were no catch data for the locations in Mexico; Seahorse Key FL; Tampa Bay, FL; or Maine.
36 Analysis I first tested for the presumed dome shaped size distribution by conducting a regression of mean body size (lo cation average of males and females) and latitude with a polynomial fit. T he degree of SSD may reflect growth rates from the penultimate to ultimate stages in different populations ; thus, I compared degree of SSD with latitude using a simple liner regress ion. I used a multiple regression model t o analyze the relationship between mean size and the ecological and environmental variables However, to control for multicollinearity among temperature, oxygen, salinity, and food abundance, I first conducted a principal components analysis ( Graham 2003 ) Principal components analysis is scale sensitive, and my variables have quite different scales; therefore I first standardized all the variables by dividing them by their sample standa rd deviations ( Ramsey and Schafer 1997 ) I then constr ucted a full model with the PC1 PC3 and their interactions as explanatory variables, and m ean size as the response variable. To accommodate the possibility of parabolic relationships between body size and these variables, I fitted these data to the above model using a quadratic function (PC1 2 PC2 2 PC3 2 ). I then ran a stepwise (mixed selection) multiple regression analysis to find the best model. Because variable selection is unbiased in principal components regression, I used all principal components i n the stepwise multiple regression ( Graham 2003 ) thus, I did not lose any explanatory power, which can occur when using only variable with high eigenvalues ( Mitchell Olds and Shaw 1987 ) Results Pattern of Body Size Distribution Horseshoe crabs show a dome shaped size distribution : the largest animals were found in South Carolina (27.1 3.2 cm), whereas the smallest were found in New
37 Hampshire (14.4 1.0 cm) to the north; and Holbox, Quintana Roo (15.7 cm) to the south (Figure 2 3 Table 2 1). F rom Hog Bay, ME to Chincoteague Bay VA the pattern just south o f this location the Gulf Stream is close to shore, which results in a 10 C increase in temperature (i.e., in North Carolina). Over this range, t he change in body size was 7.3 cm over 6.8 latitude = 1.07 cm / latitude; change in water temperature was 5.7 C = 1.3 cm / C. From Nor th Carolina to Laguna de Trminos, Campeche, Mexico, change in body size was 8.0 cm over 16.5 latitude = 0.48 cm / latitude; change in temperatur e = 4.7 C = 1.7 cm / C. There was some grouping in body size across locations, resulting in five distinct clusters (Table 2 3): 1) Maine to New Hampshire, 2) Cape Cod to Sandy Hook, NJ, 3) DE Bay to Virginia 4) North Carolina to Georgia, and 5) Florid a to Mexico. Animals in the south had a relatively higher degree of sexual size dimorphism compared to animals in the north (a notable exception was for Progreso, Yucatn, which had a very small degree of dimorphism; Figure 2 3, Table 2 1). The degree of sexual size dimorphism (a proxy for growth rates from the penultimate to ultimate stages ) was negatively related to latitude when each location is considered separately ( r 2 = 0.15, F 1, 27 = 4.8, P = 0.038; F igure 2 3, Table 2 1), or when analyzed by the average degree of SSD within a cluster (Figure 2 3, Table 2 3) Ecological and Environmental Variables Temperature increased steadily with decreasing latitude, but showed a sharp increase south of Virginia due to the presence of the Gulf Stream near shore (Figure 2 4, Table 2 2). Season length and salinity also increased steadily with decreasing latitude, whereas dissolved oxygen levels decreased with decreasing latitude (Figure 2
38 4, Table 2 2). Unfortunately, I was not able to get food abundance data for the locations in Mexico, Maine, or for Seahorse, Key and Ochlockonee B ay in FL. Food abundance was highest in the mid Atlantic states and some parts of Cape Cod, MA, but was lower in the more southern latitudes; the lowest levels were found in NH, and th e Indian River Lagoon, FL (Figure 2 4, Ta ble 2 2). Temperature, season length, and dissolved oxygen loaded on the PC1 axis (eigenvalue = 3.97, 79.4%); food abundance loaded on the PC2 axis (eigenvalue = 0.77, 15.3%); and salinity loaded on the PC3 axis (ei genvalue = 0.06, 1.2%). The stepwise multiple regression model selected PC1 3 and their quadratics, and the interactions for PC1 x PC2 and PC2 x PC3 as explanatory variables. The stepwise model was significant ( R 2 = 0.83, F 8, 10 = 6.1, P = 0.005), and PC 1, PC1 2 and PC2 were significant at =0.05, while PC3 and PC1 x PC3 were significant at =0.06 (Table 2 4 ). Contour plots (Figure 2 5 ) illustrate how body size was affected by the interactions between PC1 (temperature, season length, and dissolved oxyge n) and PC2 (food abundance), and between PC1 and PC3 (salinity). All of the ecological and environmental variables affected body size, explaining 74 % of the variation in body size with latitude (for polynomial regression variation explained = 1 residual s um of s quares / total sum of squares). Discussion I found a dome shaped distribution of adult body size in horseshoe crabs across North America. The increase in body size from Hog Bay, ME to SC was consistent with Laguna de Trminos, Campeche was consistent with a Bergma Animals from t he
39 locations in Mexico were of similar size to those in Florida, despite being farther south, resulting in a flatter parabola of size and latitude than if these locations were not included (Figure 2 3 ) This pattern may be due t o only a slight variation of water temperature in Mexico, or to some other variable for which I have not accounted The degree of SSD declines with a decrease in temperature, and this may indicate that cold temperatures negatively affect growth in this sp ecies. However, SSD may not be indicative of growth rates, thus I make very limited inference from this result. Optimal resource allocation models ( 1987 1996 Cichon 1997 Teriokhin 1999 ) examine optimal age and size at maturity. Optimal age and size at maturity is considered to be a trade off between current and future reproduction, as mediated by growth strategy. For both difference between rates of energy assimilation and metabolism) at a given size, the length of time available for growth, and the allocation of su rplus energy to reproduction. In turn, how an organism allocates surplus energy is dependent on juvenile survivorship. Thus, maturity appears at different ages and sizes, depending on the size de pendence of the production rate, season length and mortali ty These models address ecogeographic clines because tempe rature along with other variables (e.g., O 2 salinity, season length, food abundance) covaries with latitude; and the interactive effect of multiple stressful conditions can influence optimal age and size at maturity by constraining the production rate and/or by affecting survivorship for organisms with
40 temperature dependent mortality ( Cooney et al. 1983 Hanazato and Dodson 1995 Folt et al. 1999 ) A key component of how the optimal resource allocation model produces ecogeographic clines in size is the thermal sensitivity ( coupled with other factors; Chown and Gaston 2010 ) of mortality (Q 10 mort ality ) versus the thermal sensitivity ( coupled with other factors; Chown and Gaston 2010 ) of the bioenergetic parameters of production: energy assimilation and met abolism (Q 10 production ). If Q 10 mort ality is higher than for Q 10 production then optimal size will decrease with temperature. In other words, the optimal growth strategy as temperature increases is to allocate more resources to reproduction versus growth (due to increasing mortality), thus maturing earlier at a smaller size (mortality costs of continued growth do not outweigh the fecundity benefits of growing larger). In this case, the body size distribution pattern is consistent with Bergm 10 mort ality is lower than for Q 10 production then optimal size will increase with temperature. In other words, the optimal growth strategy as temperature increases is to continue to allocate more resources to growth versus reproduction (du e to decreasing mortality), thus maturing later at a larger size (fecundity benefits of growing larger outweigh the mortality costs of continued growth). In this case, the body size itional component of this relates to the length of the growing season. In cold seasonal a given minimum temperature required for growth. For longer lived organisms, thi s means that even a significant delay in maturation results only in minor increases in size, and comes at high cost with little fecundity benefits. In contrast, in warm aseasonal
41 environments, production is not limited by days above a minimum temperature. From an ultimate perspective, patterns of size variation in horseshoe crabs appear to be consistent with the optimal allocation of resources in this model. Although h orseshoe crabs are able to tolerate a wide range of salinity and temperatures ( Ehlinger and Tankersley 2004 Shuster and Sekiguchi 2009 ) the opti mal temperature range for temperature is 25 33 C ( Jegla and Costlow 1982 Laughlin 1983 Ehlinger and Tankersley 2004 ) and the optimal range for salinity is 20 30 PSS ( Jegla and Costlow 1982 Laughlin 1983 Sugita 1988 Ehlinger and Tankersley 2004 ) Within these ranges: 1) higher temperatures increase survival, growth and development ; and 2) higher salinity increases survival whereas growth and development remain the same across this salinity range ( Laughlin 1983 Ehlinger and Tankersley 2004 ) Thus, large sizes in the mid latitudinal c luster from North Carolina to Georgia may result from high er survivorship or high er production. In this area, however, oxygen levels are relatively low, temperature is relatively high (compared to further north, temperatures jump rapidly due to the Gulf S tream, but are not as extreme as f a rther south) as is salinity ( Table 2 3 ), which potentially increases survivorship. These conditions, coupled with an active season length that is prolonged to 70% of the year, and an intermediate level of food abundance may increase production ( Table 2 3 ). Thus, approaching these mid latitudes, the optimal growth strategy for horseshoe crabs may be to allocate more resources to growth than to reproduction, resulting in later maturity at a larger size due to the fecundity advantage of continued growth. Small sizes at the most northern and southern latitudes may result from low survivorship or l ow production. In the northernmost cluster from Maine to New
42 Hampshire, both annual temperature and salinity are quite low ( Table 2 3 ), potentially resulting in low survivorship. Indeed, e xperimental studies on larval horseshoe crabs show that the combin ation of low temperature and low salinity are highly detrimental to survival ( Laughlin 1983 ) Production also may be limited at these latitudes given the short season length (only 22% of the year), especially when coupled with the experimental findings that low temperatures and salinities delay growth and development ( Jegla and Costlow 1982 Laughlin 1983 Sugita 1988 Ehlinger and Tankersley 2004 ) resulting in animals that mature at earlier molts and smaller sizes ( Shuster 1979 Laughlin 1983 ) Further more, low food abundance (which my limited data suggests is the case in this cluster ; Table 2 3 ) can reduce survival and production. I n the southernmost cluster from Florida to Mexico, temperatures and salinities are high ( Table 2 3 ) potentially resulting in low survivorship. Ehlinger and Tankersley ( 2004 ) showed that e xtremely high salinity has a negative impact on survivorship, as does temperature: at high rearing temperatures (regardless of salinity) eggs failed to develop (40 C), and embryos failed to hatch into trilobite larvae (35 C). High temperature and extreme salinity also impedes growth and speeds up development resulting in smaller sizes ( Jegla and Costlow 1982 Laughlin 1983 Sugita 1988 Ehlinger and Tankersley 2004 ) These conditions also affect oxygen consumption. Laughlin ( 1983 ) found that as temperature increased, so did the rates of O 2 consumption; however within each temperature treatment, O 2 consumption decreased as salinity increased. T he Indian River Lagoon, FL population occurs in a micro tidal estuary that can have salinities up to 55 PSS ( Ehlinger et al. 2003 ) which could severely affect survival and growth of juveniles. In contrast, salinity levels in estuaries
43 on the gulf coast are highly variable and frequently low (potentially increasing oxygen consumption) because major rivers and spring discharge large amounts of freshwater into them (e.g., Seahorse Key, FL and the Suwannee River). Thus, high temperature and salinity and wide variation in salinity, and low oxygen levels, coupled with low food abundance ( Table 2 3 ) decrease s survivorship and limit s production in Florida. In the four locations in Mexico, extreme high temperature and low oxygen levels, and a con tinually high salinity may decrease survivorship and limit production for these animals. Thus, the small sizes in the northernmost and southernmost clusters may accord with the limits to production and high mortality outlined in optimal resource allocation models. As they approach these latitudes, the optimal growth strategy for horseshoe crabs may be to allocate more resources to reproduction versus growth resulting in earlier maturation at a smaller size to avoid the mortality costs of continued growth. However, I lack sufficient detail to evaluate whether thermal sensitivities (in conjunction with other variables) are higher for mortality than for production across latitudes where Proximate Hypoth eses to Explain Ecogeographic Clines One of the most difficult and contentious questions in the study of ecogeographic clines revolves around what proximate factors are responsible for constrain ing growth and development ( von Bertalanffy 1960 Strong and Daborn 1980 Cushman et al. 1993 Atkinson 1994 Perrin 1995 van der Have and de Jong 1996 Van Voorhies 1996 Atkinson and Sibly 1997 Pa rtridge and Coyne 1997 Chown and Gaston 1999 Woods 1999 van der Have 2002 Angilletta and Dunham 2003 Chown and Klok 2003 Angilletta et al. 2004a Angilletta et al. 2004b Blanckenhorn and Demont 2004
44 Makarieva et al. 2005 Walters and Hassall 2006 Gaston et al. 2008 Chown and Gaston 2010 ) Many proximate hypotheses have been proposed, and below I summarize those hypotheses and the predictions they make Few (if any) previous studies have tried to distingu ish among these proximate hypotheses using a strong inference approach ( Chown and Gaston 2010 ) Thus, I also evaluate predictions of these alternative hypotheses with respect to horseshoe cra bs given the limited data available Hypotheses based on v g rowth e quation Some e xplanations regarding ecogeographic clines are based on von ( 196 0 ) growth equation model. These growth, such that body size follows asymptotic trajectory. The Perrin Bertalanffy hypothesis states that smaller sizes in warmer temperatures are a result of less energy available for somatic growth because the processes affecting anabolism are less affected by temperature than chemical processes driving catabolism (i.e., less efficient growth; von Bertalanffy 1960 Perrin 1995 ) Others have argued that it is not the coefficients of anabolism and catabolism that matters, but rather the allometries of catabolism and anabolism, which respond in opposite ways to temperature such that growth decelerates with age in warm environments (favoring early maturation at a small size) but accelerates with age in cold environments ( Strong and Daborn 1980 Angilletta and Dunham 2003 ) Nevertheless, explanations for ecogeographic clines that are based on have been largely rejected due to theoretical problems (e.g., many species do not show asymptotic grow, no animal grows indefinitely, growth must exclude tissues involved in reproduction) and contradictory evidence ( Angilletta and Dunham 2003 Chown and Gaston
45 2010 ) Consequently, Angilletta et al. ( 2004b ) treated as a phenomenological description of growth rather than a set of mechanisms Hypotheses based on physical c onstraints The minimum metabolic rate hypothesis ( Makarieva et al. 2005 ) is a biophysical constraint model based on interactions between whole organismal and cellular oxygen supply. This hypothesis states that mass specific metabolic rate increases more rapidly with temperature than does cellular meta bolic flux If the rate of oxygen diffusion into a cell is not dependent on temperature, then at higher temperature s a cell must become smaller to cope with increasing oxygen demands (Woods 1999). Presuming a constant cell number (which has been fiercely debated; Van Voorhies 1996 Mousseau 1997 Partridge and Coyne 1997 Van Vo orhies 1997 Chown and Gaston 1999 Chown and Gaston 2010 ) declining cell size should lead to declining body size. O xygen levels in aquatic and marine environments are generally more than an order of magnitude lower than atmospher ic oxy gen levels ( Makarieva et al. 2005 ) Hence, enhancing oxygen supply in response to increases in temperature is problem atic for aquatic organisms. The minimum metabolic rate hypothesis for marine and aquatic ectotherms, and such decreases in body length with temperature (and increases with dissolved oxygen) have been found in surveys of benthic amph ipods ( Chapelle and Peck 1999 Chapelle and Peck 2004 ) Can oxygen dependent thermal tolerance explain the body size pattern I observed in horseshoe crabs ? This hypothesis predicts a constant versus variable number of cells, which remain s to be tested in horseshoe crabs. P hysiological costs in response to low oxygen conditions for horseshoe crabs include increased lactate and decreased
46 pH levels caused by metabolic acidosis ( Towle et al. 1982 ) ; higher oxygen consumption ( Laughlin 1983 ) and increased heart and book gill ventilation rates that importantly, are more severely increased in juveniles compared to larvae ( Bademan 2009 ) Measurements of whole organism metabolic rate and cellular flux across a range of temperatures are lacking. However, i ncreased oxygen demand, coupled with low levels of dissolved oxygen in warmer temperatures, suggests a constraint on growth efficiency that may be consistent with t he minimum metabolic rate model But this hyp othesis southern latitudes. Next, t he biophysical constraints hypothesis ( van der Have and de Jong 1996 ) posits that the rate of cell growth is primarily affected by protein synthesis while the rate of cell differentiation is primarily affected by DNA replication; and that these rates are controlled by independent systems of enzymes ( Angilletta et al. 2004b ) They proposed that DNA polymerase enzymes involved in D NA replication are highly temperature dependent whereas ribosomal units involved in protein synthesis are less affected by temperature This implies that in higher temperatures, development rates are faster than grow rates. Consequently, an organism rea ch es maturity faster and at a smaller size ( van der Have and de Jong 1996 Walters and Hassall 2006 ) at low latitudes. Thus In contrast, Walters & Hassall ( 2006 ) temperature threshold hypothesis states that ecogeographic clines in adult size of ectotherms are determined by the relative difference between the minimum temperature thresholds for growth and development rates. These thresholds relate to the differential activation energy constants in van der ( 1996 ) hypothesis. This hypothesis also states that selection for
47 changes in body size operates through variation in these thresholds, and that the relative benefits of ch anges in threshold values are dependent on season length (i.e., smaller size in areas with shorter season length). This means their model is also an adaptive explanation. I f temperature thresholds for development are greater than thresholds for growth, t hen patterns of size distribution should follow Bergmann's rule; yet if temperature thresholds for development are greater than thresholds for growth then size distributions should conform to converse Bergmann's rule ( Walters and Hassall 2006 ) which give this hypothesis the potential to be a unifying mechanism to explain mak es it a more likely explanation for the size distribution of horseshoe crabs than the biophysical W hile the lat t er makes no specific predictions about season length, t he former predicts increased size with season length, which is found in horseshoe crabs. Horseshoe crabs are also sexually dimorphic, which is predicted by the temperature threshold hypothesis (but not by the biophysical constraints hypothesis). T he presence of sexual size dimorphism is predicted if males and females have different thresholds for growth and differentiation due to differential phenotypic plasticity ( Teder and Tam maru 2005 ) T hreshold temperatures for growth and development are known to exist in horseshoe crabs ( e.g., molting stops if juveniles from Cape Cod are placed in 13 15 C; Shuster and Sekiguchi 2003 ) however, what those values are across locations is not known. I predict that
48 the minimum threshold values for development will be gre ater than for growth; but development than for growth. The test of this prediction in horseshoe crabs will be highly valuable for whether or not this hypothesis is truly a unifyi ng mechanism to explain responses to temperature across ectotherms. Hypotheses Regarding Food Availability to Explain Ecogeographic Clines The starvation resistance hypothesis ( Calder 1984 Cushman et al. 1993 Stillwell et al. 2007 ) that large size at higher latitudes is an adaptation to starvation (or desiccation) in that larger individuals have lower mass specific metabolic rates, and higher levels of absolute fat store than small er individuals At a proximate level, this might occur if the scaling of maintenance metabolic rate over winter has a lower exponent than the scaling of reserve storage ( Chown and Gaston 2010 ) This hypothesis is not appropriate to explain the distribution of sizes in horseshoe crabs, even in the northern locations because it predicts a positive relationship between latitude and size in seasonal climates, and a negative relationship between season length and size latitudes. Variation in q uality and quanti ty of food can influence body size ( Rosenzweig 1968 McNab 1971 Blackburn et al. 1999 Davidowitz et al. 2004 Davidowitz and Nijhout 2004 Yom Tov and Geffen 2006 Wolverton et al. 2009 Ho et al. 2010 ) This relates to t he abundant cent er hypothesis, which states that animals are found in higher abundance in the center of their range ( Brown 1984 Brown et al. 1996 ) While the validity of the abundan t center hypothesis has recently been challenged ( Sagarin et al.
49 2006 ) the hypothesis could potentially explain larger sizes in the center of a species' range ( Virgs et al. 2011 ) If horseshoe crab prey (such as bivalves and polychaetes) have similar ranges to horseshoe crabs, and if they are more abundant in the center of their range, then the large size of horseshoe crabs at mid lat itudes may be caused by an abundance of food. In contrast, my data indicate that prey is not more abundant where horseshoe crabs are largest, but is i nstead most abundant in the mid Atlantic and southern New England states (Figure 2 3, Table 2 2). While abundant center hypothesis, this does not mean that food was not influential. Animals from Cape Cod, MA to Sandy Hook, NJ were considerably larger (37%) than those from Maine and New Hampshire (Table 2 3 ) This could be due to the 28% higher annual temperature or the 36% longer season length However, changes in environmental conditions were fairly linear across groups of locations, and size between the other groups only tended to change by about 10%. What makes these loca tions different is that food abundance is the highest in these locations This suggests that animals from Cape Cod to Sandy Hook may achieve a larger than expected size due to the abundance of food, or they could increase their population sizes, which is indeed the pattern we observe ( Anderson and Shuster 2003 ) Dome Shaped Distri butions are quite unusual. A bowl shaped size distribution ( mid latitude trough ) is found in Northeastern pacific marine bivalves ( Kaustuv and Martien 2001 ) and also in damselflies of the species Enallagma cyathigerum ( Johansson 2003 ) although this pattern is likely associated with bivoltine / univoltine shifts (even though the predicted sawtooth pattern is not observed). A dome shaped distribution ( mid latitudinal pe ak)
50 has been reported in only a handful of other species including the p upae (but not adults) of the ant species Myrmica rubra ( Elmes et al. 1999 ) The snout ve nt length of common frogs, Rana temporaria ( Laugen et al. 2005 ) is greatest at mid latitudes, however, there is also a corresponding concave curve of temperatures that can explain this pattern. Geist ( 1987 ) reported a dome shaped distribution in deer and wolves, and explained the pattern based on food abundance. While Geist's study has received substantial scrutiny ( Paterson 1990 ) a newer study on the same system shows the originally reported relationship more convincingly ( Wolverton et al. 2009 ) Lastly, a recent study has found support for the abundant center hypothesis in explaining a dome shaped distribution in the mass of European badgers, Meles meles ( Virgs et al. 2011 ) While this hypothesis is not a likel y explanation for horseshoe crabs, an important aspect of t he badger study was that it examined animals over a much broader scope of their range then previous studies, and subsequently was able to identify a pattern that was not otherwise apparent. Similar to European badgers, horseshoe crabs occupy a large latitudinal range (26, 3,600 km), which may subject them to a broader range of environmental and ecological conditions than most species, and may help explain why dome shaped distributions are so rare, and why one is found in horseshoe crabs. The genetic structure of horseshoe crab populations may also contribute to this pattern. Populations of horseshoe crabs are fairly isolated due to physical barriers and the Gulf Stream ( Shuster 1979 Riska 1981 Botton and Ropes 1987 Botton and Loveland 1992 ) and there is little gene flow among them ( King et al. 2005 ) As such, low gene flow or high genetic variance might enh ance the potential for variation by local adaptation ( Blows
51 and Hoffmann 2005 Chown and Terblanche 2007 Gaston et al. 2008 ) to mechanistic factors such as the minimum temperature thresholds for growth and development. Lastly, the long development and generation times of horseshoe crabs might influence their pattern of body size distribution. Uncoupling the effects of temperature and season length on body size is difficult because both generally decline together with increasing latitude ( Arnett and Gotelli 1999 ) However, the importance of generation time relative to season length in determining whether season length or temperature has the most significant influence on size has considerable empiric al support ( Chown and Klok 2003 Blanckenhorn and Demont 2004 Chown and Gaston 2010 ) Larger species with longer development / generation times are more prone to season length constraints and tend to show c onverse Bergmann clines, whereas smaller species with shorter development / ge neration times tend to show Bergmann clines ( Chown and Klok 2003 Blanckenhorn and Demont 2004 Chown and Gaston 2010 ) Horseshoe crabs suffer from season length constraints in their northerly locations, and here they show c onverse ever, in the southern part of their range, they are not may be due to temperature e ffects. Conclusions In summary, I found that Horseshoe crabs show an unusual dome sha ped distribution in body size. Ultimate explanations for this pattern are that 1) optimal resource allocation to reproduction, mediated by survivorship, is influenced by local environmental and ecological conditio ns, and 2) optimal body size result s from tradeoffs between growth, survival, and fecundity. The dome shaped pattern has been suggested to result from optimal temperatures and salinities in the center of their range
52 ( Shuster 1957 Reynolds and Casterlin 1979 Shuster 1979 Thompson 1998 Ehlinger and Tankersley 2004 ) or from a short growth season in the north ( Anderson and Shuster 2003 ) Evidence from this study suggest that levels of dissolved oxygen and food abundance are also likely important. In a proximate sense, I infer th at the temperature threshold hypothesis is the most likely explanation for the proximate mechanism underlying the ecogeographic cline in horseshoe crabs, however we lack sufficient data to make a definitive conclusion. While less parsimonious, the dome sh aped pattern could also be explained using a combination of these hypotheses. For example, the minimum metabolic rate model pertaining to oxygen demand versus oxygen supply also may explain the small sizes in the south. The biophysical constraints model via an increased enthalpy or inactivation of cell cycle enzymes, thus constraining growth, may explain small sizes in the north. But again, we lack the data to evaluate this in a rigorous manner. This study adds further evidence to the notion that Bergma c and that constraints on growth can act multiplicatively to create a dome shaped distribution in body size ( Blanckenhorn and Demont 2004 ) Patterns of body size distribution should be further examined in other marine species with large ranges (or that occur over a broad range of environmen tal conditions) long development times, genetic population structure, and high variation in season length between known populations. More importantly, this study sets the stage for future research on the unusual size distribution of horseshoe crabs to be one of the first demonstrations of a unifying mechanism that temperature thre shold hypothesis be supported) In the future, this observational study
53 should be bolstered with experiments that manipulate temperature, oxygen, season length, and salinity to get a better understanding of minimum temperature thresholds for growth and di fferentiation, and to better understand the effects of oxygen demand and supply on growth.
54 Table 2 1. Locations, latitude, and body size for the pattern of size distribution of horseshoe crabs in North America. References can be found in appendi x A. State Location Latitude Male size Female size Average size Size ratio SD n SD n SE (F:M) Maine Hog Bay 1 44.3 14.7 na 951 17.4 na 337 16.1 na 1.184 Bagaduce River 1 44.2 13.9 na 9 17.5 na 12 15.7 na 1.259 Thomas Point Beach 1 43.5 14.8 na 26 19.2 na 14 17.0 na 1.297 New Hampshire Great Bay 2 43.0 12.8 0.9 263 16.0 1.2 147 14.4 0.05 1.258 Massachusetts Cape Cod Bay 3,4 41.8 17.4 1.6 2942 22.7 1.9 759 20.1 0.05 1.305 Nauset Estuary 3 41.8 17.5 1.7 433 23.4 2.1 256 20.5 0.13 1.337 Monomoy NWR 3 41.7 18.8 1.5 909 24.2 2.2 477 21.5 0.08 1.287 Pleasant Bay 3,4 41.7 17.3 1.2 2063 22.5 2.0 413 19.9 0.04 1.301 Rhode Island Narragansett Bay 5 41.3 18.6 a 54 24.0 a 288 21.3 na 1.290 Connecticut Milford, New Haven, Norwalk 5 41.2 19.5 b 1760 24.9 b 1145 22.2 na 1.277 New York Stony Brook 6 40.6 19.6 1.4 27 25.9 1.7 14 22.7 0.37 1.322 Inshore continental shelf 5 40.5 20.5 c 121 26.2 c 13 23.4 na 1.278 New Jersey Sandy Hook Bay 2 40.3 20.1 1.3 197 25.4 1.8 103 22.8 0.12 1.263 Delaware Delaware Bay 7,8 39.0 20.7 1.4 1233 26.6 1.9 803 23.6 0.06 1.287 Maryland Ocean City 9 38.2 21.0 na 1706 26.6 na 1210 23.8 na 1.268 Virginia Chincoteague Bay 9 37.5 20.6 1.4 3029 26.1 1.8 1751 23.4 0.03 1.266 North Carolina Inshore continental shelf 10 35.0 21.8 2.5 457 27.1 3.7 822 24.5 0.31 1.243 South Carolina Inshore continental shelf 11, 12 32.9 23.6 1.4 406 30.7 2.1 418 27.1 0.11 1.302 Georgia Skidaway Island 6 31.5 23.0 1.8 37 30.7 1.7 49 26.8 0.31 1.331 Sapelo Island 6 31.2 23.0 1.5 72 30.1 2.3 54 26.5 0.30 1.306 Florida Ochlockonee Bay 13 29.5 16.9 1.3 1552 22.3 2.2 742 19.6 0.05 1.324 Seahorse Key 6 29.0 16.1 1.1 3556 21.9 1.6 2300 19.0 0.02 1.36 Indian River Lagoon 14 28.3 13.6 na na 18.9 na na 16.3 na 1.390 Tampa Bay 15 27.5 13.6 1.0 na 18.0 1.7 na 15.8 na 1.324
55 Table 2 1. Continued. State Location Latitude Male size Female size Average size Size ratio SD n SD n SE (F:M) Yucatn San Felipe lagartos 16 21.3 14.0 1.0 55 18.8 2.0 39 16.4 0.23 1.343 Progreso 16 21.2 14.7 na 8 17.5 na 7 16.1 na 1.190 Quintana Roo Holbox 16 21.3 13.6 na 44 17.8 na 26 15.7 na 1.309 Campeche Champotn 16 19.2 16.8 na 27 22.9 na 37 19.9 na 1.363 Laguna de Trminos 16 18.5 14.2 na 333 18.7 na 395 16.5 na 1.317 1 16 References, see Appendix A Estimated values of carapace width (CW) from interocular distance (IO). Formula: a Range: males = 15.9 22.4, females = 20.1 30.0 b Range: males = 10.0 30.0, females = 15.0 37.0 c Range: males = 17.4 23.1, females = 21.9 31.2 c Males: mean IO = 8.3 0.22 cm (SE); females: mean IO = 10.2 0.75 (SE) d Males: mean IO = 7.5 0.1 cm (SE); females: mean IO = 10.4 0.21 (SE) e Males: mean IO = 9.7 0.13 cm (SE); females: mean IO = 14.0 0.15 (SE) f Males: mean IO = 7.9 0.04 cm (SE); females: mean IO = 11.1 0.05 (SE)
56 Table 2 2. Values for environmental and ecological variables hypothesized to affect the growth and survival of horseshoe crabs at 29 locations across North America. Data for annu al temperature, salinity, and dissolved oxygen were World Ocean database, and data for season length was obtained from the National Oceanic and Atmospher ic Location Latitude Temperature (C) Season length (days) Salinity (PSS) Dissolved oxygen (mg/l) Food abundance (#/m 3 ) Coordinates a Hog Bay, ME 44.3 8.037 81 32.505 7.043 na 44.125 68.125 Bagaduce River, ME 44.2 8.057 81 32.402 7.038 na 44.125 68.875 Thomas Point Beach, ME 43.5 8.435 81 32.096 6.928 na 43.625 69.875 Great Bay, NH 43 8.236 81 31.825 7.115 11 43.125 70.375 Cape Cod Bay, MA 41.8 9.569 130 32.148 6.673 1479 41.625 69.875 Nauset Estuary, MA 41.8 9.559 130 32.127 6.673 591 41.875 69.875 Monomoy NWR, MA 41.7 9.847 130 32.096 6.673 591 41.875 70.125 Pleasant Bay, MA 41.7 9.569 130 32.148 6.673 1479 41.625 69.875 Narragansett Bay, RI 41.3 10.641 133 31.54 5.994 1201 41.125 71.875 Milford, New Haven, Norwalk, CT 41.2 10.756 133 31.001 6.319 1201 41.125 72.625 Stony Brook, NY 40.6 10.979 133 32.352 6.37 1212 40.875 71.125 NY, ISCS b 40.5 12.125 133 31.694 6.272 1256 40.625 73.375 Sandy Hook Bay, NJ 40.3 11.634 135 31.654 6.321 1416 40.375 73.875 DE Bay, DE 39 10.861 168 29.431 6.2 1111 39.125 75.125 Ocean City, MD 38.2 13.017 170 31.546 6.2 747 38.375 74.875 Chincotaegue Bay, VA 37.5 13.764 170 29.612 5.725 388 37.875 75.375 NC, ISCS b 35 22.124 217 35.333 4.727 156 35.125 76.125 SC, ISCS b 32.9 22.179 264 35.729 4.949 744 32.875 79.125 Skidaway Island, GA 31.5 20.403 274 34.665 5.062 452 31.875 80.875 Sapelo Island, GA 31.2 21.206 274 34.923 4.857 525 31.375 81.125 Ochlockonee Bay, FL 29.5 22.615 320 34.884 4.899 na 29.875 84.375 Seahorse Key, FL 29 22.967 320 35.275 5.03 na 29.125 83.375
57 Table 2 2. Continued Location Latitude Temperature (C) Season length (days) Salinity (PSS) Dissolved oxygen (mg/l) Food abundance (#/m 3 ) Coordinates a Indian River Lagoon, FL 28.3 26.019 351 36.034 4.727 98 27.375 80.375 Tampa Bay, FL 27.5 24.046 355 35.794 5.573 478 27.625 82.875 San Felipe lagartos, YUC 21.3 24.407 365 36.359 4.594 na 21.625 88.375 Progreso, YUC 21.2 24.79 365 36.456 4.595 na 21.375 89.625 Holbox, ROO 21.3 25.325 365 36.189 4.569 na 21.625 87.375 Champotn, CAM 19.2 26.751 365 36.317 4.413 na 19.375 90.875 Laguna de Trminos, CAM 18.5 26.799 365 36.057 4.413 na 18.625 91.625 a Data for dissolved oxygen were not available at 1/4 th degree intervals, therefore I used the value at the nearest 1 degree entry. b Inshore continental shelf
58 Table 2 3. Average values for clusters of body size of horseshoe crabs found across locations in North America. Location Average Size (cm) Temperature (C) Season length (days) Salinity (PSS) Dissolved oxygen (mg/l) Food abundance (#/m 3 ) Maine to New Hampshire 15.8 8.2 81 32.2 7.0 na Cape Cod to Sandy Hook, NJ 21.6 10.5 131 31.8 6.4 1158 Delaware Bay to Virginia 23.6 12.5 169 30.1 6.0 749 North Carolina to Georgia 26.2 21.5 257 35.2 4.9 469 Florida to Mexico 17.3 24.9 352 35.9 4.8 288
59 Table 2 4. Results of s tepwise multiple regression model for environmental and ecological variables influencing body size. Temperature, dissolved oxygen, and season length loaded on the PC1 axis; food abundance loaded on the PC2 axis; and salinity loaded on the PC3 axis. Stepwise multiple regr ession model Effects Test F Statistic P value PC1 18.9 0.002 PC2 8.0 0.018 PC3 4.5 0.059 PC1 2 20.8 0.001 PC2 2 3.7 0.347 PC3 2 3.1 0.391 PC1 x PC3 4.6 0.057 PC2 x PC3 2.9 0.120 Parameter Estimates Estimate SE T Ratio P value Intercept 29.4 1.5 20.3 < 0.0001 PC1 4.0 0.9 4.3 0.002 PC2 3.4 1.2 2.8 0.018 PC3 4.0 1.9 2.1 0.059 (PC1+0.54) 2 1.8 0.4 4.6 0.001 (PC2+0.08) 2 0.9 0.9 1.0 0.347 (PC3+0.09) 2 3.4 3.8 0.9 0.391 (PC1+0.54) x (PC3+0.09) 4.8 2.2 2.2 0.057 (PC2+0.08) x (PC3+0.09) 4.4 2.6 1.7 0.120
60 Figure 2 1 Hypothetical scenarios that could result in the mid latitudinal peak (dashed line) in body size distribution (black curves) seen in horseshoe crabs (M. J. Angilletta, personal communication). (A) If survivorship (red curve) is high across most latitudes, then variation in size may be driven by changes in growth rates (blue curve) with latitude (i.e., temperature). (B) If juvenile survivorship is low in temperature extremes, th en variation in size may be driven by changes in juvenile survivorship across latitudes, regardless of the effects of growth rate. My data indicate that body size of horseshoe crabs is more strongly driven by growth than survivorship
61 Figure 2 2 Scaling for carapace width (CW) from interocular distance used to estimate CW for the size of horseshoe crabs from four Mexican populations.
62 Figure 2 3 A. Quadratic relationship between body size and latitude in male and female horseshoe crabs in North America B. Relationship between sexual size dimorphism (SSD) and latitude. The degree of SSD may be used as a proxy for growth rates from the penultimate to ultimate stages in different populations given that females are larger than males due to an extra year of growth (one additional molt).
63 Figure 2 4. Plots of environmental variables (temperature, salinity, dissolved oxygen, and food abundance) and body size of horseshoe crabs across North America. Circle size represents average horseshoe crab body size, color represents scale of variable.
64 Figure 2 5 Contour plots for body size of horseshoe crabs. PC1, PC2, and PC3 are principal components comprising 1) temperature, season length, and dissolved oxyge n; 2) food abundance; and 3) salinity. The following sites are indicated: Nh = Great Bay, NH; Cc = Cape Cod Bay. MA; Ne = Nauset Estuary, MA; Mr = Monomoy NWR, MA; Pb = Pleasant Bay, MA; Ri = Narragansett Bay, RI; Ct = Milford, New Haven, and Norwalk, CT; Sb = Stony Brook, NY; Ny = Inshore continental shelf, NY; Nj = Sandy Hook Bay, NJ; De = Delaware Bay, DE; Md = Ocean City, MD; Va = Chincoteague Bay, VA; Nc = Inshore continental shelf, NC; Sc = Inshore continental shelf, SC; Sk = Skidaway Island, GA; S a = Sapelo Island, GA; Ir = Indian River Lagoon, FL; Tb = Tampa Bay, FL.
65 CHAPTER 3 AN EVALUATION OF THE ULTIMATE FACTORS UNDERLYING THE EVOLUTION AND MAINTENANCE OF SEXUAL SIZE DIMORPHIS M IN HORSESHOE CRABS Introduction Across taxa, males and females of th e same species commonly have different adult body sizes. Most endotherms show a pattern of male biased sexual size dimorphism (SSD), whereas most ectotherms are female biased ( Fairbairn 2007 ) This widespread pattern is intriguing given that body size is such an influential trait ( Calder 1984 ) Many adaptive and mechanistic hy potheses have been suggested to explain SSD; most depend on disruptive selection on body size between the sexes due to differences in their reproductive roles. For example, competition for mates between males is the likely cause of male biased SSD in fallow deer, Dama dama ( McElligott et al. 2001 ) ; whereas fecundity selection likely explains the female biased SSD in spiders ( Head 1995 ) In reality, SSD is not simply the result of one selection pressure on males or females, but rather the product of a s uite of conflicting pressures on both sexes ( Moore 1990 Badyaev et al. 2000 Blanckenhorn 2000 Badyaev et al. 2001 ) Few studies have examined multiple hypotheses to explain the ult imate factors underlying SSD (but see Blanckenhorn 2007 Szkely et al. 2007 ) In this Chapter I use the hypothetico deductive method to examine the evolution of female biased SSD in North American horseshoe crabs, Limulus polyphemus Female horseshoe crabs are consistently larger than males of the same age in all populations; f or 12 locations in North America, the mean female to male size ratio is 1.29 0.04 SD ( Brockm ann and Smith 2009 ) Horseshoe crabs are an interesting species for the study of SSD because of their unusual natural history. They are long lived marine arthropod s that take many years to reach sexual maturity (9 10 years to reach maturity and a life
66 span of 17 19 years), and they have a terminal molt at the time of sexual maturation T hus they have determinant growth ( Koons 1883 S huster 1955 Sokoloff 1978 Riska 1981 Shuster and Sekiguchi 2003 ) Furthermore, the proximate cause of SSD in this species is known: females are larger due to one additional mo lt and one additional year of growth before sexual maturity, compared to males ( Smith et al. 2009a ) To understand why this pattern is adaptive, I tested seven functional hypotheses (Table 3 1) by using field studies of reprod ucing individuals, along with a comparative analysis of multiple horseshoe crab populations along the Atlantic coast of North America. Study S pecies Horseshoe crabs are divided into genetically distinct subpopulations ( King et al. 2005 ) in many parts of their range along the east and south coasts of the US, and along the north and west coasts of the Yucatan Peninsula These populations vary in a number of traits such as population size, population density, operational sex ratio, and mor pholog y ( Riska 1981 Shuster 1982 Brockmann and Smith 2009 ) Horseshoe crabs are external fertilizers, and usually nest on sandy, low energy beaches of bays and barrier islands. On the Gulf of Mexico coast of Florida, the breeding season occurs from February to May, and again from late August to October, but more northern populations have a single breeding season late in the spring. Within a breeding season, sp awning occurs during daily and nightly high tides around the new and full moons (which are the highest tides of the month). The mating system of horseshoe crabs is are a bundant for a brief time during which mating is frequent ( Thornhill and Alcock 1983 Brockmann 1990 ) Sp awning is highly synchronized both temporally ( Rudloe 1980 ) and spatially ( Penn and Brockmann 1994 ) and this often results in high nesting
67 densities. Individual females generally nest f or only one week in a breeding season, whereas males continue to nest for many weeks ( Brockmann and Penn 1992 Brockmann and Johnson 2011 ) For many populations this results in a highly male biased operational sex ratio (OSR) on spawning beaches ( Botton and Loveland 1992 ) High nesting densities and a male biased OSR are characteristics of a population with strong intrasexual competition ( Shuster and Wade 2003 ) Indeed, there is often intense competition among males as they push against and crawl over each other for acces s to desired mating positions. Male horseshoe crabs exhibit two condition dependent, alternative mating tactics ( Brockmann and Penn 1992 Brockmann 2002 ) Some males, those that are younger and in better condi tion, attach to females out at sea, and arrive on spawning beaches paired with a female in amplexus. These males usually remain with the female until she has complete d her egg laying for the year, when they detach and seek another female. Other males, th ose that are older and in poorer condition, roam the shoreline not attached to a female. These unpaired males join spawning pairs as satellites and engage in sperm competition with the attached males and other satellite males ( Brockmann 1990 Penn and Brockmann 1995 Bro ckmann 2002 ) Satellite males that achieve the most desirable positions are quite successful in fertilizing a significant proportion of eggs ( Brockmann and Penn 1992 Brockmann et al. 1994 Brockmann et al. 2000 ) On average, a satellite male gains 40% of fertilizations, though this decreases as the number o f satellites present increases. Attached and satellite males clearly differ in age, condition, and behavior; but they do not differ in size ( Penn and Brockmann 1995 Brockmann 2002 ) and there is no large size advantage for attached
68 males. However, whether there is a size advantage for satell ite males in competition with other satellites is not known. The frequent highly male biased OSR and high female nesting densities, along with the presence of alternative mating tactics in this system may contribute to the evolution and maintenance of SSD in horseshoe crabs. Adaptive Hypotheses T hat Explain SSD I evaluate multiple predictions from five alternative hypotheses to better understand the evolution and maintenance of SSD in horseshoe crabs. I recognize that these hypotheses are not mutually exclusive, and multiple selection pressures affect the female biased SSD in this species. One of the hypotheses depends on the ecological differences between males and females (i.e., the competitive displacement hypothesis), while the rest depend on differences in reproductive roles between males and females. I organize these hypotheses into three groups those that refer to: 1) natural selection favoring a divergence in size between the sexes without predicting directionality for either males or females (i.e., the competitive displacement hypothesis), 2) sexual selection favoring small male s ize (i.e., t he size agility and loading constraints hypotheses), and 3) sexual selection favoring large female s ize (i.e., the female nesting competition and fecundity advantage hypotheses). I first provide a brief overview of the hypotheses and then evaluate each w ith respect to the horseshoe crab system. Natural selection for divergence in size Ecological differences between males and females are common, but whether these differences act as selective forces for the evolution of SSD or are simply a consequence of SSD that ha ve evolved for other reasons ( e. g., sexual selection; Shine 1989b ) is unknown. However, p opulation genetics models support the hypothesis that
69 SSD could be an adaptive response to ecological selection pressures in species that are initially monomorphic ( Slatkin 1984 ) Competitive displacement hypothesis The resources used by individuals are determined in part by their body size which means that competition between the sexes for a limiting resource may favor sexual size dimorphism ( Schoener 1967 Slatkin 1984 Hedrick and Temeles 1989 Shine 1989a ) For example, larger individuals can consume larger food items than smaller individuals (or those that are not within the reach of smaller individuals), so differences in size between males and females could reduce intersexual competition for limiting re sources (e.g., food) during the breeding or non breeding season. Competitive displacement resulting in sexual dimorphism of trophic structures (e.g., bill size) has been demonstrated ( Selander 1966 Temeles et al. 2000 ) and the effects of competiti ve displacement on overall body size have been supported by empirical studies ( Madsen 1983 MacDonald 1985 Kruger 2005 ) However, this hypothesis remains controversial because dimorphism of trophic structures still can be a product of selection for large size, which in turn favors changes in bill structure. This hypothesis differs from the others in that SSD is not directly related to reproductive roles, and thus this hypothesi s does not predict the direction of SSD (only that males and females should diverge). Sexual selection for small m ale s ize Agility advantage hypothesis Male biased SSD is generally thought to be the result of sexual selection favoring larger male size as an advantage in male male interference competition for mates prior to copulation ( Darwin 1874 Stamps 1983 Warner and Schultz 1992 Weckerly 1998 Cox et al. 2003 ) Whereas, in taxa with female biased SSD, small male size is sometimes favored by female choice for smaller
70 males that better perform courtship displays ( Payne 1984 ) In contrast, pre copulatory female choice is not likely a strong selection pressure for taxa with a scramble competition mating system such as that seen in horseshoe crabs and frogs. Instead, in these taxa, selection for smaller agile males may be the result of advantages in mating competition via mate searching (e.g., lower metabolic costs of small size; ( Ghiselin 1974 Reiss 1989 ) or in gaining access to mates (e.g., sneaker morphs in alternative mating tactics ( Blanckenhorn 2000 ) In horseshoe crabs, attached and satellite males compete for access to preferred positions in a mating group. Satellite males gain the highest paternity when they: 1) achieve pos itions closer to the female, particularly those attached male, hereafter referred ( Figure 3 1; Brockmann et al. 1994 Brockmann et al. 2000 ) In turn, attached males compete with satellite males to achieve this under position, and gain higher paternity as well. A size advantage for large ma les during mate competition has not been found ( Brockmann 1990 ) but advantages for smaller males have not been explored. A smaller body size may allow attached males to better thwart satellite males from achieving the under position, or in contrast, small size may foster the success of satellite males in achieving this under position. Additionally, small size may facilitat e satellite males in reaching the 1F satellite position better than a large size might. Loading constraints hypothesis While there are no size related constraints on the ability of males to amplex with females ( Botton and Loveland 1992 Loveland and Botton 1992 Suggs et al. 2002 ) the ability of females to carry certain males may be
7 1 limited or costly. The loading constraints hypothesis states that sexual selection favors smaller size in the sex that is carried due to costs associat ed with carrying mates In some taxa males carry females (e.g., isopods; Adams et al. 1985 Hatcher and Dunn 1997 ) while in others females carry males ( e.g., amphipods, Hatcher and Dunn 1997; water striders, Fair bairn 1990 ) beaches, while spawning on shore, and while out at sea waiting for the next breeding opportunity. Female choice may favor a reduction in male size if smaller males confer lower energetic co sts to females during amplexus ( Botton and Loveland 1992 ) Females may not be able to influence which males attach to them, but they could exert choice by simply not moving to the beach to spawn when an unfavorable male is attached As a result, spawning pairs shoul d show size a ssortative mating : small males should be found with small females, intermediate sized males with intermediate size females, and large males with large females. Evidence regarding size assortative mating in horseshoe crabs is equivocal. Studi es have not found size assortative mating in Cape Cod, MA ( Pomerat 1933 Suggs et al 2002 ) Delaware Bay ( Botton and Loveland 1992 Loveland and Botton 1992 ) or Seahorse Key, FL ( Cohen and Brockmann 1983 Brockmann 1990 ) In contrast, males that were experimentally offered a choice of different sized females tended to show a self similar, size dependent preference ( e.g., small males choose to at tach to the small female, etc.; Suggs et al. 2002 ) They also found that in field populations, pairings did not occur between the smallest or largest males and females, and that size reversal in pairs (i.e., where males are larger than females) occurs only rarely (also see: Loveland and Botton 1992 Suggs et al. 2002 ) suggesting some size based mate selection at the largest and smallest
72 sizes. All previous studies have used car apace width (CW) as the measure of size, yet, if assortative mating is based on the cost of carrying males, the mass of males and females should also be considered. Therefore, I examined size assortative mating based on carapace width and mass, and also e xamined patterns of mating in the largest and smallest animals from Seahorse Key FL. Sexual s election for large female s ize Nesting competition hypothesis In the same way that mate competition between males can contribute to the evolution of larger male size, competition for optimal nesting sites between females may favor larger female body size. Female horseshoe crabs preferentially choose to spawn within specific, narrow sites on a beach that maximize the developmental success of their eggs ( Botton e t al. 1988 Penn and Brockmann 1994 Jackson et al. 2007 ) The spawning density of females is quite low in some populations, but in others density can be so high that females run into and supplant each other, and inadvertently dig up e ( Smith 2007 ) Larger femal es may have an advantage during this competition and may be better able to achieve optimal nesting sites on a beach Fecundity advantage hypothesis The capacity for increased fecundity in larger females (i.e., fecundity selection) is thought to be one of the primary selection pressures underlying the evolution of female biased SSD ( Darwin 1874 Shine 1988 Andersson 1994 Stephens and Wiens 2009 ) Increased fecundity in larger females results from an increase in space to house more eggs or embryos ( Will iams 1966 ) and from an increase in energy storage for reproduction ( Calder 1984 ) Fecundity is positively related t o body size in ectotherms ( Thornhill and Alcock 1983 Seigel and Ford 1987 Stearns 1992 Andersson 1994 Shine 2005 Cox et al. 2007 Stephens and Wiens
73 2009 ) but not in endotherms such as mammals, where fecundity is negatively related with body size ( Boyce 1988 Harvey et al. 1989 Lee et al. 1991 Purvis and Harvey 1995 ) The fecundity selection hypothesis focuses on trade offs between growth and reproduction. In horseshoe crabs, the benefit of higher fecundity achieved by growing for an additional year must outweigh the costs of that growth and the demographic effect of one lost year of breeding. Methods Competitive Displacement Hypothesis One important assumption of the competitive displacement hypothesis is that males and females eat different typ es of food, or they forage in different locations (Table 3 1). I tested this assumption by using stable isotope analyses of 15 N and 13 C. Values of 15 N indicate the trophic level at which animals are feeding: as animals feed at higher trophic levels, t he value of 15 N in their tissues increases ( Michener and Schell 1994 ) I n marine organisms, 13 C values can indicate where animals are feeding relative to the shore: inshore food webs are typically more enriched in 13 C values than offshore food webs, and so animals feeding closer to the shore would have enriched levels of 13 C in their tissues ( Michener and Sch ell 1994 ) For these analyses I used two tissue types from horseshoe crabs from Seahorse Key, FL: 1) feces, which indicate diet during the spawning season), and from a different set of animals I collected 2) claw chitin which indicates diet during the year prior to their ultimate molt into adulthood. From 10 14 March 2009, I collected spawning adult horseshoe crabs (males n = 38, females n = 19) during evening high tides on Seahorse Key, FL, and placed them in individual holding tanks at the Universit y of Florida Seahorse Key Marine Lab. After 12
74 hours, I collected feces from the tanks for stable isotope analysis As part of a larger, separate genetics study, claws and associated tissues were collected from horseshoe crabs during 2007 2009 ( Johnson and Brockmann. 2010 ) I randomly chose 39 samples from 2008 (males: n = 20; females: n = 19), and cut out a small piece (6mm x 6mm) of the claw chitin (removing any muscle t issue) for stable isotope analysis. To determine values of 15 13 samples were analyzed by the Stable Isotope Mass Spectrometry Lab in the Department of Geological Sciences at The University of Flor ida. Agility Advantage Hypothes i s The first prediction of the agility advantage hypothesis is that if smaller males have an advantage in mating competition, then average male size will correlate negatively with the mean operational sex ratio (OSR) or the m ean degree of SSD across populations (Table 3 1). I tested this prediction by conducting a literature search for values of mean operational sex ratio (number of mating males to females) and morphological data from as many populations as I could find ( n = 13 ; Table 3 2; Appendix C ). Values for OSR were square root transformed to meet the assumptions of heterogeneity of variance and a normal distribution. I then used simple linear regression s to examine the relationship between male size, OSR and degree of SSD The second and third predictions of this hypothesis (Table 3 1) deal with how body using data from the 1995 1997 and 2007 2009 mating seasons at Seahorse Key, FL. The dataset was gathered by walking along the nesting beach, locating a mating group, and collecting values for 1) mating group size, 2) mating position relative to the attached
75 male (e.g., over or under the frontal margin of the carapace; Figure 3 1a ,b), and satellite position relative to the female (e.g., 1F or non 1F position; Figure 3 1c ), 3) body size, and 4) condition of all males in the group I included condition in the following analyses because condition is thought to affect the ability of m ales to reach desired positions ( Brockmann et al. 2000 ) I assessed condition using an index based on visual inspection of the carapace. Each individual was assigned a condition score based on: 1) carapace color, which darkens as the carapace erodes ; 2) the amount of mucus present, which deters fouling organisms; and 3) the degree of pitting of the carapace, which is caused by chitinoclastic bacteria and 4) the condition of their lateral eyes, which can become covered by epibionts or deteriorate due to algal or bacterial infection (modified from previous studies; for complete methods see Brockmann and Penn 1992 Brockmann 1996 2002 ) The second prediction of the agility advantage h ypothesis is that satellite males that are able to achieve the under position (Figure 3 1) will be smaller than the attached males with which they are competing To test this prediction I ran a chi square test comparing the size and condition of satellite males that were either third prediction deals with the ability of satellite males to achieve one of the two 1F positions ( one each over the left and right in current canal s of a female ; Figure 3 1). Satellite males in the 1F position s should be smaller than those in less desirable positions. To test this prediction I examined distribution of males that reached the 1F position to determine whether more were smaller (while also accounting for condition ) Furthermore, to determine whether group composition had any effect on which males achieved that position, I ran two nominal logistic regressions; one with size of 1F males (larger or smaller), and another
76 with condition of 1F males (better / equal or w orse condition) as response variables. For both tests, the main effects of the number of 1F males (one or two), and the number of satellites (many, 4 5 or few, 1 3), and their interaction were used as explanatory variables. Loading Constraints Hypothesis If loading constraints have favored the evolution of smaller adult body sizes for males, then this should be reflected in the current mating behavior as size based mate selection (with respect to both CW and b ody mass; Table 3 1). I tested prediction s of this hypothesis using data on mated pairs from Seahorse Key, FL, during the 1995 1997, 2008, and 2009 breeding seasons ( n = 1313 pairs). The f irst prediction of this hypothesis is that there will be size assortative mating between males and females across their range of sizes. To test this prediction I ran a simple linear regression on both the CW and the mass of ma les and females in pairs. I also tested this prediction by creating a random distribution with 10,000 simulated random pairings (with replace ment) using the observed pairs from Seahorse Key (for carapace width only as I did not have the required mass data) and then compared this to the mean observed size ratio of actual pairs. Next, given that most males are substantially smaller than females a second prediction of this hypothesis is that size assortative mating will only be seen in the smallest females, and that these females choose males that are smaller on average (i.e., there is a maximum male size that is allowed by the smallest females) To test this prediction, I divided females into five ordinal categories based on the average rank and frequency of their CW: 1) 17.0 20.7 cm, 2) 20.8 21.6 cm, 3) 21.7 22.4 cm, 4) 22.5 23.3 cm, and 5) 23.4 27.2 cm; and based on the average rank and frequ ency of their mass: 1) 710 1350 g, 2) 1351 1525 g, 3) 1526 1695 g, 4) 1696
77 1920 g, 5) 1921 3040 g. I then used five simple linear regressions for each female size category for both CW and mass to compare average male size among female size ranks. The cost of carrying a male of a particular size, and hence the presence of size dependent mate choice, may depend on differences between males and females in how mass scales with size. Therefore, I examined these allometric patterns using an ANCOVA with mass as the response variable, CW as the predictor variable, and sex as a categorical variable. Nesting Competition Hypothesis The nesting competition hypothesis predicts (Table 3 1) that female size or the degree of SSD will be positively correlated with spawning density among populations. I tested this prediction by conducting a literature search for census and morphological data from as many populations as I could find ( n = 13), but I limited analysis to only those papers or reports that accurately desc ribed methods and results such that density and average female size could be accurately determined (Table 3 2; Appendix C ). These populations overlapped with those used in the OSR test, but were not the exact same ones. I ran a simple linear regression w ith density of nesting females as the explanatory variable and mean body size of females as the response variable. Values for density of nesting females were log transformed to meet the assumptions of heterogeneity of variance and a normal distribution. T he second prediction of the nesting competition hypothesis is that among tides in a single population, smaller females will be more prevalent on tides with lower nesting density to avoid competition (assuming females respond flexibly; Table 3 1). I tested this prediction using census and morphological data collected from four spawning seasons (fall 2007, spring and fall 2008, and spring 2009) at Seahorse Key, FL.
78 Censuses were conducted on a 1,000 m stretch of beach over a 30 min period at the time of the high tide. I recorded the presence of all individuals their body size (maximum carapace width) and the size of the mating group for all animals encountered. I collected census data by walking up a 1,000m stretch of beach and counting all animals and th e frequency of each mating group size (e.g., pair, pair + one satellite, lone male). From this census data, I calculated density for each tide I then used statistical software package JMP to rank density levels into 18 relative categories (e.g., level 1 = 1 50 females, level 4 = 595 775 females, level 18 = 3,865 females). I chose to use an ordinal measure of density, rather than a continuous measure because some tides had only slight variation between them. I tested this prediction using an ANOVA with size as the response variable and density level as the predictor variable. Fecundity Advantage Hypothesis The fecundity advantage hypothesis predicts (Table 3 1) that larger females will lay more eggs compared to smaller females, and that larger females wi ll contain more mature eggs within their oviducts than smaller females. Two previous studies have examined the relationship between body size and fecundity of horseshoe crabs. At Seahorse Key, FL, Schwab ( 2006 ) followed individual females ( n = 68) from the time they first arrived on nesting beaches unt il they completed egg laying and returned to the ocean. She recorded the body size of females (inter ocular distance, cm), and marked each location where females deposited clutches of eggs. Six h ours after spawning, she dug up each clutch, and measured t he volume (ml) of eggs laid ( each mL contains approximately 88 +/ 15 eggs; Brockmann 1990 ) She found a positive relationship between body size and the number of eggs laid in 2004, but not in 2005 (likely due to differences in methods; i.e., females were not obser ved for the entire length of the
79 breeding bout). In Cape Cod, MA, Leschen et al. ( 2006 ) showed that larger females carried more mature eggs within their bodies than did smaller female during the peak of the spawning season (14 May 3 June); but at the end of peak spawning (16 17 June) females contained far fewer eggs, and there was no relationship between female size and egg number. Here I test the fecundity advantage hypothesis by reexamining the data from Schwab ( 2006 ) and take potential differences in egg number througho ut the breeding season into consideration. I ran four simple linear regressions with body size (inter ocular distance in cm) as the predictor variable and the amount of eggs laid (ml) as the explanatory variable. The first two regressions were for early in the breeding season (late February end of March) in 2004 and 2005; and the second two were for late in the breeding season (early April until the end of May) in 2004 and 2005. I then use a simple quantitative model using data from the 2004 early season to explore the effects body size on fecundity and lifetime reproductive success. For this model, I used the regression equation for the relationship betwe en female size (IO) and total mL eggs laid. I then calculated the e stimated mean increase in fecundity ( m ) using the mean slope ( M m ) ; and the range of increase in fecundity ( r ) using SE of slope ( M e ) : ( 3 1 ) ( 3 2 ) w here is female size (IO), b is the intercept, c is the number of eggs per ml, and t is the number of tides on which a female returns. I then used E quation 3 1 to calculate an
80 estimated increase in lifetime reproductive success ( number of surviving offspring ) from the smallest to largest female size ( F ma x ) : ( 3 3 ) w here y is the number of breeding seasons, and is juvenile mortality rate. This simple model assumes no decline in fecundity with age, no adult mortality, and no additional cost s when reproduction is increased. Results Competitive Displacement Hypoth esis I failed to detect a difference in isotope values between the feces of males and females for 13 C (2 tailed T test: df = 56, T = 0.4, P = 0.727), or for 15 N (2 tailed T test: df = 55, T = 1.6, P = 0.128; Figure 3 2). For claw chitin, I failed to detect a difference in 15 N values between males and females (2 tailed T test: df = 37, T = 1.6, P = 0.116). However, females had statistically higher 13 C values compared to males (df = 37, T = 2.5, P = 0.015; Figure 3 2). Agility Advantage Hypotheses There was no relationship between male size and OSR ( r 2 = 0.0 1 F 1, 11 = 0. 1 P = 0. 0.731; Figure 3 3a) or between OSR and the degree of sexual size dimorphism among populations ( r 2 = 0.0 1 F 1, 11 = 0. 1 P = 0. 712; Figure 3 3 c) Within the Seahorse Key population, size did not influence the ability of a satellite male to maneuver under the attached male (Likelihood ratio: n = 54, df = 3, x 2 = 1.5, P = 0.688; Table 3 3 ) n or did size influence the ability of a satellite male to reach the most desired positions ( x 2 = 0.0, P = 1.0): an equal number of satellite males in the 1F position were larger ( n = 130) and smaller ( n = 130) than males in lower ranked positions (Table 3 4) However,
81 condition was important: more males that were in the 1F position were in better or equal condition (79%) when compared with satellite males in other positions (likelihood ratio 2 = 89.5, P < 0.0001). G roup composition affected which males achieved 1F position s. The whole nominal logistic regre ssion model for condition of 1F males was significant (df = 3, 2 =19.5, P = 0.0002). The main effect of the number of other satellites present was not influential ( 2 = 0.6, P = 0.457), but whether there were one or two 1F males present was ( 2 = 12.2, P = 0.0005), as was the interaction between these variables ( 2 = 12.0, P = 0.0005 ). The whole nominal logistic regression model for size of 1F males was significant (df = 3, 2 =19.5, P = 0.0002), as were all three other variables: the main effect of the number of other satellites present ( 2 = 8.5, P = 0.004 ), whether there were one or two 1F males present ( 2 = 12.9, P = 0.0003), and the interaction between them ( 2 = 9.0, P = 0.0028 ). In other words, when both 1F positions were occupied, there was no difference in the size or condition of males that reached those desired positions, compared to other satellite males, regardless of whether there were few or many other satellites present When only one 1F position was occupied, however, those males tended to be in better condition than other satellite males if mating group sizes were small (1 3 other satellites). But, if group size was large (4 5 other satellites present), males in the 1F position tended to be larger not smaller, than other satellites. Loading Constraints Hypothesis The regression comparing the relationship between female and male CW for all individuals was weak, but statistically significant ( r 2 = 0.02, F 4, 1309 = 28.3, P < 0.0001). However, when analyzed by female size category, I found a relationship only in the
82 smallest females (category 1; r 2 = 0.03, F 2, 267 = 8.1, P = 0.005). For mass there was also a weak but positive relationship when comparing all indiv iduals ( r 2 = 0.01, F 4, 806 = 10.5, P = 0.001). When analyzed by female size category, no relationship for categories 1 2, or 4 5 ( r 2 < 1.2, 0.281). For category 3, the relationship was not statistically significant at = 0.05 ( r 2 = 0.02, F 1, 157 = 3.3, P = 0.073), but may have caused the positive overall relationship. For the random pairing simulation, t he mean female to male size ratio SD (CW only) of the distribution generated by 10,000 random pairings ( mean = 1.366 0.0004 ) was not different ( P = 0.610 ) from the mean size ratio of actual pairs (mean = 1.3 65 0.1 2; Figure 3 5). Allometries of mass for males and females (Figure 3 6 ) were very different (ANCOVA: R 2 = 0.95, F 3, 1613 = 10,265, P < 0.0001; sex, CW and Sex x CW wer e much greater increase in mass, compared to a similar increase in CW of males. Nesting Competition Hypothesis I failed to detect a relationship between female nesting density and average female size across populations ( r 2 <0.0 3 F 1, 11 = 0. 3 P = 0. 586 ; Figure 3 3b) or between female nesting density and the degree of sexual size dimorphism ( r 2 <0.0 2 F 1, 11 = 0. 3 P = 0. 617 ; Figure 3 3 d I also failed to detect a relationship between female nesting density and size of females among tides within a season (Figure 3 7 ) for any of the years I examined (fall 2007: df = 148, T = 1.3, P = 0.186; spring 2008: F 7, 341 = 0.5, P = 0.865; fall 2008: df = 176, T = 0.2, P = 0.845; spring 2009: F 7, 332 = 1.3; P = 0.237). Fecundity Advantage Hypothesis I found a positive relationship between inter ocular distance and the volume of eggs laid early in the breeding season for both 2004 ( r 2 = 0.21, F 1, 33 = 8.8, P = 0.006)
83 and 2 005 ( r 2 = 0.07, F 1, 54 = 3.8, P = 0.056; Figure 3 8 ). However, I failed to find a relationship later in the season in either 2004 ( r 2 = 0.05, F 1, 29 = 1.4, P = 0.243) or 2005 ( r 2 = 0.04, F 1, 35 = 1.3, P = 0.255; Figure 3 8 ). The quantitative model estimat ing increases in fecundity as a function of female size showed definitive advantages for larger females (Table 3 5). The regression equation was : (3 4) where E is the t otal mL of eggs and s is female size (inter ocular distance in cm). I assumed 88 eggs per mL (Brockmann 1990), 6 8 years of reproduction ( Ropes 1961 Botton and Ropes 1988 Brockmann and Johnson 2011 ) and a mortality r ate of 0.00003 ( Botton et al. 2003 ) I calculated maximum increases by comparing the difference between the largest (18 cm) and smallest (11 cm) sizes. Across the largest size range, I found a mean increase in fecundity ( Equation 3 1) of 388% for any given number of returning tides during a single breeding season. The increase in lifetime reproductive success (Equation 3 3) varied depending on the number of tides on which a female returns ( t ) and the number of reproductive years ( y ) : when t = 3, there was an increase of 6.3 offspring for y = 6 and 8.4 offspring for y = 8; when t = 4, there was an increase of 8.4 offspring for y = 6, and 11.2 offspring for y = 8; when t = 5 there was an i ncrease of 10.5 offspring for y = 6, and 13.9 offspring for y = 8 (Table 3 5). Discussion Overall, I found little support for many of the hypotheses regarding the evolution and maintenance of sexual size dimorphism in horseshoe crab s. However, I did find support for the fecundity advantage hypotheses, and a trend for support for the loading
84 constraints hypothesis suggesting that these particular selection pressures favor larger females and smaller males. I found strong support for the fecundity advantage hypothesis. During peak spawning, larger females laid more eggs than smaller females, but by the end of the spawning season there was no relationship between female size and fecundity suggesting that females have largely spent their supply of eggs by the end of the season My simple quantitative model showed substantial increases in fecundity and lifetime reproductive success with an increase in female size. However, the estimates generated here are likely biased because I ass umed no decline in fecundity with age, no adult mortality, and no cost of increased reproduction. Previous research also lends support to this hypothesis. L arger females contain more eggs within their bodies ( Leschen et al. 200 6 ) ; larger females attract more satellite males ( Brockmann 1996 ) and satellite males are more attracted to cement models of females that were larger ( Schwab and Brockmann 2007 ) While larger females did not lay more eggs per tide ( Brockmann 1996 ) they may still have an advantage by nesting for more nights in total or have less of a decline in clutch size over the course of the season compared to smaller females. Lastly, a positive relationship between f ecundity and size has been found in the other species of horseshoe crabs ( Chatterji and Parulekar 1992 Chatterji 1995 Khan 2003 ) Taken together, these results support the fecundity advantage hypothesis that selection favors larger females because of their increased egg production. I did not find support for the predictions that males and females eat different types of food, or that they forage in different locations. The lack of a difference in 15 N values
85 between males and females in both the feces and claw chitin tissues indicate s that males and females consume foods at the same trophic level during the spawning season at Seahorse Key, FL. My evidence accords with previous studies based on gut contents that found males and females along the New Jersey coast ate qualitatively simi lar diets ( Botton 1984 ) I found no difference in 13 C values in the feces of males and females, but for claw chitin females had higher levels of 13 C. This difference was Differences between inshore and offshore food webs are supported only when values of 13 C are much larger ( 4 ) Hence, my results also indicate that male and females are not likely foraging in different locations during the year prior to adulthood at Seahorse Key, FL. One important shortcoming of this method is that stable isotopes cannot detect differences in consumption of different prey at a similar trophic level, or of the size of prey. Because females are larger, they might be able to feed on larger prey ( Botton 1984 ) However, in feeding trials, Botton ( 1984 ) found no difference between males and females in their preference of bivalve prey size. Both sexes preferred the larger Mulinia lateralis (4 10mm) to the smaller Gemma gemma (2 4mm). Taken together the evidence does not support the competitive displacement hypothesis for the evolution of SSD in horseshoe crabs. Overall, I found that condition played an important role in male male competition (within the Seahorse Key population better condition mal es were better at occupying the 1F position), however, size does not appear to be an important factor. Across populations, increased mating competition (OSR) was not correlated to smaller male size or a greater degree of SSD; and within the Seahorse Key p opulation, the body size
86 of individuals appears to be important only under the unusual circumstance (3.5% of my sample) when there is only one 1F male, but many other satellite males In these circumstances the male in the 1F position actually tended to b e larger than the other satellite males. A size advantage for large males in mate competition does not inevitably lead to male biased SSD, and in taxa with female biased SSD there can still be selection for larger males due to competition for mates ( Greenwood and Adams 1987 ) Thus, mating competit ion could actually favors large male size, as seen in taxa with male biased SSD, such as the majority of mammalian orders ( Lindenfors et al. 2007 ) and most avian species ( Szkely et al. 2007 ) as opposed to the selection for small sizes which I tested However, my tests of the agility advantage hypothesis would have revealed either a small male or large male size advantage, but I did not find evidence supporting either scenario. Hence, I find little support for the agility advantage hypotheses in the evolution of SSD in horseshoe crabs. While there were no instances of size reversal, the relation ship between the size of females and their attached males was quite low for both CW and mass, explaining only 2% of the variation. Even among females in the smallest size category (CW), where there appears to be some size dependent mating, female size explained only 3% of the variation in male size. Moreover, the distribution of observed mated pairs was not different than random. These results suggest that th ere does not tend to be size assortative mating overall, but a slight fraction of small females do have smaller males. The difference in how mass scales with size between males and females may explain m to matter for most pairings. Fo r males there is only a slight increase in mass with carapace width; whereas for females, a slight increase in
87 carapace width is accompanied by a much heavier mass. Therefore, all but the smallest females far outweigh males, and increases in costs of car rying a slightly larger male are likely to be minor. While there may be some female choice for smaller males, other factors may limit this selection. For example, t here is some evidence that smaller males have lower concentrations of sperm (D. Sasson, Pe rsonal Communication) selecting for large males and so there may be a tradeoff between the energetic cost of carrying males and fertilization success. Overall, I do not find strong support for loading constraints favoring smaller males in horseshoe crabs I did not find support for the first prediction of the nesting competition hypothesis: there was no relationship between nesting density and female size across populations. In some populations, horseshoe crabs were historically harvested for use in the fertilizer industry. Presently they are harvested as bait in eel and whelk fishing industries, and for use in the biomedical industry ( Kreamer and Michels 2009 ) Both of these industries tend to prefer harvesting larger females, and so my measures of mean female size and nesting density may not accurately reflect evolved sizes and densities in exploited populations Yet, I also did not find support for the second prediction that smaller females would be more likely to be present on less dense tides suggesting that smaller females do not favor spawning during tides when fewer females are present This could be becau se they are either not able to assess density prior to spawning, they are not able to respond flexibly to density or because nesting density simply does not affect their ability to lay eggs in the preferred locations In places like Delaware Bay, spawning densities can be much higher than at Seahorse Key, FL, and it would be interesting to see if smaller females alter the timing of spawning in high density
88 locations in DE Bay Taken together, my results do not suppo rt the nesting competition hypothesis as a likely explanation for the maintenance or evolution of SSD in this species. I did not find support for any of the hypotheses regarding selection for male size. But given that there is SSD in all populations, I ex pect some cost to being large for males. The protandry hypothesis, which I did not evaluate in this study, pertains to the body size and timing of when males undergo their final molt to adulthood. The protandry hypothesis states that there is selection f or males to mature earlier at a smaller size because of increases in fitness. This can be due to: 1) a higher lifetime reproductive success (one more year of reproduction) when maturing earlier ( Wiklund and Fagerstrom 1977 Fagerstrom and Wiklund 1982 Wiklund and Solbreck 1982 Gunnarsson and Johnsson 1990 ) ; and 2) a demographic advantage since their offspring are born earlier and start reproducing e arlier ( Stearns 1992 ) The magnitude of demographic advantages is especially strong in organisms with late ages at fir st reproduction ( Stearns 1992 ) There are few direct tests of the protandry hypothesis because manipulating size at ma turity is not feasible in most system s However, given that there does not appear to be a large male size advantage in horseshoe crabs, protandry would seem like a reasonable hypothesis. Future studies should use a dynamic state variable modeling approac h to explore the fitness consequences of m aturing earlier or later. In conclusion, I found evidence in support of the fecundity advantage hypothesis The benefits of increased fecundity favor later maturation for females at a larger size Additionally, t here may be weak selection for larger female size and smaller male size
89 due to loading constraints. The protandry benefit to males was not evaluated in this study, however, this appears to be a good hypothesis to explain the costs of males delaying maturi ty by one additional year, and may explain selection for smaller male size (minus some cost of reduced sperm concentrations). My study illustrates the importance of investigating multiple hypotheses for both males and females in understanding the ultimate factors underlying the ubiquitous pattern of sexual size dimorphism in animals.
90 Table 3 1. Hypotheses and predictions for the evolution of sexual size dimorphism in horseshoe crabs. 1 The protandry hypothesis, and the dynamic model pred icting optimal emergence time are not evaluated in this study. Males Females Both sexes Predictions Agility advantage Protandry 1 Loading constraint Nesting competition Fecundity advantage Competitive displacement Selection for smaller or larger size Smaller Smaller Smaller Larger Larger Either Males & females eat different food ( 15 N values) No No No No No Yes Males & females forage in separate areas ( 13 C values) No No No No No Yes Satellite Males in position 1 will be smaller than those in other positions Size is ( ) correlated with OSR across populations Yes No No No No No Size is ( ) correlated with SSD across populations Yes No No No No No Small satellites achieve better position relative to attached Yes No No No No No Satellites in 1F position smaller compared to other satellites Yes No No No No No Dynamic model predicts males should emerge earlier 1 Yes Yes Yes No No No Size assortative mating No No Yes No No No Smaller size ratio in spawning pairs compared to random No No Yes No No No Smaller females more prevalent on tides with lower density No No No Yes No No Density (+) correlated with size or SSD among populations No No No Yes No No Larger females contain more eggs within their bodies No No No No Yes No Larger females lay more eggs No No No No Yes No
91 Table 3 2. Data for the comparison of male size and operational sex ratio (OSR; # males : females), and female size and nesting density (# females / m2) for horseshoe crabs across 13 locations in North America. 1 15 See Appendix C for references Variance not available Â§ Male range = 15.9 22.4; Female range = 20.1 30.0 Male range = 10.0 30.0; Female range = 15.0 37.0 State Locatio n Male size n Female size n SSD OSR Density Maine 1 Hog Bay 14.7 951 17.4 337 1.184 2.8 0.057 Maine 1 Bagaduce River 13.9 9 17.5 12 1.259 1.07 0.056 Maine 1 Thomas Point Beach 14.8 26 19.2 14 1.297 1.7 0.176 Massachusetts 2, 3 Cape Cod Bay 17.4 1.6 2,942 22.7 1.9 759 1.305 2.9 0.024 Massachusetts 3 Nauset Estuary 17.5 1.7 433 23.4 2.1 256 1.337 1.6 0.032 Massachusetts 3 Monomoy NWR 18.8 1.5 909 24.2 2.2 477 1.287 1.9 0.029 Massachusetts 3 Pleasant Bay 17.3 1.2 2,063 22.5 2.0 413 1.301 5.8 0.091 Rhode Island 4, 5 Narragansett Bay 18.6 Â§ 54 24.0 Â§ 288 1.290 2.78 0.0667 Connecticut 4, 5 Milford 19.5 1,760 24.9 1,145 1.277 2.08 0.0153 Delaware & New Jersey 6, 7, 8, 9, 10 Delaware Bay 20.7 1.4 1,233 26.6 1.9 803 1.287 3.5 1.08 South Carolina 11, 12 Otter Island 23.6 1.4 406 30.7 2.1 418 1.302 1.96 0.068 Florida 13 Ochlockonee Bay 16.9 1.3 1,552 22.3 2.2 742 1.324 3.56 n/a Florida 14 Seahorse Key 16.1 1.1 3,556 21.9 1.6 2,300 1.360 2.94 0.0717 Florida 15, 16 Indian River Lagoon 13.6 n/a 18.9 n/a 1.390 n/a 0.219
92 Table 3 3. T he number of 1F satellite males that obtain ed either the under ( desired ) or the over position with respect to their body size and condition relative to the attached male Size and condition of 1F male relative to attached male Number of males in: Over position Under position Larger, better or (=) condition 5 6 Larger, worse condition 3 7 Smaller, better or (=) condition 15 14 Smaller, worse condition 2 2
93 Table 3 4. The number of satellite males that obtained the desired position near a female (1F position), with respect to their body size and condition relative to other satellites in the mating group. Size of 1F male relative to other satellite males Condition of 1F male relative to other satellite males Better or (=) Worse Larger 96 34 Smaller 108 22
94 Table 3 5. Comparison of the estimated effects of body size on fecundity for 3 5 tides and 6 8 reproductive years Estimates use data from early 2004 for the relationship between inter ocular distance (IO) and th e number of eggs laid. T otal mL of eggs laid = ( 173.5 + 18.86 6.3 x female inter o cular distance) x 88 eggs per mL ( Brockmann 1990 ) Increase in lifetime reproductive succe ss = total increase in eggs x 6 8 reproductive years x 0.00003 survival rate. Number of Tides on which a female returns to spawn Three tides Four tides Five tides Inter ocular distance (cm) Estimated Mean Estimated Range Estimated Mean Estimated Range Estimated Mean Estimated Range 11 8,972 n.a. 27,259 11,963 n.a. 36,346 14,954 n.a. 45,432 12 13,952 n.a. 33,901 18,602 n.a. 45,202 23,253 n.a. 56,503 13 18,931 n.a. 40,544 25,242 n.a. 54,058 31,552 n.a. 67,573 14 23,911 636 47,186 31,881 848 62,915 39,852 1,060 78,643 15 28,891 3,953 53,828 38,521 5,271 71,771 48,151 65,88 89,714 16 33,870 7,270 60,470 45,160 9,693 80,627 56,451 12,117 100,784 17 38,850 10,587 67,112 51,800 14,116 89,483 64,750 17,645 111,854 18 43,830 13,904 73,754 58,440 18,539 98,339 73,049 23,174 122,924 Increase in egg # 34,858 46,477 58,095 x 6 reproductive years 209,148 278,862 348,570 x 8 reproductive years 278,864 371,816 464,760 Increased success at 6 reproductive years 6.3 8.4 10.5 Increased success at 8 reproductive years 8.4 11.2 13.9
95 1 2 Figure 3 1. Mating groups in horseshoe crabs. A) T he over position of a satellite male 3 relative to the attached male. B) T he under position of a satellite male 4 relative to the attached male. C) T he desired 1F position of satellite males, 5 compared to the other satellite positions that confer less paternity compared 6 to the 1F position. F rom ( Brockmann 1990 Brockmann et al. 1994 ) 7 illustrations by Daryl Harrison. 8 9
96 10 Figure 3 2. Mean 95% CI stable isotope values for (A) claw chitin of male (n = 20) 11 and female (n = 19) horseshoe crabs from Seahorse, Key, FL. Stable isotope 12 values for (B) feces of a different subset of males (n = 38) and females (n = 13 19) from Seahorse, Key, FL. 14 15
97 16 17 Figure 3 3. A. Average size (carapace width in cm) of male h orseshoe crabs, and (C) 18 degree of sexual size dimorphism as a function of operational sex ratio in 13 19 locations across North America. B ) Average size of female horseshoe crabs 20 and (D) degree of sexual size dimorphism as a function of nesting density in 21 13 locations across North America (see Appendix C for locations and 22 sources) 23 24
98 25 26 27 Figure 3 4. Relationship between female and male sizes for all observed mating pairs of 28 horseshoe crabs collected from Seahorse Key, FL from 1995 1997 and 2008 29 2009 (n = 1,311). Regression lines are for five different female size 30 categories (1 = smallest, 5 = largest). The regression for the smallest 31 females (category 1) were significant, all others were non significant. 32 33
99 34 35 Figure 3 5. Probability density for female : male size ratios (carapace width in cm) of 36 horseshoe crabs from Seahorse Key, FL. Red dashed lines indicate the 37 observed mean size ratio SD for actual amplexed pairs found spawning on 38 the beach in 1995 1997 and 2008 2009 (n = 1313; mean = 1.365 0.122). 39 The back curve represent a distribution from 10,000 simulated random pairs 40 (n = 10,000; mean = 1.366 0.00036). The p value is the percentile to which 41 the observed mean corresponds in the simulated distribution. 42 43
100 44 Figure 3 6. A llometries of mass and carapace width for male and female horseshoe 45 crabs collected from Seahorse Key, FL in 1995 1997, and 2008 2009. 46 47 48
101 49 Figure 3 7 Relationship between female nesting density and female size of horseshoe 50 crabs across tides over f our seasons at Seahorse Key, FL. 51 52
102 53 Figure 3 8 Relationship between size and fecundity of female horseshoe crabs on 54 Seahorse Key, FL in 2004 and 2005. A B) Early in the breeding season was 55 from late February until the end of March. C D) Late in the breeding season 56 was from early Apri l until the end of May. Each mL contains approximately 88 57 +/ 15 eggs; ( Brockmann 1990 ) 58 59
103 CHAPTER 4 60 COSTS OF ALTERNATIVE MATING TACTICS TO FEEDING IN MALE HORSESHOE 61 CRABS 62 Introduction 63 Alternative mate searching tactics by males are common in competitive mating 64 systems ( Taborsky et al. 2008 ) 65 silently near larger vocalizing males and either intercept females that are attracted to 66 the callers ( Wells 1977b a 67 Robertson 1986b a ) The evolution and maintenance of such discrete alternative 68 tactics is puzzling because if one tactic is only slightly less successful than the other, it 69 should be eliminated by selection ( Brockmann 2001 ) In some cases, alternative tactics 70 are maintained as a genetic polymorphism, but in most cases they depend on the 71 individual's phenotype (e.g., body size, condition) and the circumstances in which they 72 live (e.g., population density, sex ratio; Gross 1996 Zamudio and Chan 2008 ) In order 73 for alternative tactics to be maintained at high frequencies, conditions must exist under 74 which each tactic is more successful than the other ( i.e., fitness curves must cross; 75 Brockmann and Taborsky 2008 ) This means that trade offs are inherent to alternative 76 tactics. For example, an animal simply cannot simultaneously both call and sneak, or 77 both maintain a territory an d disperse widely in search of females. As a consequence, 78 the phenotypes that maximize fitness for one tactic are different from those that 79 maximize fitness for the other tactic. Therefore, the decision about which tactic to follow 80 is based not only on a it live s 81 but also on the costs and benefits of the alternative tactic s Understanding the nature of 82 the trade offs for each tactic is vital to understanding why particular tactics take the form 83 that they do, and to understanding the evolution and maintenance of alternative mating 84
104 tactics within populations ( Brockmann et al. 2008 ) In this study, I examine the trade 85 offs associated with alternative mating tactics of male horseshoe crabs. 86 Horseshoe cra bs have a highly competitive, explosive mating system ( Brockmann 87 1990 ) in which males exhibit two condition dependent, alt ernative mating tactics 88 ( Brockmann and Penn 1992 Brockmann 2 002 ) Younger males in better condition 89 attach to females at sea and arrive on spawning beaches paired in amplexus with 90 females. The attached male remains with the female until she has completed egg 91 layi ng for the season, and then detaches and seeks another female. Unattached, older 92 males in poorer condition roam the shoreline, join spawning pairs as satellites on the 93 beach, and engage in sperm competition with attached males and other satellite males 94 ( Brockmann and Penn 1992 Brockmann et al. 1994 Brockmann 1996 2002 ) These 95 behavioral differences are not just a consequence of a male being unable to locate or 96 hold onto a female, but instead result from an evolved decision rule based on age or 97 condition ( Brockma nn 2002 ) That is, individuals maximize fitness by switching tactics 98 at a given age or conditio n (i.e., fitness curves cross). 99 Several trade offs involving differences in paternity and righting behavior have 100 been identified for each tactic. During each 1 week spawning cycle, attached males 101 normally mate with only one female, whereas satellite males may join several pairs. 102 During each mating bout in a spawning cycle, satellite males have similar paternity 103 success compared to attached males ( Brockmann et al. 2000 ) but attached males do 104 not always compete for paternity with satellites (depend ing on the number of unattached 105 males present). In contrast, satellite males must always engage in sp erm competition 106 ( Brockmann et al. 1994 Brockmann et al. 2000 ) Satellites appear at the beach to mate 107
105 more often than attached males ( Brockmann and Penn 1992 ) but they do not always 108 find a mating pair. Lastly, horseshoe crabs are often overturned on the beach, leaving 109 them vulnerable to desiccation and predation. When this happens, a ttached males are 110 better able to right themselves than satellite males ( Penn and Brockmann 1995 ) Taken 111 together, attached males have higher mating suc cess overall ( Brockmann et al. 2000 ) 112 and they are less likely to become stranded than unattached (satellite) males. 113 However, if an age or condition threshold for switching tactics has evolved that 114 maximize s fitness, then I expect that there would be compensating costs associated 115 with the attached tactic. In this study I investigate a previously unexplored, potential 116 cost to the attached male tactic: nutritional stress caused by a reduced ability to feed 117 while attached to a female. 118 Adult horseshoe cr abs feed on a variety of items (e.g., bivalves, polychaetes, 119 crustaceans) by digging into the substrate, stirring up sediment with their walking legs, 120 and grasping food and directing it to their ventral mouth with their chelae and chelicerae 121 ( Botton and Haskin 1984 Botton and Shuster 2003 ) Gnathobases (leg bases) that 122 surround the mouth ma cerate the food and also help to manipulate food into the mouth 123 where it is then drawn into the esophagus. During the breeding season, attached males 124 hold onto the posterior opisthosomal spines of the female using a modified pair of 125 pedipalps. As a resul 126 female's telson ( Brockmann 2003 ) The amount of time that males remain attached to 127 females, and the subsequent potential reduction in feeding, varies within and between 128 populations. In Florida males typically remain attached for a one week spawning cycle 129 ( mean length of attachment is 3.7 6.1 SD days; Brockmann and Penn 1992 ) but 130
106 occasionally may stay attac hed up to 51 days ( Brockmann 2003 ) In other populations 131 they can remain attached for much longer ( Shuster 19 54 ) ; for example, in New England 132 attached pairs have been observed overwintering together ( Barlow et al. 1987 Moore 133 2004 ) Thus, my reduced feeding hypothesis states that a cost to the attached tactic is 134 that males are severely restricted in their ability to feed Alternatively attached males 135 may use much less energy than satellite males, and so they may not suffer any 136 nut ritional stress when attached. 137 I tested multiple predictions and assumptio ns of the reduced feeding hypothesis 138 (Table 1). To determin e if attached males eat less than satellite males, I first conducted 139 field experiments to: 1) measure the amount of feces produced by attached and satellite 140 males, 2) detect any potential differences in food transit time s between tactics, 3) 141 compare gut fu llness and gut contents between tactics, and 4) test for nutritional stress 142 in attached males with stable isotope ( 15 N) analysis of feces. I then conducted a lab 143 experiment to confirm that starvation increases 15 N values of feces in horseshoe crabs 144 (as assumed in objective #4) This study may be the first to illuminate an explanation 145 for why the alternative reproductive tactics in this system take the form they do. 146 Methods 147 I conducted this study during 2008 2011 at the University of Florida Seahorse 148 Key Marine Laboratory. Seahorse Key is a 67 h a island that is part of the Cedar Keys 149 National Wildlife Refuge along the northwestern Gulf coast of Florida. I collected adult 150 horseshoe crabs as they initially came to the beach to spawn during evening high tides. 151 Each animal was marked with an uniquely numbered thumb tack, placed immediately 152 into a clean bucket, transported to the laboratory on the island, and placed into a 153
107 separate, randomly assigned holding tank. The holding tanks were 61 cm x 61 cm x 2 0 154 cm deep, and fed by a flow through, running seawater system. For all animals used in 155 my experiments I measured their body size (maximum carapace width in cm) and 156 as sessed their condition using an index based on visual inspection of their carapace. 157 Eac h individual was assigned a condition score based on : 1) carapace color which 158 darkens as the carapace erodes ; 2) the amount of mucus present which deters fouling 159 organisms ; and 3) the degree of pitting of the carapace which is caused by 160 chitinoclastic b acteria (modified from previous studies; for complete methods see 161 Brockmann and Penn 1992 Brockmann 1996 2002 ) Previous studies have shown 162 that attached males are in better condition than satellite males, but there are no 163 differences in carapace width (CW) between the two groups ( Brockmann and Penn 164 1992 Brockmann et al. 1994 Brockmann 1996 2002 ) ; these patterns were also 165 supported for the animals used in this study. 166 Measuring Waste Production 167 The first prediction of the reduced feeding hypothesis is that if the attached tactic 168 inhibits feeding, then attached males will not defecate at all, or will produce less than 169 satellite males (Table 4 1) I conducted an experiment to test this prediction fr om 4 17 170 October 2008, and 10 1 4 March 2009. Each experimental replicate consisted of three 171 animals collected from the beach at the same time: a female, her attached male, and 172 one satellite male associated with this pair ( n = 27 replicates and 81 individual 173 horseshoe crabs). In the running seawater system where I conducted this experiment 174 seawater is pumped into a large holding tank before entering the individual tanks and 175 water runs out of each individual tank through an overflow pipe. Hence, the relatively 176 heavy packet of waste produced by horseshoe crabs did not flow out of the individual 177
108 tanks H owever, some debri s did flow in which could impact my measure of waste 178 produced. So, in addition to the three tanks holding an attached male, a satellite male, 179 and a female, I added a fourth empty tank as a control for each replicate. I left all 180 individuals in the holdin g tanks without food for 12 h r before collecting waste. I chose 181 this time period somewhat arbitrarily because the average food passage time of 182 horseshoe crabs was not known. T his time period seems reasonable because after 12 183 h r all animals had produced w aste. After the crabs had been returned to the ocean, I 184 siphoned out all visible waste and debris from each holding tank through a fine mesh 185 plastic filter. I then rinsed off this plastic filter and collected the sample remaining on a 186 piece of filter pap er. Each sample was dried for 4 h r in an oven at 60 o C, and weighed 187 ( in g minus th e weight of the filter paper). 188 Body size was positively correlated with the amount of feces produced ( linear 189 regression: r 2 = 0.25, F 1, 78 = 26.0, P < 0.0001). Therefore, I applied a size correction to 190 my measure of the waste production: ((log was te control) / log CW)*100). I compared 191 the amount of waste (minus the amount of debris found in the control tanks) that was 192 produced among the groups with paired t tests of: sate llite attached, female attached, 193 female satellite. 194 One assumption of this waste production prediction is that differences in condition 195 between attached and satellite males do not affect assimilation efficiency, and 196 consequently the amount of feces de fecated. If this a ssumption is not correct, then 197 there should be a relationship between feces mass and condition for both males and 198 females. To test this assumption, I analyzed the influence of condition on the amount of 199 feces produced with two ANOVAs: o ne for males and a second for females. 200
109 Measuring Food Transit Time 201 My ability to detect differences in waste production assumes that food passage 202 times are the same for attached males and satellite males. F ood passage time can be 203 influenced by the size of the animal's digestive tract, by the amount type, and quality of 204 food intake ( Barboza et al. 2009 ) Thus, I test ed 205 th e assumption of equal passage times by c onducting an experiment from 28 30 206 March and from 12 2 7 April 2010 that compared passa ge times between male tactics 207 (females were also tested for comparison; Table 4 1). The passage of digesta is 208 commonly 209 (e.g., carmine red dye ) to first appear in feces ( Karasov and 210 Martinez del Rio 2007 Barboza et al. 2009 ) I collected animals for this experiment ( N 211 = 16 for attached males, 15 for satellite males, and 14 for females), and then left them in 212 the holding tanks for 12 h r before the experiment began so that all had defecated prior 213 to the start of the experiment I cut a large, fresh shrimp into 10 equal pieces (1.07 214 0.08 g) and soaked it in 2.1 g carmine re d with 1 2 mL water for 10 30 min. I fed each 215 animal by taking it out of the holding tank, turning it ventral side up on a table and 216 placing individual pieces of shrimp in its mouth, ad libitum for 20 min. I then checked 217 the animals every 3 hr until I ob served the red dye in their feces. I chose this 3 h r 218 interval based on observations from the waste production experiment. This interval was 219 not short enough to detect a difference of 1 2 h r between tactics, but was appropriate 220 for confirming that I would have detected differences in the 12 h r waste production 221 experiment. 222 I wanted to compare the transit time among my three groups, but I also wanted to 223 224
110 transit time. Therefore, I conducted an ANCOVA with transit time as the response 225 variable, status (attached, satellite or female) as the explanatory variab le, and the 226 amount of shrimp consumed and condition as covariates. I was also interested in 227 whether attached males ate more than satellite males during this feeding, so I 228 compared the amount of shrimp eaten among the three groups with an ANOVA, and 229 least squares means contrasts to identify spec ific differences among groups. 230 Stable Isotope Analysis of Feces 231 Attached males may eat and defecate less, however, they may also use much less 232 energy than satellite males because females are the ones exerting energy in locomotion 233 and digging while spawning and during circatidal movements to and from the beach 234 ) Additionally, unlike satellite males, 235 they are not required to spend energy locating multiple mating grou ps on each tide. As 236 a result, attached males may not suffer any nutritional stress from reduced feeding 237 when attached. Nutritional stress due to a period of fasting ( i.e., when feeding is 238 forgone in favor of other activities; McCue 2010 ) or starvation ( i.e., when feeding is 239 prevented due to so me extrinsic limitation; McCue 2010 ) can be inferred from s table 240 isotope values of animal tissues ( Hobson et al. 1993 Gannes et al. 1997 Gannes et al. 241 1998 Martinez del Rio and Wolf 2005 Castillo and Hatch 2007 McCue 2007 McCue 242 and Pollock 2008 Martinez del Rio et al. 2009 ) If an animal is in a negative energetic 243 balance, 15 N is preferentially retained, while 14 N is excreted. As a result, 15 N values 244 increase in tissues over time as the animal s ( McCue and Pollock 245 2008 ) Increased 15 N values during fasting or starvation occurs in a wide variety of 246 taxa and tissue types ( reviewed in McCue and Pollock 2008 ) including in the excreta of 247
111 lizards ( Castillo and Hatch 2007 ) and rattlesnakes ( McCue 2007 ) It seems paradoxical 248 that 15 N should increase in excreta of starving animals, even though 15 N is 249 preferentially retained and 14 N is excreted. However, the increase in 15 N values of 250 excreta during starvation is t hought to result from the breakdown of structural proteins 251 ( that tend to have higher 15 N values ) that progressively contribute to the pool of labile 252 proteins (i.e., those most readily metabolized to nitrogenous waste), thus becoming the 253 primary source of nitrogen in the excreta ( Castillo and Hatch 2007 McCue 2007 ) 254 Additionally, in horseshoe crabs, nitrogenous waste is converted to ammonia and 255 dumped via their book gills and coxal gland ( Towle et al. 1982 ) rather than in their 256 feces. Thus, a second prediction of the reduced feeding hypot hesis is that values of 257 15 N of any collected feces from attached males will be higher compared to satellite 258 males and females that a re able to feed freely (Table 4 1). 259 To test this prediction, I collected feces from the waste production experiment (see 260 ab ove) for stable isotope analysis (spring 2008 samples only, n = 19 for each group). I 261 removed any sand present in the samples, and then each fecal sample was ground to a 262 homogenous, fine powder using a mortar and pestle. All samples were analyzed by the 263 Stable Isotope Mass Spectrometry Lab in the Department of Geological Sciences at the 264 University of Florida to determine values of 15 I compared 265 stable isotope values among the three groups by conducting a n ANOVA, and then 266 conducte d least squares means contrasts to identify specific differences among the 267 groups. 268
112 Effect of Starvation on Fecal Stable Isotope Value 269 Inferring that differences in 15 N value between attached and satellite males are 270 due to nutritional stress requires demonstrating that fasting or starvation does indeed 271 cause an increase in the 15 N values of feces. Thus, I conducted an experimental test 272 of this prediction by feeding so me crabs and starving others (Table 4 1). On 27 March 273 2011, I collected 20 satellite males from Seahorse Key and brought them to the lab at 274 the University of Florida in Gainesville. Crabs were kept on a 12 hr light / 12 hr dark 275 cycle in a 1,360 L tank of filtered seawater with a salinity of 28 276 temperature (approximately 21 C). Partial water changes were conducted every third 277 day to alleviate nitrate build up. I hand fed all crabs a diet of freshly frozen bay scallops 278 ( Chlamys patagonica ) that were purchased from a local grocery store (Publix brand). 279 Every other day for 4 weeks, each crab was fed ad libitum for 15 min. After this 280 acclimation period, I placed each crab in a 22 L container with an oxygen bubbler for 48 281 hr. I checked the co ntainers every 3 hr and collected any feces that were produced. 282 The feces collected at this time were used to obtain the pre treatment 15 N values. I 283 then randomly chose 10 crabs to receive a feeding treatment, and another 10 crabs for 284 a starving treatme nt. For animals in the feeding treatment, I continued the same 285 feeding schedule as before. Animals in the starving treatment received no food, but 286 were handled exactly as in the feeding treatment to simulate the feeding process. After 287 4 weeks, I obtaine d feces for post treatment 15 N values, by feeding all crabs once ad 288 libitum for 15 min on the last day of the experiment. I then placed them back into 289 individual containers for 48 hr and collected all feces. Each fecal sample was placed 290 onto a coffee fi lter and dried for 4 hr in an oven at 60 o C. The samples were then 291
113 removed from the filter, ground to a homogenous powder, and analyzed at the 292 University of Florida Stable Isotope Mass Spectrometry Lab. All horseshoe crabs were 293 then returned to the beach from which they had been collected. 294 I conducted paired t tests on the feeding and starving groups separately to test 295 whether the mean difference between pre and post treat ment was different from zero. 296 I then conducted a matched pairs analysis of grouped 297 starving) as a grouping variable. This analysis performs two F tests that evaluates 298 whether the across treatment 299 that the change across the pair of responses (pr e and post treatment values) is 300 301 average response for a subject is different in the feeding and starving groups (JMP 8). 302 Gut Contents Analysis 303 The above experiments reflec t my best efforts to measure the feeding habits of 304 attached and satellite males without sacrificing animals. While these experiments can 305 show support (or not) for the reduced feeding hypothesis, they are indirect. For 306 example 1) t he amount that an anima l defecates is not always a direct reflection of diet 307 because it depends on digestive efficiency of individuals and the digestibility of the food; 308 and 2) instead of nutritional stress, increases in 15 N values may reflect differences in 309 the trophic level a t which animals are feeding : as animals feed at higher trophic levels, 310 the value of 15 N in their tissues increases ( Denir o and Epstein 1981 Michener and 311 Schell 1994 ) Therefore, a measure of gut contents was needed to fully interpret the 312 isotope results (Table 4 1). I attempted to use a non lethal lavage technique, but this 313 failed. Thus, in order to measure directly what and how much satellite and attached 314
114 males are eating, I decided to sacrifice a limited number of animals, and examine their 315 gut contents. I collecte d 10 attached males and 10 satellite males on 20 April 2011 316 while they were spawning on the evening tide. I collected these animals on the first day 317 of that particular week long spawning cycle in order to maximize the likelihood that 318 attached males would have some food in their gut. This also allowed me to directly 319 (and conservatively) compare the amount of food in the gut between attached and 320 satellite males. The animals were euthanized by immediately placin g them in a freezer 321 for 24 hr ( Botton and Ropes 1989 ) The euthanized crabs were fi xed in 10% formalin 322 for 4 5 days; I then dissected out th e digestive tract and stored it in 90% ethanol for 2 323 weeks ( Botton and Ropes 1989 ) I first cu t open the digestive tract (esophagus 324 proventriculous, mid gut and hindgut) to estimate gut fullness (e.g., a score of 100% 325 was assigned if the entire length and width of the gut was filled). Gut contents were 326 then removed by hand and placed in vials wi th 90% ethanol ( Botton and Ropes 1989 ) 327 Durin g the removal of gut contents, I specifically se parated seagrass that was 328 found in the esophagus and proventriculous (but not the lower digestive tract) from 329 other materials because it represents a low trophic level food source. Therefore, 330 differences in seagrass consumption between attached and satell ite males may inform 331 whether or not any differences in 15 N values are the result of feeding on different 332 trophic levels. All seagrass was then dried in an oven at 60 C for 4 h r and wei ghed 333 (mg). I used t tests to compare: 1) gut fullness between attach ed and satellite males, 334 and 2) amount of grass in the esophagus and proventriculous between attached and 335 satellite males. 336
115 To meet the assumptions of normality and homogeneity of variance, I log 337 transformed the values of 1) the amount of feces defecated, 2) transit times, 3) weight 338 of shrimp consumed and 4) seagrass weight in gut prior to all analyses All tests were 339 two tailed, and all variation is reported as standard error, except where noted. 340 Statistical tests were performed with JMP (version 8), and all figures were created using 341 Sigma Plot (version 11) and Adobe Illustrator (CS3). 342 Results 343 Waste Production 344 The mean amount of waste produced (controlled for body size) among the three 345 groups differed (Table 4 2, Figure 4 1). Satellite males produced 57% more was te than 346 attached males (Paired t test: T 25 = 2.7, P = 0.012), and females produced more was te 347 than either attached males (Paired t test: T 25 = 6.1, P < 0.0001) or satellite males 348 (Paired t test: T 25 = 5.4, P < 0.0001). C ondition was not related t o the amount of feces 349 produced for males ( ANOVA: F 7, 45 = 0.8, P = 0.569), or females ( ANOVA: F 7, 19 = 0.6, P 350 = 0.722). 351 Transit Time 352 The whole model ANCOVA was not significant ( F 10, 34 = 1.0, P = 0.440), and transit 353 time was not influenced by status ( F 2 = 1.3, P = 0.284), condition ( F 7 = 1.0, P = 0.421) 354 or the amount of shrimp consumed ( F 1 = 0.2, P = 0.669; Table 4 2, Figure 4 1). 355 A ttached and satellite males did not differ in the amount of shrimp eaten during the 20 356 min feeding period ( contrasts: F 1, 45 = 0.2, P = 0.679), however females ate more 357 ( contrasts: F 1, 45 = 4.1, P = 0.048) than both attached and satellite males (Tabl e 4 2). 358
116 Stable Isotope Analysis of Feces 359 The mean 15 N values for attached males were slightly higher than those for 360 satellite ma les (difference = 0.82 contrasts: F 1, 54 = 3.6, P = 0.063) and females 361 (difference = 0.99 F 2, 54 = 3.1, P = 0.024; Table 4 2 Figure 4 2). I 362 found no difference between satellite males and females (difference = 0.18 363 contrasts: F 1, 54 = 0.2, P = 0.677 ; Table 4 2 Figure 4 2). 364 Experimental Starvation 365 The mean difference between pre and post treatment was not greater than zero 366 for the feeding group (mean = 0.54 t test: T 8 = 1.4, P = 0.208), but was 367 gre ater than zero for the starving group (mean = 1.06 t test: T 9 = 6.3, P 368 = 0.0001). Analysis of grouped data shows differences in the response (pre and post 369 treatment) of 15 N values across the two treatment groups (feeding and starving) fo r 370 both the among matched pairs: F = 9.1, P = 0.008), and the within 371 matched pairs: F = 14.9, P = 0.001; Figure 4 3). 372 Gut Contents 373 Gut fullness (all contents) was 150% greater ( t test: T 18 = 4.0, P = 0.0008) f or 374 satellite males compared to attached males (Table 4 2; Figure 4 4 ). I found that 375 attached males had 200% more seagrass in the esophagus and proventriculous ( t test: 376 T 15 = 2.3, P = 0.035) than satellite males did ( Table 4 2; Figure 4 4). 377 Discussion 378 The maint enance of condition dependent alternative reproductive tactics in a 379 population depends on there being conditions under which each is more successful 380 ( Gross 1996 Brockmann and Taborsky 2008 ) Until this study the tradeoffs for the 381
117 attached tactic in male horseshoe crabs have not been obvious. Satellite males 382 produce more feces than attached males, had higher gut fullness, and had slightly lower 383 15 N values than attached males. Thus, my results support the hypothesis that 384 adva ntages of the attached tactic come at a cost of reduced feeding and nutritional 385 stress. 386 Food transit times (range: 6 to 39 hr) were well within the time that most males 387 remain attached to females ( 3.7 days 6.1 SD; Brockmann and Penn 1992 ) Coupled 388 with the fact that attached males produced some feces during the waste production 389 experiment, it appears that males and females are probably feeding while paired. 390 However, attached males only produced approximately half as much waste on average 391 as satellite males In addition, t here was no difference in transit time between attached 392 and satellite males, and condition did not influence the amount of feces produced. 393 Thus, the results from the waste production experiment demonstrate support for the 394 reduced feeding hypothesis. The gut content analysis show ing that satellite males had 395 a 150% fuller gut than att ached males also strongly supports this hypothesis. 396 Lower fecal production and an emptier gut could have been due to attached males 397 being less motivated to feed, as opposed to being due to the physical constraint of 398 being attached. But, this appears to be unlikely because the amount of food eaten by 399 attached and satellite males in the transit time experiment did not differ (although my 400 measure of consumption might not accurately reflect motivation to feed due to the rather 401 artificial conditions). Alternat ively, my results may have been due to satellite males 402 eating more recently. For example, perhaps attached males do not feed at all after 403 attaching, and that the waste I saw was what remained of their intake prior to attaching. 404
118 Nonetheless, both possibil ities suggest reduced food consumption for attached males in 405 this population. Further, I show that this reduced feeding is costly for males that are 406 using the attached tactic, even though a ttached males may have lower energetic 407 requirements than the satel 408 because they are not required to spend energy locating multiple mating groups on each 409 tide. Additionally, given the trade off between food passage time and digestive 410 efficiency ( Penry 1993 ) males may slow their food passage time when attached, and 411 increase efficiency of assimilation, thereby compensating for not eating as much as the 412 satellite males. My study was not design ed to test this, however, evidence of increased 413 15 N values for attached males refutes this hypothesis, and demonstrates nutritional 414 stress in attached males. 415 A model proposed by Martinez Del Rio and Wolf ( 2005 ) predicts that 15 N values 416 would increase with fasting time. My observations of horseshoe crabs showed t hat the 417 feces of attached males increased 15 418 males While this difference was not statistically significant, the effect size was nearly 419 20% higher for attached males, and lack of significance may be due to my relatively 420 small sample sizes. In order to attribute this difference to nutritional stress I first had to 421 show that starvation produces an increase in 15 N values in feces. Experimentally, I 422 found that a 4 week starvation increased 15 N values by 1.06 423 measurement, but the values did not change for animals that were fed. The degree of 424 enrichment that I found observationally and experimentally is comparable to the results 425 from other studies. For example, in quail that were fed a reduced food intake, blood 426 15 ( Hobson et al. 1993 ) ; in lizards, 427
119 uric acid 15 ( Castillo and Hatch 2007 ) ; in 428 daphnia, whole body tissue 15 N values increased by 5 days of starving 429 compared with controls ( Adams and Sterner 2000 ) ; and in spider hatchlings, whole 430 body tissue 15 431 initial values ( Oelbermann and Scheu 2002 ) 432 I found little evidence that differences in 15 N values between attached and 433 satellite males reflects a difference in the trophic level at which they are feeding or a 434 differen ce in their diet. Horseshoe crabs create a slurry of sediment and food when 435 they feed ( Botton and Shuster 2003 ) and it seems likely that attached males are able to 436 grab food particles missed by the females. However, I found that attached males are 437 less similar to females in 15 N values than are satellite ma les. Because attached males 438 are physically associated with females I would have expected the opposite outcome. 439 Perhaps attached males selectively feed on animal tissue (which has higher 15 N 440 values) rather than on plant matter and detritus ( which is a typical food source that has 441 ) If attached males ate less organic material 442 than satellite males or females, it could explain why attached males are higher in 15 N 443 values, and also why they are more different from females than from satellite males. In 444 contrast, I found the opposite pattern: attached males actually consumed 200% more 445 plant material than satellite m ales (perhaps seagrass is an easy source of food for 446 attached males to access ). This result, along with the finding that experimental starving 447 increase d 15 N values, indicates that the increase in 15 N values of naturally occurring 448 attached males are the result of a period of fasting, as opposed to differences in di et. 449
120 Taken together, my results demonstrate that reduced consumption of food and a 450 period of nutritional stress are costs of the attached tactic not previously considered. 451 There is evidence that in other populations males remain attached longer than in the 452 Seahorse Key population used in this study ( Shuster 1954 Barlow et al. 1987 Moore 453 2004 ) ; and in other horseshoe crab species males are more firmly attached and they 454 remain attached for long er periods ( Botton et al. 1996 Brockmann and Smith 2009 ) 455 suggesting even greater costs of attaching in those populations and species. 456 In conclusion, this is one of the first studies to use stable isotopes to investigate a 457 predicted period of nutritio nal stress in a natural population of animals. Moreover, this 458 study furthers our understanding of the trade offs in this system, and provides a key 459 piece of information that may potentially explain why these alternative tactics in 460 horseshoe crabs take the form they do. My findings show that the satellite tactic has 461 specific benefits in that these males are able to feed, whereas feeding is restricted for 462 attached males. Low energy alternative phenotypes or behaviors often evolve as a 463 release from the ener ( Brockmann and Smith 464 2009 ) and in some systems 465 energy reserves ( Taborsky 1998 Widemo 1998 Cummings and Gelineau Kattner 466 2009 ) Therefore, older, poor condition males may not be able to afford the periodic 467 fasting that accompanies being attached to a female during breeding. Investigating this 468 hypothesis is the next step to fully understanding the evolution and maintenance of 469 alternative reproductive tactics in horseshoe crabs. 470
121 Table 4 1. Summary of the predictions and assumptions of the reduced feeding hypothesis, and the methods I used to test them. Prediction Method Attached males eat less food than satellite males Examine gut fullness of wild animals Attached males defecate less than satellite males Measure waste production over 12 hr period in wild animals Assumes equal transit times Measure transit times in hand fed, wild animals Assumes eating less food is costly Test for nutritional str ess using stable isotopes ( 15N) Stable isotope values indicate nutritional stress Measure 15N of feces of wild animals Assumes fasting increases 15N values of feces Experimentally starve animals and measure 15N of feces Assumes males of both tactics have same diet Examine gut contents of wild animals
122 Table 4 2. Values of various measures for three groups of horseshoe crabs (attached males, satellite males, and females) collected from Seahorse Key, FL in 2008 2 010. Attached Males Satellite Males Females Mean SD N 95% CI Mean SD N 95% CI Mean SD N 95% CI Carapace width (cm) 16.2 0.85 43 16 16.5 16.1 1.0 43 15.8 16.4 21.4 2.1 43 20.8 22.1 Median condition 8 43 7 10 a 7 43 5 9 a 7 43 6 9 a Defecation (g, dry weight) 0.79 1.1 27 0.34 1.24 1.24 1.4 27 0.70 1.78 3.8 3 27 2.61 4.94 Shrimp consumed (g, wet weight) 0.48 0.3 16 0.31 0.65 0.44 0.3 16 0.28 0.59 0.29 0.2 16 0.16 0.42 Transit time (h) 18.9 9.1 16 14.0 24.0 17.5 6.7 15 13.8 21.2 22.8 10.7 14 16.3 29.0 Feces 15 5.0 1.3 19 4.4 5.6 4.2 1.1 19 3.7 4.7 4.0 1.6 19 3.3 4.8 Gut fullness (%) 12.4 1.2 10 8.9 17.4 31.0 1.2 10 22.2 43.4 Foregut seagrass (mg) 9.4 2.3 10 4.2 14.5 3.1 1.4 10 0 6.3 a 25% and 75% quartiles Waste minus debris in control tanks. Control tanks had an average of 0.13 0.3 g of material present (95% = 0.01 0.26).
123 Figure 4 1. Mean fecal mass SE g produced in 12 hr by wild caught horseshoe crabs (white bars; N = 19 for each group), and mean transit time SE hr of horseshoe crabs that were experimentally fed Carmine red dyed shrimp (black bars; N = 16 for attached males, 15 for satellite males, and 14 for females). Al l individuals from the three groups (A = attached males, S = satellite males, and F = females) were collected from Seahorse Key, FL. Groups not indicated by the same letter were significantly different (P = 0.05) based on least squares means contrasts.
124 Figure 4 15N stable isotopes of feces produced by three groups of horseshoe crabs (N = 19 each for each group) collected from Seahorse Key, FL.
125 Figure 4 3. A) Pre and post 15N from two groups of satellite male horseshoe crabs: one group was starved for 4 weeks, and the other group was fed scallops ad libitum for 4 weeks (N = 10 each). B) Mean 95% CI difference in 15N values for the two experimental groups (post treatment minus pre post treatment values.
126 Figure 4 4. Mean gut fullness SE % (white bars) and mean weight SE mg of seagrass found in gut (black bars) of attached and satel lite male horseshoe crabs (N = 10 each). and satellite males.
127 CHAPTER 5 SUMMARY AND CONCLUSI ONS ON PHENOTYPIC VA RIATION IN BODY SIZE AND REPRODUCTIVE BEHAVIO R Introduction I have identified ultimate and proximate explanations for phenotypic variation in horseshoe crab body size among populations and between sexes, and in the mating tactics of male horseshoe crabs. Chapter 2 In my second chapter, I found a dome shaped distribution of adult body size in horseshoe crabs across North America. The increase in body size from Hog Bay, ME to Virginia was North Caro lina to Laguna Ultimate explanations for this pattern are that 1) optimal resource allocation to reproduction, mediated by survivorship, is influenced by local environmental and ecological cond itio ns, and 2) optimal body size result s from tradeoffs between growth, survival, and fecundity. The dome shaped pattern may result from optimal temperatures salinities and oxygen concentrations in the center of their range along with a long growing se ason, and an abundant food supply. Proximately, the temperature threshold hypothesis is the most likely explanation for the mechanism underlying the ecogeographic cline in horseshoe crabs, however we lack sufficient data to make a definitive conclusion. c onverse and that constraints on growth can act multiplicatively to create a dome shaped distribution in body size ( Blanckenhorn and Demont 2004 ) Patterns of body size distribution should be further examined in other marine species with la rge ranges (or that occur over a broad range of environmental conditions) long development times, genetic population structure, and high variation in season length between
128 known populations. In the future, this observational study should be bolstered wit h experiments that manipulate temperature, oxygen, season length, and salinity to get a better understanding of minimum temperature thresholds for growth and differentiation. Chapter 3 My chapter on sexual size dimorphism found strong support for the fecun dity advantage hypotheses The benefits of increased fecundity favor later maturation for females at a larger size. In addition the protandry hypothesis predict ing that by maturing early at a smaller size, males have higher lifetime reproductive success than they would by waiting another year warrants further explor ation. A dynamic state variable model is one approach that may yield evidence regarding this hypothesis. Fecundity selection for later maturation in females, potentially coupled with selectio n for early maturation on males, appears to be an ultimate explanation for the evolution and maintenance of sexual size dimorphism in horseshoe crabs. My study illustrates the importance of investigating multiple hypotheses for both males and females in o rder to understanding the suite of pressures underlying the ubiquitous pattern of sex ual size dimorphism in animals. Chapter 4 In my fourth chapter I have shown differences in defecation (but similar retention times), and gut fullness between attached and satellite males. I also found that the feces of naturally occurring attached males had higher 15 N values compared to satellite males. I have also shown experimentally that a period of fasting results in increases in increased 15 N values. Taken togethe r, my results demonstrate that reduced consumption of food and a period of nutritional stress are additional costs of the attached tactic not previously considered. In conclusion, this study furthers our understanding of the trade offs in this system, and provides a
129 key piece of information that may potentially explain why these alternative tactics in horseshoe crabs take the form they do. My findings show that the satellite tactic has specific benefits in that these males are able to feed, whereas feedin g is restricted for attached males. Low energy alternative phenotypes or behaviors often evolve as a release of the energetic demands of ( Taborsky 1998 Widemo 1998 Cummings and Gelineau Kattner 2009 ) ( McCauley et al. 2000 ) Therefore, older, poor condition males may not be able to afford the periodic fasting that accompanies being attached to a female during breeding. Investigating this hypothesis is the next step to fully understanding the evolution and maintenance of alternative reproductive tactics in horseshoe crabs. C onclusions Overall, this research uses horseshoe crabs as a detailed system to better understand the factors that underlie the ubiquitous variation in body size and mating strategies across animal species. Because horseshoe crabs show an unusual dome shap ed distribution, they offer a unique opportunity for future research to uncover a unifying mechanism underlying ecogeographic clines in body size. Chapter 3 illustrates the importance of testing a suite of pressures, rather than simply examining one or tw o hypotheses. In chapter 4, I uncover a cost of the attached mating tactic which helps shed light on why the alternative tactics take the form they do in this species. The stable isotope techniques used in this chapter are novel, and will provide other r esearchers with a means to test for nutritional stress in other systems. This research also adds to our limited understanding of horseshoe crab biology in locations at the edge of their range ( Anderson and Shuster 2003 ) Such information is essential to the conservation of horseshoe crabs because it reveals details on the large scale constraints to the abundance and distribution of the species ( Shuster and Sekiguchi 2009 ) It seems
130 especially important to understand how environmental and ecological factors affect their distribution in light of global climate change that may affect populations of this medically and commercially important species ( Botton and Itow 2009 Sekiguchi and Shuster 2009 ) Increases in global temperature may cause a reduction in the southern range of this species, and future research on this topic is warranted. This research may also be useful to the management of horseshoe crabs that are used for commercial purposes. Fisherman use horseshoe crabs as bait, and they have a preference for taking the largest animals, generally females, but also large males. Understanding why males and females differ in size, and how costs of alternative tactics might influence reproductiv e success may also be important to management decisions ( Berkson 2009 Smith et al. 2009b )
131 APPENDIX A REFERENCES FOR TABLE 2 1. Table A 1. References for average body size of horseshoe crabs referred to in T able 2 1. Table 1 Reference # Reference Locations 1 Schaller 2002 Hog Bay, Bagaduce River, Thomas Point Beach (ME) 2 Botton and Loveland 1992 Great Bay (NH), Sandy Hook Bay (NJ) 3 James Pirri et al. 2005 Cape Cod Bay, Nauset Estuary, Monomoy NWR, Pleasant Bay (MA) 4 Shuster, Jr. 1982 Cape Cod Bay, Pleasant Bay (MA) 5 Graham et al. 2009 Narragansett Bay (RI); Milford, New Haven, Norwalk (CT); Inshore continental shelf (NY) 6 M.D. Smith, Unpub. Data Stony Brook (NY); Skidaway Island, Sapelo Island (GA); Seahorse Key (FL) 7 Shuster, Jr. et al. 1993 Delaware Bay (DE) 8 Loveland and Botton 1992 Delaware Bay (DE) 9 Walls 2001 Oc ean City (MD); Chincoteague Bay (VA) 10 NC Division of Mar. Fish. Inshore continental shelf (NC) 11 SC Office of Fish. Mgmt. Inshore continental shelf (SC) 12 Thompson 1998 Otter Island, SC 13 Rudloe 1978 Ochlockonee Bay (FL) 14 Ehlinger 2001 Indian River Lagoon (FL) 15 Gerhart 2007 Tampa Bay (FL) 16 Zaldvar Rae et al. 2009 San Felipe lagartos, Progreso (YUC); Holbox (ROO); Champotn, Laguna de Trminos (CAM)
132 APPENDIX B LIST OF TAXA USED IN ESTIMATING FOOD ABUNDANCE Table A 2. List of taxa used in the estimation of the abundance of food for horseshoe crabs in north america based on noaa trawl surveys from 1953 1999. Name Classification level Scientific classification Anthozoa Class Phylum: Cnidaria Annelida Phylum Phylum : Annelida Sipuncula Subphylum Phylum : Annelida Polychaeta Class Phylum : Annelida Echinodermata Phylum Phylum: Echinodermata Formanifera Phylum Phylum : Formanifera Mollusca Phylum Phylum : Mollusca Bivalvia Class Phylum : Mollusca Gastropoda Class Phylum : Mollusca Scaphopoda Class Phylum : Mollusca Ostracoda Class Phylum: Arthropoda Subphylum: Crustacea Copepoda Subclass Phylum: Arthropoda Subphylum: Crustacea Class: Malacostraca Isopoda Subclass Phylum: Arthropoda Subphylum: Crustacea Class: Malacostraca Amphipoda Order Phylum: Arthropoda Subphylum: Crustacea Class: Malacostraca
133 APPENDIX C REFERENCES FOR TABLE 3 2 Table A 3. References used to obtain data for body size, sexual size dimorphism, operational sex ratio, and female nesting density of horseshoe crabs refe r red to in Chapter 3, T able 3 2. T able 3 2 reference # Reference Locations 1 ( Schaller 2002 ) Hog Bay, Bagaduce River, Thomas Po int Beach; ME 2 ( Shuster 1982 ) Cape Cod Bay, Pleasant Bay; MA 3 ( James Pirri et al. 2005 ) Cape Cod Bay, Nauset Estuary, Monomoy NWR, Pleasant Bay; MA 4 ( Graham et al. 2009 ) Narragansett Bay, RI; Milford, CT 5 ( Mattei et al. 2010 ) Narragansett Bay, RI; Milford, CT 6 ( Shuster 1979 ) Delaware Bay, NJ and DE 7 ( Shuster and Botton 1985 ) Delaware Bay, NJ and DE 8 ( Loveland and Botton 1992 ) Delaware Bay, NJ and DE 9 ( Shuster et al. 1993 ) Delaware Bay, NJ and DE 10 ( Smith et al. 2002 ) Delaware Bay, NJ and DE 11 SC Office of Fish. Mgmt. Inshore continental shelf, SC 12 ( Thompson 1998 ) Otter Island, SC 13 ( Rudloe 1980 ) Ochlockonee Bay, FL 14 H.J. Brockmann, Unpublished Data Seahorse Key, FL 15 ( Ehlinger et al. 2003 ) Indian River Lagoon, FL 16 ( Ehlinger 2001 ) Indian River Lagoon, FL
134 LIST OF REFERENCES Evolution 62 :413 420. Adams, J., P. Greenwood, R. Pollitt, and T. Yonow. 1985. Loading constraints and sexual size dimorphism in Asellus aquaticus Behaviour 92 :277 287. Adams, T. S. and R. W. Sterner. 2000. The effect of dietary nitrogen content on trophic level N 15 enrichmen t. Limnology and Oceanography 45 :601 607. Alcock, J. 2009. Animal Behavior: an Evolutionary Approach. 9 edition. Sinauer Associates, Sunderland, MA. Anderson, L. I. and C. N. Shuster, Jr. 2003. Throughout geologic time: where have they lived? Pages 189 223 in C. N. S. Jr., R. B. Barlow, and H. J. Brockmann, editors. The American Horseshoe Crab. Harvard University Press, Cambridge. Andersson, M. 1994. Sexual Selection. Princeton University Press, Princeton. Angilletta, M. J., Jr. 2009. Temperature and the li fe history. Pages 157 180 in M. J. Angilletta, editor. Thermal Adaptation: a Theoretical and Empirical Analysis. Oxford University Press, New York. Angilletta, M. J., Jr. and A. E. Dunham. 2003. The temerature size rule in ectotherms: simple evolutionary e xplanations may not be general. American Naturalist 162 :332 342. Angilletta, M. J., Jr., P. H. Niewiarowski, A. E. Dunham, A. D. Leache, and W. P. history perspective with Sceloporine liz ards. American Naturalist 164 :E168 183. Angilletta, M. J., Jr., T. D. Steury, and M. W. Sears. 2004b. Temperature, Growth Rate, and Body Size in Ectotherms: Fitting Pieces of a Life History Puzzle. Integrative and Comparative Biology 44 :498 509. Arnett, A. E. and N. J. Gotelli. 1999. Bergmann's rule in the ant lion Myrmeleon immaculatus DeGeer (Neuroptera: Myrmeleontidae): geographic variation in body size and heterozygosity. Journal of Biogeography 26 :275 283. Zoology 80 :708 716. avian reptiles: turtles follow it, lizards and snakes reverse it. Evolution 57 :1151 1163. Ashton, K. G., M. C. Tracy, and A. de Queiroz. 2000 mammals? American Naturalist 156 :390 415.
135 Atkinson, D. 1994. Temperature and organism size a biological law for ectotherms? Advances in Ecological Research 25 :1 58. Atkinson, D. and R. M. Sibly. 1997. Why are organisms usually bigger in colder environments? Making sense of a life history puzzle. Trends in Ecology & Evolution 12 :235 239. Bademan, M. L. 2009. Tolerance and physiological responses of larval and juvenile horseshoe crabs Limulus polyphemus to hypoxia and anoxia. Florida Institute of Technolgy, Melbourne, FL. Badyaev, A. V., G. E. Hill, A. M. Stoehr, P. M. Nolan, and K. J. McGraw. 2000. The evolution of sexual dimorphism in the house finch: II. Population di vergence in relation to local selection. Evolution 54 :2134 3144. Badyaev, A. V., L. A. Whittingham, and G. E. Hill. 2001. The evolution of sexual size dimorphism in the house finch: III. Developmental basis. Evolution 55 :176 189. Barboza, P. S., K. L. Parker, and I. D. Hume. 2009. Integrative Wildlife Nutrition. Springer Verlag, Berlin. Barlow, R. B., M. K. Powers, H. Howard, and L. Kass. 1986. Migration of Limulus for mating: relation to lunar phase, tide height, and sunlight. Biological Bulletin 171 : 310 329. Barlow, R. B., M. K. Powers, H. Howard, and L. Kass. 1987. Vision in Limulus mating and migration. Pages 69 84 in W. F. Herrnkind and A. B. Thistle, editors. Signposts in the Sea. Florida State University Press, Tallahassee, FL. Belk, M. C. and H. fishes. American Naturalist 160 :803 808. Bergmann, C. 1847. ber die Verhltnisse der Wrme konomie der Thiere zu ihrer Grsse. Gottinger Studien 3 :595 708. Berkson, J. 2009. An integrat ive approach to horseshoe crab multiple use and sustainability. in J. T. Tanacredi, M. L. Botton, and D. R. Smith, editors. Biology and Conservation of Horseshoe Crabs. Springer, New York. Blackburn, T. M., K. J. Gaston, and N. Loder. 1999. Geographic gradi ents in body size: 5 :165 174. Blanckenhorn, W. U. 2000. The evolution of body size: what keeps organisms small? Quarterly Review of Biology 75 :385 407.
136 Blanckenhorn, W. U. 2007. Case studies o f the differential equilibrium hypothesis of sexual size dimorphism in two dung fly species. in D. J. Fairbairn, W. U. Blanckenhorn, and T. Szekely, editors. Sex, Size & Gender Roles: Evolutionary Studies of Sexual Size Dimorphism. Oxford University Press, Oxford. Blanckenhorn, W. U. and M. Demont. 2004. Bergmann and converse bergmann latitudinal clines in arthropods: two ends of a continuum? Integrative and Comparative Biology 44 :413 424 Blanckenhorn, W. U. and D. J. Fairbairn. 1995. Life history adaptatio n along a latitudinal cline in water striders. Journal of Evolutionary Biology 8 :21 41. Blanckenhorn, W. U., R. C. Stillwell, K. A. Young, C. W. Fox, and K. G. Ashton. 2006. When Rensch meets Bergmann: does sexual size dimorphism change systematically with latitude? Evolution 60 :2004 2011. Blows, M. W. and A. A. Hoffmann. 2005. A reassessment of genetic limits to evolutionary change. Ecology 86 :1371 1384. Botton, M. L. 1984. Diet and food preferences of the adult horseshoe crab Limulus polyphemus in Delawar e Bay, New Jersey, USA. Marine Biology 81 :199 207. Botton, M. L. and H. H. Haskin. 1984. Distribution and feeding of the horseshoe crab, Limulus polyphemus on the continental shelf off New Jersey. Fishery Bulletin 82 :383 389. Botton, M. L. and T. Itow. 20 09. The effects of water quality on horseshoe crab embryos and larvae. in J. T. Tanacredi, M. L. Botton, and D. R. Smith, editors. Biology and Conservation of Horseshoe Crabs. Springer, New York. Botton, M. L. and R. E. Loveland. 1992. Body size, morphologi cal constraints, and mated pair formation in 4 populations of horseshoe crabs ( Limulus polyphemus ) along a geographic cline. Marine Biology 112 :409 415. Botton, M. L., R. E. Loveland, and T. R. Jacobsen. 1988. Beach erosion and geochemical factors: influe nce on spawning success of horseshoe crabs ( Limulus polyphemus ) in Delaware Bay. Marine Biology 99 :325 332. Botton, M. L., R. E. Loveland, and A. Tiwari. 2003. Distribution, abundance, and survivorship of young of the year in a commercially exploited population of horseshoe crabs, Limulus polyphemus Marine Ecology Progress Series 265 175 184. Botton, M. L., R. E. Loveland, and T. R. Jacobsen 1988. Beach erosion and geochemical factors: influence on spawning success of horseshoe crabs ( Limulus polyphemus ) in Delaware Bay. Marine Biology 99 :325 332.
137 Botton, M. L. and J. W. Ropes. 1987. Populations of horseshoe cra bs, Limulus polyphemus on the Northwestern Atlantic Continental Shelf. Fisheries Bulletin 85 :805 812. Botton, M. L. and J. W. Ropes. 1988. An indirect method for estimating longevity of the horseshoe crab ( Limulus polyphemus ) based on epifaunal slipper sh ells ( Crepidula fornicata ). Journal of Shellfish Research 7 :407 412. Botton, M. L. and J. W. Ropes. 1989. Feeding ecology of horseshoe crabs on the continental shelf, New Jersey to North Carolina. Bulletin of Marine Science 45 :637 647. Botton, M. L. and C. N. Shuster, Jr. 2003. Horseshoe crabs in a food web: who eats whom? Pages 133 153 in C. N. Shuster, Jr., R. B. Barlow, and H. J. Brockmann, editors. The American Horseshoe Crab. Harvard University Press, Cambridge. Botton, M. L., C. N. Shuster, Jr., K. Sekiguchi, and H. Sugita. 1996. Amplexus and mating behavior in the Japanese horseshoe crab, Tachypleus tridentatus Zoological Science 13 :151 159. Botton, M. L., R. A. Tankersley, and R. E. Loveland. 2010. Developmental ecology of the Am erican horseshoe crab Limulus polyphemus Current Zoology 56 :550 562. Boyce, M. S. 1988. Evolution of Life Histories of Mammals: Theory and Pattern. Yale University Press, London. Bradford, D. E. 1990. Incubation time and rate of embryonic development in a mphibians: The influence of ovum size, temperature and reproductive mode. Physiological Zoology 63 :1157 1180. Brockmann, H. J. 1990. Mating behavior of horseshoe crabs, Limulus polyphemus Behaviour 114 :206 220. Brockmann, H. J. 1996. Satellite male groups in horseshoe crabs, Limulus polyphemus Ethology 102 :1 21. Brockmann, H. J. 2001. The evolution of alternative strategies and tactics. Advances in the Study of Behavior 30 :1 51. 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. 2003. Male competition and satellite behavior. Pages 50 82 in C. N. Shuster, Jr., R. B. Barlow, and H. J. Brockmann, editors. The American Horseshoe Crab. Harvard U niversity Press, Cambridge. Brockmann, H. J., T. Colson, and W. Potts. 1994. Sperm competition in horseshoe crabs ( Limulus polyphemus ). Behavioral Ecology and Sociobiology 35 :153 160.
138 Brockmann, H. J. and S. L. Johnson. 2011. A long term study of spawning activity in a Florida Gulf Coast population of horseshoe crabs ( Limulus polyphemus ). Estuaries and Coasts 34 :1049 1067. Brockmann, H. J., C. Nguyen, and W. Potts. 2000. Paternity in horseshoe crabs when spawning in multiple male groups. Animal Behaviour 60 :837 849. Brockmann, H. J., R. F. Oliveira, and M. Taborsky. 2008. Integrating mechanisms and function: prospects for future research. Pages 471 489 in R. F. Oliveira, M. Taborsky, and H. J. Brockmann, editors. Alternative Reproductive Tactics. An Integrat ive Approach. Cambridge University Press, Cambridge. Brockmann, H. J. and D. Penn. 1992. Male mating tactics in the horseshoe crab, Limulus polyphemus Animal Behaviour 44 :653 665. Brockmann, H. J. and M. D. Smith. 2009. Reproductive competition and sexual selection in horseshoe crabs. Pages 199 221 in J. T. Tanacredi, M. L. Botton, and D. R. Smith, editors. Biology and Conservation of Horseshoe Crabs. Springer, New York. Brockmann, H. J. and M. Taborsky. 2008. Alternative reproductive tactics and the evolu tion of alterntive allocation phenotypes. Pages 25 51 in R. F. Oliveira, M. Taborsky, and H. J. Brockmann, editors. Alternative Reproductive Tactics. An Integrative Approach. Cambridge University Press, Cambridge. Brown, J. H. 1984. On the relationship bet ween distribution and abundance of species. American Naturalist 124 :255 279. Brown, J. H., G. C. Stevens, and D. M. Kaufman. 1996. The geographic range: size, shape, boundaries, and internal structure. Annual Review of Ecology and Systematics 27 :597 623. C alder, W. A. 1984. Size, Function and Life History. Harvard University Press, Cambridge. Carmichael, R. H., D. Rutecki, B. Annett, E. Gaines, and I. Valiela. 2004. Position of horseshoe crabs in estuarine food webs: N and C stable isotopic study of foragin g ranges and diet composition. Journal of Experimental Marine Biology and Ecology 299 :231 253. Carroll, S. P. 1988. Constraints in the reproductive ecology between temerpate and tropical populations of Jadera haematoloma a mate guarding Hemipteran (Rhopal idae). Annals of the Entomological Society of America 81 :54 63. Carroll, S. P. 1991. The adaptive significance of mate guarding in the soapberry bug, Jadera haematoloma ( Hemiptera: Rhopalidae). Journal of lnsect Behavior 4 :509 530.
139 Castillo, L. P. and K. A Hatch. 2007. Fasting increases delta N 15 values in the uric acid of Anolis carolinensis and Uta stansburiana as measured by nondestructive sampling. Rapid Communications in Mass Spectrometry 21 :4125 4128. Chabot, C. C. and W. H. Watson, III. 2010. Circa tidal rhythms of locomotion in the American horseshoe crab Limulus polyphemus : underlying mechanisms and cues that influence them. Current Zoology 56 :499 517. Chapelle, G. and L. Peck. 1999. Polar gigantism dictated by oxygen availability. Nature 399 :114 1 15. Chapelle, G. and L. Peck. 2004. Amphipod size spectra: new insights in the relationship between size and oxygen. Oikos 106 :167 175. Chatterji, A. 1995. Fecundity of the Indian horseshoe crab, Tachypleus gigas (Muller) from Balramgari (Orissa). Pakistan Journal of Marine Science 4 :127 131. Chatterji, A. and A. H. Parulekar. 1992. Fecundity of the Indian horse shoe crab, Carcinoscorpius rotundicauda (Latreille). Journal of Tropical Ecology 33 :97 102. Chown, S. L. and K. J. Gaston. 1999. Exploring links between physiology and ecology at macro scales: the role of respiratory metabolism in insects. Biological Reviews 74 :87 120. Chown, S. L. and K. J. Gaston. 2010. Body size variation in insects: a macroec ological perspective. Biological Reviews 85 :139 169. Chown, S. L. and C. J. Klok. 2003. Altitudinal body size clines: latitudinal effects associated with changing seasonality. Ecography 26 :445 455. Chown, S. L. and J. S. Terblanche. 2007. Physiological div ersity in insects: ecological and evolutionary contexts. Advances in Insect Physiology 33 :50 152. Cichon, M. 1997. Evolution of longevity through optimal resource allocation. Proceedings of the Royal Society B Biological Sciences 264 :1383 1388. Cichon, M. optimal resource allocation. Evolutionary Ecology Research 2 :857 870. Cohen, J. A. and H. J. Brockmann. 1983. Breeding Activity and Mate Selection in the Horseshoe Crab, Limulus Poly phemus. Bulletin of Marine Science 33 :274 281. Conover, D. O. and T. M. C. Present. 1990. Countergradient variation in growth rate: compensation for length of the growing season among Atlantic silversides from different latitudes. Oecologia 83 :316 324.
140 Coo ney, J. D., J. J. Beauchamp, and C. W. Gehrs. 1983. Effects of temperature and nutritional state on the acute toxicity of acridine to the calanoid copepod, Diaptomus clavipes Environmental Toxicology and Chemistry 2 :431 439. Cox, R. M., M. Butler, and H. B John Alder. 2007. The evolution of sexual size dimorphism in reptiles. Pages 38 49 in D. J. Fairbairn, W. U. Blanckenhorn, and T. Szekely, editors. Sex, Size & Gender Roles: Evolutionary Studies of Sexual Size Dimorphism Oxford University Press, Oxford. Cox, R. M., S. L. Skelly, and H. B. John Alder. 2003. A comparative test of adaptive hypotheses for sexual size dimorphism in lizards. Evolution 57 :1653 1669. Cummings, M. E. and R. Gelineau Kattner. 2009. The energetic costs of alternative male reproducti ve strategies in Xiphophorus nigrensis Journal of Comparative Physiology 195 :935 946. Cushman, J. H., J. H. Lawton, and B. F. J. Manly. 1993. Latitudinal patterns in European ant assemblages: variation in species richness and body size. Oecologia 95 :30 37 optimal resource allocation? Ecology Letters 1 :5 7. Darwin, C. 1859. On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the S truggle for Life. John Murray, London. Darwin, C. 1874. The Descent of Man and Selection in Relation to Sex. John Murray, London. variation on a mechanism that controls in sect body size. Evolutionary Ecology Research 6 :49 62. Davidowitz, G. and H. F. Nijhout. 2004. The physiological basis of reaction norms: the interaction among growth rate, the duration of growth and body size. Integrative and Comparative Biology 44 :443 449. Deniro, M. J. and S. Epstein. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta 45 :341 351. Ehlinger, G. S. 2001. Species Report, Limulus polyphemus Smithsonian Marine Station at Fort Pi erce. Ehlinger, G. S. and R. A. Tankersley. 2004. Survival and development of horseschoe crab, Limulus polyphemus embryos and larvae in hypersaline conditions. Biological Bulletin 206 :87 94.
141 Ehlinger, G. S., R. A. Tankersley, and M. B. Bush. 2003. Spatial and temporal patterns of spawning and larval hatching by the horseshoe crab, Limulus polyphemus in a microtidal coastal lagoon. Estuaries 26 :631 640. Elmes, G. W., J. C. Wardlaw, M. G. Nielsen, V. E. Kipyakov, E. B. Lopatina, A. G. Radchenko, and B. Barr 1999. Site latitude influences on respiration rate, fat content and the ability of worker ants to rear larvae: a comparison of Myrmnica rubra (Hymenoptera: Formicidae) populations over their European range. European Journal of Entomology 96 :117 124. Emle n, D. J. 1994. Environmental control of horn length dimorphism in the beetle Onthophagus acuminatus (Coleoptera: Scarabaeidae). Proceedings of the Royal Society B Biological Sciences 256 :131 136. Endler, J. A. 1980. Natural selection on color patterns in P oecilia reticulata Evolution 34 :76 91. Endler, J. A. 1991. Variation in the appearance of guppy color patterns to guppies and their predators under different visual conditions. Vision Research 31 :587 608. Fagerstrom, T. and C. Wiklund. 1982. Why do males emerge before females protandry as a mating strategy in male and female butterflies. Oecologia 52 :164 166. Fairbairn, D. J. 1990. Factors influencing sexual size dimorphism in temperate waterstriders. American Naturalist 136 :61 86. Fairbairn, D. J. 2007. Introduction: the enigma of sexual size dimorphism. Pages 1 10 in D. J. Fairbairn, W. U. Blanckenhorn, and T. Szkely, editors. Sex, Size, and Gender Roles: Evolutionary Studies of Sexual Size Dimorphism. Oxford University Press, Oxf ord. Faurby, S., K. S. K. Nielsen, S. Bussarawit, I. Intanai, V. C. Nguyen, C. Pertoldi, and P. Funch. 2011. Intraspecific shape variation in horseshoe crabs: the importance of sexual and natural selection for local adaptation. Journal of Experimental Mari ne Biology and Ecology 407 :131 138. Fisher, D. C. 1984. The Xiphosurida: archetypes of Bradytely? in N. Eldrege and S. M. Stanley, editors. Living Fossils. Springer Verlag, New York. Fisher, J. A. D., K. T. Frank, and W. C. Leggett. 2010. Breaking Bergmann' s rule: truncation of Northwest Atlantic marine fish body sizes. Ecology 91 :2499 2505. Folt, C. L., C. Y. Chen, M. V. Moore, and J. Burnaford. 1999. Synergism and Antagonism among Multiple Stressors. Limnology and Oceanography 44 :864 877.
142 Gannes, L. Z., C. M. del Rio, and P. Koch. 1998. Natural abundance variations in stable isotopes and their potential uses in animal physiological ecology. Comparative Biochemistry and Physiology 119 :725 737. Gannes, L. Z., D. M. O Brien, and C. M. delRio. 1997. Stable isot opes in animal ecology: assumptions, caveats, and a call for more laboratory experiments. Ecology 78 :1271 1276. Gaston, K. J., S. L. Chown, and K. L. Evans. 2008. Ecogeographical rules: elements of a synthesis. Journal of Biogeography 35 :483 500. Gaston, K. J., S. L. Chown, and C. V. Styles. 1997. Changing size and changing enemies: the case of the mopane worm. Acta Oecologia 18 :21 26. Geist, V. 1987. Bergmann's rule is invalid. Canadian Journal of Zoology 65 :1035 1038. Ghiselin, M. T. 1974. The Ec onomy of Nature and the Evolution of Sex. University of California Press, Berkeley, CA. Gilligan, M. R. 1991. Bergmann ecogeographic trends among the triplefin blennies in the Gulf of California, Mexico. Environmental Biology of Fishes 31 :301 305. Graham, L. J., M. L. Botton, D. Hata, R. E. Loveland, and B. R. Murphy. 2009. Prosomal width to weight relationships in American horseshoe crabs ( Limulus polyphemus ): examining conversion factors used to estimate landings. Fishery Bulletin 107 :235 243. Graham, M. H. 2003. Confronting multicollinearity in ecological multiple regression. Ecology 84 :2809 2815. Greenwood, P. J. and J. Adams. 1987. Sexual selection, size dimorphism, and a fallacy. Oikos 48 :106 108. Gross, M. R. 1996. Alternative reproductive strategies and tactics: diversity within sexes. Trends in Ecology & Evolution 11 :92 98. Gunnarsson, B. and J. Johnsson. 1990. Protandry and molting to maturity in the spider Pithyohyphantes ophrygianus Oikos 59 :205 212. Hanazato, T. and S. I. Dodson. 1995. Synergist ic effects of low oxygen con centration, predator kairomone, and a pesticide on the cladoceran Daphnia pulex Limnology and Oceanography 40 :700 709. Harvey, P. H., D. E. L. Promisow, and A. F. Read. 1989. Causes and correlates of life history differences among mammals. Pages 305 318 in V. Standen and R. A. Foley, editors. Comparative Socioecology. Blackwell Scientific Publications, Oxford.
143 Hatcher, M. J. and A. M. Dunn. 1997. Size and pairing success in Gammarus duebeni : can females be too big? Animal Beha viour 54 :1301 1308. 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. and E. J. Temeles. 1989. The evolution of sexual dimorphism in animal s hypotheses and tests. Trends in Ecology & Evolution 4 :136 138. Heibo, E., C. Magnhagen, and L. A. Vllestad. 2005. Latitudinal variation in life history traits in eurasian perch. Ecology 86 :3377 3386. Ho, C. K., S. C. Pennings, and T. H. Carefoot. 2010 Is diet quality an overlooked mechanism for Bergmann's rule? American Naturalist 175 :269 276. Hobson, K. A., R. T. Alisauskas, and R. G. Clark. 1993. Stable nitrogen isotope enrichment in avian tissues due to fasting and nutritional stress: implications for isotopic analyses of diet. Condor 95 :388 394. Huey, R. B., G. W. Gilchrist, M. L. Carlson, D. Berrigan, and L. Serra. 2000. Rapid evolution of a geographic cline in size in an introduced fly. Science 287 :308 309. Hunt, J. and L. W. Simmons. 2002. Confi dence of paternity and parental care: covariation revealed through the experimental manipulation of the mating system in the beetle, Onthophagus taursus Journal of Evolutionary Biology 15 :784 795. rule, and related phenomena. Ecological Monographs 81 :349 405. Jackson, N. L., D. R. Smith, R. Tiyarattanachai, and K. F. Nordstrom. 2007. Evaluation of a small beach nouri shment project to enhance habitat suitability for horseshoe crabs. Geomorphology 89 :172 185. James Pirri, M. J. 2010. Seasonal movement of the American horseshoe crab Limulus polyphemus in a semi enclosed bay on Cape Cod, Massachusetts (USA) as determined by acoustic telemetry. Current Zoology 56 :575 586. James Pirri, M. J., K. Tuxbury, S. Marino, and S. Koch. 2005. Spawning densities, egg densities, size structure, and movement patterns of spawning horseshoe crabs, Limulus polyphemus within four coastal e mbayments on Cape Cod, Massachusetts. Estuaries 28 :296 313. Jegla, T. C. and J. D. Costlow. 1982. Temperature and salinity effects on development and early posthatch stages of Limulus Pages 103 113 in J. Bonaventura, C. Bonaventura, and S. Tesh, editors. Physiology and Biology of Horseshoe Crabs: Studies on Normal and Stressed Animals. Alan R. Liss, New York.
144 Johansson, F. 2003. Latitudinal shifts in body size of Enallagma cyathigerum (Odonata). Journal of Biogeography 30 :29 34. Johnson, S. L. and H. J. Br ockmann. 2010. Costs of multiple mates: an experimental study in horseshoe crabs. Animal Behaviour 80 : 773 782. Karasov, W. H. and C. Martinez del Rio. 2007. Physiological Ecology: How Animals Process Energy, Nutrients, and Toxins. Princeton University Pre ss, Princeton, NJ. Kaustuv, R. and K. K. Martien. 2001. Latitudinal distribution of body size in Northeastern Pacific marine bivalves. Journal of Biogeography 28 :485 493. Khan, R. A. 2003. Observations on some aspects of the biology of horse shoe crab, Car cinoscorpius rotundicauda (Latreille) on mud flats of Sunderban estuarine region. Records of the Zoological Survey of India 101 :1 2. King, T. L., M. S. Eackles, A. P. Spidle, and H. J. Brockmann. 2005. Regional differentiation and sex biased dispersal amon g populations of the horseshoe crab Limulus polyphemus Transactions of the American Fisheries Society 134 :441 465. Kingsolver, J. G. and R. B. Huey. 2008. Size, temperature, and fitness: three rules. Evolutionary Ecology Research 10 :251 268. Koons, B. F. 1883. Sexual characters of Limulus American Naturalist 17 :1297 1299. Implications for age and size at maturity. Trends in Ecology & Evolution 7 :15 19. mal initial size and adult size of animals: consequences for macroevolution and community structure. American Naturalist 147 :101 114. models explain why ectotherms grow lar ger in cold? Integrative and Comparative Biology 44 :480 493. reproduction: The Pontryagin Maximum Principle solution for the case of age and season dependent mortality. Evolutio nary Ecology Research 1 :423 441. reproduction. Theoretical Population Biology 29 :16 37. perennials with determinate growth. Evolutionary Ecology 1 :231 244.
145 Kreamer, G. and S. Michels. 2009. History of horseshoe crab harvest on Delaware Bay. in J. T. Tanacred i, M. L. Botton, and D. R. Smith, editors. Biology and Conservation of Horseshoe Crabs. Springer, New York. Kruger, O. 2005. The evolution of reversed sexual size dimorphism in hawks, falcons and owls: a comparative study. Evolutionary Ecology 19 :467 486. Laugen, A. T., A. Laurila, K. I. Jnsson, F. Sderman, and J. Meril. 2005. Do common frogs ( Rana temporaria 7 :717 731. Laughlin, R. B. 1981. Sodium, chloride, and water exchange in selected developmen tal stages of the horseshoe crab, Limulus polyphemus Journal of Experimental Marine Biology and Ecology 52 :135 146. Laughlin, R. B. 1983. The effects of temperature and salinity on larval growth of the horseshoe crab, Limulus polyphemus Biological Bullet in 164 :93 103. Lee, P. C., P. Majluf, and I. J. Gordon. 1991. Growth, weaning, and maternal investment from a comparative perspective. Journal of Zoology 225 : 99 114. Leoncini, I., Y. Le Conte, G. Costagliola, E. Plettner, A. L. Toth, M. Wang, Z. Huang, J. M. Bcard, D. Crauser, K. N. Slessor, and G. E. Robinson. 2004. Regulation of behavioral maturation by a primer pheromone produced by adult worker honey bees. Proceedings of the National Academy of Sciences of the United States of America 101 :17559 17564. Leschen, A. S., S. P. Grady, and I. Valiela. 2006. Fecundity and spawning of the Atlantic horseshoe crab, Limulus polyphemus in Pleasant Bay, Cape Cod, Massachusetts, USA. Marine Ecology 27 :54 65. Lindauer, M. 1961. Communication Among Social Bees. Harva rd University Press, Cambridge, MA. Lindenfors, P., J. L. Gittleman, and K. E. Jones. 2007. Sexual size dimorphism in mammals. Pages 16 26 in D. J. Fairbairn, W. U. Blanckenhorn, and T. Szkely, editors. Sex, Size, and Gender Roles: Evolutionary Studies of Sexual Size Dimorphism. Oxford University Press, Oxford. Loveland, R. E. and M. L. Botton. 1992. Size dimorphism and the mating system in ho rseshoe crabs, Limulus polyphemus Animal Behaviour 44 :907 916. MacDonald, D. 1985. The Encyclopedia of Mammals. Guild Publishing, London. Madsen, T. 1983. Growth rates, maturation and sexual size dimorphism in a population of grass snakes, Natrix natrix in southern Sweden. Oikos 40 :277 282.
146 Makarieva, A. M., V. G. Gorshkov, and B. L. Li. 2005. Temperature associated upper limits to body size in terrestrial poikilotherms. Oikos 111 :425 436. Malthus, T. R. 1798. An essay on the principle of population Oxfo rd World's Classics reprint, Oxford. Mangum, C. P., C. E. Booth, P. L. DeFur, H. N. A., R. P. Henry, L. C. Oglesby, and G. Polites. 1976. The ionic environment of hemocyanin in Limulus polyphemus Biological Bulletin 150 :453 467. Martinez del Rio, C. and B O. Wolf. 2005. Mass balance models for animal isotopic ecology. Pages 141 174 in M. A. Starck and T. Wang, editors. Physiological and Ecological Adaptations to Feeding in Vertebrates. Science Publishers, Enfield, New Hampshire. Martinez del Rio, C., N. Wolf, S. A. Carleton, and L. Z. Gannes. 2009. Isotopic ecology ten years after a call for more laboratory experiments. Biological Reviews 84 :91 111. Masaki, S. 1978. Climatic adaptation and species status in the lawn ground cricket II. Body size. Oecologia 35 :343 356. Mattei, J. H., M. A. Beekey, A. Rudman, and A. Woronik. 2010. Reproductive behavior in horseshoe crabs: does density matter? Current Zoology 56 :634 642. Mayer, A. G. 1914. The effects of temperature upon tropical mari ne animals. Papers Tortugas Lab 6 :1 24. Maynard Smith, J. 1958. The Theory of Evolution. Penguin Books, London. Mayr, E. 1991. One Long Argument: Charles Darwin and the Genesis of Modern Evolutionary Thought (Questions of Science). Harvard University Press Cambridge, MA. McCabe, J. and L. Partridge. 1997. An interaction between environmental temperature and genetic variation for body size for the fitness of adult female Drosophila melanogaster Evolution 51 :1164 1174. McCauley, S. J., S. S. Bouchard, B. J. Farina, K. Isvaran, S. Quader, D. M. Wood, and C. M. St Mary. 2000. Energetic dynamics and anuran breeding phenology: insights from a dynamic game. Behavioral Ecology 11 :429 436. McCue, M. D. 2007. Western diamondback rattlesnakes demonstrate physiologica l and biochemical strategies for tolerating prolonged starvation. Physiological and Biochemical Zoology 80 :25 34.
147 McCue, M. D. 2010. Starvation physiology: reviewing the different strategies animals use to survive a common challenge. Comparative Biochemist ry and Physiology, Part A 156 :1 18. McCue, M. D. and E. D. Pollock. 2008. Stable isotopes may provide evidence for starvation in reptiles. Rapid Communications in Mass Spectrometry 22 :2307 2314. McElligott, A. G., M. P. Gammell, H. C. Harty, D. R. Paini, D T. Murphy, J. T. Walsh, and T. J. Hayden. 2001. Sexual size dimorphism in fallow deer ( Dama dama ): do larger, heavier males gain greater mating success? Behavioral Ecology and Sociobiology 49 :266 272. McNab, B. K. 1971. On the ecological significance of 52 :845 854. Schistometopum thomense Evolutionary Ecology Research 8 :1049 1059. Journal of Biogeography 30 :331 351. Michener, R. H. and D. M. Schell. 1994. Stable isotope ratios as tracers in marine aquatic food webs. Pages 138 157 in K. Lajtha and R. H. Michener, editors. Stable Isotopes in Ecology and Environmental Science. Blackwel l Scientific Publications, Oxford Mitchell Olds, T. and R. G. Shaw. 1987. Regression analysis of natural selection: statistical inference and biological interpretation. Evolution 41 :1149 1161. Moore, A. J. 1990. The evolution of sexual dimorphism by sexua l selection the separate effects of intrasexual selection and intersexual selection. Evolution 44 :315 331. Moore, S. 2004. The Taunton Bay assessment. Report to Marine Resouces Committee, Maine legislature. Maine Department of Marine Resources, W. Booth bay Harbor, Maine. Morrison, C. and J. M. Hero. 2003. Geographic variation in life history characteristics of amphibians, a review. Journal of Animal Ecology 72 :270 279. Mousseau, T. A. 1997. Ectotherms follow the converse Bergmann's rule. Evolution 51 :630 632. Mousseau, T. A. and D. A. Roff. 1989. Adaptation to seasonality in a cricket: patterns of phenotypic and genotypic variation in body size and diapause expression along a cline in season length. Evolution 43 :1483 1496.
148 Nylin, S. and L. Svrd. 1991 Latitudinal patterns in the size of European butterflies. Holarctic Ecology 14 :192 202. Oelbermann, K. and S. Scheu. 2002. Stable isotope enrichment (delta N 15 and delta C 13) in a generalist predator ( Pardosa lugubris Araneae : Lycosidae): effects of prey quality. Oecologia 130 :337 344. Olalla Tarraga, M. A. and M. A. Rodriguez. 2007. Energy and interspecific body size patterns of amphibian faunas in Europe and North America: Anurans follow Bergmann's rule, urodeles its converse. Global Ecology and Bio geography 16 :606 617. Palumbi, S. R. and B. A. Johnson. 1982. A note on the influence of life history stage on metabolic adaptation: the responses of Limulus eggs and larvae to hypoxia. Pages 115 124 in J. Bonaventura, C. Bonaventura, and S. Tesh, editors. Physiology and Biology of Horseshoe Crabs: Studies on Normal and Stressed Animals. Alan R. Liss, New York. Park. 1949. Application of the converse Bergmann principle to the carabid beetle, Dicaelus purpuralus Physiological Zoology 22 :359 372. Partridge, L. and J. A. Coyne. 1997. Bergmann's rule in ectotherms: is it adaptive? Evolution 51 :632 635. Paterson, J. D. 1990. Bergmann's rule is invalid: a reply to Geist. Canadian Journal of Zoology 68 :1610 1612. Payne, R. B. 1984. Sexual selection, lek and arena behavior, and sexual size dimorphism in birds. Ornithological Monographs 33 :1 52. Penn, D. and H. J. Brockmann. 1994. Nest site selection in the horseshoe crab, Limulus polyphemus Biological Bulletin 187 :373 384. Penn, D. and H. J. Brockmann. 1995. Age bi ased stranding and righting in male horseshoe crabs, Limulus polyphemus Animal Behaviour 49 :1531 1539. Penry, D. L. 1993. Digestive constraints of diet selection. Pages 32 55 in R. W. Hughes, editor. Diet Selection: An Interdisciplinary Approach to Foragi ng Behavior. Blackwell Scientific, Oxford. Perrin, N. 1995. About Berrigan and Charnov's life history puzzle. Oikos 73 :137 139. Peters, R. H. 1983. The Ecological Implications of Body Size. Cambridge University Press, Cambridge. Pomerat, C. M. 1933. Mating in Limulus polyphemus Biological Bulletin 64 :243 252.
149 Purvis, A. and P. H. Harvey. 1995. Mammal life history evolution: a comparative test of 237 : 259 283. Ramsey, F. L. and D. W. Schafer. 1997. The Statistical Sleuth: a Course in Methods of Data Analysis. Duxbury Press, Belmont, CA. Reeve, M. W., K. Fowler, and L. Partridge. 2000. Increased body size confers greater fitness at lower experimental temperature in male Drosophila melanogaster Journal of Evolutionary Biology 13 :836 844. Reiss, M. J. 1989. The Allometry of Growth and Reproduction. Cambridge University Press, Cambridge. Reynolds, W. W. and M. E. Casterlin. 1979. Thermoregulatory behavior and diel activity of Limulus polyphemus. Pages 47 59 in E. Cohen, editor. Biomedical Applications of the Horseshoe Crab. Alan R. Liss, New York. Riska, B. 1981. Morphological variation in the horseshoe crab Limulus polyphemus Evolution 35 :647 658. Robertson, J. G. M. 1986a. Female choice, mating strategies, and the role of vocalizations in the Australian frog Uperoleia rugosa Animal Behaviour 34 :773 784. Robertson, J. G. M. 1986b. Male territoriality, fighting, and assesment of fighting ability in the Australian frog, Uperoleia rugosa Animal Be haviour 34 :763 772. Roff, D. A. 2002. Life History Evolution. Sinauer, Sunderland, MA. Ropes, J. W. 1961. Longevity of the horseshoe crab, Limulus polyphemus (L.). Transactions of the American Fisheries Society 90 :79 80. Rosenzweig, M. L. 1968. Net primar y productivity of terrestrial communities: prediction from climatological data. American Naturalist 102 :67 74. Rudloe, A. E. 1980. The breeding behavior and patterns of movement of horseshoe crabs, Limulus polyphemus in the vicinity of breeding beaches in Apalachee Bay, Florida. Estuaries 3 :177 183. Sagarin, R. D., S. D. Gaines, and B. Gaylord. 2006. Moving beyond assumptions to understand abundance distributions across the ranges of species. Trends in Ecology & Evolution 21 :524 530. Schaller, S. Y. 2002. Horseshoe crab ( Limulus polyphemus ) spawning surveys in Maine, 2001. in American Fisheries Society Annual Meeting. American Fisheries Society, Baltimore, MD.
150 Schaller, S. Y., C. C. Chabot, and W. H. Watson, III. 2010. Seasonal movements of American horsesho e crabs Limulus polyphemus in the Great Bay Estuary, New Hampshire (USA). Current Zoology 56 :587 598. Schoener, T. W. 1967. Ecological significance of sexual dimorphism in size in the lizard Anolis cospersus Science 155 :474 477. Schutze, M. K. and A. R. C larke. 2008. Converse Bergmann cline in a Eucalyptus herbivore, Paropsis atomaria Olivier (Coleoptera: Chrysomelidae): phenotypic plasticity or local adaptation? Global Ecology and Biogeography 17 :424 431. Schwab, R. L. 2006. Mating group formation and fem ale assessment by satellite male horseshoe crabs ( Limulus polyphemus ). University of Florida, Gainesville. Schwab, R. L. and H. J. Brockmann. 2007. The role of visual and chemical cues in the mating decisions of satellite male horseshoe crabs, Limulus poly phemus. Animal Behaviour 74 :837 846. Seigel, R. A. and N. B. Ford. 1987. Reproductive ecology. Pages 210 252 in R. A. Seigel, J. T. Collins, and S. S. Novak, editors. Snakes: Ecology and Evolutionary Biology Macmillan, New York. Sekiguchi, K. 1988. Ecology. Pages 50 68 in K. Sekiguchi, editor. Biology of Horseshoe Crabs. Science House, Tokyo. Sekiguchi, K. and C. N. Shuster, Jr. 2009. Limits on the glodal distribution of horseshoe crabs (Limulacea): lessons learned from two lifeti mes of observation: Asia and America. in J. T. Tanacredi, M. L. Botton, and D. R. Smith, editors. Biology and Conservation of Horseshoe Crabs. Springer, New York. Selander, R. K. 1966. Sexual dimorphism and differential niche utilization in birds. Condor 68 :113 151. Shine, R. 1988. The evolution of large body size in females a critique of darwins fecundity advantage model. American Naturalist 131 :124 131. Shine, R. 1989a. Ecological causes for the evolution of sexual dimorphism: a review of the evidence. T he Quarterly Review of Biology 64 419 461. Shine, R. 1989b. Ecological influences on the evolution of vertebrate viviparity. Life Sciences Research Report 45 :263 278. Shine, R. 2005. Life history evolution in reptiles. Annual Reviews of Ecology, Evolution and Systematics 36 : 23 46. Shuster, C. N., Jr. 1954. A horseshoe "crab" grows up. Ward's Natural Science Bulletin 28 :1 6.
151 Shuster, C. N., Jr. 1955. On morphometric and serological relationships within the Limulidae, with particular reference to Limulus pol yphemus L. Ph.D. New York University. Shuster, C. N., Jr. 1957. Xiphosura (with special reference to Limulus polyphemus ). Geological Society of America 67 :1171 1174. Limulus polyphemu s Pages 3 26 in E. Cohen, editor. Biomedical Applications of the Horseshoe Crab. Alan R. Liss, New York. Shuster, C. N., Jr. 1982. A pictorial review of the natural history and ecology of the horseshoe crab, Limulus polyphemus with reference to other Lim ulidae. Pages 1 52 in J. Bonaventura, C. Bonaventura, and S. Tesh, editors. Physiology and Biology of Horseshoe Crabs: Studies on Normal and Environmentally Stressed Animals. Alan R. Liss, New York. Shuster, C. N., Jr. and M. Botton. 1985. A contribution t o the populations biology of horseshoe crabs, Limulus polyphemus in Delaware Bay. Estuaries 8 :363 372. Shuster, C. N., Jr., C. H. P. Langrall, and R. C. Shevock. 1993. Biometric analysis of sexual dimorphism in mated pairs of Limulus polyphemus from Delaw are Bay. Shuster, C. N., Jr. and K. Sekiguchi. 2003. Growing up takes about ten years and eighteen stages. Pages 103 132 in C. N. Shuster Jr., R. B. Barlow, and H. J. Brockmann, editors. The American Horseshoe Crab. Harvard University Press, Cambridge. Shu ster, C. N., Jr. and K. Sekiguchi. 2009. Basic habitat requirements of the extant species of horseshoe crabs (Limulacea). Pages 115 129 in J. T. Tanacredi, M. L. Botton, and D. R. Smith, editors. Biology and Conservation of Horseshoe Crabs. Springer, New Y ork. Shuster, S. M. and M. J. Wade. 2003. Mating Systems and Strategies. Princeton University Press, Princeton, NJ. Slatkin, M. 1984. Ecological causes of sexual dimorphism. Evolution 38 :622 630. Smith, D. R. 2007. Effect of horseshoe crab spawning density on nest disturbance and exhumation of eggs: a simulation study. Estuaries and Coasts 30 :287 295. Smith, D. R., L. J. Brousseau, M. T. Mandt, and M. J. Millard. 2010. Age and sex specific timing, frequency, and spatial distribution of horseshoe crab spawning in Delaware Bay: insights from a large scale radio telemetry array. Current Zoology 56 :563 574.
152 Smith, D. R., M. T. Mandt, and P. D. M. MacDonald. 2009a. Proximate causes of sexual size dimorphism in horseshoe crabs ( Limulus polyphemus ) of the Delaware Bay. Journal of Shellfish Research 28 :405 417. Smith, D. R., M. J. Millard, and R. H. Carmichael. 2009b. Comparative status and assesment of Limulus polyphemus with emphasis on the New England and Delaware Bay populations. in J. T. Tanacredi, M. L. Botton, and D. R. Smith, editors. Biology and Conservat ion of Horseshoe Crabs. Springer, New York. Smith, D. R., P. S. Pooler, B. L. Swan, S. F. Michels, W. R. Hall, P. J. Himchak, and M. J. Millard. 2002. Spatial and temporal distribution of horseshoe crab ( Limulus polyphemus ) spawning in Delaware Bay: Implic ations for monitoring. Estuaries 25 :115 125. Sokoloff, A. 1978. Observations on populations of the horseshoe crab Limulus (=Xiphorsura) polyphemus Researches on Population Ecolology 19 222 236. Stamps, J. A. 1983. Sexual selection, sexual dimorphism, and teritoriality. Pages 162 204 in R. B. Huey, E. R. Pianka, and T. W. Schoener, editors. Lizard Ecology: Studies of a Model Organism. Harvard University Press, Cambridge. Stearns, S. C. 1992. The Evolution of Life Histories. Oxford University Press, Oxford. Stephens, P. R. and J. J. Wiens. 2009. Evolution of sexual size dimorphisms in emydid Evolution 63 :910 925. Stillwell, C. R. 2010. Are latitudinal clines in body size adaptive? Oikos 119 :1387 1390. Stillwell, R. C., G. E. Morse, and C. W. Fox. 2007. Geographic variation in body size and sexual size dimorphism of a seed feeding beetle. American Naturalist 170 :358 369. Stoks, R., M. De Block, W. W. Van Dooerslaer, and J. Rolff. 2006. Tim e constraints mediate predator induced plasticity in immune function, condition, and life history. Ecology 87 :809 815. Strong, K. W. and G. R. Daborn. 1980. The influence of temperature on energy budget variables, body size, and seasonal occurrence of the isopod Idotea baltica (Pallas). Canadian Journal of Zoology 58 :1992 1996. Suggs, D. N., R. H. Carmichael, S. P. Grady, and I. Valiela. 2002. Effects of individual size on pairing in horseshoe crabs. Biological Bulletin 203 :225 227. Sugita, H. 1988. Environ mental adaptations of embryos. Pages 195 224 in K. Sekiguchi, editor. Biology of Horseshoe Crabs. Science House, Tokyo.
153 Szkely, T., T. Lislevand, and J. Figuerola. 2007. Sexual size dimorphism in birds. Pages 27 37 in D. J. Fairbairn, W. U. Blanckenhorn, and T. Szkely, editors. Sex, Size, and Gender Roles: Evolutionary Studies of Sexual Size Dimorphism. Oxford University Press, Oxford. Taborsky, M. 1998. Sperm competition in fish: 'bourgeois' males and parasitic spawning. Trends in Ecology & Evolution 13 : 222 227. Taborsky, M., R. F. Oliveira, and H. J. Brockmann. 2008. The evolution of alternative reproductive tactics: concepts and questions. Pages 1 21 in R. F. Oliveira, M. Taborsky, and H. J. Brockmann, editors. Alternative Reproductive Tactics. An Integ rative Approach. Cambridge University Press, Cambridge. Teder, T. and T. Tammaru. 2005. Sexual size dimorphism within species increases with body size in insects. Oikos 108 :321 334. Temeles, E. J., I. L. Pan, J. L. Brennan, and J. N. Horwitt. 2000. Evidence for ecological causation of sexual dimorphism in a hummingbird. Science 289 :441 443. Thompson, M. 1998. Assessments of the population biology and critical habitat for the horseshoe crab, Limulus polyphemus in the South Atlantic Bight. Medical University of South Carolina, University of Charleston. Thornhill, T. and J. Alcock. 1983. The Evolution of Insect Mating Systems. Harvard University Press, Cambridge. Towle, D. W., C. P. Mang um, B. A. Johnson, and N. A. Mauro. 1982. The role of the coxal gland in ionic, osmotic, and pH regulation in the horseshoe crab Limulus polyphemus Pages 147 172 in J. Bonaventura, C. Bonaventura, and S. Tesh, editors. Physiology and Biology of Horseshoe Crabs: Studies on Normal and Stressed Animals. Alan R. Liss, New York. van der Have, T. M. 2002. A proximate model for thermal tolerance in ectotherms. Oikos 98 :141 155. van der Have, T. M. and G. de Jong. 1996. Adult size in ectotherms: temperature effect s on growth and differentiation. Journal of Theoretical Biology 18 :329 340. Van Voorhies, W. A. 1996. Bergmann size clines: a simple explanation for their occurrence in ectotherms. Evolution 50 :1259 1264. Van Voorhies, W. A. 1997. On the adaptive nature of Bergmann size clines: a reply to Mousseau, Partridge and Coyne. Evolution 51 :635 640. Virgs, E., R. Kowalczyk, A. Trua, A. de Marinis, J. G. Mangas, J. M. Barea Azcn, and E. Geffen. 2011. Body size clines in the European badger and the abundant centre h ypothesis. Journal of Biogeography 38 :1546 1556.
154 von Bertalanffy, L. 1960. Principles and theory of growth. Pages 137 259 in W. N. Nowinski, editor. Fundamental Aspects of Normal and Malignant Growth. Elsevier, Amsterdam. Walters, R. J. and M. Hassall. 200 6. The temperature size rule in ectotherms: may a general explanation exist after all? American Naturalist 167 :510 523. Warner, R. R. and E. T. Schultz. 1992. Sexual selection and male characteristics in the bluehead wrasse, Thalassoma bifasciatum mating site acquisition, mating site defense, and female choice. Evolution 46 :1421 1442. Warren, M., M. A. McGeoch, S. W. Nicolson, and S. L. Chown. 2006. Body size patterns in Drosophila inhabiting a mesocosm: interactive effects of spatial variation in temperature and abundance. Oecologia 149 :245 255. Watson, W. H., III, S. Y. Schaller, and C. C. Chabot. 2009. The relationship between small and large scale movements of horseshoe crabs in the Great Bay estuary and Limulus behavior in the laboratory. Page s 163 182 in J. T. Tanacredi, M. L. Botton, and D. R. Smith, editors. Biology and Conservation of Horseshoe Crabs. Springer, New York. Watt, C., S. Mitchell, and V. Salewski. 2010. Bergmann's rule; a concept cluster? Oikos 119 :89 100. Weckerly, F. W. 1998. Sexual size dimorphism: influence of mass and mating systems in the most dimorphic mammals. Journal of Mammalogy 79 :33 52. Wells, K. D. 1977a. The social behavior of anuran amphibians Animal Behaviour 25 :666 693. Wells, K. D. 1977b. Territoriality and mal e mating success in the green frog ( Rana clamitans ). Ecology 58 :750 762. Wenner, E. and M. Thompson. 2000. Evaluation of harvesting impacts and population trends for Limulus polyphemus in South Carolina. Page 91 in S. C. D. o. N. Resources, editor. Marine Resources Research Institute, Charleston. Whitfield, C. W., A. M. Cziko, and G. E. Robinson. 2003. Gene expression profiles in the brain predict behavior in individual honey bees. Science 302 :296 299. Widemo, F. 19 98. Alternative reproductive strategies in the ruff, Philomachus pugnax : a mixed ESS? Animal Behaviour 56 :329 336. Wiklund, C. and T. Fagerstrom. 1977. Why do males emerge before females a hypothesis to explain incidence of protandry in butterflies. Oeco logia 31 :153 158. Wiklund, C. and C. Solbreck. 1982. Adaptive versus incidental explanations for the occurrence of protandry in a butterfly, Leptidea sinapis Evolution 36 :56 62.
155 Williams, G. C. 1966. Adaptation and Natural Selection. Princeton University Press, Princeton. Withers, G. S., H. E. Fahrbach, and G. E. Robinson. 1993. Selective neuroanatomical plasticity and division of labor in the honey bee. Nature 364 :238 240. Wolverton, S., M. A. Huston, J. H. Kennedy, K. Cagle, and J. D. Cornelius. 2009. Co nformation to Bergmann's rule in white tailed deer can be explained by food availability. American Midland Naturalist 162 :403 417. Woods, H. R. 1999. Egg mass size and cell size: effects of temperature on oxygen distribution. American Zoologist 39 :244 252. Yom Tov, Y. and E. Geffen. 2006. Geographic variation in body size: the effect of ambient temperature and precipitation. Oecologia 148 :213 218. Zamudio, K. R. and L. M. Chan. 2008. Alternative reproductive tactics in amphibians. Pages 300 331 in R. F. Oli veira, M. Taborsky, and H. J. Brockmann, editors. Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge
156 BIOGRAPHICAL SKETCH Matthew Smith is a broadly trained organismal biologist generally interested i n the evolution and maintenance of phenotypic variation. His research focuses on understanding how natural selection and sexual selection influence reproductive behavior, morphology, and life history strategies. He received a Bachelor of Arts in biology f rom Earlham College in 1997. In 2004, he received a Master of Science in w ildlife s cience from the University of Arizona. He received a Doctor of Philosophy in zoology from the University of Florida in summer 2012.