MATING GROUP FORMATION AND FE MALE ASSESSMENT BY SATELLITE MALE HORSESHOE CRABS ( Limulus polyphemus ) By RACHEL L. SCHWAB A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006
Copyright 2006 by Rachel L. Schwab
To Pa. My love and thanks.
iv ACKNOWLEDGEMENTS I am ever grateful to my advisor, Jane Brockmann, for her me ntal and physical support in and out of the field and her immens e assistance during all aspects of this study. Her excitement for horseshoe crabs is contagio us and I could not have asked for a better mentor. I thank my co-advisor, Colette St. Mar y, for her helpful suggestions and her willingness to discuss my work no matter if I stopped her in the elevator, approached her at a social gathering, or called her on the weekends. These actions additionally showed her support and dedication. I also thank Tom Frazer for his positive and accepting personality that puts one at ease, fresh perspective on behavior , and careful thought and critique. This study would not have been possible without Captain Al Dinsmore and Henry Coulter, boating me to and from the island with all my supplies, assuring that I had enough fuel in the generator (to last at leas t half of the week), and providing me with much-needed rat traps. Additional thanks go to the University of Florida Marine Laboratory at Seahorse Key a nd its Director, Harvey Lillyw hite, and the Suwannee River National Wildlife Refuge for thei r support of this project. I am very appreciative for Russ McCarty w ho graciously volunteered his time and led me through every step of horseshoe cr ab model-making. I thank Ben Olaivar and Kent Vliet for their assistance in obtaining horseshoe crab carapaces and their advice on marine supplies throughout the project.
v I want to thank Nancy Targett for inviting me to join her lab at the College of Marine Studies, University of Delaware, Le wes, DE, and Kirsten Wakefield for training me in conducting horseshoe crab egg extractio ns and assisting with several other aspects of the project in Delaware. Additionally, I thank numerous gr aduate students and undergraduate visiting researchers at the University of Delaware for their help with data collection. I thank the St. Mary, Osenberg, Bolker and the Brockmann lab groups for helpful research discussions and valuable comments on the manuscript. I extend my gratitude to Jack Szczepansk i, Hope Schwab, Gary Schwab, Diana Schwab, Tobey Curtis, Amber Pitt, Jena C hojnowski and Chris Ladd for their company and assistance with data collection in the field. I am very thankful for Carrie Newsom , a wonderful research assistant who struggled with me through the ups and downs of my first season on Seahorse Key. Extra special thanks go to Kym Mitala and Danny Turbet for their invaluable assistance, support, dedication, good humor, and friendship throughout the field season and beyond. Additional thanks go to Danny for his rat-trapping skills. Lastly, I would like to thank my fami ly, Hope, Gary, and Diana Schwab, and friends for their endless support th roughout this entire experience. This project was funded in part by a gr ant from the University of Florida Foundation.
vi TABLE OF CONTENTS page ACKNOWLEDGEMENTS...............................................................................................iv LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ..x CHAPTER 1 GENERAL INTRODUCTION....................................................................................1 2 DO SATELLITE MALE HORSESHOE CRABS CHOOSE FEMALES BASED ON AN IDEAL FREE DISTRIBUTION?...................................................................5 Introduction................................................................................................................... 5 Methods........................................................................................................................ 9 Location.................................................................................................................9 Horseshoe Crab Measurements...........................................................................10 Tagging Horseshoe Crabs....................................................................................11 Population Census 2004 and 2005......................................................................11 Fecundity Measurements.....................................................................................11 Female Size Measurements.................................................................................12 Calculating the Predicted Ideal Free Distribution of Satellites...........................13 Data Analysis.......................................................................................................14 Results........................................................................................................................ .14 Fecundity Measurements.....................................................................................14 Female Size Measurements.................................................................................14 Observed Distribution Versus Predicted IFD......................................................17 Discussion...................................................................................................................19 3 HOW DO SATELLITE MALES DISTI NGUISH AMONG FEMALES OF VARIABLE FECUNDITY?.......................................................................................25 Introduction.................................................................................................................25 Methods......................................................................................................................31 Location...............................................................................................................31 Horseshoe Crab Measurements...........................................................................31 Tagging Horseshoe Crabs....................................................................................33 Data Sets..............................................................................................................33
vii Female Nesting Data...........................................................................................33 Egg Attractiveness Data......................................................................................34 Experimental Manipulations...............................................................................35 Models used during experimental manipulations.........................................37 Visual cues experiment................................................................................39 Chemical cues experiment I: Attractive eggs...............................................41 Chemical cues experiment II: Supernatant A...............................................44 Data analysis................................................................................................46 Results........................................................................................................................ .47 Female Nesting Measurements............................................................................47 Egg Attractiveness Data......................................................................................47 Visual Cues Experiment......................................................................................49 Chemical Cues Experiment I: Attractive Eggs....................................................49 Chemical Cues Experiment II: Supernatant A....................................................50 Discussion...................................................................................................................51 4 GENERAL DISCUSSION.........................................................................................59 LIST OF REFERENCES...................................................................................................66 BIOGRAPHICAL SKETCH.............................................................................................72
viii LIST OF FIGURES Figure page 1 Frequency distribution of female fecundity (mean clutch size in mL) (1 mL = 88 eggs) (n = 68)...........................................................................................................15 2 Scatter plot showing the relationship be tween female size (inter-ocular distance) (cm) and fecundity (average volum e of eggs per clutch in mL)..............................15 3 Frequency distribution of nesting female body sizes (inter-ocular distance) (cm) (n = 366) collected over 14 tides (a)........................................................................16 4 The observed and predicted frequency distribution of ma ting group sizes of nesting females based on an IFD for five tides from the 2004 census.....................18 5 This nesting female (F) has an attached male (A) plus six satellite males occupying different spawning positions...................................................................23 6 Four female horseshoe crabs (buried in the sand) nesting within 5 m of each other attract very different numbers of satellites......................................................26 7 This female horseshoe crab is nestin g in the sand partially buried up to her lateral eyes................................................................................................................28 8 The exposed carapace height....................................................................................32 9 Procedure for collecting Egg Attractiveness data....................................................36 10 An experimental arena.............................................................................................41 11 Scatter plot showing the correlation betw een female size (inter-ocular distance) (cm) and the exposed carapace height (cm).............................................................48 12 Scatter plot showing the relationship be tween female size (inter-ocular distance) and female mating group size...................................................................................48 13 Mean size (inter-ocular di stance) (cm) of females with mating groups of three or more satellite males (many) and less than three satel lite males (few).....................49 14 Mean number of satellite males to ente r the arena, approach within 10 cm, and join the egg and no egg model pa irs (n = 46, 10-minute trials)...............................51
ix 15 Mean time (sec) satellite males remain ed joined to the egg and no egg female models (n = 46, 10-minute trials).............................................................................51
x Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science MATING GROUP FORMATION AND FE MALE ASSESSMENT BY SATELLITE MALE HORSESHOE CRABS ( Limulus polyphemus ) By Rachel L. Schwab May 2006 Chair: H. Jane Brockmann Cochair: Colette M. St. Mary Major Department: Zoology The ideal free distribution (IFD) model st ates that individu als should distribute themselves among resources such as food, ovi position sites, or mates, in a way that maximizes their fitness. Satellite male horseshoe crabs, Limulus polyphemus , form spawning groups around some nesting pairs (f emale plus attached male) while leaving others to nest alone. Furthermore, females of pairs that attract satellites are larger and lay more eggs than females of pairs that do not at tract satellites. To better understand satellite male mating decisions and the formation of mating groups in the horseshoe crab, first I examined whether satellite males distribu te themselves among females of variable fecundity following an IFD. Then, I identified cues satellite males use to locate nesting pairs and assess female fecundity. I collected data on female size, fecundity and group size during the spring breeding season (February-May) 2004-2005. I combined th ese data with previous measures of
xi paternity from the same site to create a pr edicted IFD for satellites around females of variable fecundity. At a very low OSR (operati onal sex ratio), satelli te males distributed themselves among nesting females following the predicted IFD. But at high sex ratios, there were more pairs than predicted with zero or one satellite, fewer pairs with two satellites, and more pairs with three to seven satellites. Thus, satellites were undermatching the fecundity of some females and overmatching the fecundity of others. Since satellite males preferentially join large females with higher fecundity, they must be using some cue from the nesting pair to locate and evaluate female quality. I experimentally tested whether satellite males use visual cues to discriminate among large and small females and if satellites use chem ical cues from eggs to distinguish among females based directly on fecundity. I conducte d experiments in the fi eld at Seahorse Key during the 2005 breeding season, in which I obser ved the behavior of satellite males as they approached cement models of horses hoe crab pairs. Satellite males were significantly more likely to appr oach and join model pairs with large females than with small females. However, there was no differe nce in the length of time satellites remained joined to the models. When chemical cues were examined, there was no difference between the numbers of satellite males that approached or joined the model pairs that were placed over bags containing eggs and those placed over empty bags (model females were of equal size). However, of the satellites that joined models, those that joined the pairs with eggs remained attached for signi ficantly longer than those that joined the control model. In summary, this study shows that satellite male horseshoe crabs use a hierarchy of cues, first visual for longer distan ces, then chemical at shorter range to make group joining decisions.
1 CHAPTER 1 GENERAL INTRODUCTION Individuals should distri bute themselves among resources in a manner that maximizes their fitness, called an ideal fr ee distribution (IFD) (Fretwell & Lucas 1970; Goss-Custard & Sutherland 1997). When resources are equal in value, individuals should distribute themselves evenly among those resources. However, when resources are unequal then group sizes should differ, with larger groups around the higher quality resources and smaller groups or no individuals around the lower quality resources (Parker & Sutherland 1986). Horseshoe crabs, Limulus polyphemus , are an ideal system for studying the nature of group formation. Thousands of paired males and females arrive on a beach in amplexus and nest in the sand. Unattached males approach the beach alone joining the nesting pairs and participating in group spawning. Some pairs attract many unattached, satellite males while other pairs nearby on the beach remain alone, yielding highly variable mating group sizes. Differences in female quality or nesti ng location could explain this variation. Hassler (1999) found that there was no difference in nest site quality between pairs with satellites and pairs without satellites, but there was a difference in female qualities. The females of pairs nesting with satellites are, on average, larger and heavier than females nesting without satellites (Brockmann 1996). A dditionally, females nesting with satellites lay significantly more eggs than females without satellites (Hassler 1999; but see Brockmann 1996).
2 In Chapter 2 I examine whether the sate llite males distribute themselves among females of variable fecundity following the predicted IFD. To accomplish this, I combine previous paternity data (Brockmann et al. 2000 ) with data I collect on female fecundity, body size, and the distribution of mating group sizes on the beach. If the observed mating group sizes are not significantly different from those predicted by an IFD, then this suggests that satellites are able to assess the differences in female fecundity and distribute themselves freely among those females wit hout interference from other satellites. If satellite males preferentially join large females with higher fecundity, they must be using some cue from the nesting pair to locate their position and evaluate female quality. Satellites might be using vision to de tect a phenotypic indicator of fecundity such as body size (e.g., Calvo & Molina 2005; Melo & Le Clus 2005; Prado & Haddad 2005); chemoreceptors to sense an odorous by-product of other activities such as laying eggs or a specific signal such as a pheromone; hearing or touch to detect sounds or vibrations from such things as scratching and rubbing between carapaces which could reveal a femaleâ€™s fecundity; or multiple modalities to identify a combination of mating cues. Satellite males are known to use vision to identify and discriminate among femalesized objects offshore (Barlow et al. 1982). Males prefer black castings of female horseshoe crabs over similar-si ze castings of other shapes and colors (Barlow et al. 1982). Thus, it seems likely that satellites may al so use vision to assess the size of nesting females and join groups accordingly. However, when females nest they are partially buried under the sand (Brockmann 1996). This would seem to preclude the use of vision. Therefore, it seemed most likely
3 that satellites were using chemical cues alone or in addition to visual cues for discriminating among potential mates. Limulus have millions of chemoreceptor cells at chemoreceptor sites distributed around their body, but concentrated on their limbs (Patten 1894; Hayes 1985). Studies have shown that these chemoreceptors may be used for close-range and distance chemoreception (Wyse 1971; Quinn et al. 1998 ). Furthermore, Hassler and Brockmann (2001) discovered that satellites are more attracted to a model of a female horseshoe crab placed over a sponge filled with seawater fr om underneath a nesting female versus a model placed over a sponge filled with just seawater. They argue that the sponge collected whatever chemicals were being releas ed by the female and satellites were using this information to discriminate among females. Although the previous studies on vision have shown that satellite males can use vision to detect female-sized objects, it is not known whether satelli tes can use vision to assess female size (Barlow et al. 1982). Furt hermore, prior studies determined that satellites are able to discriminate among fe males based on some sort of chemical cue released during nesting but the specific source of the cue remains unknown (Hassler & Brockmann 2001). In Chapter 3 I examine the us e of visual cues and chemical cues as possible proximate mechanisms by which satel lite males detect differences in fecundity. In particular, I experimentally test whether sa tellite males use visual cues to discriminate among large and small females for mating a nd whether satellites use chemical cues located in horseshoe crab eggs to discriminate among females based directly on fecundity. These experiments eliminated the po ssibility that males might be using other sources of information such as auditory or tactile cues.
4 The ideal free distribution m odel states that a satellite male should join a spawning group when he will achieve greater reproductive success by joining a group than searching for a lone pair (Fretwell & Lucas 1970; Goss-Custard & Sutherland 1997). Consequently, if satellites follow the predicti ons of an IFD, mating group sizes should be variable with larger groups surrounding pairs with females of higher fecundity. For this to occur, satellites must be able to assess th e number of other males already participating in the group spawning and female fecundity. Th erefore, in order to better understand satellite male mating decisions and the fo rmation of mating groups, we must examine both the nature of satellite distributio n around nesting pairs and the proximate mechanisms satellites use to distribute themselves around pairs. Additionally, we must consider the costs and benefits of every d ecision due to their infl uence on the predicted outcome.
5 CHAPTER 2 DO SATELLITE MALE HORSESHOE CRABS CHOOSE FEMALES BASED ON AN IDEAL FREE DISTRIBUTION? Introduction Distributional patterns of animals are cen tral to our understa nding of behavioral ecology. The Ideal Free Distribu tion (IFD) predicts that in dividuals should distribute themselves in such a way that use of avai lable resources will ma ximize their fitness (Fretwell & Lucus 1970). If res ources in the environment are patchy but of equal quality, individuals should distribute themselves even ly among those resources. If resources are unequal, the distribution of individuals will be uneven, with the higher quality resources attracting more individuals and the lower quality resources attracting fewer. As a consequence, the resulting fitn ess of individuals in the popula tion will be equal (Parker & Sutherland). The basic model of IFD makes th ree main assumptions: 1) all individuals have equal competitive abilities; 2) individuals are able to move anywhere within their environment at no cost; and 3) individuals know the relative quality of every resource patch (Fretwell & Lucus 1970). If these assump tions are supported, th e said distributions should occur. Although the IFD model is most often used to describe the dist ribution patterns of animals with respect to foraging behavior (e.g., Ward et al. 2000; Jackson et al. 2004; Ramp & Coulson 2004), it has also been used to examine the distribution of individuals with respect to other resour ces such as oviposition site s (Whitman 1980; Korona 1990; Blanckenhorn et al. 2000) or mates (Nishi da 1993; Huhta et al. 1998; Widemo 1998;
6 Blanckenhorn et al. 2000). IFD may also be us eful to explore the distribution of males relative to females in species that form mating groups. In species with external fertilization, a mating group can be defined as one female that releases eggs and multiple males that release sperm, which compete to fertilize those eggs (Roberts et al. 1999; Prado & Haddad 2003; Byrne & Roberts 2004). In many species, males are thought to incur considerable costs from mating es pecially when sperm competition is high (Dewsbury 1982; Nakatsuru & Kramer 1982; Ol sson et al. 1997; Wedell et al. 2002). Thus, males should benefit by accu rately assessing female quali ty, selecting to mate with the female that would provide the greatest repr oductive fitness. If a male is following an IFD, and resources, in this case females, are unequal in quality, then either choosing to join a large group and engage in high levels of sperm competition for a higher quality mate or avoiding high levels of sperm compe tition and mate with a less desirable female, could both result in equal ne t fitness for that male. The horseshoe crab, Limulus polyphemus , is an ideal species for studying IFD and itâ€™s predictions about the formation of va riable mating group sizes. At high tide on the new and full moons during the spring and summ er, paired males and females arrive on a beach in amplexus and nest in the sand. So me nesting couples (female plus attached male) remain as pairs whereas others only a meter or two away are encircled with as many as ten or more unattached, satellite males (Brockmann 1996). Within minutes the female lays a clutch of eggs and the attach ed male and any participating satellite males release their aquatic, free-swimming sperm fertilizing the eggs externally. Next, the female, attached male, and any satellite male s push ten to twenty centimeters forward in the sand (Brockmann 1990, 1996) and the female lays a second clutch of eggs and the
7 males release a second round of sperm. Nesti ng continues through an average of three to five clutches before the male-female pair l eaves the beach and the satellites disperse (Brockmann 1996). Females, together with th eir attached male, may nest again over the next few days during a one-week period ar ound the time of the new or full moon, but rarely across tidal cycles (Cohen & Brockmann 1983; Ba rlow et al. 1986). Males commonly return to the beach across many tides either attached to a new mate or as satellites. Both satellite and attached males that are part of large mating groups participate in strong sperm competition to fertilize eggs. If no satellites are present, the attached male fertilizes 100% of the eggs in a nest (Bro ckmann et al. 1994; Brockm ann et al. 2000). In a pair with one satellite, on aver age, the attached male fertiliz es about 60% and the satellite male fertilizes approximately 40% of the eggs in the nest (Brockmann et al. 1994; Brockmann et al. 2000). When tw o satellites join a pair, th ey achieve approximately the same proportion of fertilizations at ab out 40% each, significantly reducing the fertilization success of the att ached male (Brockmann et al. 2000). However, when a third or forth satellite male joins a group, the mean fertilization success of all the satellites drops to approximately 20% (Brockmann et al. 2000). These results suggest that it should be adva ntageous for a satellite male to join a pair with no or few satellites rather than join a pair with many satellites. However, large mating group sizes up to as many as four to seven satellites are commonly observed (Brockmann 1996; pers. obs. from populat ion census 2004, 2005). Furthermore, mating group sizes are uneven and do not follow a poisson distribution (which would be indicative of random group formation) (Bro ckmann 1996). More pairs than expected fail
8 to attract satellites or attract many satell ites and fewer pairs than expected have intermediate numbers of sate llites (Brockmann 1996). On aver age, mating groups of four to seven are associated with 13 % of the pa irs on the beach and at the same time 51 % of pairs have no satellites (Brockmann 1996). If satellites could achieve greater reproductive success by spreading out among pairs instead of clumping around some pairs and leaving others alone, why are these vast disc repancies in mating group size observed? A number of hypotheses, consistent with th e predictions of IFD, might be posited to explain why satellite male horseshoe cr abs might choose to join large mating groups and engage in intense sperm competition rather than join a pair with no satellites and avoid sperm competition. One possibility is that pairs that attract satellites might nest in higher quality environments than pairs that do not attract satellites, ultimately increasing the proportion of eggs in a nest that survive. However, Hassler (1999) found no differences in the hatching or developmental rate s of eggs that developed in the nest sites of females with and without satellites, sugge sting no differences in nest site quality. A second possibility is that females that at tract satellites are of higher quality than females that do not attract satellites. This woul d mean that satellites that mate with higher quality females have the potential to achieve greater overall reproductive success despite an increase in sperm competition when compared with satellites that mate with lower quality females. Female quality may be measured as higher quality eggs or higher fecundity. Hassler (1999) found th at there were no significant differences in the hatching success, developmental rates, hatching rates, or energy content of eggs from females with and without satellites. This m eans that there are no observabl e differences in egg quality
9 between females that attract many satellites as compared with those that attract no satellites. Alternatively, Hassler (1999) also examined the quantity of eggs laid by females with and without satellites and found that females with satellites lay on average 65% more eggs than females without satellite s (but see Brockman n 1996). Hassler (1999) went on to show that in spite of the incr eased sperm competition, satellite males could increase the number of eggs they fertili zed by associating with these highly fecund females that were surrounded by satellite male s. Using this argument she accounted for males joining groups with one satellit e male but not larger groups. In this study I collected data on female fecundity, body size, and the distribution of mating group sizes on the beach. I combined th ese data with previous paternity data (Brockmann et al. 2000) to de termine if satellite male horseshoe crabs distribute themselves among females of variable fec undity following an IFD. Furthermore, I discuss the possible influence of the proportion of satellite males to females on the beach and operational sex ratio (OSR) on satellite male distribution around nesting females. This study provides insight into satellite ma le mating decisions and the formation of mating groups in the horseshoe crab. Methods Location This study was conducted on a 1-km stre tch of low energy beach on the south beach shore of Seahorse Key, (SHK), Levy C ounty, FL. Seahorse Key is an island on the west coast of Florida in the northern Gulf of Mexico approximately 4 km SW of Cedar Key, FL (29' N 83' W). It is part of the Lower Suwannee National Wildlife Refuge and is the location of the Univ ersity of Florida Marine Labo ratory. Each spring thousands
10 of horseshoe crabs nest from about 2 h befo re until 2 h after th e high tides during the weeks of new and full moons. I collected data on the relations hip between female fecundity and female body size on alternati ng weeks between 6 March and 2 May 2004. I examined the frequency distribution of mati ng group sizes during alte rnating weeks from 24 March to 23 April 2005. Additionally, I conducted a census documenting the operational sex ratio (OSR) and also the propor tion of nesting female s to satellite males on the beach (satellite ratio) during altern ating weeks from 5 March 2004 to 3 May 2004 and 23 February 2005 to 26 May 2005. Horseshoe Crab Measurements The inter-ocular distance (IO) of the lateral eyes was measured on nesting females and attached males. This is the distance betw een peaks of the inter-o cular spines located on the dorsal side of the la teral eyes. IO correlates we ll with body size (weight and carapace width) and can be measured without disturbing the nesti ng activity (Brockmann & Penn 1992). Inter-ocular distan ce was measured to the nearest 1 mm using a measuring tape. Mating group size of nesting females was r ecorded. Mating group size is defined as the total number of satellite males in contact with a female and her attached mate. For example, a nesting female with just an attached male has a mating group of zero. A nesting female with an attached male and three other males (sat ellites) surrounding her has a mating group size of three. A male is said to have joined a group when he comes in physical contact with the nesting female, att ached male, or another male in the group, and remains there for at least 10 sec.
11 Tagging Horseshoe Crabs Every animal in the study was tagged fo r documentation and to assure that no animal was used more than once. Tags were uniquely numbered plastic labels (1 cm x 2.5 cm) attached on the lower right side of th e prosoma with a flat-backed thumbtack. The tags remain attached to the horseshoe cr ab for a few weeks to a few months and horseshoe crabs do not seem to be adversel y affected by the tags (Cohen & Brockmann 1983; Brockmann 1990). Population Census 2004 and 2005 Each tide of the study (2004: n = 57; 2005: n = 89), from the time of the predicted maximum high tide until 0.5-1 h after (the time when the most horseshoe crabs are on the beach), I walked once along a 1 km section of the beach recording the mating group size of all nesting pairs as well as the number of unattached males and the number of lone females. Fecundity Measurements To determine the relationship between fema le fecundity (average clutch size) and female size, I conducted fiel d observations on individual pa irs of nesting females. I walked along the beach during the high tide until I found both a nesting female with no satellite males and a nesting female with at l east one satellite male located within 5 m of each other at approximately the same hei ght on the beach. I separately marked each clutch of eggs laid by both females by placing a pair of wire flags on either side of the prosoma-opisthosomal hinge of each female. When a clutch was laid as indicated by the female moving forward in the sand, I placed a se cond pair of flags at the new location. This continued until either a female finished laying eggs and returned to the ocean or until I marked a total of five clutches per female (Cohen and& Brockmann 1983;
12 Brockmann 1990). I recorded the date, time, an d number of satellites present around the nest each time the female laid a clutch of eggs. If a female finished laying eggs and returned to the ocean before I marked at leas t three clutches, I discarded that paired trial and it was not used for the analyses. Record ing a minimum of three clutch sizes per female enabled me to calculate an accurate av erage clutch size. Prior to data analysis, average clutch size was calculated by taking th e total volume of eggs laid per nest and dividing that by the total number of clutches . I individually tagged each nesting female and attached male and measured the femaleâ€™s size (IO). I returned to the flagged nests during the following low tide and carefully uncove red each clutch of eggs. I sieved each clutch with a plastic mesh sieve (1 mm mesh size), in seawater to remove excess sand and debris and measured and recorded the vol ume of eggs per clutch using a graduated cylinder. I placed the fertilized eggs back into their respec tive hole and covered the nest back with sand so that the e ggs would survive to hatching. Di sturbing eggs in this manner has no demonstrable effect on hatching succe ss (Hassler 1999). I measured 68 nests (34 nests of females with satellites and 34 wit hout satellites), over 18 high tide events (11 daytime tides and 7 nighttime tides), from 6 March to 2 May 2004. Female Size Measurements I used the south beach of SHK to measure the body size of nesting females on alternating weeks from 24 Marc h to 23 April 2005 to create a frequency distribution of female sizes on the beach. I began the female nesting measurements by walking along the beach at the high tide line during a high tide wh en females were nesting. I approached a nesting female (i.e. at leas t partially buried in the sand) , noted her mating group size, measured her inter-ocular distance (IO), and ta gged her. I repeated this procedure on as many nesting females as possible during a hi gh tide, across 14 high tides events (8
13 daytime tides and 6 nighttime tides), amounting to a total of 366 individual females (4 â€“ 69 females per high tide event). Calculating the Predicted Ideal Free Distribution of Satellites I calculated the expected f ecundity (average clutch si ze) distribution of nesting females on an average tide using the observed distribution of female body sizes (n = 366) across the breeding season and the function relating body size and fecundity (fecundity = -34.35 + 3.91 * female size). The function was determined by a Linear Regression test (SPSS 12.0 software). As the size distribution of females was not observed to be different among tides, I chose five sample tides (three from the 2004 census and two from the 2005 census), each with a different proportion of satellite males to females (satellite ratio), ranging from ratios of 1:2.5 to 2.5:1 (and an associated OSR which included all males â€“ attached, lone, and satellite). Using the expect ed fecundity distribution, I calculated the corresponding paternity distribution by multiplying the average percent fertilization success of the 1st, 2nd, 3rd, and 4th satellite male to join a mating group (taken from Brockmann et al. 2000) with the expected fecundity of each female for all five tides. This provided me with an expected fertili zation success for every satellite up to and including four around all nesting females on e ach sample tide. Finally, for each sample tide, I manually determined the distribution of satellites around th e females of varying sizes (and thus, varying fecund ity) that would be predicted if satellites distributed themselves following an IFD based on their ow n expected fertilization success. If the observed distribution of sate llites does not differ signifi cantly from the predicted distribution of satellites I can conclude that the satellite ma le horseshoe crabs are joining mating groups following the IFD for that OSR / sex ratio of satellites to females.
14 Data Analysis I examined the difference in the quantity of eggs laid by females with and without satellites using a Paired Samples T-test ( SPSS 12.0 software). I compared the predicted distribution of satellites around nesting females based on an IFD to the observed distribution of satellites for each sample tide chosen from the census (n = 5) using the Kolmogorov-Smirnov Two-Sample test (Siegel & Castellan 1988). Results Fecundity Measurements Fecundity was highly variable among fema le horseshoe crabs nesting on Seahorse Key (mean clutch size = 17.8 mL 10.9 S.D. ; Fig. 1), though there was a significant positive relationship between female body size (IO ) and female fecundity (average clutch size) (Linear Regression: r2 = 0.11, n = 68, p = 0.006; fecundity = -34.35 + 3.91 * female size; Fig. 2). Furthermore, females w ith satellites (mating group size of 1) laid significantly more eggs than females without satellites (mating group size of 0) (PairedSamples T-test: t = 2.10, df = 33, p = 0.043). Fe males with satellites had a mean clutch size of 20.32 12.54 mL, whereas females that di d not attract any satellites had mean clutch size of 15.20 8.31 mL. Female Size Measurements Body size (inter-ocular dist ance) was variable among ne sting females. Nesting females (n = 366) had a mean inter-ocular distance of 13.59 cm 1.1 S.D. (Fig. 3a). The frequency distribution of female body sizes re mained approximately the same across all tides (n = 14) (Fig. 3b-f).
15 0102030405060Fecundity (mL) 0 4 8 12 16Frequency Mean = 17.7591 Std. Dev. = 10.86707 N = 68 Figure 1. Frequency distribution of female f ecundity (mean clutch size in mL) (1 mL = 88 eggs) (n = 68). Fecundity = -34.35 + 3.91 * Female Size R-Square = 0.11 12.013.014.015.0Female Size (cm) 0.0 20.0 40.0 60.0F e c u n d i t y ( m L ) Figure 2. Scatter plot showing the relati onship between female size (inter-ocular distance) (cm) and fecund ity (average volume of eggs per clutch in mL). Linear Regression: r2 = 0.11, n = 68, p = 0.006.
16 11.012.013.014.015.016.017.0Female Size (cm) 0 10 20 30 40 50 60Frequency Mean = 13.585 Std. Dev. = 1.1039 N = 366 11.012.013.014.015.016.017.0Female Size (cm) 0 1 2 3 4 5 6Frequency Mean = 13.266 7 Std. Dev. = 1.05 7 N = 24 11.012.013.014.015.016.017.0Female Size (cm) 0 2 4 6 8 10 12Frequency Mean = 13.4823 Std. Dev. = 0.97552 N = 62 11.012.013.014.015.016.017.0Female Size (cm) 0 2 4 6 8 10 12Frequency Mean = 13.2485 Std. Dev. = 1.02031 N = 68 Figure 3. Frequency distribution of nesti ng female body sizes (int er-ocular distance) (cm) (n = 366) collected over 14 tides (a ). Examples of the distribution of female sizes for five of the 14 tides sampled (b-f). Frequency distributions were similar for each tide. (a) (d) OSR = 2.00 (c) OSR = 2.79 (b) OSR = 1.90
17 11.012.013.014.015.016.017.0Female Size (cm) 0 2 4 6 8 10Frequency Mean = 13.5574 Std. Dev. = 1.10958 N = 54 11.012.013.014.015.016.017.0Female Size (cm) 0 1 2 3 4 5Frequency Mean = 13.7731 Std. Dev. = 1.25938 N = 26 Figure 3. Continued Observed Distribution Versus Predicted IFD When the number of satellite males to nesting females (satellite ratio) was extremely female biased (1:2.5), satellite males distributed themselves among nesting females following the predicted IFD (Kolmogorov-Smirnov test: D316,316 = 0.095, p > 0.05; Fig. 4a). At a slightly less female biased satellite ratio (1:2), satellites did not follow the IFD, although the observed distribution was only slightly differe nt than predicted (Kolmogorov-Smirnov test: D151,151 = 0.159, p < 0.05). A few more females than predicted had one satellite and a few less female s than predicted had zero satellites or two satellites. All other gr oup sizes were similar to what was predicted by an IFD (Fig. 4b). As satellite males became more common on the beach, mating group sizes deviated significantly from the predicted IFD (Kolm ogorov-Smirnov test: satellite ratio = 1:1, D193,193 = 0.254, p < 0.001; satellite ratio = 2:1, D513,513 = 0.417, p < 0.001; satellite ratio = 2.5:1, D333,333 = 0.363, p < 0.001). More females than predicted failed to attract any satellites or attracted one sa tellite, fewer females than predicted had mating groups with two satellites, and more than predicted ha d groups larger than two (Fig. 4c-e). (e) OSR = 3.21 (f) OSR = 2.24
18 01234567Mating Group Size 0 20 40 60 80 100 120Frequency Observed Predicted 01234567Mating Group Size 0 50 100 150 200 250 300Frequency Observed Predicted 01234567Mating Group Size 0 20 40 60 80 100 120Frequency Observed Predicted 01234567Mating Group Size 0 100 200 300 400 500Frequency Observed Predicted Figure 4. The observed and predicted freque ncy distribution of mating group sizes of nesting females based on an IFD for five tides from the 2004 census. Each tide represents a different ratio of satel lite males to nesting females. When the satellite ratio was low (1:2.5) the sa tellites followed the IFD (KolmogorovSmirnov test: D316,316 = 0.095, p > 0.05) (a). For all other ratios, the satellites did not distribute themselves around females following the IFD (KolmogorovSmirnov test: D151,151 = 0.159, p < 0.05 (b); D193,193 = 0.254, p < 0.001 (c); D513,513 = 0.417, p < 0.001 (d); D333,333 = 0.363, p < 0.001 (e)). (a) (b) satellite ratio = 1:2 OSR = 2:1 satellite ratio = 1:2.5 OSR = 2.5:1 satellite ratio = 1:1 OSR = 3:1 (c) satellite ratio = 2:1 OSR = 3.5:1 (d)
19 01234567Mating Group Size 0 50 100 150 200 250Frequency Observed Predicted Figure 4. Continued Discussion If satellite males were distributing th emselves among spawning pairs according to an IFD, then the resultin g group sizes should provide the maximum reproductive fitness for all associated satellite males (Fre twell & Lucas 1970). When the predicted distribution is uneven, females (or mate locat ions) must be unequal in quality, and the largest groups should form around the highest quality females (or mate locations) (Parker & Sutherland 1986). In the horseshoe crab system, mating group sizes are uneven and do not follow a poisson distribution (which woul d indicate random choice), thus suggesting that there is not an equal fitness benef it for satellite males to join every group (Brockmann 1996). This study confirmed that fema le horseshoe crabs are very variable in fecundity, one potential measure of quali ty. Additionally, just as Hassler (1999) documented, I found that females that attracte d satellite males laid more eggs than females that did not attract sate llites suggesting that satellites are able to detect female fecundity or a correlate of fecundity and choose to join groups accordingly. But, in general, satellite male distri bution on a given tide did not follow the predicted IFD. (e) satellite ratio = 2.5:1 OSR = 4:1
20 On all high tides other than those with a very low OSR on the beach, more females than predicted attracted zero or only one satellite, fewer females than expected attracted two satellite males. This suggests that sate llite males were undermatching females (Gray & Kennedy 1994). In other words, satellite males were underestimating the fecundity of some females resulting in fewer satellite jo ining females with expected high fecundity and more satellites joining fe males with low fecundity. Although the under use of higher quality females seems surprising, it is a commonly observed divergence from IFD predictions (Tregenza 1995). Furthermore, at high OSRs, more females than predicted attract ed large group sizes of greater than two satellites indicating that overmatching also occurred. Thus, although satellite males were underestimating the fec undity of some females (with smaller group sizes), they were also overestimating the fec undity of others. These findings, when taken together, suggest that it is ei ther difficult for satellites to assess the potential reproductive fitness gained by joining a group or there ar e some situations where they may change their decision rules and base their joining decisions on factors othe r than fecundity and group size. Large deviations from the predicted IFD have been often reported (e.g., Gray 1994; Blanckenhorn et al. 2000; Shochat et al. 2002; Jackson et al. 2004; Ramp & Coulson 2004) and there are many conditions that pred ict its occurr ence (Kennedy & Gray 1993; Gray & Kennedy 1994; Tregenza 1995). Deviations are likely to occur when individuals cannot assess all possible resources, due to time constraints or locati on constraints. In the horseshoe crab system, satellite males have limited time to search the beach for females. Females nest during a 2-3 h time span and ar e widely distributed along the beach at the
21 high tide line, frequently in windy conditions (Cohen & Brockmann 1983; Brockmann 1990; Penn & Brockmann 1994). If satellites are extremely choosy and forgo opportunities to join spawning females with mo derate fitness benef its in search of females with high fitness benefits they may compromise their spawning opportunity. Additionally, satellites in shallow water ar e frequently flipped over onto their backs due to wind induced turbulence (Brockmann 1990). Satellites are able to right themselves using their telson although this maneuve r takes time and energy (Brockmann & Penn 1992; Penn & Brockmann 1995). Also, satellite s are in general older and in worse condition overall than attached males and have a harder time righting themselves (Brockmann & Penn 1992; Penn & Brockmann 1 995). When satellites are part of a group they are able to hold on to the female, attach ed male, or another satellite male making it easier for them to resist the force of the waves and wind (Penn & Brockmann 1995). Thus, it may be more beneficial for a satellite to join a group of any female, regardless of fecundity or group size, than to take the time and energy to search the beach for a female with higher fecundity or a smaller mating group. Furthermore, undermatching of females is predicted to occur when there is interference among competitors with differe ntial competitive abil ities (Fretwell 1972; Gray 1994; Moody & Houston 1995). Although sate llites in groups struggle to maintain their position around a female, it is not known wh ether they actively prevent others from joining their spawning group. If this were th e case then the str onger satellites could thwart the weaker satellites' attempts to jo in a group. However, th is possibility seems unlikely for several reasons. The average fert ilization success of the first and second satellites surrounding a female is equal, with every additional sate llite decreasing the
22 fertilization success of all other satellites (Brockmann et al. 2000). Therefore, there would be no reason for the first satellite male to try to prevent a second male from joining because the second male has no measurable effect on the success of the first, thus failing to explain the lack of groups with two sate llites. Additionally, every satellite after two should try to prevent additional satellites fr om joining their group since all subsequent satellites to join do decrease the fertilization success of all preceding satellites. If males that are better competitors were able to accomplish this, large group sizes should not occur or occur less frequently than predicted by an IFD. However, more females than predicted had large mating groups of three to seven. Another possibility for the discrepancy between the observed group sizes and the group sizes predicted by an IF D for the population of horseshoe crabs on Seahorse Key is that there is great variati on in fertilization success among all satellites and satellite fertilization success is strongly influenced by the position where a male is located around a female when spawning (Brockmann et al. 2000; Fig. 5). Position one is over the border between the prosoma and opisthosoma and may be on either the right or left sides of the female (1F), attached male (1A), or satellite male (1S). Position three is over the terminal spines and tail of the attached male (3A) or sa tellite male (3S). Position four refers to all other positions around the female (4F), attached male (4A), or satellites (4S) (Brockmann et al. 2000). Position 1F is strongly pref erred over all other positions and the average percentage of fertilizations of satellite ma les occupying that position (whether there were one or two satellites in the mating group) is approximately 40%. But, if there were two males in position 1F, one of the two males co mmonly achieves much mo re fertilizations than the other (Brockmann et al. 2000). For the purposes of this study, I did not take
23 position into consideration, consequently giving the first and second satellite male to join a group about 40% of the ferti lizations. Thus, I made the a ssumption that the first and second satellite do equally well at fertiliz ing the eggs, potentially inflating the reproductive fitness gain for one of the two satellites in groups of two, thereby over estimating the predicted frequency of groups of two and under estimating the predicted frequency of groups of one or zero. Figure 5. This nesting female (F) has an at tached male (A) plus six satellite males occupying different spawning positions. Th ree satellite males are located in position one, one on the right side of the female (1F), one on the left side of the female (1F), and one on the left si de of a satellite (1S) on the border between the prosoma and opisthosoma. Two satellites are in position three, over the tail and terminal spines of sate llites (3S). The last satellite is in position four, located on the anterior side of the femaleâ€™s prosoma (4F). Now that it is known that the distribution of satellite males among nesting females deviates from what is predicted by an IFD, there are several other aspects of the spawning behavior that need to be addressed before one is able to fully understand the mating decisions of satellite males and the formati on of mating groups in the horseshoe crab. In the future, one should directly assess sate llite male competitive ability to examine whether satellites with greater competitive ability join spawning groups with higher quality females. Additionally, it is possible th at satellites are better at accurately assessing 1F F A 4F 1S 1F 3S 3S
24 female quality when females are clumped (minimizing time and location constraints) resulting in an observed distribut ion that follows closer to the predicted IFD. Therefore, it would be interesting to look at the differe nce between group sizes on tides when females are clumped on the beach (possibly due to certain weather conditi ons, water currents, etc.) versus spread out along the entire beach or by simply examining the distribution of satellites across a small section of the beach versus the entire beach. Finally, a dynamic model needs to be made to test whether a sa tellite should stay with a female or leave a group and search for a better female, taking into consideration the costs and benefits of staying or searching. The state and envir onmental variables that could influence a satelliteâ€™s decision include his condition, the quality of the nesting female, the number of females on the beach from which a satellite w ould be able to choose, the distribution and density of groups on the beach, and the weather conditions (for example, it is more costly to leave a group and search for another group on windy days because of the likelihood of being flipped over), among others.
25 CHAPTER 3 HOW DO SATELLITE MALES DISTINGUISH AMONG FEMALES OF VARIABLE FECUNDITY? Introduction Thousands of male and female horseshoe crab pairs approach the beaches along eastern United States and the Yucatan peni nsula every spring and summer to mate. A mating pair consists of a fecund female and a male that has attached himself to the posterior spines of that female in order to participate in fertilizations. Some pairs nest alone, whereas other pairs perhaps only a meter or two away, are surrounded with as many as ten unattached, satellite males (Bro ckmann 1996; Fig. 6). Furthermore, mating group sizes are uneven and non -random (Brockmann 1996; Schw ab Ch. 2). More pairs than expected fail to attract satellites or at tract large groups of sate llites, and fewer pairs than expected attract intermediate numbers of satellites (Brockmann 1996; Schwab Ch. 2). Therefore, some nesting pairs seem to be preferred over others, possibly due to these pairs being of higher quali ty. Additionally, satellites are somehow locating and subsequently choosing to join these pairs for group spawning. Females nesting with satellites are, on aver age, larger and heavier than pairs nesting alone (Brockmann 1996; Hassler 1999; Schwab Ch. 2). Additionally, Hassler (1999) and Schwab (Ch. 2) found that satellite males are choosing to join groups with females that lay more eggs. If these are the cues that sa tellite males use to discriminate among pairs, how do unattached males swimming in the oc ean along the shoreline, frequently on
26 moonless nights, locate pairs and assess the size and fecundity of the females nesting in the sand on the beach? Figure 6. Four female horseshoe crabs (burie d in the sand) nesti ng within 5 m of each other attract very different numbers of satellites. From bottom to top, female #1 has five satellite males in addition to her attached male, female #2 only has an attached male, female #3 has two sa tellites plus her attached male, and female #4 only has an attached male. Vision is known to be an important modality used by lone male horseshoe crabs for identifying mates offshore (Barlow et al. 1982; Powers et al. 1991; Herzog et al. 1996). The visual system of Limulus seems to have exploited every possible adaptation for increasing retinal sensitivity (Barlow et al . 1989). More specifi cally, through endogenous circadian rhythms, horseshoe crabs anticipate the onset of darkness, and anatomical and physiological changes take place in the ey e (including an increase in lateral eye sensitivity of one million times daytime levels paralleling the million-fold decrease in ambient light intensity after dark) which allow them to see as well, if not better, at night than in the day (Barlow et al. 1982; Atherton et al. 2000). Barlow et al. (1982) conducted the first study of vision as a mechanism for mate identification in horseshoe cr abs. They found that male hor seshoe crabs approached and attempted to mate with black castings of female horseshoe crabs under water more #1 #2 #3 #4
27 frequently than female castings painted gray or white and cement cubes and hemispheres regardless of color. Additionally, males blinded by eye coverings neither approached the castings nor attempted to mate with them. Optic nerve fiber firing rates in horseshoe crab eyes dramatically increase when crab-like objec ts move within visual range at about the speed of a horseshoe crab (15cm/s) demonstr ating the eyeâ€™s sensitivity to objects having the same size, contrast, and moti on as potential mates (Passaglia et al. 1997; Barlow et al. 2001). Herzog et al. (1996) determined that ma les are highly sensitive to contrast and will orient toward Limulus -like objects at distances up to 1.2 m in clear water. Furthermore, studies have shown that shallow water and wind-driven overhead waves, common features of horseshoe crab mating habitats , create strobic beams of light that when reflected off low-contrast moving objects, such as newly molted females, enhance neural signals which improves their response to obj ects underwater (Pa ssaglia et al. 1997). Unlike pairing with unattached females unde rwater, joining already spawning pairs on the shore has some different features. When females nest they are partially buried under the sand (Brockmann 1996; Fig. 7). This would seem to make it difficult for unattached males to use visual assessment to gauge female size from offshore. Additionally, females nest above the tide line, whereas lone males frequently approach females from underwater, further raising th e difficulty for searching males to see potential mates due to the light refraction at the air-water interface (but see Schuster et al. 2004). Therefore, it seems unlikely that body size is the only cue used by satellite males to locate and assess nesting females. Many arthropods (Krieger & Breer 1999) use chemical cues for identifying and locating potential mates (e.g., Me lville et al. 2003; Fukaya et al. 2004a; De Cock &
28 Matthysen 2005). Furthermore, it has been shown that a few arthropods and some reptiles and amphibians even possess the ability to us e chemical cues for detecting mate quality (including such things as size, fecundity, or dominance status), and subsequent evaluation and discrimination among potential mates (e.g., Marco et al. 1998; Bushmann & Atema 2000; Shine et al. 2003). Is it possible that male horseshoe crabs also have the ability to locate females and evaluate female size and/or fecundity using chemical signals? Figure 7. This female horseshoe crab is ne sting in the sand partially buried up to her lateral eyes. In addition to the attached male, she has attracted three satellite males, all participating in group spawning. Limulus possess millions of chemoreceptive ce lls distributed among chemoreceptor sites located in the spines of the coxal gnat hobases of each walking leg, the spines of the chilarial appendages, and the chelae of a ll the limbs (Patten 1894; Hayes 1985). Studies have shown that these chemoreceptors may be used for both close-range and distance chemoreception (Wyse 1971; Quinn et al. 1998). For example, horseshoe crabs have the ability to generate both incoming and outgoing currents by beating their book gills causing water flowing towards and away from their bodies (Quinn et al. 1998). Quinn et al. (1998) found that book gill activity can cr eate an incoming current pulling water from up to 18 cm away directly over the chemosensory organs. Thus, male horseshoe crabs
29 may possess the ability to use chemical cues for locating and assessing females for mating. Though horseshoe crabs have advanced chem osensory abilities, very few studies have actually examined the use of chemical cues in mating decisions. Hassler and Brockmann (2001) discovered that a sponge filled with seawater taken from below a nesting female placed under a cement model of a female was more attractive to males than a sponge filled with seawater. Presumab ly, the sponge collected whatever chemical compound was being released from under th e nesting female allowing unattached (satellite) males to use the chemical cue, in addition to the visual cue of the model, to locate and discriminate between â€œfemalesâ€ for mating. Although the specific chemi cal compound or compounds to which male horseshoe crabs are attracted is unknown, Hassler and Brockmann (2001) hypothesized that it is a proteinaceous compound of high molecular we ight. Horseshoe crab chemoreceptors are sensitive to amino acid soluti ons and steroids extracted from their prey (i.e. clams) (Barber 1956; Hayes & Barber 1982). Furthermore, Ferrari a nd Targett (2003) discovered a proteinaceous compound of high molecular we ight (>10kDa) in the fertilized eggs and hemolymph of female horseshoe crabs that caused chemotaxis in mud snails and eels, two organisms that are highly a ttracted to dead horseshoe crab s. Finally, horseshoe crab eggs are known to produce a substance that attracts sperm and there is strong evidence that this chemical also induces the acroso mal reaction allowing sperm attachment to the egg (Shoger & Bishop 1967). Taken together, th ese studies indicate that horseshoe crabs release and respond to chemical compounds in seawater, and it is likely that these chemicals are present in horseshoe crab eggs.
30 In this study I explore the use of visual and chemical cues as possible proximate mechanisms by which satellite males make mating decisions. I expect, based on the Limulus sensory system and the characteristics of the marine environment, that satellite male horseshoe crabs may use visual cues for longer-range sensory detection and chemical cues for shorter-range detection, leadin g to a combination of cues used to locate and evaluate females for mating. Prior to conducting a manipulative experiment testing whether satellites prefer to mate with large over small females, I collected observational data to confirm the positive relationship between female body size, expos ed carapace height, and mating group size. If unattached males are unable to view the entire body of a nesting female, perhaps they are able to see the section of the carapace that is not buried under the sand and interpolate body size. Secondly, I used female cement models to test directly whether satellite males choose to join groups surrounding large vers us small females. Then, I documented whether satellite males approach nesting females pr ior to or after the first clutch is laid in order to help support or refute the possibility that male horseshoe crabs are attracted to compounds in fertilized eggs. Finally, I expe rimentally examined the use of chemical cues located in a clutch of horseshoe crab e ggs by looking at satellite male preferences to female models placed over a bag of fertilized horseshoe crab eggs versus female models placed over empty bags, and female models pl aced over a sponge fille d with fertilized egg extract (supernatant A) versus female models placed over a sponge filled with seawater. Taken together, these studies will enable me to determine the proximate mechanisms by which satellite male horses hoe crabs locate and discriminate among nesting females.
31 Methods Location Two field sites were used in this study: the south beach of Seahorse Key (SHK), Levy County, FL and Bowers Beach, Kent County, DE. Seahorse Key is an island off the west coast of Florida in the northern Gulf of Mexico approximately 4 km SW of Cedar Key, FL (29' N 83' W). Bowers Beach (39' N 75' W) is a low energy, course sand and gravel beach located on the Delaware Bay approximately 56 km north of Lewes, DE, the location of the College of Marine Studies, University of Delaware. Horseshoe crabs approach and nest along the b eaches of the Delaware Bay at the peak of the high tides, until about 3 h after, from the late spring until the early summer. At SHK crabs nest from 2 h before to 2 h after th e maximum high tide during the early spring high tides (Rudloe 1980; Cohen & Broc kmann 1983; Barlow et al. 1986). Horseshoe Crab Measurements A suite of measurements were taken on fe male and male horseshoe crabs for this study. The inter-ocular distance (I O) of the lateral eyes was measured on nesting females and attached males. This is the distance betw een peaks of the inter-ocular spines located on the dorsal side of the la teral eyes. IO correlates we ll with body size (weight and carapace width) and can be measured wit hout disturbing the nes ting (Brockmann & Penn 1992). The exposed carapace height is the distance on a nesting female from the point where the edge of the prosoma-opisthosom a border meets the sand, up along the border to the hinge, and down the prosoma-opisthosom a border to the point where the carapace meets the sand on the other side (Fig. 8). Females who are entirely buried under the sand while nesting have an exposed height of zero. Inter-ocular distance and exposed height were measured to the nearest 1 mm using a measuring tape.
32 Figure 8. The exposed carapace height is the distance on a nesting female from the point where the edge of the prosoma-opist hosoma border meets the sand, up along the border to the hinge, and down the prosoma-opisthosoma border to the point where the carapace meets the sand on the other side. The condition of the lateral eyes and gills of satellite males were visually examined and recorded. A satelliteâ€™s eye condition was considered â€œgoodâ€ if both eyes were smooth with the entire eye visible and considered â€œpoorâ€ when one or both eyes were rough to the touch and partially blocked by fouling organisms (such as barnacles, bryozoans, hydroids, tunicates, slipper limpets, or other invertebrates or algae) or when both eyes were rough and one or both eyes were complete ly covered by fouling organisms, soft to the touch, or missing entirely. Gill condition wa s considered â€œgoodâ€ when the gills were completely intact with no worms on the surface or the underside and considered â€œpoorâ€ when the gills were completely intact with a minor amount of worms attached to either the surface or the underside or partially shredded or missing a nd covered by parasites and a significant amount of worms. Mating group size is defined as the total number of males in contact with a female, her attached male, or any other males that ar e in contact with the pair. For example, a nesting female with just an attached male has a mating group of zero. A nesting female with an attached male and three other ma les (satellites) surrou nding her has a mating group size of three. A male is said to have joined a group when he comes in physical
33 contact with the nesting female, attached male , or another male in the group, and remains there for at least ten seconds. Tagging Horseshoe Crabs Every animal in the study was tagged fo r documentation and to assure that no animal was used more than once. Tags were uniquely numbered plastic labels (1 cm x 2.5 cm) attached on the lower right side of th e prosoma with a flat-backed thumbtack. The tags can be inserted without disturbing the nesting animals and tags remain attached to the horseshoe crabs for a few weeks to a few months and horseshoe crabs do not seem to be adversely affected by the tags (C ohen & Brockmann 1983; Brockmann 1990). Data Sets To examine some possible proximate mech anisms that satellite male horseshoe crabs use to locate and discriminate among fe males I collected observational data on the size of nesting females and their mating group size (Female Nesting Data) and the relative time during nesting when unatt ached males approach females (Egg Attractiveness Data). In addi tion, I carried out three expe rimental manipulations: (1) Visual Cues Experiment, (2) Chemical Cues Experiment I: Attractive Eggs, and (3) Chemical Cues Experiment II: Supernatant A. Female Nesting Data I used the south beach of SHK to examine the relationships among female body size, exposed carapace height, and mati ng group size on alterna ting weeks from 12 March to 23 April 2005. I began the female nesting measurements by walking along the beach at the high tide line during a high tide wh en females were nesting. I approached a nesting female (i.e. at leas t partially buried in the sand) , noted her mating group size, measured her IO, and tagged her. I repeated this procedure on as many nesting females as
34 possible during a high tide, across 14 high tide events (8 daytime tides and 6 nighttime tides), during which time a total of 367 indivi dual females were identified and measured (4 â€“ 69 females per high tide event). I used Pearsonâ€™s Correlation test (SSPS 12.0 software) to analyze the hypotheses that the larger the female the greater the exposed height, the greater the exposed height the larger the group size , and the larger the female the more satellites in her group. I used square root transformed data for exposed carapace height and mating group size prior to analyses to meet the assumptions of the statistical tests. Additionally, I used an Independent Samples T-test to examine whether females with large mating group sizes have a greater inter-ocular distance and exposed carapace height than females with small group sizes. If these relationships we re significant this would support the hypothesis that satellite males use female size or a correlate of female size to locate and/or discri minate among females for mating. Egg Attractiveness Data I collected observational data on nesting pairs and groups to identify when during the egg laying process satellite males approach females. This would help me to identify if satellite males use chemical cues that are re leased from the eggs or oviducal fluid to locate and/or evaluate females for mating. I walked along the beach during the high tide until I came upon a female and her attached male crawling out of the ocean onto the beach. After a minute or two, the female dug into the sand and began to lay eggs. I marked the location of each clutch of eggs w ith wire flags until the female finished laying and returned to the ocean or until I marked a total of five clutches (Cohen & Brockmann 1983; Brockmann 1990; Fig. 9). If a female finished laying and returned to the ocean before I marked at least three clutches I did not use the data in analyses. By only marking nests of females that I observed approach the beach and begin ne sting, I was able to
35 identify the laying order for each clutch of eggs within a nest (i.e. the first clutch, second clutch, etc.). I recorded the date, the time at which she approached the beach (approach time), the time that the female began digging into the sand (nest time), and the time that each clutch was laid. Additionally, I recorded the time that any satellite joined and left the mating pair/group. Finally, I tagged the nes ting female and attached male. I returned to the flagged nests during the following low tide and retrieved each clutch of eggs. I measured 111 nests, over 27 high tides (14 da ytime tides and 13 nighttime tides), on alternating weeks from 25 Febr uary to 25 May 2005 at Seah orse Key, FL. I used a ChiSquare test (SPSS 12.0 software) to examine when during the egg-laying process (before or after the first clutch of eggs was laid) satellite males approach and join a nesting female. If satellite males approach nesting fe males significantly more often after the first clutch of eggs were laid than before the first clutch was laid, then this is consistent with the hypothesis that nesting female s release a chemical cue locat ed in the eggs or oviducal fluid that attracts satellite males. If ther e were no difference between the number of satellite males that approach and join a nesti ng female before and after the first clutch is laid then it is probable that satellite males do not soley rely on the use of chemical cues located in the eggs or oviducal fluid to lo cate and/or discriminate among females. This does not negate the possibility that fema les release chemical cues from another unidentified source th at attract satellite males for mating. Experimental Manipulations I conducted three manipulative experiment s to help determine the proximate mechanism(s) by which satellite males locate and/or evaluate nesting females: (1) the Visual Cues Experiment, (2) the Chemical Cu es Experiment I: Attractive Eggs, and (3) the Chemical Cues Experiment II: Supernatant A. In the Visual Cues Experiment I
36 Figure 9. Procedure for collecting Egg Attrac tiveness data. A mating pair that emerged from the ocean and started to nest in the sand (a). The two flags were positioned on either side of her carapace aligned with the border between her prosoma and opisthosoma at the moment that she laid the first clutch. With every subsequent clutch she laid, two mo re flags were placed into the sand marking each clutch. After laying five cl utches the pair left the beach and went back into the ocean (b). I returned to the nest during the low tide to dig up each clutch of eggs marked by the pairs of flags (c). compared the response of satell ites to large and small models of females in a series of field trials to test the hypothesis that satelli te male horseshoe crabs are using visual cues, specifically female size, to locate and/or assess females for mating. The Chemical Cues Experiment I: Attractive Eggs was used to examine the hypothesis that satellite males are using chemical cues located in a clutch of e ggs to locate and/or assess females for mating. In this experiment I compared the responses of satellite males to female models presented with horseshoe crab eggs versus female m odels without eggs in an experiment on the beach. If satellite males were more attracted to female horseshoe crab models placed over a bag of sieved horseshoe crab eggs than female horseshoe crab models over an empty bag then satellites must be using a chemical cue(s) located in the eggs to locate and/or discriminate among females for mating. But even if this is the case, the exact compound(s) in the eggs that are attractive to satellite ma les is still unknown. I conducted the Chemical Cues Experiment II: Supernatan t A to test the hypot hesis that satellite males are attracted to chemical compounds loca ted in horseshoe crab egg extract, thus attempting to narrow the possibilities for the primary source of the chemical cues. (b) (c) (a)
37 Models used during exp erimental manipulations I used two kinds of cement models for th e experiments: carapaces of dead crabs filled with cement and 100% cement models made from horseshoe crab molds. To make the models for the Supernatant A Experiment I collected four similar-sized carapaces from dead females and three similar-sized ca rapaces from dead males off the shore of Bowers Beach, DE. I thoroughly cleaned the ca rapaces with water and removed all the legs and the gills from the ventral side leavi ng just the hard dorsal shell surface. Once the hollowed-out carapaces dried completely, I fill ed the concave ventral cavity with equal parts of cement and mortar mix and allowed th e cement to dry for at least 48 h. Once dry, the cement models sat immersed in tubs of wa ter for 24 h to help facilitate leaching of chemicals prior to the start of the experi ment. Finally, I labeled each model with an individual identification letter so I was able to record and rotate the models used in each trial. In order to construct the models used in the Visual Cues and Attractive Eggs Experiments, I collected and cleaned one la rge female horseshoe crab carapace (IO = 14 cm), one small female horseshoe crab car apace (IO = 11.5 cm) and one average sized male carapace (IO = 9 cm) from the south b each on Seahorse Key, FL. I laid the carapace dorsal-side up on a wooden board and plugged a ll of the edges of the carapace that did not touch the board surface with non-sulfur clay to create a seal. I constructed a clay lip approximately 8 cm high around the edge of th e board. I painted a ve ry thin impression coat of liquid silicone rubber mixed with a sl ow-drying catalyst (10:1) (Silicones, Inc.) (1st coat) onto the entire exposed outer surface (including the sections plugged with clay) of the carapace. This technique meant that the surface of the cast would create an accurate, detailed impression of the dorsal surface of the carapace. Twenty-four hours
38 later I mixed a second batch of liquid silicone rubber with a quick-drying rapid catalyst (10:1) and poured (2nd coat) and painted (3rd and 4th coat) it over the firs t coat of silicone rubber to result in a strong rubber cast appr oximately 2.5 cm thick. The clay lip that I formed around the edge of the wooden board he lped to prevent the final coats of liquid silicone rubber from flowing beyond the board before they dried. After the rubber cured I made a holder for the rubber cast by immersing a section of air conditioning filter material slightly larger than the rubber cas t into liquid plaster-of -paris until it was completely saturated. I neatly placed the thick, wet, filter rectangle over the top surface of the rubber cast. This created a hard, supportive holder for the cast once the filter dried. I rested the rubber cast in the holder and filled it with wet cement mix making sure that the edges were an even height so the model would lie flat on the sand, but allowing for a slightly concave center. I al lowed 24 to 48 h for the cement to dry, lifted the rubber cast filled with the cement model out of the moth er-mold, and slowly p eeled the rubber away from the model, leaving an impression of the dorsal surface of a horseshoe crab in the cement. I painted the entire surface of th e cement model with aquarium-safe, black, epoxy paint. By painting the model with an aquarium-safe paint, I created a nonhazardous barrier between the cement and the field environment preventing potentially harmful chemicals from leaching out of th e cement model while it was immersed in seawater. Additionally, Barlow et al. (1982) showed that satellite males can visually detect and have a greater attraction towards female horseshoe crab models painted black than those painted gray or white due to the high contrast between the black model and the light, sandy surroundings. This procedure was repeated five times each with the large female carapace, the small female carapace, and the male carapace making a total of 15
39 cement horseshoe crab models. Just as with the other models, I labeled, rotated, and recorded the models used in each trial. Visual cues experiment Cement model set-up. I began an experimental trial by walking along the shoreline when females were nesting and sa tellite males were pr esent until I came upon an empty area of sand where no pairs were pr esently nesting but with at least a few satellite males visible in th e water within 25 m. I placed a large female model down on the sand at the edge of the high tide line perp endicular to the shorel ine with the prosoma facing up beach in a natural ne sting position. The prosoma of the female model was out of the water above the tide lin e, but the female's opisthosoma and the male were partially submerged. Directly behind the female model I placed a male model in a natural attached position. I placed a small female model with a male down on the sand in the same position 1 m away, measured from the center of the prosoma of the large female model to the center of the prosoma of the small fe male model (Fig. 9a). Although nesting pairs frequently nest with their carapaces partially buried in th e sand, I did not recreate that effect because of the impossi bility of maintaining a unifo rm layer of sand over both models due to waves unpredictably washi ng away any sand cove ring the models. I alternated the left and right positions of the large and small models with every experimental trial to account for any possible side biases (such as might arise from tidal currents). Additionally, I varied which large and small female models and male models I used with each trial to account for any possible differences among the models. Experimental arena. To create an experimental arena around the models, I measured 1.2 meters from either side of the la rge and small female models parallel to the shoreline and 1.2 meters perpe ndicular to the shoreline towards the ocean thus making an
40 arena 3.4 x 1.2 m around the models, which I mark ed off with wire flags (Fig. 9a). I positioned identical large plastic tubs on the s hore 2 m directly landward of the large and small female models to be used as a holding tank for any satellite males that joined the models during the trial. This design allowe d me to distinguish among the males that joined each of the models. If there were any ne sting pairs, satellites, or lone males located within the arena I gently removed them from the arena and placed them into the ocean away from the experimental site immediatel y prior to beginning a trial. I observed the arena by standing on the shore directly between the two plastic tubs. Prior to the start of the trial I noted any field conditions, such as wind, that could affect the outcome. Procedure for visual cues trials. Each trial lasted 10 mi n. I recorded the number of satellite males that entered the arena, the side on which they entered and all actions of each male from the moment when he entere d the arena until he exited. I recorded the number of males that came within 10 cm of each female model (with the front of its prosoma pointed in the direction of the model) and the number that joined each model pair (Fig. 10c). A male was recorded as Â“farÂ” when he entered the arena, Â“nearÂ” when he approached a female model within 10 cm, and Â“joinÂ” when he touches a female model or her attached male model with the front of his prosoma. I also recorded the total length of time that a male was on each side of the arena and the total length of time a male remained joined with a model. Satellite males approached and entered the arena either by walking along the beach or by swimming along the shoreline into the arena (Fig. 10b). I allowed the males to walk freely about the aren a and cross from one side to the other, but I removed the male, placing it in the designated tub, and ceased recording that male if he joined, unjoined, crossed over to the other side of the arena or if he exited the arena.
41 As soon as the 10-min trial ended, I tagged and recorded the eye condition of each male that had joined a model in order to examine possible di fferences among the satellite males that joined the large and small female model pairs. I released the satellite male back into the ocean after it was tagged and reco rded. I conducted 55 trials (1 Â– 7 trials per high tide), over 12 high tide events (6 daytime tides and 6 nighttime tides), on alternating weeks from 26 March to 12 Apr il 2005, at Seahorse Key, FL. Figure 10. Two sets of model pairs (large fema le model plus an attached male and small female model plus an attached male) we re placed in a natu ral nesting position 1 m apart on the beach. An experiment al arena of 3.4 x 1.2 m surrounding the models was marked off with wire flag s (a). Satellite males approached the models by walking along the beach in to the arena or by swimming along the shore and entering the arena from the ocean (b). A male joined a model pair when he comes in physical contact with the female model, attached male model (c), or another satellite male that has joined the group, and remained there for at least 10 sec. Chemical cues experiment I: Attractive eggs Egg collection. I began the Attractive Eggs Experiment by walking along the beach until I came upon a nesting female, preferably with at least one satell ite male. I carefully and slowly lifted the anterior portion of th e nesting femaleÂ’s prosoma up out of the sand while trying not to disturb any satellites. Most of the time, the nesting female abandoned her nest after being lif ted out of the sand and returned to the ocean, but was frequently seen nesting later in the tide in another section of the beach. If the female did not immediately abandon her nest, I held the front of the nesting female up out of the way (a) (c) (b)
42 and dug with my hand into the sand approxi mately 10 cm directly below where the female had been nesting to feel for any eggs th at the female might have laid. If I felt eggs I scooped into the sand with a la rge plastic petri dish to re move the clutch of eggs. I sieved the eggs to remove all excess sand and debris. If there were any eggs that seemed to be part of a clutch laid by another female from earlier that tide or a previous tide I carefully removed those eggs from the sample eggs and reburied them. I usually collected eggs from females with satellites because fema les with satellites lay more eggs than those without (Schwab Ch. 2). Unfortunately, on days with few satellites or early in a tide it was difficult to find many females with sa tellites necessitating th e use of eggs from females without satellites. I continued egg co llection until I collected a total of 50 mL of recently laid eggs (under 2 h old) from at leas t two different nests. I filled a plastic mesh bag (10 cm x 20 cm) (sewn from plastic window screening) with th e collected eggs and sealed the bag with a rubber band. I continue d this procedure until I collected multiple bags filled with 50 mL of eggs. All egg bags that were not used immediately in a field trial were stored in a large petri dish on th e beach in the shade. Bags containing eggs were only used to hold eggs and the empty bags that were used as controls were never filled with eggs for any trial. Those not used for a field trial during that tide were put back in a hole dug in the sand in the area of other horseshoe crab nests. Hassler (1999) showed that eggs can be removed from a ne st, replaced into another nest in the beach, and survive to at least th e trilobite larval stage. I chose to use fertilized eggs rather than unfertilized eggs to be used for this study for several reasons. First, it is very difficult to obtain unfertilized e ggs without sacrificing females. Electro-ovulation has been used (Brockmann et al. 2000), but only small
43 numbers of eggs (< 50) are ever released (p ers. comm. H. J. Broc kmann). It is possible that not all eggs are fertilized immediately after laying and unfertili zed eggs retain their ability to be fertilized for at least 40 min (Brockmann et al. 2000). Additionally, horseshoe crabs exhibit polyspermy (Shoger & Brown 1970; Brown & Humphreys 1971). Up to one million sperm may go through ac rosomal reactions with one egg and the factors that determine which spermÂ’s DNA enter the egg nucleus remain unknown (Brown & Knouse 1973). Furthermore, unattach ed males move back and forth over the depressions in the sand left by recently de parted pairs (Cohen & Brockmann 1983). Thus, it seemed very reasonable th at satellite males would be attracted to fertilized eggs. Cement model set-up. With bags of eggs in hand, I began an experimental trial by walking along the beach until I came upon an ar ea of sand where no pairs were presently nesting but with at least a fe w satellite males identified within 25 m. I set up the models and experimental arena just as in the Visual Cues Expe riment except that two large female models were used. Im mediately prior to beginning th e experiment, I lifted up the posterior portion of one female model and pl aced one bag filled with eggs. I placed an identical empty bag underneat h the posterior portion of the other female model. Experimental trials. Each trial lasted 10 min and I r ecorded the same data as in the Visual Cues Experiment. The satellite male s were treated in the same way as in the Visual Cues Experiment alt hough gill condition was recorded in place of eye condition. I conducted 46 trials (1 Â– 10 tria ls per high tide), over 9 high ti de events (8 daytime tides and 1 nighttime tide), on alte rnating weeks from 9 April to 11 May 2005, at Seahorse Key, FL.
44 Chemical cues experiment II: Supernatant A Egg collection. Eggs were collected on Bowers Beach, DE during the week of 16 May to 22 May 2004, one hour after the hi gh tide by walking along the beach until I came across a slight indent in the sand indicati ng a female had nested th ere. At that spot I dug into the sand with my hands and scooped al l freshly laid eggs directly into 50mL plastic tubes with lids. After many tubes were filled with eggs from multiple nests, I immediately transported them back to the lab where they were stored for 2 weeks in the tubes at -80 C. Horseshoe crab eggs can be stored at -80 C for an indefinite length of time and should resume all original properties when thawed (pers. comm. K. Wakefield). Egg extraction. I performed the egg extraction on 30 and 31 May 2004 in a laboratory at the College of Ma rine Studies, University of Delaware, Lewes, DE. I began the egg extraction by retrieving 10 tubes of eggs out of the -80 C freezer. I picked out as many rocks, pebbles, and debris from the e gg clusters as possible using tweezers and my fingers. Following the protocol described by Ferrari and Targett (2003), I weighed out 250g of eggs and ground the egg aliquot using a mortar and pestle until frothy. I poured the crushed eggs into a flask filled with 500 mL 0.05 M Tris HCl buffer (pH 7.5) and set it to stir overnight, in the dark at 4 C. The following day I filter ed the stirred egg extract through two cheesecloths into centrifuge tubes and centrifuged the extract for 10 min at 10,000 rpm. This separated the extract into su pernatant (crude extract), pellet (sand particles and cellular debris), and lipid. I carefully poured off all of the supernatant into 50mL plastic tubes and stored them at 4 C until directly before the start of the Supernatant A Experiment (12 Â– 15 days later) . Storage of Supernatant A egg extract for
45 two weeks at 4 C had no noticeable effect on its attr activeness to the mud snail (pers. comm. K. Wakefield). Cement model set-up. I conducted the Supernatant A Experiment to examine if satellite males are more attracted to the horse shoe crab egg extract than to a seawater control. I walked along the beach when females were nesting and lone males were present until I came upon a section of sand (approximately 2 m) that was not occupied by any nesting female. I set up the cement models and the arena just as in the previous experiments. I filled one clean, 760-mL pl astic Tupperware container with 50 mL horseshoe crab egg extract and 450 mL fresh seawater. I filled a second 760 mL plastic container with 500 mL fresh seawater to serv e as a control. I submerged for 30 seconds one piece of sponge (10 x 6 x 4 cm) into the pl astic container with the egg extract dilution and another piece into the container holding seawater. Immediately following submersion I carefully lifted the egg extr act sponge and seawater spong e out of the solutions and simultaneously placed them under the female prosoma of the treatment and control models, respectively. Experimental trials. I recorded the number of males that approached each model, the time within the trial that one a pproached, and the direc tion (left, right, or behind) from which the male approached fo r 10 min. An approach was counted when a male swam within 5 cm of the model pair wi th the anterior porti on of the satelliteÂ’s prosoma facing the model and remained there for 10 sec. After this 10 sec the male was removed from the arena and placed inside the appropriate holding container to be tagged and measured once the trial finished. I also noted the number of touches to each model by satellite males during the tria l and any unusual weather cond itions that could influence
46 which model the satellites approached. The sate llites that approached a model pair were treated the same as they were in the previous experiments. When the trial was finished I tagged and released all the males that approa ched either of the two models. I conducted 18 trials (3 Â– 6 trials per tide), over 4 tides (all nighttime tides), from 15 June to 18 June 2004, at Bowers Beach, DE. For each trial I alternated the side of the arena on which the treatment sponge and control sponge were placed and which models I used for the trial. New sponges were used for each trial within a night, but sponges were re-used on subsequent nights. Â“Egg extractÂ” sponges were never used as Â“seawaterÂ” sponges and vice versa. Data analysis I compared the total number of satellite ma les that entered the arena, came near, and joined each model during each 10 mi n trial using a Wilcoxon Signed-Ranks test. If satellite males entered the large female modelÂ’s side of the arena, came near, or joined the large female model pair significantly mo re and/or for a signifi cantly longer length of time than the small female model pair than I can conclude that satellite males are using visual cues to discriminate among females fo r mating. If satellite ma les entered the side of the arena with the female model placed over the bag of eggs, came near, or joined that model pair significantly more or for a longer le ngth of time than the female model placed over an empty bag of eggs then satellite male s are using chemical cues in horseshoe crab eggs to locate and/or evaluate nesting female s. If satellite males approached the female model pair over a sponge filled with egg extract significantly more than the female model pair over a sponge filled with s eawater than I can conclude th at satellite males are using chemical cues located in the supernatant of egg extract to locate and/or discriminate among females for mating. Additionally, I used a Chi-Square test to examine if the
47 lateral eye and book gill condition of the satellite males that approached the large female model versus the small female model and th e female model placed over a bag of sieved eggs versus the female model placed over an empty bag, respectively, affected satellite male choices when joining a model pair. All statistical analyses were performed using SPSS 12.0 software. Results Female Nesting Measurements Larger females had more exposed carapace than smaller females, even when both were partially buried (Pearson correlation: r = 0.327, n = 366, p < 0.001; Fig. 11). Females with larger exposed carapace height were surrounded by more satellite males, i.e. had a larger mating group size (Pear son correlation: r = 0.104, n = 366, p = 0.047). Interestingly, female size and mating gr oup size were not significantly correlated (Pearson correlation: r = 0.077, n = 366, p = 0.142; Fig. 12). A more de tailed analysis of the data, however, indicated that females with many satellites ( 3 satellites) were significantly larger than females with few or no satellites, with mean inter-ocular distances of (mean Â± S.E.) 14.0 Â± 0.17 and 13.5 Â± 0.06, respectively (Independent Samples T-test: t = 2.56, df = 55.4, p = 0.013; Fig. 13). Furthermore, females with satellites had larger exposed carapace he ights than females without satellites (Independent Samples T-test : t = 2.0, df = 352.9, p = 0.046). Egg Attractiveness Data The first satellite male to join a nesting pair was equally likely to approach and join a pair prior to the female laying the first clut ch (n = 28) as after the female has laid the first clutch (n = 33) (Chi-square test: 2 = 0.410, df = 1, n = 61, p = 0.522).
48 Female Size (Inter-ocular Distance) (cm)18 17 16 15 14 13 12 11 10Exposed Height (cm)40 30 20 10 0 Figure 11. Scatter plot showing the corre lation between female size (inter-ocular distance) (cm) and the exposed carapace height (cm). Pearson correlation: r = 0.327, n = 366, p < 0.001. Female Size (Inter-ocular Distance) (cm)18 17 16 15 14 13 12 11 10Mating Group Size10 8 6 4 2 0 Figure 12. Scatter plot show ing the relationship between female size (inter-ocular distance) and female mating group size (the number of males excluding the attached male that is participating in the mating). Pearson correlation: r = 0.077, n = 366, p = 0.142.
49 44322 N =Mating Group SizeFew Satellites Many SatellitesMean Female Size (cm)14.2 14.0 13.8 13.6 13.4 Figure 13. Mean size (inter-ocu lar distance) (cm) of females with mating groups of three or more satellite males (many) and less than three satellite males (few). Standard error bars are given. Indepe ndent Samples T-test: t = 2.56, df = 55.4, p = 0.013. Visual Cues Experiment Satellite males were significantly more likely to enter the arena (Wilcoxon Signed Ranks: z54 = 2.44, p = 0.015), approach within 10 cm (z54 = 2.14, p = 0.032), and join (z54 = 2.43, p = 0.015) the large female model (IO = 14.0 cm) pairs than the small female models (IO = 11.5 cm). Among the satellites that joined a model pair, there was no difference between the average (Wilcoxon Signed Ranks: z54 = 1.27, p = 0.204) or the total time (z54 = 1.86, p = 0.063) that satellites remain ed attached to the large female models and the small female models. Over th e 55 experimental trials, 67% (N = 187) of the satellite males that appr oached either the large or small female model had good eye condition and 33% had poor eye condition. There was no difference in the joining decisions made by satellite males with good and poor eye condition (Chi-square test: 2 = 0.145, df = 1, n = 187, p > 0.200). Chemical Cues Experiment I: Attractive Eggs There was no difference between the number of satellite males that entered the arena (Wilcoxon Signed Ranks: z45 = -1.56, p = 0.119) or joined (z45 = -1.41, p = 0.159)
50 the female model with eggs versus the fema le model without eggs (control). However, significantly more satellites approached within 10 cm of the female model with eggs than the control (Wilcoxon Signed Ranks: z45 = -1.97, p = 0.049; Fig. 14). Of the satellites that joined either model, those that joined the fe male model with eggs remained attached for significantly longer than thos e that joined the control model (Wilcoxon Signed Ranks: average time attached, z45 = -2.82, p = 0.005; tota l time attached, z45 = -2.31, p = 0.021; Fig. 15). Of the satellite males that joined either the model with eggs or the control model, 85% (n = 196) had good gill condition and 15% had poor gill condition. Gill condition had no effect on the joining decisions made by satellite males (Chi-square test: 2 = 0.269, df = 1, n = 196, p > 0.200). Chemical Cues Experiment II: Supernatant A Satellite males were equally likely to join the female models over sponges containing supernatant A extr act and female models over sponges containing seawater (Wilcoxon Signed Ranks: z18 = -0.875, p = 0.381). However, unlike the previous experiments where identical models were used, these models were not identical. I performed a post-hoc analysis on female mode l inter-ocular distance to examine if the female carapaces used for these trials were truly about equal in size or if my size approximations between the supernatant A female model carapaces and the seawater female model carapaces were in actuality not equal. I ndeed, the carapaces randomly assigned to be placed over sponges containing supernatant A were significantly smaller than the carapaces placed over sponges cont aining seawater (Wilcoxon Signed Ranks: z18 = -2.77, p = 0.006).
51 46 46 4646 46 46 N =Location Around ModelFar Near JoinAverage No. Satellties Around Models7 6 5 4 3 2 1Model TreatmentNo Egg Model Egg Model Figure 14. Mean number of satellite males to enter the arena, a pproach within 10 cm, and join the egg and no egg model pair s (n = 46, 10-minute trials). A male was recorded as Â“farÂ” when he entered the arena, Â“nearÂ” when he approached a female model within 10 cm, and Â“joinÂ” when he touches a female model or her attached male model with the front of his prosoma. Standard error bars are given. Wilcoxon Signed Ranks: z45 = -1.56, p = 0.119 (far); z45 = -1.97, p = 0.049 (near); z45 = -1.41, p = 0.159 (join). 4646 N =Female Model TreatmentNo Egg EggAverage Time Satellites Joined (sec)140 120 100 80 60 40 20 Figure 15. Mean time (sec) sate llite males remained joined to the egg and no egg female models (n = 46, 10-minute trials). St andard error bars are given. Wilcoxon Signed Ranks: z45 = -2.82, p = 0.005. Discussion This study suggests that satelli te males are using both visual and chemical cues in joining decisions. When the use of visual cu es, specifically female size, was examined, satellite males approached and attempted to attach or join groups with large female *
52 horseshoe crab models placed in a mating position on the sand more frequently than small female models. As satellites approach ed the models from a minimum distance of 1.2 m either by crawling out of the water onto the sand or on the beach along the tide line towards the model, it is clear that satellite males, underwater or on dry land, can visually detect and evaluate female size from this distance. However, eye condition did not have an effect on the model that the satellites chose. Therefore, it seems likely that the measures that I used to assess eye condition ma y not reflect the ability of horseshoe crabs to see mate-like objects. Although satellites can easily view the entire size and sh ape of the female models in the Visual Cues Experiment allowing them to make the appropriate decision, they are not able to view the entire female when she is actually nesting. I confirmed the positive relationship among female body size, exposed carapace height, and mating group size, showing that it is probable th at satellites visually detect a nesting femaleÂ’s exposed carapace height, as opposed to her entire body, to make their decisions about which females to join. From these results it seems quite evident that satellite males are using visual cues to locate and evaluate female size, choosing to mate with larger females (with higher fecundity) over smaller females (Hassler 1999; Schwab Ch. 2). If this were the sole cue used by satellites in mating d ecisions however, one would also expect satellite males to remain with the larger female models longer than the smaller models. But, this is not the case. When I looked at a satelliteÂ’s decision to remain with a female model, there was no difference in the average or total length of tim e a satellite male remained joined to the
53 large or small female model. Therefore, satellite males may be using other sensory modalities in addition to visual cu es when making mating decisions. Hassler and Brockmann (2001) found that satellites are using chemical cues in mating decisions, but they did not identify th e source of the cues. I proposed that eggs may be a possible source of the cue because fertilized horseshoe crab eggs have been found to be attractive to mud snails and eels, indicating th at eggs indeed release a chemoattractant, possibly also attractive to other horseshoe crabs (Ferarri & Targett 2003). Additionally, if the chemical cue were pro portional to the quantity of eggs laid this would allow males to directly assess an indicator of female fecundity. Although more satellites approached and atte mpted to mate with the female models over eggs and the female models with sponges filled with egg extract (supernatant A) than the female models over empty bags a nd sponges filled with seawater, respectively, the differences were not stat istically significant. However, in the Chemical Cues Experiment I: Attractive Eggs , more satellite males came near (within 10 cm) the female models over eggs than the control models. A dditionally, of those that did join the group attempting to mate with either model, satelli tes remained joined to the female model over eggs much longer than the control. These re sults suggest that horse shoe crab eggs do exude chemical cues that satellite males us e in mating decisions, but satellites may not have the ability to detect these cues from a di stance due to a lack of sensitivity in their chemoreceptors or because of the nature of the their nesting environment. Consequently, once satellites have joined a group, presumably through visual detection of female size, they may be using chemical cues located in the eggs to decide if it is worth remaining with a female or leaving that group a nd searching for a higher quality female.
54 Furthermore, when actual nesting was observe d, satellites were equally likely to join a group whether or not a female had laid a cl utch of eggs, additi onally supporting this hypothesis. The findings from this study, in combination, suggest th at satellite male horseshoe crabs use a hierarchy of cues, firs t visual (i.e. body size)Â—f or longer distances, then chemical (i.e. compounds released from the eggs)Â—for shorter range communication, to make mating decisions. The use of multiple mating cues for mating decisions has been found in several vertebrate (e.g., Crapon de Caprona & Ry an 2000; Shine & Mason 2001; Hankinson & Morris 2003; Partan et al. 2005; Shine 2005) and invertebrate species (e.g., Rovner & Barth 1981; Bushmann & Atema 2000; Acquistapace et al. 2 002; Fukaya et al. 2004a; Fukaya et al. 2004b; Raethke et al. 2004), frequently in aquatic environments. For example, male crayfish, Austropotamobius pallipes , did not respond when presented with female odor alone, although when female odor was presented along with the visual stimulus of a dead female crayfish, male s spent more time in stereotypical mating behaviors such as locomotion and the intermediate posture (Acquistapace et al. 2002). Also, male turtle-headed sea snakes, Emydocephalus annulatus , that live in shallowwater habitats, use visual cues of size to lo cate potential mating partners from a distance, and use pheromones in addition to vision once they have moved within physical contact of a female (Shine 2005). The use of multiple mating cues and/or sensory modalities can be explained by the behavioral ecology and living environment of the study species. It is predicted to be beneficial (1) when mate quality is highly variable, thus allowing for a better overall assessment of quality; (2) when an individual has limited time for mate assessment or the
55 costs of mating may be high because perceiving multiple cues allows for a more accurate assessment of mate quality; and (3) when the mating environment is variable allowing individuals to pay attention to different cues from different distances and under different conditions (review in Candolin 2003). In the horseshoe crab system female quality is highly variable (Brockmann 1996; Schwab Ch. 2). There is great variation in the quantities of eggs laid by females during one tide and across the mating season (C ohen & Brockmann 1983; Brockmann 1996; Hassler 1999; Schwab Ch. 2). Additionally, fe male horseshoe crabs in FL nest for only approximately a 2-3 h span during the high tides, on the weeks of a new or full moon, from March to May, and only under certain fa vorable environmental conditions (i.e. low to moderate wind speed, wind direction towa rds land to assist the rising and brief maintenance of the high tide; Rudloe 1980; Cohen & Brockmann 1983; Barlow et. al 1986; Brockmann unpublished data). Therefore, satellite males are presented with a limited amount of time for mate searching and assessment favoring as many mating cues as possible in order to aid in efficient mati ng decisions. As females are partially buried (so body size is partially concealed) and seawat er flows periodically over the nesting pair with every wave, opportunities to locate and assess the exposed height of the female are limited as are opportunities to detect chemi cal compounds that the female is releasing. Although water assists in the transport of chemical compounds, females are not always entirely submerged underwater and the dispersa l potential of released chemicals may be restricted. As a consequence, the ability fo r males to detect chemical cues may be compromised. Additionally, turbul ent water and tidal currents mean that seawater quickly disperses along the shore, also hindering the ability of males to locate the source of a
56 chemical cue (pers. comm. H.J. Brockmann) . The aforementioned aspects of horseshoe crab nesting behavior and the environment w ould clearly favor multiple cues to enhance mating opportunities. Although it is likely that satellite males use a hierarchy of cues, first visual to locate and assess females and their size (as an indicato r of fecundity), then chemical to directly assess fecundity once the males have approache d, this does not negate the possibility that chemical cues are used alongside of visual cues earlier on in the mate choice process. It is possible that satellites are using chemical cu es from another source other than eggs in longer-range detection (or that I failed to present a sufficient quantity of eggs to elicit a long-range response). Hassler and Brockmann (2001) found that satellite male horseshoe crabs were more attracted to and remained longer with female models placed over a sponge containing seawater coll ected from underneath a nesting female than models over a sponge containing seawater alone. In th eir experiment, Hassler and Brockmann eliminated the potential of visual, auditory, and tactile cues eliciting a difference between models just as I did. However, they were not able to definitively de termine the source of the cues because the seawater collected from underneath a nesting female may contain chemical compounds from the eggs, oviducal fluid, hemolymph, or even chemical compounds released from the a ttached male or other males in the group (but this is unlikely). Consequently, differen ces in the attractiveness of th e models may be a result of a wide variety and concentr ation of compounds that were collected in the sponge containing water from undern eath a nesting female. On the other hand, Hassler and Brockmann (2001) constructed their models by filling similar-sized dead carapaces with cem ent. It is possible that the approaching
57 satellites could detect the slight size differences between the models. If this were true and the models chosen to be paired with the sponge containing seawater from underneath a female were on average slightly larger than the models paired w ith sponges containing just seawater, just as in the Chemical Cues II: Supernatant A experiment, then it is plausible that the satellites we re not making a decision solely based on chemical cues, but using the visual cue of size as well. Thus, it is possible that satellites may be using chemical cues either from the eggs or a nother source (such as oviducal fluid that is released with the eggs) in addition to visu al cues during initial stages of female assessment (as a source of long-range communi cation), although this is unable to be confidently determined though the previous experiments. It is known that satellite male horseshoe crabs preferentially join spawning groups with larger, more fecund females than smalle r females that lay fewer eggs (Brockmann 1996; Hassler 1999; Schwab Ch. 2), but the m echanisms with which males detect these differences remained elusive until now. This study confirmed that satellites initially use visual cues to locate and evaluate females ba sed on body size from a distance of at least 1.2 m, but rely additionally on chemical cues released by horseshoe crab eggs when making the decision to remain with that female or search for a female with higher fecundity. Thus, satellite male s appear to use a hierarchy of cues, first visual, then chemical, to make mating decisions. Future stud ies should be designed to directly test the hierarchical multiple cue hypothesis. Additi onally, more studies need to be conducted examining the mating cues used by other spec ies that move between habitats for mating, such as from water to land, and the mechanisms that individuals use to detect these cues,
58 to determine whether this system is unique in its use of cues or if this is fairly common in these kinds of systems.
59 CHAPTER 4 GENERAL DISCUSSION The ideal free distribution (IFD) model pred icts that individuals should distribute themselves around resources such as food (e .g., Ward et al. 2000; Jackson et al. 2004; Ramp & Coulson 2004; Veerana goudar et al. 2004) or mates (e.g., Nishida 1993; Huhta et al. 1998; Widemo 1998; Blanckenhorn et al. 2000) in a manner that maximizes their overall fitness (Fretw ell & Lucas 1970; Goss-Custard & Sutherland 1997). If resources are unequal in quality and indi viduals are following an IFD, larger groups of individuals should form around high quality resources a nd smaller groups should form around low quality resources resulting in equal fitness across all individuals (Parker & Sutherland 1986). An unusual feature of the horseshoe cr ab mating system is the large groups of satellites that form around some nesting pa irs and not around others nearby on the beach (Brockmann 1996). Additionally, satellites are more likely to form groups around pairs with females that are larger (Brockmann 1996) and lay more eggs (Hassler 1999) than around pairs with smaller, less fecund females. Although this would sugge st that satellites assess fecundity and distribute themselves around nesting pairs following the predicted IFD, in general, this is not the case. When the ratio of satellite males to nesting females on the beach (satellite ratio) was low (1:2.5), satellites di stributed themselves among females following the predicted IFD. However, as the number of males on the beach increased, group sizes differed significantly from that predicted by the IF D. Satellites seemed to under estimate the fecundity of some females and over estimate the fecundity of others resulting in more
60 pairs than predicted with zero or one satellite , fewer pairs with two satellites, and more pairs with three to seven satellites. Conse quently, as the satell ite ratio on the beach increased, the resulting distribution appeared more even than when there were few animals on the beach. Although satellite male horseshoe crabs failed to distribut e themselves among nesting pairs in a manner predicted by the IF D based on fecundity, they did seem able to assess fecundity or a correlate of fecundity (Brockmann 1996; Hassler 1999; Schwab Ch. 2). To better understand mating group formati on and the cues used by males to identify groups, I examined the mechanisms by wh ich satellites make mating decisions. Satellite male horseshoe crabs appear to be locating and discriminating among nesting pairs using both visual cues of body size and chemical cues located in horseshoe crab eggs. Satellites were more attracted to large than to small models of nesting females although there was no difference in the length of time that satellites remained with each model. Additionally, there was no difference in the number of satellites that joined female models over bags containing horseshoe crab eggs when compared with female models over empty bags. But, of those satellites that joined, satellites remained joined to the models over bags filled with eggs signifi cantly longer than to the models over empty bags. Thus, visual cues seem to be most impor tant in the initial l ong-distance location and attraction of a satellite to a pa ir, and chemical cues aid in a satelliteÂ’s decision to remain with a pair once he has approached. While it does not appear that all satellite male horseshoe crabs are always making the most advantageous mating decisions when it comes to joining spawning groups, they have evolved a sensory system that allows them to assess multiple mating cues to make
61 the best decisions possible. The use of mu ltiple cues to assess fecundity provides a satellite with a more accurate assessment of female fecundity because each signal alone provides information about female fecundity with some error (review in Candolin 2003). When multiple signals are assessed the error is reduced. Because satellite males have the ability to use both visual and chemical cues to assess fecundity we would expect that they s hould be capable of distributing themselves following an IFD to maximize their fertilization success. However, there are many constraints and trade-offs that satellit e males face when engaging in locating and assessing potential mates. Each constraint alone may only cause noise in the distribution, but when acting in concert they could contri bute to a deviation from an IFD. First, although body size is positively correlated with fecundity, there is high variability among female fecundity at a given body size (Cohen & Brockmann 1983; Brockmann 1996; Hassler 1999; Schwab Ch. 2). The use of visual and chemical cues when assessing fecundity helps to form an accurate assessm ent, but satellites may still make suboptimal decisions. Additionally, nesting occurs in an unpredictable environment partially on land and partially underwater. Multiple cues help to facilitate female assessment and increase the information gained while in the water a nd on land. However, frequently females are partially buried restricting visi bility, and the water at the tide line is often churning limiting the ability to detect directionality and potentially causing chemical cues to dissipate quickly. Furthermore, satellites ha ve limited time to make these assessments (Rudloe 1980; Cohen & Brockmann 1983; Barl ow et al. 1986). The use of back-up signals of fecundity may reduce the time it take s for satellites to inspect females, but it does not eliminate the potential need to decide whether to stay with a low quality female
62 and fertilize a small proportion of eggs or for go mating with that female to search for a higher quality female that may or may not be found in hopes of fertilizing a greater quantity of eggs. This trade-off between re maining with a female and searching for a better female is at the core of mate assessment and it necessitates having perfect knowledge of all the groups on the beach and unlimited time for spawning. Consequently, unless these ideal conditions are met, it is highly unlikely that satellite males will be able to follow an IFD in most situations. Taken together, despite the benefit of using multiple sensory cues in mate assessment, these individual and environmental constraints make it harder to assess fecundity and riskier to take the time to search for a higher quality female, causing difficulty in making the optimal mating decision, corroborating our lack of support for gr oup sizes following the predicted IFD. Besides environmental and individual constr aints, there are a few other issues that may have contributed to the observed deviation from the IFD. This study assumes that the first and second satellite male to join a spawning pair achieve approximate the same percentage of fertilizations, leading to the predicted distribution of many pairs with two satellites and very few pairs with one. My model predicted that ma ting group sizes of one should only occur when there is not a second satellite available to join that group. In reality, however, the first and second sate llite may not achieve approximately equal fertilization success. If this is the case, I would expect to see more groups with only one satellite and fewer groups with two leading to a closer ma tch with what was observed. The first male to join a spawning pair is al most assured to fertil ize a large proportion of the eggs. Although the average fertilizati on success is documented to be 40.5 Â± 4.3 % when there are two males around a female, ther e is marked variability, one of the two
63 satellites frequently achieves significantly greater paternity (56.4 Â± 5.2 %) than the other satellite (11.9 Â± 2.0 %) (Brockma nn et al. 2000). Thus, there is less uncertainty in success for a male that is the only satellite spawning w ith a pair. This may cause more satellites to prefer to be the first satellite on a pair rather than having an equal preference for the first and second satellite position as our model a ssumes, thereby increas ing the number of pairs with group sizes of one a nd decreasing the number of pa irs with group sizes of two, just as we observed. There are several other aspects of spawni ng behavior that wo uld provide greater insight into the rules and mechanisms used by satellite male horseshoe crabs to make mating decisions. This study found that as th e number of satellite males on the beach changed, the distribution of gr oup sizes changed. As the sate llite ratio increased, there were fewer groups with zero, one, or two sate llites and more groups with three or more satellites leading to a distribution that appe ars more even than what was observed at a lower OSR. This could be due to the fact th at there are so many more males on the beach than females that almost all of the females, regardless of quality, attract multiple satellite males. When the first sate llite males approach the beach they should distribute themselves following an IFD as best as possibl e just as they did when the sex ratio was very low. As the tide progresses and more satellites approach the beach there will be fewer females left without any satellites and more females than expected with large group sizes just as we observed. Additionally, when there are la rge numbers of males on the beach in relation to females, the competition for females becomes even greater. Therefore, the cost of leaving a female in s earch of a higher quality female is greater. This potentially causes males to relax thei r mating standards and jo in spawning groups of
64 lower quality females, leading to fewer females with small group sizes. We need to specifically test these predictions by direc tly controlling the number of females and satellite males that approach the beach and observing the resulting di stributions. Also, it is unknown if the distribution of group sizes around females differs with changes in the distribution of females on the b each (i.e. when females are cl umped or evenly spread out across far distances). It is possible that satel lite males change their decision rules and are less selective about jo ining spawning groups when females are more spread out on the beach because it is harder to assess multiple fema les. If this is the case, the predicted IFD should change with changes in female di stribution and this sh ould be taken into consideration when modeling the predicted IF D. In particular, the more spread out females are across the beach, the more ra ndomly distributed the satellites may be. Furthermore, it is relatively unknown what e nvironmental and state variables (besides female fecundity and group size) a satellite assesses when deci ding whether to stay with a female or search for a higher quality female . Potential variables besides fecundity and group size could include the sa telliteÂ’s condition, the number of females on the beach from which a satellite would be able to c hoose, the distribution and density of groups on the beach, and the weather conditions. A dyna mic model needs to be made including these variables to determine which are th e most important in spawning decisions. Following that, the modelÂ’s predictions need to be tested in the field. Once we know which other variables satellites consider wh en deciding to join a group, we need to identify how satellites are assessing these va riables. Satellites may be using visual and chemical cues, just as in assessing fecundit y, or they may be using other cues such as auditory or tactile. Finally, my work shows that satellite males use a hierarchy of cues,
65 first visual then chemical, to assess fecundity from different distances and reduce error. But, future studies should directly test this multiple cues hypothesis and specifically examine the benefits of using multiple cues rather than one cue in mate assessment.
66 LIST OF REFERENCES Acquistapace P., Aquiloni L., Hazlett B.A., & Gherardi, F. 2002. Multimodal communication in crayfish: sex reco gnition during mate search by male Austropotamobius pallipes. Canad. J. Zool. Rev. Canad. Zool ., 80 , 2041-2045. Atherton, J.L, Krutky, M.A., Hitt, J.M., Dodge, F.A., & Barlow, R.B. 2000. Optic nerve responses of Limulus in its natural habitat at night. Biol. Bull ., 199 , 176-178. Barber, S.B. 1956. Chemoreception and proprioception in Limulus . J. Exp. Zool. , 131 ,5174. Barlow, R.B., Chamberlain, S.C., & Lehm an, H.K. 1989. Circadian rhythms in the invertebrate retina. In Facets of Vision (Ed. by D.C. Stavenga and R.H. Hardie), pp. 257-280. New York: Springer-Verlag. Barlow, R.B., Hitt, J.M., & Dodge, F.A. 2001. Limulus vision in the marine environment. Biol. Bull ., 200 , 169-176. Barlow, R.B., Ireland, L.C., & Kass, L. 1982. Vision has a roll in Limulus mating behavior. Nature , 296 , 65-66. Barlow, R.B., Powers, M.K., Howard , H., & Kass, L. 1986. Migration of Limulus formating: relation to lunar pha se, tide height, and sunlight. Biol. Bull. , 171 , 310329. Blanckenhorn, W.U., Cornelia, M. & Re uter, M. 2000. Are dung flies ideal-free distributed at their ovi position and mating site? Behav. , 137 , 233-248. Brockmann, H.J. 1990. Mating beha vior of horseshoe crabs, Limulus polyphemus . Behaviour , 114 , 206-220. Brockmann, H.J. 1996. Satellite ma le groups in horseshoe crabs, Limulus polyphemus . Ethology , 102 , 1-21. Brockmann, H.J., Colson, T., & Potts, W. 1994. Sperm competition in horseshoe crabs ( Limulus polyphemus ). Behav. Ecol. Sociobiol. , 35 , 153-160. Brockmann, H.J., Nguyen, C., & Potts, W. 2000. Paternity in horseshoe crabs when spawning in multiple-male groups. Anim. Behav. , 60 , 837-849.
67 Brockmann, H.J. & Penn, D. 1992. Male ma ting tactics in the horseshoe crab, Limulus polyphemus . Anim. Behav. , 44 , 653-665. Brown, G.G. & Humphreys, W.J. 1971. Sperm-egg interactions of Limulus polyphemus with scanning electron microscopy. J. Cell Biol. , 51 , 904-907. Brown, G.G. & Knouse, J.R. 1973. Effects of sperm concentration, sperm aging, and other variables on fertiliza tion in the horseshoe crab, Limulus polyphemus . Biol. Bull. , 144 , 462-470. Bushmann, P.J. & Atema, J. 2000. Chemically mediated mate location and evaluation in the lobster, Homarus americanus . J. Chem. Ecol. , 26 , 883-899. Byrne, P.G. & Roberts, J.D. 2004. Intrase xual selection and group spawning in quacking frogs (Crinia Georgiana). Behav. Ecol. , 15 , 872-882. Calvo, D. & Molina, J.M. 2005. Fecundity-bod y size relationship and other reproductive aspects of Streblote panda (Lepidoptera: Lasiocampidae). Ann. Entomol. Soc. Amer. , 98 , 191-196. Candolin, U. 2003. The use of multiple cues in mate choice. Biol. Rev. , 78 , 575-595. Cohen, J.A. & Brockmann, H.J. 1983. Breedi ng activity and mate selection in the horseshoe crab, Limulus polyphemus . Bull. Mar. Sci. , 33 , 274-281. Crapon de Caprona, M.D.C. & Ryan, M.J. 1990. Conspecific mate recognition in swordtails, Xiphohorus nigrensis and X. pygmaeus (Poecilidae) Â– olfactory and visual cues. Anim. Behav. 39 , 290-296. De Cock, R. & Matthysen, E. 2005. Sexual communication by pheromones in a firefly, Phosphaenus hemipterus (Coleoptera: Lampyridae). Anim. Behav. , 70 , 807-818. Dewsbury, D. 1982. Ejaculate cost and male choice. Am. Nat. , 119 , 601-610. Ferrari, K.M. & Targett, N.M. 2003. Chem ical attractants in horseshoe crab, Limulus polyphemus , eggs: The potential for an artificial bait. J. Chem. Ecol. , 29 , 477-496. Fretwell, S.D. 1972. Populations in a Seasonal Environment . Princeton, New Jersey: Princeton University Press. Fretwell, S.D. & Lucas, H.J. Jr. 1970. On territorial behavior and other factors influencing habitat distribution in birds. Acta Biotheor. , 19 , 16-36. Fukaya, M., Akino, T., Yasuda, T., Yasui, H., & Wakamura, S. 2004a. Visual and olfactory cues for mate orientation be haviour in male white-spotted longicorn beetle, Anoplophora malasiaca . Ent. Exper. Appl. , 111 , 111-115.
68 Fukaya, M., Arakaki, N., Yasui, H., & Wakamura, S. 2004b. Effect of colour on male orientation to female pheromone in the black chafer Holotrichia loochooana loochooana . Chemoecol. , 14 , 225-228. Goss-Custard, J.D. & Sutherland, W.J. 1997. Behavioral Ecology: An Evolutionary Approach (4th ed.) (Ed. By Krebs, J.R. and Davies, N.B.), pp. 373-395. Oxford: Blackwell Sci. Gray, R.D. 1993. Sparrow, matching and the id eal free distribution Â– can biological and physiological approaches be synthesized. Anim. Behav. , 2 , 411-423. Gray, R.D. & Kennedy, M. 1994. Perceptual co nstraints on optimal foraging Â– a reason for departure from the ideal free distribution. Anim. Behav. , 47 , 467-471. Hankinson, S.J. & Morris, M.R. 2002. Sexual se lection and species recognition in the pygmy swordtail, X iphophorus pygmaeus : conflicting preferences. Behav. Ecol. Sociobiol. , 51 , 140-145. Hassler, C. 1999. MS Thesis. Satellite male groups in horseshoe crabs (Limulus polyphemus ) : How and why are males choosing females? Department of Zoology, Univ. of Florida, Ga inesville, Florida. Hassler, C. & Brockmann, H.J. 2001. Eviden ce for use of chemical cues by male horseshoe crabs when locating nesting females ( Limulus polyphemus ). J. Chem. Ecol. , 27 , 2319-2335. Hayes, W.F. 1985. Chemorecepto r sensillum structure in Limulus . J. Morphol. , 119 , 121142. Hayes, W.F. & Barber, S.B. 1982. Peripheral synapses in Limulus chemoreceptors. Comp. Biochem. Physiol. , 72A , 287-293. Herzog, E., Powers, D.M.K., & Barlow, R.B. 1996. Limulus vision in the ocean day and night: effects of image size and contrast. Vis. Neurosci ., 13 , 31-41. Huhta, E., Jokimaki, J., & Rahko, P. 1998. Di stribution and reproductive success of the Pied Flycatcher Ficedula hypoleuca in relation to forest patch size and vegetation characteristics; the effect of scale. Ibis , 140 , 214-222. Jackson, A.L., Ranta, E., Lundberg, P., Ka itala, V., & Ruxton, G.D. 2004. Consumerresource matching in a food chain when bot h predators and prey are free to move. Oikos , 106 , 445-450. Kennedy, M. & Gray, R.D. 1993. Can ecological theory predict the distribution of foraging animals Â– a critical analysis of experiments of the ideal free distribution. Oikos , 68 , 158-166.
69 Korona, R. 1990. Travel costs and the IF D of ovipositing female flour beetles, Tribolium confusum . Anim. Behav. , 40 , 186-187. Krieger, J. & Breer, H. 1999. Olfact ory reception in invertebrates. Science , 286 , 720-723. Marco, A., Chivers, D.P., Kiesecker, J. M., & Blaustein, A.R. 1998. Mate choice by chemical cues in western redback ( Plethodon vehiculum ) and DunnÂ’s ( P. dunni ) salamanders. Ethology , 104 , 781-788. Melo, Y.C. & Le Clus, F. 2005. Growth and reproduction of the pelagic goby Sufflogobius bibarbatus off the Orange Rive r, Southern Africa. African J. Marine Sci ., 27 , 265-273. Melville, J.M., Tallarovic, S. K., & Brownell, P.H. 2003. Evidence of mate trailing in the giant hairy desert scorpion Hadrurus arizonensis (Scorpionida luridae). J. Insect Behav. , 16 , 97-115. Moody, A.L. & Houston, A.I. 1995. Interf erence and the ideal free distribution. Anim. Behav. , 49 , 1065-1072. Nishida, T. 1993. Spatial relationships be tween mate acquisition probability and aggregation size in a gr egarious coreid bug ( Colpula lativentris ): A case of the ideal free distribution unde r perceptual constraints. Res. Popul. Ecol. , 35 , 45-56. Nakatsuru, K. & Kramer, D. 1982. Is sperm cheap? Limited male fertility and female choice in the lemon tetra ( Pisces, Characidae ). Science , 216 , 753-755. Olsson, M., Madsen, T., & Shine, R. 1997. Is sperm really so cheap? Costs of reproduction in male adders, Vipera berus . Proc. R. Lond. B , 264 , 455-459. Parker, G.A. & Sutherland, W.J. 1986: Ideal fr ee distributions when individuals differ in competitive ability: phenotype-limited ideal free models. Anim. Behav. , 34 , 12221242. Partan, S., Yelda, S., Price, V., & Shimizu, T. 2005. Female pigeons, Columba livia , respond to multisensory audio/video pla ybacks of male courtship behaviour. Anim. Behav. , 70 , 957-966. Passaglia, C., Dodge, F.A., Herzog, E., Jack son, S., & Barlow, R.B. 1997. Deciphering a neural code for vision. Proc. Natl. Acad. Sci. , 94 , 12649-12654. Patten, W. 1894. On the morphology and physiol ogy of the brain and sense organs of Limulus . Quart. J. Micr. Sci. , 35 , 1-96. Penn, D. & Brockmann, H.J. 1994. Nest-sit e selection in the horseshoe crab, Limulus polyphemus . Biol. Bull ., 187 , 373-384.
70 Penn, D. & Brockmann, H.J. 1995. Age-based st randing and righting in male horseshoe crabs, Limulus polyphemus . Anim. Behav. , 49 , 1531-1539. Powers, M.K., Barlow, R.B., Jr., & Kass, L. 1991. Visual performance of horseshoe crabs day and night. Visual Neurosci ., 7 , 179-189. Prado, C.P.A. & Haddad, C.F.B. 2003. Testes size in leptodactylid frogs and occurrence of multimale spawning in the genus Leptodactylus in Brazil. J. Herp. , 37 , 354-362. Prado, C.P.A. & Haddad, C.F.B. 2005. Sizefecundity relationships and reproductive investment in female frogs in th e Pantanal, South-western Brazil. Herp. Journal , 15 , 181-189. Quinn, E., Paradise, K., & Atema, J. 1998. Juvenile Limulus polyphemus generate two water currents that contact one proven and one putative chemoreceptor organ. Biol. Bull. , 195 , 185-187. Raethke, N., MacDiarmid, A.B., & Montgomer y, J.C. 2004.The role of olfaction during mating in the southern temperate spiny lobster Jasus edwardsii . Horm. Behav. , 46 , 311-318. Ramp, D. & Coulson, G. 2004. Small-scale patch selection and consumer-resource dynamics of eastern grey kangaroos. J. Mammal. , 85 , 1053-1059. Roberts, J.D., Standish, R.J., Byrne, P. G., & Doughty, P. 1999. Synchronous polyandry and multiple paternity in the frog Crinia Georgiana (Anura: Myobatrachidae). Anim. Behav. , 57 , 721-726. Rovner, J.S. & Barth, F.G. 1981. Vibrator y communication through living plants by a tropical wandering spider. Science , 214 , 464-466. Rudloe, A.E. 1980. The breeding behavior and pa tterns of movement of horseshoe crabs, Limulus polyphemus , in the vicinity of breeding b eaches in Apalachee Bay, Florida. Estuaries , 3 , 177-183. Schuster, S., Rossel, S., Schmidtmann, A., Jage r, I., & Poralla, J. 2004. Archer fish learn to compensate for complex optical distorti ons to determine the absolute size of their prey. Current Biol . 14 , 1565-1568. Shine, R. 2005. All at sea: aquatic life modifi es mate-recognition modalities in sea snakes ( Emydocephalus annulatus, Hydrophiidae). Behav. Ecol. Sociobiol. , 6 , 591-598. Shine, R. & Mason, R.T. 2001. Courting male garter snakes ( Thamnophilis sirtalis parietalis ) use multiple cues to identify potential mates. Behav. Ecol. Sociobiol. , 49 , 465-473.
71 Shine, R., Phillips, B., Waye, H., LeMaster, M., & Mason, R.T. 2003. Chemosensory cues allow courting male garter snakes to assess body length and body condition of potential mates. Behav. Ecol. Sociobiol. , 54 , 162-166. Shoger, R.L. & Bishop, D.W. 1967. Sper m activation and fertilization in Limulus polyphemus . Biol. Bull. , 133 , 485. Shoger, R.L. & Brown, G.G. 1970. Ultrastructura l study of sperm-egg interactions of the horseshoe crab, Limulus polyphemus L. (Merostomata: Xiphosura). J. Submic. Cytol. , 2 , 167-179. Siegel, S. & Castellan, N.J. Jr. 1988. Non Parametric Statistics for the Behavioral Sciences . New York: McGraw Hill Inc. Tregenza, T. 1995. Building on the ideal free distribution. Adv. Ecol. Res. , 26 , 253-307. Shochat, E., Abramsky, Z., Pinshow, B., & Whitehouse, M.B.A. 2002. Densitydependent habitat selection in migrator y passerines during st opover: what causes the deviation from IFD? Evol. Ecol. , 16 , 469-488. Veeranagoudar, D.K., Shanbhag, B.A., & Sa idapur, S.K. 2004. Foraging behavior in tadpoles of the bronze frog Rana temporalis : Experimental evidence for the ideal free distribution. J. Biosci. , 29 , 201-207. Ward, J.F., Austin, R.M., & MacDonald, D. W. 2000. A simulation model of foraging behavior and the effect of predation risk. J. Anim. Ecol. , 69 , 16-30. Wedell, N. Gage, M.J.G., & Parker, G.A. 2002. Sperm competition, male prudence and sperm-limited females. Trends Ecol. Evol. , 17 , 313-320. Whitman, T.G. 1980. The theory of habitat selection examined and extended using Pemphigus aphids. Amer. Natur ., 115 , 449-466. Widemo, F. 1998. Competition for females on leks when male competitive abilities differ: empirical test of a model. Behav. Ecol. , 9 , 427-431. Wyse, G.A. 1971. Receptor organization and function in Limulus chelae. Z. Vergl. Physiol. , 73 , 249-273.
72 BIOGRAPHICAL SKETCH Rachel L. Schwab was born on September 13, 1979, and grew up in the suburbs of Philadelphia. RachelÂ’s love for animals a nd nature began when she was young and grew through her participation in out door activities and camping trip s with her family. Rachel enrolled in as many science classes as po ssible during high school. She went on to earn her B.A. in biology with a concentration in ecology, evolution, and behavior from Skidmore College, Saratoga Springs, NY, in 2001. During college Rachel spent a semester abroad in Australia where sh e studied marine and rainforest ecology, volunteered at Cairns Tropi cal Zoo, and conducted research on the husbandry of seahorses, Hippocampus kuda , at Reef HQ Aquarium. When she was a senior in college Rachel worked as a laboratory assistant to Corey Freeman-Gallant studying the genetic mating system of Savannah Sparrows, Passerculus sandwichensis . Additionally, in the summers during college, Rachel worked as a Conservation Education Intern at the Philadelphia Zoo and an Animal Keeper at the Elmwood Park Zoo. Prior to entering graduate school Rachel worked as a dolphi n trainer at Kewalo Basin Marine Mammal Lab in Honolulu, HI, assisting with laborat ory studies on dolphin cognition and providing full-time care for the dolphins. She also worked as a research assistant studying the evolution and development of segmentation genes in polychaete annelids with Elaine Seaver at the University of Hawaii. Fina lly, Rachel moved back to the mainland and worked as an andrologist at the Reproductive Science Institute in Wayne, PA, to broaden her laboratory skills and scientific know ledge. She entered graduate school at the
73 University of Florida, Gainesville, FL, in the fall of 2003 under the advise of H. Jane Brockmann and Colette St. Mary. Rachel completed her M.S. studying the mating behavior of horseshoe crabs, Limulus polyphemus , in the spring of 2006 and plans to pursue a career in science education.