Aspects of Antipredation in Panulirus Argus and Panulirus Guttatus: Behavior, Morphology, and Ontogeny Peter Edward Bouwma

Home , Automimicry, Caribbean reef octopus, Dauphin Island Sea Lab, Palaemon serratus, Palibythus, Panulirus

Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2006 Aspects of Antipredation in Panulirus Argus and Panulirus Guttatus: Behavior, Morphology, and Ontogeny Peter Edward Bouwma

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected]

THE FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

ASPECTS OF ANTIPREDATION IN PANULIRUS ARGUS AND PANULIRUS

GUTTATUS: BEHAVIOR, MORPHOLOGY, AND ONTOGENY.

PETER EDWARD BOUWMA

A Dissertation submitted to the Department of Biological Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Fall Semester, 2006

The members of the Committee approve the dissertation of Peter E. Bouwma defended on October 4, 2006.

______William F. Herrnkind Professor Directing Dissertation

______David E. Thistle Outside Committee Member

______Joseph Travis Committee Member

______Walter R. Tschinkel Committee Member

______Thomas A. Houpt Committee Member

Approved:

______Timothy S. Moerland, Chair, Department of Biological Science

The office of Graduate Studies has verified and approved the above named committee members.

For Mom, Dad, Andy, Laura, Doc, and Dee.

I couldn’t have done it without you.

iii ACKNOWLEDGEMENTS

This work would not have been possible without the wisdom, patience, editing skills, and knowledge of lobster biology and behavioral ecology from my major professor, Bill Herrnkind. I can only hope to contribute to this discipline a fraction of what he has over the last 40 years. I also thank Walter Tschinkel, Joseph Travis, David Thistle, and Thomas Houpt for their contributions to this work as the members of my doctoral committee. This research was supported by funds from the Florida State University Department of Biological Science, the Sigma Xi Scientific Research Society, the Aylesworth Foundation for the Advancement of Marine Science, Florida State University Office of Research, the Office of Science Teaching Activities in the FSU Department of Biological Science, the Florida Institute of Oceanography, the Florida Fish and Wildlife Conservation Commission, Florida Sea Grant, and Old Dominion University. I would also like to thank the Keys Marine Laboratory and the Florida State University Coastal and Marine Laboratory for invaluable logistical support at all stages of this project. My work in the Florida Keys would not have been possible without many friends at the Keys Marine Laboratory and elsewhere. Cindy Lewis and Lonny Anderson have both helped me immeasurably throughout my tenure in the Keys. I will forever owe them both a debt of gratitude for everything they have done. I also thank Amy Parsons for all her help in the Keys. Fernando Briones, Steve Heiney, Chris Catlett, Lisa Tipsword, Jon Fajens, Chris Humphries, and Kevin McCarthy all provided sage advice, logistical support, and camaraderie; I thank them all for their help. Tracy Zeigler, Andrew Hannes, David Cleveland, Mary Alice Coffroth, Tom Walcott, Adrienne Sloan, Frank Trampus, Sean Kinane, and Darren Parsons all provided both logistical support and friendship during my tenure at KML. I would also like to thank the Lobster Crews at both Old Dominion University and Clemson University. My time in the Keys would not have been possible without the support of Mark Butler and his students Jason Goldstein, Donald Behringer, Scott Donahue, Jennifer Lear, Tom Dolan, and David Cook. Michael Childress and Adrianna

iv Zito were also invaluable in the planning and execution of my work in the Keys. John Hunt, Bill Sharp, Tom Matthews, Rodney Bertelson, and Lynn Cox at the Fish and Wildlife Research Institute in Marathon, FL provided additional logistical support. Allison George assisted in the care and training of triggerfish. Arley and Laura Bouwma assisted me with data collection in the field. Kari Lavalli was an invaluable resource concerning lobster/triggerfish interactions and loaned me a camcorder when I thought all else was lost. David Mann loaned me the hydrophone that changed my life. Many others, whose names I can no longer recall, also contributed to success of my work in the Keys; I thank you all! I would also like to thank the EERDG group in the Department of Biological Science at FSU, as well as other FSU faculty and graduate students for their help in the planning and analysis of this work. I also thank Judy Bowers for saving me from myself more times than I can count. Andrew Bouwma provided important statistical support. Kevin Smith wrote an event recorder program to help in the analysis of my videotapes. Kent Smith, Frank Stevenson, Steve Wilson, Barbara Shoplock, Maurizio Tomaiuolo, and Kim Young all assisted me in catching triggerfish. Mark Daniels, Bobby Henderson, and Dennis Tinsley helped care for triggerfish held at the FSUCML. Finally, I thank Denise Akob for her infinite patience and caring during both my data analysis and the composition of this work.

v TABLE OF CONTENTS

List of Tables ...... vii List of Figures...... ix Abstract...... xi

INTRODUCTION ...... 1

1. ANTIPREDATOR SOUND, BEHAVIOR, AND WEAPONRY IN THE SPINY LOBSTERS PANULIRUS ARGUS AND PANULIRUS GUTTATUS...... 9

2. APOSEMATISM BY WEAPONRY AND SOUND? RETALIATORY DEFENSIVE BEHAVIOR IN CARIBBEAN SPINY LOBSTER PANULIRUS ARGUS ...... 33

3. SOUND PRODUCTION IN CARIBBEAN SPINY LOBSTER PANULIRUS ARGUS AND ITS ROLE IN ESCAPE DURING PREDATORY ATTACK BY OCTOPUS BRIAREUS ...... 55

4. THE ONTOGENY OF ANTI-PREDATOR RESPONSES TO ALARM ODOR IN CARIBBEAN SPINY LOBSTER PANULIRUS ARGUS...... 73

CONCLUSIONS...... 91

REFERENCES ...... 95

BIOGRAPHICAL SKETCH ...... 111

vi LIST OF TABLES

Table 1.1. (a) Descriptions of lobster defensive actions and (b) triggerfish actions recorded during attacks by gray triggerfish Balistes capriscus on Panulirus argus and P. guttatus individuals...... 26

Table 1.2. (a) Mean lobster defensive actions per minute of encounter (± SE) for P. argus and P. guttatus and (b) results of two-way ANOVAs comparing means for time to subdue and lobster defensive actions (tailflips, whips, pirouettes) with fish individual and species type as factors (plus the interaction of fish individual X species type). N = 15 for P. argus and 14 for P. guttatus for all ANOVAs. Number of different fish: N = 6 for both species...... 27

Table 1.3. Measurements of lobster antennal weaponry as predicted by body size (Carapace Length) and species (Panulirus argus or P. guttatus) in multiple linear regression models. Separate models were fit for (a) the width of the antenna flagellum 10 cm from the base (Wflag10), (b) antenna flagellum cross-sectional area near base (XSflag), and (c) antenna base cross-sectional area (XSbase). DF = degrees of freedom, Type III SS = type three sums of squares, F = F statistic...... 28

Table 1.4. Measurements of lobster body size as predicted by carapace length and species (Panulirus argus or P. guttatus) in multiple linear regression models. Separate models were fit for (a) the width of the anterior carapace (Wac), (b) width of the abdomen rd (Wab), and (c) width of the 3 walking leg (W3wl). DF = degrees of freedom, Type III SS = type three sums of squares, F = F statistic...... 29

Table 2.1. (a) Descriptions of a subset of individual Panulirus argus defensive actions (during attacks only) and (b) triggerfish (Balistes capriscus) attack behaviors recorded during encounters in both the aposematic trials and the triggerfish choice experiment....50

Table 2.2. (a) Number of defensive actions per minute of encounter and results of t-tests comparing means for stridulating (N = 12) and muted (N = 13) lobsters during encounters with triggerfish. (b) Triggerfish attack behaviors per minute of encounter and results of t- tests comparing means for fish attacking stridulating (N = 12) versus muted (N = 13) lobsters...... 51

Table 3.1. Numbers of individuals escaping from octopuses at least once, individuals escaping at least once without injury, individuals grasped on the antenna on the 1st attempt, individuals captured and killed on the first attempt, and individuals which lost an antenna through autotomy before capture. Results of G-tests comparing frequencies between stridulating and muted lobsters...... 70

vii Table 4.1. Results of the regression model for distance moved on lobster body size (carapace length in mm)...... 88

viii LIST OF FIGURES

Figure 1.1. (a) Defensive actions per minute by Panulirus argus and P. guttatus during encounters by triggerfish. (b) The occurrence of stridulation during P. argus and P. guttatus defensive actions against triggerfish. Stridulation occurrence is reported as the percent of times each type of action was accompanied by stridulation...... 30

Figure 1.2. Anterior comparison photograph of P. argus (right) and P. guttatus (left) with arrows indicating the stridulating organ and the 2nd antennae...... 31

Figure 1.3. (a) Antennna flagella width 10 cm from the base, (b) antenna flagella cross- sectional area (π*L/2*W/2) near the base, (c) antenna base cross-sectional area, (d) anterior carapace width, (e) abdomen width, and (f) the width of the 3rd walking leg plotted on body size (carapace length) for P. argus and P. guttatus ...... 32

Figure 2.1. Caribbean spiny lobster Panulirus argus in an alert defensive posture (antennae spread, high on legs, abdomen and telson extended) shortly before lunging at a grey triggerfish Balistes capriscus...... 52

Figure 2.2. (a) Numbers of naïve triggerfish (Total fish = 30; 15 against stridulating lobsters, 15 against muted lobsters) which successfully attacked and subdued muted or stridulating spiny lobsters by the end of the 1st day (2 total hours) and by the end of the 5th day (10 total hours) during the attack trials. Asterisks indicate significant differences between fish attacking stridulating versus muted lobsters (* P ≤ 0.05; ** P ≤ 0.01, *** P ≤ 0.001). (b) Choice of which lobster to attack and subdue by experienced triggerfish when given a choice between stridulating and muted individuals of the same size (N = 20) ...... 53

Figure 2.3. (a) Triggerfish behaviors per minute of encounter during attacks against stridulating and muted lobsters. (b) Defensive behaviors per minute of encounter exhibited by lobsters during attacks by triggerfish. N = 12 for stridulating lobsters, N = 13 for muted...... 54

Figure 2.4. (a) Minutes of encounter before a kill (N = 12 for stridulating, N = 13 for muted), minutes after introduction of the triggerfish to the arena before the first attack (N = 15 for stridulating and muted), and minutes of encounter before a kill in the choice experiment (N = 9 for stridulating, N = 11 for muted). Asterisk indicates significance at the 0.05 level. (b) Minutes of encounter before a kill for all experienced and inexperienced triggerfish (N = 20 for experienced and inexperienced fish) ...... 55

Figure 3.1. (a) Sequence of screen captures showing an octopus 1) stalking, 2) reaching, and 3) grasping at a lobster. The lobster 4) successfully escapes. (b) Successful capture of a lobster by an octopus ...... 71

Figure 3.2. Percent of lobsters escaping from octopuses at least once, lobsters escaping at least once without injury, lobsters grasped on the antenna on the 1st attempt, lobsters captured and killed on the first attempt, and lobsters which lost an antenna through autotomy before capture. Asterisks indicate significant differences between stridulating and muted lobsters (* P ≤ 0.05; ** P ≤ 0.01, *** P ≤ 0.001) ...... 72

Figure 3.3. Numbers of lobsters stridulating and silent at 10 second intervals after capture by an octopus (N = 15 lobsters) ...... 73

Figure 4.1. (a) Percent of nomadic phase and post-algal P. argus individuals moving from dens or staying in response to alarm odor. (b) Percent of nomadic phase and post- algal phase lobsters moving from dens or staying when sheltering alone or with conspecifics. Asterisks indicate significant differences in frequencies (* P ≤ 0.05; ** P ≤ 0.01, *** P ≤ 0.001)...... 89

Figure 4.2. Distance moved (m) versus body size (mm carapace length) by nomadic and post-algal phase P. argus individuals abandoning natural dens in response to conspecific alarm odor ...... 90

Figure 4.3. Seconds spent by P. argus algal phase juveniles in locomotion, moving without locomoting, with antennae in a forward position, and in an elevated posture during control and test (conspecific alarm odor) periods. Asterisks indicate significant differences in means between control and test periods (* P ≤ 0.05; ** P ≤ 0.01, *** P ≤ 0.001) ...... 91

x ABSTRACT

Spiny lobsters (Family Palinuridae) are large, diverse, and abundant marine crustaceans, which have conquered tropical, subtropical, and temperate coastal waters around the globe despite strong predation pressure. The mechanisms and function of antipredation strategies for most species in this highly successful taxon, encompassing behavior, morphology, and life-history characteristics, are poorly understood, particularly against natural predators. I investigate mechanisms of antipredation in spiny lobster Panulirus argus in the open during the day, at night, and while sheltering diurnally in natural dens. I also examine the function of putatively defensive acoustic signals produced by P. argus during diurnal attacks by piscine predators and while escaping octopuses at night. I also compare and contrast the mechanism and survival value of antipredator behavior and morphology between sympatric Panulirus argus and P. guttatus. Finally, I investigate ontogenetic changes in defensive behavior by diurnally sheltered P. argus to chemically-mediated predator cues. Nearly 40 species of spiny lobsters produce a characteristic sound (termed stridulation), speculated to deter predation. The occurrence and efficacy of stridulation has not been documented quantitatively during encounters with natural predators. I examined sound production in the sympatric spiny lobsters Panulirus argus and P. guttatus during attacks by their common predator, gray triggerfish Balistes capriscus, to determine if lobsters produce sound during defense, how stridulation integrates with behavioral and morphological defenses, and how interspecific differences in sound production relate to efficacy in repelling predators. Both lobster species stridulated coincident with specific defensive actions during triggerfish attack. In P. argus, stridulation occurred both during antennal lunging and during escape attempts (rapid retreat by tailflips). Panulirus guttatus stridulated only coincident with tailflips and did not lunge. Same-sized individuals of P. guttatus were subdued ~3 times more quickly on average than P. argus. The two species differed also in the relative size of the primary defensive weapons, the spinose 2nd antennae, which were far more robust in P. argus,

xi particularly at larger body sizes. These results suggest that stridulation is an integral component of aggressive defense and escape behavior in spiny lobsters. The timing of sound production during aggressive, retaliatory defensive behavior (lunging) by P. argus suggests an aposematic role for stridulation against triggerfish. Using staged encounters of P. argus with B. capriscus, I examined whether stridulation, coincident with thrusting spines during aggressive defense, functions aposematically or simply renders a defending lobster more difficult to subdue without playing an aposematic role. I demonstrate, by disabling the stridulating organ in some lobsters (muting), that sound plays a vital role in defense against inexperienced (naïve) triggerfish, resulting in fewer successful attacks in subsequent encounters. Choice experiments with triggerfish that previously bypassed defenses and consumed lobsters show that experienced attackers do not choose muted lobsters over stridulating individuals. I propose that stridulation by P. argus against triggerfish is aposematic, as part of a multi-modal display, advertising the lobster’s spiny defenses to predators. It is widely documented that sound production in P. argus and other spiny lobsters accompanies grasping of the carapace or other disturbance by human captors. Additionally, stridulation accompanies tailflip escape attempts during attacks by triggerfish. Although sound production during daytime attacks does not appear to increase survival against triggerfish, stridulating during escape may be more effective against grasping predators like octopus. Here, I investigate P. argus defensive behavior during nighttime encounters with Caribbean reef octopus Octopus briareus to determine whether P. argus stridulate during octopus attacks, how stridulation is used along with other defensive behavior (e.g. tailflips), and whether sound production improves survival in stridulating individuals. Lobsters stridulate both during grasping attacks by octopus and during escape attempts after being captured. Stridulating lobsters are also more likely to escape from attacking octopuses and remain uncaptured longer during encounters. I suggest that improving the efficacy of tailflip escapes against octopus may have been the function for which the stridulating organ initially evolved in the Stridentes clade of the Palinuridae. Benthic stages of P. argus reside in shelters during the day as a primary means of antipredation. However, when an active predator approaches and/or successfully attacks a

xii nearby conspecific, these individuals must decide whether to emigrate quickly from the area or remain in shelter (dens or macroalgae) and rely on crypticity, defensive behavior, or the presence conspecifics to avoid attack, injury, or death. In this study, I examine how the three benthic juvenile phases of Caribbean spiny lobster Panulirus argus respond to exposure to fresh conspecific body fluid and how antipredator behavior, particularly the decision to stay or leave the area, changes during ontogeny. Additionally, I examined how the presence of conspecifics affects the decision to stay or leave in gregarious juvenile stages of P. argus. Although all size classes of P. argus respond to alarm odor, the decision to stay or leave dens changes unexpectedly with increasing body size and in the presence of conspecifics. Once shelters were abandoned, body size was a strong indicator of distance traveled in response to alarm odor. This indicates that Panulirus argus undergo an ontogenetic shift in defensive behavior, more frequently leaving dens in response to alarm odor and traveling across open substrate during the day, but only after reaching a body size at which they can effectively defend against predators.

xiii INTRODUCTION

Spiny lobsters (family Palinuridae) are among the most diverse, widespread, abundant, socially complex, and largest of marine crustaceans (Lipcius & Eggleston, 2000). They are nearly ubiquitous in tropical, subtropical, and some temperate coastal waters of North and South America, Africa, India, the Mediterranean, Europe, the Orient, Australia, New Zealand, and the Pacific Islands (Kanciruk, 1980; Holthuis, 1991; Lipcius & Eggleston, 2001; Ptacek et al., 2001). The approximately 50 species and 11 genera of palinurid lobsters (Patek et al., 2006) also live in diverse habitats, ranging from deep, open, particulate substrates at >300 m depth to vegetated nursery areas, hard substrates, and coral reefs in the littoral zone as shallow as 1 m (George & Main, 1967; Kanciruk, 1980; Holthuis, 1990; George, 2005; Butler et al., 2006). Full grown spiny lobsters range from 1-10 kg and 300-500 mm total length (not including antennae), with some species attaining >11 kg and 600 cm (e.g., Sagmariasus verrauxi; Phillips et al., 1990; Holthuis, 1991; Wahle & Fogarty, 2006). Palinurids are also among the fastest growing marine crustaceans, some tropical species reaching one kilogram in weight 2-4 years after settlement (Butler et al., 2006). A recent study suggests that growth in Caribbean spiny lobster Panuliurs argus can be even more rapid, some individuals reaching 85% of maximum size (4-5 kg) within 4 years after settlement (Maxwell, 2006). Numerous palinurid species are so abundant that they support local, artisanal or large-scale commercial fisheries nearly everywhere they occur (Lipcius & Eggleston, 2000; Groeneveld et al., 2006; Phillips & Melville-Smith, 2006). Spiny lobsters support some of the largest, most continuously exploited, and economically valuable fisheries worldwide (Lipcius & Eggleston, 2000). This demonstrates that many palinurid species across multiple genera have been able to thrive in a diversity of marine habitats (Kanciruk, 1980). Their success can in part be attributed to a high reproductive capacity (up to 2 million eggs in a single clutch) (Bertelson & Matthews, 2001), some of the longest larval periods among marine invertebrates (up to 24 months), with a high dispersal potential (Phillips et al., 2006), and the capacity of the puerulus settlement stage for predator avoidance and habitat

1 selectivity (Acosta & Butler, 1999; Phillips et al., 2006). Yet, much of what allows spiny lobsters to attain rapid growth and large body size in such great numbers occurs in the benthic life phase (Kanciruk, 1980; Butler et al., 2006). Young lobsters, generally less than 10 mm in total length at settlement, must somehow acquire resources and grow to adulthood over the span of just a few years while avoiding natural mortality from a diverse array of enemies, including: piscine predators such as groupers, porgies, triggerfish, snappers, toadfish, wrasses, eels, sharks, and rays (Heydorn, 1969; Kanciruk, 1980; Howard, 1988; Smith & Herrnkind, 1992; Lozano-Alvarez & Spanier, 1997; Sadovy & Eklund, 1999; Schratweiser, 1999; Cruz & Phillips, 2000; Herrnkind et al., 2001; Barshaw et al., 2003; Young & Winn, 2003; Booth, 2006; Groeneveld et al., 2006); molluscs, including some boring whelks (Munro, 1974; Barkai & McQuaid, 1988) and octopuses (Heydorn 1969; Joll, 1977; Kanciruk, 1980; Cruz & Phillips, 2000; Booth, 2006; Groeneveld et al., 2006); portunid crabs (Smith & Herrnkind, 1992); sea turtles (Munroe, 1974); seals (Heydorn, 1969; Berry, 1971; MacDonald, 1982; Wickens, 1996); and dolphins (Munroe, 1974). These predators may be diurnally or nocturnally active and present a wide range of hunting tactics (Butler et al., 2006). Those spiny lobsters that undergo ontogenetic shifts in habitat, especially as juveniles, must confront different suites of predators at each life-history stage or new habitat (Butler et al., 2006). Therefore, it is not surprising that predation is the primary source of natural population mortality, particularly during the juvenile period (Kanciruk, 1980; Butler & Herrnkind, 2000, Butler et al., 2006). Despite strong predation pressure, immense numbers of spiny lobsters survive the predator gauntlet each year to replenish fisheries around the world. This suggests that each species has evolved effective antipredation to counteract multiple predators and strategies at each ontogenetic stage. Antipredation in the Palinuridae is mediated by both behavior and morphology (Kanciruk, 1980; Barshaw et al., 2003; Childress & Jury, 2006). All spiny lobsters are morphologically similar, with numerous forward-pointing spines covering the anterior carapace and long 2nd antennae, used as weaponry (Kanciruk, 1980; Barshaw et al., 2003; George, 2005; Childress & Jury, 2006). Shallow-water species also have highly calcified exoskeletons, strong legs, and large rostral horns protecting the eyes, while deep-water species generally have weaker legs, less-calcified

2 exoskeletons, and sunken eyes beneath low, flattened horns (George, 2005). Deep-water palinurids presumably face less predation pressure, although virtually nothing is known of their ecology, life history, or behavior (George & Main, 1967; George, 2005). Among lobster morphological defenses generally, shallow-water palinurids appear to fall midway between homarid (clawed) lobsters with major chelae on the 1st appendages and the achelate scyllarid lobsters, which lack defensive weaponry (Barshaw et al., 2003). Shell thickness in these palinurids also lies midway between the thin shells of clawed lobsters and the thick, highly puncture-resistant armor of scyllarids (Barshaw et al., 2003). All spiny lobsters are also endowed with a broad tailfan and powerfully muscled, elongated abdomen characteristic of the macruran body type, which allows for rapid (but short-distance) retreat (Kanciruk, 1980; Spanier et al., 1991; Patek et al., 2006). Although body spines, antennal weaponry, thick armor, and tailflip escape may offer protection from small, weak, or slow predators, particularly when the lobster attains large size, quick (e.g. triggerfish, octopus) or powerful (e.g. nurse sharks, goliath grouper) predators are able to overcome morphological defenses (Kanciruk, 1980; Butler et al., 2006; Childress & Jury, 2006). Because spines and tailflips alone are inadequate to confront the multitude of predators that spiny lobsters face in the varied habitats they exploit or pass through, antipredator behavior is essential for survival, particularly in smaller-bodied juvenile phases (Kanciruk, 1980; Childress & Jury, 2006). However, other than generalized predator avoidance strategies shared across the Palinuridae, the mechanisms and function of antipredator behaviors during the benthic period are poorly defined for most species. Most knowledge of palinurid antipredation comes from experiments or rigorous observations on Panulirus argus, spotted lobster P. guttatus, European/Mediterranean spiny lobster Palinurus elephas, and New Zealand spiny lobster Jasus edwardsii, although anecdotal accounts of putatively antipredator devices abound in the literature for numerous other species (Butler & Herrnkind, 2000; Butler et al., 2006; Childress et al., 2006). Overall, it appears that lobster antipredator behavior employs three primary strategies 1) crypticity to reduce detection, 2) avoidance behaviors and sheltering to reduce encounters with predators, and 3) defensive and escape behaviors to avoid capture (Childress & Jury, 2006).

3 Most early benthic stage juveniles are cryptic in appearance and use crevice shelters or dense foliage to avoid detection or deter attack (Butler & Herrnkind, 2000). Although crypticity becomes increasingly less effective as individuals grow larger, lobsters adopt additional means to avoid encountering predators (Childress & Jury, 2006). In most species, large juvenile and adult lobsters reduce encounters with diurnal predators by sheltering in deep crevices during the day and foraging in the open only at night (Butler & Herrnkind, 2000; Butler et al., 2006; Childress & Jury, 2006). Caribbean spiny lobster Panulirus argus, which are highly sensitive to chemical stimuli (Derby et al., 1997; Berger & Butler, 2001), also presumably reduce encounters by avoiding shelters infused with chemical cues from predators or dead conspecifics (Berger & Butler, 2001; Parsons & Eggleston, 2005; Briones-Fourzan et al. 2006). Despite these avoidance behaviors, spiny lobsters still encounter dangerous predators, both during daytime sheltering and solitary nocturnal foraging in the open away from immediate refuge (Smith & Herrnkind, 1992; Childress, 1995). All spiny lobsters possess the capacity to attempt to escape a predator’s attack or grasp by tailflipping (Kanciruk, 1980; Spanier et al., 1991; Childress & Jury, 2006). How and when exposed lobsters use tailflips after encountering predators in nature, during the day or at night, is less clear. Briones-Fourzan et al. (2006) determined in the laboratory that both Panulirus argus and P. guttatus initiated tailflips within 0.3 seconds after a physical strike on the antennae during daytime trials. Although P. guttatus responded comparatively quicker, traveled farther, and for greater duration (Briones- Fourzan et al. 2006), it is not clear whether these differences affect survival against natural predators. Numerous authors (Lindberg 1955; Heydorn 1969; Berry 1971; Herrnkind et al., 2001; Parsons, 2005; Barshaw et al., 2003; Briones-Fourzan et al. 2006) report tailflipping by lobsters that were disturbed by fish, octopus, or (mostly) humans, both in the wild and in the laboratory, during the day. Fewer studies document tailflips by spiny lobsters against natural predators during nighttime encounters (Cobb, 1980). While residing diurnally in dens, lobsters may use tailflips to retreat deeper into shelter (Lindberg, 1955; Heydorn, 1969; Berry, 1971; Cobb, 1980; Kanciruk, 1980; Briones- Fourzan et al., 2006) or create a murky cloud of sand or mud, presumably obscuring the lobster from visual predators (Heydorn, 1969; Berry, 1971). Many lobsters also use

4 strong legs to brace against the walls of a shelter or cling to rocky substrate to avoid being extracted (Kanciruk, 1980; Briones-Fourzan, 2006). Again, most of these observations involve interactions with human divers rather than responses to attacks by natural predators. Some species of spiny lobster, including the spotted lobster Panulirus guttatus, are rarely observed during daytime at the entrance to dens or in the open (Holthuis & Loesch, 1967; George, 1972; Sharp et al., 1997). This cryptic, secluded behavior along with tailflipping and clinging/bracing against physical extraction, apparently comprise the entirety of their antipredator strategy (Briones-Fourzan, 2006). Some other lobster species are visible and potentially vulnerable to predators during the day, typically residing in shallow crevices or at the entrances to dens (Lindberg, 1955; Holthuis & Loesch, 1967; Heydorn, 1969; Berry, 1971; Cobb, 1980; Pillai et al., 1985; Briones- Fourzan et al., 2006). Additionally, many species exhibit long-distance migrations or other movements exposing individuals on open bottom during the daytime (Herrnkind, 1980; George, 2005; Childress & Jury, 2006). Some of these highly exposed species display additional antipredator behavior (in addition to tailflips) to counteract diurnal predators in the open. Spiny lobsters counteract attacks by whipping, prodding, thrusting, and/or lashing their antennae at predators, thus parrying the attacker (Cobb, 1980; Kanciruk, 1980; Parsons, 2005; Briones-Fourzan et al., 2006; Childress & Jury, 2006). Some species are highly aggressive, pre-emptive defenders, using lunges, in which the cephalothoracic and antennal spines are driven into the attacker by rapid, forward- directed tailfan strokes (Barshaw et al., 2003; Parsons, 2005). Large lobsters may also aggressively attack predators with the robust and often spinose front legs, as has been observed in response to human divers (Lindberg, 1955; Kelly et al., 1999). Often, highly exposed spiny lobsters exhibit group defensive behaviors, in addition to individual defense, including single-file queuing (Herrnkind, 1967; Bill & Herrnkind, 1976; MacDonald et al., 1984) and coordinated defensive formations (Kanciruk & Herrnkind, 1978; Kelly et al., 1999; Herrnkind et al., 2001). One of the most widely reported aspects of spiny lobster antipredation may also be the least well understood. Spiny lobsters of the strident clade produce a characteristic loud sound when disturbed and coincident with certain defensive actions against

5 predators (George & Main, 1967; Cobb, 1980; Patek & Oakley, 2003). This stridulatory burst is also triggered by grasping and prodding by humans, the situation in which it is most commonly reported (Moulton, 1957; George & Main, 1967; Atema & Cobb, 1980; Kanciruk, 1980; Patek & Oakley, 2003; Childress & Jury, 2006). The evolution of a sound-producing organ ~200 Mya in the Palinuridae split the taxon into the Stridentes (sound producers) and Silentes (George & Main, 1967; Patek et al., 2006). Because no current or extinct species in the Stridentes has lost the acoustic organ, stridulation appears to be highly conserved (George & Main, 1967; Patek et al., 2006). Nearly every reported observation of stridulation is associated with physical disturbance, leading most authors to assume an antipredator function (Moulton, 1957; George & Main, 1967; Mulligan & Fischer, 1977; Patek, 2001; Patek & Oakley, 2003). However, few studies document the co-occurrence of stridulation with other events during encounters with natural predators (Lindberg, 1955; Berry, 1971; Cobb, 1980) and none has previously demonstrated an antipredatory effect, communication, or any other functional consequence of stridulating (Childress & Jury, 2006). In summary, the literature indicates variation in palinurid antipredator strategies among both species and ontogenetic stages within species. However, essential questions remain as to the degree to which antipredator behaviors correspond to particular ecological circumstances for each species or ontogenetic stage. In addition, the proximate mechanisms of antipredation in this group and the survival value of putative antipredator behaviors, such as lobster stridulation, are poorly understood, particularly in a comparative context. This makes evolutionary comparisons difficult. In this study, I investigate mechanisms of antipredation in juvenile spiny lobster Panulirus argus during three ecologically important circumstances: in the open during the day, while sheltering in natural dens during the day, and in the open at night. I also examine the functional significance of stridulating by P. argus during defense against daytime attacks by triggerfish and while tailflipping during escape from octopuses at night. I also compare and contrast the use of stridulation, defensive actions, and weapon morphology between sympatric Panulirus argus and P. guttatus and examine how their respective antipredatory actions affect survival during attacks by triggerfish. Additionally, I investigate ontogenetic changes in behavioral responses by diurnally

6 sheltered juvenile P. argus to chemical cues from injured conspecifics. In examining these aspects of antipredation in P. argus and P. guttatus, I hope to advance understanding of the evolution of antipredator behavior in the Palinuridae, as well as to address important behavioral ecological questions concerning the function of stridulation. In determining the functional consequences of lobster stridulation, I also address important questions concerning the function of defensive acoustics generally. While defensive sounds have been described in many different animal taxa, few studies document their functional significance broadly (Högstedt, 1983; Klump & Shalter, 1984), or particularly, in marine systems (Myrberg, 1981; Childress & Jury, 2006). Because lobsters may produce sound in several distinct defensive contexts in captivity and can be silenced easily without invasive procedures (Moulton, 1957; Patek, 2002), they represent a promising organism with which to experimentally test hypotheses about the function of antipredator sound in the marine environment. For example, recent work on aposematic animals advocates an expanded definition for aposematic defenses (conspicuous signals advertising defended prey) to include brightly colored, mechanically defended spiny prey, using odors and sounds as additional conspicuous signals (Rowe, 2002; Inbar & Lev-Yadun, 2005, Speed & Ruxton, 2005), in addition to well known poisonous or distasteful chemical defenses. Lobsters, with often brightly-colored spines and stridulatory sound produced during aggressive lunges versus triggerfish (as seen during pilot experiments), may exemplify this unusual (but perhaps not uncommon) type of aposematism. Additionally, the role of sound in distress calls (after capture) produced by many marine animals is not well demonstrated empirically (Myrberg, 1981; Childress & Jury, 2006). Stridulation during nocturnal tailflip escapes from octopus, similar to distress calls in some birds and insects (Masters, 1979; Klump & Shalter, 1984), may function separately from aposematism, potentially startling these predators.

ORDER OF PRESENTATION

Chapter two describes how differences between Panulirus argus and P. guttatus in the use of stridulation, defensive behaviors, and the morphology of antennal weaponry relates to survival in attacks by triggerfish predators. Chapter three tests whether

7 stridulation operates aposematically against triggerfish, facilitating learning to avoid suffering a lobster’s spiny defenses. Chapter four examines how stridulation is used to facilitate escape from octopus during nighttime encounters. Chapter five describes an ontogenetic shift in defensive behavior by juvenile P. argus from strictly retreat into shelter to abandoning a den where a conspecific suffers a successful attack, sensed as body fluid chemical cues.

8 CHAPTER 1

ANTIPREDATOR SOUND, BEHAVIOR, AND WEAPONRY IN THE SPINY

LOBSTERS PANULIRUS ARGUS AND PANULIRUS GUTTATUS

INTRODUCTION

Evidence of sound produced as a predator deterrent is common among terrestrial animals, including birds (e.g. Curio, 1978; Högstedt, 1983; Klump & Shalter, 1984; Neudorf & Sealy, 2002), mammals (e.g. Sherman, 1977; Russ et al., 1998; Caro et al., 2004; Randall, 2001), and arthropods (e.g. Alexander, 1967; Smith & Langley, 1978; Masters, 1979; Marshall et al., 1995). In marine systems, despite a wide variety of sound producing taxa, antipredator sound is known primarily from a few species of fish and crustaceans. Some 35 taxa of fish will produce sound in response to prodding, capture, or disturbance by humans, yet few studies document either the production or efficacy of these putative antipredator sounds during defense against natural predators (Myrberg, 1981). For example, cod (Gadus morhua) produce grunts when penned by conger eels (Brawn, 1961) and clicks when approached by harp seals, which may cause the seals to abort an attack (Vester et al., 2004). Squirrelfish, soldierfish, and damselfish also produce sound, apparently directed at potential predators, as part of a defensive display (Myrberg, 1981; Mann & Lobel, 1998). Among marine crustaceans, the best-known sound-producers are the snapping shrimps (Alpheidae). In alpheids, sound is produced along with a powerful water jet caused by rapidly closing the large, modified snapping claw. This action is commonly exhibited in agonistic interactions against conspecifics, other alpheids, as well as other species and, in some instances, to kill prey (Hazlett & Winn, 1962; Schmitz & Herberholz, 1998; Duffy et al., 2001). The role of sound in these actions as distinct from the water jet is unclear because the two are produced simultaneously by the same mechanism (Versluis et al., 2000). Clawed American lobsters Homarus americanus also

9 produce sound and carapace vibrations using internal movements of the antennal muscles, but without a specialized sound-producing organ (Fish, 1966; Henninger & Watson, 2005). These sounds and vibrations are produced when humans startle, grasp, or threaten H. americanus and, although apparently operating to deter predation, their behavioral significance remains unclear (Henninger & Watson, 2005) Many spiny lobster species (Palinuridae) also produce sounds long presumed to act in defense. Palinurid lobsters produce sound from an acoustic organ located bilaterally at the base of the large second antennae, anterior to the eyes (Moulton, 1957; Hazlett & Winn, 1962; Patek, 2001; 2002). Because spiny lobsters have a long history of interactions with human fishers, producing sound when handled, references to spiny lobster sound appear as early as the Greek classical literature (Moulton, 1957). Descriptions of spiny lobster sound, commonly called stridulation (Moulton, 1957), first appeared in the scientific literature in the 1800's (Mobius, 1867; Parker, 1878), followed by detailed study of the anatomy of the stridulating organ (Moulton, 1957; Smale, 1975; Meyer-Rochow & Penrose, 1974) and descriptions of the acoustic signal (Hazlett & Winn, 1962; Meyer-Rochow & Penrose, 1976; Mulligan & Fischer, 1977). Despite scattered attention given this phenomenon, the unique mechanism of sound production was only recently elucidated (Patek, 2001; 2002). The acoustic signal is produced when a small piece of exoskeleton (the plectrum), articulated with the antenna base, is drawn posteriorly over a raised ridge on the antennular plate (the file). The soft, elastic underside of the plectrum creates friction as it interacts with the surface of the file, causing the plectrum to alternately stick and slip as it moves up the file. Sound is produced with each stick-and-slip of the plectrum, resulting in a broadband, pulsed signal ranging from ~50-30,000 Hz (Mulligan & Fischer, 1977; Patek, 2002). The 11 genera and some 50 species (including coral lobsters) (Holthuis, 1991; Patek & Oakley, 2003) of palinurid lobsters occupy the coastal shelves of every continent except Antarctica. Many species are large and numerically abundant enough to support highly valuable commercial and artisanal fisheries (Lipcius & Eggleston, 2000). The evolution of the palinurid acoustic organ splits the Palinuridae into two monophyletic taxa: the sound-producing "stridentes" (8 genera, 39 species) and the non-sound producing "silentes" (3 genera, 11 species) (George & Main, 1967; Patek & Oakley,

10 2003). George & Main (1967) proposed, based on morphological and fossil evidence, that this split in the Palinuridae occurred some 200 MYA, and the ability to produce sound has been retained in all extant stridentes taxa. Although the sound-producing mechanism in part defines the taxonomy of the Palinuridae and persists in some 39 extant species, we know virtually nothing about the consequence of spiny lobster sound in the wild. Many authors have speculated that lobster stridulation functions in antipredation: as a defensive action (e.g. as a startling signal), as an alarm call to conspecifics, or as a warning directed at attacking predators. Most of this speculation derives from anecdotal accounts of interactions in the field (Lindberg, 1955; Moulton, 1957; Berry, 1971; Smale, 1975) and experiments in the lab (Meyer-Rochow & Penrose, 1974; 1976; Mulligan & Fischer, 1977; Meyer-Rochow et al., 1982; Patek, 2001; 2002) in which humans elicited sound production by lobsters. Lindberg (1955) observed in the field that Panulirus interruptus stridulate only when contacted by a fish entering the same den and not when threatened without contact, with few other details about the encounters. The only experimental study to date that describes sound production during interactions with natural predators reports that young juvenile Panulirus longipes (cygnus) lobsters do not produce sound during attacks by octopus but do so during agonistic encounters with conspecifics (Berrill, 1976). Therefore, although lobster stridulation may function in antipredation, we currently have insufficient information indicating if and when lobsters produce sound against natural predators, or what influence it might have on the outcome of an attack. In this study, I investigated the production of sound by two sympatric congeners, Panulirus argus and P. guttatus during defensive actions elicited by attacks from a common piscine predator, the gray triggerfish Balistes capriscus. I examined three questions: (1) Do lobsters produce sound during attacks by triggerfish in the defensive context, (2) how does stridulation integrate with behavioral and morphological defenses and, (3) are there intraspecific differences in the use of sound and relative effectiveness of defensive behavior between these two species? I demonstrate that both P. argus and P. guttatus produce sound in association with defensive actions but differ in the way in which sound is used and in the relative effectiveness of defense during attacks by triggerfish.

MATERIALS AND METHODS

Experimental system The lobsters studied, Panulirus argus and P. guttatus, both commonly stridulate when captured or prodded by human divers in the wild. Although these morphologically similar species shelter in crevices during the day and forage under cover of night as a primary means of avoiding diurnal predators (Butler & Herrnkind, 2000), they otherwise exhibit dissimilar lifestyles and, therefore, may use stridulation differently in antipredation. The Caribbean spiny lobster, P. argus, has a complex benthic life-history, with three behaviorally and ecologically distinct ontogenetic habitat shifts before reaching adulthood. In Florida, postlarval P. argus settle from the plankton into nearshore hardbottom areas, far from adult habitat on the reef (Marx & Herrnkind, 1985). Initial stages of P. argus (5-20 mm Carapace Length) are cryptic, living in dense macroalgae where they are substantially protected from predators (Smith & Herrnkind, 1996). After ~3 months, young lobsters move out of the algae and begin to shelter in crevices as "post-algal" phase juveniles (20-40 mm CL). A final juvenile ontogenetic shift to a "nomadic" stage is made at approximately 40-50 mm CL (Butler & Herrnkind, 2000). Nomadic sub-adults emigrate from nursery habitat out to the reef and often travel many kilometers (20-40) to do so (Warner et al., 1977; Gregory & Labisky, 1986). In some regions, nomadic P. argus also begin to participate in annual or episodic mass migrations, where lobsters travel by the thousands, day and night, across shelterless substrate in long, single-file queues (Herrnkind, 1969; 1980; 1985). The increased movement by nomadic P. argus, during migratory and non-migratory periods, exposes these pre-reproductive individuals to a heightened risk of encountering diurnal predators, especially triggerfish, away from shelter (Kanciruk, 1980). In contrast, spotted lobster, Panulirus guttatus, postlarvae settle directly into adult habitat on the reef (Sharp et al., 1997). Cryptic P. guttatus rarely emerge during the day, instead sheltering deep in crevices. They do not respond to disturbance by moving to the front of dens as does P. argus (Sharp et al., 1997; P. Bouwma, pers. obs.) Additionally, spotted lobsters are not migratory and may never leave the reef patch on which they

12 settled, so they are unlikely to face the same elevated diurnal predation risk as nomadic phase and adult P. argus (Sharp et al., 1997). Gray triggerfish, Balistes capriscus (Balistidae), co-occur with both lobster species throughout their respective geographic ranges in the Caribbean. Balistids are generalist predators of hard-shelled invertebrates (mollusks, crustaceans, echinoderms, etc.) and use a variety of behaviors to capture, access, and process different types of prey (Wainwright & Friel, 2000). Triggerfish are known predators of spiny lobsters in the field (Kanciruk, 1980; Barshaw et al., 2003; Herrnkind; unpublished data) and the laboratory (Lozano-Alvarez & Spanier, 1997; Herrnkind et al., 2001). Gray triggerfish are strictly diurnal and form social hierarchies with conspecifics that are maintained by posture, color patterns, and perhaps sound production (Cleveland, 2002). B. capriscus can produce sound, like many other balistids, by fluttering the pectoral fins against an exterior membrane just posterior of the gill opening (Salmon et al., 1968). Because triggerfish produce sound during interactions with conspecifics and their sounds occupy much of the same frequency spectrum as stridulation (Hazlett & Winn, 1962b, Salmon et al., 1968; Mulligan & Fischer, 1977), it is inferred that the fish can hear the lobsters’ auditory signal. Previous work on encounters between P. argus and balistids (Kanciruk, 1980; Herrnkind et al., 2001; Parsons, 2005) indicates that P. argus stand their ground when attacked on open terrain during the day, retaliating aggressively with the large, spinose 2nd antennae to repel attacks by triggerfish. Both P. argus and P. guttatus also possess a rapid escape mechanism, the tailflip, in which the lobster swims rapidly backward, propelled by the tailfan and powered by the strong, but quickly fatigued, muscles of the abdomen (Cobb & Wang, 1985). Although effective for rapid escape into nearby shelter, tailflips are of limited duration and distance (Spanier et al., 1991). When handled by humans, lobsters of both species are known to stridulate usually, but not always, coincident with tailflips.

Experimental protocol Nomadic phase Panulirus argus (35-65 mm CL) and comparatively sized P. guttatus were caught by hand near the Keys Marine Laboratory (KML), Long Key, FL,

13 USA. Both species were held in flow-through seawater tanks and periodically (every 3 days) fed frozen squid and shrimp to satiation. All lobsters in this study were determined to be in the intermolt stage by physical inspection of the carapace. Individuals of either species were used only once in these experiments. Gray triggerfish Balistes capriscus (22-30 cm standard length, N=8) were captured from the Northern Gulf of Mexico (GOM), near the Florida State University Marine Laboratory and in the Florida Keys (Southern GOM) (FFWCC Permit #03SR-053 for fish and lobsters). Fish were transported to the Keys Marine Laboratory and held individually in large, flow-through seawater enclosures (protocol #0105, approved by the Florida State University Animal Care and Use Committee). Triggerfish were fed to satiation every 2-3 days with frozen squid or shrimp, and they also foraged on naturally occurring small fish and snails in their enclosures when not being used in experiments. Lobsters of both types and the triggerfish acclimated quickly to captivity (within days), feeding normally and exhibiting the full range of social behaviors (Herrnkind et al., 2001; Cleveland, 2002). Staged encounters between lobsters and triggerfish were carried out from September - December, 2003 in a 1 m deep, 8 m X 12 m flow-through seawater enclosure at the KML. Lobsters of both species were tethered to an anchored plastic rod with fishing swivels, attached ventrally to a plastic-coated wire wound around the posterior carapace, between the 3rd and 4th walking legs. Tethering in this way allowed lobsters a full range of motion of the antennae (the primary defensive weapons) and did not hinder pirouetting (rotating the body to face the triggerfish) but kept individuals from using arena walls for shelter. Previous experiments showed that tethered Panulirus argus use the same behavioral repertoire and frequency of defending actions against triggerfish as exhibited by free-moving lobsters (Parsons, 2005). Preliminary experiments with similarly tethered P. guttatus and triggerfish indicated that this species also behaved comparably to free-moving conspecifics. Lobsters were given a 30-minute acclimation period in the arena before triggerfish were introduced. Encounters lasted until the fish attacked and successfully disabled a lobster. Unresponsive fish were removed from the arena after 30 minutes to be used in trials on subsequent days, until they were motivated to persist in an attack. Consequently, time since last feeding (hunger state) for triggerfish varied somewhat

14 between trials but did not differ significantly between trials of each lobster species (t = - 1.63, P = .115). Attacked lobsters were considered disabled if one eye was bitten off, 3 legs were removed from one side, or the fish bit into the basal joint of the antennae. Lobsters injured as such are unable to effectively defend vulnerable areas of the body on one or both sides from triggerfish. After disabling a lobster, the fish was allowed to quickly kill (usually less than 5 minutes) and consume it. Past experiments (Herrnkind et al., 2001; Parsons, 2005; PEB, pers. obs.) indicated that a triggerfish's motivation to subsequently attack lobsters depends on being allowed to feed to satiation on its prey after expending considerable effort to disable it. Intraspecific comparisons of defensive actions were conducted with lobsters size- matched (within 2.5 mm CL) with congeners, which resulted in nearly-identical size distributions of lobsters facing triggerfish for P. argus and P. guttatus (N=15 for P. argus, N=14 for P. guttatus). Size-matched pairings of individuals mostly faced the same triggerfish individual (12 of 14 matchings) in consecutive trials, but on different days. Individuals of P. guttatus generally faced a triggerfish predator before a size- matched P. argus faced the triggerfish a few days later. However, all triggerfish had attacked a similarly-sized P. argus in other experiments before making an attack on a P. guttatus in this study. As P. guttatus defensive behavior encompasses a subset of the same defensive actions exhibited by P. argus, we did not consider it necessary to sacrifice additional P. guttatus to assure experience by triggerfish with its lesser repertoire of defensive actions. We also compared the occurrence of stridulation between the two species using 13 P. argus and 12 P. guttatus. Triggerfish predators were used multiple times in this experiment. Five different fish were used against P. argus, with those 5 fish plus one additional fish against P. guttatus. Results from previous studies (Herrnkind et al., 2001; Parsons, 2005; Bouwma, Chapter 3) show that P. argus lobsters display all defensive actions (lunges, whips, pirouettes, tailflips) during attacks by all triggerfish individuals. Other experiments (Bouwma, Chapter 3) also suggest that the experience level of fish does not influence how long it takes a triggerfish to subdue a lobster. However, to assure that multiple use of predators did not bias the data comparing the lobster species, we compared tailflips, whips, pirouettes, and time to subdue an individual lobster between species by two-way

15 ANOVA with lobster species and triggerfish individual as factors. All triggerfish in these experiments faced lobsters at least twice. Triggerfish individuals faced P. guttatus and P. argus equally during both the first two trials and in subsequent trials (Pearsons χ2 = 0.58, P= 0.445). We examined the potential for differences in attack efficiency (potentially influencing time to subdue the prey) in early trials and later trials by comparing time-to-subdue during the first two trials with that in subsequent trials by t- tests, regardless of lobster type. Each trial was recorded with a digital video camera (Sony DCRTRV9) with an attached hydrophone (HTI 96min). Tapes were reviewed later to determine lobster antipredator effort (defensive and escape actions), triggerfish attack behaviors (Table 1.1), and the occurrence of lobster stridulation, both during attacks and while a fish was not actively engaged with the lobster. The morphology of defensive weaponry (2nd antennae) relative to body size was also compared between the two lobster species over a shared size range (20-65 mm CL, P. argus N = 29, P. guttatus N = 35). Length and width measurements were taken using a digital caliper at the base of the antennal flagella and of the last segment of the antennal peduncle. Cross-sectional area of the flagella and peduncle were then determined by modeling both as an oval (π*L/2*W/2). The width of the antennae 10 cm from the base was also measured. Antennal size measurements were taken on both antennae and averaged when possible. In order to determine if other morphological features relevant to defense differed between species, measurements of the anterior carapace width (near the insertion of the 2nd antennae), width of the 3rd walking leg (1st and 2nd walking legs not measured as these are enlarged in adult male P. guttatus), and the width of the abdomen were also taken. The three antennal size and other morphological measurements were all incorporated into separate multiple regression models, with carapace length and an indicator variable (for species identity) as predictors of each measurement. In these models, significance (at P <0.05) of the species variable would indicate a different intercept for P. argus and P. guttatus, while significance of the interaction between species and CL would indicate separate slopes for each species.

16 RESULTS

Brief, intense bursts of sound by both lobster species were observed during certain defensive actions and in response to bites by triggerfish. Lobsters stridulated only while the predator was attacking (within 1 antenna-length), but not upon initial detection of the fish (lobster pointed antennae), as it approached (beyond 2 antenna- lengths), or other times when the fish was not attacking. On average, P. argus (N = 13) and P. guttatus (N = 12) stridulated 9.84 and 9.18 times, respectively, per min of encounter, with 80.2% and 79.9% of stridulations associated with defensive actions, 19.3% and 15.9% with triggerfish bites, and the few other incidents with no particular event. Defensive actions most commonly associated with stridulation were lunges (only for P. argus), tailflips, and whips (Fig. 1.1a). Bite attempts (with or without injury) on the antennae, followed by release, and bites followed by pulling on the antennae (those which did not elicit tailflips) also often elicited stridulation (Fig. 1.1b). Whips were the most common defensive behavior in P. argus and P. guttatus (51.8% and 40.9% of total defensive actions respectively), yet only a small percentage were accompanied by stridulation. The likelihood of producing sound with whips did not differ significantly between species (t = 0.55, P = 0.588). The most striking difference between P. argus and P. guttatus defensive behavior and stridulation was lunging – only the former exhibited this action. In P. argus, the initial action observed as a triggerfish approached to within 1 antenna length was a lunge with stridulation. All P. argus lobsters lunged within 1 min of initiation of attack by triggerfish. Lunging is a rapid, highly aggressive, often pre-emptive, defensive action in which the lobster swings both antennae forward from a spread position while driving the entire body ahead, sometimes greater than one body length (10-20 cm), with rapid up- down flexes (carangiform kicks) of the tailfan (discernible only in slow motion). During lunges, the antennae were typically directed towards the fish's eyes and forebody. Although we detected no injuries to the fish, triggerfish typically responded to lunges by aborting and turning upward or pausing an attack, swimming beyond lunging range before resuming an attack, or sometimes ceasing further attack. P. argus stridulated during 86.6% of lunges, indicating that sound is not caused by the lunging action, but is

17 produced in addition to it. Lunges by P. guttatus were never observed in encounters with triggerfish. Panulirus argus and P. guttatus stridulated similarly in conjunction with 80.9% and 69.3% of tailflips, respectively (t = 1.28, P = .210). Tailflips occurred throughout encounters for both lobster species, although P. guttatus exhibited tailflips more in the first min of an encounter (13.2) than P. argus (7.6). Except for one small (41.0 mm CL) P. argus which tailflipped 37 times (most for either species) in the first minute, this difference would have been statistically significant. Both species stridulated when grasped by the antennae and either pulled forward or released by the fish. P. argus stridulated significantly more frequently when bitten and released (t = 2.31, P = 0.031) than did P. guttatus. Triggerfish subdued P. guttatus individuals three times more quickly than P. argus (5.5 vs. 16.2 min; Table 1.2). In general, P. argus was more aggressive in defense than P. guttatus, often moving forward on the tether to whip or lunge at an attacking fish. Defensive behavior was similar to that reported for Palinurus elephas against gray triggerfish Balistes carolinensis (Barshaw et al., 2003). All P. guttatus individuals consistently retreated to the end of the tether while fending the fish with the antennae or while tailflipping. However, with the exception of lunging, all other defensive actions were observed in both species and none differed in frequency (per minute of encounter) between species (Table 1.2). None of the observed lobster defensive behaviors (whips, tailflips, pirouettes) or the time to subdue differed significantly due to the triggerfish individual making the attack (Table 1.2). Additionally, the average time to subdue lobsters (regardless of species) during the first two encounters by a fish versus was equivalent to that for later trials (10.5 and 11.7 min respectively; t = 0.41, P = 0.689). This indicates that the fish did not become more efficient at subduing lobsters as the experiment progressed. Morphological measurements of the defensive weaponry (cross-sectional area of the peduncle, base of the antennal flagella, and width of the flagella 10 cm from the base) all varied differently with body size in P. argus than P. guttatus (Fig. 1.3a, b, c). The multiple regression models for all three antennal size measurements included different slopes and intercepts for the two species (Table 1.3). Although small individuals of both

18 species had similar antennal sizes, antennal cross-section increased in proportion to body size at a higher rate in P. argus, meaning that larger P. argus had visibly more robust antennae (see Fig. 1.2) from the base to 10 cm forward than did comparably-sized P. guttatus (Fig. 1.3a, b, c). Larger cross-sectional area provides comparatively more area for muscle attachment, while larger-circumference oval tubes theoretically should be more resistant to bending (PEB, pers. obs.; Alexander, 1983; Patek & Oakley, 2003). We also did not observe any qualitative differences in exoskeleton thickness between P. argus and P. guttatus antennae. Although it is immediately obvious while handling large individuals of each species that P. argus antennae are stiffer than those of same sized P. guttatus, I did not measure force resistance differences. However, I did observe that P. argus was more successful at fending away the triggerfish than P. guttatus. Anterior carapace width, the width of the 3rd walking leg, and abdomen width did not vary differently with body size between species (Fig. 1.3d, e, f). Although each of these correlated significantly with body size (CL), all three multiple regression models included the same intercept and slope for P. argus and P. guttatus (Table 1.4). This agrees with prior observations that, other than the antennae, the general body morphology (posterior of the antennae) in P. argus and P. guttatus is very similar between equivalently-sized individuals.

DISCUSSION

Stridulation and defensive behavior These results show that at least two spiny lobster species produce sound in conjunction with defense against a common, ubiquitous predator. In Panulirus argus, stridulation appears to operate in two distinct contexts, the first of which is aggressive, preemptive, or retaliatory. At the beginning of most encounters with gray triggerfish, P. argus repeatedly combined short, intense bursts of sound with lunging behavior. In other experiments with unrestrained P. argus, I observed individuals on multiple occasions to pursue a retreating triggerfish, lunging repeatedly at them. Lunging is not unique to P. argus, as at least one other species of spiny lobster, Palinurus elephas, has been documented to lunge at triggerfish predators in the field (Barshaw et al., 2003).

19 However, it is not clear whether P. elephas, a sound producing species, also combines stridulation with lunges or if the antennae are comparably robust to P. argus. Some other species of spiny lobster probably lunge at predators, but there are currently few studies of palinurid anti-predatory defensive behavior on which to comment. The triggerfish in these experiments persisted in attacking solitary P. argus and were usually successful despite vigorous lobster defense. High attack success reflected previous lobster-killing experience of our captive gray triggerfish. The fish were caught in the northern Gulf of Mexico where lobsters are rare. Upon transport to the laboratory, most fish had to be food deprived and required repeated encounters with P. argus individuals before a successful attack was made. Following a number of successful attacks and consumption of the prey, a triggerfish was more likely to persist attacking in subsequent encounters (PEB, pers. obs.). In a separate study, single lobsters tethered in the field in the northern Gulf of Mexico were more likely to survive, or survive longer, during attacks by triggerfish (WFH, unpublished data). Therefore, the aggressive defense seen in P. argus may be more effective against less-experienced triggerfish, as might occur across the range of varying natural situations. The 16.9 min (mean, 26.1 min maximum) handling time required to subdue a P. argus might lead a generalist predator like a triggerfish to give up and search for more easily accessible prey. Further work is needed to determine whether the aggressive defense and associated sound production initially displayed by P. argus is most effective at deterring relatively inexperienced triggerfish. In many areas, lobsters probably encounter triggerfish with varying experience with lobsters. In such circumstances, aggressive defense plus sound production may "buy" the lobster time to find a suitable shelter when confronted by a naïve attacker, as suggested by Barshaw et al. (2003). In other experiments, we have observed unfettered lobsters moving about the large seawater enclosure during attacks by the triggerfish, presumably searching for shelter (P. Bouwma, pers. obs.). During mass migrations, when nomadic-phase and adult lobsters cross long, shelterless stretches of substrate, migrants travel in queues up to 60 members strong. However, 2% of individuals become separated from queues, moving solitarily until chancing upon cohorts (Herrnkind et al., 2001). Groups of 5 or more lobsters defend more effectively against attacking triggerfish than

20 solitary individuals (Herrnkind et al., 2001). Yet individuals caught in the open may be able to deter attack by strong individual defense long enough to rejoin a group. Stridulation also appears to play a role in escape for both P. argus and P. guttatus. The primary fleeing action for spiny lobsters is the tailflip, which is seen in all decapod crustaceans with the macruran body type (Cobb & Wang, 1985; Spanier et al., 1991). Both species stridulated during the great majority of tailflips, indicating that a combination of the tailflip plus stridulation is typically involved in escape. Lobsters of both species also stridulated (to a lesser extent), without an associated tailflip, when the antennae were grasped by a triggerfish. Because the stridulatory organ is attached to the antennae, the auditory and vibrational components of stridulation may act to startle or irritate predators that grasp the antennae during encounters. To bite off an eye, triggerfish usually first shortened the antennae of either species considerably by serially biting chunks away. Shortened antennae decreased the area of the body the lobster could effectively defend, which usually resulted in the eventual debilitating loss of an eye. By stridulating when the antennae are grasped, lobsters may increase the longevity of these crucially important defensive weapons in encounters with triggerfish. Lobsters, particularly P. argus, tend to reside in crevices with the antennae exposed. If a predator like a triggerfish were to grasp these exposed antennae to pull a lobster out of a den, tailflips plus stridulation may allow lobsters to escape deeper into the den without injury or resorting to autotomy of an antenna. We casually note that lobsters held in captivity without predators grow much longer antennae after a molt than is typically found among wild lobsters of the same size. Barshaw et al. (2003) noted that slipper lobsters (Family Scyllaridae) occasionally caused triggerfish to pause in an attack when they exploded off the substrate in bouts of tailflipping and suggested that such explosive, unpredictable flight (protean behavior) may startle the fish. The addition of loud stridulation by spiny lobsters to this sudden, unexpected flight might enhance a startle effect, allowing lobsters to swim rapidly into shelter. However, in areas where shelter is not available, tailflipping may be of limited utility against the very agile triggerfish, since lobsters are not likely to escape the visual range of the fish during the daytime in clear waters. The sudden burst of sound and rapid escape might be more effective against nocturnal predators such as octopus or

21 nurse sharks. At night, any startle response or hesitation by a predator probably allows lobsters to quickly move out of its sensory range. Because all spiny lobsters are primarily active nocturnally, a startling role for loud stridulation is a promising possibility.

Function of stridulation To understand the evolution of sound production in the Palinuridae, it is vitally important to determine the functional significance of this behavior. This study is an initial step, establishing that two sympatric, yet behaviorally distinct, species both produce similar sounds during defense against a common predator. Previous authors have had to speculate about the function of stridulation without knowing the context of sound production exhibited against natural predators. I can now provide some insight on potential functional hypotheses. Based on work done primarily in terrestrial systems, there are six main, non- mutually exclusive hypotheses for the function of antipredator sound: Calls may (1) elicit help from conspecifics, (2) alert conspecifics of danger, (3) deter pursuit by a predator (e.g. alerts the predator that the prey is aware of its presence), (4) attract another predator to attack or interfere, (5) startle the predator (pause in attack or loosen its grip on captured prey), or (6) be aposematic displays paired with chemical or mechanical defense to deter future attacks (Sherman, 1977; Curio, 1978; Masters, 1979; Högstedt, 1983; Klump & Shalter, 1984; Caro, 1995; Randall, 2001). One of the most commonly suggested functions for stridulation is intraspecific communication (hypotheses 1 and 2: Lindberg, 1955; Berry, 1971; Smale, 1975; Berrill, 1976; Atema & Cobb, 1980; Kanciruk, 1980; Meyer-Rochow et al., 1982), although these hypotheses are not well supported by available evidence. In most situations where animals call to conspecifics for help or warning, there is usually some degree of relatedness between callers and receivers (kin selection) or a long-term association between them (reciprocity) to explain potentially altruistic behavior by either party (e.g. Sherman, 1977; Atema & Cobb, 1980; Randall, 2001). Although spiny lobsters are gregarious and shelter together, individuals are not likely to be related. Spiny lobsters have an exceedingly long planktonic dispersing larval period (est. 6-9 months; Phillips &

22 Sastry, 1980; Phillips et al., 2006), and, subsequently, juveniles are individually highly nomadic thus obviating long-term associations (Herrnkind et al., 2001). Lobsters may be able to detect waterborne sound from a calling lobster (PEB, unpublished data), yet hearing thresholds in other crustaceans (shrimp) are relatively high (Lovell et al., 2005), indicating that calls would need to be produced close by and loudly, limiting any communicatory function. Our data also fail to support that stridulation is a warning or call for help. These calls are usually produced as soon as a predator is detected (Klump & Shalter, 1984), yet both P. argus and P. guttatus were silent upon detection of the fish, turning to face it, and pointing the antennae towards it. This evidence also does not support the pursuit-deterrent hypothesis (3), which requires that a prey signal an approaching predator before an attack begins (Caro, 1995), which these lobsters clearly do not. Our data suggest that sound is directed either at an attacking predator to startle it (5) or as an aposematic display (6), or perhaps to attract an additional predator to interfere (4). As mentioned above, stridulation along with tailflipping as a startle mechanism seems a promising function, as does predator attraction. Triggerfish produce sounds in interspecific communication within the same auditory range as stridulation (Salmon et al., 1968), suggesting that they might be attracted to an actively stridulating lobster. Gray triggerfish form dominance hierarchies and often fight over food (Cleveland, 2002). Additionally, Berry (1971) noted that moray eels seemed to be attracted to lobster stridulation. This raises the possibility that other piscine predators may be attracted to stridulation and interfere long enough with each other to allow the lobster time to escape. Another possibility is the use of stridulation as part of an aposematic display. This is a common feature of other arthropods that produce defensive sounds, usually warning predators of unpalatability or noxiousness (Alexander, 1967; Marshall et al., 1995; Rowe & Guilford, 1999). Although lobsters are palatable, the aggressive thrust of spines at an attacking fish may make lobsters sufficiently unprofitable to discourage further attacks.

Interspecific differences and evolution Panulirus argus, for which heightened daytime predation risk is a feature of nomadic juvenile and adult stages, survived much longer in encounters with triggerfish

23 than did Panulirus guttatus. The defensive repertoire of P. argus consisted of aggressive defense, followed by escape attempts as compared to the retreat-only defense seen in P. guttatus. The antennae of P. guttatus were also smaller in cross-sectional area than its congener for comparably-sized individuals. Although it is possible that there were differences in exoskeleton thickness that made P. guttatus antennae as strong as P. argus, our observations of the flexibility and effectiveness of the antennae in fending triggerfish away, suggests that P. guttatus antennae bent with much less force. This suggests that thrusting and parrying as a deterrent may not be an important function of the antennae in this species. It was remarkable that other morphological features of the body presumably important for aggressive defense (lunging) were nearly identical between P. argus and P. guttatus, suggesting that robust antennal weaponry may be the only additional morphological tool necessary for lunging as opposed to larger (and presumably stronger) legs, carapace, or abdomen. The dimensions of P. argus antennae were similar to those of P. guttatus at sizes below ~40 mm CL. This is approximately the size at which P. argus makes an ontogenetic shift to the nomadic juvenile stage, begins to participate in long-distance migrations, and begins to experience heightened diurnal predation risk. Because Panulirus guttatus is rarely observed away from shelter during the day and likely never faces diurnal predation risk for which visually-directed antennal defense is appropriate at any body size, these lobsters may not benefit from increased antennal size with increasing body size. Additionally, less robust, more flexible antennae may be more useful for lobsters such as P. guttatus that reside deep in crevices, thus easing passage by the long antennae through the interstices of the reef. Morphological and genetic phylogenetic studies indicate that P. argus and P. guttatus may have evolved from an "P. argus-like" ancestor (George & Main, 1967; Ptacek et al., 2001; Patek & Oakley 2003). This suggests that the ancestor to the Panulirus genus may have incorporated stridulating into both aggressive defense and escape. However, the defensive context in which the stridulating organ appeared in ancestral palinurids, ~190-200 My before the Panulirus species diverged from an Indo- Pacific ancestor, is not clear. Several species in the Palinuridae, both stridentes and silentes, are known to defend aggressively. Palinurus elephas (stridentes) lunges at

24 attacking triggerfish, although it is not clear whether sound was produced during these actions (Barshaw et al., 2003). In the laboratory, Palinurus delagoae, forms defensive aggregations in the open similar to those seen in Panulirus argus during active group defense against triggerfish (Berry, 1971). Kelly et al. (1999) reported that New Zealand spiny lobster Jasus edwardsii (silentes) in aggregations on open substrate were very aggressive, approaching divers and using the antennae and anterior walking legs as weaponry. Although few other quantitative studies document palinurid defensive behavior, the size/strength of spiny lobster defensive weaponry tends to correspond similarly to body size in many palinurids (Patek & Oakley, 2003). Our finding that antennal size is linked to the nature of defensive behavior, suggests that retaliatory defense may be a common feature of many palinurid taxa. On the other hand, tailflipping escape is a feature of all palinurid lobsters, inherited from an ancestor to all lobster types with the macruran body plan (Patek et al., 2006), and stridulation may have evolved initially to complement this existing escape mechanism. This indicates that more work on spiny lobster defensive behavior will be needed before we can infer the defensive context in which stridulation initially evolved.

Table 1.1. (a) Descriptions of lobster defensive actions and (b) triggerfish actions recorded during attacks by gray triggerfish Balistes capriscus on Panulirus argus and P. guttatus individuals.

Action Description (a) Lobster Lunge Aggressive defense where lobster swings the antennae forward from a spread position while simultaneously driving the body forward using a kick of the tail (P. argus only). Tailflip Escape behavior where the lobster retreats rapidly by swimming, propelled by strokes of the tailfan and muscles of the abdomen. Whip Lobster strikes the fish with either one or both antennae without an associated forward movement of body or simply pushes the fish away using the antennae. Point Lobster directs the antennae at a fish located greater than 1 antenna length away without making contact. Pirouette Lobster turns in place while keeping the antennae and anterior of body directed at attacking fish. Rear Back Lobster leans backwards, tucking the abdomen under cephalothorax while directing the antennae over the carapace and abdomen, without facing the fish.

(b) Triggerfish Bite Attempt Fish attempts to bite the antennae, legs, abdomen or eyes with or without making contact with lobster.

Circles Fish swims in a circle around a lobster within 2 antenna lengths.

26 Table 1.2. (a) Mean lobster defensive actions per minute of encounter (± SE) for P. argus and P. guttatus and (b) results of two-way ANOVAs comparing means for time to subdue and lobster defensive actions (tailflips, whips, pirouettes) with fish individual and species type as factors (plus the interaction of fish individual X species type). N = 15 for P. argus and 14 for P. guttatus for all ANOVAs. Number of different fish: N = 6 for both species. (a)

Actions P. argus P. guttatus

Lobster Behaviors Lunges 1.4 (± 0.2) 0.0 Tailflip 5.4 (± 2.1) 10.0 (± 1.8) Whip 11.9 (± 1.2) 10.1 (± 1.1) Point 0.2 (± 0.1) 2.1 (± 0.4) Pirouette 2.9 (± 0.4) 2.7 (± 2.5) Rear back 0.1 (± 0.1) 0.3 (± 0.1)

(b)

Time To Subdue DF SS F P

Fish Individual 5 335.324 2.169 0.106 Lobster Species 1 440.370 14.243 0.002 Species X Fish Ind. 5 117.146 0.758 0.592

Tailflips DF SS F P

Fish Individual 5 50.475 0.524 0.7549 Lobster Species 1 1.674 0.0869 0.7718 Species X Fish Ind. 5 112.208 1.1648 0.366

Whips DF SS F P

Fish Individual 5 13.352 0.056 0.998 Lobster Species 5 583.470 2.457 0.075 Species X Fish Ind. 1 59.181 1.246 0.280

Pirouettes DF SS F P

Fish Individual 5 2.936 0.144 0.979 Lobster Species 5 42.853 2.098 0.116 Species X Fish Ind. 1 2.782 0.681 0.421

27 Table 1.3. Measurements of lobster antennal weaponry as predicted by body size (Carapace Length) and species (Panulirus argus or P. guttatus) in multiple linear regression models. Separate models were fit for (a) the width of the antenna flagellum 10 cm from the base (Wflag10), (b) antenna flagellum cross-sectional area near base (XSflag), and (c) antenna base cross-sectional area (XSbase). DF = degrees of freedom, Type III SS = type three sums of squares, F = F statistic.

2 (a) Wflag10: N = 62, adjusted R = 0.93, F = 267.42, P = <.0001 Source DF Type III SS F P CL 1 35352.812 769.97 <.0001 Species 1 2751.709 59.93 <.0001 CL x Species 1 8062.037 175.59 <.0001 Parameter Estimate Standard Error t-value P Intercept -2.344 0.236 -9.95 <.0001 CL 0.099 0.005 20.62 <.0001 P. guttatus 2.009 0.297 6.76 <.0001 P. argus 0 . . . CL x P. guttatus -0.061 0.006 -9.78 <.0001 CL x P. argus 0 . . . 2 (b) XSflag: N = 64, adjusted R = 0.96, F = 566.64, P = <.0001 Source DF Type III SS F P CL 1 2209.546 769.90 <.0001 Species 1 172.021 59.94 <.0001 CL x Species 1 503.954 175.60 <.0001 Parameter Estimate Standard Error t-value P Intercept -4.121 1.162 -3.54 0.0008 CL 0.296 0.025 11.44 <.0001 P. guttatus -15.061 1.945 -7.74 <.0001 P. argus 0 . . . CL x P. guttatus 0.542 0.041 13.25 <.0001 CL x P. argus 0 . . . 2 (c) XSbase: N = 64, adjusted R = 0.98, F = 867.59, P = <.0001 Source DF Type III SS F P CL 1 16143.961 1494.86 <.0001 Species 1 580.774 53.78 <.0001 CL x Species 1 1831.773 169.61 <.0001 Parameter Estimate Standard Error t-value P Intercept -11.744 2.255 -5.21 <.0001 CL 1.018 0.050 20.24 <.0001 P. guttatus -27.674 3.773 -7.33 <.0001 P. argus 0 . . . CL x P. guttatus 1.035 0.079 13.02 <.0001 CL x P. argus 0 . . .

28 Table 1.4. Measurements of lobster body size as predicted by carapace length and species (Panulirus argus or P. guttatus) in multiple linear regression models. Separate models were fit for (a) the width of the anterior carapace (Wac), (b) width of the abdomen (Wab), rd and (c) width of the 3 walking leg (W3wl). DF = degrees of freedom, Type III SS = type three sums of squares, F = F statistic.

2 a. Wac: N = 62, adjusted R = 0.98, F = 942.28, P = <.0001 Source DF Type III SS F P CL 1 3238.332 2515.56 <.0001 Species 1 0.089 0.07 0.7939 CL x Species 1 1.148 0.89 0.3488 Parameter Estimate Standard Error t-value P Intercept 0.711 1.055 0.67 0.5027 CL 0.703 0.021 32.95 <.0001 P. guttatus 0.344 1.311 0.26 0.7939 P. argus 0 . . . CL x P. guttatus -0.026 0.028 -0.94 0.3488 CL x P. argus 0 . . . 2 b. Wab: N = 61, adjusted R = 0.96, F = 511.18, P = <.0001 Source DF Type III SS F P CL 1 2159.869 1347.02 <.0001 Species 1 0.229 0.14 0.7068 CL x Species 1 0.906 0.56 0.4554 Parameter Estimate Standard Error t-value P Intercept 2.402 1.220 1.97 0.0538 CL 0.585 0.024 23.90 <.0001 P. guttatus 0.566 1.498 0.38 0.7068 P. argus 0 . . . CL x P. guttatus -0.023 0.031 -0.75 0.4554 CL x P. argus 0 . . . 2 c. W3wl: N = 59, adjusted R = 0.92, F = 229.30, P = <.0001 Source DF Type III SS F P CL 1 69.654 639.97 <.0001 Species 1 0.001 0.01 0.9301 CL x Species 1 0.015 0.14 0.7082 Parameter Estimate Standard Error t-value P Intercept -0.658 0.318 -2.07 0.0434 CL 0.102 0.006 16.06 <.0001 P. guttatus 0.035 0.395 0.09 0.9301 P. argus 0 . . . CL x P. guttatus 0.003 0.008 0.38 0.7082 CL x P. argus 0 . . .

29 (a) P. argus 14.0 P. guttatus 11.5 10.4

12.0 10.0

10.0 er Minute p 8.0 5.8

6.0

3.1 4.0 2.7

Defensive Actions Defensive Actions 2.0 1.3 0.1 0.2 0.0 0.0 LungesLunges Tailflips Tailflips Whips Whips Pirouettes Pirouettes Rear Rear Back (b)

100.0

90.0 80.9 76.4

80.0 69.4 70.0

60.0

50.0

40.0

30.0

20.0 11.4 7.1

Percent of Actions with stridulation 10.0 0.0 0.0 0.0 0.1 0.0 0.0 LungesLunge Tailflips Tailflip Whips Whip Pirouettes Pirouette Rear Rear Back

30 Panulirus argus Panulirus guttatus

Stridulating organ 2nd Antennae

Figure 1.2. Anterior comparison photograph of P. argus (right) and P. guttatus (left) with arrows indicating the stridulating organ and the 2nd antennae.

31 (a) P. argus (d) P. guttatus 4.50 50.00 4.00 45.00 3.50

40.00 ) 3.00

mm 2.50 35.00 (

2.00 30.00 1.50

Width 25.00 1.00 0.50 20.00

0.00 15.00 25.00 35.00 45.00 55.00 65.00 25.00 35.00 45.00 55.00 65.00 (b) (e) 40.00 45.00

35.00 40.00 30.00 35.00 25.00

20.00 30.00

15.00 25.00

)

2 10.00 20.00

mm 5.00 (

0.00 15.00 25.00 35.00 45.00 55.00 65.00 25.00 35.00 45.00 55.00 65.00 (c) (f) 100.00 7.00 90.00 6.00 80.00 70.00 Cross sectional area Cross sectional 5.00 60.00 50.00 4.00 40.00 3.00 30.00

20.00 2.00 10.00 0.00 1.00 25.00 35.00 45.00 55.00 65.00 25.00 35.00 45.00 55.00 65.00

Carapace Length (mm)

32 CHAPTER 2

APOSEMATISM BY WEAPONRY AND SOUND? RETALIATORY DEFENSIVE BEHAVIOR IN CARIBBEAN SPINY LOBSTER PANULIRUS ARGUS

INTRODUCTION

Various taxa exhibit conspicuous, aposematic signals against predators, including bright, contrasty, visual displays; noxious odors; acoustic signals; or some combination thereof in multimodal displays (Wallace, 1867; Poulton, 1890; Rothschild, 1961; Edmunds, 1974; Gittleman & Harvey, 1980; Endler, 1988; Dunning & Kruger, 1995; Mallet & Joran, 1999; Rowe & Guilford, 1999; Lindström et al., 2001; Exnerova et al., 2003; Summers & Clough, 2001; Mappes et al., 2005). In these studies, for a species to be aposematic, 1) it must be well-defended or unpalatable such that it is unprofitable to predators (or mimic another species that is), 2) it must advertise its defenses (real or otherwise) through conspicuousness, and 3) it or other similarly conspicuous individuals must benefit from predator avoidance after an encounter with a defended prey. In most well described examples of aposematism and mimicry, visually conspicuous signals are paired with unpalatability or other adverse chemical defenses (Mallet & Joran, 1999; Summers & Clough, 2001; Ruxton et al., 2004; Mappes, et al, 2005). However, unprofitability may also be mediated through spines or other mechanical defenses, which potentially endanger the predator or otherwise increase the handling time required to subdue the prey (Marshall et al., 1995; Montealegre-Z, 2004; Inbar & Lev-Yadun, 2005). Although mechanical unpalatability is often combined with visually conspicuous patterns or colors (Inbar & Lev-Yadun, 2005), conspicuous signals can exploit other sensory modalities, such as acoustics or chemosenses. This suggests that palatable, cryptically-colored, but mechanically-defended prey potentially might advertise their defenses only upon discovery with sounds or chemicals and still benefit from learned predator avoidance. Recently, a theoretical study indicated that mechanical

33 defenses may have been important in the evolution of some warning displays (Speed & Ruxton, 2005); however, aposematism based on mechanical defenses by cryptic, palatable organisms has received little attention (Inbar & Lev-Yadun, 2005). Although numerous plants and animals exhibit defensive spines, which often contrast against background body coloration as a visual display (Lev-Yadun, 2003a; 2003b; Inbar & Lev-Yadun, 2005; Speed & Ruxton, 2005), some spiny animals produce additional signals that may be important, additional components in aposematic displays. For example, many insects that defend by visually conspicuous spines (Inbar & Lev- Yadun, 2005) also produce conspicuous sounds, including some wasps and katydids (Montealegre-Z, 2004; Inbar & Lev-Yadun, 2005; Hauglund, 2006). Additionally, many species of spiny fishes produce sounds when captured, prodded, or otherwise disturbed (Myrberg, 1981; Inbar & Lev-Yadun, 2005). Several spiny crustaceans produce sound as well, particularly spiny lobsters (Palinuridae) and some crabs (Moulton, 1957; George & Main, 1967; Popper et al., 2001; Patek & Oakley, 2003). Acoustic defense per se, and particularly an aposematic function for sound in conjunction with mechanical defenses, has yet to be demonstrated in these taxa (Myrberg, 1981; Popper et al., 2001; Childress & Jury, 2006). Therefore, the question remains whether a combination of mechanical, auditory, and/or visual cues is effective in aposematic defense. The Stridentes lineage of spiny lobsters possess a putatively aposematic combination of mechanical and acoustic devices. All palinurid lobsters have numerous, robust, forward-pointing spines covering the anterior carapace, antennal peduncles, and the long flagella of the second antennae (Phillips et al., 1980; Holthuis, 1991). When confronted by predators, lobsters pirouette to face the attack and adopt an alert defensive posture, with the antennae spread and pointed forward, poised high on the walking legs, and with the telson and uropods extended (Fig. 2.1) (Mulligan & Fischer, 1977; Cobb, 1980). In addition to the defensive posture, many spiny lobsters are also brightly colored with contrasty spines (Holthuis, 1991). To the human eye, these color patterns appear cryptic amidst reef habitat; however, they become more conspicuous against less visually-complex backgrounds, such as sandy substrata where a lobster is most vulnerable to attack (Kanciruk 1980). During attacks, using the spiny carapace and antennae, lobsters keep fish from accessing less well-armored portions of the body

34 (Herrnkind et al., 2001; Barshaw et al., 2003; Bouwma, Chapter 1). In addition to these retaliatory tactics, some lobsters also actively pre-empt attacks by fish. Pre-emptive actions in spiny lobsters range from lashing by the antennae to rapid, violent lunging, during which the antennae are whipped at an approaching fish's eyes while the entire body is driven forward using quick forward kicks of the tailfan (Herrnkind et al., 2001; Barshaw et al., 2004; Childress & Jury, 2006). Strident palinurids, including Caribbean spiny lobster Panulirus argus produce a characteristic loud rasping sound when grasped or prodded by humans as well as during attacks by natural predators (George & Main, 1967; Kanciruk, 1980; Bouwma, Chapter 1). This sound, termed stridulation, is produced by a highly specialized organ located bilaterally at the base of the 2nd antennae (Moulton, 1957; Mulligan & Fischer, 1977; Patek, 2001; 2002). The evolution of sound production (~200 MYA) split the Palinuridae into two groups, the sound-producing Stridentes (8 genera, 40 species) and the Silentes (3 genera, 10 species), which lack the stridulating organ (George & Main, 1967). Although lobster sound has been of scientific interest since the mid-19th century, until recently, little experimental data or other compelling evidence from encounters with natural predators existed to demonstrate if, when, or why lobsters use sound in the wild (Moulton, 1957; Childress, & Jury, 2006). Recently, Bouwma & Herrnkind (Chapter 1) found that, in encounters with triggerfish predators (Balistes capriscus), spiny lobster Panulirus argus used stridulation in two different defensive contexts. Sound was paired both with lunges and during tailflip escape attempts. Of these two defensive actions, preemptive or retaliatory defense with lunging is the most likely to be aposematic, because triggerfish do not confront the thrusting spines (and potential injury) during tailflips (PEB, pers. obs.). Although Panulirus argus have many vertebrate and invertebrate predators (Kanciruk, 1980; Smith & Herrnkind, 1992; Cruz & Phillips, 2000; Butler et al., 2006), none are lobster specialists. For example, durophagous triggerfish feed on a variety of hard-shelled prey such as long-spined urchins, sand dollars, some gastropods, a variety of actively-defending spiny and clawed crustaceans, and highly-armored crustaceans such as slipper lobsters (Scyllaridae) (Wainwright & Friel, 2000; Barshaw et al. 2004). I have also observed triggerfish in captivity learn to corner and capture small fish (sheepshead

35 minnows Cyprinodon variagatus; PEB, pers. obs.). Based on descriptions in the literature and my own observations, it is apparent that Balistes capriscus readily investigate and learn to attack and access novel prey. By meeting curious but naïve generalists like triggerfish with aggressive lunges plus aposematic stridulation, lobsters may exploit the predator’s excellent learning ability and "teach" them to avoid lobsters in subsequent encounters. Interestingly, because lobsters are not chemically defended, they are entirely palatable once mechanical defenses are bypassed or overcome. This raises the possibility that, were triggerfish to occasionally encounter poorly defended lobsters (e.g., those in soft post-molt stage, previously injured, etc.), thus leading to a successful attack, the negative association acquired through aposematism might break down. Successful fish might then learn that the nutritional reward from successfully attacking a lobster outweighs the costs, resulting in experienced fish persisting in attacks despite the lobsters' noisy, but now ineffective defensive efforts. Alternatively, if stridulating is not aposematic and simply increases the effectiveness of other defensive actions, like lunging, the acoustic cue should still have some defensive value, even against experienced fish. In this study, I report results from staged encounters of Caribbean spiny lobster Panulirus argus with grey triggerfish Balistes capriscus undertaken to determine: 1) does stridulation in spiny lobsters, coincident with thrusting spines, function along with visual cues in aposematic defense, or 2) does stridulating during defense per se make a lobster more difficult to attack and subdue without playing an aposematic role? We demonstrate, by disabling the stridulating organ in some lobsters (muting), that sound plays a vital role in defense against naïve triggerfish, resulting in fewer successful attacks. However, in choice experiments with triggerfish that have previously bypassed defenses and eaten lobsters, we show that these experienced attackers do not choose muted lobsters over stridulating individuals. I propose that stridulation by P. argus against triggerfish is aposematic when paired with visual defensive cues, as part of a multi-modal display advertising the lobster’s spiny defenses.

36 METHODS

Sub-adult Panulirus argus (35-65 mm CL) individuals were collected by hand near the Keys Marine Laboratory in Long Key, FL, USA (FFWCC Permit #03SR-053 for lobsters and triggerfish). Lobsters were held communally in flow-through seawater tanks and periodically (every 3 days) fed frozen squid and shrimp to satiation. All lobsters in this study were determined to be in the intermolt stage by physical inspection of the carapace. Each lobster was used only once per experiment, and surviving individuals were marked on the tailfan and released away from the capture site. During encounters with triggerfish, lobsters were tethered to an anchored plastic rod with a 15 cm chain of fishing swivels, attached ventrally to a plastic-coated wire wound around the posterior carapace, between the 3rd and 4th walking legs. Tethering in this way allowed lobsters a full range of motion of the antennae (the primary defensive weapons) and to pirouette (rotating the body to face attacking fish) and make other short-range movements but kept individuals from using arena walls for shelter. Previous experiments have shown that tethered Panulirus argus use the same behavioral repertoire and frequency of defending actions against triggerfish as exhibited by free-moving lobsters (Parsons, 2005). Some lobsters were muted in these experiments by removing part of the stridulating organ, which is made up of two parts, a movable plectrum and flap which translates posteriorly along a fixed file on the antennular plate to produce sound. I clipped off the small plectrum and flap entirely, leaving the rest of the stridulating organ intact (Moulton, 1957). Few lobsters responded behaviorally to this manipulation and all individuals rapidly healed (stopped bleeding) at the site of injury within a few hours. Lobsters were muted at least one day before encountering a triggerfish. All lobsters, muted and stridulating, were handled extensively on multiple occasions during capture, sorting, measuring, and preparation for trials. Because removal of plectra closely mimicked these manipulations, particularly measurement and trial preparation (tethering, removal and trimming of other appendages, etc.), additional sham controls were not performed on stridulating animals.

37 Aposematism Trials

We randomly assembled two groups of 15 similar-sized, naïve triggerfish Balistes capriscus captured by (barbless) hook and line at 10 m depth from the northern Gulf of Mexico near Dog Island, Franklin County, Florida. Spiny lobsters are rare in this area and typically are large adults (est. >100 mm carapace length) found further offshore in depths of 20 m or greater (WFH, pers. obs.). Fish were transported in aerated seawater to the Keys Marine Laboratory, Long Key, Florida and held communally in large, flow- through seawater pens (protocol #0105, approved by the Florida State University Animal Care and Use Committee). Fish were individually marked in the dorsal and caudal fins using Visual Implant Elastomer tags (Northwest Marine Technologies) to allow recognition of individuals in the communal holding pens. Both groups of fish were fed to satiation every 2 days with frozen squid or shrimp, and they also foraged on naturally occurring small fish and snails in their enclosures when not being used in experiments. We did not allow fish any contact with lobsters prior to experimentation. Triggerfish acclimated quickly to captivity (within days), feeding normally and exhibiting the full range of social behaviors (Herrnkind et al., 2001; Cleveland, 2002). From May – August, 2004, after at least 1 month in captivity, fish in both groups were given two short-duration staged encounters with tethered lobsters in a large, (1 m deep, 8 m X 12 m) flow-through seawater enclosure at the KML. During these "training" sessions, each triggerfish experienced 5-21 lunges: one group experienced only stridulating lobsters, the other interacted exclusively with muted lobsters. Muted and stridulating lobsters lunged at fish equally during these training sessions (t=2.05, P=.306). Lobsters were then removed from the enclosure before any attacks could escalate. The ratio of triggerfish size to lobster size did not differ between the two groups of fish during training sessions (t=0.24, P=.814). After the second training encounter, each triggerfish was removed from the arena and fed to satiation. Two days later, triggerfish were given 2-hour long opportunities to attack and subdue (disable) a tethered lobster on 5 consecutive days. Lobsters presented during the "attack" trials were ~5 mm CL smaller than those the fish experienced during training sessions. Triggerfish:lobster size ratios during these trials did not differ between groups (t<.001, P>.999). If no successful attack was made, we presented smaller lobsters (~2.5

38 mm CL less) for the 3rd and 4th trial periods and then decreased the lobster size again (~2.5 mm CL less) for the final period. Additionally, triggerfish were not fed during these trials, increasing their hunger state for each consecutive trial. For unsuccessful triggerfish, after 10 total hours of interaction time, the lobster was quickly euthanized to assure that the fish were hungry enough to consume the remains. For successful attacks, lobsters were considered disabled (trial finished) once one eye was bitten off, 3 legs were removed (autotomized when bitten) from one side, or if the fish bit into the basal joint of the antennae. Lobsters injured as such are unable to effectively defend vulnerable areas. Each trial was recorded using a digital video camera (Sony DCRTRV9) with an attached hydrophone (HTI 96min). Lobster survival was compared between groups after the 1st and 5th attack opportunities using G-tests. Additionally, we compared (t-test) the time elapsed in minutes between the beginning of the trial and the first attack for both groups of fish. Tapes were also reviewed to assess lobster antipredator effort (defensive and escape actions) and triggerfish attack behaviors (see Table 2.1). To determine any possible effects of removing the plectrum on lobster antipredator effort, we compared the occurrence and frequency of defensive actions (per minute of encounter), as well as survival times, for muted lobsters to stridulating lobsters. Because too few stridulating lobsters were successfully attacked in this study to allow for sufficient statistical power, we combined data from the 7 stridulating lobsters in the above trials that perished with that from 5 other lobsters that faced 5 different non-naive triggerfish (otherwise same protocol) from a prior experiment (Bouwma, Chapter 1). The combined data were compared to that from the muted lobsters using t-tests. Triggerfish attack behaviors and effort (circles and bite attempts) were also compared in this way.

Triggerfish choice experiment

To determine if stridulating makes a lobster more difficult for experienced triggerfish (successfully attacked lobsters in prior encounters) to subdue, we performed the following choice experiment based on Parsons (2005). She found that triggerfish preferentially attacked a lobster with reduced defenses when given a choice; i.e., one antenna missing as opposed to two intact. We hypothesized that, if noisy defense is more effective than the same actions unaccompanied by stridulation, experienced triggerfish

39 would preferentially attack muted lobsters over stridulating lobsters when given a choice between the two. In this experiment, we tethered (as above) size-matched lobsters (within 2 mm CL) 1.5 m apart in the experimental arena. Because lobsters are commonly caught with missing perieopods and clipped antennae, individuals were also matched, one day prior to use, in pereiopod number (by inducing autotomy) and for the length of the 2nd antennae (flagella trimmed to match). One individual of each pairing was then muted (as above) by removing the plectra. Thirty minutes prior to encountering the triggerfish, lobsters were introduced to the experimental arena and allowed to acclimate. Position in the arena (right or left) for stridulating and silent animals was randomly assigned by coin toss. Additional triggerfish (N=20) were caught (see above for care information) from the northern Gulf of Mexico and the Florida Keys, followed by transportation to KML. After at least 1 month of acclimation, fish were allowed to attack and subdue at least one lobster of each type (stridulating and muted) prior to experimentation. Choice trials were then conducted from May – December, 2003 and during August, 2004. Each fish was introduced into the experimental enclosure and allowed to interact with both tethered lobsters. Only trials where the fish engaged both lobsters sufficiently to induce lunging by both individuals (including audible stridulation in sound-producing animals) were included in this experiment. Additionally, we used only data from trials in which a fish successfully attacked and subdued one of the lobsters. Although we originally intended to determine individual fish choice across multiple encounters, time constraints only allowed this for a subset of triggerfish. Therefore, we considered only the first trial for each fish in our choice analysis (χ2 test). To compare survival times for muted and stridulating lobsters, we averaged the attack time required to disable each type of lobster per fish to include in our analysis (t-test).

RESULTS

40 Aposematism Trials Upon a triggerfish’s approach, lobsters of both types adopted the defensive posture with the antennae spread wide in a vigilant, “ready-to-strike” position (Fig. 2.1). Triggerfish which subsequently moved within striking range (1 antenna length) were initially confronted with lunges by lobsters of both types, during both training and attack periods. Despite active retaliatory defense by lobsters, triggerfish confronting muted lobsters were statistically (Pearson’s χ2 = 6.946, P = 0.008) more likely to attack and subdue (66.7% of fish) a lobster at the first opportunity (Day 1) than fish that had experienced stridulating lobsters (20%). Even by Day 5, increasingly hungry triggerfish were still less likely (Pearson’s χ2 = 5.68 , P = 0.017) to have attacked stridulating lobsters (46.7%) than muted lobsters (86.7%). Triggerfish that experienced stridulating lobsters also took longer to initially approach a lobster within antenna range (attack distance) than fish conditioned with muted lobsters (55.8 vs. 14.9 min; t = 2.18, P = 0.043) after the training sessions. All triggerfish which did not make kills, immediately began to feed on euthanized lobsters following completion of their last trial. In the expanded data set comparing the actions of triggerfish attacking muted or stridulating lobsters, the frequency of attack behaviors (bite attempts and circles) did not differ between the two groups of fish (Fig. 2.2; Table 2.2). Both stridulating and muted individuals also displayed the full array of defensive behaviors during attacks by experienced and inexperienced triggerfish. Comparing all defensive actions revealed no differences in frequency of occurrence for defensive behaviors between the two groups (Fig. 2.3; Table 2.2). Additionally, the time it took triggerfish to disable muted individuals did not differ from stridulating lobsters (13.83 vs. 10.97 min; t = 0.82, P = 0.42) (Fig. 2.2); i.e., both were similarly difficult to subdue.

Triggerfish Choice Experiment

The "experienced" triggerfish (those that had previously subdued a lobster) did not prefer one type of lobster over another; 11 of the 20 fish chose stridulating lobsters while 9 chose muted lobsters (χ2 = 0.2 , P = 0.655). Most fish made several passes (within ~1/2 m) at each lobster before initiating an attack on one or the other. Although some fish attacked both lobsters, no fish subdued both and most ignored the surviving

41 individual entirely once the conquered lobster was consumed. The time required to subdue lobsters in these trials did not differ between muted and stridulating lobsters (12.12 min for muted, 16.26 min for stridulating; t = 0.98, P = 0.34; Fig. 2.2). Interestingly, the time required by experienced fish to subdue lobsters in these trials, once a choice was made, also did not differ from those inexperienced fish that made successful attacks (14.40 min for 20 experienced fish, 10.85 for 20 inexperienced; t = 1.17, P = 0.25; Fig. 2.2; because time-to-subdue durations did not differ between stridulating and muted lobsters, lobsters were grouped in this analysis based on experience only, regardless of the ability to stridulate).

DISCUSSION

Stridulation by Panulirus argus against attacking triggerfish appears to be aposematic. By combining stridulatory sound with visual cues (color patterns, defensive posture) and aggressive lunges (thrusting spines), learning improves by naïve triggerfish to avoid lobster defenses. Although all naïve triggerfish were initially curious and approached within lunging range of tethered lobsters during training sessions, in the subsequent attack periods, fish paired with stridulating lobsters were significantly less likely to persist (and eventually failed) in attacks than those facing muted lobsters. Triggerfish experienced with stridulating lobsters were also more reluctant to initiate attacks on lobsters after training sessions. Because lobsters rarely stridulated when triggerfish were greater than one antenna length away, except during occasional lunges at distant fish, it appears that the visual image of a lobster - not the acoustic cue - is subsequently recognized by triggerfish to avoid spiny defenses. Stridulation only increases the likelihood that triggerfish will learn this association. One study with domestic chicks (Gallus gallus domesticus) (Rowe, 2002) indicated a similar effect of sound. Chicks were given a choice to peck at two different colors of paper, one with food underneath and one without a reward. For half of these chicks, a simple tone (sound) was played when they pecked at unrewarded pieces of paper. Although the paper colors in this experiment were not typically aposematic, Rowe (2002) found that, when sound was played, chicks learned more quickly to discriminate

42 which pieces of paper covered food. This suggests that sound can be used by avian predators to enhance learning of a visual warning signal from prey, without playing any defensive role per se (Rowe, 2002). Our data indicate that sound can also enhance the effect of visual aposematic cues in mechanically defended prey. In P. argus, the visual cue appears to be the image of a lobster in its vigilant defensive posture (Fig. 2.1), although we did not specifically test for any differential effects of P. argus coloration on triggerfish learning. During the course of a trial, even fish that generally avoided the tethered lobster occasionally approached within one antenna length when the lobster was crouched or facing in the opposite direction. Regardless, without stridulation, the visual signal did not facilitate learning by the fish as well as when lobsters produced sound during attacks. A number of studies suggest that, in avian predators, the additional acoustic or odor signals in multimodal aposematic displays primarily exploit innate biases against typically aposematic colors and speed the learning of these visual displays (Schuler & Roper, 1992; Rowe & Guilford, 1999; Jetz et al., 2001; Rowe & Skelhorn, 2005). It is possible that stridulating, along with the lobster’s color patterns or other visual features, exploited hidden biases in the triggerfish against brightly-colored prey. However, in many other species of fish, naïve individuals readily sample unpalatable prey with typical aposematic color patterns such as nudibranchs (Guilford & Cuthill, 1990; Tullrot & Sundberg, 1991; Tullrot, 1994; Penney, 2004; Long & Hay, 2006), suggesting that fish may lack innate biases for warning coloration and avoid defended prey primarily through experience (Tullrot & Sundberg, 1991). Additionally, we observed that many of the triggerfish repeatedly approached and bit at the brightly colored orange and red dive weights with contrasting lettering which anchored the tethering apparatus until the paint was nearly all removed. We have also observed triggerfish approaching and biting at nearly any conspicuous object introduced to their environment, including sticks, hydrophones, dive weights, buoys, and wire ties (PEB, pers. obs.). This suggests that triggerfish are initially attracted to visually conspicuous objects as potential prey items, not wary of them. The timing of lobster stridulation during aggressive defense appears to be unique among animals which produce defensive sounds. Unlike many other acoustically

43 aposematic animals (Klump & Shalter, 1983), sound is not produced as a warning to an approaching predator until after it has physically engaged the lobster. P. argus also do not wait until "capture" (grasping of the antennae, legs, or biting the carapace) to stridulate, which is typical of protean behavior used to startle predators (Masters, 1979; Hogstedt, 1983; Klump & Shalter, 1983; Smith, 1986; Driver & Humphries, 1988). Instead, sound is generally produced first during pre-emptive, aggressive lunges, which may or may not contact the fish. The timing of sound production by lobsters is fundamentally similar to when insects stridulate when disturbed or handled (Masters, 1979), which may be important for enhancing learning by predators of these aposematic prey (Claridge, 1974; Rowe, 2002). However, lobsters may produce sound only when predators are close because of the fish’s hearing abilities Other than triggerfish, piscine predators of nomadic and adult P. argus include nurse sharks, sting rays, and large groupers (Smith & Herrnkind, 1992; Sadovy & Eklund, 1999; Butler & Herrnkind, 2000). Without swimbladders or other accessory hearing structures to convert sound pressure waves to displacement, most teleost fish and all elasmobranchs can detect acoustic cues only through particle motion, which attenuates within a few meters of the source, unless the signal is very loud or a low frequency (Casper & Mann, 2006). Although groupers and triggerfish both posses swim bladders (Moulton, 1958; Sadovy & Eklund, 1999; PEB pers. obs.), it is not yet known how well they detect sound. Were these predators primarily sensitive to particle motion like sharks and rays, it might explain why lobsters don't appear to use stridulation as a warning to distant predators, but instead as a learning tool during a close attack when the sound can potentially be detected best. Spiny lobsters appear to have evolved a type of aposematic defense that does not incorporate chemical unpalatability. Although some authors (Inbar & Lev-Yadun, 2005; Speed & Ruxton, 2005) have proposed that unprofitability mediated by spines may be common in aposematic animals, this study documents an example of this type of aposematism. It should not be surprising that such an aposematic mechanism would evolve in the Crustacea. Many terrestrial arthropods are chemically defended (Edmunds, 1974; De Cock & Matthysen, 2001; Laurent et al., 2005); however, with one known exception (Luckenbach & Orth, 1990), crustaceans do not appear to have developed

44 chemical unpalatability (Luckenbach & Orth, 1990; Pawlik, 1993; Becker & Wahl, 1996). Although no crustaceans are described as aposematic, despite the prevalence of these types of displays in other marine invertebrates and terrestrial arthropods (Guilford & Cuthill, 1990; Mappes et al., 2005), mechanical unpalatability may be an advantage. Unprofitability mediated by substantial armor and spiny defenses can be evaluated by predators at a distance and is an honest representation of an individual's defensive capacity (Inbar & Lev-Yadun, 2005; Speed & Ruxton, 2005). Mechanically defended prey also can be sampled by predators and released unharmed, which has been suggested to improve the likelihood of evolving aposematic defenses under individual selection (Wiklund & Jarvi, 1982; Guilford & Cuthill, 1990; Tullrot, 1994; Speed & Ruxton, 2005). The effectiveness with which lobsters repelled naïve triggerfish and discouraged future attacks, without being injured, suggests that one reason chemically defended crustaceans are rare is that spines are an evolutionarily more accessible and/or effective route to unpalatability for this taxon. Other than their inability to produce sound, muted lobsters generally did not differ from stridulating lobsters in the execution of defensive actions. This suggests that the tendency to initiate and persist in more attacks against muted lobsters by naïve fish was not the result of another change in the lobster’s defense. In the choice experiment, stridulation also did not appear to increase the defensive ability of stridulating lobsters over silent individuals. Experienced triggerfish in other experiments have been shown to prefer less-defended individuals (those missing an antenna from one side, Parsons, 2005). We anticipated that, were lobsters of unequal defensive ability because one could not stridulate, the triggerfish would choose to attack and disable muted lobsters. The lack of a choice by the triggerfish in this experiment suggests that stridulating lobsters did not have a greater defensive capacity than muted lobsters. This supports our assertion that the role of stridulation in aggressive defense (against triggerfish) is to facilitate learning by the naïve predator, rather than act as a repellant per se. The function of stridulation produced coincident with other defensive actions such as tailflips, grasping of the antennae, or other handling is still unclear. The timing of sound production suggests a startling role in these instances (Driver & Humphries, 1988), although perhaps not effective individually against triggerfish or other diurnal piscine

45 predators. However, we have also observed (PEB, WFH pers. obs.) that lobsters in multiply-occupied natural dens can be instigated simultaneously to tailflip once a single, disturbed individual begins tailflipping. A large number of co-denning lobsters tailflipping and stridulating simultaneously upon disturbance is potentially loud and 'annoying' enough to deter a diurnal predator like a triggerfish. Stridulation associated with tailflips could also function during nocturnal foraging. Panulirus argus forage at night away from shelter and encounter octopus, rays, and nurse sharks primarily under circumstances where visually directed, aggressive defense is not possible. Although both octopus and elasmobranchs likely do not hear well at distance in the range of frequencies produced by lobsters (Hanlon & Budelmann, 1987; Casper et al., 2003; Casper & Mann, 2006), particle motion and vibration from stridulation should be detectable when close or in contact with lobsters. Other than during attacks by triggerfish (Bouwma, Chapter 1), sound production has most commonly been observed in spiny lobsters while gripping the carapace during handling (Moulton, 1957; Kanciruk, 1980) which superficially mimics grasping by predators like octopus. Western Australian Panulirus cygnus has been observed tailflipping with stridulation in response to contact from octopuses; however, the role of stridulation during encounters with octopus to this point is undetermined (Cobb, 1980). Aggressive defense in Panulirus argus does not appear to be triggered by triggerfish attacks only. We have observed P. argus individuals striking at a simple model (a grey-painted 2-liter soda bottle; PEB & WFH, unpublished data) and at various types of fish near dens in the field (PEB, pers. obs.). However, the use of aggressive defense may be somewhat predator-size specific. Unless a lobster individual is very large, P. argus generally do not strike at divers (simulating a large predator) who venture within antennal range, even when molested in the open (WFH, pers. obs.). We have observed that these lobsters usually attempt to fend off collecting equipment (sticks and nets), followed by rapid retreat with tailflips. Additionally, it is questionable what the efficacy of aggressive antennal defense by a 1-2 kg lobster would be against a 3 m nurse shark or 300 kg goliath grouper. We have observed a ~90 mm CL P. argus in the open during the migratory period repelling an attack by a ~1 m diameter loggerhead turtle (WFH, pers. obs.). This lobster did not aggressively lunge at the turtle, but used tailflips

46 and fended off the turtle with the antennae until it escaped into shelter after ~10 min of interaction (WFH, pers. obs.). Although other P. argus defensive responses to large, engulfing or crushing predators like groupers or elasmobranchs are not documented in the field, suggesting that aposematic, aggressive defense may have evolved primarily to confront relatively small predators. Very large, well-armored, and aggressive adult lobsters (>120 mm CL, ~2-5 kg) are defensively formidable and likely have few natural predators (Butler et al., 2006). Some studies indicate that P. argus is one of the fastest growing crustaceans in the Caribbean and reach a very large size in a just a few years (Butler et. al., 2006; Maxwell, 2006). Defensive behavior is arguably less important for larger lobsters as armor and large size per se probably provide refuge from most predators. Although very large adult lobsters have been observed lunging at a diver (WFH, pers. obs.), juveniles transitioning from a non-migratory, crevice-dwelling existence to nomadism are far more vulnerable and likely benefit more from aggressive defense than larger conspecifics. Aggressive defense and aposematism may allow these individuals to transition earlier than their size alone would allow to the nomadic phase, which is necessary to migrate the 10-30 km from nursery habitat to the reef where all mating occurs. This phenomenon might be widespread in the stridentes, as similar ontogenetic migrations over similar or greater distances occur in other spiny lobster species, including Panulirus cygnus, P. ornatus, Palinurus delagoae, and P. gilchristi (Herrnkind, 1980; Butler, et. al., 2006). Further work is needed to determine how sound and spines are utilized in defense against piscine predators in these species. Triggerfish are unique among spiny lobster predators due to their exceptional maneuverability and ability to use a small mouth to devour a lobster piecemeal rather than all at once. Triggerfish jaws are fused to the cranium, which results in greater bite forces than other lobster predators, allowing triggerfish alone to punch through the strongly armored lobster carapace and abdomen (Wainwright & Friel, 2000; Santini & Tyler, 2003). Additionally, the triggerfish's attack strategy of disabling lobsters by sequentially removing defensive weapons, legs, or by targeting the eyes, means that lobsters may not have size refuge from a persistent or skilled triggerfish. Although we have observed very large, experienced triggerfish fail to engage large adult P. argus

47 (~100 mm CL) (P. Bouwma, pers. obs.), Atlantic triggerfish (Balistes carolinensis) appeared to learn to preferentially attack the largest lobster in a tethered group (K. Lavalli, pers. obs.). Therefore, even otherwise well-protected, large lobsters might benefit from triggerfish learning to avoid spiny defenses through aposematism. This suggests that the selective pressure to develop and maintain effective aggressive defense might persist throughout the life of a lobster, not just during earlier juvenile stages. Although this study demonstrates the first evidence for a function of sound production in a stridulating lobster, why stridulating originally evolved is not clear. The ~21 species of the Panulirus genus appear to have evolved from a "P. argus-like" ancestor around 2 Mya (George & Main, 1967; Holthuis, 1991; Ptacek et al., 2001; Patek et al., 2006). Because we see similar aggressive defense in other stridentes species (Palinurus elephas), it seems likely that the ancestral form of Panulirus also possessed aggressive, retaliatory defense and associated defensive weaponry as well as the tailflip escape response. However, there is evidence to suggest that escape rather than retaliatory defense may have been the context in which stridulation initially evolved much earlier (~200 Mya). It appears that the stridentes arose from a deep-water ancestor, similar to the extant genera Puerulus and Linuparus, after which some species moved into shallow- water habitats (George & Main, 1967). Visually-directed, aggressive defense should be less useful than escape behaviors in deep-water, low-light habitats. Additionally, extant deep-water palinurid lobsters lack the well-calcified, robust exoskeleton and strong legs of shallow-water species (George, 2005). On the other hand, all stridentes lobsters appear to stridulate while attempting to escape with tailflips after being grasped, poked, or prodded by humans (Moulton, 1957; George & Main, 1967; Holthuis, 1991). This suggests that the stridulating organ and its use may have evolved initially to aide in escape (Chapter 3). Incorporating stridulation into the aggressive defense that we now see in P. argus may have followed later with the invasion of shallow waters and the associated increase in predation risk.

48 Table 2.1 (a) Descriptions of a subset of individual Panulirus argus defensive actions (during attacks only) and (b) triggerfish (Balistes capriscus) attack behaviors recorded during encounters in both the aposematic trials and the triggerfish choice experiment.

Action Description (a) Lobster defense Aggressive Defense Lunge Aggressive defense where lobster swings the antennae forward from a spread position (in a “ready-to-defend” posture) while P. simultaneously driving the body forward using a kick of the tail ( argus only).

Whip Lobster strikes the fish with either one or both antennae without an associated forward movement of body or simply pushes the fish away using the antennae. Pirouette Lobster turns in place while keeping the antennae and anterior of body directed at attacking fish.

Escape Tailflip Behavior where the lobster retreats rapidly by swimming, propelled by strokes of the tailfan and muscles of the abdomen.

(b) Triggerfish Bite Attempt Triggerfish attempts to bite the antennae, legs, abdomen or eyes with or without making contact with lobster. Circles Triggerfish makes at least ¾ of circuit around the tethered lobster within no more than 2 antenna lengths

49 Table 2.2. (a) Number of defensive actions per minute of encounter and results of t-tests comparing means for stridulating (N = 12) and muted (N = 13) lobsters during encounters with triggerfish. (b) Triggerfish attack behaviors per minute of encounter and results of t- tests comparing means for fish attacking stridulating (N = 12) versus muted (N = 13) lobsters.

Actions Stridulating Muted t P (a) Lobster Lunges 2.1 (± 0.3) 2.8 (± 0.4) 0.44 0.668 Tailflip 5.4 (± 1.2) 5.7 (± 1.2) 0.19 0.855 Whip 24.9 (± 2.6) 27.7 (± 1.9) 0.89 0.384 Pirouette 3.6 (± 0.7) 3.5 (± 0.4) 0.02 0.982

(b) Triggerfish Bite Attempts 5.9 (± 0.8) 4.3 (± 0.5) 1.71 0.103 Circles 4.0 (± 0.7) 4.2 (± 0.5) 0.25 0.802

51 (a) (b) Stridulating 14 Muted ** 13

d 12 11 ** 10 10 9

8 7

4 3 Lobster Individuals Kille 2

0 AfterAfter Day Day 1 1 After After Day Day 5 5 TriggerfishFish Choice Choice

52 (a)

Stridulating 8

r Muted 7 5.85 6 4.17 4.26 5 3.96

er Minute of Encounte 3 p

1 Behaviors 0 Circles Bite Attempts Circles Bite Attempts (b) Stridulating Muted 27.71 30 r 24.86

15 er Minute of Encounte p 10 5.40 5.71 3.55 3.53 5 2.07 2.81 Behaviors

0 Lunges Tailflips Whips Pirouettes Lunges Tailflips Whips Pirouettes Figure 2.3. (a) Triggerfish behaviors per minute of encounter during attacks against stridulating and muted lobsters. (b) Defensive behaviors per minute of encounter exhibited by lobsters during attacks by triggerfish. N = 12 for stridulating lobsters, N = 13 for muted.

53 (a) (b) Stridulating Experienced 80 Muted 55.8 Inexperienced 70

r * 60

30 14.93 16.26 Minutes of Encounte 13.8 14.4 20 10.76 12.1 10.85 10

0 TimeTime To Kill-Aposematism To Kill TimeTime BeforeTo First First Attack- Attack TimeTime To Kill-Taster's To Kill Choice Time ToTime Kill-Experienced To Kill vs. AposematismTrials Trials AposematismAposematism Trials Trials Choice Trials ExperiencedInexperienced vs. Inexperienced Fish Figure 2.4. (a) Minutes of encounter before a kill (N = 12 for stridulating, N = 13 for muted), minutes after introduction of the triggerfish to the arena before the first attack (N = 15 for stridulating and muted), and minutes of encounter before a kill in the choice experiment (N = 9 for stridulating, N = 11 for muted). Asterisk indicates significance at the 0.05 level. (b) Minutes of encounter before a kill for all experienced and inexperienced triggerfish (N = 20 for experienced and inexperienced fish).

54 CHAPTER 3

SOUND PRODUCTION IN CARIBBEAN SPINY LOBSTER PANULIRUS ARGUS AND ITS ROLE IN ESCAPE DURING PREDATORY ATTACK BY OCTOPUS BRIAREUS

INTRODUCTION

Many animals produce auditory distress signals when captured and/or handled by humans, presumably in an antipredator context. For example, various birds produce distinctive distress calls when captured in mist nets and handled by humans during banding (Perrone & Paulson, 1979; Högstedt, 1983; Klump & Shalter, 1984; Chu, 2001; Neudorf & Sealy, 2002). Sound producing insects such as cicadas, wasps, true bugs, and beetles also produce defensive stridulation when handled or restrained (Alexander, 1967; Smith & Langley, 1978; Masters, 1978; 1980; Schilman et al., 2001). Young crocodiles are also known to make alarm calls when captured and handled by humans (Staton, 1978; Romero, 1983; Gorzula, 1985). Thirty-five different fish taxa produce distress calls in response to capture or other anthropogenic disturbance (Myrberg, 1981). Some 39 species of spiny lobsters (Palinuridae) are also well known to produce sounds, commonly called stridulation, when handled (George & Main, 1967; Kanciruk, 1980; Holthuis, 1991, Patek & Oakley, 2003; Childress & Jury, 2006). Most of these sounds are assumed to deter predators. Various functional hypotheses for distress calls have been suggested, including: calling for help from conspecifics, alerting conspecifics to danger, attracting another predator to interfere with the attacker, aposematism, and startling the predator, thus facilitating escape (Klump & Shalter, 1983; Smith, 1986; Driver & Humphries, 1988). Excepting insect stridulation and crocodile alarm calls, the effectiveness and functional mechanisms of distress calls are not well understood experimentally (Myrberg, 1981; Hogstedt, 1983; Klump & Shalter, 1984; Childress & Jury, 2006).

55 Although most of the animal taxa listed above also produce sounds in contexts other than antipredation (Alexander, 1967; Campbell, 1973; Myrberg, 1981; Kroodsma & Byers, 1991), spiny lobster sounds appear to operate primarily in an antipredatory context. Berrill (1976) noted stridulation by small (1st and 2nd benthic stage) Panulirus cygnus during intraspecific aggressive interactions between conspecifics. No response was observed when these sounds were replayed to P. cygnus (Berrill, 1976). Additionally, Mulligan & Fischer (1977) noted stridulation during aggressive interactions by adult Panulirus argus, although it is unclear whether sound was produced with tailflips elicited by these interactions or as part of “threat” displays similar to those seen in small P. cygnus. In a brief abstract, Mercer (1973) suggests that Palinurus elephas female stridulate to attract males. However, this observation was never documented further (Atema & Cobb, 1980). All other described instances of lobster stridulation stem from putatively antipredator behavior elicited by human or natural predators (Lindberg, 1955; Moulton, 1957; Hazlett & Winn, 1962a,b; Heydorn, 1969; Berry, 1971; Takemura, 1971; Meyer-Rochow & Penrose, 1974; Smale, 1975; Meyer-Rochow & Penrose, 1976; Mulligan & Fischer, 1977; Cobb, 1980; Patek 2001, 2002; Patek & Oakley, 2003). In many of the instances described in these studies, lobsters stridulated when held or restrained by humans, and sound was often associated with tailflip escape attempts. Spiny lobsters produce sound in a way unique in the animal kingdom (Patek, 2001). Sound emanates from a specialized stridulating organ located at the base of the long 2nd antennae, which are also the lobster’s primary defensive weapons. The stridulating organ consists of two parts, a movable plectrum and flap attached to the last segment of the antennal peduncle, and the rigid file, located between the eyes on the antennular plate (Moulton, 1957; Holthuis, 1991; Patek, 2002). The underside of the plectrum is made of soft, elastic material and the file, while macroscopically smooth, is covered in microscopic scales facing the same direction (Meyer-Rochow & Penrose, 1976; Patek, 2002). To produce sound, a lobster draws the plectrum up the file by moving the antenna base posteriorly. Friction between the soft underside of the plectrum and the anteriorly projecting scales of the file cause the moving plectrum to alternately stick and slip, producing a pulse of sound with each slip (Patek, 2001; 2002).

56 Consequently, in addition to sound, when lobsters stridulate, the entire flagella of the antenna and the carapace can also be felt vibrating (Moulton, 1957; PEB pers. obs.). George and Main (1967) suggested that the stridulating apparatus first appeared ~200 Mya and split the Palinuridae into two groups, the stridentes (9 genera) and the silentes (3 genera, 9 species). Although it is not possible to determine exactly why stridulating evolved in ancestral palinurids, until recently, the only data pertaining to the use of stridulation in extant taxa came from the aforementioned interactions with humans. In studies with natural predators, we determined that two different species of lobsters, Panulirus argus and Panulirus guttatus, stridulate during attacks by a natural predator, the grey triggerfish, Balistes capriscus (Bouwma, Chapter 1). Both lobster species stridulate during tailflip escape attempts and when a triggerfish bit and pulled at the antennae, which usually elicited tailflips (Bouwma, Chapter 1). Panulirus argus also stridulated coincident with aggressive lunging defense, in which the lunging lobster drove the antennae, with its forward-pointing spines, at the fish’s eyes while propelling the body forward with the tailfan (Bouwma, Chapter 1.). We also found that stridulation during aggressive defense in P. argus functions as part of an aposematic defensive strategy against triggerfish, teaching naïve individuals not to attack in future encounters (Bouwma, Chapter 2). Although stridulating was effective in the context of aggressive defense, we were unable to determine a function for stridulation during tailflips or when triggerfish grasped and pulled the antennae (Bouwma, Chapter 2). However, P. argus face a variety of other predators with different hunting techniques, particularly at night while individuals forage away from shelter. Bonnethead sharks, nurse sharks, sting rays, toadfish, bonefish, permit, and octopus all prey on nocturnally active P. argus in the open (Smith & Herrnkind, 1992). Tailflips may be more effective for escape against these predators under cover of darkness than against diurnal, highly visual, strong-swimming triggerfish. Coincident with tailflips, stridulation may then facilitate escaping surprise attacks by some nocturnal predators. Of the nocturnal predators of P. argus, we chose to examine lobster responses to octopus. Lobsters actively choose not to shelter in areas with octopuses nearby and avoid dens emanating chemical cues from octopuses (Berger & Butler, 2001). Additionally,

57 Berry (1971) and Cobb (1980) observed lobsters (Panulirus homarus, P. cygnus) leaving dens after the entrance of an octopus. Observations in the laboratory indicate that octopuses catch rapidly retreating, tailflipping lobsters by grasping at the antennae and, once caught, grip the carapace and abdomen with their arms (PEB pers. obs.). It seems likely that such grasping at the antennae would cause a tailflip and stridulation response similar to that observed when triggerfish bite and pull the antennae. Cobb (1980) also noted some stridulation by P. cygnus lobsters at night when tailflipping to escape octopus. By gripping the carapace, the way lobsters are typically held, humans may be mimicking the grip of an octopus after capture, thus eliciting stridulation, tailflips, and other escape efforts. Were a lobster to escape the grasp of an octopus, the likelihood of surviving the encounter seems high as octopuses likely can not match the swimming speed of a tailflipping lobster. Studies on hearing in cephalopods indicate that, while octopuses may not hear the pressure oscillations in sound waves very well, they are quite sensitive to low frequency (1-200 Hz) waterborne vibrations associated with underwater sound (Hanlon & Budelmann, 1987; Williamson, 1988; Packard et al., 1990). In this study, we examine sound production by Panulirus argus during nighttime encounters with Caribbean reef octopus Octopus briareus. We address the following three questions: 1) do P. argus lobsters stridulate when attacked at night by octopus, 2) how is stridulation used in conjunction with other antipredator behavior such as tailflips, and 3) does sound production confer greater survival for stridulating lobsters compared to experimentally muted individuals. We demonstrate that lobsters produce stridulation both during grasping attacks by octopus and during escape attempts after being captured. We also show that stridulating lobsters are more likely to successfully escape from attacking octopuses and remain uncaptured much longer during encounters.

MATERIALS AND METHODS

Panulirus argus lobsters were collected by hand from shallow hardbottom areas near the Keys Marine Laboratory, Long Key, FL from May – August 2005 (FFWCC Permit #04SR-082). Individuals were held communally in large, flow through seawater tanks and fed periodically with a mixture of frozen squid and shrimp. All lobsters used in

58 this study were determined to be in intermolt stage by physical inspection of the carapace. To disable the stridulating apparatus, I clipped and removed the small, movable plectrum and flap from the base of the antennae at least 1 day prior to testing and allowed these individuals to heal in the communal tanks. Before testing, lobsters were visually inspected to assure that the clipped area had healed and no sound could be produced. Although stridulating lobsters were not clipped, each was handled similarly on the day prior to testing when they were measured, sexed, and visually inspected for injury before use the following night. Previous experiments (Bouwma, Chapter 2) indicate that muted lobsters otherwise exhibit the same defensive behaviors and survival times versus predatory grey triggerfish (Balistes capriscus). Octopus briareus were captured from baited crab traps near the Keys Marine Laboratory from May – August 2005. Octopuses were weighed and held individually in flow-through seawater enclosures with a piece of plastic pipe for shelter. Octopuses used in these trials ranged in total weight from 40 to 150 g. Each individual was subject to natural day/night cycles, and most were seen out and active in their enclosures after dark. Since newly captured octopuses were often reluctant to attack lobsters, each octopus was given at least 3 opportunities to attack and consume a live P. argus in their enclosures prior to testing. Octopuses that did not acclimate to captivity sufficiently to attack a lobster were released after a few weeks, away from the original capture site. Octopus were only fed live lobsters prior to and during experimental trials. Encounters between octopus and lobsters were carried out between 22:00 and 05:00 the next day in 600-liter oval, flow-through seawater tanks. Octopus were introduced to the experimental enclosures for acclimation at least 30 minutes prior to testing. Tanks were illuminated with dim red light from 25-watt light bulbs placed 2 m above the tanks. This provided sufficient light for digital video recording (Sony DVC1000) but did not disrupt octopuses and lobsters from engaging in typical nighttime activity. Trials were started when a lobster was introduced to the experimental tank by gently lowering it into the side of the tank opposite the octopus. Encounters were then videotaped from above while the observer hid behind a dark screen. Trials lasted until either the octopus successfully captured the lobster or did not attempt to catch the lobster

59 after 30 minutes. Octopuses that did not engage lobsters were returned to their enclosures without feeding to try again the next evening. Octopuses used in these trials all made lobster captures 2-4 days after their last (previous) meal. Because lobsters occasionally escaped an octopus, even minutes after capture, I videotaped escape attempts by lobsters for 3 minutes after they were grasped. By that time, octopus venom usually had taken effect, and lobster tailflips and other struggling had stopped. No lobster escaped following 3 minutes after capture. Octopuses were then allowed to consume their prey, which had already perished. To minimize the effects of octopus size and experience on capture efficiency, sized-matched pairs (within 2 mm CL) of stridulating and muted lobsters faced the same octopus predator when possible. Hunger state was the same (2-4 days) in both trials for each octopus. Each of 13 octopuses was used twice, once for each type of lobster. Four additional octopuses attacked only one type of lobster but escaped before the second trial. In total, 15 muted and 15 stridulating lobsters were tested against octopuses between June and August, 2005. Lobster size varied from 32.0 – 52.5 mm CL (mean 39.6 mm CL) and did not differ between size-matched muted and stridulating lobsters (t=2.05, P=0.41). Videotapes were reviewed to determine the number of reaches, grasps, and captures by the octopuses during encounters. A reach consisted of rapidly extending the arms toward a lobster, usually after contact. A grasp occurred when the octopus made contact with a retreating lobster and held on sufficiently well that the arm was pulled in the direction of the lobster’s travel. Captures were grasps from which lobsters did not escape within 10 seconds. Lobsters were scored for the number of tailflip bouts, any stridulation produced, other defensive behaviors, and successful escapes from an octopus’s grasp or capture. Attacks were defined to start when octopuses rose off the substrate and crawled or jetted towards the lobster and finished when the octopus successfully captured the lobster or made no attempt to catch it for more than 10 seconds. We then determined both the occurrence and timing of stridulation behavior in relation to defensive actions both during encounters and after the lobster had been captured. I compared, using Fisher’s Exact tests, the number of lobsters from each group escaping a grasp and those captured on the first reach. I also compared by t-test the total attack time required for each type of lobster to be captured.

RESULTS

During escape attempts from an attacking octopus, lobsters stridulated only coincident with tailflips. Some, but not all, of these tailflips were elicited by grasps of the antenna or body. During every grasp, lobsters stridulated on average 4.3 (± 0.4 SE) times before escaping with tailflips. Lobsters also stridulated coincident with tailflips during reaches where no contact was made, but not as frequently (27.1 % ± 3.9%). Usually, a lobster stridulated one time when the arm of the octopus apparently touched the body or antenna, which also elicited a tailflip. However, on subsequent tailflips in the same bout (there were often multiple strokes of the tailfan), lobsters were generally silent. Because we were unable to precisely count the subsequent strokes of the tailfan in the dim light, we notated them as 1 bout. Therefore, our estimate of the percentage of tailflips with stridulation is probably overestimated. Stridulating lobsters were statistically more likely than muted lobsters to escape at least once from the grasp of an octopus (Table 3.1; Fig. 3.2). All octopuses caught lobsters by first grasping the ends of one or both antennae and then wrapping additional arms around parts of the body. Consequently, a number of lobsters escaped by autotomizing one or both antennae when they were grasped. Stridulating lobsters were also more likely to escape at least once without autotomizing antennae (Table 3.1). However, there was no difference in the likelihood of injury between stridulating and muted lobsters. Additionally, 66.7 % of muted and 60.0% of stridulating lobsters were equally as likely to have the antennae grasped by an octopus on its first reach attempt. However, nearly 47% of muted lobsters were caught, did not escape, and were subsequently killed on the very first reach attempt as compared to none of the stridulating animals. Although three stridulating lobsters were caught by octopuses on the first try, all escaped after 65.0 (± 31.1) seconds, forcing the octopus to make a second catch. Only one of these escapees suffered major injuries (5 missing legs). Octopuses also took significantly longer to successfully capture stridulating lobsters than muted lobsters (21.6 ± 6.5 seconds for muted, 72.5 ± 11.1 seconds for stridulating) (t = 3.98, df = 23, P < 0.001). Octopuses initially attacked lobsters by

61 swimming close, spreading the arms and tentacles around the lobster and trying to grasp (primarily the antennae) the lobster as it tailflipped (Fig. 3.1a). During subsequent attacks, lobsters generally did not allow the octopuses to approach as closely, and catches were mostly made by long reaches of the tentacles. Presumably exhausted lobsters, after many tailflipping circuits around the tank (~1.8 m per circuit), were occasionally caught by octopuses grasping the carapace and/or abdomen. After an initial unsuccessful grasp attempt, octopuses were forced to pursue retreating lobsters, which often swam around the entire circumference of the tank once or more while being chased. After the long chase, lobsters usually struggled only feebly when eventually captured, and none subsequently succeeded in escaping the octopus. Octopuses appeared to identify lobsters visually in the dim red illumination and use visual cues to hunt them. In pilot trials, with infrared illumination only, octopuses were unable to successfully stalk and catch lobsters but instead seemed to blunder into them while moving around the tank. Regardless of the type of illumination (infared or dim red), lobsters did not respond to the initial approach of an octopus until after contact was made by a reaching arm. After capture (Fig. 3.1b), lobsters stopped all movements for several seconds before beginning to struggle to escape. In lobsters with intact stridulating organs, struggling began with increasingly more frequent stridulations without tailflips 8.8 (± 0.8) seconds after capture (Fig. 3.3). Lobsters alternated left and right antennae in producing the sound rather than using both simultaneously. At least 10 of the 15 muted lobsters also attempted to stridulate when held by octopuses, which we observed by the characteristic back-and-forth movements of the antenna bases used to produce sound in intact individuals. On average, stridulating lobsters produced sound for 91.8 (± 10.7) seconds after initiating sound production. Once lobsters stopped stridulating, typically at the same time other escape attempts stopped, they generally were silent for the rest of the encounter (mean 79.7 ± 10.6 seconds).

DISCUSSION

62 The occurrence and timing of stridulation by lobsters during octopus attacks suggest a defensive role primarily at the beginning of tailflip-mediated escape attempts. Lobsters stridulated during tailflips every time contact was made by the octopus but rarely during subsequent tailflipping (in the same bout) after a successful escape. This suggests that, in the open at night, stridulation is primarily elicited by contact and is not produced automatically with each pulse of the tailfan. Cobb (1980) noted a similar use of sound during tailflips by Panulirus cygnus, stridulation initially, followed by silent tailflips. Lobsters were particularly sensitive to contact made on an antennal flagellum; i.e., in every instance, contact there elicited a tailflip with stridulation. That lobsters initially allowed octopuses to close in and spread their arms around them before exploding off the substrate with tailflips suggests they did not visually detect approaching octopuses or did not recognize them as dangerous. Berry (1971) and Cobb (1980) noted in the laboratory and the field that P. homarus and P. cygnus typically pointed the antennae at and backed away from octopuses after visually identifying them during the day. We never observed this behavior in P. argus under red illumination at night. After an initial tailflip escape, lobsters were more responsive to octopuses and often tailflipped away before the approaching predator could make a reach attempt. In these later attacks, lobsters may have been responding to subtle hydrodynamic cues, visual cues like shadows from the swimming octopus, or other mechanical cues (Wilkins et al., 1996; Briones-Fourzan et al., 2006). In earlier daytime experiments against triggerfish (Bouwma, Chapter 1), I found that P. argus lobsters stridulated in >80% of tailflips, more frequently than with the 27.1% of tailflips we observed here. During triggerfish encounters, we observed stridulation with nearly every stroke of the tailfan during bouts, not just at the beginning of a tailflip bout as against octopus. Part of this difference may result from the physical restraint by tethers used in the triggerfish study, which only allowed defending individuals to move within a 30-cm radius. Although Parsons (2005) found that tethered lobsters did not differ from free-moving lobsters in the execution of defensive actions, she did not measure stridulation. During grasps by octopuses, lobsters appeared to stridulate with each stroke of the tailfan until breaking free, and then were silent as they swam away. Tethered lobsters could not break free of restraint, which may explain why

63 they stridulated with a greater percentage of tailflip strokes. It is not surprising that lobsters would remain silent after escape during nighttime encounters, as continued stridulation would potentially provide directional cues as to where they came to rest for the original predator or any other nearby. Tailflipping lobsters also commonly change directions before coming to a rest (PEB, pers obs) and any additional information, auditory or otherwise, indicating a new location might ruin this deception and increase the likelihood of being recaptured. Most octopuses were able to grasp lobsters of both types by the antennae on the first attempt. In the wild, a slow-moving octopus may only get one chance to stalk and successfully snare a lobster. Our data suggest that stridulating, tailflipping lobsters are highly effective at escaping grasps during nighttime attacks by octopuses. Even stridulating lobsters that were caught on the first try were able to eventually escape and tailflip away. Only after repeated attempts in the enclosed tank were octopuses able to catch these increasingly wary lobsters and for some, only when the lobster tired and could no longer tailflip effectively. Although octopuses are well-known opportunistic predators of crustaceans (Hanlon, 1983; Mather, 1991), this study may in part explain why there are so few descriptions of octopus predation on strident spiny lobsters in the wild, other than on constrained animals (e.g. in traps or tethered) (Joll, 1977; Cobb, 1980; Boyle, 1997, Barshaw et al., 2003). Lobsters in this study were difficult to capture for octopuses with experience catching lobsters, in relatively small tanks, and with a visual advantage for the octopus. This suggests that predation on stridulating lobsters in the open under natural conditions by octopus may be rare. A denned lobster confronted by an octopus might well be caught (Berry, 1971; Cobb, 1980). Perhaps this explains why lobsters avoid otherwise suitable dens near octopuses (Berger & Butler, 2001) and lobsters leave dens when octopuses enter (Berry, 1971; Cobb, 1980); the chance of being attacked by an octopus without being able to escape might be too high to risk. Remarkably, after only disabling the stridulating organ, under the same conditions, and with mostly the same octopus predators, lobsters were much less effective at escaping, with over 40% caught and killed on the first reach. This strongly suggests that stridulating is a vital component of tailflip escape against octopus, despite the fact that octopus predation on free-moving lobsters in the wild is likely to be rare. Because

64 octopus and lobsters reside in similar habitats, in hardbottom as juveniles and the reef as adults, and both forage nocturnally, they likely encounter each other with some frequency. Octopus briareus, like all octopuses, is a generalist (Hanlon, 1983; Jereb et al., 2005), and our experience suggests that they are likely to stalk and attempt to catch lobsters. Cobb (1980) observed octopuses attempting to catch lobsters at night in a large pool in the laboratory, without any success. While collecting lobsters at night, we have also observed an O. briareus grasp at and elicit tailflips from a free-moving P. argus (PEB, pers. obs.). The ever-present risk of encountering octopuses offers one explanation for why P. argus has retained stridulation with tailflips, despite the fact that successful catches by octopus might be rare – any silent individuals emerging in the population are much more likely to be killed. Additionally, lobsters and octopus co-occur throughout tropical and temperate seas (Berry, 1971; Cobb, 1980; Holthuis, 1991; Voigt, 1998; Groeneveld et al., 2006), suggesting that this type of strong predation pressure may not be just a local phenomenon but is likely widespread in Stridentes lobsters. Conversely, this suggests that shallow-water Silentes species such as Jasus edwardsii, which also co- occur with octopus (Holthuis, 1991; Voigt, 1998), will be caught more frequently when encountering octopuses away from dens. Heydorn (1969) observed octopus feeding on Jasus lalandii, a silentes species, but it is not clear whether the lobster was caught in the open or while constrained. Additionally, octopus take tethered J. edwardsii (Oliver et al., 2005) and are the primary cause of mortality in trap-caught J. edwardsii (Joll, 1977; Brock & Ward, 2003). However, the extent of octopus predation on free-moving Jasus spp. remains largely unknown. The timing of sound production, just as the octopus makes contact, suggests that stridulation may be a type of protean display, startling grasping octopuses (Driver & Humphries, 1988), as has been suggested by many authors (Lindberg, 1955; Moulton, 1957; Berry, 1971; Meyer-Rochow & Penrose, 1976; Mulligan & Fischer, 1977; Cobb, 1980; Kaciruk, 1980; Patek, 2001). Startle responses by fish might include a pause in attack or loosened grip on the lobster, thereby increasing the effectiveness of tailflips. However, we did not observe any pauses by attacking octopuses while grasping at either muted or stridulating lobsters. In this experiment, we observed grasps as when octopuses held on sufficiently well that the arm was pulled in the direction of the lobster’s travel.

65 Were octopuses to loosen their grip in response to stridulation, we expected fewer grasps of stridulating lobsters. In fact, we recorded no differences in responses by octopuses to stridulating lobsters as compared to muted individuals. Stridulating lobsters only appeared to be more difficult for the octopus to hold once grasped. One explanation for this might be the role of antenna flagellum-borne vibration associated with stridulating as opposed to the effects of sound per se, i.e. water-borne particle motion or pressure waves. For any object moving against another surface, such as an antenna against an octopus’s skin, the amount of kinetic friction between it and the surface is less than when that object is stationary (Giancoli, 1998). This is cited as the reason why it is easier to keep a heavy object moving than it is to start it moving from a stationary position (Giancoli, 1998). Because of vibration, the static force of friction between a lobster’s antenna and an octopus’s arm should be less than when the antenna is not vibrating. Unless sufficient force is provided to compensate for this decrease in friction, the antenna is presumably more likely to slip from an octopus’s grasp. By timing stridulation with tailflips, lobsters may be taking advantage of the substrate-borne vibrational component of sound production to escape. Panulirus argus lobsters appear to use the stridulating organ in two separate defensive contexts, as an aposematic display during daytime aggressive defense (Chapter 2) and to aid escape during tailflips at night. However, it is most parsimonious to suggest that the stridulating apparatus initially evolved to enhance the effectiveness of tailflips. All crustaceans with the extended abdomen typical of the macruran body plan (shimps, crayfish, lobsters, etc.) inherited the morphological capacity for tailflip escape (Holthuis, 1991; Spanier et al., 1991; Patek et al., 2006). George & Main (1967) and George (2005) also suggested a general evolutionary trend in the Palinuridae of moving from relatively stable ancestral conditions in deeper waters, where more primitive lobster taxa are still found (e.g. Puerulus, Linuparus), to more varied and fluctuating conditions in shallow- water habitats, where the most advanced palinurid taxa are found today (e.g. Panulirus, Justitia). It is parsimonious to assume that stridulation evolved initially as part of an escape strategy in the low-light conditions typical of deeper-water habitats (similar to nocturnal conditions). Presumably, this was followed by the development of visually- directed, aggressive defense later, when palinurids began to invade shallower, tropical

66 waters. A recent study also suggests a Paleozoic origin of the Octopoda, meaning that octopuses in some form or another existed for at least 50 My before the stridentes clade of palinurids diverged ~200 Mya (George & Main, 1967; Strugnell et al., 2006). Assuming that the stridulating organ did not evolve for another function, as yet unknown, which no longer exists in extant species, it seems most reasonable to suggest that stridulation evolved for facilitating tailflip escapes from grasping predators like octopus. If so, the stridulating organ evolved primarily to vibrate the antenna, with sound production as a by-product. Stridulation was likely incorporated later into aggressive, lunging defense. Interestingly, the two defensive actions coincident with stridulation, tailflips and lunging, both involve use of the tail to propel the body at or away from a predator. In lunges, the body is propelled forward by short, up-down (carangiform) kicks by the tailfan, while tailflips use repeated, full, downward strokes of the tail to move the body backwards. Once stridulation was associated with the tailflip mechanism, the basic physiological and mechanical machinery were probably present for the development of lunges with stridulation once visually adept palinurids moved into shallower water and began facing diurnal predators like triggerfish. However, until we have more information concerning defensive behavior in the extant Palinuridae, we can only speculate on how the use of the stridulating organ has evolved. Every intact lobster in this experiment also stridulated after capture by octopuses. Even muted lobsters attempted stridulations, although sound was not produced. Unlike earlier sound production while being chased, lobsters stridulated consistently for an average of over 90 seconds and much of the time without associated tailflips. This contradicts the observations of Berrill (1976), who reported no stridulation by early benthic stage P. cygnus while being eaten by fish or octopus. It is not known whether older P. cygnus also do not stridulate after capture. Our results with P. argus suggest a different reason for stridulating during captures than earlier in the encounter. After capture, stridulating might alert conspecifics of danger, call other lobsters to help, or attract an additional predator. Except in a den or other aggregation of lobsters, calling to conspecifics by a solitary lobster seems unlikely because, unless conspecifics are very close, they probably can not detect the signal (Lovell et al., 2005; Bouwma, unpublished data). Additionally, all attempts to record lobster responses to stridulation experimentally

67 have failed thus far (Atema & Cobb, 1980). However, piscine predators of lobsters or octopuses might be able to detect stridulation from some distance and interfere. Berry (1971) observed in South Africa that “Moray eels are attracted to the stridulation of spiny lobsters [Panulirus homarus] and even if not actually present in a shelter with them, may appear and attack an intruder.” Moray eels are also known to be predators of octopus where they occur together in the Caribbean (Young & Winn, 2003). Morays are commonly seen co-denning with spiny lobsters but apparently prey upon them only rarely (Kanciruk, 1980; Cruz & Phillips, 2000; Young & Winn, 2003). Were eels attracted to lobster stridulation after a capture, the resulting encounter between eel and octopus might allow a captured lobster to escape. The nature of the eel-lobster-octopus interaction has yet to be described; therefore, whether stridulation functions in this third way, to attract predators, remains unknown. An analog to palinurid stridulation during handling by humans is the low- frequency “buzzing” vibrations produced by American clawed lobster Homarus americanus when picked up or handled (Fish, 1966; Atema & Voigt, 1995; Henninger & Watson, 2006). In clawed lobsters, sound is produced by simultaneously contracting antagonistic remoter and promoter muscles at the base of the second antennae, producing vibrations in the carapace (Henninger & Watson, 2006). Like palinurids, which alternate stridulating with left and right side stridulating organs over extended periods, clawed lobsters also alternate left- and right-side pairs of muscles in producing a series of vibrations (Henninger & Watson, 2006). Although octopus are not described as a predator of Homarus americanus, the European lobster Homarus gammarus, which vibrates when handled similarly to H. americanus (K. Lavalli, pers. comm.) occur in the Mediterranean and Atlantic coast of Europe, where they are potentially prey for co- occurring octopus (Holthuis, 1991; Barshaw et al., 2003, van der Meeren, 2005). Were octopus to attack and capture H. gammarus, eliciting vibrations, it might suggest that an escape function for vibration might exist in homarid lobsters, analogous to stridulation in palinurids.

68 Table 3.1 Numbers of individuals escaping from octopuses at least once, individuals escaping at least once without injury, individuals grasped on the antenna on the 1st attempt, individuals captured and killed on the first attempt, and individuals which lost an antenna through autotomy before capture. Results of Fisher’s exact tests comparing frequencies between stridulating and muted lobsters.

N = 15 (per group) Stridulating Muted Fisher’s Exact (P-value) Individuals which:

Escaped at least 14 8 0.017 once, Escaped at least 10 4 0.028 once without injury, Were grasped by 9 10 0.275 the antenna on 1st attempt, Were caught and 0 6 0.008 killed on 1st reach attempt, Suffered injury 6 4 0.227 before capture

69 (a)

1. 2.

3. 4.

(b)

70 Stridulating

100 Muted 93.3** 90 80 * 70 66.7 60.0 60.0 60 53.3 ** 50 40.0 40.0 40 26.7 26.7 Percent of Lobsters 30 20 10 0.0 0 At Least One Escape Without Grasped Antenna Caught and Killed Injury Before At leastEscape one At leastInjury one OnAntenna 1st Attempt CaughtOn 1st Reach and InjuryCapture before escape escape w/o grasped on killed on 1st capture injury 1st attempt reach attempt

Stridulating 16 Not Stridulating

Number of Lobsters 4

0 5 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 Seconds After Capture

Figure 3.3. Numbers of lobsters stridulating and silent at 10 second intervals after capture by an octopus (N = 15 lobsters).

72 CHAPTER 4

THE ONTOGENY OF ANTI-PREDATOR RESPONSES TO ALARM ODOR IN CARIBBEAN SPINY LOBSTER PANULIRUS ARGUS

INTRODUCTION

The likelihood of encountering predators and not surviving is one of the most important factors shaping prey behavior and morphology (Edmunds, 1974; Sih et al., 1985; Sih, 1987; Vermeij, 1987; Lima, 1998; Sih et al., 1998). Antipredator strategies are diverse, including chemical defenses (Edmunds, 1974; Guilford & Cuthill, 1991; Pawlik, 1993; Summers & Clough, 2001; Laurent et al., 2005); circumstantial adjustment in the timing of hatching or metamorphosis (Sih & Moore, 1993; Warkentin, 1995; Benard, 2004); morphological adaptations including large body size, weaponry, defensive spines, armor, and cryptic coloration (Vermeij, 1982; Werner & Gilliam, 1984; Cloudsley-Thompson, 1995; Sih, 1987; Caro et al., 2001; Barshaw et al., 2003; Lass & Spaak, 2003; Inbar & Lev-Yadun, 2005); defensive behaviors such as migration, nocturnal foraging, fleeing, aggressive defense, aggregation, and sheltering (Hamilton, 1971; Zaret & Suffern, 1976, Kanciruk, 1980; Lampert, 1989; Bollens et al., 1992; Cloudsley-Thompson, 1995; Wooster & Sih, 1995; Caro et al., 2001; Lingle & Pellis, 2002; Barshaw et al., 2003; Stankowich & Blumstein, 2005; Childress & Jury, 2006); or a strategy employing some combination of chemical, developmental, mechanical, and/or behavioral defenses. Antipredation processes are often plastic, changing to reflect the intensity of local predation (Sih, 1987; Lima & Dill, 1990; Lima, 1998). Although chemical and mechanical defenses, as well as life-history traits can change as long-term responses to predation risk, animals generally rely on behavior to contend with more rapid, moment-to-moment fluctuations in risk (Sih, 1987; Lima & Dill, 1990). Commonly, predation risk quickly rises when a predator moves into an area or begins actively hunting and is visually, chemically, or mechanically detected by nearby prey (Sih, 1987; Lima & Dill, 1990; Lima, 1998; Stankowich & Blumstein, 2005). Vulnerable

73 prey must then make a crucial response: flee (often abandoning feeding, mating, and/or sheltering opportunities) or stay and rely on aggressive defense (Sih, 1987; Lingle & Pellis, 2002; Barshaw et al., 2003; Wasson & Lyon, 2005), shelter (Lima, 1993; Lehtiniemi, 2005; Stankowich & Blumstein, 2005), the presence of conspecifics (Bertram, 1978; Pulliam & Caraco, 1984; Herrnkind et al., 2001; Lavalli & Spanier, 2001), or crypsis (Sih, 1987; Eilam, 2005) to survive. Evolution is expected to select for individuals that either stay or leave an area after detecting a predator based upon which of the options provides the greatest net fitness benefit at the particular size, age, and/or life stage of the prey. For example, larger and/or older experienced individuals are more likely to survive a physical encounter with a predator because of greater defensive ability (Werner & Gilliam, 1984; Lingle & Pellis, 2002; Wasson & Lyon, 2005; Butler et al., 2006). Additionally, the number of different predator species potentially able to successfully attack a given prey decreases with growth (Sogard, 1997; Mittelbach & Persson, 1998; Honma et al., 2006). Conversely, as individuals increase in body size, they are both less likely to find appropriately-sized shelter should they leave a familiar area (Caddy, 1986) and benefit less from crypsis (Sih, 1987; Blanckenhorn, 2000). Therefore, body size, particularly during the rapid growth period, should influence the outcome of the decision to stay or leave an area in response to a predator. However, animals typically undergo changes in morphology, habitat use, or reproductive state as they grow, which may also help determine this crucial decision (Wilbur, 1980; Magnhagen, 1991; Butler et al., 2006). It is not clear whether increasing body size during ontogeny, without the influence of these other factors, alters the response of whether to stay or leave when a predator is detected. Spiny lobsters (Palinuridae) represent an excellent model to address this question. After settling from a prolonged planktonic larval stage, most palinurid species live as benthic juveniles across a wide size range (~6-75 mm carapace length) and coexist in the same habitat (Bulter et al., 2006). Palinurid lobsters generally undergo only slight, gradual, morphological changes during the juvenile period (Butler & Herrnkind, 2000; Bulter et al., 2006). For example, the Caribbean spiny lobster Panulirus argus, passes through three behaviorally distinct but morphologically similar ontogenetic stages as juveniles in shallow vegetated and hard-bottom habitat (Butler & Herrnkind, 2000). In

74 South Florida, USA, new recruits settle into regionally abundant foliose macroalgae at about 5-6 mm CL (Marx & Herrnkind, 1985a; b; Butler & Herrnkind, 2000). After approximately three months of sheltering and foraging within vegetation, algal phase individuals shift to a post-algal stage at about 20 mm CL and shelter in nearby sponge, rock, or coral crevices (Childress & Herrnkind, 1996; 2001). A change in color pattern, but not morphology, coincides with the shift in behavior (Bulter et al., 2006). Algal phase lobsters often delay the shift from living in macroalgae to sheltering in crevices slightly in the presence of predators (Childress & Herrnkind, 2001). Post-algal phase lobsters, which over several molts acquire the adult coloration, leave the den to forage at night, usually within ten meters of shelter (Butler & Herrnkind, 2000). At about 40 mm CL, post-algal phase lobsters make another ontogenetic shift to the nomadic phase (Butler & Herrnkind, 2000). Nomadic phase P. argus (40-75 mm CL) shelter during the day alongside post-algal juveniles but forage over greater distances (~100 m) and travel up to 30 km over a period of several months to a year, moving out of nursery habitat to the reef, where they attain sexual maturity (Herrnkind, 1980; Butler et al., 2006). Tethering experiments in nursery habitat show that, as lobsters grow larger, predation risk decreases, particularly when lobsters are exposed away from shelter (Butler et al., 2006). Although behavioral responses to predators by early stage P. argus within macroalgae have not been observed, their small size, seclusive behavior and outline disruptive color patterns, which camouflage them in vegetation, suggest a defensive strategy of crypsis (Childress & Jury, 2006). However, older juvenile P. argus, when attacked, display several distinct antipredator behaviors. For example, human disturbance of nomadic stage P. argus in crevices causes a lobster to initially fend off the attacker with the spiny antennae and, if the disturbance continues, to retreat deeply into the crevice using tailflips and then wedge tightly (Kanciruk, 1980). If caught in the open by predators, such as during diurnal mass migrations (Herrnkind, 1969), lobsters either swim rapidly in retreat using tailflips (although they fatigue quickly) or aggressively fight back, lunging and parrying with the robust, spinose 2nd antennae (Herrnkind, et al., 2001; Childress & Jury, 2006; Bouwma, Chapter 1). When attacked as a group in the open, P. argus also form closely packed, outward-facing defensive “rosettes” and defend collectively (Kanciruk, 1980; Herrnkind et al., 2001; Childress & Jury, 2006). To

75 summarize, juvenile P. argus exhibit an array of alternative escape strategies, when confronted by predators that change during ontogeny (Childress & Jury, 2006). However, numerous aspects of escape behavior, triggering stimuli, and relationship to size and life-stage, as well as the influence of size per se, remain to be examined. Indirect evidence suggests that older juveniles and adult spiny lobsters, when sheltering in a den during the day, respond to attack by abandoning that den the following night and leave the immediate area. While capturing P. argus in the Florida Keys, we have observed that often, when a nomadic phase lobster is disturbed in a group of conspecifics, other individuals in the den leave within several minutes (PEB, pers. obs.). Herrnkind et al. (1975) documented that adult P. argus were much more likely to move from a den after being handled in that den than lobsters that were closely approached but only observed in dens. Lindberg (1955) noted that adult P. interruptus, when “gripped, held, or pulled in any way,” in the field would back out of a den and tailflip away after a few minutes. South African fishers commonly use a dead octopus tied to a pole thrust into the den to coax Panulirus homarus to leave shelter (Berry, 1971). George (1972) noted that fishermen in the New Hebrides islands (now the Republic of Vanuatu) used a similar technique with a simulated predator on a pole to coax Panulirus penicillatus lobsters to leave dens. Although none of these observations include responses to natural predators without artifact (Frid & Dill, 2002), they suggest that leaving a den in response to similar disturbances during predator attacks might be common in the genus Panulirus. To determine if and how the decision to stay or leave after detecting a predator changes with the different juvenile stages of Panulirus argus, it is important to use a predator cue that conveys imminent danger regardless of prey size. A nearly ubiquitous cue in aquatic systems, which indicates immediate predation danger, is incorporated in the suite of chemicals emanating from a recently killed conspecific, often termed “alarm odor” (Kats & Dill, 1998; Dicke & Grostal, 2001). Alarm odor can be produced from injuries caused by a variety of predators, large and small (Kats & Dill, 1998). Because alarm odor is only released when a prey is injured or killed, it is a reliable indicator of an active predator nearby feeding on conspecifics (Kats & Dill, 1998; Dicke & Grostal, 2001). Although immediate or overt responses to alarm odor are not described in palinurid lobsters, there is evidence that alarm odor from injury or death influences

76 shelter selection in P. argus, which avoid dens infused with the odor of a dead conspecific (Briones-Fourzan et al. 2006; Parsons & Eggleston, 2005). Additionally, many other marine and freshwater crustaceans respond strongly to conspecific alarm cues (Hazlett & Schoolmaster, 1998; Hazlett et al., 2000; Wisenden et al., 2001; Lass & Spaak, 2003; Mima et al., 2003). In this study we examine 1) how the three ontogenetic phases of Caribbean spiny lobster Panulirus argus respond to exposure to fresh conspecific body fluid and 2) how antipredator behavior, particularly the decision to stay or leave the area, changes during ontogeny. Additionally, we examined how the presence of conspecifics affects the decision to stay or leave in gregarious juvenile P. argus. We show that, although all size classes of P. argus respond to alarm odor, the decision to stay or leave dens changes in an unexpected way with increasing body size and in the presence of conspecifics. We also demonstrate that body size is a strong predictor of distance traveled once a lobster abandons shelter in response to alarm odor.

MATERIALS AND METHODS

Field experiments We tested the response of post-algal and nomadic phase Panulirus argus to lobster body fluid in natural dens at three nearshore sites in Florida Bay within 15 km of the Keys Marine Laboratory, Long Key, FL between July and December, 2003. All sites were typical of hardbottom lobster nursery habitat with surrounding seagrass in 1-2 m depth (Smith & Herrnkind, 1992; Herrnkind et al., 1997). Approximately 30 minutes before testing, divers carefully marked up to 4 dens in sponges, under sea whips, in solution holes, or in artificial structure (concrete blocks, abandoned traps, etc.) with small, removable buoys at least 5 m apart. One lobster in each den was chosen as a focal individual and was visually estimated to be either post-algal or nomadic. Lobsters in each ontogenetic stage were then subdivided into 2 categories: 1) denning with conspecifics or 2) denning alone. For all lobsters sharing dens, regardless of life-history stage, co-denning conspecifics ranged in number from 1 – 7 (mean = 2.2, median = 2) and occupied a size range of 20-70 mm CL.

77 Lobsters in dens were then given 60 ml of either a seawater control, drawn from near the den, or a conspecific body fluid odor, delivered through a short (5 cm) rubber tube by a syringe. The conspecific body fluid odor (hereafter called alarm odor) was freshly prepared by crushing a single lobster (45-55 mm CL) in 0.5 L of seawater, simulating an attack by a predator such as a nurse shark or triggerfish. Although spiny lobsters are otherwise difficult to euthanize quickly, lobsters killed in this way died instantly. This mixture was then filtered through fine-mesh netting to remove large particles and loaded into the 60 mL syringes immediately before testing. Separate syringes were used for alarm and control odors. Alarm odor prepared from P. argus typically became darker and thickened slightly after ~ 45 minutes after preparation; therefore, we only used odor prepared within 40 minutes of testing. Since each replicate required ~ 10 minutes to complete, this allowed a maximum of only 4 treatments per alarm odor preparation. In order to minimize sacrificed individuals, we randomly assigned the first set of 4 focal individuals on a particular day to either the alarm odor solution or the seawater control. A second set of 4 individuals on the same day was then given the alternate treatment. This was done to reduce differences between treatments because, environmental conditions such as temperature, sunlight, turbidity, current speed and direction, and wave action varied from day to day. Additionally, to minimize the likelihood of contamination between treatments, we moved against the current, testing subsequent individuals upstream from prior tests. Sets of control and alarm tests performed on the same day were separated by at least 25 m. Test solutions (control or alarm odor) were introduced to a lobster by discharging the syringe within 5 cm of the antennules (primary chemosensory organs) over a ten- second time period. Detection of either cue was verified by an increase in antennule flicks. Care was taken not to physically disturb either a focal lobster or nearby conspecifics by approaching each lobster from out of sight (behind the den) and only displaying the clear syringe to denning individuals. Once the odor was introduced, we retreated to just within visual range of the den (5-10 m, depending on visibility) and observed the lobster's behavior for 5 minutes. Focal lobsters exiting dens were followed and observed from 5-10 m away. After 5 minutes, focal individuals were captured, sexed, measured, checked for molt condition and disease, and removed from the site so

78 they would not be tested twice. An individual's location after 5 minutes was marked and the distance moved from the original den was measured. Lobsters infected with late- stage viral disease, known to be less active (Butler et al., 2006), are easily recognized and were not tested. The likelihood of lobsters of different ontogenetic stages leaving dens in response to alarm odor was compared using G-tests. Additionally, within each ontogenetic stage, the likelihood of leaving was compared for lobsters denning communally versus those denning alone. We also determined if the distance each individual traveled was a function of body size by regressing carapace length on distance moved. Too few premolt and postmolt individuals were recovered to statistically test the effects of molt condition on the likelihood of movement.

Laboratory experiments Newly-settled, algal phase Panulirus argus individuals, too cryptic to observe in the field, were tested in laboratory tanks for responses to alarm odor both within the typical shelter of benthic macroalgae and in the open. Recently-settled (less than 1 month on the benthos) P. argus (6 - 10 mm CL) were acquired from floating Witham collectors (on which swimming postlarvae settle and metamorphose) near the Keys Marine Laboratory in May 2002 and August 2005. Lobsters were held communally in 75 L tanks with flow-through seawater and artificial structure (specify) and were regularly fed minced shrimp and squid. To test for responses by algal-phase lobsters within macroalgae, several 5m long, ~ 500 L flow-through seawater troughs were set up with three ~1 L clumps of macroalgae Laurencia spp. collected near the Keys Marine Laboratory. Each clump was examined to insure that no algal phase individuals were already in residence. The center clump was covered completely with a mesh cage. After determining body size, a single individual was added to the center algal clump by gently lifting one side of the cage and allowing the lobster to tailflip back into the algae. Each lobster was then allowed 8 hours to acclimate to the algal clump, within the mesh cage. After acclimation, cages were removed carefully so as not to disturb the algae clumps, the water flow turned off, and lobsters were given 60 ml of either a control seawater odor or an alarm odor. Both

79 control and alarm substances were released using syringes over a 10-second time period all around and into the algal clump. Clumps were then observed for 5 minutes from 2 m away to determine if lobsters responded to the odor by moving from shelter out over the open tank bottom to the other clumps. Because it was not possible to observe algal-phase lobster responses to alarm odor directly within an algal clump, due to their small size and crypticity, we also observed them in the open. First stage (6-7 mm CL) algal phase P. argus were transferred to small, shelterless experimental arenas with flow-through seawater. After a 1-day acclimation period in these experimental arenas, the flow was turned off to the arenas and each individual received 2 mL of either a seawater control or an alarm odor cue prepared as before. Lobster behavior was recorded from above the test tank by a digital video camera (Sony DSC-TRV9) for a 5-minute test period. Treatments were randomly assigned to each lobster, which were used once, then released. The following behaviors were later recorded from videotapes using an event timer computer program (courtesy K. Smith): 1) Time spent with the telson extended or tucked, 2) time spent in locomotion, 3) time spent moving legs and antennae without locomoting, 4) time spent completely still, and 5) antennal position. The durations of each type response were compared between control and treatment groups using t-tests (N = 24).

RESULTS

In the field, 78.4% of nomadic phase P. argus individuals (N=37) treated with alarm odor responded by exiting shelter within minutes and walking quickly away from the den (Fig. 4.1). This is likely an underestimate of the total proportion of lobsters that exited dens in response to alarm odor. We were unable to track non-focal, co-resident lobsters in a den, which were typically absent when we returned to the focal den after 5 minutes. Focal individuals that left dens typically did not enter the nearest available den (only 10.4%) but traveled as far as 19.3 meters (mean 9.9 m, median 10.5 m) during the 5-minute test period. Additionally, many nomadic phase individuals (46.6%) were captured in the open while still traveling away from the original den. All of the focal lobsters passed within a meter of seemingly appropriate dens without attempting to enter

80 them. Nomadic phase lobsters were significantly more likely to leave dens than post- algal lobsters (Pearson’s χ2 = 12.959, P = .003). Only one nomadic phase lobster (1 of 31, 1.5 m dist.) and one post-algal phase lobster (1 of 28; 0.2 m dist.) responded to the seawater control by exiting the den, indicating that the fluid delivery action and human presence were insufficient to elicit the observed response to alarm odor. Only 34.5 % of post-algal phase P. argus, 40 mm CL and smaller (N=29), abandoned dens in response to alarm odor. Of the 10 individuals that moved, 30% moved to the next available den, and 70% had found shelter within 5 minutes of receiving the odor. Most of the post-algal phase lobsters that moved (70%) were also larger than 30 mm CL. While observing non-moving lobsters in dens, we also noted that a variety of small fish (too small to be predators of post-algal or nomadic phase P. argus), including snappers, Bermuda chubs, and pinfish, were attracted to the site of alarm odor release. Snappers (Lutjanus griseus), known juvenile lobster predators, attracted to P. argus body fluid were usually observed with dark stripes running up diagonally from the snout, through the eye, commonly called “feeding” stripes. This eye-stripe was similar to that seen when feeding captive snappers in the laboratory (PEB, pers. obs.). Lobsters remaining in dens typically lashed their antennae at these fish. The linear regression analysis yielded a significant regression for distance moved with lobster size (Table 4.1). The regression model indicated that the slope was significantly different than zero, while the intercept was not (Table 4.1). Larger lobsters typically traveled longer distances than smaller animals (Fig. 4.2), moving an additional 0.3 m further for every mm increase in carapace length. Body size accounted for 31% of the variance in distance moved. All but two post-algal phase individuals that left dens traveled less than 5 meters. Nearly 80% of nomadic phase lobsters leaving dens traveled farther than 5 meters. Interestingly, after leaving the den, all moving lobsters appeared to avoid contact with conspecifics. Only one lobster (post-algal phase) of the 39 that moved entered an occupied den after leaving the original den. Additionally, although groups of P. argus often travel in queues in the open (Herrnkind, 1969), no queues were observed when multiple lobsters moved from a den, even though these animals often left the den within seconds of each other and were commonly seen walking close together (1-2 m).

81 Nomadic phase lobsters were more likely to leave dens after detecting alarm odor in the presence of conspecifics than when alone (Pearson’s χ2 = 4.194, P = .0405). On the other hand, the likelihood that post-algal lobsters exited or stayed in dens did not statistically differ in the presence of conspecifics (Pearson’s χ2 = 2.885, P = .0862). We observed on two occasions, when nomadic and post-algal phase lobsters (stage visually estimated) were co-denning, that nomadic phase lobsters left a shared den in response to alarm odor while adjacent post-algal phase individuals stayed. Algal phase lobsters in the large laboratory tanks did not leave algae clumps for the open or travel the short distance to another algae clump upon release of either the alarm odor (N=16) or seawater control (N=14). In the smaller arenas without algae, algal phase lobsters increased their time spent in locomotion from a mean of 25.1 ± 10.9 (S.E.) to 91.2 ± 19.7 seconds (t = 2.93, P= .006) (Fig. 4.3). Additionally, algal phase lobsters increased their time spent performing other, non-locomotory movements from 70.4 ± 16.3 to 185.6 ± 19.0 seconds (t = 4.60, P = .00004) (Fig. 4.3). None of the durations of postures or antennal positions differed significantly between control and test animals.

DISCUSSION

Algal, post-algal, and nomadic phase Panulirus argus individuals all detect and respond to alarm odor, albeit in different ways. Although this study is the first to demonstrate an immediate alarm response by a spiny lobsters to conspecific body fluids, other evidence indicates that conspecific alarm signals are an important signal for selecting shelters. Briones-Fourzan et al. (2006) note that fishermen in the Mexican Caribbean remove tails from lobsters while still at sea but avoid discarding the cephalothoraxes back into the water because they claim this reduces subsequent catches at that site. These authors also demonstrate by an overnight Y-maze shelter choice experiment, that P. argus individuals avoid dens with water emanating from a head tank containing a killed conspecific (Briones-Fourzan et al. 2006). Along with our results, this indicates that alarm odor not only elicits an immediate move from a den in nomadic P. argus but may also make a return to an abandoned den unlikely.

82 Our results also strongly support the existence of an ontogenetic shift in the stay- or-leave response to alarm odor in juvenile P. argus at the transition from the post-algal to nomadic phase (~40 mm CL). Nomadic phase juveniles abandoned dens and left the immediate vicinity much more frequently than post-algal phase lobsters. This result was puzzling considering that nomadic phase lobsters tethered in shelter show significantly lower predation rates than animals tethered in the open (Eggleston et al., 1990). Additionally, we have noted elsewhere (Herrnkind et al., 2001; Bouwma, Chapter 1) that when nomadic P. argus are attacked in the open during the day by grey triggerfish (Balistes capriscus), individual lobster are usually defeated by bites to the eyes or exposed carapace, areas of the body that are well-protected in shelter. On the other hand, nomadic phase P. argus are very effective at keeping triggerfish at bay for 10-16 minutes on average when attacked (PEB, unpublished data). In hardbottom habitat with numerous structures (sponges, coral heads, rock crevices), 10-16 minutes potentially is long enough to find an appropriate shelter. Regardless, lobsters frequently suffer bleeding wounds early in an attack (Bouwma, Chapter 1), which raises the question of why a response to leave a den and travel alone across open substrate during the day, something rarely seen in crevice-dwelling crustaceans, would evolve in response to alarm odor. The nature of information contained in alarm odor might help explain the decision to leave in juvenile P. argus. Conspecific alarm odor indicates 1) that a potentially dangerous predator is nearby and 2) that a predator has already injured or killed a conspecific (Kats & Dill, 1998). A predator is potentially very dangerous to nearby conspecifics if it injures a lobster in a shelter, despite its defendability. Even if a predator is sated after killing a single lobster, other predators may be attracted to the lobster body fluid (Mathis et al., 1995; Chivers et al., 1996). During laboratory experiments, we noticed that large (2-3 m) nurse shark Ginglymostoma cirratum, held in the adjacent downstream section of the channel enclosure at the Keys Marine Lab, responded to discarded lobster alarm fluid by rapid swimming and circling (PEB, pers. obs.). This resembled their response to the fish, shrimp, and squid fed to them, suggesting that freshly prepared lobster body fluid triggers search for food. A large nurse shark, with its strong suction and crushing capacity, can successfully attack lobsters in dens during the

83 day (Childress, 1995). In this case, immediately leaving the area potentially reduces predation risk by comparison, despite exposure. Further work is necessary to determine whether solitary nomadic phase P. argus staying in dens, despite detecting alarm odor, suffer higher mortality than lobsters abandoning dens. We did not expect nomadic phase P. argus, exposed to alarm odor to leave dens when conspecifics were present. Groups of three lobsters, or more, potentially benefit from both statistical dilution and cooperative defensive behavior, which can be very effective versus predators like grey triggerfish Balistes capriscus (Herrnkind et al., 2001). Although dilution decreases risk whenever other potential victims are nearby, cooperative defense may be less effective against large predators like nurse sharks, rays, etc. than against triggerfish. Denning with conspecifics may also offset individual crypticity, if bigger groups of uninjured conspecifics produce a larger chemical signature (Ratchford & Eggleston, 1998) or are easier to identify visually (Wrona & Dixon, 1991; Uetz & Hieber, 1994). If so, abandoning a multiply-occupied den and leaving the area may decrease the likelihood of encountering a predator. By contrast, the presence of agitated conspecifics did not affect the likelihood of abandoning dens by post-algal phase P. argus. Even when co-denning nomadic conspecifics left in response to alarm odor, post-algal phase lobsters did not leave the den more often than when denning alone. In mesocosm experiments, Childress (1995) found that post-algal phase lobsters more frequently occupied single dens after a nurse shark predator killed a nearby conspecific. This suggests that post-algal P. argus prefer to be alone and in dens when predators are active nearby. After conspecifics have all left the den, there would seem to be little additional incentive for these individuals to move. Body size explained about one third of the variance in distance moved by P. argus once they exited the den. Although larger lobsters in general can walk faster than smaller lobsters, in this study, the individual that traveled farthest measured only 42 mm CL. Distance traveled likely also depended on the time spent walking versus pausing, how soon after receiving the cue an individual exited its den, if it traveled in a straight line, and how soon it stopped (if it stopped at all). The fact that larger nomadic phase lobsters traveled farther than post algal counterparts reflect the general movement patterns of juvenile lobsters. During nightly foraging, post-algal lobsters are typically

84 restricted (<25m, Butler & Herrnkind, 2000) and individuals may reside in an area for weeks or months. Nomadic phase P. argus tend to move longer distances as part of ~15- 30 km (or longer) ontogenetic and mass migrations from the bay to reef and likely move into and out of new, unfamiliar areas frequently. Algal phase P. argus within algal clumps in the laboratory failed to leave shelter in response to the alarm cue; however, individuals tested without shelter increased locomotion compared to control periods. These latter individuals were likely seeking shelter, the primary defensive strategy for P. argus at this size. Considering that algal phase lobsters are rarely seen outside of shelter during the daytime, these results are not surprising. In fact, this stage is so cryptic and elusive that, despite extensive research on P. argus before the mid 1980’s, scientists were unsure where this early benthic juvenile stage lived until 1983 (Marx, 1983). Panulirus argus suffers a particularly high mortality rate during the algal stage, with less than 5% of individuals surviving to make the transition to the post-algal phase (Butler et al., 1997; Herrnkind et al. 1997). Predation is strongly inferred to be the primary source of this high mortality rate (Butler & Herrnkind, 2000), primarily by fishes such as those attracted to the release of alarm odor in the field (Smith & Herrnkind, 1992). This strongly suggests that any algal phase P. argus caught in the open or leaving macroalgal shelter in response to alarm odor would likely encounter a wide array of lobster predators (Smith & Herrnkind, 1992) attracted to the odor and ready to feed. Tethering experiments of algal phase lobster away from shelter support this finding by indicating that these individuals are much more vulnerable to predation than those remaining within the algae (Herrnkind & Butler, 1986). In Panulirus argus increasing age and/or body size apparently can influence the decision to stay in a den or leave when confronted with alarm odor in, particularly in the crevice-dwelling post-algal and nomadic stages. Both life-stages live in the same habitats, share shelter, and face the same guild of predators. Their defensive morphology does not change, other than enlargement (Bouwma, Chapter 1), neither denning or social behavior change, and neither stage is reproductive. The only substantial change between these two juvenile stages is body size and the increased passive and active defensive capacity it engenders. It is not known whether adults, which live in crevices on the reef, leave dens in response to alarm odor. Considering that Herrnkind’s (1975) observations

85 of lobsters switching dens after handling were of adults, it seems plausible that these individuals would also respond strongly to alarm odor. However, very large adult lobsters likely have few predators, which begs the question of when and if they stop responding to alarm cues. Further work will need to be done to address this question. The response by lobsters to alarm odor resembles their response to handling stress by humans (capturing, prodding, grasping the carapace, etc.) (Herrnkind, 1975). This suggests that the leaving response in nomadic P. argus might be triggered by high stress in general, mediated by either mechanical or chemical cues, but not visual cues alone. If so, the large-scale movements by P. argus in response to recreational fishing pressure are not surprising. Recreational fishing such as during the annual 2-day sport lobster season in the Florida Keys results in high injury rates for nomadic phase lobsters. In choice experiments, lobsters were less likely to choose dens in a Y-maze receiving water from a head tank containing an injured conspecific (Parsons & Eggleston, 2005). Additionally, Parsons & Eggleston (2006) found, in mesocosms, that injury to a conspecific at an autotomy joint increased the likelihood that conspecifics switch shelters by the following day. Our results suggest that alarm odor is released causing immediate movement by lobsters out of shelters and not necessarily to the next available crevice. Repeated disturbance, likely during high, sustained fishing pressure over the two-day period might result in lobsters moving out in the open, without nearby conspecifics, such as during mass migrations. Such a situation might result in higher mortality rates for sub-legal (i.e., juvenile) individuals.

86 Table 4.1. Results of the regression model for distance moved on lobster body size (carapace length in mm).

N = 39, adjusted R2 = 0.310 Source DF SS MS F P Model 1 401.55 401.55 18.075 <0.001 Error 37 821.97 22.22 Total 38 1223.51

Parameter Estimate Standard Error t-value P

Intercept -5.120 3.221 -1.59 0.120 Slope 0.278 0.066 4.25 0.001

87 (a)

100 Move

90 Stay

80 ***

Percent of Lobsters Lobsters Percent of 30

0 NomadicNomadic PhasePhase Post-algal Post-algal Phase Phase

(b)

100 Move

90 Stay * 80

Percent of Lobsters Lobsters Percent of 30

0 Alone Conspecifics Alone Conspecifics Alone Conspecifics Alone Conspecifics Nomadic Phase Post-algal Phase Figure 4.1. (a) Percent of nomadic phase and post-algal P. argus individuals moving from dens or staying in response to alarm odor. (b) Percent of nomadic phase and post- algal phase lobsters moving from dens or staying when sheltering alone or with conspecifics. Asterisks indicate significant differences in frequencies (* P ≤ 0.05; ** P ≤ 0.01, *** P ≤ 0.001).

88 20.0

18.0

16.0

14.0 in 5 min ) m ( 12.0

10.0

8.0

6.0 Distance Traveled Distance Traveled

4.0

2.0

0.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0

Carapace Length (mm)

Figure 4.2. Distance moved (m) versus body size (mm carapace length) by nomadic and post-algal phase P. argus individuals abandoning natural dens in response to conspecific alarm odor.

89 300 Control

Alarm Odor 229.4 250 209.9 *** 185.6 170.1181.4 200

150 Seconds ** 91.2 100 70.4

50 25.1

0 Locomotion Locomotio n Movements Movement s AntennaeAntennae Forward Forward ElevatedElevated PosturePosture

90 CONCLUSIONS

Antipredator life-styles in the Palinuridae Predators can dramatically shape the evolution of prey life-styles, affecting, among other traits, life-history, morphology, and behavior (Sih, 1987). My results suggest that the benthic stages of the sympatric spiny lobsters Panulirus argus and P. guttatus have evolved very different life-styles for countering threats from the same suite of predators. Crypticity/retreat: Panulirus guttatus exhibits a cryptic life-style, which reduces predator encounters. By settling directly from the water column to the reef, Panulirus guttatus avoids most of the increased risk P. argus faces during its long ontogenetic migrations. Additionally, P. guttatus likely avoids most encounters with diurnal predators by residing deep within dens and rarely venturing into the open during the day. In adopting this cryptic, avoidance-only lifestyle, P. guttatus has also apparently lost (or failed to acquire) the morphological and defensive capacity to effectively defend itself in the open. However, these lobsters do retain strong legs (Briones-Fourzan et al., 2006) and a well-calcified exoskeleton (PEB, pers. obs.), which may be important for escaping predators. When attacked by triggerfish, P. guttatus only attempt escape, failing to utilize the aggressive, lunging defense exhibited by P. argus, which provides greater survival time and opportunity for evasion. Morphologically, the smaller, more flexible antennae of P. guttatus seem functionally inferior for retaliatory defense, although possibly superior for moving about the interstices of the reef. P. guttatus also use stridulation only in escaping a predator’s grasp. Furthermore, spotted lobsters grow more slowly and do not attain the high degree of size refuge from predators seen in P. argus. Confrontation/preemption: Panulirus argus have adopted a riskier, more exposed life-style, which includes more frequent encounters with predators but has evolved mechanisms to survive these encounters. Subadults migrate 10-50 km offshore over several months and longer to oceanic reefs from nursery habitats (Herrnkind, 1980). Adults reside by day near the den opening and move across open substrate in daylight as well as nightly (Butler et al., 2006). Unlike P. guttatus, P. argus appears to have

91 retained/attained behavioral and morphological tools to counter the increased risk from predators that such behavior entails. P. argus more effectively defend through retaliation with the antennae against attacks by diurnally-hunting triggerfish. By incorporating acoustic aposematism into their defensive repertoire, P. argus also reduces subsequent attacks by a triggerfish after its initial ineffective encounter. In addition to effective antipredator behaviors, P. argus also have substantial morphological defenses, especially the robust, spine-laden and tough 2nd antennae, which typically must be bitten away piecemeal by an attacking triggerfish before a lobster can be subdued. The effectiveness of withstanding attack by retaliatory defense in the open may explain why nomadic phase P. argus risk leaving shelter where a conspecific is injured. That is, when lobsters are attacked by predators in the open, aggressive defense and tailflip escapes probably “buy” enough time to locate and enter other shelter. In addition to solitary defense, Herrnkind et al. (2001) show that aggregations of P. argus defending cooperatively are even more effective in repelling attacks by triggerfish. Although their respective antipredator life-styles and tactics presumably allow both P. argus and P. guttatus to survive predators during the juvenile period and reproduce, there may be evolutionary advantages to the riskier life-style typified by P. argus. This species grows much larger than P. guttatus, has a greater geographic range (one of the largest among crustaceans), and is far more abundant. P. argus is fished throughout its range and, on the southwestern coast of Cuba, supports the 2nd largest lobster fishery in the world. On the other hand, P. guttatus is too uncommon to support a major commercial fishery. The lesser reproductive capacity (due to small female body size), lower abundance, and limited range of habitats suggest that P. guttatus pays a greater price, in terms of fitness, for its type of antipredation than does P. argus. The two antipredator life-styles exhibited here apparently represent opposite ends of a continuum for antipredation in the Palinuridae. However, the information required to assemble complete descriptions of antipredator life-styles does not exist for most palinurid species. Descriptions of other bold, abundant, shallow-water species exposed at the entrances to dens, on open substrate, or elsewhere during the daytime besides P. argus (P. interruptus Lindberg, 1955; P. homarus Berry, 1971, P. cygnus Cobb, 1980; P. penicillatus Holthuis & Loesch, 1967; P. marginatus MacDonald et al., 1984; Jasus

92 edwardsii Kelly et al., 1999; Palinurus elephas Giacalone et al., 2006) are understandably more common in the literature than descriptions of “shyer” lobsters such as P. guttatus (Sharp et al., 1997) and P. gracilis (Holthuis & Loesch, 1967). Other than morphological characteristics (poorly calcified exoskeleton, weak legs), virtually nothing is known about the antipredator life-styles of deep-dwelling species of Linuparus, Puerulus, Palinustus, and Palibythus (George, 2005). Of particular interest are species with life-styles and antipredation features falling between the extremes exhibited by P. argus and P. guttatus, i.e. lobsters with robust morphological defenses but less exposure- prone behaviors or life-histories, or bold, highly-exposed species with less-robust weaponry. I propose further behavioral study of such species in the field and in large, naturalistic enclosures to better assess how morphological defenses, antipredator behaviors, and life-history traits have evolved in palinurid antipredator life-styles.

Spiny lobster stridulation: a special case

Sound production by Panulirus argus juveniles played a role in survival during encounters with two very different predator types, diurnally-active triggerfish and nocturnally-hunting octopus. That stridulation had an effect on survival largely explains why the organ has been so highly conserved in strident palinurids (George & Main, 1967; Patek & Oakley, 2003). My results suggest that shallow-water palinurid species of the silentes clade, which lack the stridulating organ (e.g., Jasus edwardsii, Sagmariasus verrauxi, and J. lalandii), are at a defensive disadvantage compared to stridentes lobsters when defending in the open. Alternatively, J. edwardsii and S. verrauxi may compensate for the lack of the acoustic device by attaining both very large size and extremely robust spination (Butler et al., 2006). However, maximum body size per se does not address antipredation by juveniles of these species and of J. lalandii, which does not reach as large a size. Although temperate species in general have fewer predatory species to contend with, octopus and fish predators are still likely to be important in both Jasus and Sagmariasus (Booth, 2006). The next step in determining how stridulating evolved in the Palinuridae will require empirical studies of defensive behavior and associated sound production (or lack thereof) against natural predators for other species and genera in silent, strident, shallow-,

93 and deep-water taxa. In revealing the mechanisms and function of this important acoustic defensive tool in the ecologically variable Stridentes species, and how Silentes lobsters cope without it, the study of lobster stridulation should also provide insight into the evolution and defensive role of acoustic devices generally, in marine systems and elsewhere.

94 REFERENCES

Acosta CA, Butler MJ IV (1999) Adaptive strategies that reduce predation on Caribbean spiny lobster postlarvae during onshore transport. Limnol Oceanogr 44: 494-501.

Alexander RD (1967) Acoustical communication in arthropods. Annu Rev Entomol 12: 494-526.

Alexander RM (1983) Animal Mechanics. Seattle, WA: University of Washington Press.

Atema J, Cobb JS (1980) Social Behavior, pp. 409-450. In: Cobb JS, Phillips BF edits. The Biology and Management of Lobsters, Volume I: Physiology and Behavior. New York: Academic Press.

Atema J, Voigt R (1995) Behavior and sensory biology, pp. 313-348. In: Factor JR edit. The Biology of the Lobster, Homarus americanus. New York: Academic Press.

Barkai A, McQuaid C (1988) Predator-prey role reversal in a marine benthic ecosystem. Science 4875: 62-64.

Barshaw DE, Lavalli KL, Spanier E (2003) Offense versus defense: responses of three morphological types of lobsters to predation. Mar Ecol Prog Ser 256: 171-182.

Becker K, Wahl M (1996) Behaviour patterns as natural antifouling mechanisms of tropical marine crabs. J Exp Mar Biol Ecol 203: 245-258.

Benard MF (2004) Predator-induced phenotypic plasticity in organisms with complex life histories. Ann Review Ecol Evol Syst 35:651-673.

Berger DK, Butler MJ IV (2001) Octopuses influence den selection by juvenile Caribbean spiny lobster. Mar Freshw Res 52: 1049-1054.

Berrill M (1976) Aggressive behaviour of post-puerulus larvae of the western rock lobster Panulirus longipes (Milne Edwards). Aust J Mar Freshw Res 27: 83-88.

Berry PF (1971) The spiny lobsters (Palinuridae) of the east coast of southern Africa: distribution and ecological notes. S Afr Assoc Mar Biol Res Bull 27: 1-23.

Bertelson RD, Matthews TR (2001) Fecundity dynamics of female spiny lobster (Panulirus argus) in a south Florida fishery and Dry Tortugas National Park lobster sanctuary. Mar Freshw Res 52: 1559-1565.

95 Bertram BCR (1978) Living in groups: predators and prey, pp. 64-94. In, Behavioural Ecology: An Evolutionary Approach, Edits. Krebs, J. R. & Davies, N. B. Oxford: Blackwell Scientific.

Bill RG, Herrnkind WF (1976) Drag reduction by formation movement in spiny lobsters. Science 193: 1146-1147.

Blanckenhorn WU (2000) The evolution of body size: What keeps organisms small? Q Rev Biol 75:385-407.

Bollens SM, Frost BW, Thoreson DS, Watts SJ (1992) Diel vertical migration in zooplankton: Field evidence in support of the predator avoidance hypothesis. Hydrobiologia 234:33-39.

Booth JD (2006) Jasus species, pp. 340-358. In: Phillips, B. F. edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

Boyle PR (1997) Octopus interactions with crustacean fisheries, pp. 125-129. In: Lang, MA, Hochberg FG, Ambrose RA, Engle JA edits. Workshop on the Fishery and Market Potential of Octopus in California. Washington, D.C.: Smithsonian Institution Office of the Provost Scientific Diving Program.

Brawn VM (1961) Sound production by the cod Gadus morhua. Behaviour 18, 239–255.

Briones-Fourzan P, Perez-Ortiz M, Lozano-Alvarez E (2006) Defense mechanisms and antipredator behavior in two sympatric species of spiny lobsters, Panulirus argus and P. guttatus. Mar Biol 149:227-239.

Brock TJ, Ward TM (2003) Maori octopus (Octopus maorum) bycatch and southern rock lobster (Jasus edwardsii) mortality in the South Australian rock lobster fishery. Fish Bull 102: 430-440.

Butler MJ IV, Herrnkind WF (2000) Puerulus and Juvenile ecology, pp 276-301. In: Phillips BF, Kittaka J edits. Spiny Lobsters: Fisheries and Culture 2nd Ed. London, UK: Blackwell Scientific Publications.

Butler MJ IV, Herrnkind WF (2000) Puerulus and Juvenile ecology, pp 276-301. In: Phillips, BF; Kittaka, J., Edits. Spiny Lobsters: Fisheries and Culture 2nd Ed. London, UK: Blackwell Scientific Publications.

Butler MJ IV, Herrnkind WF, Hunt JH (1997) Factors affecting the recruitment of juvenile Caribbean spiny lobsters dwelling in macroalgae. Bull Mar Sci 61: 3-19.

Butler MJ IV, Steneck RS, Herrnkind WF (2006) Juvenile and Adult Ecology, pp. 263- 309. In: Phillips, B. F. edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

96 Caddy JF (1986) Modeling stock recruitment processes in Crustacea - Some practical and theoretical perspectives. Can J Fish Aquatic Sci 43:2330-2344.

Campbell HW (1973) Observations on acoustic behavior of crocodilians. Zoologica 58: 1-11.

Caro, TM (1995) Pursuit-deterrence revisited. TREE 10: 500-503.

Caro TM, Graham CM, Stoner CJ, Vargas JK (2004) Adaptive significance of antipredator behaviour in Artiodactyls. Anim Behav 67: 205-228.

Casper BM, Lobel PS, Yan HY (2003) The hearing sensitivity of the little skate, Raja erinacea: A comparison of two methods. Environ Biol Fish 68: 371-379.

Casper BM, Mann DA (2006) Evoked potential audiograms of the nurse shark (Ginglymostoma cirratum) and the yellow stingray (Urobatis jamaicensis). Environmental Biology of Fishes 76:101-108.

Childress MJ (1995) The ontogeny and evolution of gregarious behavior in juvenile Caribbean spiny lobster, Panulirus argus. Dissertation, Florida State University, Tallahassee, FL.

Childress MJ, Herrnkind WF (1996) The ontogeny of social behaviour among juvenile Caribbean spiny lobsters. Anim Behav 51:675-687.

Childress MJ, Herrnkind WF (2001) Influence of conspecifics on the ontogenetic habitat shift of juvenile Caribbean spiny lobsters. Mar Freshw Res 52:1077-1084.

Childress MJ, Jury SH (2006) Behavior, pp. 78-112. In: Phillips BF edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

Chivers DP, Brown GE, Smith RJF (1996) The evolution of chemical alarm signals: Attracting predators benefits alarm signal senders. Am Nat 148:649-659.

Chu M (2001) Heterospecific responses to scream calls and vocal mimicry by phainopeplas (Phainopepla nitens) in distress. Behaviour 138: 775-787.

Claridge MF (1974) Stridulation and defensive behavior in the ground beetle, Cychrus caraboides (L). J Entomol A 49:7-15.

Cleveland DW (2002) Factors influencing the establishment of dominance hierarchies of the grey triggerfish Balistes capriscus. Masters thesis. Southwest Texas State University, San Marcos, TX.

Cloudsley-Thompson JL (1995) A review of the anti-predator devices of spiders. Bull Br Arachnol Soc 10:81-96.

97 Cobb JS (1980) Behavior of the Western Australian spiny lobster, Panulirus cygnus (George), in the field and laboratory. Austr J Mar Freshw Res 32: 399-409.

Cobb JS, Wang D (1985) Fisheries biology of lobsters and crayfishes. In: Provenzano, A. J. edit. The Biology of Crustacea, Vol. 10, pp. 167-247. New York: Academic Press.

Cruz R, Phillips BF (2000) The artificial shelters (pesqueros) used for the spiny lobster (Panulirus argus) fisheries in Cuba, pp. 400-419. In: Phillips, BF; Kittaka, J., Edits. Spiny Lobsters: Fisheries and Culture 2nd Ed. London, UK: Blackwell Scientific Publications.

Curio E (1978) Adaptive significance of avian mobbing. 1. Teleonomic hypotheses and predictions. Z Tierpsychol 48: 175-183.

De Cock R, Matthysen E (2001) Do glow-worm larvae (Coleoptera: Lampyridae) use warning coloration? Ethology 107: 1019-1033.

Derby CD, Steullet P, Horner AJ, Cate HS (1997) The sensory basis of feeding behavior in the Caribbean spiny lobster, Panulirus argus. Mar Freshw Res 52: 1339-1350.

Dicke M, Grostal P (2001) Chemical detection of natural enemies by arthropods: An ecological perspective. Ann Rev Ecol Syst 32: 1-23.

Driver PM, Humphries DA (1988) Protean behaviour: the biology of unpredictability. Oxford: Clarendon Press.

Duffy JE, Morrison CL, Macdonald KS (2001) Colony defense and behavioral differentiation in the eusocial shrimp Synalpheus regalis. Behav Ecol Sociobiol 51:488–495.

Dunning DC, Kruger M (1995) Aposematic sound in African moths. Biotropica 27: 227- 231.

Edmunds M (1974) Defense in Animals. White Plains, New York: Longman Publishing Group.

Eilam D (2005) Die hard: A blend of freezing and fleeing as a dynamic defense - implications for the control of defensive behavior. Neurosci Biobehav Rev 29:1181-1191.

Eggleston DB, Lipcius RN, Miller DL, Cobacetina L (1990) Shelter scaling regulates survival of juvenile Caribbean spiny lobster Panulirus argus. Mar Ecol Prog Ser 62:79-88.

Endler JA (1988) Frequency-dependent predation, crypsis and aposematic coloration. Philos Trans R Soc London [Biol] 319: 505-523.

98 Exnerova A, Landova E, Stys P, Fuchs R, Prokopova M, Cehlarikova P (2003) Reactions of passerine birds to aposematic and nonaposematic firebugs (Pyrrhocoris apterus; Heteroptera). Biol J Linnean Soc 78: 517-525.

Fish JF (1966) Sound production in the American lobster Homarus americanus. Crustaceana 11: 105.

Frid A, Dill L (2002) Human-caused disturbance stimuli as a form of predation risk. Conserv Ecol 6: Art. No. 11.

George RW (1972) South Pacific Islands – Rock Lobster Resources. Rep S Pac Is Fish Dev Agency Rome UN, FAO: 1-42.

George RW (2005) Evolution of life cycles, including migration, in spiny lobsters (Palinuridae). N Z J Mar Freshw Res 39: 503–514.

George RW, Main AR (1967) The evolution of spiny lobsters (Palinuridae): A study of evolution in the marine environment. Evolution 21: 803-820.

Giacalone VM, D’Anna OG, Pipitone C, Badalamenti, F (2006) Movements and residence time of spiny lobsters, Palinurus elephas released in a marine protected area: an investigation by ultrasonic telemetry. J Mar Biol Ass UK 86: 1101-1106.

Giancoli, D. C. 1998. Physics principles with applications, 5th ed. Upper Saddle River, New Jersey: Prentice Hall.

Gittleman JL, Harvey PH (1980) Why are distasteful prey not cryptic? Nature 286: 149– 150.

Groeneveld JC, Goni R, Latrouite D (2006) Palinurus species, pp. 385-411. In: Phillips BF edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

Gorzula S (1985) Are caimans always in distress? Biotropica 17: 343-344.

Gregory DL Jr., Labisky RF (1986) Movements of the spiny lobster Panulirus argus in South Florida. Can J Fish Aquatic Sci 43: 2228-2234.

Guilford T, Cuthill I (1991) The evolution of aposematism in marine gastropods. Evolution 45: 449-451.

Hamilton WD (1971) Geometry for the Selfish Herd. J Theor Biol 31: 295-311.

Hanlon RT (1983) Octopus briareus, pp. 251-265. In, Boyle P edit. Cephalopod life cycles. Vol.1. London: Academic Press.

Hanlon RT, Budelmann B -U (1987) Why cephalopods are probably not “deaf.” Am Nat 129: 312-317.

99 Hauglund K, Hagen SB, Lampe HM (2006) Responses of domestic chicks (Gallus gallus domesticus) to multimodal aposematic signals. Behavioral Ecology 17:392-398.

Hazlett BA, Winn HE (1962a) Characteristics of a sound produced by the lobster Justitia longimanus. Ecology 43: 741-742.

Hazlett BA, Winn HE (1962b) Sound production and associated behavior of Bermuda crustaceans (Panulirus, Gonodactylus, Alpheus, and Synalpheus). Crustaceana 4: 25-39.

Hazlett BA, Bach CE, McLay C, Thacker RW (2000) A comparative study of the defense syndromes of some New Zealand marine Crustacea. Crustaceana 73: 899-912.

Hazlett BA, Schoolmaster DR (1998) Responses of cambarid crayfish to predator odor. J Chem Ecol 24: 1757-1770.

Henninger HP, Watson WH III (2005) Characterization of Carapace Vibrations and Waterborne Sound Production and their Underlying Mechanism in the American Lobster, Homarus americanus. J Exp Biol, In Press.

Herrnkind WF (1969) Queuing behaviour of spiny lobsters. Science 164: 425-1427.

Herrnkind WF, Butler MJ IV (1986) Factor regulating settlement and microhabitat use by juvenile spiny lobsters, Panulirus argus. Mar Ecol Prog Ser 34: 23-30.

Herrnkind WF, VanderWalker J, Barr L (1975) Population dynamics, ecology, and behavior of spiny lobster, Panulirus argus, of St. John, U.S. Virgin Islands: Habitation and pattern of movements. Results of the Tektite Program, Vol. 2. Sci Bull Nat Hist Mus Los Angeles Cty 20:31-45.

Herrnkind WF (1980) Spiny lobsters, patterns of movement, pp. 349-407. In: Cobb JS, Phillips BF edits. The Biology and Management of Lobsters, Volume I: Physiology and Behavior. New York: Academic Press.

Herrnkind WF (1985) Evolution and mechanisms of mass single-file migration in spiny lobster: synopsis. Contrib Mar Sci (Special Symposium Volume) U Texas 27: 197-211.

Herrnkind WF, Butler MJ, Hunt JH (1997) Can artificial habitats that mimic natural structures enhance recruitment of Caribbean spiny lobster? Fisheries 22: 24-27.

Herrnkind WF, Childress MJ, Lavalli KL (2001) Cooperative defense and other benefits among exposed spiny lobsters: inferences from group size and behaviour. Mar Freshw Res 52: 1113-1124.

Heydorn AEF (1969) Notes on the biology of Panulirus homarus and on length/weight relationships of Jasus lalandii. S Afr Div Sea Fish Invest Rep 69: 1-26.

100 Högstedt G (1983) Adaptation unto death: Function of fear screams. Am Nat 121: 562- 570.

Holthuis LB (1991) FAO Species catalogue, Vol. 13. Marine lobsters of the world. Food and Agriculture Organization of the United Nations, Rome, Italy.

Holthuis LB, Loesch H (1967) The lobsters of the Galapagos islands (Decapoda: Palinuridae). Crustaceana 12: 214-224.

Honma A, Oku S, Nishida T (2006) Adaptive significance of death feigning posture as a specialized inducible defence against gape-limited predators. Proc R Soc Lond [Biol] 273: 1631-1636.

Howard RK (1988) Fish predators of western rock lobster (Panulirus cygnus George) in nearshore nursery habitat. Austr J Mar Freshw Res 39: 307-316.

Inbar M, Lev-Yadun S (2005) Conspicuous and aposematic spines in the animal kingdom. Naturwissenschaften 92:170–172.

Jereb P, Roper CFE, Vecchione M (2005) Introduction, pp. 1–19. In, Jereb P, Roper CFE edits. Cephalopods of the world. An annotated and illustrated catalogue of species known to date. Volume 1. Chambered nautiluses and sepioids (Nautilidae, Sepiidae, Sepiolidae, Sepiadariidae, Idiosepiidae and Spirulidae). FAO Species Catalogue for Fishery Purposes. No. 4, Vol. 1. Rome: FAO.

Jetz W, Rowe C, Guilford T (2001) Non-warning odors trigger innate color aversions: as long as they are novel. Behav Ecol 12: 134-139.

Joll LM (1977) The Predation of Pot-caught Western Rock Lobster (Panulirus Longipes cygnus) by Octopus. Department of Fisheries and Wildlife, Western Australia. Report No. 29.

Kanciruk P (1980) Ecology of juvenile and adult Palinuridae (spiny lobsters), pp. 59-96. In: Cobb JS, Phillips BF edits. The Biology and Management of Lobsters, Volume I: Physiology and Behavior. New York: Academic Press.

Kanciruk P, Herrnkind WF (1978) Mass migration of spiny lobster, Panulirus argus (Crustacea: Palinuridae): behavior and environmental correlates. Bull Mar Sci 208: 601-623.

Kats LB, Dill LM (1998) The scent of death: Chemosensory assessment of predation risk by prey animals. Ecoscience 5: 361-394.

Kelly S, MacDiarmid AB, Babcock RC (1999) Characteristics of spiny lobster, Jasus edwardsii, aggregations in exposed reef and sandy areas. Mar Freshw Res 50: 409-416.

101 Klump GM, Shalter MD (1984) Acoustic behaviour of birds and mammals in the predator context. II. Functional significance and evolution of alarm signals. Z Tierpsychol 66: 206-226.

Kroodsma DE, Byers BE (1991) The function(s) of bird song. Am Zool 31: 318-328.

Lampert W (1989) The adaptive significance of diel vertical migration of zooplankton. Funct Ecol 3: 21-27.

Lass S, Spaak P (2003) Chemically induced anti-predator defences in plankton: a review. Hydrobiologia 491: 221-239.

Laurent P, Braekman J-C, Daloze D (2005) Insect chemical defense. Top Curr Chem 240: 167–229.

Lavalli KL, Spanier E (2001) Does gregarious behavior function as an anti-predator mechanism in Mediterranean slipper lobster, Scyllarides latus? Mar Freshw Res 52: 1133–1143.

Lehtiniemi M (2005) Swim or hide: Predator cues cause species specific reactions in young fish larvae. J Fish Biol 66: 1285-1299.

Lev-Yadun S (2003a) Weapon (thorn) automimicry and mimicry of aposematic colorful thorns in plants. J Theor Biol 224: 183–188.

Lev-Yadun S (2003b) Why do some thorny plants resemble green zebras? J Theor Biol 224: 483–489.

Lima SL (1993) Ecological and evolutionary perspectives on escape from predatory attack: a survey of North-American birds. Wil Bull 105: 1-47.

Lima SL (1998) Stress and decision making under the risk of predation: Recent developments from behavioral, reproductive, and ecological perspectives. Adv Stud Behav 27: 215-290.

Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68: 619-640.

Lindberg RG (1955) Growth, population dynamics, and field behavior in the spiny lobster Panulirus interruptus. Univ Calif Publ Zool 59: 157-248.

Lindström L, Rowe C, Guilford T Pryazine odour makes visually conspicuous prey aversive. Proc R Soc Lond [Biol] 268: 159-162.

Lingle S, Pellis SM (2000) Fight or Flight? Antipredator behavior and the escalation of coyote encounters with deer. Oecologia 131: 154-164.

102 Lipcius RN, Eggleston DB (2000) Introduction: Ecology and Fishery Biology of Spiny Lobsters, pp. 1-41. In: Phillips BF, Kittaka J edits. Spiny Lobsters: Fisheries and Culture 2nd Ed. London, UK: Blackwell Scientific Publications.

Long JD, Hay ME (2006) Fishes learn aversions to a nudibranch’s chemical defense. Mar Ecol Prog Ser 307: 199-208.

Lovell JM, Findlay MM, Moate RM, Yan HY (2005) The hearing abilities of the prawn Palaemon serratus. Comp Biochem Physiol, A 140: 89– 100.

Lozano-Alvarez E, Spanier E (1997) Behaviour and growth of captive spiny lobsters (Panulirus argus) under the risk of predation. Mar Freshw Res 48: 707-713.

Luckenbach MW, Orth R.J (1990) A chemical defense in Crustacea? J Exp Mar Biol Ecol 137: 79-87.

MacDonald CD (1982) Predation by Hawaiian monk seals on spiny lobsters. J Mammal 63: 700.

MacDonald CD, Jazwinski SC, Prescott JH (1984) Queuing behavior of the Hawaiian spiny lobster Panulirus marginatus. Bull Mar Sci 35: 111-114.

Magnhagen C (1991) Predation risk as a cost of reproduction. TREE 6: 183-185.

Mallet J, Joran M (1999) Evolution of diversity in warning color and mimicry: polymorphisms, shifting balance, and speciation. Ann Rev Ecol Syst 30: 201-233.

Mann DA, Lobel PS (1998) Acoustic behavior of the damselfish Dascyllus albisella: behavioral and geographic variation. Environ Biol Fish 51: 421–428.

Mappes J, Marples N, Endler JA (2005) The complex business of survival by aposematism. TREE 20: 598-603.

Marshall SD, Thoms EM, Uetz GW (1995) Setal entanglement: an undesirable method of stridulation by a neotropical tarantula (Araneae: Theraphosidae). J Zool Lond 235: 587-595.

Marx JM (1983) Aspects of microhabitat use by young juvenile spiny lobster, Panulirus argus. Masters Thesis, Florida State University, Tallahassee, FL.

Marx JM, Herrnkind WF (1985a) Macroalgae (Rhodophyta: Laurencia spp.) as habitat for young juvenile spiny lobsters, Panulirus argus. Bull Mar Sci 36:423-431.

Marx JM, Herrnkind WF (1985b) Factors regulating microhabitat use by young juvenile spiny lobsters, Panulirus argus - food and shelter. J Crust Biol 5:650-657.

Masters WM (1979) Insect defensive stridulation: Its defensive role. Behav Ecol Sociobiol 5: 187-200.

103 Masters WM (1980) Insect disturbance stridulation: Characterization of airborne and vibrational components of the sound. J Comp Physiol [A] 135: 259-268.

Mather JA (1991) Foraging, feeding and prey remains in middens of juvenile Octopus vulgaris (Mollusca: Cephalopoda). J Zool 224: 27-39.

Mathis A, Chivers DP, Smith RJF (1995) Chemical alarm signals: Predator deterrents or predator attractants? Am Nat 145:994-1005.

Maxwell KE (2006) Neurolipofuscin is a measure of age in the Caribbean spiny lobster, Panulirus argus, in Florida. Masters Thesis, Georgia State University, Atlanta, GA.

Meyer-Rochow VB, Penrose JB (1974) Sound and sound emission apparatus in puerulus and postpuerulus of the western rock lobster (Panulirus longipes). J Exp Zool 189: 283-289.

Meyer-Rochow VB, Penrose JB (1976) Sound production by western rock lobster Panulirus longipes (Milne Edwards). J Exp Mar Biol Ecol 23: 191-209.

Meyer-Rochow VB, Penrose JB, Oldfield BP, Bailey WJ (1982) Phonoresponses in the rock lobster Panulirus longipes (Milne Edwards). Behav Neural Biol 34: 331- 336.

Mima, A.; Wada, S. & Goshima, S. 2003. Antipredator defence of the hermit crab Pagurus filholi induced by predatory crabs. Oikos 102:104-110.

Mittelbach, G. G. & Persson, L. 1998. The ontogeny of piscivory and its ecological consequences. Canadian Journal of Fisheries and Aquatic Sciences 55:1454- 1465.

Mobius K (1867) Ueber die Entstehung der Tönge, welche Palinurus vulgaris mit den äusseren Fühlern hervorbringt. Arch f Naturgeschift 33: 73-75.

Montealegre-Z F, Morris GK (2004) The spiny devil katydids, Panacanthus Walker (Orthoptera: Tettigoniidae): an evolutionary study of acoustic behaviour and orphological traits. Syst Entomol 29: 21-57.

Moulton JM (1957) Sound production in the spiny lobster Panulirus argus (Latreille). Biol Bull 113: 286-295.

Moulton JM (1958) The acoustical behavior of some fishes in the bimini area. Biol Bull 114: 357-374.

Mulligan BE, Fischer RB (1977) Sounds and behavior of spiny lobster Panulirus argus (Latreille, 1804) (Decapoda, Palinuridae). Crustaceana 32: 185-199.

104 Munro JL (1974) The biology, ecology, exploitation and management of Caribbean reef fishes: crustaceans (spiny lobsters and crabs). Res Rep Zool Dept Univ West Indies 3, Part IV: 57.

Myrberg AA Jr. (1981) Sound communication and interception in fishes, pp. 395–426. In: Tavolga WN, Popper AN, Fay RR edits. Hearing and Sound Communication in Fishes. New York, Springer.

Neudorf DL, Sealy SG (2002) Distress calls of birds in a neotropical cloud forest. Biotropica 34: 118-126.

Oliver MD, Stewart R, Mills D, Macdiarmid AB, Gardner C (2005) Stock enhancement of rock lobsters (Jasus edwardsii): timing of predation on naive juvenile lobsters immediately after release. N Z J Mar Freshw Res 39: 391-397.

Packard A, Karlson HE, Sand O (1990) Low frequency hearing in cephalopods. J Comp Physiol [A] 166: 501-505.

Parker TJ (1878) Note on the stridulating organ of Palinurus vulgaris. Proc Zool Soc London 1878: 442-444.

Parsons A (2005) I can beat you one-handed: spiny lobster self defense after loss of an antenna. Masters Thesis, Florida State University, Tallahassee, FL.

Parsons DM, Eggleston DB (2005) Indirect effects of recreational fishing on behavior of the spiny lobster Panulirus argus. Mar Ecol Prog Ser 303:235-244.

Parsons DM, Eggleston DB (2006) Human and natural predators combine to alter behavior and reduce survival of Caribbean spiny lobster. J Exp Mar Biol Ecol 334: 196-205.

Pawlik JR (1993) Marine Invertebrate Chemical Defenses. Chem Rev 93: 1911-1922.

Patek SN (2001) Spiny lobsters stick and slip to make sound - These crustaceans can scare off predators even when their usual armour turns soft. Nature 411: 153-154.

Patek SN (2002) Squeaking with a sliding joint: mechanics and motor control of sound production in palinurid lobsters. J Exp Biol 205: 2375-2385.

Patek SN, Oakley TH (2003) Comparative tests of evolutionary trade-offs in a palinurid lobster acoustic system. Evolution 57: 2082-2100.

Patek SN, Feldmann RM, Porter M, Tshudy D (2006) Phylogeny and Evolution, pp. 113- 145. In: Phillips BF edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

105 Penney BK (2004) Individual selection and the evolution of chemical defence in nudibranchs: experiments with whole Cadlina luteomarginata (Nudibranchia: Doridina). J Mollusc Stud 70: 399–402

Perrone M, Paulson DR (1979) Incidence of Distress Calls in Mist-Netted Birds. Condor 81: 423-424.

Phillips BF, Melville-Smith R (2006) Panulirus species, pp. 359-384. In: Phillips BF edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

Phillips BF, Sastry AN (1980) Larval Ecology, pp. 11-57. In: Cobb JS, Phillips BF edits. The Biology and Management of Lobsters, Volume II: Ecology and Management. New York: Academic Press.

Phillips BF, Cobb JS, George RW (1980) General Biology, pp. 1-72. In: Cobb JS, Phillips BF edits. The Biology and Management of Lobsters, Volume I: Physiology and Behavior. New York: Academic Press.

Phillips BF, Booth JD, Cobb JS, Jeffs AG, McWilliam P (2006) Larval and postlarval ecology, pp. 231-262. In: Phillips BF edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

Pillai CS Gopinatha, Mohan M, Kunhikoya KK (1985) Observations on the lobsters of Minicoy Atoll. Indian J Fish 32: 112-122.

Popper AN, Salmon M, Horch KW (2001) Acoustic detection and communication by decapod crustaceans. J Comp Physiol [A] 187: 83-89.

Poulton EB (1890) The colours of animals. Their meaning and use. Especially considered in the case of insects. London: Kegan Paul, Trench, Trübner.

Ptacek MB, Sarver SK, Childress MJ, Herrnkind WF (2001) Molecular phylogeny of the spiny lobster genus Panulirus (Decapoda: Palinuridae). Mar Freshw Res 52: 1037-1048.

Pulliam HR, Caraco T (1984) Living in groups: is there an optimal group size? pp. 122- 147. In, Krebs JR, Davies NB Edits. Behavioural Ecology: An Evolutionary Approach, 2nd Ed. Oxford: Blackwell Scientific.

Randall JA (2001) Evolution and function of footdrumming as communication in mammals. Am Zool 41: 1143-1156.

Ratchford SA, Eggleston DB (1998) Size- and scale-dependent chemical attraction contribute to an ontogenetic shift in sociality. Anim Behav 56: 1027-1034.

Romero GA (1983) Distress call saves a Caiman c. crocodilus hatchling in the Venezuelan llanos. Biotropica 15: 71.

106 Rothschild M (1961) Defensive odours and Mullerian mimicry among insects. Trans R Entomol Soc Lond 113: 101–121.

Rowe C (2002) Sound improves visual discrimination learning in avian predators. Proc R Soc Lond [Biol] 269: 1353-1357.

Rowe C, Guilford T (1996) Hidden colour aversions in domestic clicks triggered by pyrazine odours of insect warning displays. Nature 383:520-522.

Rowe C, Guilford T (1999) The evolution of multimodal warning displays. Evol Ecol 13: 655-671.

Rowe C, Skelhorn J (2005) Colour biases are a question of taste. Anim Behav 69: 587- 594.

Russ JM, Racey PA, Jones G (1998) Intraspecific responses to distress calls of the pipistrelle bat, Pipistrellus pipistrellus. Anim Behav 55: 705-713.

Ruxton GD, Sherratt TN, Speed MP (2004) Avoiding attack: the evolutionary ecology of crypsis, warning signals and mimicry. Oxford: Oxford University Press.

Sadovy Y, Eklund A-M (1999) Synopsis of Biological Data on the Nassau Grouper, Epinephelus Striatus (Bloch, 1792), and the E. itajara (Lichtenstein, 1822). United States Department of Commerce, NOAA Technical Report NMFS 146, and FAO fisheries synopsis 157, pp. 1-65. Seattle, Washington.

Santini F, Tyler JC (2003) A phylogeny of the families of fossil and extant tetraodontiform fishes (Acanthomorpha, Tetraodontiformes), Upper Cretaceous to recent. Zool J Linnean Soc 139: 565-617.

Salmon M, Winn HE, Sorgente N (1968) Sound production and associated behavior in triggerfishes. Pac Sci 22: 11-20.

Schilman PE, Lazzari CR, Manrique G (2001) Comparison of disturbance stridulations in five species of triatominae bugs. Acta Trop 79: 171-178.

Schmitz B, Herberholz J (1998) Snapping behavior in intraspecific agonistic encounters in the snapping shrimp (Alpheus heterochaelis). J Biosci 23: 623-632.

Schratwieser J (1999) The impact of resident and transient predators on the population dynamics of juveniles Caribbean spiny lobster (Panulirus argus) in Florida Bay, Florida. Masters Thesis, Old Dominion University, Norfolk, VA.

Schuler W, Roper TJ (1992) Responses to warning coloration in avian predators. Adv Stud Behav 21: 111-146.

Sharp WC, Hunt JH, Lyons WG (1997) Life history of the spotted spiny lobster, Panulirus guttatus, an obligate reef-dweller. Mar Freshw Res 48: 687-698.

107 Sherman PW (1977) Nepotism and the evolution of alarm calls. Science 4310: 1246- 1253.

Sih A (1987) Predators and Prey Lifestyles: An Evolutionary and Ecological Overview, pp 203-224. In: Predation: direct and indirect impacts on aquatic communities. Hanover, NH: University Press of New England.

Sih A, Moore RD (1993) Delayed Hatching of Salamander Eggs in Response to Enhanced Larval Predation Risk. Am Nat 142: 947-960.

Sih A, Crowley PH, McPeek MA, Petranka JW, Strohmeier K (1985) Predation, competition, and prey communities. Ann Rev Ecol Syst 16: 269-311.

Sih A, Englund G, Wooster D (1998) Emergent impacts of multiple predators on prey. TREE 13: 350-355.

Smale M (1975) The warning squeak of the Natal rock lobster. S Afr Assoc Mar Biol Res Bull 11: 17-19.

Smith KN, Herrnkind WF (1992) Predation on early juvenile spiny lobsters Panulirus argus (Latreille): influence of size and shelter. J Exp Mar Biol Ecol 157: 3-18.

Smith RJF (1986) Evolution of alarm signals: role of benefits of retaining group members or territorial neighbors. Am Nat 128: 604-610.

Smith RL, Langley WM (1978) Cicada stress sound: an assay of its effectiveness as a predator defense mechanism. Southwest Nat 23: 187-196.

Spangler HG, Manley DG (1978) Sounds associated with mating behavior of a mutillid wasp. Ann Entomol Soc Am 71: 389-392.

Spanier E, Weihs D, Almog-Shtayer G (1991) Swimming of the Mediterranean slipper lobster. J Exp Mar Biol Ecol 145: 15-31.

Speed MP, Ruxton GD (2005) Warning displays in spiny animals: One (more) evolutionary route to aposematism. Evolution 59: 2499–2508.

Sogard SM (1997) Size-selective mortality in the juvenile stage of teleost fishes: A review. Bull Mar Sci 60: 1129-1157.

Staton MA (1978) Distress calls of crocodilians: whom do they benefit? Am Nat 112: 327-332.

Stankowich T, Blumstein DT (2005) Fear in animals: a meta-analysis and review of risk assessment. Proc R Soc Lond [Biol] 272: 2627-2634.

Strugnell J, Jackson J, Drummond AJ, Cooper A (2006) Divergence time estimates for major cephalopod groups: evidence from multiple genes. Cladistics 22: 89-96.

108 Summers K, Clough ME (2001) The evolution of coloration and toxicity in the poison frog family (Dendrobatidae). Proc Natl Acad Sci U S A 98: 6227-6232.

Tullrot A (1994) The evolution of unpalatability and warning coloration in soft-bodied marine-invertebrates. Evolution 48: 925-928.

Tullrot A, Sundberg P (1991) The conspicuous nudibranch Polycera quadrilineata: aposematic coloration and individual selection. Anim Behav 41: 175-176.

Uetz GW, Hieber CS (1994) Group size and predation risk in colonial web-building spiders; analysis of attack abatement mechanisms. Behav Ecol 5: 326-333.

van der Meeren GI (2005) Potential of ecological studies to improve survival of cultivated and released European lobsters, Homarus gammarus. N Z J Mar Freshw Res 39: 399-424.

Vermeij GJ (1982) Unsuccessful predation and evolution. Am Nat 120: 701–720.

Vermeij GJ (1987) Evolution and escalation: an ecological history of life. Princeton University Press, Princeton, NJ.

Versluis M, Schmitz B, von der Heydt A, Lohse D (2000) How snapping shrimp snap: through cavitating bubbles. Science 289: 2114-2117.

Vester HI, Folkow LP, Blix AS (2004) Click sounds produced by cod (Gadus morhua). J Acoust Soc Am 115: 914-919.

Voight JR (1998) An overview of shallow-water Octopus biogeography, pp. 549–559. In: Voss NA, Vecchione M, Toll RB, Sweeney MJ edits. Systematics and Biogeography of Cephalopods. Washington, D. C.: Smithsonian Institution Press.

Wahle RA, Fogarty MJ (2006) Growth and development: understanding and modeling growth variability in lobsters, pp. 1-44. In: Phillips, B. F. edit. Lobsters: Biology, Management, Aquaculture, and Fisheries. Oxford: Blackwell.

Warkentin KM (1995) Adaptive plasticity in hatching age: a response to predation risk trade-offs. Proc Natl Acad Sci U S A 92: 3507-3510.

Warner RE, Combs CL, Gregory DR (1977) Biological Studies of the Spiny Lobster, Panulirus argus (Decapoda: Palinuridae), in South Florida. Proc Gulf Caribb Fish Inst 29: 166-183.

Wainwright PC, Friel JP (2000) Effects of Prey Type on Motor Pattern Variance in Tetraodontiform Fishes. J Exp Biol 286: 563–571.

Wallace AR (1867) Untitled. T Entomol Soc Lond y 1864–1869 and March. III:lxxx– lxxxi.

109 Wasson K, Lyon BE (2005) Flight or fight: flexible antipredatory strategies in porcelain crabs. Behav Ecol 16: 1037-1041.

Werner EE, Gilliam JF (1984) The ontogenetic niche and species interactions in size- structured populations. Ann Rev Ecol Syst 15: 393-425.

Wickens PA (1996) Fur seals and lobster fishing in South Africa. Aquatic Conservation- Marine and Freshwater Ecosystems 6: 179-186.

Wiklund C, Jarvi T (1982) Survival of distasteful insects after being attacked by naive birds - a reappraisal of the theory of aposematic coloration evolving through individual selection. Evolution 36:998-1002.

Wilbur HM (1980) Complex Life-Cycles. Ann Rev Ecol Syst 11: 67-93.

Wilkens LA, Schmitz B, Herrnkind WF (1996) Antennal responses to hydrodynamic and tactile stimuli in the spiny lobster Panulirus argus. Biol Bull 191: 187-198.

Williamson R (1988) Vibration sensitivity in the statocyst of the northern octopus Eledone cirrosa. J Exp Biol 134: 451-454.

Wisenden BD, Pohlman SG, Watkin EE (2001) Avoidance of conspecific injury-released chemical cues by free-ranging Gammarus lacustris (Crustacea:Amphipoda). J Chem Ecol 27: 1249-1258.

Wooster D, Sih A (1995) A review of the drift and activity responses of stream prey to predator presence. Oikos 73: 3-8.

Wrona FJ, Dixon RWJ (1991) group-size and predation risk: a field analysis of encounter and dilution effects. Am Nat 137: 186-201.

Young RF, Winn HE (2003) Activity patterns, diet, and shelter site use for two species of moray eels, Gymnothorax moringa and Gymnothorax vicinus, in Belize. Copeia 2003: 44-55.

Zaret TM, Suffern JS (1976) Vertical migration in zooplankton as a predator avoidance mechanism. Limnol Oceanogr 21: 804-813.

110 BIOGRAPHICAL SKETCH

PETER E. BOUWMA

Education Ph.D., Department of Biological Science (2006), Florida State University, Tallahassee, FL. Thesis Advisor: Professor William F. Herrnkind. GPA (4 scale): 3.96

B.S., Biology with Honors (2000), Calvin College, Grand Rapids, MI. GPA (4 scale): 3.784

Research and Employment Experience Post-Doctoral Associate: Department of Biological Sciences, Clemson University (2006- 2007). Post Doctoral Advisor: Michael Childress.

Graduate Research Assistant: Department of Biological Science, Old Dominion University (2002, 2003). Funded by the Florida Fish and Wildlife Conservation Commission. Research Topic: Field Surveys of the Nearshore Hard-bottom Habitat of the Florida Keys. Principle Investigator: Professor Mark J. Butler IV.

Graduate Research Assistant: Department of Biological Science, Florida State University (2001). Funded by Florida Sea Grant. Research topic: Management of Spiny Lobsters in South Florida Based on Postlarval Supply and Juvenile Dynamics. Principle Investigator: Professor William F. Herrnkind.

Graduate Teaching Assistant: Department of Biological Science, Florida State University (2000 – 2006). Introductory Biology Laboratory (BSC 2010L), Animal Diversity Laboratory (BSC 2011L), Animal Behavior (ZOO 4513), The Biology of Higher Marine Invertebrates (ZOO 4204C).

Research Experience for Undergraduates Fellow: University of Michigan Biological Station, (1999). Research Topic: The Integration of Multiple Predator Cues in the Crayfish, Orconectes propinquus. Research Mentor: Professor Brian A. Hazlett.

Undergraduate Research Assistant: Departments of Entomology and Zoology, University of Wisconsin – Madison (1998). Funded by NSF Research Experience for Undergraduates grant to Professor Robert L. Jeanne. Research Topic: Adult Mortality Rates of the Social Wasp, Polybia occidentalis. Principle Investigator: Professor Robert L. Jeanne.

Fellowships and Academic Awards

111 Outstanding Teaching Assistant Award (2006). Florida State University, Program for Instructional Excellence.

Best Student Presentation, Honorable Mention (2005). Marine Benthic Ecology Meeting, Williamsburg, VA.

Jack Gramling Award in Marine Science (2004). Florida State University, Department of Biological Science.

Sigma Xi Grant-in-Aid of Research (2004). Sigma Xi, The Scientific Research Society.

Margaret Y. Menzel Endowed Award (2003). Florida State University, Department of Biological Science.

Dissertation Research Grant Fall (2003). Florida State University, Office of Graduate Studies.

Aylesworth Scholarship Award (2003 – 2004). Aylesworth Foundation for the Advancement of Marine Sciences, Florida Sea Grant College Program.

Keys Marine Lab Graduate Student Grant-in-Aid (2002, 03, 04, 05). Florida Institute of Oceanography, Florida Fish and Wildlife Conservation Commission.

Honorable Mention (2001). National Science Foundation Graduate Research Fellowship Program.

Teaching Assistantship (2000-2006). Florida State University, Department of Biological Science.

Dean's Scholarship (1996-2000). Calvin College.

Two-Ten Scholarship (1997-2000). Two-Ten International Footwear Foundation.

Dean's List (1996-2000). Calvin College.

Outreach Saturday-at-the-Sea Instructor (2002, 2004, 2005). Florida State University Office of Science Teaching activities.

Teaching Experience Graduate Teaching Assistant, Department of Biological Science, Florida State University. ZOO 4204C: (1X) The Biology of Higher Marine Invertebrates (co-instructor). ZOO 4513: (2X) Animal Behavior. BSC 2011L: (8X) Animal Diversity Laboratory. BSC 2010L: (3X) Introductory Biology Laboratory.

112 Guest Lecturer, Department of Biological Science, Florida State University. ZOO 4513: (3X) Animal Behavior.

Publications Bouwma AB, Bouwma PE, Nordheim EV, Jeanne RL (2003) Adult mortality rates in young colonies of a swarm-founding social wasp. J Zoo 260: 11-16.

Bouwma AB, Bouwma PE, Nordheim EV, Jeanne RL (2003) A Cost of Swarm Founding in a Tropical Social Wasp: Adult Mortality Increases with Distance Emigrated. J Insect Behav 16: 439-452.

Bouwma P, Hazlett BA (2001) Integration of Multiple Predator Cues in the Crayfish Orconectes propinquus. Anim Behav 61:771-776.

Bouwma PE, Bouwma AM, Jeanne RL (2000) Social wasp swarm emigration: males stay behind. Ecol Ethol Evo 12: 35-42.

News Articles "Decoding Spiny Lobsters' Violin-like Screech." John Roach, National Geographic News, July 28, 2004 URL: http://news.nationalgeographic.com/news/2004/07/0728_040728_spinylobsters.ht ml.

Presentations Dissertation Public Defense (4 October, 2006). Department of Biological Science, Florida State University, Tallahassee, FL, USA. ”Aspects of Antipredation in Panulirus argus and Panulirus guttatus: Behavior, Morphology, and Ontogeny.” Peter E. Bouwma

Marine Benthic Ecology Meeting (12 March, 2006). Quebec City, Quebec, Canada. "Contenders and Born Losers: A comparison of spiny lobster defensive behavior and morphology." Peter E. Bouwma & William F. Herrnkind.

Marine Benthic Ecology Meeting (7 April, 2005). Williamsburg, VA, USA. "Spiny lobsters combine weaponry with sound to "teach" predators not to attack" Peter E. Bouwma & William F. Herrnkind

Ecology and Evolution Seminar (4 March, 2005), Department of Biological Science, Florida State University, Tallahassee, FL, USA. "The Mysterious Squawk of the Spiny Lobster: Does sound production help lobsters survive encounters with predators?" Peter E. Bouwma & William F. Herrnkind.

Invited Colloquium Talk (3 August, 2004). Dauphin Island Sea Lab, Dauphin Island, AL, USA. "The Mysterious Squawk of the Spiny Lobster: Does sound production help lobsters repel predators?" Peter E. Bouwma.

113 7th International Conference and Workshop on Lobster Biology and Management (10 February, 2004). Hobart, Tasmania, Australia. "The antipredator function of sound production by the spiny lobster, Panulirus argus." Peter E. Bouwma & William F. Herrnkind.

Marine Benthic Ecology Meeting (28 March, 2003). Mystic, CT, USA. "Harried by Hemolymph: Assessment of predation risk by spiny lobsters." Peter E. Bouwma & William F. Herrnkind.

Natural History Seminar (23 March, 2001). Department of Biological Science, Florida State University, Tallahassee, FL, USA. "3½ weeks: The Costs of Swarm Emigration and Colony Founding in the Swarm-Founding Social Wasp Polybia occidentalis." Peter E. Bouwma, Andrew M. Bouwma, Robert L. Jeanne & Eric V. Nordheim.

Departmental Seminar (31 March, 2000). Biology Department, Calvin College, Grand Rapids, MI. "Scared Hemolymphless: Responses of the crayfish Orconectes propinquus to multiple predator inputs." Peter E. Bouwma & Brian Hazlett.

Research Seminar (August, 1999). University of Michigan Biological Station, Pellston, MI. "Scared Hemolymphless: Responses of the crayfish Orconectes propinquus to multiple predator inputs." Peter E. Bouwma & Brian Hazlett.

Departmental Seminar (30 April, 1998). Biology Department, Calvin College, Grand Rapids, MI. "Colony size and mortality in a social wasp." Peter E. Bouwma, Andrew M. Bouwma & Robert L. Jeanne.

114