California State University, Northridge

CALIFORNIA STATE UNIVERSITY, NORTHRIDGE

Leg Armaments and Alternative Mating Behavior in Pristoceuthophilus Camel Crickets

A thesis submitted in partial fulfillment of the requirements

For the degree of Master of Science

in Biology

Lauren P. Conroy

May 2014 The thesis of Lauren P. Conroy is approved:

______

James N. Hogue, Ph.D. Date

______

Paul S. Wilson, Ph.D. Date

______

David A. Gray, Ph.D., Chair Date

California State University, Northridge

ii ACKNOWLEDGEMENTS

I am extremely grateful to my thesis advisor Dr. Dave Gray for his support and encouragement throughout this entire process. Thanks for always being available and receptive when I endlessly pestered you with questions and concerns about my project.

Under your advisement, I have learned a tremendous amount about the scientific process.

I am also grateful to my thesis committee members, Drs. Jim Hogue and Paul Wilson.

Thank you both for posing thought-provoking questions about my thesis, and for your meticulous editing and spell-chekcing. Jim, thanks for the photographs of cricket junk.

I could not have run my projects without the help of many field assistants who were willing to go into the woods at night. I thank Amanda Lindgren, Andrea Haberkern,

Sydney Hunter, Thom Kendall, Melanie Farias, Tian Huang, and especially Tom Chen, who helped me on more occasions than I can count. I thank my lab-mates Tom Chen,

Mark Oliva, and Nick Gutierrez for discussing my project with me and for their bizarrely comforting camaraderie. I am also grateful to my parents for encouraging a love of learning throughout my whole life and enabling me to pursue a Master’s degree.

The background information provided by Dr. Ted Cohn of the University of

Michigan (UM) was instrumental to this study. Mark O’Brien, also from UM, was very helpful by sharing Pristoceuthophilus specimens. I thank my funding sources: the

Orthopterists’ Society and the CSUN Departments of Graduate Studies and Biology.

Finally, I would like to thank the entire Biology department here at CSUN. The faculty have contagious passion for biology. My fellow grad students have been instrumental to my mental and emotional health, and without their support I could not have made it through this process.

iii TABLE OF CONTENTS

Signature Page ...... ii

Acknowledgements ...... iii

List of Figures ...... v

List of Tables ...... vi

Abstract ...... vii

Chapter 1: Forced Copulation as a Conditional Alternative Mating Strategy in Pristoceuthophilus marmoratus ...... 1 Introduction ...... 1 Experiment 1: Condition Dependence of Somatic and Sexually Selected Traits ...... 6 Methods...... 6 Results ...... 9 Bridge from Exp.1 to Exp. 2 ...... 13 Experiment 2: Predictors of Forced Copulation ...... 14 Methods...... 16 Results ...... 18 Discussion ...... 23

Chapter 2: Interspecific Comparisons of Armaments and Behavior in Pristoceuthophilus ...... 28 Introduction ...... 28 Methods...... 33 Results ...... 37 Discussion ...... 51

Works Cited ...... 56 Appendix ...... 60

iv LIST OF FIGURES

Figure 1. Condition dependence, somatic traits ...... 10

Figure 2. Condition dependence, hind leg traits ...... 11

Figure 3. Mating script of P. marmoratus ...... 19

Figure 4. Logistic regressions, body size and armaments ...... 19

Figure 5. Comparisons of mated vs. unmated males ...... 21

Figure 6. Male pseudosternites ...... 30

Figure 7. Pristoceuthophilus phylogeny ...... 31

Figure 8. Measurement of hind leg traits ...... 34

Figure 9. RMA slopes for spine length (all species combined) ...... 38

Figure 10. RMA slopes for flange area (all species combined) ...... 39

Figure 11. RMA slopes for 1° tibial deflection (all species combined) ...... 40

Figure 12. RMA slopes for 2° tibial deflection (all species combined) ...... 41

Figure 13. RMA slopes for femur area (all species combined) ...... 42

Figure 14. RMA slopes for femur length (all species combined) ...... 43

Figure 15. Comparisons of body and leg traits (all species) ...... 45

Figure 16. RMA regressions for spine length (species separate; APPENDIX) ...... 60

Figure 17. RMA regressions for flange area (species separate; APPENDIX) ...... 61

Figure 18. RMA regressions for 1° tibial deflection (species separate; APPENDIX)...... 62

Figure 19. RMA regressions for 2° tibial deflection (species separate; APPENDIX)...... 63

Figure 20. RMA regressions for femur area (species separate; APPENDIX) ...... 64

Figure 21. RMA regressions for femur length (species separate; APPENDIX)...... 65

v LIST OF TABLES

Table 1. Diet components ...... 7

Table 2. RMA slopes of leg traits, 2012 ...... 9

Table 3. RMA slopes of leg traits, 2013 ...... 18

Table 4. Logistic regressions ...... 18

Table 5. Mating outcomes...... 21

Table 6. RMA slopes of leg traits (5 spp.) ...... 37

Table 7. ANOVA comparisons (5 spp.) ...... 44

Table 8. Male-male fights ...... 47

Table 9. Male-female trials ...... 48

Table 10. Power analysis ...... 48

Table 11. Dual-purpose armaments, presence/absence ...... 48

Table 12. Tukey’s HSD tests, body and leg comparisons (APPENDIX) ...... 66

Table 13. Tukey’s HSD tests, copulation duration (APPENDIX) ...... 67

vi ABSTRACT

Leg Armaments and Alternative Mating Behavior in Pristoceuthophilus Camel Crickets

Lauren P. Conroy

Master of Science in Biology

In the animal kingdom, males of many species possess striking weaponry used in intrasexual competition for access to females. Until recently, there were no known cases of male weaponry being used against females in sexual coercion. However, in the camel cricket Pristoceuthophilus marmoratus, males use modified hind legs (modifications consist of femoral spines and bent tibiae) not only to fight with each other, but also to trap females and force them to copulate. As male hind leg armaments are positively allometric, the largest males would be best equipped to force-copulate, although they should be the most attractive to females. In order to resolve this potential paradox and determine which males are most likely to force-copulate, I manipulated male body size with diet and performed mating trials. The use of forced copulation was conditional on male body size, with smaller males being more likely to attempt coercive mating. The finding is in accord with the literature on male predictors of forced copulation and suggests an evolutionary exaptation of hind leg armaments in this species, wherein an intrasexual fighting weapon took on a secondary function of sexual coercion. In addition to this study, I performed a comparative survey of armaments, examining fighting and mating behavior in four morphologically similar congeners (P. ‘Huachuca summer,’ P.

vii arizonae, P. ‘Madera,’ and P. ‘Mt. Pinos’) to determine whether hind leg armaments also serve intra- and intersexual functions in these species. Intrasexual leg fighting occurred in all species for which trials were performed, while hints of sexual coercion occurred in two species (P. ‘Huachuca summer’ and P. ‘Mt. Pinos’), suggesting additional cases of a uniquely dual-purpose armament.

viii CHAPTER 1: FORCED COPULATION AS A CONDITIONAL ALTERNATIVE MATING STRATEGY IN PRISTOCEUTHOPHILUS MARMORATUS

Introduction

Anisogamy, or the differential investment in gamete size by males and females, sets the stage for selection favoring differences in reproductive strategies between the sexes (Andersson 1994, Alcock 2009). Because sperm are small and inexpensive to produce, they greatly outnumber the large, energetically costly eggs able to be fertilized in a population (Alcock 2009). Males maximize their fitness by fertilizing as many eggs as possible; thus, male fitness should steadily increase with the number of matings procured (Bateman 1948). However, females typically require only one mating (per lifetime for semelparous organisms/per breeding season for iteroparous organisms) to fertilize all of their eggs, and are selected to be choosy about with whom they mate.

Therefore, males and females differ in their optimal number of matings (Trivers 1972), with multiple matings often lowering female fecundity and longevity because of physical injury during copulation/courtship (Clutton-Brock & Parker 1995), reduced time available for foraging (Kokko 2005), or the toxic side effects of sperm competition

(Bateman 1948, Wigby & Chapman 2005) (but see Arnqvist & Nilsson 2000 on the benefits of multiple mating for females). Recently mated females typically show reduced interest in courting males (Andersson 1994, Haley and Gray 2012).

Because females must be choosy about their mates, males must compete for access to fertile females. Males may directly fight each other with armaments for access to females (intrasexual selection), or they may compete for females’ attention by

1 producing elaborate displays of songs or ornamentation (gender-biased selection)

(Andersson 1994). Both armaments and ornaments are usually condition-dependent

(Andersson 1986, Berglund et al. 1996, Rowe & Houle 1996, Møller & Petrie 2002,

Bonduriansky 2007), meaning that they are expressed to different degrees between individuals, based on each individual’s condition. Condition is a function of genotype and environmental influences, such that some genotypes are better suited to process extrinsic factors (resources, stressors, etc.) than others. Individuals with these favorable genotypes amass higher levels of residual energy that may be allocated to secondary traits

(Andersson 1986). In addition, sexually selected traits are typically positively allometric, meaning that for an increase in body size, there is a disproportionate increase in armament/ornament size because an increase in relative trait size results in greater marginal fitness gains for large individuals than small individuals (Bonduriansky & Day

2003). Armaments, ornaments, and songs reveal information about a male’s quality, which may indicate his “good genes,” ability to provide direct benefits, or Fisherian attractiveness (Alcock 2009).

Displays of songs and ornaments, while often necessary to attract a female, come with substantial risks. Many song-producing organisms are targets for acoustically orienting predators, parasites, and parasitoids. In one extreme example, a population of field crickets (Teleogryllus oceanicus) in Kauai was parasitized by parasitoid flies so heavily that a tegminal mutation that renders males silent spread to near fixation (Zuk et al. 2006). Ornamental traits also attract unwanted attention from predators, as classically demonstrated in guppies (Poecilia reticulata). In predator-rich streams, male guppies

2 have reduced their colorful spotting compared to males in predator-poor streams (Endler

1980).

Males use numerous strategies to circumvent the costs and risks of sexual displays. The use of these alternative mating strategies is often contingent on a male’s condition (Alcock 2009). Satellite males, commonly found in acoustically displaying species, are usually small individuals that cannot front the expense of producing a high quality song attractive to females (Cade 1980). Such males wait near calling males and intercept females homing in on a calling male. Sneaking and forcing behaviors do not necessarily occur near a displaying male. Another way male guppies can reduce predation risk is to forgo their mating display, a conspicuous dance that emphasizes their color spots, and simply swim toward females and thrust their intromittent organ into a female’s gonopore (Evans et al. 2003). In sagebrush crickets (Cyphoderris strepitans), males lacking the fleshy hind wings that serve as nuptial gifts use a specialized abdominal pinching structure, or “gin trap,” to restrain females and force copulation

(Sakaluk et al. 1995). Forced copulation is also common in waterfowl, where it is employed by unpaired and presumably unattractive males (Cheng et al. 1982, McKinney et al. 1983).

Alternative explanations for forced copulation exist. Early, albeit somewhat farfetched, hypotheses include satisfaction of male sexual urges or simply abnormal male behavior (McKinney et al. 1983). The former could be a byproduct of male over- eagerness, whereas the latter could stem from genetic abnormalities. However, both of these hypotheses suggest that the use of forced copulation would be random with respect to male quality, which is generally not the case (Clutton-Brock & Parker 1995). Allen

3 and Simmons (1996) posited that females assess male vigor while being subdued prior to forced copulation, implicating female choice. However, other studies show that females being forced to copulate suffer reduced fecundity and longevity (Thornhill 1980, McLain and Pratt 1999), implying a lack of female choice. Kokko (2005) suggested a twist on

Fisherian selection, in that females should prefer coercive males in order to produce coercive sons. However, given that forced copulation is typically employed by a minority of males in a population as an alternative mating strategy (McLain & Pratt 1999,

Evans et al. 2003), Kokko’s hypothesis seems unlikely. (For species in which copulation is always forceful and unsolicited by females, see bed bugs in Stutt & Siva-Jothy 2001, water striders in Arnqvist & Rowe 2002, and bushcrickets in Vahed & Carron 2008.)

Low quality males that have little chance of enticing a female should be most likely to force copulation (Cade 1980).

The camel cricket Pristoceuthophilus marmoratus (Rhaphidophoridae) presents a potentially paradoxical situation with regard to male quality and forced copulation. In this species, males possess highly modified hind legs, with two large ventral spines on an enlarged femur and a markedly bent tibia. Males use these modified hind legs to fight with other males (Haley and Gray 2012). However, Haley and Gray (2012) showed that males also use their hind legs to trap females and subsequently force them to mate. In a forced copulation, a male grabs and holds a female in the crook of one hind leg and then forces his genitalia into her genital opening. The forced copulation position looks very different from the voluntary copulation stance, where the female climbs up onto the male’s back. Females often struggle against hind leg submission by males, and sometimes escape before genital coupling occurs. Like other forms of armament, the

4 male hind leg modifications in P. marmoratus (femoral spines, thickened femur, and tibial bend) are positively allometric (Haley and Gray 2012), meaning that they are expressed to a disproportionate degree in large individuals and nearly unnoticeable in small individuals. The allometry of these armaments creates a situation in which the largest and presumably most attractive males are best equipped to coerce females to mate.

Although Haley and Gray (2012) noted that non-virgin females were more likely to be forced to mate than virgin females, they did not investigate which males were most likely to be forcers. The goals of the present study were to: (1) test the condition dependence of male body size and leg armaments, (2) determine if male condition is a predictor of forced copulation, and (3) evaluate if female preference is correlated with male quality. I hypothesize that low quality males will be most likely to be forcers, despite having relatively inferior coercive structures. The results of this study should shed light on a seemingly paradoxical case of alternative mating strategies.

5 Experiment 1: Condition Dependence of Somatic and Sexually Selected Traits

Rationale

Diet manipulations were performed to test the condition dependence of somatic and sexually selected traits. Males on a high-quality diet were expected to grow larger and express sexually selected traits to a greater degree than males on a low-quality diet.

Methods

Collection and Rearing

In early July 2012, young undifferentiated nymphs of P. marmoratus were collected from Malibu Creek State Park (Los Angeles County, California, USA; 34.10°N,

118.73°W, elev. 200 m). Collection was accomplished by placing oatmeal piles along the dry bed of Malibu Creek approximately one hour before sunset and returning to the piles half an hour after sunset. Captured nymphs were returned to the lab and randomly placed on one of two diet treatments, a high-quality (HQ) diet and a low-quality (LQ) diet. The diets (Table 1) were taken from a study by Patton (1967) on house crickets (Acheta domesticus). In Patton’s study, the diets promoted similar survivorship but large differences in growth and mass increases. Individuals assigned to the diet treatments did not differ in pre-diet initial mass (HQ: 13.08 ± 0.76 mg, LQ: 13.29 ± 0.76 mg; t = -0.20, df = 140, p = 0.84).

Individuals were reared in a 27°C growth room on a 12:12 h dark/light reversed photoperiod. Each individual was housed in a 500 mL plastic tub with cotton-plugged

6 water vial, ad libitum food, and egg carton pieces for shelter. Eclosion to adulthood began in early October 2012 and ended in late December.

Table 1. Components and % composition of the high quality and low quality diets.

Components High Quality Diet (%) Low Quality Diet (%)

Brewer’s yeast 5 0 Corn meal 35 0 Liver Powder 5 0 Milk, dried skim 15 0 Soybean meal 10 50 Wheat middlings 30 50

Composition Protein 20.4 23.8 Carbohydrate 47.0 36.2 Fat 3.2 4.9 (Water 29.4 35.1)

Male-Female Mating Trials

Following the final molt, diet-manipulated virgin males were randomly paired with lab-reared virgin females (all raised on a diet of Purina Cat Chow under the same laboratory conditions) to test whether condition would have an effect on the likelihood of using forced copulation. (Using virgins allowed us to control for each individual’s mating history, but ultimately might not have been the best choice; see Results below.)

Due to high rates of early juvenile mortality from tachinid flies, and because approximately half of the initial 142 undifferentiated juveniles were female, only 28 males (HQ: n = 13, LQ: n = 15) matured to adulthood and were used in mating trials.

Individuals were paired up for mating trials approximately 4-6 days after the adult molt.

Mating trials were conducted in the growth room during the dark portion of the cricket photoperiod. Each male-female pair was placed in a 14 x 8 x 10 cm plastic tub

7 under a red light, to minimize light disturbance. Each trial was filmed by a high definition HDMI video camera (Canon Vixia HV30) for 1 hour. No observer was present during the trial. Each individual was used in only one trial.

Measurements of Condition and Armature Investment (Experiments 1 and 2)

Following mating trials, males were preserved in ethanol so that their pronotums and hind legs could be measured (male mass had been measured immediately pre-trial).

Pronotum length is isometric with body size (Haley and Gray 2012) and was measured with digital calipers. Male condition was determined by taking the residuals from a linear regression of mass on pronotum length (r2 = 0.82, p < 0.001). Male hind legs were photographed, and ImageJ was used to measure three armament traits: length of proximal femoral spine, femur area, and angle of tibial deflection (calculated as 180° minus the angle of the tibial bend). Each pronotum and hind leg trait was measured twice, and the average of the two values was used. Measures of allometric slopes were made by regressing the three armament traits on pronotum length using Reduced Major Axis

(RMA) regression (after square root-transforming femur area and arcsine square root- transforming the angle of tibial deflection).

Statistical Analyses

Analyses were performed in SYSTAT 13, with the exception of RMA regression, which was accomplished with Bohonak & Van Der Linde’s (2004) online software

(http://www.kimvdline.com/professional/rma.html). Transformations were made when data did not meet the assumptions of normality and homogeneity of variance.

8 Results

Condition Dependence of Somatic and Armament Traits

Males on the high-quality diet were significantly heavier (t = 3.03, df = 27, one- tailed p = 0.003), larger (t = 2.07, df = 27, p = 0.024), and in better condition (t = 2.27, df

= 27, p = 0.016) than males on the low-quality diet (Figure 1). (One-tailed tests were used because males on the high-quality diet were expected to be superior to males on the low-quality diet.)

Hind leg armaments were also condition dependent. Armaments were defined as traits with allometric slopes significantly greater than 1 (Table 2). Not only did HQ males have absolutely larger femoral spines (t = 3.07, df = 27, p = 0.002) and thicker femurs (t = 3.20, df = 27, p = 0.002; Figure 2A, C), they also invested relatively more in armaments when controlling for body size (relative spine length: t = 2.86, df = 27, p =

0.004; relative femur area: t = 3.29, df = 27, p = 0.001; Figure 2D, F). Only the angle of tibial deflection trait was not affected by diet (absolute angle: t = 0.24, df = 27, p = 0.41; relative angle: t = -0.23, df = 27, p = 0.42; Figure 2B, E).

9 Table 2. RMA regressions of hind leg traits on body size for HQ and LQ males (n = 28). Variables were log-transformed. Traits with RMA slopes significantly different from 1 are allometric. Positively allometric traits are considered armaments. Stastical significance is denoted with an asterisk, as in all subsequent graphs.(sqrt = square root, asin = arcsine)

Trait RMA slope r2 p Allometry ± SE

Spine length 5.24 ± 0.81 0.38 <0.01* Positive Asin sqrt (° of tib. deflection) 3.12 ± 0.51 0.31 <0.01* Postive Sqrt (femur area) 1.08 ± 0.11 0.72 0.47 Weakly Positivea aAlthough the RMA slope for sqrt femur area was not significantly different from 1 in this data set, Haley and Gray (2012) showed this trait to be positively allometric with body size (RMA slope 1.16 ± 0.08, n = 23). Additional data in this chapter also show a positively allometric slope for this trait (1.06 ± 0.03, n = 101; see Table 3). For this reason, femur area will be considered weakly positively allometric and will be included in subsequent comparisons of armaments of HQ and LQ males.

A 0.4 B 5 0.35 4

0.3

0.25 3 0.2 2 Mass Mass (g) 0.15 0.1 1 0.05 0 Pronotumlength (mm) 0 HQ LQ HQ LQ

C 0.03

0.02

0.01

0 Condition -0.01

-0.02 HQ LQ

Figure 1. Males on the high quality (HQ) diet (n = 13) were heavier (A), bigger (B), and in better condition (C) than males on the low quality (LQ) diet (n = 15). Means ± SE are shown

(as in all subsequent graphs).

10 0.1 0.025

0.08 0.02

0.06 0.015

0.04 0.01 (cm/mm)

Spinelength (cm) 0.02 0.005 Spinelength/pronot 0 0 HQ LQ HQ LQ

50 10

40 /mm) 8 °

30 6

) ° ( 20 4

10 Angle tibial of 2 deflection/pronot (

Angle Angle tibial of deflection 0 0 HQ LQ HQ LQ

0.35 0.08

0.3 0.07 0.06 0.25 0.05 0.2

/mm) 0.04 0.15 2

0.03 (cm

0.1 0.02 Femur Femur area/pronot Femur Femur area (sq. cm) 0.05 0.01 0 0 HQ LQ HQ LQ

Figure 2. Condition dependence of hind leg armature traits. Males on the HQ diet (n = 13) not only had larger femoral spines (A) and greater femoral area (C) than LQ males (n = 15), they also invested more in femoral spines and femoral area when controlling for body size (D and F). Diet did not affect either the absolute or body size-controlled angle of tibial deflection (B and E).

11 Male-Female Mating Trials

Mating trials (n = 28) were performed to determine whether condition has an effect on the likelihood of males forcing copulation. However, of the 28 trials, there was only one trial in which voluntary copulation occurred and two trials in which the male leg-squeezed the female but did not successfully force her to copulate. No trials resulted in successful forced copulation.

12 BRIDGE FROM EXPERIMENT 1 TO EXPERIMENT 2

What could explain this general lack of mating behavior? Individuals were used

4-6 days after the adult molt, which may have been before they were fully sexually mature. Lab-rearing the crickets from a young age may have perturbed their normal development and subsequent adult behavior. Pristoceuthophilus camel crickets are very sensitive to lab conditions; they even refuse to oviposit in the lab, despite being offered a number of different substrates (pers. obs.). The mating trials (1 hour each) may not have been long enough to capture the behavior of these slow-moving individuals. Finally, the use of virgin males and females might not have been the ideal choice, given the higher propensity for males to force non-virgin females to mate than virgin females (Haley and

Gray 2012).

Having demonstrated the condition dependence of male somatic and armament traits in 2012, I redid mating trials in 2013 (see Experiment 2 below) to determine predictors of forced copulation. To maximize the chances of observing forced copulation, I used (presumably) non-virgin males and females that had been caught as adults in the wild (precluding a manipulation of male condition) and increased the length of the mating trials.

13 Experiment 2: Predictors of Forced Copulation

Rationale

In order to identify factors that predict forced copulation, I compared traits of males at three key points in the scripted mating behavior (Figure 3) of Pristoceuthophilus marmoratus (Haley & Gray 2012). After a mating invitation from a male, in which the male spreads his hind legs and pushes his abdomen back toward the female, the female can either voluntarily mount the male or not (Branch Point 1). After this point, a rejected male can either attempt forced copulation by grabbing a female with a hind leg or not

(Branch Point 2); those that attempt forced copulation can either be successful or not if the female escapes (Branch Point 3). Any differences between males at any branch point will shed light on which males force copulation.

Additionally, I evaluated female preference for male traits by offering females two males consecutively (no-choice trials). The results of these trials should confirm whether demonstrably “less desirable” males are also the most likely to force copulation.

Figure 3. Mating script of P. marmoratus, showing the path to successful forced copulation. Differences between males at each branch point will be analyzed in order to determine which males are most likely to force-copulate.

15 Methods

Collection and Housing

In early October 2013, adult males and females of P. marmoratus were collected from Malibu Creek State Park using oatmeal piles. Mating history was unknown, but all individuals could have been non-virgins, which is biologically relevant and promotes statistical power because non-virgin females are more likely to be forced to mate than virgins (Haley and Gray 2012). After collection, individuals were isolated in individual

500 mL plastic tubs with water vials, cat food, and egg cartons. Post-capture isolation was at least two weeks to insure complete sexual maturity and to eliminate individuals that were harboring tachinid larvae. Individuals were housed in a 27°C room on a 12:12 h dark/ light reversed photoperiod.

Male-Female Mating Trials

Adult males and females were randomly paired for mating trials to determine whether male condition (un-manipulated) or other predictors (body size, overall degree of armaments = “armature score”) influence the likelihood of forced copulation. Each male was only used once, but females were used in two trials each to determine if they prefer any male traits (no-choice trials). There were at least two days between each female’s 1st and 2nd trial, and no oviposition took place in the interim. Indeed, no oviposition ever took place. Mating trials were conducted at 22°C during the dark portion of the cricket photoperiod. Each trial was preceded by a 30-minute acclimation period to allow the crickets to adjust to the colder temperature (which more accurately simulated autumn nighttime conditions). Each male-female pair was placed in a 14 x 8 x 10 cm plastic tub under a red light, to minimize light disturbance.

16 Each trial was three hours long, but up to six trials (in six tubs) were conducted simultaneously. Trials were not filmed but directly observed. I noted cricket behavior while minimizing movement and noise. I recorded the following outcomes: no copulation attempted by the male, successful voluntary copulation (where the female was mounted on top of the male and not being restrained by the male’s hind legs), unsuccessful forced copulation (where the male grabbed the female with one or both hind legs but the female struggled and escaped before genital coupling), or successful forced copulation (where the male restrained the female with one or both hind legs while the female was on the ground or on her side during genital coupling). A total of 101 mating trials were performed.

17 Results

Male-Female Mating Trials

Of the 101 males in the mating trials, 14 obtained voluntary copulation (VC), 4 successfully forced copulation (FC), 15 attempted forced copulation but did not succeed

(AFC), and 68 neither voluntarily copulated nor attempted to force copulation (NC).

Using this mating script of P. marmoratus (Figure 3), I analyzed whether there were differences between males at different points along the path. Logistic regressions were used to compare body size, condition, and armature score of males at each node.

Armature score was determined by combining three positively allometric armament traits

(Table 3) into one trait (Factor 1 in a Principal Components Analysis; eigenvalue =

2.648, 88.3% of variance explained; component loadings: spine length = 0.96, femur area

= 0.93, angle of tibial deflection = 0.93).

There were no differences between males that obtained a voluntary copulation (n

= 14) versus those that did not (n = 87) (Branch Point 1; Table 4). However, when males could either attempt a forced copulation (n = 19) or not (n = 68) (Branch Point 2),

“forcers” were significantly smaller (Figure 4A) and had smaller armaments (Figure 4B) than “non-forcers.” There was no difference in the trait “condition” (measured as residual mass) between forcers and non-forcers. At the last point in the path (Branch

Point 3), there were no differences between successful forcers (n = 4) and unsuccessful forcers (n = 15).

18 Table 3. RMA regressions of hind leg traits on body size for P. marmoratus males (n= 101). Variables were log-transformed. Traits that are positively allometric are considered armature. (Sqrt = square root)

Trait RMA slope ± SE r2 p Allometry

Spine length 4.67 ± 0.29 0.71 <0.01* Positive Asin sqrt (1° tib. deflection) 2.35 ± 0.13 0.76 <0.01* Positive Sqrt (femur area) 1.06 ± 0.03 0.94 0.05* Weakly Positive

Table 4. Parameter estimates from univariate logistic regressions. McFadden’s Rho-Squared is given. Categories assigned a state value of 1 are listed first for each comparison. One-tailed p- values are shown.a (FC = forced copulation)

Comparison Variable Parameter Z ρ2 p Estimate ± SE

Vol. cop. (n=14) vs. no vol. cop. (n=87) Body size 0.03 ± 0.7 -0.05 <0.01 0.481 (Branch Point 1) Condition 6.83 ± 9.8 -0.70 <0.01 0.243 Armature 0.68 ± 0.8 -0.83 <0.01 0.204 Score Attempted FC (n=19) vs. no attempted FC (n=68) Body size -1.12 ± 0.6 -1.82 0.04 0.035* (Branch Point 2) Condition 3.61 ± 8.6 -0.42 <0.01 0.337 Armature -1.34 ± 0.8 -1.77 0.03 0.039* Score Successful FC (n=4)vs. unsuccessful FC (n=15) Body size 0.01 ± 1.5 -0.01 <0.01 0.498 (Branch Point 3) Condition 1.55 ± 21 -0.08 <0.01 0.470 Armature -0.21 ± 1.5 -0.14 0.01 0.444 Score aA priori predictions justify the use of one-tailed p-values. At Branch Point 1, males that attained voluntary copulation were expected to be superior to males that did not voluntarily copulate. At Branch Point 2, males that attempted forced copulation were expected to be inferior to males that did not. At Branch Point 3, males that successfully forced copulation were expected to be superior to males that were unsuccessful.

forcers" forcers" 0 =

- "Forcers" "Forcers" = "Non 1, 0 2 2.5 3 3.5 4 4.5 5 Pronotum length

forcers forcers 0 =

- "Forcers" "Forcers" = Non 1,

0 -1.5 -1 -0.5 0 0.5 1 Armature score

Figure 4. Logistic regressions comparing traits of males that attempted to force copulation (“forcers”: n = 19, “State 1”) versus males that did not (“non-forcers”: n = 68, “State 0”). Smaller males (A) and males with a lesser degree of armature (B) were significantly more likely to be “forcers” than “non-forcers.”

20 Female Preference

Each female was used in two mating trials to determine if female preferences for male traits exist. Only females that survived to the second trial (n = 44; some females died in the interim between trials but never during trials) produced analyzable comparisons. The observed outcomes for each female are shown in Table 5. Of the observed outcomes, only “1st trial voluntary copulation, 2nd trial no voluntary copulation” had both sufficient sample size (n = 10 females) and biological relevance. Females that did not re-mate after their first mating can be considered unwilling to trade their first mate’s sperm for another’s.

I used paired t-tests to compare each female’s 1st and 2nd trial males (although I realize that that this is statistically questionable). When females mated with the 1st male but not the 2nd male, the 1st male was significantly heavier than the 2nd male (t = 2.41, df

= 9, one-tailed p = 0.02) (Figure 5A). (One-tailed p-values were used because 1st males were expected to be superior to 2nd trial males.) There were non-significant but collectively suggestive trends indicating the same pattern with respect to male body size

(t = 1.50, df = 9, p = 0.086), condition (t = 1.56, df = 9, p = 0.078), and armature score (t

= 1.16, df = 9, p = 0.14) (Figure 5B, C, D).

I used binomial sign tests to compute the probabilities of the observed differences

(+/-) between the 1st trial males and the 2nd trial males. The 1st trial males were significantly heavier (p = 0.011) and marginally significantly bigger (p = 0.055) and more heavily armed (p = 0.055) than the 2nd trial males (condition was not significantly different, p = 0.38). Taken as a whole, these trends indicate that females tended not to trade down after mating with a comparatively larger male.

21 Table 5. Frequencies of observed outcomes of each female’s 1st and 2nd trials. Bold outcomes were subjected to analyses of female preference.

Female’s 1st trial Female’s 2nd trial Frequency (# of females)

No cop No cop 16 Vol cop Vol cop 1 Att. forced cop Att. forced cop 1 Forced cop/att. forced cop No cop 2 No cop Forced cop/att. forced cop 10 No cop Vol cop 4 Vol cop No cop* 7 Vol cop Forced cop/att. forced cop* 3

*Females that mated with their 1st trial male but whose 2nd trial either resulted in no copulation or forced copulation/attempted forced copulation (1st trial vol cop, 2nd trial no vol cop) will be considered females unwilling to trade down from their first mate (n = 10). This is the only biologically meaningful category that can be analyzed.

0.5 5

0.4 4.5

4 0.3 3.5 0.2 Mass Mass (g) 3 0.1 2.5

0 Pronotumlength (mm) 2 1st trial male 2nd trial male 1st trial male 2nd trial male

0.04 1 0.02 0.5

0 -0.02 0 -0.04 -0.5 Condition -0.06 Armature Armature score -1 -0.08 -0.1 -1.5 1st trial male 2nd trial male 1st trial male 2nd trial male

st nd Figure 5. Paired t-tests comparing 1 and 2 trial males that were matched with females (n = 10) that mated with the 1st male and did not mate with the 2nd male. 1st trial males were significantly nd heavier (A) than 2 trial males. This trend was non-significant but collectively suggestive for body size (B), condition (C), and armature score (D).

22 Discussion

Forced Copulation as a Conditional Alternative Mating Strategy

I elucidated the nature of an alternative mating strategy in P. marmoratus.

Experiment 1 demonstrated that male somatic traits, such a body size, and sexually selected armaments are positively condition dependent. Experiment 2 showed that forced copulation is employed by males with small bodies and small armaments, and as such is a negatively condition dependent alternative mating strategy. These findings are in accord with the body of literature on alternative mating strategies (Cade 1980, Dawkins 1980,

Cheng et al. 1983, McKinney et al. 1983, Andersson 1994, Clutton-Brock & Parker 1995,

Alcock 2009).

By comparing traits of males at each branch point in the scripted mating pathway of P. marmoratus in Experiment 2, it is possible to determine the conditions under which forced copulation is likely to happen. There were no differences between males that were able to attain a voluntary copulation or not (Branch Point 1). However, when males could decide to try to force a copulation or not (Branch Point 2), “forcers” had significantly smaller bodies and armaments (both of which are condition dependent traits) compared to “non-forcers.” Although it was slightly troubling that male condition, measured as residual mass, was not a significant predictor of forced copulation, Gray and

Eckhardt (2001) showed that this method is not always the best measure of actual male condition (in their study, measured as male fat reserve).

I interpret the differences between males at Branch Point 2 as follows: smaller males, whose mating invitations were rebuffed, were “making the best of a bad job”

(Dawkins 1980) and seizing their only possible mating opportunity. There are several

23 possible reasons why larger males chose not to force at this same branch point. These males could have decided to simply wait for another mating opportunity with a different female in the future instead of spending the energy to chase and subdue an unwilling female (females do struggle and try to resist forced copulation attempts). Another explanation could be that these males might have been reluctant to antagonize a female

(in the trial) that might later consent to copulation. In either case, larger males were most plausibly banking on future reproduction, whereas for smaller males, the chances of future reproduction were probably minimal. Because of this, smaller males’ best chance of reproducing would be by forcing any female at hand (or at leg, as it were).

There were no differences between males at Branch Point 3 (whether an attempted forced copulation is successful or not). This lack of a difference was likely due to the very small sample size of the successful forcers (n = 4). It could be that larger forcers would be more successful than smaller forcers because they would be better at catching and subduing a female.

The results of the female preference analysis dovetail with the finding that smaller males are forcers. Females significantly refused to trade down to a lighter male after mating with a heavier male. This trend was non-significantly but suggestively repeated for male body size, condition, and armament score. Female unresponsiveness to smaller males after mating with larger males provides a selective pressure that would favor small males that try to force copulation.

My study was designed to resolve the paradox of a positively allometric coercive structure (the male hind leg). Although the largest males were the best equipped to coerce, they were not actually the ones performing forced copulation. Despite having

24 small armaments, small males attempted and succeeded at forcing females to mate.

Hence, we can conclude that the primary selective pressure for large armaments comes from intrasexual selection. Hind leg squeezing was most likely exapted (sensu Gould &

Vrba 1982) for mating coercion after initial evolution as a male-male agonistic trait.

Disadvantages of Forced Copulation for Males

When conditionally used, forced copulation is an alternative mating strategy because it often results in lower fitness compared to voluntary copulation (Thornhill

1980, McLain & Pratt 1999). Thornhill (1980) showed that only half of forced copulations resulted in sperm transfer in Panorpa scorpionflies, whereas all voluntary copulations led to sperm transfer. In the heteropteran Neacoryphus bicrucis, sexual coercion lowered female fecundity (number of eggs laid; McKlain and Pratt 1999).

McLain and Pratt hypothesized that the energetic costs of female fleeing and resisting behavior could have reduced their fecundity. They also suggested that female cryptic choice (sensu Eberhard 1996) could have lowered female investment in the progeny of coercive males.

Females of Pristoceuthophilus marmoratus have so far been unwilling to oviposit in a laboratory setting, despite being offered a variety of egg-laying substrates. Because of this, I have been unable to investigate whether males that force copulation have fewer offspring than males that attained voluntary copulation. However, the rate of successful completion of voluntary versus forced copulation (14 VC versus 4 FC) and the fact that forced copulation is employed by small males suggest that forced copulation is not the most successful strategy.

25 Despite its relatively lower fitness payoff, forced copulation is still a viable option for low quality males. These males would likely have no offspring otherwise. For this reason, forced copulation persists in numerous taxa as a conditional alternative strategy.

Forced Copulation in a Gift-less and Display-less Species

According to Thornhill (1980), forced copulation should be most likely in species where males provide direct benefits, such as nuptial gifts, territories, or nesting sites. In such species, not all males are able to provide these benefits, and thus have nothing to offer a female in exchange for a voluntary copulation. In Thornhill’s 1980 study of scorpionflies, males that could not acquire a dead insect or produce a salivary mass for females to eat during mating were likely to force copulation (by grasping females with a specialized abdominal structure). In many waterfowl, forced copulation is an alternative mating strategy employed by unpaired males that do not defend territories (Cheng et al.

1983, McKinney et al. 1983). Forced copulation also occurs in species whose mating displays are highly conspicuous to predators (Evans et al. 2003).

Unlike these gift-providing and display-producing species, P. marmoratus males do not offer any direct benefits or signal conspicuously. Although many other male ensiferans provide their mates with large, externally-attached offerings (a sperm-filled spermatophore + an edible spermatophylax) (Vahed 1997, Lehmann & Lehmann 2000), copulation in P. marmoratus consists of insertion of eversible male genitalia without any nuptial gift. Although P. marmoratus is parasitized by a tachinid parasitoid (species undetermined but most likely Dichocera lyrata Williston, see O’Hara & Gray 2004), the fly does not home in on any male song (camel crickets do not chirp) and parasitizes juveniles, females, and adult males indiscriminately (L. P. Conroy unpubl. data).

26 Courtship invitations by males consist of them spreading their legs and pushing their abdomen toward the female, actions that occur at close range.

Displays and gifts are apparently absent in P. marmoratus. But male fighting is common, occurring 60% of the time when two males are paired (Haley and Gray 2012).

Fight outcome may influence a male’s chance of mating, with larger males being more likely to win a fight and subsequently mate. In such cases, smaller “loser” males might force copulation to attain any mating at all. In my study, I did not include male fight outcome as a predictor of forced copulation, and forced copulation still occurred without any fight taking place. With the apparent irrelevance of male fights and lack of nuptial gifts and mating displays, in P. marmoratus male quality (body size) alone appears to be under strong enough female selection to render some males sufficiently unattractive that they pursue forced copulation.

The way(s) in which females assess male quality in this species are currently unknown. Visual inspection of male size seems unlikely in this nocturnal species, although perhaps antennation is sufficient for females to evaluate male physique. Haley and Gray (2013) showed preliminary evidence that females are attracted to the pheromones produced by tubercles on the male abdomen. Although pheromones have the potential to indicate a male’s genotypic quality in this species, other morphologically similar Pristoceuthophilus species lack tubercles altogether. Future research should involve further pheromone analyses (gas chromatography), and perhaps experimental manipulation of females’ antennae, in order to tease apart female assessment in a species where differences in male quality are enough to generate an alternative mating strategy.

27 CHAPTER 2: INTERSPECIFIC COMPARISONS OF ARMAMENTS AND BEHAVIOR IN PRISTOCEUTHOPHILUS

Introduction

Agonistic interactions are well documented among conspecific male individuals vying for access to females. Often, these fights are facilitated by male-specific weapons that are absent in females. Examples of these weapons are found in a wide range of animal taxa, with horns, antlers, spines, and tusks arming the ranks of rhinoceroses and rhinoceros beetles alike. The fighting advantage conferred by these weapons maintains their expression in a population; thus, armaments are primarily a result of intrasexual selection (Andersson 1994, Kotiaho 2002, Alcock 2009, Kim et al. 2011), although they often also serve as a basis for female assessment of male quality (Berglund et al. 1996).

Although armaments play a major role in male-male combat, there is no evidence in the literature of their use in male-female interactions. Harassment of females and mating coercion occur in many species, but they are typically mediated by brute force associated with larger male body size (McKinney et al. 1983, Wilgers et al. 2009) or by specialized coercive structures, such as the abdominal “gin trap” of sagebrush crickets

(Cyphoderris strepitans) (Sakaluk et al. 1995) and grasping antennal hooks in water striders (Rheumatobates rileyi) (Khila et al. 2012), that lack a role in male-male interactions.

Haley and Gray’s (2012) study of Pristoceuthophilus marmoratus Rehn, 1904 was the first to reveal a male armament (i.e., the male hind leg) that functions in both intrasexual and intersexual interactions. The novelty of this “dual-purpose” structure

28 warranted further exploration as to how widespread this trait might be among congeners.

Pristoceuthophilus marmoratus is one of twenty species that share the male hind leg armaments previously described (enlarged femur, femoral spines, bent tibia). These species were grouped in the unofficial subgenus “Nucifractor,” or nutcracker, in reference to the way that males leg-squeeze a person’s finger while being handled (T. J.

Cohn pers. comm.). In fact, P. marmoratus is comparatively less armed than many other species in this subgenus, which additionally possess a ventral triangular flange on the hind tibia and a secondary tibial bend dorsal to the flange. These two additional traits presumably contribute to leg-squeezing. Most of these nutcracker species are undescribed, although T. H. Hubbell and Cohn amassed a great store of morphological and geographic range descriptions about them (which they never published before they died). When Haley and Gray’s study was published, it was unknown whether any of these other nutcracker species exhibit male-male hind leg fights and/or hind leg-mediated mating coercion.

The scope of my study was to compare armaments of five species (P. ‘Huachuca summer,’ P. arizonae, P. ‘Madera,’ P. marmoratus, and P. ‘Mt. Pinos’), as well as to observe mating and fighting behavior for four of the five species (all except for P. arizonae) to investigate whether armaments are also “dual-purpose” in other species of

Pristoceuthophilus. Of the twenty nutcracker species, I initially selected three on the basis of accessibility and my ability to identify them. Collection sites were recommended by T. J. Cohn (pers. comm.). In 2013, I collected the undescribed P. ‘Mt. Pinos,’ (“P. dasyglossus” in Hubbell and Cohn’s unpublished notes) from McGill Campground on

Mt. Pinos in Los Padres National Forest, Kern County, California. For the other two

29 species, I went to Arizona, where Cohn claimed only two species of Pristoceuthophilus existed: a low elevation species, the undescribed P. ‘Madera,’ (“P. leptodrilus” in

Hubbell and Cohn’s notes) and a high elevation species, P. arizonae Hebard 1935. In

2013, I collected P. ‘Madera’ from Madera Canyon in the Santa Rita Mountains in

Coronado National Forest. A preliminary collecting trip in late September 2012 to high elevation Reef Townsite Campground in the Huachuca Mountains in Coronado National

Forest yielded a handful of adult male P. arizonae. However, on a return trip to the same site in July 2013, I discovered adult male Pristoceuthophilus crickets that were a drastically different color from the 2012 P. arizonae specimens (the 2012 P. arizonae specimens were a dull gray, whereas the 2013 July specimens were a deep mahogany- brown). As Cohn had informed us that only two species of Pristoceuthophilus occurred in Arizona, we concluded that we had discovered a species unknown to either him or

Hubbell, which I refer to as P. ‘Huachuca summer.’ Unfortunately, Cohn died in 2012, so we could not consult his opinion on the matter. However, additional morphological

(Figure 6; also see Results) as well as genetic comparisons (D. A. Gray unpubl. data;

Figure 7) support the distinctness of the two phenologically isolated morphs from Reef

Townsite Campground.

A B

D C

Figure 6. Caudal view of adult male pseudosternites from (A) published P. arizonae description (Hebard 1935), (B) P. arizonae reference specimen from University of Michigan Museum of Zoology, (C) P. arizonae specimen collected from Reef Townsite in October 2012, (D) putative new species P. ‘Huachuca summer’ collected from Reef Townsite in July 2013, and (E) P. ‘Madera’ (for comparison). Although similar, the pseudosternites of P. arizonae and P. ‘Huachuca summer’ are consistently distinguishable in that P. arizonae’s two bumps (see arrows) are more symmetrically rounded than P. ‘Huachuca summer’s, and the groove between them is shallower in P. arizonae. Photos by J. N. Hogue.

Figure 7. Phylogeny of the Pristoceuthophilus species used in this study, based on maximum likelihood analysis of 560 base pairs of the 16s ribosomal RNA gene (mitochondrial). Two samples were taken from each species. P. ‘Huachuca summer’ is genetically distinct from P. arizonae and is cladistically less closely related to P. arizonae than to the three other species. Farallonophilus cavernicolus Rentz 1972, another camel cricket in tribe Pristoceuthophilini, is the outgroup.

32 Methods

Collection and Housing

Collection of all five species was accomplished by nocturnal checks of oatmeal piles. Adults were collected for all species but P. ‘Mt. Pinos,’ of which juveniles were collected. From July 25-26, 2013, P. ‘Huachuca summer’ adults were collected from mixed conifer forest in Reef Townsite Campground in the Huachuca Mountains in

Coronado National Forest (Cochise County, Arizona; 31.43°N, 110.29°W, elev. 2180 m).

(Juvenile P. arizonae were also collected at the same time and at the same oatmeal piles as P. ‘Huachuca summer’ adults. Only one P. arizonae individual survived to adulthood in the lab in November 2013.) On September 29, 2012, P. arizonae adults were collected from the exact same trail in Reef Townsite Campground. Pristoceuthophilus ‘Madera’ adults were collected from October 3-5, 2013 from riparian woodland in Madera Canyon in the Santa Rita Mountains in Coronado National Forest (Pima County, Arizona;

31.71°N, 110.87°W, elev. 700 m). From October 1-2, 2013, P. marmoratus adults were collected from riparian woodland in Malibu Creek State Park in the Santa Monica

Mountains (Los Angeles County, California; 34.10°N, 118.73°W, elev. 200 m). (N.B., these were the same individuals from Experiment 2 in Chapter 1.) Pristoceuthophilus

‘Mt. Pinos’ early juveniles were collected from July 1-2, 2013 from mixed conifer forest in McGill Campground on Mt. Pinos in Los Padres National Forest (Kern County,

California; 34.81°N, 119.10°W, elev. 2300 m).

Upon collection, crickets were individually housed in a 27°C growth room on a

12:12 h dark: light reversed photoperiod. Each individual was placed in a 500 mL plastic tub with a cotton-plugged water vial, ad libitum Purina Cat Chow, and egg carton pieces

33 for shelter. Pristoceuthophilus ‘Mt Pinos’ juveniles were reared collectively in communal tubs (50-100 crickets per tub) until adulthood, at which point they were isolated in individual tubs.

Measurement of Body Size, Armaments, and Allometry

Male mass was measured immediately before behavioral trials (see below).

Measurements of pronota and hind leg traits were made after crickets were preserved in ethanol. Pronotum length was measured with digital calipers. Hind legs were photographed and analyzed in ImageJ. Six hind leg traits were measured, three of which are positively allometric in P. marmoratus (length of proximal femoral spine, femur area, and 1° angle of tibial deflection, one of which is isometric in P. marmoratus (femur length), and two of which are absent in P. marmoratus but are presumed to be armaments in the other four species (flange area, 2° angle of tibial deflection; Figure 8). Each pronotum and hind leg trait was measured twice, and the average of the two values was used in analyses. Measures of allometric slopes were made by regressing the six traits on pronotum length using Reduced Major Axis (RMA) regression (after square root- transforming areas and arcsine square root-transforming angles).

Figure 8. Hind leg of a P. ‘Huachuca summer’ adult male showing the traits measured. 1 = femur length, 2 = spine length, 3 = femur area, 4 = flange area, 5 = 1° angle of tibial deflection, 6 = 2° angle of tibial deflection. P. marmoratus lacks the flange (4) and the 2° tibial deflection (6).

Behavioral Trials: Male-Male Fights and Male-Female Mating

Behavioral trials took place at least two weeks after field collection to insure full sexual maturity and elimination of individuals harboring tachinid larvae. In each species, male-male fighting trials and male-female mating trials were performed (except for P. marmoratus, where male-male fighting data were taken from Haley and Gray’s 2012 study). Male-male trials consisted of two males plus one female for incentive, while male-female trials consisted of one male and one female. Individuals were randomly grouped or paired for trials. Each individual was used only once in a male-male trial or a

35 male-female trial, but some individuals were used in both a male-male trial and a male- female trial (although they were never matched with the same individual(s) from the previous trial).

Behavioral trials were conducted at 22°C during the dark portion of the crickets’ photoperiod. Each trial was preceded by a thirty-minute acclimation period to allow the crickets to adjust to the colder temperature (which more accurately simulated autumn nighttime conditions). Each trial took place in a 14 x 8 x 10 cm plastic tub under a red light, to minimize light disturbance.

Each trial was three hours long, but up to six trials (in six tubs) were conducted simultaneously. Trials were not filmed but directly observed. I noted the crickets’ behavior while minimizing movement and noise. I recorded whether or not a fight took place (male-male trials only) and the presence or absence of any matings, voluntary or forced (all trials). In addition, copulation duration was recorded.

Statistical Analyses

Analyses were performed in SYSTAT 13, with the exception of RMA regression, which was accomplished with Bohonak & Van Der Linde’s (2004) online software

(http://www.kimvdline.com/professional/rma.html). Transformations were made when data did not meet the assumptions of normality and homogeneity of variance.

36 Results

Measurements of Armament Allometry

RMA regression slopes for each trait and species are shown in Table 6. Spine length and 1° tibial deflection were positively allometric for all species. The two traits absent in P. marmoratus but present in the other four species, flange area and 2° tibial deflection, were also positively allometric. Somewhat unexpectedly, femur area was positively allometric only in P. ‘Madera’ and P. marmoratus. Finally, femur length was isometric in all species.

The steepness of allometric slopes varied among species (see Table 6, Figures 9-

14, although slopes were not statistically compared). Generally, P. ‘Huachuca summer,’ the smallest-bodied species (Figure 14), tended to have the greatest allometric slopes for armament traits. (See Appendix Figures 16-21, for individual regression graphs of each trait for each species.)

37 Table 6. RMA regressions of hind leg traits on body size for P. ‘Huachuca summer’ (n = 51), P. arizonae (n = 12), P. ‘Madera’ (n = 13), P. marmoratus (n = 101), and P. ‘Mt. Pinos’ (n = 61). Variables were log-transformed. If an RMA slope is significantly different from 1, a trait is considered allometric. Traits that are positively allometric are considered armaments. Species are listed in order of increasing body size (see Figure 14). (asin = arcsine, sqrt= square root, tib. = tibial)

Trait Species RMA slope r2 p Allometry ± SE

Spine length P. ‘Huachuca summer’ 6.24 ± 0.48 0.71 <0.01* Positive P. arizonae 3.66 ± 0.68 0.56 <0.01* Positive P. ‘Madera’ 5.28 ± 0.70 0.81 <0.01* Positive P. marmoratus 4.67 ± 0.29 0.71 <0.01* Positive P. ‘Mt. Pinos’ 2.53 ± 0.18 0.69 <0.01* Positive Sqrt flange area P. ‘Huachuca summer’ 4.05 ± 0.33 0.67 <0.01* Positive P. arizonae 2.77 ± 0.68 0.40 0.05* Positive P. ‘Madera’ 3.85 ± 0.57 0.76 <0.01* Positive P. ‘Mt. Pinos’ 2.40 ± 1.15 0.76 <0.01* Positive Asin sqrt (1° tib. deflection) P. ‘Huachuca summer’ 6.23 ± 0.56 0.61 <0.01* Positive P. arizonae 2.09 ± 0.53 0.36 0.06 Weakly positive P. ‘Madera’ 3.04 ± 0.47 0.74 <0.01* Positive P. marmoratus 2.35 ± 0.13 0.76 <0.01* Positive P. ‘Mt. Pinos’ 3.62 ± 0.35 0.43 <0.01* Positive Asin sqrt (2° tib. deflection) P. ‘Huachuca summer’ 6.44 ± 0.68 0.45 <0.01* Positive P. arizonae 2.48 ± 0.54 0.54 0.05* Positive P. ‘Madera’ 2.41 ± 0.60 0.32 0.05* Positive P. ‘Mt. Pinos’ 3.62 ± 0.44 0.14 <0.01* Positive Sqrt femur area P. ‘Huachuca summer’ 1.13 ± 0.08 0.74 0.11 Isometric P. arizonae 1.07 ± 0.14 0.82 0.63 Isometric P. ‘Madera’ 1.2 ± 0.1 0.92 0.07 Weakly positive P. marmoratus 1.06 ± 0.03 0.94 0.05* Positive P. ‘Mt. Pinos’ 1.03 ± 0.04 0.91 0.46 Isometric Femur length P. ‘Huachuca summer’ 0.94 ± 0.07 0.72 0.40 Isometric P. arizonae 0.89 ± 0.12 0.82 0.38 Isometric P. ‘Madera’ 1.17 ± 0.17 0.77 0.34 Isometric P. marmoratus 1.02 ± 0.04 0.91 0.62 Isometric P. ‘Mt. Pinos’ 1.02 ± 0.05 0.84 0.69 Isometric

3.1

2.6

P. 'Huachuca summer' 2.1 P. arizonae P. 'Madera'

P. marmoratus ln (spine length) 1.6 P. 'Mt. Pinos’

1.1

0.6 0 0.5 1 1.5 2 2.5 3 ln (pronotum length)

Figure 9. RMA regression slopes for ln (spine length) on ln (pronotum length). Color coding for each species is shown in the legend. The horizontal length of each line corresponds to the range of pronotal lengths in each species. The dashed line is the line of isometry (slope = 1). Spine length is positively allometric in all species.

39 2.4

1.9

1.4 P. 'Huachuca summer' P. arizonae P. 'Madera' ln (sqrt flange area) P. 'Mt. Pinos’ 0.9

0.4 0 0.5 1 1.5 2 ln (pronotum length)

Figure 10. RMA regression slopes for ln (sqrt flange area) on ln (pronotum length). Color coding for each species is shown in the legend. The horizontal length of each line corresponds to the range of pronotal lengths in each species. The dashed line is the line of isometry (slope = 1). Flange area is absent in

P. marmoratus but positively allometric in the other four species.

40 5.4

4.9

4.4

3.9

3.4 P. 'Huachuca summer'

tibial deflection))

° P. arizonae 2.9 P. 'Madera' P. marmoratus 2.4 P. 'Mt. Pinos’

ln(asin sqrt(1 1.9

1.4

0.9 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 ln (pronotum length)

Figure 11. RMA regression slopes for ln (asin sqrt (1° tibial deflection)) on ln (pronotum length). Color coding for each species is shown in the legend. The horizontal length of each line corresponds to the range of pronotal lengths in each species. The dashed line is the line of isometry (slope = 1). 1° tibial deflection is positively allometric in all species.

41 5

4.5

3.5

P. 'Huachuca summer'

tibial deflection)) 3

° P. arizonae P. 'Madera' 2.5 P. 'Mt. Pinos’

2 ln(asin sqrt (2

1.5

1 0 0.5 1 1.5 2 2.5 3 3.5 4 ln (pronotum length)

Figure 12. RMA regression slopes for ln (asin sqrt (2° tibial deflection)) on ln (pronotum length). Color coding for each species is shown in the legend. The horizontal length of each line corresponds to the range of pronotal lengths in each species. The dashed line is the line of isometry (slope = 1). 2° tibial deflection trait is absent in P. marmoratus but positively allometric in the other four species.

42 2

1.8

1.6 P. 'Huachuca summer' P. arizonae 1.4 P. 'Madera'

P. marmoratus ln (sqrt femur ln (sqrt femur area) P. 'Mt. Pinos’

1.2

1 0.7 0.9 1.1 1.3 1.5 1.7 ln (pronotum length)

Figure 13. RMA regression slopes for ln (sqrt femur area) on ln (pronotum length). Color coding for each species is shown in the legend. The horizontal length of each line corresponds to the range of pronotal lengths in each species. The dashed line is the line of isometry (slope = 1). Femur area is weakly positively allometric only in P. ‘Madera’ and P. marmoratus; it is isometric in the other three species.

43 2.9

2.7

2.5

2.3 P. 'Huachuca summer' P. arizonae 2.1

ln (femur ln(femur length) P. 'Madera' P. marmoratus 1.9 P. 'Mt. Pinos’

1.7

1.5 0.5 1 1.5 2 ln (pronotum length)

Figure 14. RMA regression slopes for ln (femur length) on ln (pronotum length). Color coding for each species is shown in the legend. The horizontal length of each line corresponds to the range of pronotal lengths in each species. The dashed line is the line of isometry (slope = 1). Femur length is isometric in all species.

44 Comparisons of Body Size and Relative Armament Size among Species

One-way ANOVAs were used to compare the following traits among all five species: mass, pronotum length, and body-size controlled hind leg traits (spine length/pronotum, flange area/pronotum, 1° tibial deflection/pronotum, 2° tibial deflection/pronotum, femur area/pronotum, and femur length/pronotum). Dividing hind leg traits by pronotum length allowed me to compare the relative investment in these traits among species. All traits were significantly different between species (Table 7).

Tukey’s Honestly Significant Difference tests were used to compare each pair of species post hoc (Appendix, Table 12). Although no general trends prevailed for all species for body size-controlled hind leg traits, the largest species, P. ‘Mt. Pinos,’ had the weakest 1° and 2° tibial deflection. Of particular relevance to my proposal that P. ‘Huachuca summer’ is a different species than syntopic but phenologically isolated P. arizonae were the significant differences between these species in body size (Fig. 15B), relative spine length (Fig. 15C), relative flange area (Fig. 15D), relative femur area (Fig. 15G), and relative femur length (Fig. 15H). Pristoceuthophilus ‘Huachuca summer’ was smaller than P. arizonae and had smaller relative hind leg traits (except for relative femur length, which was larger in P. ‘Huachuca summer’).

Table 7. Single-factor ANOVAs comparing traits of the 5 species. Hind leg traits are body size- controlled. See Table 12 in Appendix for between-species-pairs comparisons (Tukey’s HSD).

Trait F df p

Mass 71.89 4, 236 <0.001* Pronotum (body size) 50.17 4, 236 <0.001* Spine length/pronotum 4.36 4, 236 0.002* Flange area/pronotum 13.97 3, 137 <0.001* 1° angle of tibial deflection/pronotum 8.34 4, 236 <0.001* 2° angle of tibial deflection/pronotum 17.49 3, 137 <0.001* Femur area/pronotum 66.01 4, 236 <0.001* Femur length/pronotum 6.45 4, 236 <0.001*

45 0.5 D C B B 4 B B 0.4 A C 3 0.3 BC 0.2 2 Mass Mass (g) A AB 0.1 1 0.0 0 P. 'Hua' P. ariz P. P. mar P. 'MP' Pronotum length (mm) P. 'Hua' P. ariz P. P. mar P. 'MP' 'Mad' 'Mad'

C 0.03 BD BCD D 0.0016 B 0.0014 B AC ACD A 0.0012 AB

0.02 0.001 /mm)

2 0.0008 0.01 0.0006 0.0004

length (cm/mm) length(cm/mm) 0.0002

0.00 length(cm 0 Spinelength/pronotum P. P. ariz P. P. mar P. Flange area/pronotum P. 'Hua' P. ariz P. P. 'MP' 'Hua' 'Mad' 'MP' 'Mad'

E 14 AC F 7 A 12 ACD 6

ACD AD

10 5 A A /mm)

° 8 B /mm)

° 4

tibial

6 tibial

1 ° 3

4 2 B length(

length( 2

2 deflection/pronotum

0 deflection/pronotum 1 P. P. ariz P. P. mar P. 0 'Hua' 'Mad' 'MP' P. 'Hua' P. ariz P. 'Mad' P. 'MP'

0.1 0.3 BD BCD AC BCD

G D H A

0.08 BC C B 0.25

0.06 A

/mm) /mm) 2 0.04 0.2

0.02

length (cm/mm) length(cm/mm) length(cm

0 0.15 Femur Femur area/pronotum P. P. ariz P. P. mar P. Femur length/pronotum P. P. ariz P. P. mar P. 'Hua' 'Mad' 'MP' 'Hua' 'Mad' 'MP'

Figure 15. Comparisons of body and leg traits among species of Pristoceuthophilus. Letters above bars indicate significantly different species. Species are listed according to increasing mass (A) and body size (B). Hind leg traits (C-H) are controlled for body size. (C-F) are the only consistently positively allometric traits among all species. Means ± SE are shown. (P. ‘Hua’ = P. ‘Huachuca summer’ (n = 51), P. ariz = P. arizonae (n = 12), P. ‘Mad’ = P. ‘Madera’ (n = 13), P. mar = P. marmoratus (n = 101), P. ‘MP’ = P. ‘Mt. Pinos’ (n = 61))

46 Behavioral Trials

Behavioral trials were performed for all species except P. arizonae, which was excluded because of small sample size. Individuals were collected as adults in all species except P. ‘Mt. Pinos,’ for which I used lab-reared individuals that had been caught as juveniles. Male-male fights were not performed for P. marmoratus (data on fights were taken from Haley and Gray 2012), although male-female trials were performed for this species (as part of Experiment 2 in Chapter 1).

Results of the male-male trials for each species are presented in Table 8. Fights occurred when males either mutually or singly squeezed each other with one or both hind legs. Fights took place in all species. Because incentive females were also present during these trials, voluntary copulation and coercion attempts also took place in certain instances. Voluntary copulation was frequent in P. ‘Huachuca summer’ male-male trials, although less common in trials of other species. Females were leg-squeezed in all species except P. ‘Madera,’ although no successful forced copulations took place in species other than P. marmoratus.

Male-female trials (Table 9) did not resolve whether hind legs successfully serve a second function (mating coercion) in species of Pristoceuthophilus other than P. marmoratus. Voluntary copulation was most relatively frequent in P. ‘Huachuca summer.’ Females were leg-squeezed by males in P. ‘Huachuca summer’ and P. marmoratus, but successful forced copulation only occurred in P. marmoratus. No voluntary mating occurred in P. ‘Madera,’ and in P. ‘Mt. Pinos’ voluntary copulation took place in only 1 of 69 trials.

47 I performed a power analysis to determine whether the lack of successful coercion in species other than P. marmoratus might have stemmed from small sample sizes rather than an actual biological absence of function (Table 10). The number of male-female trials in P. ‘Huachuca summer’ and P. ‘Madera’ was not great enough to detect successful forced copulations, assuming a rate of forced copulation at least as low as in P. marmoratus (which in my study was only 4 out of 101 trials). The number of male- female trials in P. ‘Mt. Pinos’ exceeded the minimum requirement (25 trials), but no successful forced copulations were observed in this species. However, I attribute this result, and the general lack of mating and fighting behavior in this species, to artefacts of lab rearing (see Discussion).

Combined, the behavioral trials provide evidence of a male-male agonistic function in all species for which trials were performed, and strong suggestion of a sexual coercion function in P. ‘Huachuca summer’ and P. ‘Mt. Pinos’ (Table 11). Thus, armaments most likely serve dual purposes in P. ‘Huachuca summer’ and P. ‘Mt Pinos’ as well as in P. marmoratus.

Table 8. Male-male trials: frequencies of fighting, copulatory, and coercive behaviors performed. Bold values are used to indicate presence of a behavior (as in subsequent graphs).

No. of trials with… Species No. of trials Male- Voluntary Females leg- Successful male copulation squeezed by forced cop. fights males (attempted forced cop.)

P. ‘Huachuca 21 11 15 4 0 summer’ P. ‘Madera’ 16 2 4 0 0 P. marmoratus* 20 12 2 2 2 P. ‘Mt. Pinos’ 47 1 2 2 0

*P. marmoratus male-male trials were performed by Haley and Gray (2012). However, I performed the male-female trials of 2013 for this species.

48 Table 9. Male-female trials: frequencies of mating and coercive behaviors performed.

No. of trials with… Species No. of Voluntary Females leg-squeezed Successful trials copulation by males (attempted forced cop. forced cop.)

P. ‘Huachuca 16 11 3 0 summer’ P. ‘Madera’ 11 0 0 0 P. marmoratus 101 14 15 4 P. ‘Mt. Pinos’ 69 1 0 0

Table 10. Power analysis determining whether there were enough male-female trials to detect forced copulation.

Species No. of Proportion of Number of male- Minimum male- trials with forced female trials needed number of female cop. in P. to show effect (forced trials trials marmoratusᶧ* cop.) reached?

P. ‘Huachuca 16 4/101 (~1/25) 25 no summer’ P. ‘Madera’ 11 4/101 (~1/25) 25 no P. ‘Mt. Pinos’ 69 4/101 (~1/25) 25 yes

ᶧ This comparison assumes that the rate of forced copulation in other species (if it exists) is at least as high as the rate in P. marmoratus.

*My rate of forced copulation in P. marmoratus differed from Haley’s (she saw 7 successful forced copulations in 25 male-female trials), but I will use my rate for the sake of consistency and because it is more conservative.

Table 11. Presence or absence of fighting and coercive behaviors in the 5 species. Question marks (indicating no positive results as of yet) are most likely the product of sample sizes (see Table 10) and the artefacts of lab-rearing nymphs from a young age (P. ‘Mt. Pinos’).

Species Do males Do males Do males Are hind leg fight with leg-squeeze successfully force armaments dual- hind legs? females? females to mate? purpose?

P. ‘Huachuca Yes Yes ? Probably summer’ P. arizonae ? ? ? ? P. ‘Madera’ Yes ? ? ? P. marmoratus Yes Yes Yes Yes P. ‘Mt. Pinos’ Yes Yes ? Probably

49 General Observations of Voluntary Copulatory Behavior

Voluntary copulatory behavior in all species was similar to that previously recorded for P. marmoratus (Haley and Gray 2012). Males invited females to mount them by spreading their hind legs and pushing their abdomen back towards the female.

Females often antennated the posterior portion of male abdomen at this point.

Copulation duration, recorded as the time from the beginning to the end of genital coupling, was highly variable among species (F4, 45 = 113.3, p <0.001). Mean ± SE voluntary copulation durations were as follows (species listed from smallest-bodied to largest-bodied): P. ‘Huachuca summer,’ 17.1 ± 0.8 minutes (n = 26); P. ‘Madera,’ 145.7

± 18.0 min. (n = 4); P. marmoratus, 62.8 ± 5.5 min. (n = 14); P. ‘Mt. Pinos,’ 14.1 ± 0.5 min. (n = 3). Copulation durations were significantly different between all species except between P. ‘Huachuca summer’ and P. ‘Mt. Pinos’ (Tukey’s HSD; see Appendix, Table

13). There did not appear to be a trend between body size and copulation duration (see listing of copulation durations from smallest to largest species).

50 Discussion

Likely New Species P. ‘Huachuca summer’

According to the leading Pristoceuthophilus expert, the late Ted Cohn, only two species of Pristoceuthophilus occur in Arizona, P. arizonae and P. ‘Madera.’ In this study, I provide evidence that Arizona-dwelling P. ‘Huachuca summer’ is morphologically and genetically distinct not only from syntopic P. arizonae and sympatric P. ‘Madera,’ but also from all other species of Pristoceuthophilus in this study.

P. ‘Huachuca summer’ had a recognizably different pseudosternite and unusually high allometric slopes for armaments traits. P. ‘Huachuca summer’ also differed from P. arizonae and P. ‘Madera’ in relative investment in several armament traits.

Based on the multitude of these differences, we posit that P. ‘Huachuca summer’ is an undescribed and previously unknown species. However, we cannot rule out the possibility that P. ‘Huachuca summer’ might be a species previously known to Cohn

(although not known to occur in Arizona), documented somewhere in his unpublished notes as a species other than P. arizonae or P. ‘Madera.’ In order to resolve this issue, it would be necessary to sort through Cohn’s voluminous records and specimens at the

University of Michigan Museum of Zoology. But we can say with certainty that P.

‘Huachuca summer’ is neither P. arizonae nor P. ‘Madera.’

Comparisons of Allometries and Armaments

Traits that were previously shown to be positively allometric (and therefore armaments) in P. marmoratus were also positively allometric in the four other species in this study (with the exception of femur area, which was inconsistently positively

51 allometric). The two new traits introduced in this study, flange area and 2° tibial deflection, also proved to be positively allometric, implicating their role in leg-squeezing.

Biomechanical analyses would be necessary to show whether these two additional traits improve leg-squeezing efficacy. The most parsimonious explanation for the absence of these two traits in P. marmoratus would be a loss of these traits from the ancestral state in this relatively young species (see Figure 7). Interestingly, P. marmoratus, engages in frequent male-male fights (about 60% of male-male encounters ended in a fight; Table 8), and is the only species shown to successfully force-copulate. Somewhat unsettlingly, it would appear that the least armed species (in terms of total number of armaments) experiences the greatest selection for agonistic and sexually antagonistic leg use. More behavioral trials in the other species of Pristoceuthophilus are necessary to resolve this unintuitive conclusion.

Species showed variation in allometric slopes of armament traits. Allometric slopes tended to be highest for P. ‘Huachuca summer,’ the smallest species. Steeper allometric slopes of armaments indicate greater marginal fitness payoffs for larger individuals in a population (Bonduriansky & Day 2003, Kodric-Brown et al. 2006). In addition, allometric slopes may be increased by the strength of sexual selection on a trait

(Bonduriansky & Day 2003). The high rate of male-male fighting in P. ‘Huachuca summer’ (about 50% of male-male encounters ended in a fight; Table 8) seems to support a strong sexual selection pressure, although the same fails to hold true for P. marmoratus, which exhibited much weaker allometric slopes despite engaging in male-male fights at a higher rate than P. ‘Huachuca summer.’ Allometric slope steepness was not consistent among armaments for other species.

52 Relative armament investment also showed little in the way of trends among species. The only somewhat consistent pattern was that P. ‘Mt. Pinos’ tended to have the smallest relative armament investment (for spine length and 1° and 2° tibial deflection).

Cohn (pers. comm.) was of the opinion that P. ‘Mt. Pinos’ was the most highly armed of all Pristoceuthophilus species, but his observation was probably understandably biased by the large body size of P. ‘Mt. Pinos.’

Dual-Purpose Armaments in other Pristoceuthophilus?

Before Haley and Gray’s (2012) study, no one knew what the elaborately modified hind legs of male Pristoceuthophilus were used for. Haley and Gray’s work demonstrated not only that the legs function in male-male fighting, as might be expected, but also that the legs facilitate mating coercion. The apparent uniqueness of such a “dual- purpose armament” among all animals prompted me to see how widespread this trait might be among morphologically similar congeners.

My study revealed that male hind legs serve the first purpose, male-male fighting, in three additional species of Pristoceuthophilus. Leg-mediated sexual coercion attempts were present in two species, P. ‘Huachuca summer’ and P. ‘Mt. Pinos,’ but never ended in successful forced copulation. Because of this, I cannot conclusively say that hind leg armaments are dual-purpose in these species. However, because of low rates of forced copulation in P. marmoratus (4 out of 101 trials), small sample sizes in P. ‘Huachuca summer’ and in P. ‘Madera’ probably precluded successful observation of this behavior.

Sample sizes in P. ‘Mt. Pinos’ were very large: 47 male-male trials and 69 male-female trials. Yet out of all of these trials, only three voluntary copulations, two male leg- squeezes of females, and one male-male fight occurred.

53 The general lack of mating and fighting behavior in P. ‘Mt. Pinos’ is most likely explained by the fact that, unlike the other species in this study, P. ‘Mt. Pinos’ individuals were collected from the field as juveniles rather than as adults. Lab rearing appeared to have detrimental effects on this species: approximately half of females exhibited broken ovipositors as adults. Males appeared externally normal, but perhaps were hormonally abnormal or completely deterred by female deformities.

Pristoceuthophilus species are sensitive to laboratory conditions. Unlike Gryllus field crickets, which oviposit in the lab without hesitation, Pristoceuthophilus crickets have never laid eggs in my lab, despite being offered a number of egg-laying substrates (sand, potting soil, dirt from the place they were collected). It appears that nymphal development also does not occur normally in a laboratory setting in Pristoceuthophilus

(see also Experiment 1 of Chapter 1).

Future Directions

Although my study provided evidence that dual-purpose armaments are not unique to P. marmoratus, future studies with larger sample sizes (and using only adult individuals) will be necessary to confirm the first hints of sexual coercion indicated in other species here. Ideally, behavioral trials should be performed for all twenty

(described and undescribed) nutcracker species of Pristoceuthophilus. Given the death of

Pristoceuthophilus expert Ted Cohn, his failure to publish his copious notes on the many undescribed species, and the widespread distribution of these Pristoceuthophilus species

(throughout the American northwest and southwest), this is a daunting task. However, doing so would make it possible to answer many questions about this intriguing dual- purpose armament. Possible questions include: Is leg-mediated sexual coercion truly an

54 exaptation of male-male agonism? Do rates of leg use in male-male fighting versus sexual coercion differ among species, and if so, how is selection on armament allometries and investment affected? Answering these questions would help elucidate a unique phenomenon in sexual selection.

55 WORKS CITED

Alcock, J. 2009. Animal Behavior. Sinauer Associates: Sunderland.

Allen, G.R. & Simmons, L.W. 1996. Coercive mating, fluctuating asymmetry and male mating success in the dung fly Sepsis cynipsea. Animal Behaviour, 57: 737-741.

Andersson, M. 1986. Evolution of condition-dependent sex ornaments and mating preferences: sexual selection based on viability differences. Evolution, 40: 804- 816.

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

Arnqvist, G. & Rowe, L. 1995. Sexual conflict and arms races between the sexes: a morphological adaptation for control of mating in a female insect. Proceedings of the Royal Society of London B, 261: 123-127.

Arnqvist, G. & Nilsson, T. 2000. The evolution of polyandry: multiple mating and female fitness in insects. Animal Behaviour, 60: 145-164.

Arnqvist, G. & Rowe, L. 2002. Correlated evolution of male and female morphologies in water striders. Evolution, 56: 936-947.

Bateman, A.J. 1948. Intra-sexual selection in Drosophila. Heredity, 2: 349-368.

Berglund, A., Bisazza, A. & Pilastro, A. 1996. Armaments and ornaments: an evolutionary explanation of traits of dual utility. Biological Journal of the Linnean Society, 58: 385-399.

Bohonak, A.J. & van der Linde, K. 2004. RMA: Software for Reduced Major Axis regression, Java version. Website: http://www.kimvdlinde.com/professional/rma.html

Bonduriansky, R. 2007. The evolution of condition-dependent sexual dimorphism. The American Naturalist, 169: 9-19.

Bonduriansky, R. & Day, T. 2003. The evolution of static allometry in sexually selected traits. Evolution, 57: 2450-2458.

Cade, W. 1980. Alternative male reproductive behaviors. Florida Entomologist, 63: 30- 45.

Cheng, K.M., Burns, J.T. & McKinney, F. 1982. Forced copulation in captive mallards (Anas platyrhynchos): II. Temporal factors. Animal Behaviour, 30: 695-699.

56 Cheng, K.M., Burns, J.T. & McKinney, F. 1983. Forced copulation in captive mallards. III. Sperm competition. The Auk, 100: 302-310.

Clutton-Brock, T.H. & Parker, G.A. 1995. Sexual coercion in animal societies. Animal Behaviour, 49: 1345-1365.

Dawkins, R. 1980. Good strategy or evolutionarily stable strategy? In: Sociobiology, Beyond Nature/Nurture?, pp. 331-367. Westview Press.

Eberhard, W.G. 1996. Female Choice: Sexual Selection by Cryptic Female Choice. Princeton University Press: Princeton.

Endler, J.A. 1980. Natural selection on color patterns in Poecilia reticulata. Evolution, 34: 76-91.

Evans, J.P., Pilastro, A. & Ramnarine, I.W. 2003. Sperm transfer through forced matings and its evolutionary implications in natural guppy (Poecilia reticulata) populations. Biological Journal of the Linnean Society, 78: 605-612.

Gould, S.J. & Vrba, E.S. 1982. Exaptation - a missing term in the science of form. Paleobiology, 8: 4-15.

Gray, D.A. & Eckhardt, G. 2001. Is cricket courtship song condition-dependent? Animal Behaviour, 62: 871-877.

Haley, E.L. & Gray, D.A. 2012. Mating behavior and dual-purpose armaments in a camel cricket. Ethology, 118: 49-56.

Haley, E.L. & Gray, D.A. 2013. Abdominal tubercles of adult male camel crickets, Pristoceuthophilus marmoratus Rehn (Orthoptera: Rhaphidophoridae), produce cues attractive to females. Journal of Insect Behavior, 26: 804-811.

Hebard, M. 1935. Studies of Orthoptera in Arizona. Transactions of the American Entomological Society, 61: 111-153.

Hruskova-Martisova, M., Pekar, S. & Bilde, T. 2010. Coercive copulation in two sexually cannibalistic camel-spider species (Arachnida: Solifugae). Journal of Zoology, 282: 91-99.

Khila, A., Abouheif, E. & Rowe, L. 2012. Function, developmental genetics, and fitness consequences of a sexually antagonistic trait. Science, 336: 585-589.

Kim, H., Jang, Y. & Choe, J.C. 2011. Sexually dimorphic male horns and their use in antagonistic behaviors in the hornheaded cricket Loxoblemmus doenitzi (Orthoptera: Gryllidae). Journal of Ethology, 29: 435-441.

57 Kodric-Brown, A., Sibly, R.M. & Brown, J.H. 2006. The allometry of ornaments and weapons. PNAS, 103: 8733-8738.

Kokko, H. 2005. Treat ‘em mean, keep ‘em (sometimes) keen: evolution of female preferences for dominant and coercive males. Evolutionary Ecology, 19: 123-135.

Kotiaho, J.S. 2002. Sexual selection and condition dependence of courtship display in three species of horned dung beetles. Behavioral Ecology, 13: 791-799.

Lehmann, G.U.C. & Lehmann, A.W. 2000. Spermatophore characteristics in bushcrickets vary with parasitism and remating interval. Behavioral Ecology and Sociobiology, 47: 393-399.

McKinney, F., Derrickson, S.R. & Mineau, P. 1983. Forced copulation in waterfowl. Behaviour, 86: 250-294.

McLain, D.K. & Pratt, A.E. 1999. The cost of sexual coercion and heterospecific sexual harassment on the fecundity of a host-specific, seed-eating insect (Neacoryphus bicrucis). Behavioral Ecology and Sociobiology, 46: 164-170.

Møller, A.P. & Petrie, M. 2002. Condition dependence, multiple sexual signals, and immunocompetence in peacocks. Behavioral Ecology, 13: 248-253.

O’Hara, J.E. & Gray, D.A. 2004. Two new Orthopteran hosts of North American Polideini (Diptera: Tachinidae). Entomological News, 115: 171-172.

Patton, R.L. 1967. Oligidic diets for Acheta domesticus (Orthoptera: Gryllidae). Annals of the Entomological Society of America, 60: 1238-1242.

Rowe, L. & Houle, D. 1996. The lek paradox and the capture of genetic variance by condition dependent traits. Proceedings of the Royal Society B, 263: 1415-1421.

Sakaluk, S.K., Bangert, P.J., Eggert, A.K., Gack, C. & Swanson, L.V. 1995. The gin trap as device facilitating coercive mating in sagebrush crickets. Proceedings of the Royal Society B, 261: 65-77.

Stutt, A.D. & Siva-Jothy, M.T. 2001. Traumatic insemination and sexual conflict in the bed bug Cimex lectularius. Proceedings of the National Academy of Sciences, 98: 5683-5687.

Thornhill, R. 1980. Rape in Panorpa scorpionflies and a general rape hypothesis. Animal Behaviour, 28: 52-59.

Trivers, R.L. 1972. Parental investment and sexual selection. In: Sexual Selection and the Descent of Man, pp. 136-179. B. Campbell Ed.

58 Vahed, K. 1997. Copulation and spermatophores in the Ephippigerinae (Orthoptera: Tettigoniidae): prolonged copulation is associated with a smaller nuptial gift in Uromenus rugosicollis Serville. Journal of Orthoptera Research, 6: 83-89.

Vahed, K. & Carron, G. 2008. Comparison of forced mating behavior in four taxa of Anonconotus, the Alpine bushcricket. Journal of Zoology, 276: 313-321.

Wigby, S. & Chapman, T. 2005. Sex peptide causes mating costs in female Drosophila melanogaster. Current Biology, 15: 316-321.

Wilgers, D.J., Nicholas, A.C., Reed, D.H., Stratton, G.E. & Hebbets, E.A. 2009. Condition-dependent alternative mating tactics in a sexually cannibalistic wolf spider. Behavioral Ecology, 20: 891-900.

Zuk, M., Rotenberry, J.T. & Tinghitella. R.M. 2006. Silent night: adaptive disappearance of a sexual signal in a parasitized population of field crickets. Biology Letters, 2: 521-524.

59 APPENDIX

This appendix contains supplementary materials (Figures 16-21, Tables 12-13) referenced in the main text. Figures 16-21 are RMA regressions of male leg traits on body size (after log-log transformation) for each species, with individual data points shown (unlike the corresponding Figures 9-14 in the main text). Tables 12-13 show

Tukey’s HSD values for comparisons of leg traits (Table 12) and copulation duration

(Table 13) among species. The Tukey’s HSD values in Table 12 are displayed graphically in Figure 15, whereas the Tukey’s HSD values in Table 13 are described verbally on page 50.

60 A 3 B 3 2.8

2.5 2.6

2 2.4 2.2 1.5 2 1.8

1 ln (spine ln (spine length) ln (spine length) 1.6 0.5 1.4 1.2 0 1 0.7 0.9 1.1 1.3 0.9 1.1 1.3 ln (pronotum length) ln (pronotum length)

C 2.8 D 3

2.5

2.3 2

1.8 1.5

ln (spine ln (spine length) ln (spine length) 1.3 0.5

0.8 0 0.9 1.1 1.3 0.9 1.1 1.3 1.5 ln (pronotum length) ln (pronotum length)

E 2.6 2.4

2.2 2 Figure 16. RMA regressions of ln (spine 1.8 length) on ln (pronotum length) for (A) P. 1.6 ‘Huachuca summer’ (n = 51), (B) P. arizonae 1.4 (n = 12), (C) P. ‘Madera’ (n = 13), (D) P. ln (spine length) marmoratus (n = 101), and (E) P. ‘Mt. Pinos’ 1.2 (n = 61). The dashed line is the line of 1 isometry, while the solid line has the slope 0.8 and intercept specified by the RMA 1 1.2 1.4 1.6 regression. Spine length is positively ln (pronotum length) allometric in all species.

61 A 2.5 2.4

2.2

2 2

1.5 1.8 1.6

1 1.4 ln (sqrt flange area) ln (sqrt flange area) 1.2 0.5 1 0 0.8 0.7 0.9 1.1 1.3 0.9 1.1 1.3 ln (pronotum length) ln (pronotum length)

2.5 2.4 2.3

2.2

2.1 1.9 2 1.7 1.8 1.5 1.6 1.3 1.4

1.1 ln (sqrt flange area) ln (sqrt flange area) 1.2 0.9 0.7 1 0.5 0.8 0.9 1.1 1.3 1 1.2 1.4 1.6 ln (pronotum length) ln (pronotum length)

Figure 17. RMA regressions of ln (sqrt flange area) on ln (pronotum length) for (A) P. ‘Huachuca summer,’ (B) P. arizonae, (C) P. ‘Madera,’ and (D) P. ‘Mt. Pinos.’ The dashed line is the line of isometry, while the solid line has the slope and intercept specified by the RMA regression. Flange area is absent in P. marmoratus but positively allometric in the other four species.

62 6 2.6

5.5

5 2.4 4.5 4 2.2

3.5

tibial deflection))

° 2 ° 3 2.5 1.8 2

1.5 1.6 ln(asin sqrt 1 ln(asin sqrt (1 0.7 0.9 1.1 1.3 0.9 1.1 1.3 ln (pronotum length) ln (pronotum length)

2.8

2.6 3 2.4 2.2 2.5 2

1.8 2

tibial deflection)) tibial deflection))

° ° 1.6 1.4 1.5 1.2

1 1

ln(asin sqrt (1 ln(asin sqrt (1 0.9 1.1 1.3 0.9 1.1 1.3 1.5 ln (pronotum length) ln (pronotum length)

3.5

2.5

tibial deflection)) 1.5

° Figure 18. RMA regressions of ln (asin 1 sqrt (1° tibial deflection)) on ln (pronotum length) for (A) P. ‘Huachuca summer,’ (B) 0.5 P. arizonae, (C) P. ‘Madera,’ (D) P.

0 marmoratus, and (E) P. ‘Mt. Pinos.’ The ln(asin sqrt (1 1 1.2 1.4 1.6 dashed line is the line of isometry, while ln (pronotum length) the solid line has the slope and intercept specified by the RMA regression. 1° tibial deflection is positively allometric in all species.

63 5.5

2.2

5 4.5 2 4 1.8 3.5

1.6

tibial deflection))

° ° 3 1.4 2.5 2 1.2

1.5 1 ln(asin sqrt (2 ln(asin sqrt (2 0.7 0.9 1.1 1.3 0.9 1.1 1.3 ln (pronotum length) ln (pronotum length)

2.4

4.5

2.2 2 4 1.8 3.5

1.6 3

tibialdeflection)) tibial deflection))

1.4 ° ° 2.5 1.2 1 2

0.8 1.5 ln (asin (2 ln sqrt (asin ln(asin sqrt (2 0.9 1.1 1.3 1 1.2 1.4 1.6 ln (pronotum length) ln (pronotum length)

Figure 19. RMA regressions of ln (asin sqrt (2° tibial deflection)) on ln (pronotum length) for (A) P. ‘Huachuca summer’, (B) P. arizonae, (C) P. ‘Madera,’ and (D) P. ‘Mt. Pinos.’ The dashed line is the line of isometry, while the solid line has the slope and intercept specified by the RMA regression. 2° tibial deflection is absent in P. marmoratus but positively allometric in the other four species.

64 1.8 1.7 1.7

1.6

1.6 1.5 1.5 1.4 1.3 1.4 1.2

1.1 1.3

ln (sqrt femur area) femur ln (sqrt ln (sqrt femur ln (sqrt femur area) 1 1.2 0.9 0.8 1.1 0.7 0.9 1.1 1.3 0.9 1.1 1.3 ln (pronotum length) ln (pronotum length)

1.8 2 1.9

1.7

1.8 1.6 1.7 1.6 1.5 1.5 1.4

1.4 1.3

ln (sqrt femur area) femur ln (sqrt ln (sqrt femur ln (sqrt femur area) 1.3 1.2 1.1 1.2 1 0.9 1.1 1.3 0.9 1.1 1.3 1.5 ln (pronotum length) ln (pronotum length)

1.95

1.85 1.75 1.65 Figure 20. RMA regressions of ln (sqrt femur 1.55 area) on ln (pronotum length) for (A) P. 1.45

ln (sqrt femur ln (sqrt femur area) ‘Huachuca summer,’ (B) P. arizonae, (C) P. ‘Madera,’ (D) P. marmoratus, and (E) P. ‘Mt. 1.35 Pinos.’ The dashed line is the line of isometry, 1.25 while the solid line has the slope and intercept 1 1.2 1.4 1.6 specified by the RMA regression. Femur area is ln (pronotum length) positively allometric only in P. ‘Madera’ (C) and P. marmoratus (D); it is isometric in the other three species.

65 2.5 2.4 2.4

2.3 2.3

2.2 2.2 2.1 2 2.1 1.9

1.8 2

ln (femur ln length) (femur ln (femur ln(femur length) 1.7 1.9 1.6 1.5 1.8 0.7 0.9 1.1 1.3 0.9 1.1 1.3 ln (pronotum length) ln (pronotum length)

2.5 2.7

2.4 2.5

2.3 2.3 2.2 2.1

2.1

ln (femur ln length) (femur ln (femur ln(femur length) 2 1.9

1.9 1.7 0.9 1.1 1.3 0.9 1.1 1.3 1.5 ln (pronotum length) ln (pronotum length)

2.7

2.5

2.3

2.1 Figure 21. RMA regressions of ln (femur

ln (femur ln(femur length) length) on ln (pronotum length) for (A) P. 1.9 ‘Huachuca summer,’ (B) P. arizonae, (C) P. ‘Madera,’ (D) P. marmoratus, and (E) P. ‘Mt. 1.7 Pinos.’ The dashed line is the line of 1 1.2 1.4 1.6 isometry, while the solid line has the slope and intercept specified by the RMA ln (pronotum length) regression. Femur length is isometric in all species.

66 Table 12. Tukey’s HSD comparisons of body and leg traits between species (Pariz = P. arizonae, PHua = P. ‘Huachuca summer,’ PMad = P. ‘Madera,’ Pmar = P. marmoratus, PMP = P. ‘Mt. Pinos’).

Trait Species 1 Species 2 Difference p

Mass Pariz PMP -0.286 <0.001* Pariz PHua 0.017 0.985 Pariz PMad -0.082 0.205 Pariz Pmar -0.145 <0.001* PMP PHua 0.303 <0.001* PMP PMad 0.204 <0.001* PMP Pmar 0.141 <0.001* PHua PMad -0.099 0.009* PHua Pmar -0.162 <0.001* PMad Pmar -0.063 0.172 Pronotum Pariz PMP -0.578 <0.001* Pariz PHua 0.448 0.004* Pariz PMad -0.064 0.995 Pariz Pmar -0.313 0.083 PMP PHua 1.026 <0.001* PMP PMad 0.515 <0.001* PMP Pmar 0.265 0.001* PHua PMad -0.512 <0.001* PHua Pmar -0.761 <0.001* PMad Pmar -0.249 0.224 Spine length/pronotum Pariz PMP 0.006 0.072 Pariz PHua 0.008 0.005* Pariz PMad 0.002 0.957 Pariz Pmar 0.007 0.017* PMP PHua 0.002 0.516 PMP PMad -0.004 0.396 PMP Pmar 0.001 0.926 PHua PMad -0.006 0.055* PHua Pmar -0.001 0.905 PMad Pmar 0.005 0.158 Flange area/pronotum Pariz PMP 0 0.953 Pariz PHua 0.001 <0.001* Pariz PMad 0 0.394 PMP PHua 0.001 <0.001* PMP PMad 0 0.426 PHua PMad 0 0.121 1° angle of tibial deflection/pronotum Pariz PMP 5.466 <0.001* Pariz PHua 2.678 0.137 Pariz PMad 2.521 0.401 Pariz Pmar 3.373 0.021* PMP PHua -2.788 <0.001 PMP PMad -2.945 0.056* PMP Pmar -2.093 0.006* PHua PMad -0.157 1 PHua Pmar 0.695 0.821 PMad Pmar 0.852 0.933

67 Trait (Table 12 Cont.) Species 1 Species 2 Difference p

2° angle of tibial deflection/pronotum Pariz PMP 4.133 <0.001* Pariz PHua 1.629 0.133 Pariz PMad 1.482 0.391 PMP PHua -2.504 <0.001* PMP PMad -2.65 0.001* PHua PMad -0.147 0.997 Femur area/pronotum Pariz PMP -0.021 <0.001* Pariz PHua 0.007 0.004* Pariz PMad -0.005 0.622 Pariz Pmar -0.009 0.024* PMP PHua 0.028 <0.001* PMP PMad 0.016 <0.001* PMP Pmar 0.012 <0.001* PHua PMad -0.012 <0.001* PHua Pmar -0.015 <0.001* PMad Pmar -0.003 0.748 Femur length/pronotum Pariz PMP -0.014 0.005* Pariz PHua -0.017 <0.001* Pariz PMad -0.019 0.002* Pariz Pmar -0.01 0.097 PMP PHua -0.003 0.696 PMP PMad -0.005 0.645 PMP Pmar 0.004 0.327 PHua PMad -0.002 0.981 PHua Pmar 0.007 0.014* PMad Pmar 0.009 0.096

Table 13. Tukey’s HSD comparisons of voluntary copulation duration between species (PHua = P. ‘Huachuca summer,’ PMad = P. ‘Madera,’ Pmar = P. marmoratus, PMP = P. ‘Mt. Pinos’).

Species 1 Species 2 Difference p

PMP PHua -3 0.985 PMP PMad -131.59 <0.001* PMP Pmar -48.675 <0.001* PHua PMad -128.59 <0.001* PHua Pmar -45.675 <0.001* PMad Pmar 82.915 <0.001*