Sexual behavior in North American of the genera Magicicada and

by

John Richard Cooley

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Biology) The University of Michigan 1999

Doctoral Committee:

Professor Richard D. Alexander, Chairman Professor George Estabrook Professor Brian Hazlett Associate Professor Warren Holmes John R. Cooley © 1999 All Rights Reserved To my wife and family

ii ACKNOWLEDGMENTS

This work was supported in part by the Horace H. Rackham School of Graduate Studies, the Department of Biology, the Museum of Zoology (UMMZ), the Frank W. Ammerman Fund of the UMMZ Division, and by Japan Television Workshop Co., LTD. I thank the owners and managers of our field sites for their patience and understanding:

Alum Springs Young Life Camp, Rockbridge County, VA.

Horsepen Lake State Wildlife Management Area, Buckingham County, VA.

Siloam Springs State Park, Brown and Adams Counties, IL.

Harold E. Alexander Wildlife Management Area, Sharp County, AR.

M. Downs, Jr., Sharp County, AR.

Richard D. Alexander, Washtenaw County, MI.

The University of Michigan Biological Station, Cheboygan and Emmet Counties, MI. John Zyla of the Battle Creek Cypress Swamp Nature Center, Calvert County, MD provided tape recordings and detailed distributions of Magicicada in Maryland. Don Herren, Gene Kruse, and Bill McClain of the Illinois Department of Natural Resources, and Barry McArdle of the Arkansas State Game and Fish Commission helped obtain permission to work on state lands. Keiko Mori and Mitsuhito Saito provided funding and high quality photography for portions of the 1998 field season. Laura Krueger performed the measurements in Chapter 7. Artwork in Chapters 7 and 8 is by Melissa Anderson, Dan Otte, and John Megahan.

iii I thank Charles L. Remington for introducing me to the complexities of Magicicada, Oren Hasson, Tom Moore, Mark O’Brien, and Dan Otte for helpful discussion, and Chris Simon, for her pioneering genetic surveys of 13- year cicadas and our conversations about cicadas. Andrew Richards has been instrumental in the project to map the distributions of periodical broods in central Illinois. Although little of that information is included in this thesis, Andrew’s meticulous work and careful thinking have been an important component of all our various and ongoing cicada projects. My development and maturation as a graduate student were strongly influenced by the members of my doctoral committee, Richard D. Alexander, George Estabrook, Brian Hazlett, and Warren Holmes. None of this work could have been possible without the nearly 40 years of effort of Richard D. Alexander. I thank Dick for reigniting my early interest in . Most of the techniques, approaches, and questions in this work can be traced back to Dick, and the discovery of the new Magicicada species and the female wing flick signal are the legacies of his pioneering bioacoustical studies and his admonition to watch and measure everything. It was at Dick’s suggestion that we followed the behaviors of individually marked females kept separate from males except while under direct observation, allowing us to observe the complete male-female rapprochement sequence. What makes Dick uniquely unique is that he has an uncanny ability, and a proven track record, for being right about how things work. Dick has been tireless in his support, profound in his suggestions, and unwavering in his commitment to seeing that this work meets high quality standards. My friends David C. Marshall and Deborah Ciszek have also provided irreplaceable support, input, and collaboration. The fieldwork in this thesis is the joint collaborative effort of David and myself. The pronoun “I” in this thesis is but a formality; all “I’s” are truly “We’s.” As a collaborative team, we have accomplished far more than we could have separately, and it has been a joy to work with as insightful,

iv careful, and thorough a partner as David. This collaboration has been an important part of my own intellectual development. My thanks would not be complete without considering my family. I thank my parents, Rickie and Bill Cooley, for valuing education above all else, my brothers Bill and Jamie for their interest and technical expertise, my grandmothers, Frances S. Cooley for her determination, and Jane H. Yager for her creativity and patience, and my grandfathers, William H. Cooley for teaching me practical ways to construct things, and Richard S. Yager, for instilling in me a love of the strange, the natural, and the absurd. My wife, Louanne Reich Cooley, is a constant source of support. She designed an inexpensive and portable flight cage, contributed her knowledge of northern Michigan to the studies of Okanagana, and has been incredibly patient and understanding throughout this process. Last, I thank my study organisms. They lead bizarre lives—but then, so do we!

v PREFACE

Sexual behavior in North American cicadas of the genera Magicicada and Okanagana

by John Richard Cooley

Chairman: Richard D. Alexander.

The dense, raucous, mass emergences of North American periodical cicadas (Magicicada spp.) have attracted scientific study since the 17th century. Periodical cicadas are unique in their combination of long life cycles (13- or 17- years) and synchronous development, leading to periodical emergences of adults. Although their long life cycles make it difficult to conduct a longitudinal study of any one periodical cicada population, because populations in different regions of eastern North America are divided into asynchronous “broods,” or year classes, in almost any year, it is possible to study Magicicada. Over four years, I collected natural history information and studied the mating behaviors of 13- and 17- year Magicicada and two species in a non-periodical , Okanagana.

vi Chapter 1 is a review chapter, and it owes a major intellectual debt to my experiences as a co-author of a book chapter1 about the evolution of insect mating systems. As I became involved in that project, and as I started to construct my thesis, I realized that I would have difficulty discussing unless I clarified for myself, at least, the nature of mate choice. In this chapter, I review the reasons, starting with asymmetries in parental investment, why mates should be choosy. I continue with discussions of how sexual conflicts of interest shape mating systems, and the kinds of choice mechanisms and criteria invertebrates are most likely to employ. I conclude with some predictions about possible mechanisms and functions of mate choice in periodical cicadas.

Chapter 2 is largely a methodological chapter. I present data demonstrating that the methods I used to mark and confine cicadas were not likely to cause mortality or behavioral changes that would bias my studies.

Chapter 3 is a brief examination of “seminal plugs” in Magicicada. These structures are usually left behind in the female genital opening after mating. I demonstrate that these structures contain DNA and are likely composed of dried ejaculate, not other specially constructed or secreted materials. Seminal plugs do not prevent future mating attempts, but I suggest that they may be the products of antagonistic coevolution between females who are time-limited and males who can reduce threats to their paternity by imposing time-costs on females’ future mating attempts.

1 Alexander, R. D., D. C. Marshall, and J. R. Cooley. 1997. Evolutionary Perspectives on Insect Mating. Ch. 1 in B. Crespi, J. Choe, eds. The evolution of mating systems in and arachnids. Cambridge University Press.

vii Chapter 4 tests two hypotheses for female multiple mating in Magicicada: Females remate either to replenish depleted supplies or to effect postcopulatory mate choice. I demonstrate that interrupting a female’s mating tends to promote remating, while matings with an inappropriate, conspecific mate do not. Further, mated females do not actively seek additional mating opportunities, nor are they sexually attractive to males. Interrupted females, however, solicit second matings. These results are most consistent with the hypothesis that females remate for replenishment, and they disprove the hypothesis that females use postcopulatory choice mechanisms to make species-level mating decisions.

Chapter 5 documents a previously unknown female wing flick signal, used by females to signal sexual receptivity. No such signal has previously been reported in North American cicadas, although similar signals are known in some Australian and New Zealand species. This chapter documents the signal in different Magicicada species and also describes a bizarre effect (behavioral bisexuality) of a disease in males, as well as a specialized, competitive signal males use to “jam” the acoustical signals of interloping rivals. The newly described behaviors allow formulation and testing of a hypothesis for the two-part calling songs of Magicicada species such as M. septendecim . The two-part structure of these songs facilitate female discrimination of an individual male’s call against a background chorus.

Chapter 6 describes a new 13- year species, Magicicada neotredecim. M. neotredecim male calling songs and female mate acceptance criteria are distinct from those of the similar, synchronic and partially sympatric species, M. tredecim. Perhaps the most compelling piece of evidence that M. neotredecim and M. tredecim are separate species is that M. neotredecim has undergone reproductive character displacement as a result of its interactions with M. tredecim. The existence of this new species suggests two general hypotheses for speciation in Magicicada, the “nurse brood” hypothesis, and

viii the “canalization” hypothesis. Each could explain the origin of M. neotredecim, and each leads to specific predictions that may someday clarify how M. neotredecim came to be. This chapter ends with a prediction, based on the “nurse brood” hypothesis and patterns of reproductive character displacement, suggesting what other undiscovered Magicicada species might be like and where they might be found.

Chapter 7 explores the possibility that females discriminate among conspecifics on the basis of phenotypic symmetry. Symmetry has been promoted as a universal and reliable mate choice criterion. I begin by reviewing asymmetry theory in order to clarify the nature of the relationship between symmetry and mating success. I conclude that asymmetry has no special status as a mate choice criterion, but rather, it is like any other condition-dependent trait. I present the results of a study of symmetry and mating success in periodical cicadas demonstrating that males with more symmetrical forelegs are most likely to mate. However, the differential success of some males is more likely due to indirect mate discrimination than to evolved female mate choice mechanisms.

Chapter 8 compares the behaviors of Magicicada to members of another genus, Okanagana. I present the results of a 3- year biogeographic survey of the distributions of and O. rimosa in northern Michigan. I then describe the mating behaviors of both species, which involve females approaching stationary calling males, and contrast them to those of Magicicada, which involve males searching for females. From this contrast, I suggest hypotheses for how the unique of Magicicada may have evolved from mating systems such as those of Okanagana.

ix TABLE OF CONTENTS

DEDICATION...... ii

ACKNOWLEDGMENTS ...... iii

PREFACE...... vi

LIST OF TABLES ...... xii

LIST OF FIGURES ...... xv

LIST OF APPENDICES ...... xviii

CHAPTER

I. MATE CHOICE IN INVERTEBRATES, WITH SPECIAL REFERENCE TO PERIODICAL CICADAS (MAGICICADA SPP.) ...... 1

II. EXPERIMENTAL METHODS IN STUDIES OF PERIODICAL CICADAS: EXAMINATIONS OF POSSIBLE METHODOLOGICAL BIASES IN THE STUDY OF PERIODICAL CICADAS (MAGICICADA SPP.) ...... 30

III. MATE GUARDING AND THE NATURE AND POSSIBLE FUNCTIONS OF “SEMINAL PLUGS” IN PERIODICAL CICADAS (MAGICICADA SPP.)...... 53

IV. FEMALE MULTIPLE MATING AND MATE CHOICE IN A PERIODICAL CICADA, ...... 70

V. SEXUAL SIGNALING IN PERIODICAL CICADAS, MAGICICADA SPP...... 97

x VI. REPRODUCTIVE CHARACTER DISPLACEMENT AND ALLOCHRONIC SPECIATION IN PERIODICAL CICADAS WITH A NEW SPECIES, MAGICICADA NEOTREDECIM...... 152

VII. ASYMMETRY AND MATING SUCCESS IN A PERIODICAL CICADA, MAGICICADA SEPTENDECIM (HOMOPTERA: ) ...... 202

VIII. THE GEOGRAPHIC DISTRIBUTION AND MATING BEHAVIORS OF TWO CICADA SPECIES (OKANAGANA SPP.) IN NORTHEAST MICHIGAN ...... 235

xi LIST OF TABLES

Table

1.1. Mate discrimination and its alternatives ...... 27

2.1. Reported effects of methods for marking individual insects...... 46

2.2. Results of “seminal plug” analysis ...... 48

2.3. Cage density resampling results...... 49

2.4. Results of pretarsal clipping experiment...... 50

2.5. Summary of resampling results comparing remating frequencies in small bag cages to remating frequency in large flight cage...... 51

3.1. Study sites 1995- 1997...... 65

3.2. Mating durations in Magicicada septendecim, Okanagana rimosa, and O. canadensis...... 66

3.3. Results, female mating interruption experiment...... 67

3.4. Results, resampling algorithm to evaluate whether females with seminal plugs were unlikely to remate ...... 68

4.1. Mating durations in Magicicada septendecim...... 90

4.2. Mating interruption experiment: Likelihood of remating first or ovipositing first, and presence/absence of sperm plug ...... 91

4.3. Counts of eggnests in mating interruption experiments ...... 92

4.4. Results of playback/ mating status experiment...... 93

xii 4.5. Mating interruption and female wing flick response to playbacks of male calls ...... 94

4.6. Census data from female attractiveness experiment...... 95

5.1. Traits distinguishing Magicicada species...... 129

5.2. The sexual sequence in Magicicada species...... 130

5.3. Study Sites, 1995- 1998 ...... 131

5.4. Length of teneral period, in days ...... 132

5.5. Mated and unmated female M. septendecim responses to 2- minute playbacks of male calls...... 133

5.6. Wing flick responses of females to a series of alternating con- and hetero- specific calls ...... 134

5.7. Male M. septendecim responses to simulated wing-flick signals ...... 135

5.8. Male M. –decim movement direction following artificial wing flick signal...... 136

5.9. Experiments on male chorusing behavior...... 137

5.10. M. septendecim wing-flick signals responses to frequency-modified playbacks of male song...... 138

5.11. Contrasts of female wing flick responses to playbacks of complete artificial calling songs and partial artificial calling songs ...... 139

5.12. Effects of background chorus intensity on female wing flick responses to playbacks of complete artificial calling songs and partial artificial calling songs ...... 140

5.13. Effects of M. -decim “interference buzz.”...... 141

6.1. Traits distinguishing Magicicada neotredecim and other Magicicada species...... 184

6.2. Relationship of abdomen color to male call pitch and female pitch preference...... 185

xiii 6.3. Dominant frequencies of Magicicada septendecim chorus recordings in the UMMZ sound library...... 186

7.1. Tests for normality of signed, scaled asymmetries...... 224

7.2. Tests for measurement error and handedness ...... 225

7.3. Numbers of cicadas used in analysisof symmetry and mating success ...... 226

7.4. Pearson pairwise correlation matrix for character asymmetry...... 227

7.5. Pearson pairwise correlation matrix for character size...... 228

7.6. Pearson pairwise correlation matrix for character sizes and asymmetry....229

7.7. Comparison of asymmetry of mated and unmated males for each character...... 230

7.8. Results of tymbal rib asymmetry analysis ...... 231

7.9. Results of resampling algorithm to evaluate whether males with wing asymmetries were as likely to mate as normal males...... 232

8.1. Lengths of Okanagana canadensis calling and silent bouts...... 256

8.2. Mating durations in Okanagana and Magicicada septendecim...... 257

8.3. Responses of male cicadas to presentation of model cicada...... 258

xiv LIST OF FIGURES

Figure

1.1. Distribution of male traits in relation to female threshold acceptance value...... 28

1.2. Comparison of threshold acceptance criteria and best-of-n choice ...... 29

2.1. Survivorship of marked and unmarked cicadas...... 52

3.1. Photograph of gel showing RAPD-PCR amplification...... 69

4.1. Numbers of eggnests laid by female M. septendecim...... 96

5.1. Stylized sonogram of male call/female wing flick coutship duet for M. -decim ...... 142

5.2. Sonogram of male M. -decim call and female wing flick response...... 143

5.3. Sonogram of male M. -cassini call and female wing flick response ...... 144

5.4. Sonogram of male M. -decula call and female wing flick responses ...... 145

5.5 Comparison of male responses to a model female that moves only...... 146

5.6. Male responses to actions of model female that moves and makes sounds ...... 147

5.7. Sonogram of response of M. -decim male with Stage I cicadina infection to an M. -decim call ...... 148

5.8. Sonogram of response of M. -cassini male with Stage I Massospora cicadina infection to an artificial M. -cassini call...... 149

xv 5.9. Responses of female M. septendecim to playbacks of artificial, pure tone calls and portions of calls against an artificial background chorus ....150

5.10. Sonogram of male M. -decim call and interference buzz of nearby male...... 151

6.1. Geographic distribution of the seven periodical cicada species ...... 187

6.2. Comparison of typical 17- year Magicicada chorus with 13- year chorus containing M. tredecim and M. neotredecim...... 188

6.3. Spectrogram showing a two-banded, mixed-species chorus ...... 189

6.4. Air temperature effects on M. -decim chorus frequency ...... 190

6.5. Sonogram of male M. neotredecim calls illustrating consistency of individual male calling song pitch...... 191

6.6. Power spectrum of mixed M. neotredecim and M. tredecim chorus with accompanying frequency histograms of male song pitches and female pitch preferences ...... 192

6.7. Relative proportions of M. neotredecim and M. tredecim estimated from chorus recordings of the 1998 emergence of Magicicada 13-year Brood XIX...... 193

6.8. Geographic variation in dominant chorus pitch of M. neotredecim, showing reproductive character displacement in and near the region of overlap with the low-pitched M. tredecim...... 194

6.9. Schematic representation of calling song displacement in Brood XIX M. tredecim and M. neotredecim ...... 195

6.10. Geographic variation in dominant chorus pitch of M. tredecim, suggesting weak reproductive character displacement...... 196

6.11. Box plots of male calling song pitch, right wing length, thorax width, and first sternite width of M. tredecim (Sharp Co., AR) and M. neotredecim (Sharp Co., AR. and Piatt Co., IL)...... 197

6.12. Formation of an incipient Magicicada species by “nurse brood facilitation” of life cycle mutants...... 198

xvi 6.13. A model of Magicicada speciation using nurse brood facilitation and reinforcement of premating isolation...... 199

6.14. Two hypotheses for the origins of Midwestern 13- year cicadas ...... 200

6.15. A model of Magicicada life cycle evolution via canalization of a climate-induced life cycle shift...... 201

7.1. Characters used in analysis of phenotypic symmetry...... 233

7.2. Comparisons of cumulative asymmetry scores for mated and unmated males ...... 234

8.1. Locations of study sites, Okanagana canadensis, O. rimosa 1996- 1998 ...... 259

8.2. 1996 distribution map of O. rimosa and O. canadensis ...... 260

8.3. 1997 distribution map of O. rimosa and O. canadensis ...... 261

8.4. 1998 distribution map of O. rimosa and O. canadensis ...... 262

8.5. Location of O. rimosa study site in 1998...... 263

8.6. Sonogram of male Okanagana canadensis call...... 264

8.7. Sonogram of male Okanagana rimosa call ...... 265

8.8. Sketches of lateral view of typical Okanagana male genitalia...... 266

8.9. Sketches of lateral view of typical Magicicada male genitalia ...... 267

xvii LIST OF APPENDICES

Appendix

A. C++ code for mating system simulation...... 269

B. Pascal program code for comparing experimental cage densities to densities of free cicadas observed on cage-sized branches...... 287

C. 1995 Cage mating experiments...... 290

D. C++ code for resampling program to evaluate remating frequencies in large and small cages ...... 291

E. C++ code for seminal plug resampling algorithm ...... 302

F. Male frequency collection, 1998 Brood XIX ...... 313

G. Playback experiments, 1998 Brood XIX ...... 316

H. Locality, chorus pitch data, 1998...... 317

I. Holding jig for symmetry measurements...... 319

J. Symmetry and size measurements for mated and unmated cicadas ...... 320

K. Code for resampling algorithm to evaluate whether males with asymmetrical wing venation were over- or under-represented among males that did not mate in mating experiments ...... 322

L. Miscellaneous field notes on Okanagana...... 326

xviii CHAPTER I

MATE CHOICE IN INVERTEBRATES, WITH SPECIAL REFERENCE TO PERIODICAL CICADAS (MAGICICADA SPP.)

Abstract

Most invertebrates have limited adult experiences and mating opportunities. Thus, studies of invertebrate mate choice require a different approach from many vertebrate-based mate choice models, which implicitly allow to make use of past experiences and multiple mating opportunities. Careful consideration of the benefits realizable from mate discrimination as well as the ways in which discrimination may take place can reveal whether observed mating patterns are manifestations of evolved mate choice mechanisms or whether they are incidental effects of other behaviors. For example, for those insects lacking male parental investment and multiple mating opportunities, female choice criteria would not include material benefits, and mate choice mechanisms would be unlikely to involve comparisons. Alternatively, if existing male variation has minimal fitness consequences for females, then the costs of discrimination may outweigh the benefits, and any observed mating biases may be incidental rather than

1 the results of mate choice. With this in mind, I describe a series of studies designed to evaluate whether female periodical cicadas (Magicicada spp.) choose mates and if so, what choice criteria they are likely to use.

Introduction

Every aspect of behavior, life history, and ecology provides unique raw materials for the operation of . Many vertebrates have extended, juvenile periods in social contact with conspecifics, multiple opportunities for observing all aspects of adult behavior before becoming sexually mature, multiple breeding opportunities, and the possibility of gaining proficiency at acquiring mates. Insects, in contrast, often have single breeding opportunities (i.e., they are semelparous), no experience with conspecifics prior to mating, and few opportunities for rectifying errors. Life history patterns common among invertebrates often have no vertebrate parallels; thus, vertebrate- based models of sexual selection and mate choice (Bradbury 1981, 1985, Höglund and Alatalo 1995) may not accurately characterize invertebrate mating systems.

Rather than revise or modify vertebrate mate choice models, this review provides a general discussion of how the opportunities for mate choice and the benefits from choosing create the selective pressures shaping a given species’ mating system. This general discussion of mate choice leads to specific predictions for how invertebrates should be expected to choose mates or mating opportunities and concludes with a discussion of mate choice in periodical cicadas (Magicicada spp.), whose lek-like mating aggregations suggest a history of strong sexual selection.

A caveat

Mate choice itself may be difficult to observe directly, so behaviorists usually measure correlates of choice, such as polygyny. The hazards of assuming that observed mating patterns such as extreme polygyny necessarily result from choice are illustrated

2 by the following example: Suppose that in a population of 50 males and 50 females, males are always sexually receptive, females become receptive sequentially (thus the operational sex ratio is always 50 males:1 female), and females mate with the first male encountered. This scenario, written as a C++ computer program (Appendix A) and run for 10,000 iterations, leads to the following results: First, counting the maximum number of matings achieved by any male in each iteration, the 95th percentile is 5 matings. That is, in 9,500 of the iterations, at least one male achieved 5 matings and was notably more polygynous than others. In rare cases, males received up to 9 matings, and invariably, some males received no matings. At first glance, the population appears to be characterized by a 50:50 sex ratio, because a random sample of individuals drawn from the population would contain equal numbers of males and females. Thus, it would be tempting to conclude that the mating skew results from some form of mate choice. However, the operational sex ratio is male-biased, because at any given time only one female is receptive. The skewed operational sex ratio causes some males, by chance, to be much more successful than others even though females mate indiscriminately (see Boggs 1995 for a similar scenario). Some males were more successful than others simply because they were, by accident, in the right place at the right time.

This example, in which a strongly male-biased operational sex ratio leads to polygyny, highlights what should be the first question in any study of mate choice: Are the animals involved behaving such that they receive fitness benefits from participating in mate choice, or are some patterns better explained as spurious or incidental effects? The answer to this question lies in understanding the benefits possible from choice and the ways in which animals might obtain them.

Mate choice

The distinction between mate choice and its alternatives corresponds to Darwin’s original distinction between sexual and natural selection (Table 1.1; Darwin 1859, 1871,

3 Arnold 1983). As with Darwin’s forms of selection, mate choice and its alternatives are not easily separable. Mate choice results from individual actions that reduce the probability of mating with some members of the opposite sex (Parker 1983a). Mate choice includes both direct and indirect choice (see below), or all circumstances in which the identity of a mating partner affects fitness, either through genetic materials provided, or through potential mates’ differential ability to deliver resources or appear at times or locations most appropriate for mating (see Halliday 1983). In choosy species, members of the chosen sex can take actions that increase their own chances of mating success. The alternatives to mate choice are context choice and haphazard mating. Context choice occurs when fitness benefits result from restricting mating to particular contexts, but potential mates have no differential abilities to appear in the appropriate context and cannot take actions to increase their mating success; thus mating is independent of partner identity. Under haphazard mating, differences in mating context or partner identity have no influence on fitness.

Why choose mates at all?

Each sex strives to defend its post-zygotic investment against preemption or usurpation by members of the other sex (see Wiley and Poston 1996). Post-zygotic investment is the “parental investment” of Trivers (1972; also Alexander and Borgia 1979) and refers to any investment that is irretrievably committed to a particular zygote or collection of zygotes. In contrast, pre-zygotic investment is mating effort or resources committed to gametes (or collections of gametes) or to the production and/or delivery of gametes before fertilization. As with any artificial dichotomy, it is not always easy to distinguish pre- and post-zygotic investment. Even the term “post-zygotic investment” is misleading; the terms “investment in zygotes,” or “zygotic investment” are more accurate, because they refer to investment that has its most profound effects after fertilization, although it may be dispensed at any time before or after zygote formation

4 (Alexander and Borgia 1979). Variations in zygotic investment directly affect offspring quality. Thus, parental care by either sex is investment in zygotes, because it is delivered to particular offspring. Such care may actually be dispensed before the zygote is formed; for instance, egg yolk is investment by a female in her zygotes, because variation in the amount or quality of the yolk affects a particular zygote’s likelihood of success. In contrast, male courtship is mating effort, or investment in gametes, not zygotes, because different levels of investment affect only the likelihood that a male will, by delivering gametes, produce a zygote, not the success of any particular zygote. Like any resource, the zygotic investment of one sex is available for use and exploitation by members of the other.

Control of fertilization

Each sex strives to control the act of fertilization in order to defend its zygotic investment against preemption or usurpation. The costs of both finding additional mates and of zygotic investment influence sexual choosiness, but these factors are not independent. As members of one sex increase their zygotic investment, members of the other sex seek them out, so high zygotic investment is correlated with low mate search costs (Parker 1983a). Thus, the sex with the greater zygotic investment has the greater incentive to protect its investment and avoid mating indiscriminately, because its reproductive output is limited more by cost of the investment than by the cost of finding mating partners, while the sex with the lesser zygotic investment is limited by its ability to find willing mates and cannot afford to forego mating opportunities.

In many sexual species, females initially have the greater zygotic investment; they produce large, well-provisioned gametes specialized for post-zygotic success (Parker et al. 1972, Alexander and Borgia 1979). Delivery of resources to gametes makes the resources available to zygotes immediately after fertilization (Alexander and Borgia 1979). Associated with gamete provisioning is physical control and protection of the

5 gametes, probably for the following reasons: Females that evolved to dedicate greater and greater amounts of resources to zygotes also evolved to protect their investment by increasing physical control over them, such as by retaining them within the body after fertilization. As females increased their control, they gained confidence of the maternity of the zygotes they invested in and benefited by investing even earlier, in gametes. Thus, for many females, the distinction between prezygotic and zygotic investment is blurred, because investment in gametes translates directly into investment in zygotes. As females gained control over gametes, such as by retaining them inside their bodies for as long as possible, they also increased their control over the fertilization process.

Conflicts of interest between the sexes

The sexes’ relative levels of post-zygotic investment (which, as noted above, are related to confidence of parenthood) affect the degree of conflict over mating decisions. If the strategies of the sexes are identical, e.g., if both sexes have substantial zygotic investment, or if neither sex invests substantially and therefore neither discriminates, then there is little conflict over investment (Trivers 1972, Alexander and Borgia 1979, Parker 1979). When one sex invests more in zygotes than the other invests, the sexes will be in conflict over mating decisions (Parker 1979). Sexual conflict will, in each sex, favor the evolution of adaptations for controlling fertilization and thwarting the ability of the other sex to do so (see Rice 1996). Because the optimal mating strategy for each sex is contingent upon the other’s strategy, the sexes evolve in response to each other, with both sexes using a variety of strategies to gain control over fertilization (Parker 1979, 1983b, Alexander et al. 1997). Females with internal fertilization have an immediate control advantage; male responses to this advantage include genitalia that place sperm in close proximity to a female’s eggs (Parker 1970, Alexander et al. 1997), contributed substances that alter female physiology (see Lum 1961; Riemann and Thorson 1969; Burnet et al. 1973; Leopold 1976; Scott 1986; Chen et al. 1988; Kingan et al. 1993; Gromko,

6 Newport, and Kortier 1984; Gromko, Gilbert, and Richmond 1984, Wolfner 1996; Eberhard 1997 pp. 37- 40), or any number of other physiological or behavioral adaptations. Each sexes’ adaptations manipulate the other’s mating costs and benefits to effect control over fertilization. For instance, a male might restrict access to some valuable resource (such as a nesting site), trading resource access for matings. If the cost to a female of losing control of fertilization is less than the cost to her of losing access to the resource, then she will accept the trade. Each sexes’ adaptations to gain control may promote counter adaptations, leading to evolutionary races in which a wide range of strategies and possible counter strategies contribute to the complexity of the mating system (Parker 1983b). By inference, any species characterized by complex sexual interactions during sexual rapprochement likely exhibits some form of mate choice, because elaborate sexual interactions suggest that members of the chosen sex can take actions to improve their odds of mating. It is the mutual contingency of the sexes’ mating strategies, a consequence of mate choice, that leads to complex sexual adaptations (Alexander et al. 1997).

Selection and mate choice

Mate choice is an adaptation for producing genetically superior offspring, acquiring direct phenotypic benefits convertible into reproductive output, or both. Genetic benefits from choice fall into two non-exclusive categories: (1) Arbitrary benefits or (2) Offspring quality benefits. Arbitrary benefits consist of the genes coding for ornaments or other traits enhancing sexual attractiveness. Offspring quality benefits include the genetic materials necessary for producing offspring well equipped to survive and reproduce.

For example, in a given habitat, male Colias with the most appropriate enzyme variants not only gain from efficiencies in flight and foraging, but they are also more able to produce long and successful courtship displays (Watt et al. 1986). Females

7 mating with the most actively displaying males produce offspring that realize flight and foraging benefits and sons that produce more attractive sexual displays (Weatherhead and Robertson 1979; but see Kirkpatrick 1985). In this case, choosy females gain genetic benefits with aspects of both arbitrary and offspring quality benefits.

Females may also discriminate among males in order to receive phenotypic benefits. Such benefits include time and effort conserved by mating with easily located males, acquisition of male-provided nuptial resources, protection from injury or other harm, and reduced costs, or at least minimization of losses, associated with acquiescence to persistent male harassment. To the extent that male ability to provide phenotypic benefits is heritable, female choice for such benefits will impose selection on males, and thus phenotypic and genetic benefits are not sharply distinct.

Mechanisms of mate choice: Direct vs. indirect

Mate choice involves either direct assessment of potential mates or indirect biases (Wiley and Poston 1996, Sullivan 1989). In species with female choice, direct choice requires individual recognition and memory of male characteristics, and it requires females to perceive their different mating options. Indirect choice requires that the highest quality males have stereotypical characteristics, and that females act in ways that discriminate against males not exhibiting such qualities even though they may not have perceived the different mating options available to them.

Females may favor a particular male indirectly, rather than directly, on the basis of several possible cues: 1. Females may be attracted to the loudest or most easily located signals apart from any comparisons among those signals (The “passive attraction” hypothesis; Parker 1983a). 2. Females may accept only those males that have dominated in a contest with others. Male-male contests, in effect, provide résumés of the competitors’ lifetime achievements (Borgia 1979). A female may be absent for the contest itself, but the behavior of the contestants after the contest restricts her mating options. For

8 example, Burk (1983) discusses how losers in male-male contests in the field cricket Teleogryllus oceanicus reduce their calling; thus, females are more likely to mate with the victors. Moore and Moore (1988) present similar results for the cockroach Nauphoeta cinerea. 3. Females may favor males on the basis of some trait or resource. A female may feed on a copulatory gift or secretion, breaking off the mating when the gift is gone and thus disfavoring males with inadequate or poor-quality gifts. Thornhill (Thornhill and Alcock 1983) discusses how male scorpionflies providing inferior nuptial gifts are less successful as mates, because females stop mating once they have consumed the gift. 4. Females may engage in behaviors that discriminate among males according to their relative performance with respect to some environmental cue. Females may accept only males who, because of particular sensory, locomotory, or thermoregulatory abilities, are able to be in the right place at the right time. Honeybee mating swarms may be an example of such mate discrimination (see Winston 1987).

Others have classified these possibilities differently; for instance, Lloyd’s “passive filtering” would include aspects of 1 and 4 above; his “passive selection” is most similar to 2; “active filtering includes 1 and 4; while “active selection” includes aspects of 2, 3, and 4 (Lloyd 1979).

At the crudest level, all females exercise some degree of “indirect” mate choice by successfully mating only with living, sexually mature males (Parker 1979). It is trivial to point out that selection thus favors females who make this distinction, but this distinction is hardly the result of any specially evolved or elaborated adaptation. To the extent that more subtle male variation is heritable and has fitness consequences for females, selection can reinforce both sexes’ phenotypic cues and predictable behaviors involved in pairing, leading to evolved mate choice mechanisms.

Mate choice criteria

Mate choice criteria fall on a continuum of two extremes: (1) Females may, actively or passively, make relative comparisons among available mating partners (“best- of-n” choice, Janetos 1980), or (2) Females may accept as mates males who meet or exceed internally generated criteria (“threshold” or “minimal criterion” choice; Janetos

9 1980). Each of these models leads to distinct predictions about the kinds of behaviors males and females should exhibit (Janetos 1980, Alexander et al. 1997). Note that mate choice criteria do not constrain the evolution of mate choice mechanisms (either direct or indirect), and all four permutations of mate choice mechanisms and choice criteria are theoretically possible.

Direct relative comparisons: Best-of-n choice. Females exercising best-of-n choice seek to identify and mate with the best available male in a group of n- males. Best-of-n choice criteria are relative, open-ended, and require a comparison of potential mates. As long as such relative comparisons do not exact prohibitive costs, and as long as female fitness covaries with incremental changes in the resources or genetic materials provided by mates, selection will favor females using best-of-n choice. Best- of-n choice is functionally equivalent to threshold choice (see below) in which the acceptance criteria are changeable and influenced by the kinds of males a female encounters. The mechanisms involved in best-of-n choice need not be complex: Females could choose the best mate of those available by setting their acceptance threshold to the value of the first male encountered, raising it after encountering a higher quality male, and mating with the highest quality male when the threshold becomes fixed or remains stable after repeated encounters. Male-male contests could also restrict sexual access of all but the best males present. Females may require potential partners to simultaneously perform tests of skill to reveal relative quality. High costs, in terms of time and effort, and limited opportunities for making comparisons make short-lived, semelparous animals such as insects unlikely to exercise best-of-n choice (Alexander et al. 1997).

“Indirect” or “non-simultaneous” relative comparisons. Males, through active, intrasexual competition, can sort themselves in ways that limit sexual or resource access by lesser quality individuals (Cox and LeBoeuf 1977) and thus females’ mating options will be restricted as if they were exercising best-of-n choice. Females

10 making indirect, relative choices do not require simultaneous tests of skill; instead, they create situations in which males compete with each other before mating, and they will mate readily even if only one male is present.

Some “lekking” insects may provide examples of indirect best-of-n choice: If male insects form aggregations through which receptive females fly, and if males within an aggregation chase, catch, and mate with the females, these behaviors place premiums on male flight and perceptual capabilities and cause male mating success to covary with male quality. Lloyd (1979) points out how, by imposing sexual selection on such traits, females will cause males to compete to demonstrate their prowess (p. 337). If more than one male is present in such a “lek,” it provides an arena in which male competitive interactions sort males on the basis of quality. Insects that engage in multi-male courtship chases, such as honeybees (see Winston 1987 p. 207), midges (McLachlan and Cant 1995), some (Sullivan 1981, Pliske 1975) and syrphid flies (see Chapman 1954, Cooley “hotrod” model, unpublished) may all have mating systems characterized by this kind of choice.

Threshold or minimal criterion choice. Females exercising threshold choice discriminate against a class of potential mates that fail to meet minimal acceptance criteria (Janetos 1980). Males meeting the criteria are all equally acceptable, regardless of differences among them. Females acquire the choice criteria without regard to the characteristics of the male population, and once acquired, the criteria are fixed and unchangeable. The feature most distinguishing best-of-n from threshold choice is that under threshold choice, all acceptable males are equally acceptable independent of other qualities, and differences among acceptable males have no appreciable fitness consequences for females, while under best-of-n choice, in a given group of males, there is one best mate, because each male’s acceptability as a mating partner depends both on his own qualities and on the characteristics of the population at large. Some mate choice

11 mechanisms, such as “one step” or movable threshold mechanisms may appear to involve threshold choice, but from the standpoint of sexual selection they are more similar to best-of-n than to threshold choice because they involve relative comparisons and “thresholds” generated only after females sample the male population (Janetos 1980).

If threshold choice criteria exclude a large proportion of the males in the population, then females with lower acceptance thresholds will locate mates more readily than females using arbitrarily high criteria (Fig. 1.1). This increased efficiency may provide phenotypic benefits compensating any loss in mate quality, promoting an increase in the proportion of females with lower criteria in future generations. Thus, arbitrarily high criteria are unlikely to persist, and threshold choice is unlikely to lead to the extremes of sexual selection possible under best-of-n choice.

If male variation with respect to mate choice criteria is heritable, and if female choice criteria are unchanging and uniform across all females, then over the course of several generations, unless mutation rates are extremely high, threshold-based mate choice should diminish the proportion of unacceptable males, because all offspring will have fathers that met or exceeded the choice criteria. The exception, in which a long history of threshold choice would not lead to the elimination of unacceptable variants, is the case of species discrimination by threshold choice, because all the unacceptable mates would belong to a different species. Within a species, males subjected to threshold choice would be selected to exceed the acceptance threshold to the smallest extent that ensures that they will mate and that their offspring also exceed the threshold. After several generations of threshold choice, females would face a pool of potential mates most of whom slightly exceed the acceptance threshold, but among whom there would be few consequential differences. Thus, differences in male quality could be trivial in the face of the direct advantages females might gain from choosing mating times and locations that maximize their reproductive output. In species for which this is true, there

12 would be little evidence of mate choice; instead, females would mate so as to maximize direct benefits (such as associated with time or place of mating) to themselves while the identity of their mate would be largely irrelevant.

Many studies of insect behavior seem to suggest threshold choice. Scorpionflies (Hylobittacus apicalis ) obtaining nuptial gifts above a certain size threshold are accepted by females, while those offering smaller gifts are rejected or allowed only brief copulations with little or no chance of paternity (Thornhill 1980, Thornhill and Alcock 1983 p. 367). Females do not differentiate among males providing acceptable gifts, but appear to reject mates on the basis of a fixed gift size criterion or threshold.

Moore and Moore (1988) demonstrated that female roaches (Nauphoeta cinerea ) use absolute, rather than relative, preferences when mating. Male roaches formed stable dominance hierarchies, and females appeared to make best-of-n choices, because they decreased courtship time and increased time as male rank increased. However, female behavior was independent of the number of males present, and females deprived of any previous experience with males mated with the first male presented. Thus, females could not have made relative comparisons among dominant and subordinate males, and their mating behavior appears to involve an acceptance threshold, which all males in the experiment met or exceeded.

The action of an absolute mate acceptance threshold in Nauphoeta cinerea does not explain fully why female behavior was different for males of different dominance rank. Some aspects of male behavior provide clues: For example, subordinate males generally delay longer before initiating courtship (Moore and Breed 1986), courtship and mating duration are influenced by quantity and quality of male pheromone (Breed et al. 1980, Moore 1988), and these effects are independent of male-male competition. Furthermore, when females are presented with mating partners sequentially instead of simultaneously, they do not appear to discriminate (Breed 1983). The behavior of

13 females presented with solitary males is evidence for threshold choice; however, superimposed on this choice is another factor: Inherent qualities of dominant males apparently cause them to be more easily accepted or perceived by females, independent of the action of female acceptance criteria.

Passive attraction. “Passive attraction” (Parker 1983a) blurs the distinction between relative and threshold choice mechanisms. Under the passive attraction hypothesis, females may have an acceptance threshold, and males act in ways that ensure they will meet or exceed that threshold. If males employ mate attraction signals or complex courtship, the more readily females perceive such signals, the more likely they are to mate (Parker 1983a). Thus, females may tend to approach signaling males who are more obvious without directly making a relative comparison. For example, if two males were to sing at equal volume, one near a female and one far, then we would predict that the female should approach the nearest one. Similarly, if two males were at equal distance from a female, but one produced a louder signal than the other, the female should approach the louder male. This “passive attraction” can appear to be best-of-n choice if two males court simultaneously, because the female will respond only to the more apparent male. In reality, the female perceives only one of the males, and is not actually choosing. Animals assorting by passive attraction can appear as though they are selecting mates, but any choice is indirect at best. It is possible that the mate sorting qualities of female perception are incidental and provide no selective benefits over and above the benefits of being able to detect mates at all. Males, however, will be subject to selection, to the extent that male ability to make use of appropriate advertisement times, locations, channels, and intensities is heritable (Alexander et al. 1997). Although passive attraction in some ways sorts males relatively, it does not necessarily lead to male-female antagonistic coevolution because the sexes do not evolve strategies in response to each other. Instead, one sex evolves strategies to cope with the unchanging problem of being distinguishable from its surroundings or with the problem of prevailing in competition

14 with members of its own sex. However, if increased male display intensity thwarts female interests, intersexual competition becomes a more important selective force, and, rather than passive attraction, the mating system may be better characterized by a sexual conflict-based model such as chase-away selection (Holland and Rice 1998).

Jang and Greenfield (1996) demonstrated that female wax moths (Achroia grisella) were most likely to orient towards louder male signals, male signals with more rapid pulses, and male signals with longer pulses. Thus, the more evident a male’s signal, the greater the chance of a female response. Sometimes both a mate acceptance threshold and “passive attraction” seem to be involved in mate choice, as suggested by the example of Nauphoeta cinerea, (see above), as well as by male wing amputation experiments in fruit flies (Drosophila melanogaster ; Robertson 1982), in which the greater the area of wing removed, the longer the courtship duration before copulation. Male Drosophila without wings rarely mated, but as long as males had some wing area remaining and were able to produce some song, they eventually mated. A threshold, in which females require males to produce some song before mating, could explain why wingless males failed to mate, and female difficulty in perceiving quiet males could explain why partially amputated males courted longer before copulating.

Conflicts of interest and timing of mate choice

Mate choice could occur at any stage of the sexual sequence, including after copulation. It is in the interests of the choosy sex (usually the female) to effect mate choice as early in the mating sequence as possible. Early in the sequence, she retains substantial control over the mating decision; however, she also has little information on which to base her actions (Alexander et al. 1997). Thus, females should seek to cause males to reveal information about themselves before males have an opportunity to obtain a mating by force. Because the effectiveness of forcing increases as male and female come into closer proximity and more intimate contact, it is in the male’s interests either

15 to: (1) convince the female to commit to him early in the sequence, or (2) delay the female’s final mating decision until he approaches and has a high probability of compelling a mating in the event the female rejects him. Males should be selected, in so far as is possible, to refrain from including in calling songs or other sexual advertisements any information that could jeopardize their ability to approach or be approached closely by females. Thus, there is a conflict of interest between male and female over the nature and timing of information to be transmitted, and there is also a conflict over when in the sequence the female should make her final mating decision.

Multiply-mating females could take actions that affect paternity after they have engaged in one or more matings. However, if a female must copulate in order to gather information to make her mating decision, she places copulating males in a strong position to force fertilization upon her (Sivinski 1984). Because of the threat of forcing, postcopulatory mate choice should be restricted to rare cases in which females can, with certainty, separate copulation success and fertilization success (Eberhard 1985). Females exercising such choice should posses special adaptations for storing, manipulating, or discarding sperm (see Hellreigel and Ward 1998), or they should exhibit a strict pattern of last-male sperm priority so that they can retain control over fertilization by restricting sexual access.

General predictions about mate choice mechanisms

Females choosing the best-of-n available mates can favor rare, novel mutants for no reason other than their novelty or extremeness. The sexual selection advantages of such traits can cause them to spread even if they are costly in terms of natural selection. In contrast, sexual selection under threshold choice sets a lower limit for mate acceptability, and there is little or no mating advantage to exceeding that limit; for equally acceptable males facing threshold choice, novelty is at best irrelevant, and, since any costs are unopposed by sexual selection, novelty may be a liability (Fig. 1.2). One

16 exception is that threshold choice can favor novel traits that enhance a male’s ability to be detected or noticed (passive attraction); females favoring such traits are usually not favoring novelty per se, but augmentation of existing mate location and recognition mechanisms instead. For similar reasons, in species where the only benefits from mate choice are genetic, threshold choice is unlikely to evolve because only best-of-n choice can grant rare, novel mutant males a selective advantage allowing the spread of the mutation. Thus, best-of-n choice is a likely prerequisite of “runaway” evolution and the evolution of sexual ornaments.

Invertebrates’ general lack of male resource investment would seem to place a premium on females’ abilities to discriminate among potential mates (Emlen and Oring 1977, Alexander and Borgia 1979, Andersson 1994, Höglund and Alatalo 1995). Any benefits from discrimination are likely to be genetic, and obtainable only through some form of relative comparison; further, invertebrate life history characteristics make it likely that such comparisons are direct. Even if males provide no nuptial resources, and even if females obtain no genetic advantages for their offspring by selecting mates, females could still materially benefit by mating at a time or location optimal for their own interests, or choosing a mating context.

Mate choice in periodical cicadas

13- and 17- year periodical cicadas (Magicicada spp.1) provide an unparalleled opportunity to examine mechanisms of mate choice. Male Magicicada join mixed- species singing aggregations attractive to sexually receptive females. Males in these aggregations engage in chorusing (or “sing-fly”) behavior (Alexander 1968, 1975), in which they produce a small number of calling phrases (1- 3 in the M. –cassini and M.

1 For convenience the Magicicada cognate species groups are referred to using the following shorthand: M-decim: M. septendecim (17), M. tredecim (13), and M. neotredecim (13); M -cassini: M. cassini (17) and M. tredecassini (13); M. -decula: M. septendecula (17) and M. tredecula (13).

17 –decim species or a short calling bout in the M. –decula species) and then fly to another calling perch. Upon hearing a male’s calling phrase, a receptive female flicks her wings, producing a visual and acoustical signal similar to that reported in other cicada species (Chapter 5). Females signal only when they are mature and ready to mate, most signaling females copulate within several minutes, and most male mating attempts in the absence of a signal fail. A male perceiving a wing-flick signal ceases sing-fly behavior and begins courtship, engaging in an acoustical duet with the signaling female while moving towards her. Most females mate only once (Chapter 4).

Alexander (1975) considered periodical cicada male singing aggregations to be non-resource based leks, in which females have opportunities to discriminate among mates. Because unmated female Magicicada eventually oviposit normal-appearing, presumably unfertilized eggs (Graham and Cochran 1954, pers. obs.), it is unlikely that females visiting male aggregations receive male donated resources (or gain access to resources other than sperm) necessary for reproduction. Magicicada females could choose among mates at several stages of the sexual sequence: (1). They could choose during early stages of courtship when males call and females answer. (2). They could choose during late stages of courtship when males and females approach and achieve physical contact. (3). They could exercise postcopulatory choice by remating with higher quality males should they become available.

Any benefits females receive as a direct consequence of discriminating among potential mating partners must be genetic (because males provide no material resources), or must include phenotypic benefits other than male-donated gifts, such as reductions in the immediate costs of copulation by reducing time and effort wasted in locating appropriate mates or reducing risks of disease, damage, or predation. Alternatively, instead of choosing among potential partners, female cicadas may gain more from mating to

18 maximize direct phenotypic benefits to themselves, such as by selecting an appropriate time and safe location for mating independent of male identity.

Female discrimination against heterospecific males. The Magicicada species M. -decim, M. -cassini, M. –decula are sympatric over most of their ranges (Alexander and Moore 1962). Because females respond sexually only to conspecific male calls (Chapters 5, 6), hybrid matings are unlikely. Females avoiding hybrid matings conserve time and resources that would otherwise be wasted in fruitless matings. If female Magicicada mate choice behaviors evolved in the context of hybrid avoidance, and if the differences in the Magicicada species’ sexual behaviors result in part from past episodes of character displacement (Chapter 6), threshold choice mechanisms, based upon species-typical behaviors such as courtship, would be the simplest and least costly means for females to discriminate against inappropriate mates. Because there is little or no sexual conflict of interest over species identity if individuals of neither sex realize appreciable fitness benefits from cross-species matings, if females discriminate only against heterospecifics, males should be selected to cooperate instead of deceiving females.

Female discrimination against diseased males. Magicicada are vulnerable to an entomophagous fungal infection (Massospora cicadina Peck; Soper 1974, Soper et al. 1976). It is possible that the tendency or predisposition to become infected is inheritable. Infected cicadas are contagious and, in advanced stages of the infection, sterile. One simple way that females could reject infected mates is to fail to respond to their calls, or, if the fungus does not become contagious until males cease normal calling, to refuse to mate with silent males.

Even if there are no genetic differences among males concerning ability to resist or survive infection, the direct phenotypic benefits to females of avoiding infected males are substantial. Females should avoid contagious, infected males at all costs and should

19 reject infected males as early in the mating sequence as possible. Even if males exhibit no heritable genetic variation for susceptibility, females could benefit from discriminating against infected males because the fungus is invariably fatal, it reduces infected individuals’ fitness, it is avoidable, and females thus receive fitness payoffs from avoiding it. If this is an important influence on the Magicicada mating system, then females can avoid a high phenotypic cost by forcing males to engage in behaviors, such as joining singing aggregations and participating in complex courtship, likely to reveal male infection status and help females identify risky matings.

Unless female choice criteria are such that they exclude only sterile, fungus- infected males (who, on their own behalf, have no reproductive interest and thus no conflict of interest with the females) then male and female interests will conflict. Specifically, until infected males become sterile, they should strive to disguise their infection, while females should seek to cause males to court in ways that expose infections (Alexander et al. ms). For example, if females use male song as an “uncheatable honest indicator” of male quality, then they should be reluctant to mate unless their suitors sing, they should place a premium on song power output and the endurance of singing males, and they should discriminate against males with unusual or atypical songs characteristic of infected males. If infected and contagious males remain acceptable to females, then, perhaps because of the low frequency of infection, it is likely that females gain no net selective advantages from expending any effort to discriminate against rare infected males.

Female choice of mating time and location. Females may receive direct phenotypic benefits by restricting mating to optimal times and places. For instance, females may minimize exposure to predation by avoiding small populations and visiting only the densest aggregations (see Hamilton 1971; Alexander and Moore 1962, Williams et al. 1993). Females may also minimize risks by mating early in the day and terminating

20 copulation before cooler evening temperatures force them to remain relatively immobile until morning, or females may avoid mating until they have found a secure perch from which they are unlikely to fall; copulating pairs on the ground are relatively immobile and likely exposed to greater risk of predation from nocturnal rodents (pers. obs.). As long as males and females face different costs for finding mature, healthy mates, their interests over timing and location of mating will differ. If male periodical cicadas typically encounter only one or a few potential mates during their lifetimes, then if they encounter one at a suboptimal time or location, they will persist in courtship. Females, in turn, will be selected to resist forced matings. The net result will be that male and female daily activity patterns will differ, with females tending to become unreceptive before males stop chorusing, and with decreases in activity related to perceived costs of locating future mates. If female choosiness is related primarily to mating context, partner identity will be relevant only if males have heritable differences in their abilities to appear in the appropriate context.

21 Literature Cited

Alexander, R. D. 1968. . Pp. 167- 216 in: T. A. Sebeok (ed). Communication.Techniques of Study and Results of Research. (Indiana University Press: Bloomington, IN). Alexander, R. D., 1975. Natural selection and specialized chorusing behavior in chorusing insects. Pp. 35-77 in: D. Pimentel, ed. Insects, Science, and Society . (Academic Press, New York). Alexander, R. D., and T. E. Moore. 1962. The evolutionary relationships of 17-year and 13-year cicadas, and three new species. (Homoptera: Cicadidae, Magicicada). Misc. Publ. no. 121. Univ. of Michigan Museum of Zoology, Ann Arbor, MI . Alexander, R. D., and G. Borgia. 1979. On the origin and basis of the male-female phenomenon. Pp. 417- 440 in: M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects. (Academic Press: New York). Alexander, R. D., D. C. Marshall, and J. R. Cooley. 1997. Evolutionary Perspectives on Insect Mating. Pp. 1-31 in B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Alexander, R. D., J. R. Cooley, and D. C. Marshall. (ms). A specialized fungal parasite (Massospora cicadina) modifies the behavior of periodical cicadas (Homoptera: Cicadidae: Magicicada). Andersson, M. 1994. Sexual selection . (Princeton University Press: Princeton NJ). Arnold, S. J. 1983. Sexual selection: The interface of theory and empiricism. Pp. 67- 107 in: P Bateson, ed. Mate Choice . (Cambridge University Press). Boggs, C. L. 1995. Male nuptial gifts: Phenotypic consequences and evolutionary implications. Ch. 10 in: S. R. Leather and J. Hardie, eds. Insect Reproduction . New York: CRC Press. Borgia, G. 1979. Sexual selection and the evolution of mating systems. Pp. 19- 80 in M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects . (Academic Press: New York). Bradbury, J. W. 1981. The evolution of leks. Pp. 138- 169 in: R. D. Alexander and D. W. Tinkle, eds. Natural selection and social behavior . (Chiron Press: New York). Bradbury, J. W. 1985. Contrasts between insects and vertebrates in the evolution of male display, female choice, and . Fortschr. Zool. 31: 273-289.

22 Breed, M. D. 1983. Cockroach mating systems. Ch. 12 in: D. T. Gwynne and G. K. Morris, eds. Orthopteran mating systems: Sexual competition in a diverse group of insects . (Westview Press, Boulder Colorado). Breed, M. D., S. K .Smith, and B. G. Gall. 1980. Systems of mate selection in a cockroach species with male dominance hierarchies. Anim. Behav. 28: 130-134. Burk, T. 1983. Male aggression and female choice in a field cricket (Teleogryllus oceanicus ): The importance of courtship song. Ch. 5 in: D. T. Gwynne and G. K. Morris, eds. Orthopteran mating systems: Sexual competition in a diverse group of insects . (Westview Press, Boulder Colorado). Burnet, B., K. Connolly, M. Kearney, and R. Cook. 1973. Effects of male paragonial gland secretion on sexual receptivity and courtship behaviour of female Drosophila melanogaster. J. Ins. Physiol. 19: 2421- 2431. Chapman, J. A., 1954. Studies on summit-frequenting insects in western Montana. Ecology 35: 41-49. Chen, P. S., E. Stumm-Zollinger, T. Aigaki, J. Balmer, M. Bienz, and P. Böhlen. 1988. A male accessory gland peptide that regulates reproductive behavior of female D. melanogaster. Cell 54: 291- 298. Cox, C. R., and B. J. LeBoeuf. 1977. Female incitation of male competition: A mechanism in sexual selection. Am. Nat. 111: 317-335. Darwin, C. 1859. On the origin of species . A facsimile of the first edition, with an introduction by Ernst Mayr (Cambridge: Harvard University Press, 1964). Darwin, C. 1871. The descent of man, and selection in relation to sex. With an introduction by J. T. Bonner and R. M. May (Princeton: Princeton University Press, 1981) Eberhard, W. G. 1985. Sexual Selection and Animal Genitalia. (Harvard University Press: Cambridge, MA). Eberhard, W. G. 1997. Cryptic female choice. Pp. 32- 57 in B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Emlen, S. T., and L. W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science (Wash. D. C. ) 197: 215- 223. Graham, C., and A. B. Cochran. 1954. The periodical cicada in Maryland in 1953. J. Econ. Ent . 47: 242- 244. Gromko, M. H., M. E. A. Newport, and M. G. Kortier. 1984. Sperm dependence, of female receptivity to remating in Drosophila melanogaster . Evolution 38: 1273- 1282. Gromko, M. H., D. G. Gilbert, and R. C. Richmond. 1984. Sperm transfer and use in the multiple mating system of Drosophila. In: R. L. Smith, ed. Sperm competition and the evolution of animal mating systems . (New York: Academic Press).

23 Halliday, T. R., 1983. The study of mate choice. Pp. 3- 32 in: P Bateson, ed. Mate Choice . (Cambridge University Press). Hamilton, W. D. 1971. Geometry for the selfish herd. J. Theor. Biol. 31: 295-311. Hellriegel B., and P. I. Ward. 1998. Complex female reproductive tract morphology: Its possible use in postcopulatory female choice. J. Theo. Biol. 190: 179-186. Höglund, J. and R. V. Alatalo. 1995. Leks . (Princeton University Press: Princeton NJ). Holland, B., and W. R. Rice. 1998. Perspective: Chase-away sexual selection: Antagonistic seduction versus resistance. Evolution 52: 1-7. Janetos, A. C. 1980. Strategies of female mate choice: A theoretical analysis. Behav. Ecol. Sociobiol 7: 107-112. Jang, Y., and M. D. Greenfield. 1996. Ultrasonic communication and sexual selection in wax moths: Female choice based on energy and asynchrony of male signals. Anim. Behav. 51: 1095- 1106. Kingan, T. G., P. A. Thomas-Laemont, and A. K. Raina. 1993. Male accessory gland factors elicit change from “Virgin” to “Mated” behaviour in the female corn earworm moth Helicoverpa zea. J. Exp. Biol. 183: 61- 76. Kirkpatrick, M. 1985. Evolution of female choice and male parental investment in polygynous species: The demise of the “sexy son.” Am. Nat. 125: 788-810. Leopold, R. A. 1976. The role of male accessory glands in insect reproduction. Ann. Rev. of Emtomol. 21: 199-221. Lloyd, J. E. 1979. Sexual selection in luminescent beetles. Pp. 293- 342 in: M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects . (Academic Press: New York). Lum, P. T. M. 1961. The of some Florida mosquitoes. II. The male accessory glands and their role. Ann. of the Ent. Soc. of Amer. 54: 430- 433. McLachlan, A., and M. Cant. 1995. Small males are more symmetrical: Mating success in the midge Chironomous plumosus L. (Diptera: Chironomidae). Anim. Behav. 50: 841- 846. Moore, A. J. 1988. Female preferences, male social status, and sexual selection in Nauphoeta cinerea. Anim. Behav. 36: 303- 305. Moore, A. J., and M. D. Breed. 1986. Mate assessment in a cockroach, Nauphoeta cinerea Anim. Behav. 34: 1160- 1165. Moore, A. J., and P. J. Moore. 1988. Female strategy during mate choice: Threshold assessment. Evolution 42: 387- 391. Parker, G. A., 1970. Sperm competition and its evolutionary consequences in the insects. Biological Reviews 45: 525- 567.

24 Parker, G. A., 1979. Sexual selection and sexual conflict. Pp. 123- 167 in M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects . (Academic Press: New York). Parker, G. A., 1983a. Mate quality and mating decisions. Pp. 141-166 in: P Bateson, ed. Mate Choice , (Cambridge University Press). Parker, G. A. 1983b. Arms races in evolution—and ESS to the opponent-independent costs game. J. Theor. Biol . 101: 619-648. Parker, G. A., R. R, Baker, and V. G. F. Smith. 1972. The origin and evolution of gamete dimorphism and the male-female phenomenon. J. Theor. Biol. 36: 529-553. Pliske, T. E. 1975. Courtship behavior of the Monarch , Danaus plexippus L. Ann. Ent. Soc. of Amer. 68: 143- 151. Rice, W. R. 1996. Sexually antagonistic male adaptation triggered by experiental arrest of female evolution. Nature (Lond.) 381: 232-234. Riemann, J. G., and B. J. Thorson. 1969. Effect of male accessory material on oviposition and mating by female house flies. Ann. Ent. Soc. Am. 62: 828- 834. Robertson, H. M. 1982. Female courtship summation in Drosophila melanogaster. Anim. Behav. 30: 1105- 1117. Scott, D. 1986. Inhibition of female Drosophila melanogaster remating by a seminal fluid protein (esterase 6). Evolution 40: 1084- 1091. Sivinski, J. 1984. Sperm in competition. Ch. 3 in: Smith, R. L., ed. 1984. Sperm competition and the evolution of animal mating systems . (Academic Press: New York). Soper, R. 1974. The genus Massospora entomopathogenic for cicadas. Part I. of the genus. Mycotaxon 1:13-40. Soper, R., A. J. Delyzer, and L. F. R. Smith. 1976. The genus Massospora entomopathogenic for cicadas. Part II. Biology of Massospora levispora and its host Okanagana rimosa, with notes on Massospora cicadina on the periodical cicadas. Ann. Ent. Soc. Amer . 69: 89- 95. Sullivan, B. K. 1989. Passive and active female choice: A comment. Anim. Behav. 37: 692- 694. Sullivan, R. T. 1981. Insect swarming and mating. Florida Entomologist 64: 44- 65. Thornhill, R. 1980. Mate choice in Hylobittacus apicalis (Insecta: ) and its relations to some models of female choice. Evolution 34: 519- 538. Thornhill, R.A., and J. Alcock. 1983. The evolution of insect mating systems. (Cambridge: Harvard University Press). Trivers, R. L. 1972. Parental investment and sexual selection. In: B. Campbell, ed. Sexual selection and the descent of man, 1871- 1971 . (Aldine: Chicago).

25 Watt, W. B., P. A. Carter, and K. Donohue. 1986. Females’ choice of “good genotypes” as mates is promoted by an insect mating system. Science (Wash., D. C.) 233: 1187- 1190. Weatherhead, P. J., and R. J. Robertson. 1979. Offspring quality and the polygyny threshold: “The sexy son hypothesis.” Am. Nat. 113: 201-208. Wiley, R. H., and J. Poston. 1996. Indirect mate choice, competition for mates, and coevolution of the sexes. Evolution 50: 1371- 1381. Williams, K. S., K. G. Smith, and F. M. Stephen. 1993. Emergence of 13-yr periodical cicadas (Cicadidae: Magicicada): phenology, mortality, and predator satiation. Ecology 74: 1143- 1152. Winston, M. L. 1987. The biology of the honey bee. (Harvard University Press: Cambridge, MA). Wolfner, M. F. 1996. Tokens of Love: Functions and regulation of Drosophila male accessory gland products. Insect Biochem. Molec. Biol. 27: 179- 192.

26 Table 1.1. Mate discrimination and its alternatives. Identity of Kind of choice mate matters Benefits Selection mate choice yes genetic, material sexual context choice no material natural none no none none

27 Number of males

n, number of 28 males excluded

Male quality, q

t, threshold value

Figure 1.1. Distribution of male traits in relation to female threshold acceptance value, t. Females do not accept males of quality q < t. As t decreases, number of males excluded from mating (n) also decreases.

l, limit of male variation

t1, female acceptance threshold

t2, female acceptance threshold 29

A B C

number of males q, male quality

Figure 1.2. Comparisons of threshold mate acceptance criteria (A, B), and best-of-n choice criteria (C). Increasing width of line (A, B, or C) indicates increasing probability that a male of quality q will be accepted as a mate. (A) Under threshold choice, females do not accept males with quality q < t1 , excluding some males from mating. (B) At a lower value of t (t2), number of males excluded also decreases. Under threshold choice, all males of quality q > t have an equal probability of mating. (C) Under best-of-n choice, as male quality q increases, so does probability of being accepted as a mate. Best-of-n choice can lead to runaway evolution, because a mutant male with quality q > l (limit of male variation), is potentially more attractive than any other male in the population. CHAPTER II

EXPERIMENTAL METHODS IN STUDIES OF PERIODICAL CICADAS: EXAMINATIONS OF POSSIBLE METHODOLOGICAL BIASES IN THE STUDY OF PERIODICAL CICADAS (MAGICICADA SPP.)

Abstract

Periodical cicadas (Magicicada spp.) are ideal subjects for studies of mating behavior because adults can be kept and mated in captivity. However, for the results of captive studies to be applicable to natural populations, the methods used must not bias or alter mating behavior. For instance, marking allows focal-individual studies, but chemical effects of marking could cause premature death or reduced likelihood of mating. Further, cages could bias sexual behaviors, since confined females are limited in their abilities to escape persistent, harassing males. In this paper we demonstrate that the methods we use to mark and cage periodical cicadas are harmless. Our marking methods have no measurable effects, the densities in our cages are not unrealistically higher than densities of free cicadas on branches, and we uncover no evidence that matings observed in cages result

30 from forcing, or that smaller, denser cages increase males’ abilities to force matings upon sexually unreceptive females.

Introduction

Our long-term studies of periodical cicada (Magicicada spp.) mating behavior include experiments and observations in which we confine individually marked periodical cicadas in closed mesh cages (Chs. 3, 4, 5, 6, 7, 8). Increased mortality or morbidity are possible undesirable effects of such manipulation. Other effects could be more subtle; in mate choice studies, cage and marking effects could create false positive results by causing cicadas to be unusually active. Conversely, if manipulation disturbs cicadas or causes them to expend significant effort attempting to escape cages or remove marks, mate choice experiments could yield false negative results. Perhaps the most insidious problem of cage studies is that they may prevent or at least alter normal sexual rapprochement behaviors by placing members of both sexes in close, unsolicited proximity. To assess the seriousness of these problems in our studies of periodical cicadas, we present data addressing changes in mortality and sexual behavior due to marking or cage confinement artifacts and discuss ways of minimizing biases.

Methods and Results

Individual marking

The ease of use of different marking methods and their appropriateness for particular study designs are well reviewed (Southwood 1978, Walker and Winewriter 1981, Kearns and Inouye 1993). Above and beyond logistical concerns, the possible effects on the study subjects further constrain the choice of marking method. Commonly, studies making use of marking techniques assume no ill effects. Table 2.1 summarizes literature reporting effects of individual marking (not mass-marking) methods on study organisms. We located sources using published reviews and electronic databases

31 (Zoological Record, Wilson Indexes to Journal Articles, and Biological Abstracts) and included only papers containing data specifically addressing possible mortality or morbidity due to marking. Of several hundred papers examined, we found only 20 fitting these criteria. Table 2.1 is neither exhaustive nor prescriptive, but rather supplements earlier reviews (Southwood 1978, Walker and Winewriter 1981, Kearns and Inouye 1993). Few general trends are apparent, except that marks applied topically to sclerotized parts seemed to have few deleterious effects, while dusts, powders, or, in the case of soft-bodied insects, some solvents, must be used cautiously. The lack of general trends means that controlled testing of a marking method is the only way to ensure its appropriateness for any given study.

Individual marking has proven effective in our research on periodical cicadas (Magicicada septendecim (L.)), allowing us to track the behaviors of individual insects or classes of insects as they emerge, mature, mate, and die (Chs. 3, 4, 5, 6, 7, 8). One marking technique we have used is to apply colored enamel paint dots to the dorsal thoracic surface. Because published reports contain no clear information on the toxicity of the solvents in these paints (xylene, xylol and n- propoxy propanol) we conducted experiments to determine whether this marking method has adverse effects on our study animals.

Methods. We compared the survivorship and mating activity of marked and unmarked adult Magicicada septendecim in a recently logged clearing along Route 639 in the Horsepen Lake State Wildlife Management Area, Buckingham County, VA, in mid- May 1996. We placed two 1 x 2 m black fiberglass mesh bags over living vegetation and stocked each cage with 10 unmarked and 10 marked mature, unmated female cicadas. The cicadas were marked or handled before being put in the cages, and were assigned to the cages in no particular order.1 We marked cicadas by using a flat toothpick to apply one

1In this design, the 40 female cicadas, not the cages, are replicates; the females were divided into two cages for convenience and to prevent severe overcrowding. The experiments do not test for a cage effect, but rather test for effects of two different treatment regimes on the females. The two treatment regimes are entirely independent of the cages and cage assignments.

32 yellow and two blue dots of Testors’ lead-free gloss enamel (Testors Corp., Rockford IL, 61104) to the dorsal surface of the thorax. The primary solvents in Testors’ enamel paints are xylol and n- propoxy propanol, which are toxic to humans in high doses. We used no anesthetic during marking, and we were careful to prevent paint from contacting the wing articulations or the articulation between the head and thorax. The total surface area covered by paint was between 6 and 10 mm2. To allow matings, we placed ten sexually mature adult males collected from an active chorus into each cage. Cicada densities in these cages were thus high (see below), but intentionally so, to expose any differential survival ability of marked and unmarked cicadas. From May 23- June 4, we censused the cages 8 times by counting and removing dead cicadas and recording whether or not each was marked or showed evidence of a mating sign (or “seminal plug;” White 1973). We scored females exhibiting seminal plugs as having mated; those lacking the mark as not having done so. We preserved all cicadas in 70% ethanol, and at the final census, we removed all remaining cicadas.

Results. We compared the survivorship of the marked and unmarked cicadas using Cox regression analysis, implemented with the PHREG function in SAS (SAS Institute Inc., 1996). This technique allowed us to include individuals who survived through to the end of our observations. The small difference in survivorship curves for painted and unpainted individuals was not statistically significant (Wald Chi-Square-2.32, 1 d.f., P £ 0.13). If anything, marked cicadas seemed to survive longer than unmarked cicadas, although this pattern is not statistically significant. The two cages do appear to have different mortality rates (Wald Chi-Square=14.02, 1 d.f., P £ 0.0002), attributable to the death of substantial amounts of vegetation in one cage, although the increased mortality in this cage was distributed uniformly among marked and unmarked cicadas.

33 A chi-square goodness-of-fit analysis of seminal plug presence/absence data reveals no significant pattern. Marked and unmarked females were equally likely to exhibit this evidence of mating. Data are summarized in Table 2.2 and Figure 2.1.

Cage artifacts

In our studies of periodical cicada behavior, we make extensive use of mesh bag cages. Cages allow recovery of marked individuals, facilitate experiments with manipulated individuals, and permit isolation from free cicadas so that observations may be restricted to cicadas whose entire adult histories are known. Our cages consist of fiberglass hardware cloth folded over living vegetation suitable for feeding or oviposition, giving an approximate volume of 200 liters. We are careful to construct cages so that the highest points, in which cicadas tend to congregate, have few corners or narrow spaces that might trap or cause injury. In experiments where we watch and record cicada behaviors, we typically stock cages with 10- 20 cicadas; when we are less concerned about the details of behavior (such as our paint marking experiments), or for storing cicadas, we typically place 20- 40 cicadas per cage.

Each sex appears subject to sex-specific cage mortality. Females, whether mated or not, oviposit in cage branches, sometimes to a degree that kills or weakens the vegetation. This problem can be alleviated by moving ovipositing females to fresh branches on a regular basis. Males, especially those caged near unmated females, tend to damage themselves by struggling with each other or by attempting to escape. The severity of this effect can be reduced by housing males at low densities somewhat away from females. Since both of these problems could severely compromise long-duration experiments, we have kept cage experiments short and maintained and stocked cages in a manner that minimizes these effects.

34 Of greater concern are the ways in which cages may alter cicada sexual behavior by unnaturally confining males and females in close proximity, perhaps promoting male sexual coercion. For example, females confined with males in cages are denied the opportunity to participate in normal rapprochement behaviors. If critical components of the mating sequence occur during rapprochement, such as if females normally accept or reject a courting male before he comes within one meter of her, females in small cages would not be able to engage in normal mate rejection behaviors. Observations do not support the view that male forcing behavior is generally successful in periodical cicadas; of approximately 100 courtships of caged cicadas observed in 1996, three appeared to involve male forcing, and only one of these three resulted in copulation. Nevertheless, it is still possible that cages may facilitate male forcing behaviors, and thus confinement in cages may compromise mating experiments.

We used three approaches to evaluate the possibility of a cage effect. First, to determine whether our experimental cage densities are unrealistically high, we compared the densities in our cages to densities of natural cicada aggregations. Because of the difficulties of sexing uncaged cicadas, our analysis concerns only cicada densities and not sex ratios within cages. Second, we caged both normal and physically handicapped males with receptive females and examined the relative mating success of each class of male. Comparing the success of normal and handicapped males provides information about forcing behaviors and whether such behaviors might be exacerbated by cage studies. Last, we compared female remating frequencies in a large 4m x 4m x 4m screen “flight cage” to frequencies in small bag cages. Mated females are generally unreceptive (Chapter 4) so extreme polyandry in this experiment would indicate that confinement reduces the effectiveness of female resistance. If cages affect rapprochement or escape behaviors, then the severity of a cage effect should depend on the size of the cage. Demonstrating similar remating frequencies in large and small cages would suggest the absence of a cage effect.

35 Cage density surveys; methods. On May 14 and 17, 1996, an observer (JRC) walked along Virginia Route 639, an unpaved road with little traffic, and counted the numbers of cicadas visible on 39 branches of approximately the same size as those inside our mesh bag cages. We developed a computer resampling algorithm (Appendix B) to evaluate the extremeness, relative to our branch scan data, of specified cage densities. Given a hypothetical cage density, the algorithm randomly picks, with replacement, a value from the 1996 branch scan data and calculates whether the picked value is more or less extreme than the specified density by comparing the distances to the scan data mean. The algorithm repeats these calculations 10,000 times and reports the percentage of iterations in which the picked value was more extreme than the specified value. This percentage is analogous to a P-value indicating the likelihood of, by chance alone, picking a value from the data set more extreme than the specified cage density. The algorithm calculates such P- values for many hypothetical cage densities, generating 95% confidence intervals for the 1996 scan data.

Results. Our branch survey counts ranged from 3 to 27 (mean 12.61, ± 5.6). The resampling algorithm suggests that densities of up to 25 cicadas per caged branch are not extreme in comparison to the 1996 scan data (Table 2.3) and are comparable to natural cicada densities on similar branches. Greater densities are atypical of 1996 scan data, and behavioral data from such cages may be suspect.

Pretarsal clipping experiment; methods. Cicadas have a pair of sharp terminal claws (or pretarsal ungues; Snodgrass 1993) on each leg. Claw damage reduces cicadas’ ability to grasp vegetation or other cicadas. Damage to the foreleg claws appears to increase mortality, because cicadas apparently use their front legs almost as antennae to investigate the characteristics of a surface before venturing onto it; cicadas with damaged front claws tend to remain motionless, perpetually clawing at the surface in front of them.

36 Cicadas with claws other than the front pair removed appear to have reduced gripping abilities but are still mobile.

When a male cicada attempts to force a copulation, he climbs on the back of a female and must resist her attempts to dislodge him. A female so threatened flaps her wings vigorously, and is usually successful in preventing mating, unless the male is able to immobilize her wings or insert his genitalia before being dislodged (JRC pers. obs.). Given the chance, males will attempt to mount and copulate, but successful forced matings require substantial male effort. In contrast, in a mating that involves normal courtship and apparent mate acceptance by the female, the male meets little or no resistance when he attempts to mount.

If males are manipulated by removing some of their claws, they should be less successful at some of the behaviors required for forcing matings. If handicapped and normal males have equal access to females, but normal males mate more frequently, forcing behaviors are likely an important male tactic, perhaps encouraged by cage studies. Alternatively, no difference in mating frequencies would suggest that male forcing behaviors are not generally successful.

To evaluate whether bag cages allow males to trap females and force matings upon them, we placed twelve mature unmated female cicadas into three hardware cloth cages, four females per cage. Cages enclosed living vegetation, providing access to live branches for feeding and oviposition. Into each cage we also placed eight mated males obtained by separating mating pairs. Males can mate more than once; we used mated males to establish that all males were acceptable as mates prior to the manipulations. Four of these eight males had their two midleg pretarsal claws removed with a fingernail clipper, while the other four were handled similarly but left intact. We censused the cages regularly and recorded the identities of copulating males. Copulations are distinguishable from other

37 postures, because even though body orientations vary, the male genitalia are clearly engaged.

Claw removal had observable effects on the cicadas’ mobility and agility, but no evident immediate effects on mortality. If forcing behaviors are prevalent in bag cages, then unmanipulated and fully agile cicadas should obtain a greater proportion of matings than should males whose claws have been removed and who are thus at a disadvantage.

Results. The experiment ran for approximately 71 hours, or until all but two females copulated. The first matings occurred 40 hours after the experiment began, and the last matings at 68 hours. No rematings were allowed. Data were pooled across replicates, and pooled results were analyzed using a two- tailed Fisher’s Exact Test. Manipulated (clipped) and unmanipulated cicadas were statistically equally likely to obtain matings (Table 2.4). For comparison, a data set of this size showing a significant treatment effect would need to have two or fewer males mate in one treatment and eight or more in the other. With the caveat that this experiment involves small sample sizes, it does not support the view that male forcing behaviors are generally prevalent or successful, because males whose ability to force matings was impaired were at no disadvantage compared to normal males. Thus, we have no evidence that male forcing behaviors are a significant complication in cage experiments.

Cage size effects; methods. Unconfined, singly mated female Magicicada do not exhibit normal female behaviors indicating sexual receptivity (Chapter 5). If such females are generally able to avoid encountering males and remating, then a cage effect may have its most observable impact on female remating frequencies. A cage effect would also likely involve a relationship between the cage volume, or cage surface-to -volume ratio, and the severity of the effect, because in the smallest cages, females should be least able to escape unsolicited courtships and males should be most able to find females.

38 One way to test for the operation of a cage effect, then, is to compare remating behavior in small cages with behavior in large cages. In 1995, we stocked one large cage (the “flight cage”) and many smaller cages (“bag cages”) with cicadas. To determine whether remating frequencies were higher in the bag cages than expected from flight cage data we used a Monte-Carlo computer resampling algorithm to compare remating frequencies in the flight cage and bag cage data.

During the 1995 Brood I periodical cicada emergence at Alum Springs, Rockbridge County, Virginia, we stocked a 4m x 4m x 4m nylon mesh cage (our “flight cage”) with 79 adult female Magicicada septendecim, 86 adult male M. septendecim, 14 adult female M. cassini, and 10 adult male M. cassini. Magicicada nymphs tend to emerge from the ground and undergo their final molt in the evening (Maier 1982); the morning after ecdysis, newly emerged cicadas (“teneral” cicadas) are readily identifiable by their pale color, soft bodies, and by their position low in the vegetation. We captured teneral cicadas, individually marked them, placed them in the cage, and recorded the dates, times, and individuals’ identities for all matings. Based on day of tenerality (capture), we assigned each cicada to a “day-cohort” useful in estimating the onset of sexual maturity. We recorded all observed mating pairs in the cage. Out of 79 female M. septendecim, only one mated with a male M. cassini ; in addition, one female M. cassini mated with a M. septendecim male. Because cross-species interactions appear minimal, M. cassini will not be considered further in discussions of this experiment. Of the 78 female M. septendecim that mated, most did so by the eighth day of their adult lives, and 13 mated more than once before ovipositing.

Although the flight cage experiments ran for 20 days, because cicadas in the cage belonged to several day-cohorts, their behavior is best recorded with respect to day from tenerality, not day of first cage stocking. Therefore, only information on the behavior of 78 female cicadas for a period from tenerality to 12 days of adulthood was available. Not all of the cicadas belonging to later cohorts have information available about the later days

39 (days 8-12) of their adult lives, because the experiment was terminated less than 12 days after the latest cohorts were collected. It is not known whether cicadas in later cohorts would have remated, because they may only have had enough time to mate once. Thus, calculations of remating frequency in the flight cage are conservative, because any errors in the collection or interpretation of the flight cage data can lead only to underestimates of remating frequencies. This conservative figure for female remating provides the maximum potential contrast to cicada behavior in the smaller cages.

Of the five bag cage experiments conducted in 1995, only two allowed remating and are appropriate for this analysis. The design and results of each are summarized in Appendix C. Although the flight cage data covers a period of 12 days, and these experiments ran for 11 days, these time periods are comparable, since the number of days is less important than day quality (in terms of weather, etc.). In any case, all but one mating in the flight cage occurred before the mating cicada’s tenth day from tenerality, well before the end of the flight and bag cage experiments.

A cage effect that scales with cage size should manifest itself as higher remating frequencies in the small bag cages than in the large flight cage. We compared mating frequencies using a C++ resampling algorithm (Appendix D). The algorithm resamples the flight cage data on singly-and doubly- mated females to create sets of sizes comparable to the bag cages (n= 8 or n= 4 as appropriate), and it tallies the number of females in the resampled data that mated twice. The program iterates this resampling 10,000 times to generate a frequency distribution, and thus percentiles, of the outcomes of the resampling.

Results. The second replicate in the first experiment I may show a cage effect, since half of its 8 females remated, which is outside the 95th percentile of the resampling algorithm, and more that twice the remating frequency in any other cage (Table 2.5). Interpretation of these results requires caution, because two of the rematings in this cage are unusual, and unlike any others in these experiments. One of the four rematings in this cage

40 occurred after a mating involving conflict between two males. In this particular instance, two males struggled with each other, both attempting to insert their genitalia into the female. It was not possible to determine whether either male was successful in copulating. When the cage was next observed, one hour later, a period too short to be considered a normal mating (Chapter 4), the female was alone and not mating. Approximately 20 hours later, she engaged in a normal mating.

Another of the rematings in this replicate was the only post-oviposition remating in this study. This study was not designed to evaluate post-oviposition remating, which may occur for reasons distinct from pre-oviposition rematings (Chapter 4). If pre- and post- oviposition remating are treated as different phenomena, and if the post-oviposition remating is therefore eliminated from consideration, the results in this replicate are not significantly different from flight cage results. We scored both of these unusual situations as matings, but neither would have met the stricter criteria typical of our other experiments. If either one of the unusual matings in replicate 2 is discounted, then this replicate does not have a remating frequency higher than that in the flight cage.

Conclusions

We have uncovered no evidence that careful marking with enamel paints affects cicadas’ mortality or likelihood of mating, nor is there any evidence that cicada densities in our experimental cages are unrealistically high. The pretarsal clipping experiment suggests that males prevented from forcing are no less successful than normal males, and that male forcing strategies are not a significant factor in bag cage experiments. The cage size resampling algorithm demonstrates the absence of a cage effect whose severity scales with cage size. We have thus uncovered no evidence that our methods significantly alter cicada behavior in ways that would generate false or misleading results.

41 From observing cicada behavior in cages, however, we can add some notes of caution. In 1997, we placed female M. septendecim and male M. cassini in cages away from a male M. septendecim chorus. Under normal circumstances, we observed few sexual interactions among these cicadas. When we played male M. septendecim song, however, female M. septendecim responded with a receptivity signal (Chapter 5). This signal, which is similar in all Magicicada species, apparently incited the male M. cassini males in the cages to increase their sexual activities, and in some cases, a heterospecific male successfully engaged in a cross-species mating by mounting and copulating with a female who had responded to conspecific playbacks.

In another experiment, we constructed several dozen model cicadas from blocks of wood, painted them black, and attached them to the branches of a tree enclosed in a large flight cage. We stocked the cage with male M. septendecim and M. cassini. Normally, the males paid no attention to the models, but if we produced a simulated female wing-flick signal in time with a male call, males directed increased attention to the blocks, in some cases mounting and copulating. Male M. cassini in particular seemed to increase their sexual activity after we produced a simulated female wing-flick signal, and one male’s excitement seemed to have a synergistic effect on the others.

The experimental results in this paper demonstrate the effectiveness of female resistance, and that cages do not cause unreceptive females to mate. However, close confinement of several males and females in cages might compromise mate choice experiments, because if one female in a cage produces a wing-flick signal, indicating sexual receptivity, all the males in the cage may increase their sexual activity. In small, densely stocked cages, males can reach a mating frenzy, in which they interfere with and attempt to replace other courting males. Our observations indicate that these problems are more severe in M. cassini than in M. septendecim, so it appears advisable to avoid using the former species in mating experiments in which several males are confined in one cage.

42 Cage experiments in general require close observation and careful design to ensure that male-male competition does not unduly influence mating. Criteria for scoring matings must be designed so that unusual situations, such as post-oviposition mating, male-male competition, and interloping attempts, do not contribute to false positive results.

43 Literature Cited Brussard, P. F., 1971. Field techniques for investigations of population structure in a “ubiquitous” butterfly. Journal of the Lepidopterists’ Society . 25: 22- 29. Caprio, M. A., D. Miller, and E. Grafius. 1990. Marking adult Colorado potato beetles, Leptinotarsa decemlineata (Coleoptera: Chrysomelidae), using paper labels. Great Lakes Entomologist 23: 13-18. Davey, J. T. 1956. A method of marking isolated adult locusts in large numbers as an aid to the study of their seasonal migrations. Bull. Ent. Res . 46: 797-802. Dobson, R. M., J. W. Stephenson, J. R. and Lofty. (1958). A quantitiative study of a population of wheat bulb fly, Leptohylemia coarctata (Fall.), in the field. Bull. Ent. Res . 49: 95-111. Dobson, R. M., and M. G. Morris. 1961. Observations on emergence and life-span of wheat bulb fly, Leptohylemia coarctata (Fall.), under field cage conditions. Bull. Ent. Res. 51: 803- 821. Dreyer, H., and J. Baumgartner. 1997. Adult movement and dynamics of Clavigralla tomentosicollis (Heteroptera: Coreidae) populations in cowpea fields of Benin, West Africa. J. Econ. Entomol. 90: 421-426. Fales, J. H., O. F. Bodenstein, G. D. Mills, Jr., and L. H. Wessel. 1964. Preliminary studies on face fly dispersion. Ann. Ent. Soc. Amer . 57: 135-137. Gallepp, G. W. and A. D. Hasler. 1975. Behavior of larval caddisflies (Brachycentrus spp.) as influenced by marking. Am. Midl. Nat. 93: 247-254. Gangwere, S. K., W. Chavin, and F. C. Evans. 1964. Methods of marking insects, with especial reference to Orthoptera (Sens. Lat.). Ann. Ent. Soc. Amer . 57: 662-669. Garcia, M. A., and L. M. Paleari. 1990. Marking cassinidae (Coleoptera: Chrysomelidae) larvae in the field for population dynamics studies. Ent. News 101: 216-218. Greenslade, P. J. M. 1964. The distribution, dispersal, and size of a population of Nebria brevocillis (F.) with comparative studies on three other Carabidae. J. Anim. Ecol . 33: 311- 333. Hager, B. J., and F. E. Kurczewski. 1986. Nesting behavior of Ammophila harti (Fernald)(Hymenoptera: Sphecidae). American Midland Naturalist 116: 7- 24. Hart, D. D. and V. H. Resh. 1980. Movement patterns and foraging ecology of a stream caddisfly . Can. J. Zool . 58: 1174-1185. Hunter, P. E., 1960. Plastic paint as a marker for mites. Ann. Ent. Soc. Amer. 53: 698.

44 Kearns, C. A. and D. W. Inouye, 1993. Techniques for pollination biologists . (University Press of Colorado: Niwot, CO). Maier, C. T. 1982. Obervations on the seventeen-year periodical cicada, Magicicada septendecim (: Homptera: Cicadidae). Ann. Ent. Soc. Amer. 75: 14- 23. Mascanzoni, D., and H. Wallin. 1986. The harmonic radar: a new method of tracing insects in the field. Ecol. Ent . 11: 387-390. Mead-Briggs, A. R. 1964. Some experiments concerning the interchange of rabbit fleas, Spilopsyllus cuniculi (Dale), between living rabbit hosts. J. Anim. Ecol . 33: 13-26. Paulson, G. S., and R. D. Akre. 1991. Role of predaceous ants in pear Psylla (Homoptera: ) management: estimating colony size and foraging range of Formica neoclara (Hymenoptera: Formicidae) through a mark-recapture technique. J. Econ. Entomol. 84: 1436-1440. Rieske, L. K. and K. F. Raffa, 1990. Dispersal patterns and mark-and-recapture estimates of two pine root weevil species, Hylobius pales and Pachylobius picivorus (Coleoptera: Curculionidae), in Christmas tree plantations Environmental Entomology 19: 1829- 1836. SAS Institute Inc., 1996. SAS/STAT® Software: Changes and Enhancements through Release 6.11. Cary, NC: SAS Institute Inc. Sempala, S. D. K. 1981. The ecology of Aedes (Stegomyia) africanus (Theobald) in a tropical forest in Uganda: mark-release-recapture studies on a female adult population. Insect Sci. Application 1: 211-224. Snodgrass, R. E. 1993. Principles of Insect Morphology (Cornell University Press: Ithaca, NY). Southwood, T. R. E., 1978. Ecological Methods, with particular reference to the study of insect populations. (Chapman and Hall: London). Stuart, R. J. 1986. Use of polyester fibers to mark small leptothoracine ants (Hymenoptera: Formicidae). J. Kans. Ent. Soc . 59: 566-568. Walker, T. J., and S. A. Winewriter. 1981. Marking techniques for recognizing individual insects. Florida Entomologist 64: 18- 29. White, J. 1973. Viable hybrid young from crossmated periodical cicadas. Ecology 54: 573- 580.

45 changes None (Hemiptera) (Coleoptera) (Coleoptera) None (Coleoptera) None (Coleoptera) None (Diptera) (Diptera) None Toxic (Homoptera) (Orthoptera) None (Trichoptera) None (Coleoptera) None (Orthoptera) (mite) None (Hymenoptera) None (Diptera) Toxic (Coleoptera) None† (Lepidoptera) None (Hymenoptera) None§ (Coleoptera) spp. (Trichoptera) Behavioral Erebia epipsodea Erebia Hylobius pales Pachylobius picivorus Leptinotarsa decemlineata Dicosmoecus gilvipes Laelaspis georgiae Ammophila harti Nebria brevicollis Formica neoclara Lepthohylemia coarctata Lepthohylemia coarctata Brachycentrus Musca autumnalis Charidotis punctatostriata Clavigralla tomentosicollis Leptinotarsa decemlineata Magicicada septendecim Melanoplus sanguinipes Leptinotarsa decemlineata gregaria Schistocerca paper labels affixed with cynoacrylate gluepaper labels affixed None Monolite Yellow in wood rosin and kerosene)Yellow Monolite None 1958 1961 Oil-based paint paint (nitrocellulose) Lacquer-based 1990 by Acetone wash, followed 1990 with cynoacrylate glue Paper labels affixed 1990 based) Enamel paint (xylol, propanol Reported effects of methods for marking individual insects, grouped by general method and organized chrono- organized method and grouped by general individual insects, for marking of methods effects Reported . 1964 Fingernail polish (acetone based) et al. et al. et al. et al. et al. et al Rieske and Raffa 1990Rieske and Raffa Acrylic paint Caprio Hart and Resh 1980 with wire or elastic band Plastic tags affixed Brussard 1971Caprio “Sharpie” marker (xylol solvent) Fales Dreyer and polish (acetone based) Fingernail Greenslade 1964Akre 1991Paulson and based) Enamel paint (petroleum-solvent Dobson Enamel paint (petroleum-solvent based) Dobson Gallepp and Hasler 1975 correction fluid Typewriter Garcia and Paleari 1990 Fingernail polish (acetone based) Baumgartner 1997 Hunter 1960Hager and Kurczewski 1986 Enamel paint (xylol, propanol based) Enamel paint (xylol, propanol based) Table 2.1. Table each grouping. logically within ReferenceThis study Marker propanol based) Enamel paint (xylol, Taxon Effect* Caprio 1956Davey, or Toner (Rubine Oil-based paint

46 Toxic None‡ moderate treatment spp. (Coleoptera) None (Blattaria) None for (Siphonaptera) None (Orthoptera)(Orthoptera) (Orthoptera)(Orthoptera) None‡ None‡ in Toxic Harpagoxenus (Diptera) spp., Spilopsyllus cuniculi Harpalus niger, P. melanarius, Pterostichus rufipes, Carabus granulatus Aedes africanus Acheta domesticus Leptothorax (Hymenoptera)Acheta domesticus Acheta domesticus Acheta domesticus Periplaneta americana None Radar tags combined with Duco paint combined with Duco spray oxide Titanium dry pigmentOil and Cellulose paints, doses large 1964Aniline dyes, or Melted crayon, colored ink, 1964 Metallic dust 1964 or clipping amputation Wing 1964 with glue Colored thread or paper attached 1964 pigment in oil spray, Fluorescent powder, et al. et al. et al. et al. et al. * “None:” no effects reported; “Toxic:” increased mortality reported; “Injury:” unspecified damage or mortality noted. increased mortality reported; “Injury:” reported; “Toxic:” * “None:” no effects § Pers. Comm. noted. † Dispersal attributable to handling effects of limited use. ‡ Method proved to be impermanent and Gangwere Mead-Briggs 1964 1986 Wallin Mascanzoni and clipping Tarsal Gangwere Gangwere Table 2.1 (Continued) Table Sempala 1981Gangwere powder Fluorescent Stuart 1986 tied around body Polyester fiber Gangwere

47 Table 2.2. Numbers of marked and unmarked females showing a “seminal plug,” or evidence of mating, and c 2 comparison of seminal plug frequency in marked and unmarked cicadas. Equal numbers of marked and unmarked females were assigned to two cages, with an excess of males. Total females: Total number of females of each category present. Seminal plug present: total number of females at end of experiment with seminal plug. Marking does not affect likelihood of mating.

Total “seminal plug” females present c 2(2- tail) P

Cage A Marked 10 6 Unmarked 10 7 Total 20 13 0.038 P>>0.1 (n.s.)

Cage B Marked 10 5 Unmarked 10 6 Total 20 11 0.045 P >>0.1 (n.s.)

Combined cages Marked 20 11 Unmarked 20 13 Total 40 24 0.083 P >>0.1 (n.s.)

48 Table 2.3. Cage density resampling results, 10,000 iterations. Algorithm determines whether a specified cage density is greater than densities of cicadas on actual branches by resampling, with replacement, branch density data and generating percentiles. Cage densities above 25 are significantly greater than naturally-occurring densities on branches of similar size.

Hypothetical cage density P 12 0.85 14 0.82 16 0.51 25 0.053 26 0.028*

* indicates significance

49 Table 2.4. Results of pretarsal clipping experiment. Mating likelihoods of normal males and males with center tarsal claws removed were compared by confining 4 unmated females in each of 3 cages, and adding 4 normal and 4 clipped males to each cage. Counts are of normal and tarsal clipped males mating in each cage. 10 total matings occurred. In the combined replicates, the number of normal and tarsal clipped males mating do not differ (Fisher’s Exact 2- Tail P £ 0.656). Males mating Cage Normal Clipped 1 3 1 2 2 1 3 2 1 Combined 6 4

50 n size equal 0.029 0.130 0.396 0.396 0.515 0.515 0.515 P ≤ ≤ ≤ ≤ ≤ ≤ ≤ # Iterations with Observed # Resampling = rematings is the size of the population in small bag cages (8 for experiment 1; 4 for experiment 2). The algorithm 2). for experiment 1; 4 for experiment population in small bag cages (8 is the size of the n Summary of resampling results comparing remating frequencies in small bag cages to remating frequency in large flight to remating frequency in large frequencies in small bag cages results comparing remating Summary of resampling Experiment 1: (A)†(A)§BCExperiment 2: D 8E 8 8 are included as first matings (see text). A † If both questionable matings in replicate 8 is included as a first mating (see text). A § If only one questionable mating in replicate 4 3 4 2 4 2 10,000 10,000 1 1 10,000 10,000 10,000 290 10,000 1299 3960 3960 5150 5150 F 4 1 10,000 5150 Replicate # Females rematings iterations observed rematings Table 2.5. Table draws samples of The algorithm cage. flight mated females from large starts with data on singly- and doubly- Algorithm cage. from this population, where from this population, than or mated females in the sample is greater which the number of multiply of resampling iterations in then tallies the number bag cages. observed rematings in the small to the number of

51

12

10

8

6

Number Living 4 52

2

0 6/1/96 6/2/96 6/3/96 6/4/96 5/27/96 5/28/96 5/29/96 5/30/96 5/31/96 5/11/96 5/12/96 5/13/96 5/14/96 5/15/96 5/16/96 5/17/96 5/18/96 5/19/96 5/20/96 5/21/96 5/22/96 5/23/96 5/24/96 5/25/96 5/26/96 Date

Figure 2.1. Survivorship of marked and unmarked cicadas in two cages. Triangular and square markers indicate different cages, filled markers indicate marked treatment, and open markers indicate unmarked treatment. Cicadas in the two cages differed in survivorship, but within a cage, marking did not affect survivorship. CHAPTER III

MATE GUARDING AND THE NATURE AND POSSIBLE FUNCTIONS OF “SEMINAL PLUGS” IN PERIODICAL CICADAS (MAGICICADA SPP.)

Abstract

After mating, a female periodical cicada (Magicicada spp.) usually exhibits a mass, or “seminal plug,” blocking her genital opening. This mass contains DNA and appears to be composed of dried seminal fluid rather than being specially secreted or constructed. Although the “seminal plug” does not prevent additional matings, it may be part of a male strategy to make remating so costly, in terms of time, that mated females will be sexually unreceptive. Evidence that time is valuable to female Magicicada includes the limited nature of oviposition space, the monopolization of oviposition space by the first females to find it, and the apparent space-conserving nature of Magicicada eggnests. Thus, “seminal plugs” may be part of a male paternity-assurance strategy even though they do not prevent matings.

53 Introduction

As long as a given male’s sperm remain in a female’s reproductive tract, it is not in his interests to facilitate future mating by that female. However, because females may seek opportunities to remate in order to acquire resources or manipulate offspring paternity (Chapter 4), the sexes may be in conflict over remating. In species where males’ paternity is threatened by opportunities for polyandry, selection will favor males that evolve methods for blocking or preventing female remating, such as by inserting a “mating plug,” or barrier, into the female genital opening. Even a current tendency for females not to remate could be the direct result of past male efforts to devalue females’ remating incentives. For example, by inserting a mating plug, a male imposes costs on his mate and reduces her potential benefits from remating by forcing future mating efforts to include time spent removing or circumventing the plug. If females are time-limited, even if mating plugs are not absolute barriers to remating, females with such plugs may benefit from avoiding costly second matings by becoming sexually unreceptive after their first mating.

Mated female periodical cicadas (Magicicada spp.) usually exhibit a whitish or yellowish ovoid mass lodged in their genital openings (White 1973). Although the name applied to this structure, “seminal plug,” implies that it is an adaptation for mate-guarding, there are no detailed studies of its function. Magicicada seminal plugs could be the incidental results of ejaculate desiccation upon exposure to air. However, it is also possible that males use seminal plugs to protect their paternity, because under certain conditions female Magicicada remate (Chapter 4).

In other insect species, seminal plugs may function in biasing paternity by causing ejaculate to be retained within the spermatheca, by excluding additional sperm during the period that previously acquired sperm may be used (or stored for future use), or by permanently excluding additional sperm (Parker 1970, Boorman and Parker 1976, Leopold

54 1976, Ehrlich and Ehrlich 1978, Thornhill and Alcock 1983: 339, Alcock 1994, Polak et al. 1998). Extended copulation may serve similar functions (Alcock 1994, Parker 1970). Functional seminal plugs are often constructed of materials other than ejaculate. For example, honeybee (Winston 1987) and harvester ant (Hölldobler 1976) males leave their genitalia as plugs. In some Diptera, “sperm plugs” appear to be derived from spermatophore materials (see Neilsen 1959, Lum 1961, Kotrba 1996, Yasui 1997), while in many Lepidoptera, males apply sealing substances to the female genital opening after inserting a spermatophore (Labine 1964, Matsumoto 1987, Ehrlich and Ehrlich 1978).

Determining whether “seminal plugs” are male paternity-assurance adaptations or whether they are instead functionless consequences of matings is not a simple task. Two general predictions distinguish between functional and nonfunctional explanations for seminal plugs. (1) If a female insect can be caused to mate and acquire a seminal plug, and if she is prevented from acquiring a normal sperm supply but does not remate, the plug itself may be an effective deterrent to matings; alternatively, if such females show no decrease in likelihood of remating, remating propensity is likely controlled by factors such as spermathecal fullness. (2) If males use special materials or processes to construct or insert seminal plugs, the costs of doing so must be offset by some functional benefit, such as the ability of plugs to block matings; if, on the other hand, seminal plugs are incidental results of ejaculate drying, their functional significance and adaptive nature would be less certain.

I used several approaches to assess whether male Magicicada behave in ways that protect their paternity and whether seminal plugs are part of a paternity-assurance strategy. First, I evaluated whether mating durations are far in excess of what would be required for complete insemination; an excess would imply that males use mating duration as a form of mate-guarding. I also interrupted matings to determine whether seminal plugs are deposited at the conclusion of the normal mating sequence and whether they are produced

55 when males are removed suddenly and without warning. I determined whether seminal plugs deter or prevent subsequent matings, and I analyzed the substances contained in them to evaluate whether they are specially constructed or secreted. I confined my research to the 17- year species Magicicada septendecim, but since M. cassini and M. septendecula, as well as the 13- year species M. tredecim, M. neotredecim, M. tredecassini, and M. tredecula have similar mating systems and exhibit seminal plugs, the results are likely general to the genus Magicicada.

Methods and Results

Mating duration

In 1995 (study sites listed in Table 3.1), I captured and individually marked (Cooley et al. 1998, Chapter 2) teneral females and allowed them to mature in ca. 200 liter single-sex storage cages. Cages were constructed by folding a piece of hardware cloth over living hardwood vegetation. When the females were mature, I placed 5- 10 in newly constructed cages with fresh vegetation, and to each cage I added males captured from those active in nearby choruses. To establish starting and stopping times for each observed copulation, I monitored cages continuously or scanned them every 10-15 minutes, or ad lib scanned when I heard male courtship sounds. Mating durations are thus, in some cases, estimates ± 15 minutes. In 1996- 1998, I conducted similar experiments on 8 individuals belonging to two species of the genus Okanagana to provide comparative data (4 each O. canadensis and O. rimosa). I observed these cicadas continuously during courtship and mating and examined females for the presence of “seminal plugs” after mating.

Results. For the 61 Magicicada mating pairs I observed, mating duration averaged 838.8 (± 1074) minutes. Since poor weather may extend matings over days (unpub. data), if 17 matings lasting more than 10 hours are excluded, the remaining 43

56 matings had an average duration of 243.4 (± 121.48) minutes (Table 3.2). Matings in Okanagana were much shorter, with average durations under 30 minutes (Table 3.2).

Seminal plug deposition

In 1996, I collected 65 unmated, teneral female Magicicada septendecim and allowed them to mature as described above. When they were mature, I allowed them to mate with males caught from those present singing and flying in nearby choruses. I interrupted 38 of the resulting matings after 1 or 2 hours by suddenly, grasping the base of the male’s genitalia, gently working male and female genitalia apart, and noting the presence or absence of a “seminal plug.”

Results. Although interrupted matings could result in seminal plugs, indicating that the plug may be formed if the male transfers at least some ejaculate, interruption significantly decreases the likelihood that a seminal plug will be present (Table 3.3). All seminal plugs were clearly distinguishable from the absence of a plug, but plugs resulting from an interrupted copulation were more likely to have a soft, glossy appearance than plugs from uninterrupted matings. Because some females involved in terminated copulations exhibited sperm plugs, the insertion or creation of this plug requires either (1) no special actions or processes, or (2) that the male engage in processes that he can initiate and complete with little or advance preparation.

When I separated copulating pairs, in several cases I noted a small, whitish drop of ejaculate clinging to the males’ genitalia. In one case, the male continued ejaculating after separation, and the ejaculate quickly dried into a solid, whitish mass reminiscent of seminal plugs. This dried mass had the same color and consistency a typical seminal plug, except that it was of a more uniform consistency; the distal end of seminal plugs (the end visible in the genital opening) tended to be drier and more solid than the soft, moist proximal end.

57 These observations of drying ejaculate demonstrate that it can harden into a solid mass without additions or manipulations by the male.

Female remating

I allowed the 36 of the mating-interrupted females the opportunity to remate with males recently captured from the nearby chorus. Of the 36 females, 11 exhibited seminal plugs. When given the opportunity to remate, 20 of the 36 females in our mating interruption experiment remated; 4 of these 20 females exhibited seminal plugs. I compared remating frequencies of females with and those without seminal plugs using both a Chi-squared goodness of fit test and a computer resampling algorithm (Appendix E). The resampling algorithm randomly draws 20 “remating” samples from a population composed of 11 females with mating plugs and 25 without (total 36). The algorithm tallies the number of females in each sample with seminal plugs and repeats the process 100,000 times to generate percentiles. These percentiles can be used to determine whether the number of remating females with seminal plugs is smaller than expected by chance, as would be predicted if the seminal plug prevented copulation.

Females with seminal plugs were not underrepresented among remating females (Table 3.4). The seminal plug is therefore not a barrier to remating.

Seminal plug composition

In 1997, I captured approximately 25 mated female M. septendecim and, using forceps, removed seminal plugs from their genital openings. I preserved the plugs in 70% ethanol. To isolate high molecular weight DNA, I used Qiagen tissue preparation columns (Qiagen corp., Cat. 29304) to extract DNA from preserved plugs or from 70% ethanol preserved adults collected in 1996. I used a modified version of the tissue protocol included with the kit:

58 1. I removed and macerated a leg from each adult cicada, or I combined several seminal plugs into one microfuge tube. 2. I added 180 ul Qiagen buffer ATL and 20 ul Proteinase K stock solution without vortexing. I incubated the solution for a minimum of 24 hours at room temperature. 3. Without vortexing, I added 200 ul Qiagen buffer AL and incubated at 70°C for at least 10 min. 4. Without vortexing, I added 210 ul 100% ethanol.

5. I washed and eluted the DNA twice with warm distilled water as per instructions.

I amplified extracted DNA using a Randomly Amplified Polymorphic DNA oligonucleotide primer (RAPD; Haymer 1994, Hadrys et al. 1992) obtained from the University of British Columbia, with adult cicadas as positive controls and distilled water as our negative control in all amplifications. After screening approximately 200 other 10- base primers, I found that the primer sequence (CCT GCG CTT A) reliably produces a characteristic double or single band in M. septendecim with a minimum of amplification artifacts. I visualized amplified DNA on 1% agarose gels stained with ethidium bromide, photographed the gels, and noted the presence or absence of bands.

Results. The DNA extraction and amplification demonstrates that seminal plugs contain DNA (Fig. 3.1), a finding consistent with the hypothesis that they are formed of dried ejaculatory material instead of special-purpose substances. Qualities of the ejaculate alone can account for the properties of seminal plugs. As I extracted the seminal plugs, I noted that the proximal ends of most plugs were bifurcated, suggesting that they extended into paired spermathecae. The texture of the plugs was consistent with the hypothesis that they were formed by progressive drying of ejaculatory fluid, starting from the more exposed, distal end.

59 Discussion

The properties of seminal plugs can be accounted for by the properties of ejaculate alone. Males do not appear to employ special substances or structures other than ejaculate to construct seminal plugs; thus, they are not unambiguously male paternity assurance adaptations, although the hardening properties of ejaculate may be the product of selection for ejaculate to form barriers to remating. The remarkably long mating durations and “seminal plugs” of Magicicada may be parts of a male paternity-protection strategy. Mating interruption experiments (Chapter 4) have demonstrated that females allowed to mate for more than 2 hours tend to be sexually unreceptive and unlikely to remate; the longer average mating durations (ca. 5 hours) typical of Magicicada may be a form of male mate guarding (see Thornhill and Alcock 1983:340, Alcock 1994). Furthermore, the hooks and complex structure of male Magicicada genitalia (Chapter 8) may allow males to extend mating duration contrary to female interests, allowing the ejaculate to harden and making future matings costly or difficult. In contrast to Magicicada, Okanagana have short mating durations, simple genitalia, lack a seminal plug, and, presumably, have a greater confluence of interest between the sexes over mating duration.

Paradoxically, although Magicicada have long lifespans, time may be the adults’ greatest enemy. For example, since females appear to avoid constructing new eggnests directly over pre-existing ones, oviposition space in dense emergences may be limited and available only to the first finder. The ideal eggnest density probably represents a complex balance between overly dense eggnests, leading to “flagging,” or death of twigs (and the eggs therein), and overly sparse eggnests, which might be vulnerable to predators or parasitoids (Marlatt 1923). Of these competing sources of egg mortality, flagging is certain to occur on densely oviposited branches, while the risks from ovipositing in sparsely used branches are uncertain; thus females may be better off to err on the side of ovipositing in an area of sparse eggnests than to risk having their eggnests “flag.” The prevalence of

60 flagging may thus be weak evidence suggesting the scarcity of suitable oviposition space and the lack of oviposition options for many females.

The structure of Magicicada eggnests also supports the claim that oviposition space is limited. Magicicada eggnests are bifurcated, unlike those of other cicadas ovipositing in live wood, such as Okanagana. Bifurcated eggnests take up approximately as much linear space on a branch as do single eggnests, so Magicicada eggnests contain approximately twice as many eggs as Okanagana nests. Magicicada’s bifurcated eggnests may be an adaptation to ensure complete oviposition and monopolization of unused twigs when limited oviposition space is available on a “first come, first serve” basis.

Female- female competition for finding oviposition space may place a premium on efficient use of time and ability to locate and take maximum advantage of unused twigs. Males may use the currency of time to coerce females to acquiesce to male interests. Even without being an absolute barrier to copulation, seminal plugs may make future matings costly in terms of time, because future mates would have to remove or circumvent the seminal plug. By engaging in lengthy copulations, obstructing the female reproductive tract, and thereby ensuring that any future copulations would also be lengthy, males may reduce females’ potential benefits from remating.

One hypothesis for how Magicicada male and female interests in mating duration have evolved is as follows: Because males that monopolized the paternity of their mates’ offspring sired more offspring and were more successful than males whose sperm was diluted or discarded because of remating, some males increased their control of paternity by increasing mating duration. Over evolutionary time, as males first began to extend matings, the sexes may have been in conflict over the time costs of mating. As males escalated these costs beyond a certain point, and as males began to employ more and more effective forcible restraining devices, such as complex genitalia and ejaculate that forms seminal plugs (or any other strategy that functions to increase intermating interval), male

61 and female interests would have become more congruent, in that females could not afford to permit additional matings. Those females acceding to male interests in monopolizing paternity would have had an advantage over those engaging in multiple lengthy matings, so females would have evolved to be unreceptive to multiple matings. Thus, by the process described above, a seminal plug that did not prevent mating may have been part of a male strategy to negate any advantages females could realize from remating. This hypothesis would be disproven by findings that (1) seminal plugs provide no impediment to future matings and males have no adaptations for evading or removing them; or (2) seminal plugs are found in similar cicada species (such as Okanagana spp.) in which there is no evidence of male-female conflict over mating duration.

62 Literature Cited Alcock, J. 1994. Postinsemination associations between males and females in insects: The mate-guarding hypothesis. Ann. Rev. Entomol . 39: 1-21. Boorman, E., and G. A. Parker. 1976. Sperm (ejaculate) competition in Drosophila melanogaster, and the reproductive value of females to males in relation to female age and mating status. Ecological Entomology 1: 145- 155. Cooley, J. R., Hammond, G. S., and D. C. Marshall. 1998. The effects of enamel paint marks on the behavior and survival of the periodical cicada, Magicicada septendecim (L.) (Homoptera) and the lesser migratory grasshopper, Melanoplus sanguinipes (F.) (Orthoptera). Great Lakes Entomologist 31: 161-168. Ehrlich, A. R., and P. R. Ehrlich. 1978. Reproductive strategies in the butterflies: I. Mating frequency, plugging, and egg number. J. Kans. Ent. Soc. 51: 666- 697. Hadrys, H., M. Balick, and B. Schierwater. 1992. Applications of random amplified polymorphic DNA (RAPD) in molecular ecology. Molecular Ecology . 1: 55-63. Haymer, D. S. 1994. Random amplified polymorphic DNAs and microsatellites: What are they, and can they tell us anything we don’t already know. Ann. Ent. Soc. Amer. 87: 717- 722. Hölldobler, B. 1976. The behavioral ecology of mating in harvester ants (Hymenoptera: Formicidae: Pogonomyrmex ). Behav. Ecol. Sociobiol . 1: 405- 423. Kotrba, M. 1996. Sperm transfer by spermatophore in Diptera: new results from the Diopsidae. Zool. J. Linn. Soc . 117: 305- 323. Labine, P. A. 1964. Population biology of the butterfly, Euphydryas editha I. Barriers to multiple inseminations. Evolution 18: 335- 336. Leopold, R. A. 1976. The role of male accessory glands in insect reproduction. Ann. Rev. Emtomol. 21: 199-221. Lum, P. T. M. 1961. The reproductive system of some Florida mosquitoes. II. The male accessory glands and their role. Ann. Ent. Soc. Amer. 54: 430- 433. Marlatt, C. L. 1923. The periodical cicada. U.S.D.A. Bureau of Entomology Bulletin 71, 183 pp. Matsumoto, K. 1987. Mating patterns of a sphragis-bearing butterfly, japonica Leech (Lepidoptera: Papilionidae), with descriptions of mating behavior. Res. Pop. Ecol. 29: 97- 110. Nielsen. E. T. 1959. Copulation of Glyptotendipes (Phytotendipes) paripes Edwards. Nature (London) 184: 1252- 1253. Parker, G. A. 1970. Sperm competition and its evolutionary consequences in the insects. Biological Reviews 45: 525- 567.

63 Polak, M., W. T. Starmer, and J. S. F. Barker. 1998. A mating plug and male mate choice in Drosophila hibisci Brock. Anim. Behav. 56: 919- 926. Thornhill, R. and J. Alcock. 1983. The evolution of insect mating systems . (Harvard University Press: Cambridge, MA) White J. 1973. Viable hybrid young from crossmated periodical cicadas. Ecology 54: 573- 580. Winston, M. L. 1987. The biology of the honey bee . (Harvard University Press: Cambridge, MA) Yasui, Y. 1997. Sperm competition and the significance of female multiple mating in the predatory mite Parasitus fimetorum. Experimental and Applied Acarology 21: 651- 664.

64 199519961997 I II III1996- 1998 Siloam Springs Lake SWF Horsepen Siloam Springs SP UMBS BuckinghamAdams Brown, Rockbridge VA IL VA Logged site Emmet Cheboygan/ Clearing MI Logged site Open field Study sites 1995- 1997. Table 3.1. Table SpeciesMagicicada septendecim YearOkanagana canadensis/ O. rimosa Brood Location County State Characteristics

65 Table 3.2. Mating durations in Magicicada septendecim and Okanagana. O. canadensis mating lasted approximately 4 times as long as the average of the others, so two averages, one with this long mating, and one without, are calculated for this species.

Species Duration (min: sec) O. rimosa 19:00 17:00 22:09 17:04 Average 18:48 ± 2:25

O canadensis 80:00 19:06 15:09 23:18 Average 34:23 ± 30:35 Average w/o outlier 19:11 ± 4:05

Magicicada septendecim 195 315 195 330 295 442 167 297 235 235 235 195 310 500 60 124 305 111 354 295 124 331 440 440 310 64 220 444 263 295 354 360 360 250 263 310 65 526 148 91 310 60 405 Average (n= 43) 243:24 ± 121:48

66 Table 3.3. Results, female mating interruption experiment. For each treatment (disrupt 1 hr., disrupt 2 hr., combined 1 and 2 hr disruption, and no disrupt), numbers of females without seminal plugs (No SP), with firm seminal plugs (Firm SP), and with soft seminal plugs (Soft SP) were counted. There were no statistical differences between 1 and 2 hour mating disruptions, so these treatments were combined. Females whose matings were interrupted were less likely to exhibit seminal plugs than were females with uninterrupted matings (Fisher’s Exact 2- tail, P < 0.001). Soft seminal plugs were more prevalent in interrupted matings than in uninterrupted matings (Fisher’s Exact 2- tail P < 0.001).

Mating Total interruption females No SP Firm SP Soft SP Firm + Soft 1 hr. 13 8 1 4 5 2 hr. 25 17 6 2 8 £ 2 hr. (combined) 38 25 7 6 13

Not interrupted 27 2 25 0 25

67 Table 3.4. Numbers of females with and without seminal plugs (SP) remating when given free access to males. The presence of a seminal plug did not affect female likelihood of remating: If rematings were expected to be evenly distributed between females with and without plugs, 6 females with plugs would be expected to remate. 4 did : X 2 P > 0.10. The data were also analyzed with a resampling algorithm, which drew a sample of 11 females from the “without SP” data: In 100,000 iterations of the algorithm, the proportion of iterations in which 4 or fewer plugged females were included in the sample was 0.2214. Remating Not remating Females with SP 4 7 Females without SP 16 9

68 Seminal Plug Seminal Plug Adult Adult Adult

69 H 2 O

Figure 3.1. Photograph of gel showing RAPD-PCR amplification with primer (CCT GCG CTT A). Adults and seminal plugs show characteristic amplification products, demonstrating that seminal plugs contain DNA. If plugs did not contain DNA, they would show no amplification product. Gel is 1% agarose stained with EtBr, negative control is distilled water, positive controls are adult cicadas. CHAPTER IV

FEMALE MULTIPLE MATING AND MATE CHOICE IN A PERIODICAL CICADA, MAGICICADA SEPTENDECIM

Abstract

Explanatory hypotheses for polyandry in the periodical cicada Magicicada septendecim fall into two categories: 1. Females remate because of inadequate insemination; and 2. Females remate to effect mate choice. I gave mated females the opportunity to remate after manipulating both mating duration and mate quality. Females whose first matings were terminated prematurely tended to remate, while females allowed complete, undisturbed matings with inappropriate, heterospecific males did not. After participating in a complete, undisturbed mating, females were no longer sexually receptive nor were they attractive to males. Thus, when placed in situations where they were either inadequately inseminated or mated with low-quality mates, only females in the former manipulations mated multiply, suggesting that polyandry in M. septendecim is best understood as the result of sperm limitation rather than as an evolved mate choice strategy.

70 Introduction

Multiple mating is sometimes considered to be evidence that females exercise postcopulatory mate choice (Eberhard 1985, 1996, 1997). Selection favors the evolution of mate choice mechanisms when individuals discriminating among potential mates produce genetically superior offspring, acquire phenotypic benefits specific to a given mate and convertible into reproductive output, or both. Postcopulatory mate choice is a possibility for females that mate multiply and for which insemination and fertilization are separated in time (e.g.; Eberhard 1985, Birkhead and Møller 1993, LaMunyon and Eisner 1993, Eberhard 1996, 1997), although some models of postcopulatory choice have been subjected to strong criticism (Alexander et al. 1997). Alternatively, females may remate to acquire sufficient sperm (or other resources) to complete their reproduction, in which case, mate choice need not be invoked. Magicicada septendecim 1 females sometimes remate (Chapter 2, Alexander 1968), although anecdotal evidence suggests that remating is not common: During the 1998 emergence of periodical cicadas, mating pairs became increasingly rare as the emergence progressed, the opposite of the pattern expected if females typically remate. To evaluate whether female M. septendecim remating is an evolved mate choice mechanism, I first discuss the general benefits females might obtain by remating and whether such benefits involve mate choice. Although, in order to identify selective contexts of remating, the benefits of polyandry are presented as if they belong to strict categories, this is undoubtedly an oversimplification, because the different kinds of benefits are likely not exclusive. I conclude by presenting the results of manipulations designed to evaluate the causes of polyandry in M. septendecim.

Remating as bet-hedging. Multiply mating females may gain from potential increases in offspring diversity or from bet hedging against genetically incompatible sires

1 This paper specifically concerns Magicicada septendecim. However, its findings may be applicable to other species in this genus since the sexual behaviors of all Magicicada species share general similarities.

71 (Walker 1980, Smith 1984, Watson 1991, Zeh and Zeh 1996, 1997, Zeh 1997, Brown 1997a, Tregenza and Wedell 1998). If females remate for diversity or bet-hedging, most or all females in a population should mate multiply, mate quality should not necessarily increase with each successive mating (see Watson 1991), nor should mating with a lower- quality male necessarily trigger remating. If females remate only to increase offspring diversity, sperm displacement must be incomplete, or, if complete, females must oviposit between matings to allow multiple paternity.

Remating for replenishment. Limited female sperm storage capacity, variable ejaculate volumes, limits to males’ ability to fully inseminate in a single mating, and disturbances during mating may promote serial polyandry because remating allows females to replenish or supplement inadequate sperm supplies (Boorman and Parker 1976, Walker 1980, Gromko et al. 1984, Ward et al. 1992). Females remating to replenish depleted supplies are expected to oviposit or engage in other sperm- or resource- depleting activities between successive matings. Some female dipterans, including Drosophila (Pyle and Gromko 1978, Gromko and Markow 1993), Ceratitis (Nakagawa et al. 1971, Whittier and Shelley 1993) and Rhagoletis (Neilson and McAllan 1965), homopterans (Nilaparvata lugens; Oh 1979), and heteropterans (Riptortus, Sakurai 1996, 1998a, b) experience reversible decreases in rates of ovipositing normal, fertilized eggs (considered an index of fecundity and fertility), several days after mating. In Riptortus and Nilaparvata, coincident with decreasing oviposition rate is a return to sexual behaviors typical of unmated females. In each of the above examples, fertility and fecundity returned to high levels subsequent to remating.

Remating and phenotypic benefits. Just as females may remate to replenish depleted sperm, if remating increases female resource access, females may remate to obtain direct phenotypic benefits. Mating males may provide nuptial gifts, spermatophores, (Morris 1979, Sakaluk and Cade 1982, Thornhill and Alcock 1983, Thornhill 1984,

72 Gwynne 1986, Goulson et al. 1993, Kaitala and Wiklund 1995, Brown 1997a, b, 1999) increased predator protection, or access to limited resources (as in resource-based leks; Alexander 1975, Hoglund and Alatalo 1995). Promiscuous females could collect phenotypic benefits in the same manner as any acquirable resource; here, the price of acquisition is permitting male sexual access. Female “foraging” for male-provided resources may be at least partially responsible for polyandry in some Lepidoptera (Rutowski 1984, Svärd and McNeill 1994, Karlsson 1996,) and perhaps in some Orthoptera (Sakaluk and Cade 1982, Brown 1997b) and Hymenoptera (Alcock et al. 1977). Phenotypic benefits from polyandry also include reduced costs, or at least minimization of losses, associated with acquiescence to persistent male harassment (Trillmich and Trillmich 1984, Clutton-Brock and Parker 1995, Jormelainen and Merilaita 1995, Arnqvist 1997, Crean and Gilburn 1998), although an effective female response to this male tactic is to leave the harassing male’s vicinity (see Alexander and Moore 1962, Stone 1995). Females remating only to replenish depleted supplies, and not to effect mate choice, would be expected to remate when they run low on supplies and not necessarily whenever a high quality male becomes available.

If male variations affecting ability to produce, acquire, or defend high-quality resources or nuptial gifts are heritable, then direct phenotypic benefits from choosing males with superior resources could be accompanied by genotypic benefits through jointly- produced offspring, and female remating could be a form of postcopulatory mate choice.

Remating and genetic benefits. If females remate to improve the genetic constitution of their offspring, then they are exercising mate choice. Genetic benefits from multiple mating include the advantages of cuckolding a genetically inferior parental male (Westneat et al. 1990, Birkhead and Møller 1993, Kempenaeres et al. 1992), and, under restricted conditions, benefits from sperm competition within the female, such that male offspring sired by especially competitive sperm might themselves produce competitive

73 sperm (Harvey and May 1989, Curtsinger 1991). If females remate to improve the genetic quality of their offspring, no complex mechanisms need be invoked; for instance, females with “last male” sperm priority using “threshold criteria” (Janetos 1980; Alexander et al. 1997) to choose mates could, by raising their choice threshold after each mating, “revise” their previous choices by mating with successively higher quality males (see Gabor and Halliday 1997). If males are readily available and not costly to acquire, then choosy females first mated with inferior males should be most likely to remate, and females should grant all their paternity to their best mate, wasting none on inferior ones. If remating tendency is independent of mate quality, replenishment seems a more likely explanation.

Bias against heterospecifics. Bias against heterospecifics is often an example of indirect mate choice (Chapter 1) rather than an evolved mate choice mechanism. Females benefit from such biases by obtaining appropriate genetic materials for their offspring. One form of such bias is conspecific sperm precedence, in which, regardless of mating order, few hybrid offspring are produced when a female is mated with a series of con- and hetero- specifics. Conspecific sperm precedence is reported in flour beetles (Wade et al. 1994), grasshoppers (Bella et al. 1992), sunflowers (Rieseberg et al. 1995), ground crickets (Howard and Gregory 1993, Gregory and Howard 1994; Howard, Gregory, Chu, and Cain 1998; Howard, Reece, Gregory, Chu, and Cain 1998), and several Drosophila species (Price 1997). If precopulatory behaviors allow a significant proportion of heterospecific matings, females could recover from such mistakes by remating (see Jamart et al. 1995). However, this pattern is not necessarily evidence of an evolved post-copulatory mate choice mechanism. Instead, it may result from the inferior competitive ability of heterospecific sperm and seminal products in the environment of the female reproductive tract, especially if the interactions between female and ejaculate are shaped by a long history of coevolutionary antagonism over control of fertilization (Rice 1996, Price 1997, Alexander et al. 1997, Rice 1998).

74 Cross-mated females capable of postcopulatory mate discrimination would be expected to remate before ovipositing unless sperm priority was first in/first out, there was no sperm mixing, and there was no way to remove stored spermathecal sperm except through fertilization. Otherwise, the sperm dilution or replacement caused by subsequent matings prior to oviposition would be to the female’s advantage, since it would minimize the production of hybrid offspring, each of which represents a loss of potential fitness. If females mated to heterospecific males do not actively seek subsequent matings with conspecifics, then it is unlikely that they have evolved to exercise postcopulatory mate choice.

Polyandry in M. septendecim

Precopulatory behaviors may provide female periodical cicadas an efficient first defense against heterospecific males; precopulatory species discrimination based on song characteristics has been demonstrated (Chapter 6), and, in addition, 78 of 80 female M. septendecim in one of our cage experiments avoided heterospecific matings, a figure comparable to data reported in White (1973). However, if, in spite of precopulatory discrimination, heterospecific matings are a significant threat to females, then it is possible that remating behaviors evolved to allow mate choice revision. Females could also remate to discriminate among conspecific sires. To evaluate whether female remating is likely to function in mate choice, I conducted mating experiments to determine the contexts in which females remate, playback experiments to evaluate female sexual receptivity after mating, and experiments to evaluate female sexual attractiveness after mating.

Female remating experiments

For M. septendecim, at the present time, we are unable to identify a class of inferior conspecifics or determine whether, within a species, males vary in ways likely to promote female mate choice. One way to cause females to remate is to interrupt their first matings.

75 Interruption causes remating either because it simulates a mating attempt by an inferior male or because it results in incomplete insemination and an inadequate sperm supply. The mate choice through remating hypothesis predicts that females should remate immediately in response to mating with an “inferior” male and grant paternity only to later mates and that the length of time allowed before interruption should not influence likelihood of remating. In contrast, the replenishment hypothesis predicts that interrupted females will oviposit, and as they deplete their sperm stores, they will remate; thus females allowed the least amount of time for their first matings should be the most likely to remate immediately. Furthermore, if females have an evolved mechanism for improving their circumstances after mating with an inappropriate mate, then females mated with heterospecifics should have high pre-oviposition remating frequencies.

Methods. I studied Brood II periodical cicadas in a large open field near Horsepen Lake State Wildlife Management Area, Buckingham County, VA, during an emergence lasting from late April through early June 1996. Vegetation in the field consisted primarily of oak, maple, and tulip tree stump sprouts from a logging operation several years before. The field was surrounded by a mixed forest composed of mature hardwoods, primarily oaks and maples, and planted pines.

The morning after they molt, newly emerged cicadas (“teneral” cicadas) are readily identifiable by their pale color, soft bodies, and by their position low in the vegetation. I collected unmated, teneral female cicadas daily and placed them into fiberglass mesh storage cages one meter in length and approximately 60 cm in diameter placed over living vegetation. After females had matured in the storage cages without access to males, I marked them with either dots of enamel paint or with symbols drawn onto their wings with laboratory markers. Neither marking method has adverse effects on cicadas (Chapter 2; Cooley et al. 1998). I caught males from those present singing and flying in the area of the experiments and marked them in the same manner as the females.

76 I placed 5- 10 mature females and fewer males (to discourage usurpation or harassment) in nylon tulle cages similar to the storage cages. I stocked these mating cages between 8 A. M. and 12 noon, and most matings occurred shortly thereafter. To establish a starting time for each observed copulation, I monitored cages continuously or scanned them every 10-15 minutes, stopping the scan only when I had seen every cicada in the cage. Alternatively, I ad lib scanned when I heard male courtship songs. Starting times are thus, in some cases, estimates ± 15 minutes. Matings whose beginning times could not be established within this degree of accuracy were used for the “no interruption” treatment (see below). Any cicadas failing to mate the same day they were placed in mating cages were discarded.

Mated female periodical cicadas usually exhibit a mating sign or “seminal plug” in their genital opening (White 1973), although this structure is not always present and does not prevent subsequent matings (Chapter 3). A seminal plug appears as a whitish or yellowish ovoid mass when the female’s ovipositor sheath is gently pulled open. The plug appears to be composed of dried seminal fluid (Chapter 3), and its presence is thus an indicator that the male has transferred at least some ejaculate. In all experiments, I noted the presence or absence of a seminal plug after each mating.

I interrupted conspecific matings after 15 min., 1 hour, 2 hours, or not at all by, at the appropriate time, quickly removing mating pairs from the mating cages, grasping the base of the male genitalia, and gently working male and female genitalia apart. Once separated, males were immediately preserved in 70% ethanol. In addition, I mated female M. septendecim with male M. cassini by confining them together in small cages. I left mating pairs not to be disrupted undisturbed until they stopped mating, at which time I preserved the males. I checked females for the presence of a seminal plug immediately after disruption or cessation of mating.

77 After they had mated once, or after their first mating had been disrupted, females were placed into cages where they had access to conspecific males and appropriate oviposition space. I surveyed these cages at least once per hour and immediately placed females observed ovipositing into separate oviposition cages, one female per cage. In addition, I subjected some females to a treatment in which their mating was interrupted after two hours and they were placed directly into oviposition cages. Once placed in oviposition cages, females were permitted no further mating opportunities.

Oviposition cages were either surplus mating cages or fiberglass screen cages approximately one meter in length and 30 cm in diameter. I placed all oviposition cages over oak, maple, or beech vegetation that had been thoroughly inspected for the absence of eggnests and that had a minimum of one meter of woody branch suitable for oviposition. I distributed females among oviposition cages haphazardly with respect to tree species, cage type, and cage location and removed each female from its oviposition cage when it died, became moribund, or at the end of the experiment. I left all oviposition cages intact and closed until the branch broke due to excessive oviposition (“flagging”) or until the end of the experiment, at which time I removed, counted, labeled, and preserved all eggnests in 70% ethanol.

Results. Normal, uninterrupted matings lasted an average of 243.4 minutes (Table 4.1). Females whose first matings were interrupted after 15 minutes or one hour were most likely to remate before ovipositing, those interrupted after two hours were equally likely to remate or not, and those not interrupted usually did not remate before ovipositing (Table 4.2). Females whose matings were disrupted within 2 hours of starting also tended to lack a seminal plug. Female M. septendecim first mated with M. cassini males did not remate before ovipositing. One female whose mating was uninterrupted and apparently normal did remate before ovipositing.

78 There is a statistical difference in number of pre-remating eggnests constructed by females interrupted at two hours and females who were not interrupted (Table 4.3, Fig. 4.1). However, there was no difference between the total number of eggnests produced by heterospecific-mated, mating-interrupted females allowed to remate, and uninterrupted females.

Female sexual receptivity

Upon hearing a male’s calling phrase, a receptive female flicks her wings with a quick motion (Chapter 5); this signal causes the male to continue courtship and attempt to copulate. If females solicit second matings as a form of mate choice, then we would expect them to employ many of the same courtship behaviors, such as signaling, as in first matings, because non-signaling females would have difficulty obtaining mates.

Mating status and receptivity. I identified a high quality recording of Magicicada septendecim song in the University of Michigan Museum of Zoology Sound Library. I imported this recording into a Macintosh IIcx computer using Mac Recorder hardware and Sound Edit software. I removed extraneous background noise, isolated a single male song, and confirmed that females signal in response to playbacks of this song (Chapter 5).

In 1997, at Siloam Springs State Park, Brown and Adams Counties, Illinois, (Brood III), I collected 44 unmated teneral female M. septendecim and allowed them to mature in single-sex storage cages. Half (22) were mated to males captured from the chorus, half were left unmated. Mated and unmated females into two cages, and in each of 4 subsequent days, I played recorded M. septendecim calling songs to the surviving females for 2 minutes and recorded the number mated and unmated females responding.

Results. None of the mated females responded to playbacks, while at least half of the unmated females responded each day with wing flicks (Table 4.4).

79 Mating completeness and receptivity. In June 1997, I conducted experiments combining the approaches of the female remating experiments and the playback experiments to determine how mating interruption would affect the likelihood that a female would solicit a second mating. As before, I allowed the cicadas to mature in fiberglass cages and then marked them individually prior to use in experiments. After the cicadas were mature, I allowed them to mate, disrupting the matings after 15 min, 1 hour, 2 hours, or not at all. I discarded the males and placed the females in cages without access to males. I visited female cages at least twice daily, played artificial male calling songs, and recorded female responses.

Results. Females whose matings were disrupted were likely to respond positively to male song, while normally mated females never responded. All disruption treatments resulted in an equal likelihood of female response (Table 4.5). These results are compatible with the 1996 experiments, in which all disruption treatments led to an elevated incidence of remating.

Female sexual attractiveness

Our playback experiments demonstrate that normally mated females are not sexually receptive to courting males. Males’ willingness to court mated and unmated females is potentially a separate issue (for a possible example see Gromko and Markow 1993). If there is even a slight chance that a mated female will remate, such as if she were to remate to effect mate choice, then males should find mated females sexually attractive and court them. If, on the other hand, female remating is an unusual event, one that happens only under special circumstances, males should seek unmated females instead of wasting time and effort courting mated and unreceptive females, and they should not behave as though mated females are sexually attractive.

80 Methods. Unmated female M. septendecim confined in cages signal to courting males outside. Cages containing signaling females thus are attractive to males, and males seeking mates congregate on the outsides of such cages. I evaluated the relative sexual attractiveness of mated and unmated females by confining them in cages and censusing the males observed on the outsides of the cages.

I erected three large, cylindrical, white tulle cages 1.5 m in diameter and 2 m tall over stump sprouts, all at the same distance from the forest edge in our Horsepen Lake, VA study site. I collected unmated and mated females in the same manner as for the playback experiments and placed 13 mated females in one cage, 13 unmated females of the same age in another, and nothing in the third cage. The cages were scanned at intervals no greater than one hour over the course of a two-day period and recorded the number of males on the exterior surface of each cage. At each census, all males were colllected from the exterior of the cages and removed to a location at least 15 m distant.

Results. Because the males at each census time were unlikely to be the same, each census is treated as an independent data point. Significantly greater numbers of males were observed on the cage containing unmated females than on the empty cage, demonstrating that it was the females in the cage, not the cage itself, that attracted courting males. Significantly fewer males visited the cage containing mated females than the cage containing unmated females, indicating that males are not strongly sexually attracted to mated females. (Table 4.6).

Discussion

Mating aggregations suggest the action of sexual selection (Emlen and Oring 1977, Alexander and Borgia 1979, Andersson 1994, Höglund and Alatalo 1995). Alexander (1975) considered periodical cicada male singing aggregations to be non-resource based leks. In such leks, males neither control nor have exclusive access to resources (other than

81 gametes) required by females, and thus females visit the lek only for the purpose of mating. Although female periodical cicadas occasionally remate, and although the situations faced by females in these experiments would seem to promote postcopulatory mate choice, the data provide scant evidence of a postcopulatory mate choice mechanism2. If females were able to accomplish mate choice by remating, they might be expected to detect the inappropriateness of a heterospecific mate or the inferiority of an abbreviated ejaculate and remate as quickly as possible. However, in our experiments, only females whose first matings were extremely brief were attractive to males, actively solicited matings, and reliably remated before ovipositing.

The readiness of mating interrupted or cross-mated females to oviposit makes evolved postcopulatory mate choice doubtful. The failure of cross-mated females to remate before ovipositing suggests that females are unable to evaluate male quality after the mating act has commenced or that they realize no benefits from doing so. Thus, any species- identification mechanisms of periodical cicada mating behavior must be precopulatory. Magicicada’s complex, species-specific courtship makes it unlikely that females are routinely in close proximity to heterospecifics or that they normally mate with them. The failure of female postcopulatory behavior to discriminate against mates as unsuitable as heterospecifics and the demonstrated post-mating change in both female receptivity and sexual attractiveness indirectly support the importance of precopulatory discrimination in Magicicada.

For periodical cicada females, mating is not known to provide any phenotypic benefits, either through ejaculatory nutrients or nuptial gifts, nor do males restrict female access to resources (other than sperm). Phenotypic benefits through reduced costs from

2 One female whose mating was uninterrupted and apparently normal did remate before ovipositing; this observation has no real significance, but it is consistent with female remating behavior being an adaptation to discriminate against poor quality mates, because at least some females who were randomly paired with males would be expected to remate because at least some would have been presented with a better alternative to their first mating.

82 acquiescence to male harassment may be applicable, but as noted elsewhere, females can simply fly away from persistent courting males. Furthermore, because the only male behavior that effectively blocks the access of other males is copulation (pers. obs), mating with a persistent male would only immunize female cicadas from courtship during the actual mating act, would give the female no additional feeding time (feeding may be carried out alone or simultaneously with many other behaviors, including copulation; pers. obs. and similar to ladybird beetles, see Obata 1988) and could only cost oviposition time. Another direct cost may concern male health; in Magicicada, contagious and fatal infections of Massospora cicadina Peck, an entomophagous fungus, produce a class of potential mates best avoided and with whom any interactions carry a risk of infection (Soper 1974, Soper et al. 1976, Alexander et al. ms.). Yet female polyandry cannot be explained as an infection-avoidance tactic or as a tactic to avoid genetically susceptible males because females should be selected to identify and avoid infected males early in the courtship sequence and avoid excessive intimate contact. Lastly, the fact that few females mate multiply suggests that there are not likely to be significant direct phenotypic benefits promoting remating.

Females must evidently require some threshold quantity of sperm before ovipositing; once they use up a substantial quantity of stored sperm, they may remate. Evidence in support of this statement includes the results that females interrupted at two hours produced, on average, fewer eggnests before remating than the total number produced by females who were not interrupted (Table 4.3). However, the experiments also indicate that females interrupted at two hours and deprived of further mating opportunities are capable of producing as many eggnests as uninterrupted females. Possibly, when a female’s sperm stores fall below some threshold volume, she will remate if given the opportunity; yet regardless of whether she has access to mating opportunities or not, she will continue to oviposit for as long as possible. This pattern would be consistent with the existence of an evolved mechanism that caused females to replenish sperm

83 supplies well before exhausting their spermathecal contents, perhaps a form of bet-hedging against difficulties of efficiently finding a mate. Alternatively, if it is never difficult for females to find mates, females may have no evolved mechanisms to stop oviposition. Females may maladaptively oviposit unfertilized eggs (Graham and Cochran 1954) after exhausting their sperm stores because the ready availability of receptive mating partners means that females are never in the position of lacking sperm; thus we would expect no history of selection on females for suspension of oviposition pending sperm replenishment.

84 Literature Cited Alcock, J., G. C. Eickwort, and K. R. Eickwort. 1977. The reproductive behavior of Anthidium maculosum (Hymenoptera: Megachilidae) and the evolutionary significance of multiple copulations by females. Behav. Ecol. Sociobiol. 2: 385- 396. Alexander, R. D., and G. Borgia. 1979. On the origin and basis of the male-female phenomenon. In: M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects. ( Academic Press: New York). Alexander, R. D. 1968. Arthropods. Pp. 167-216 in: T. A. Sebeok, ed. Animal Communication. Techniques of Study and Results of Research . (Indiana Univ. Press: Bloomington, IN). Alexander, R. D., 1975. Natural selection and specialized chorusing behavior in chorusing insects. Pp. 35-77 in: D. Pimentel, ed. Insects, Science, and Society. (Academic Press, New York). Alexander, R. D., D. C. Marshall, and J. R. Cooley, 1997. Evolutionary Perspectives on Insect Mating. Pp. 1- 31 in B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Alexander, R. D., J. R. Cooley, and D. C. Marshall. (ms). A specialized fungal parasite (Massospora cicadina) modifies the behavior of periodical cicadas (Homoptera: Cicadidae: Magicicada). Alexander, R. D., and T. E. Moore. 1962. The evolutionary relationships of 17-year and 13-year cicadas, and three new species. (Homoptera: Cicadidae, Magicicada). Misc. Publ. no. 121. Univ. of Michigan Museum of Zoology, Ann Arbor, MI . Andersson, M. 1994. Sexual selection . (Princeton University Press: Princeton NJ). Arnqvist, G. 1997. The evolution of water strider mating sytems: Causes and consequences of sexual conflicts. Pp. 146- 163 in B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Bella, J. L., R. K. Butlin, C. Ferris, and G. M. Hewitt. 1992. Asymmetrical homogamy and unequal sex ratio from reciprocal mating-order crosses between Chorthippus parallelus subspecies. Heredity 68: 345- 352. Birkhead, T. and A. Møller. 1993. Female control of paternity. TREE 8: 100- 104. Boorman, E., and G. A. Parker. 1976. Sperm (ejaculate) competition in Drosophila melanogaster, and the reproductive value of females to males in relation to female age and mating status. Ecological Entomology 1: 145- 155. Brown, W. D. 1997a. Female remating and the intensity of female choice in black-horned tree crickets, Oecanthus nigricornis. Behavioral Ecology 8: 66-74.

85 Brown, W. D. 1997b. Courtship feeding in tree crickets increases insemination and female reproductive life span. Anim. Behav. 54: 1369-1382. Brown, W. D. 1999. Mate choice in tree crickets and their kin. Ann. Rev. Entomol. 44: 371- 396. Clutton-Brock, T. H., and G. A. Parker. 1995. Sexual coercion in animal societies. Anim. Behav. 49: 1345- 1365. Cooley, J. R., G. S. Hammond, and D. C. Marshall. 1998. The effects of enamel paint marks on the behavior and survival of the periodical cicada, Magicicada septendecim (L.) (Homoptera) and the lesser migratory grasshopper, Melanoplus sanguinipes (F.) (Orthoptera). Great Lakes Entomologist 31: 161-168. Crean, C. S., and A. S. Gilburn. 1998. Sexual selection as a side-effect of sexual conflict in the seaweed fly, Coelopa ursina (Diptera: Coelopidae). Anim. Behav. 56: 1405- 1410. Curtsinger, J. W. 1991. Sperm competition and the evolution of multiple mating. Am. Nat. 138: 93-102. Eberhard, W. G. 1985. Sexual Selection and Animal Genitalia. (Harvard University Press: Cambridge, MA). Eberhard, W. G., 1997. Cryptic female choice. Pp. 32- 57 in B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Eberhard. W. G. 1996. Female control: Sexual selection by cryptic female choice . (Princeton University Press: Princeton, NJ). Emlen, S. T., and L. W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science (Wash. D.C. ) 197: 215- 223. Gabor, C. R., and T. R. Halliday. 1997. Sequential mate choice by multiply mating smooth newts: Females become more choosy. Behav. Ecol. 8: 162- 166. Goulson, D., M. C. Birch, and T. D. Wyatt. 1993. Paternal investment in relation to size in the deathwatch beetle, Xestobium rufovillosum (Coleoptera: anobiidae) and evidence for female selection for large mates. J. Ins. Behavior 6: 539- 547. Graham, C., and A. B. Cochran. 1954. The periodical cicada in Maryland in 1953. J. Econ. Ent . 47:242-244. Gregory, P. G., and D. J. Howard. 1994. A post-insemination barrier to fertilization isolated two closely related ground crickets. Evolution 48: 705-710. Gromko, M. H., and T. A. Markow. 1993. Courtship and remating in field populations of Drosophila. Anim. Behav. 45: 253- 262. Gromko, M. H., M. E. A. Newport, and M. G. Kortier. 1984. Sperm dependence of female receptivity to remating in Drosophila melanogaster. Evolution 38: 1273- 1282.

86 Gwynne, D. T. 1986. Courtship feeding in katydids (Orthoptera: Tettigoniidae): investment in offspring or in obtaining fertilizations? Am. Nat. 128: 342- 352. Harvey, P. H., and R. M. May. 1989. Out for the sperm count. Nature (London) 337: 508- 509. Höglund, J. and R. V. Alatalo. 1995. Leks . (Princeton University Press: Princeton NJ). Howard, D. J., and P. G. Gregory. 1993. Post-insemination signaling systems and reinforcement. Phil. Trans. Roy. Soc. Lond. B. Biol. Sci. 340: 231-236. Howard, D. J., M. Reece, P. G. Gregory, J. Chu, and M. L. Cain. 1998. The evolution of barriers to fertilization between closely related organisms. Pp. 279-288 in D. J. Howard and S. H. Berlocher, eds. Endless Forms: Species and Speciation . (Oxford University Press: New York). Howard, D. J., P. G. Gregory, J. Chu, and M. L. Cain. 1998. Conspecific sperm precedence is an effective barrier to hybridization between closely related species. Evolution 52: 511-516. Jamart, J. A., P. Caseres, M. C. Carracedo, and R. Piñero. 1995. Consequences of homo- and heterospecific rapid remating on the fitness of Drosophila melanogaster females. J. Ins. Physiol . 41: 1019-1026. Janetos, A. C., 1980. Strategies of female mate choice: A theoretical analysis. Behav. Ecol. Sociobiol. 7: 107-112. Jormelainen, V. and S. Merilaita. 1995. Female resistance and duration of mate-gurading in three aquatic peracarids (Crustacea). Behav. Ecol. Sociobiol . 36: 43- 48. Kaitala, A. and C. Wiklund. 1995. Female mate choice and mating costs in the polyandrous butterfly Pieris napi (Lepidoptera: Pieridae) J. Ins. Behavior 8: 355- 363. Karlsson, B. 1996. Male reproductive reserves in relation to mating system in butterflies: A comparative study. Proc. Roy. Soc. Lond. B. 263: 187-192. Kempenaers, B., G. R. Verhayen, M. Van der Broeck, T. Burke, C. Van Broeckhoven, and A. A. Dhondt. 1992. Extra-pair paternity results from female preference for high-quality males in the blue tit. Nature (London) 357: 494-496. LaMunyon, C. W. and T. Eisner. 1993. Postcopulatory sexual selection in an arctiid moth (Utethesia ornatrix) Proc. Nat. Acad. Sci. USA 90: 4689- 4692. Morris, G. K. 1979. Mating systems, paternal investment, and aggressive behavior of acoustic orthoptera. Florida Entomologist 62: 9- 17. Nakagawa, S., G. J. Farias, D. Suda, R. T. Cunningham, and D. L. Chambers. 1971. Reproduction of the Mediterranean fruit fly: Frequency of mating in the laboratory. Ann. Ent. Soc. Amer. 64: 949-950. Neilson, W. T. A., and J. W. McAllen. 1965. Effects of mating on fecundity of the Apple Maggot, Rhagoletis pomonella (Walsh). Canadian Entomologist 97: 276- 279.

87 Obata, S. 1988. Mating refusal and its significance in females of the ladybird beetle, Harmonia axydris . Physiol. Entomol . 13: 193- 199. Oh, R. J. 1979. Repeated copulation in the brown planthopper, Nilaparvata lugens Stål (Homoptera: Delphacidae). Ecological Entomology 4: 345- 353. Price, C. S. C. 1997. Conspecific sperm precedence in Drosophila. Nature (London) 388: 663-666. Pyle, D. W., and M. H. Gromko. 1978. Repeated mating by female Drosophila melanogaster: The adaptive importance. Experientia 15: 449- 450. Rice, W. R. 1996. Sexually antagonistic male adaptation triggered by experimental arrest of female evolution. Nature (London) 381- 232- 234. Rice, W. R. 1998. Intergenomic conflict, interlocus antagonistic coevolution, and the evolution of reproductive isolation. Pp. 261-270 in D. J. Howard and S. H. Berlocher, eds. Endless Forms: Species and Speciation . (Oxford University Press: New York). Rieseberg, L. H., A. M. Desrochers, and S. J. Youn. 1995. Interspecific pollen competition as a reproductive barrier between sympatric species of Helianthus (Asteraceae). Am. J. Bot. 82: 515-519. Rutowski, R. L. 1984. Production and use of secretions passed by males at copulation in Pieris protodice (Lepidoptera: Pieridae). Psyche 91: 141- 152. Sakaluk, S. K., and W. H. Cade. 1982. Orthopteran Mating Systems: sexual competition in a diverse group of insects. (Westview Press: Boulder, CO). Sakurai, T. 1996. Multiple mating and its effect on female reproductive output in the bean bug Reptortus clavatus (Heteroptera: Alydidae). Ann. Ent. Soc. Amer. 89: 481- 485. Sakurai, T. 1998a. Variation in time to sperm depletion and oviposition patterns in females of Riptortus clavatus (Heteroptera : Alydidae). Ann. Ent. Soc. Amer. 91: 737-740. Sakurai, T. 1998b. Receptivity of female remating and sperm number in the sperm storage organ in the bean bug, Riptortus clavatus (Heteroptera : Alydidae). Res. Popul. Ecol. 40: 167- 172. Smith, R. L. 1984. Human sperm competition. Ch. 19 in: Smith, R. L., ed. Sperm competition and the evolution of animal mating systems . (Academic Press: New York). Soper, R. 1974. The genus Massospora entomopathogenic for cicadas. Part I. Taxonomy of the genus. Mycotaxon 1:13-40. Soper, R., A. J. Delyzer, and L. F. R. Smith. 1976. The genus Massospora entomopathogenic for cicadas. Part II. Biology of Massospora levispora and its host Okanagana rimosa, with notes on Massospora cicadina on the periodical cicadas. Ann. Ent. Soc. Amer . 69:89-95.

88 Stone, G. N. 1995. Female foraging responses to sexual harassment in the solitary bee Anthophora plumipes. Anim. Behav . 50: 405- 412. Svärd, L., and J. N. McNeil. 1994. “Female benefit, male risk” polyandry in the true armyworm, Pseudaletia unipunctata. Behav. Ecol. Sociobiol . 35: 319- 326. Thornhill, R. 1984. Alternative female choice tactics in the scorpionfly Hylobittacus apicalis (Mecoptera) and their implications. Amer. Zool. 24: 367- 383. Thornhill, R. and J. Alcock. 1983. The evolution of insect mating systems . (Harvard Unversity Press: Cambridge, MA). Tregenza, T., and N. Wedell. 1998. Benefits of multiple mates in the cricket Gryllus bimaculatus. Evolution 52: 1726-1730. Trillmich, F. and K. G. K. Trillmich. 1984. The mating systems of pinnipeds and marine iguanas: convergent evolution of polygyny. Biol. J. Linn . Soc. 21: 209- 216. Wade, M. J., H. Patterson, N. W. Chang, and N. A. Johnson. 1994. Postcopulatory, prezygotic isolation in flour beetles. Heredity 72: 163- 167. Walker, W. F. 1980. Sperm utilization strategies in nonsocial insects. American Naturalist 115: 780- 799. Ward, P. I., J. Hemmi, and T. Roosli. 1992. Sexual conflict in the dung fly Sepsis cynipsea. Functional Ecology 6: 649-653. Watson, P. J. 1991. Multiple paternity as genetic bet-hedging in female sierra dome spiders, Linyphia litigiosa (Linyphiidae). Anim. Behav. 41: 343- 360. Westneat, D. F, P. W. Sherman, and M. L. Morton. 1990. The ecology and evolution of extra-pair copulations in birds. Current Ornithology 7: 331- 369. White J., 1973. Viable hybrid young from crossmated periodical cicadas. Ecology 54: 573- 580. Whittier, T. S., and T. E. Shelley. 1993. Productivity of singly vs. multiply mated female Meditteranean fruit flies, Ceratitus capitata. J. Kans. Ent. Soc. 66: 200- 209. Zeh, J. 1997. Polyandry and enhanced reproductive success in the harlequin-beetle-riding pseudoscorpion. Behav. Ecol. Sociobiol . 40: 111- 118. Zeh, J. A., and D. W. Zeh. 1996. The evolution of polyandry I: Intragenomic conflict and genetic incompatibility. Proc. R. Soc. Lond. B. 263: 1711- 1717. Zeh, J. A., and D. W. Zeh. 1997. The evolution of polyandry II: Post-copulatory defences against genetic incompatibility. Proc. R. Soc. Lond. B. 264: 69-75.

89 Table 4.1. Mating durations (minutes) in Magicicada septendecim.

195 315 195 330 295 442 167 297 235 235 235 195 310 500 60 124 305 111 354 295 124 331 440 440 310 64 220 444 263 295 354 360 360 250 263 310 65 526 148 91 310 60 405 Average (n= 43) 243:24 ± 121:48

90 Table 4.2. Mating interruption experiment: Likelihood of remating first or ovipositing first, and presence/absence of seminal plug. All P-values Bonferroni corrected for 18 comparisons. £ 1hr category includes combined data from 15 minute/ 1 hr. disruption; £ 2 hr category includes combined data from all disruption treatments.

Treatment Remate first Oviposit first SP No SP Disrupt at 15 min 14 0 0 14 Disrupt at 1 hr. 11 1 1 9 Disrupt at 2 hr. 14 11 6 19 Not disrupted Conspecific (M. septendecim) mate 1 31 26 2 Heterospecific (M. cassini) mate 0 5 5 0

Comparison P (Fisher’s Exact 2- Tail) Remating Frequency 15 min. vs. not disrupt < 0.001* 1 hr. vs. not disrupt < 0.001* 2 hr. vs. not disrupt < 0.001* £1 hr. vs. not disrupt < 0.001* £2 hr. vs. not disrupt < 0.001* 15 min. vs. 1 hr. < 1.000 15 min. vs. 2 hr. < 0.040* 1 hr. vs. 2 hr. < 0.220 M. cassini vs. not disrupt < 1.000 Presence of seminal plug 15 min. vs. not disrupt < 0.001* 1 hr. vs. not disrupt < 0.001* 2 hr. vs. not disrupt < 0.001* £1 hr. vs. not disrupt < 0.001* £2 hr. vs. not disrupt < 0.001* 15 min. vs. 1 hr. < 1.000 15 min. vs. 2 hr. < 1.000 1 hr. vs. 2 hr. < 1.000 M. cassini vs. not disrupt < 1.000 * indicates statistically significant difference.

91 Table 4.3. Counts of female M. septendecim eggnests in mating interruption experiments. Conspecific mates are M. septendecim; heterospecific mates are M. cassini. Eggnests Additional after first eggnests after Total Mate mating Remated second mating eggnests Uninterrupted mating: Conspecific 29 no - 29 Conspecific 31 no - 31 Conspecific 14 no - 14 Conspecific 26 no - 26 Conspecific 34 no - 34 Conspecific 22 no - 22 Conspecific 23 no - 23 Conspecific 34 no - 34 Conspecific 1 no - 1 Average 23.8 ± 10.7 23.8 ± 10.7

Heterospecific 45 no - 45 Heterospecific 32 no - 32 Heterospecific 29 no - 29 Heterospecific 97 no - 97 Average 50.8 ± 31.6 50.8 ± 31.6

Mating interrupted at 2 hr: Conspecific 9 yes 0 9 Conspecific 23 yes 3 26 Conspecific 0 yes 7 7 Conspecific 10 yes 20 30 Conspecific 9 yes 20 29 Conspecific 0 yes 20 20 Conspecific 14 yes 12 26 Average 9.29 ± 8.0 11.7 ± 8.6 21.0 ± 9.5

Conspecific 5 no - 5 Conspecific 35 no - 35 Conspecific 41 no - 41 Average 27.0 ±19.3 27.0 ± 19.3

Conspecific 39 no possibility - 39 Conspecific 51 no possibility - 51 Conspecific 35 no possibility - 35 Conspecific 10 no possibility - 10 Conspecific 87 no possibility - 87 Conspecific 30 no possibility - 30 Conspecific 33 no possibility - 33 Average 40.7 ± 23.8 40.7 ± 23.8

Comparison vs. uninterrupted conspecific mating (Wilcoxon Signed Ranks): Pre-remating eggnests, conspecific mating interrupted at 2 hours with possibility of remating: P £ 0.018 Total eggnests, conspecific mating interrupted at 2 hours with subsequent remating: P £ 0.128 Total eggnests, conspecific mating interrupted at 2 hours without subsequent remating: P £ 0.593 Total eggnests, conspecific mating interrupted at 2 hours without possibility of remating: P £ 0.090 Total eggnests, heterospecific mating interrupted at 2 hours with possibility of remating: P £ 0.068

92 Table 4.4. Mated and unmated female M. septendecim responses to 2 minute playbacks of male calling song. Females that produced at least one wing flick in response to playbacks of male songs were scored as responding positively. Day Female n Response P (+) (-) (Fisher’s Exact 2- tailed) 1 Mated 22 0 22 Unmated 17 9 8 £ 0.001 2 Mated 22 0 22 Unmated 16 8 8 £ 0.001 3 Mated 20 0 20 Unmated 15 7 8 £ 0.001 4 Mated 18 0 18 Unmated 14 7 7 £ 0.001

93 Table 4.5. Mating interruption and female M. septendecim wing flick responses to playbacks of male calls. Mating-interrupted females were likely to signal in response to male calls, and all disruption treatments were equally likely to cause females to remain responsive. Number of Number of females responding females Treatment with wing flick signal not responding Disrupt at 15 min. 8 1 Disrupt at 1 hr. 8 3 Disrupt at 2 hr. 5 4 Not disrupted 0 17

Comparison P (Fisher’s Exact 2- tail)

15 min. vs. not disrupt £ 0.001* 1 hr. vs. not disrupt £ 0.001* 2 hr. vs. not disrupt £ 0.002* 15 min. vs. 1 hr. £ 0.591 15 min. vs. 2 hr. £ 0.294 1 hr. vs. 2 hr. £ 0.642

* indicates statistically significant difference.

94 Table 4.6. Census data from female attractiveness experiment. The cage containing unmated females had a significantly greater number of visiting males than did the empty cage (Wilcoxon Signed Ranks test, Z= -3.320, P£ 0.001), and significantly fewer males visited the cage containing mated females than the cage containing unmated females (Wilcoxon Signed Ranks test, Z= 3.331, P£ 0.001).

Males found on Males found on Males found on cage containing cage containing empty Time unmated females mated females cage Day 1 1:08 1 0 0 1:37 2 0 0 2:27 2 0 0 2:39 0 0 0 3:20 4 0 0 4:09 3 0 0 4:31 2 0 0 5:15 0 0 0 Day 2 9:00 5 0 1 9:31 5 0 0 10:01 6 2 1 10:30 4 0 0 10:50 5 0 0 11:07 2 0 0 11:17 0 0 0 11:30 2 0 0 11:42 2 0 1

95 with 0.068 0.090 0.128 0.018 ² ² ² ²

P P P P ) M. cassini Heterospecific ( Mate (control), mating-interrupted females allowed to Not Permitted To remate Total M. septendecim gned Ranks): nt remating: ossibility of remating: terrupted females not premitted to remate, and mated bility of remating: bility of remating: Conspecific mating Permitted to remate Pre-remating Control ). 0 70 60 50 40 30 20 10 M. cassini . Numbers of eggnests laid by normally-mated female Figure 4.1 remate (pre-remating eggnests and total eggnests), mating in heterospecifics ( Comparison vs. uninterrupted conspecific mating (Wilcoxon Si Pre-remating eggnests, conspecific mating interrupted with p Total eggnests, conspecific mating interrupted with subseque Total eggnests, conspecific mating interrupted without possi Total eggnests, heterospecific mating interrupted with possi

96 CHAPTER V

SEXUAL SIGNALING IN PERIODICAL CICADAS, MAGICICADA SPP.

Abstract

We describe the nature and timing of a previously unknown female sexual receptivity signal in Magicicada. Females produce “wing-flick” signals in response to male calls with a timing that is species-specific. Males perceive both acoustical and visual components of wing-flick signals. Chorusing males perceiving female signals alter their behavior by increasing the number of calls between flights and localizing their mate search. A male receiving repeated nearby responses to his calls approaches the signaling female while calling and begins late-stage courtship behaviors leading to copulation. Using the wing-flick signal as an assay of female receptivity, we demonstrate the functional significance of the complex, two-part structure of male calls in M. septendecim, M. tredecim, and M. neotredecim. These species’ calls include a main element and a downslur; calls without downslurs are unlikely to provoke female responses against a background chorus, but such call fragments may elicit responses in the absence of a background chorus. Normally, only receptive females produce wing-flick signals; however, a fungal pathogen evidently increases its odds of transmission by causing infected males to signal. We also report a novel form of male-male acoustic competition:

97 Males engaged in close-range or protracted courtships obscure the downslurs of arriving potential interlopers with “interference buzzes,” reducing the likelihood of a female response and thus increasing the likelihood that interlopers will depart and search elsewhere.

Introduction

Detailed studies of insect communication have provided valuable insights into topics as diverse as insect systematics (e.g. Alexander 1957a; Walker 1962, 1963), speciation (e.g. Henry 1985, Otte 1992, Butlin and Ritchie 1994, Alexander et al. 1997), development and inheritance (e.g. Fulton 1933, Bigelow 1960, Olvido and Mousseau 1995, Shaw 1996), conflicts of interest and deception (e.g. Lloyd 1966, Lloyd 1979, Spooner 1968, Otte 1974; Alexander et al. 1997), and sexual selection (e.g. Alexander 1975, Thornhill and Alcock 1983, Ritchie 1996). The “singing insects” (sound-producing members of the Orthoptera and Cicadidae) have played a central role in this research. Many key contributions have resulted from long-term synthetic studies of individual species or species groups (e.g. Alexander and Moore 1962, Walker 1964, Lloyd 1966, Spooner 1968, Lloyd 1979, Otte and Alexander 1983, Otte 1994). One such group, the periodical cicadas of eastern North America (Magicicada spp.1, Table 5.1), has attracted interest because of its extraordinary combination of unique traits, including long life cycles (13 or 17 years), synchronized development, periodical adult emergences, near-perfect sympatry and synchrony of multiple related species, dense populations and loud, conspicuous male sexual behaviors. Periodical cicada aggregations served as the primary model for the concept of the non-resource-based lek (Alexander 1975), and they offer an important opportunity to compare and contrast mating aggregations in insects and vertebrates

1 Because the songs and behaviors of the 13- and 17- year cognate species are generally similar, they are abbreviated herewith as M. –decim, M. –cassini, and M. –decula.

98 (Bradbury 1981, 1985). Although their remarkable characteristics have made Magicicada a focus of research on ecology, life cycle evolution, and behavior (Williams and Simon 1995), central questions about Magicicada mating behavior and communication remain unanswered.

Calling male periodical cicadas aggregate to form choruses that attract sexually receptive females (Alexander and Moore 1958, 1962). Teneral (adult immature) and ovipositing females are also found in these aggregations, although they may not be specifically attracted to them. Magicicada chorusing consists of “sing-fly” behavior, or short (ca. 3-15 s) bouts of singing alternated with local flights of variable length (up to several meters; Alexander and Moore 1962, Alexander 1975, Dunning et al. 1979). Under some conditions males appear to engage in local mate search behavior involving extremely short flights or walking between calls. Upon locating a potentially receptive female, a chorusing male may begin courtship, which involves a stereotypical series of behaviors and distinctive acoustical and visual signals (Alexander and Moore 1962; Alexander 1968, 1975; Dunning et al. 1979; Table 5.2, Fig. 5.1). Courtship may be followed quickly by copulation or may last hours and include long periods of waiting (Alexander 1968). These observations have led to the suggestion that periodical cicada females may be extremely choosy, imposing strong sexual selection on males (Alexander 1975) and that males locate females by progressively narrowing their search as they shift from chorusing to local mate search to courtship. The cues causing males to identify a suitable courtship target and switch from chorusing (not directed at a specific individual) to courtship behavior (directed at an individual female), or from extended courtship to copulation, have until now remained unknown (Alexander 1968). The function of male calling song has also remained poorly understood; Alexander (1975) noted that, because females are attracted to choruses of many individuals yet no males within choruses employ silent searching tactics, the calling song must play a role in individual males’ mating success.

99 We describe a visual and acoustical female receptivity signal, hereafter referred to as a “wing-flick.” Receptive Magicicada females flick their wings with a quick motion in timed response to male calls; males perceiving the signal respond by immediately dropping out of the chorus and engaging in courtship behaviors directed at the signaling female. The discovery of this signal has facilitated the development and testing of new hypotheses for the adaptive nature of Magicicada chorusing behavior and the evolution of male calling and courtship songs. The discovery also provides new information about the transmission of a specialized fungal parasite of Magicicada, Massospora cicadina Peck. Although normal males have never been observed making wing-flick signals, males infected with the fungus sometimes produce wing-flick signals in response to the calls of other males; this behavioral modification is presumably a fungal adaptation to increase spore transmission (see also Alexander et al. ms). In addition, we report acoustical jamming signals produced by Magicicada males in close-range competition for receptive females, a form of intrasexual mating competition not previously reported in insects.

General materials and methods

Periodical cicadas in different regions emerge in different years; the populations of different regions are called “broods.” We first observed female Magicicada signaling to males in Brood I in 1995. We continued our studies of sexual signaling in 1996-1998 (Table 5.3) on both 13-and 17-year cicadas, concentrating on the M. -decim species but including the others when sufficient numbers were available. We used a Macintosh computer and Canary software (Cornell Bioacoustics Laboratory) for acoustical analyses, and Sound Edit software (MacroMedia) for modification of songs and synthesis of artificial sounds. Playback equipment consisted of a Sony WM-D6C cassette player or a Macintosh powerbook computer connected to a Radio Shack SA-10 amplifier driving a 3” midrange speaker for M. -decim calls or a tweeter for M. -cassini calls. We maintained playback call intensity at natural levels, ca. 75 db at 20 cm. In all years, we kept cicadas in captivity, but

100 within their natural environment, by placing them in ca. 200 liter cages made by enclosing living vegetation with black fiberglass screen or white nylon tulle. Some observations were completed in larger 4x4x4 m “flight cages” placed over living woody vegetation. Many playback experiments were conducted using 22x24x22 cm screen test chambers.

Magicicada nymphs usually emerge and undergo their final molt in the evening (Maier 1982). The next morning, newly emerged “teneral” adults are identifiable by their dull color, soft bodies (especially diagnostic in the ovipositor), and by their position low in the vegetation. Although mated females commonly have a hardened white seminal plug in the genital opening (White 1973), this plug is occasionally absent in mated females and is therefore an unreliable indicator. To ensure that we used only unmated females in our playback experiments, we collected teneral females early each morning and stored them in single-sex cages.

Female responses to male calls

The nature of the female signal and the “rapprochement duet”

In 1995-1998, we observed interactions between male and female Magicicada. We noted whether females produced wing-flick signals and how the responses were timed in relation to male calls, using sonograms from audio-and videotape to measure timing. We also recorded female M. –decim and M. -cassini responses to playbacks of male calling song and how males responded to natural and simulated wing-flick signals.

Results. We have documented female wing-flick signals in all Magicicada species except M. tredecula, which was not present in sufficient numbers to study in 1998. In all species studied, a female signals in response to a calling conspecific male by moving her wings in a single, quick motion which produces a broad-frequency sound of approximately < 0.02 s duration. The motion and sound are similar to wing flutters produced in response to disturbance, except that wing flutters consist of multiple wing movements. The signal’s

101 visual and acoustical components appear unspecialized but the timing of the signal in relation to the male’s call is species-specific; the signal therefore consists of both the wing- flick and the timing of its delivery. In M. septendecim, females produce the signal an average of 0.387 ± 0.106 s after the end of the male calling phrase (Fig. 5.2). The delay in M. cassini, 0.705 ± 0.112 s, is nearly twice as long (Fig. 5.3). These species’ 13-year counterparts appear to have similar or identical signal timing. M. septendecula females produce multiple wing-flicks in the brief silences between subphrases during the second part of a male’s two-part call (Fig. 5.4).

In all species, a receptive female produces wing-flick signals in response to each of the male’s calls as he approaches, engaging in a “rapprochement duet” with the calling male (Fig. 5.1). This duet continues until the male switches to CII courtship calling. We have not determined what stimuli cause the male to switch from the duet to CII calling; he usually does so once within approximately 1-15 cm of the female, perhaps upon making close visual contact. In M. -decim the female does not respond during the male’s CII song, which is a continuous stream of shortened calling phrases without silent gaps; if the male stops calling, the female responds with a wing-flick after the appropriate delay. M. -decim males continue CII calling until reaching the female; once positioned next to the female, the male switches to CIII courtship calling and attempts to copulate. The rapprochement duet in M. -cassini is similar to that in M. -decim, except that females sometimes signal during male CII calling. In M. septendecula there may not be a homologous CII courtship song; after the rapprochement duet the male may proceed directly to CIII courtship and mounting. Females do not signal during CIII in any Magicicada species studied. The context of the signal, and male reactions to it, suggest its purpose: Sexually receptive females signal to obtain mates.

102 The relationship of the female response to sexual receptivity in M. septendecim

To demonstrate that immature females are not sexually receptive and do not signal, in 1998, we six daily collections of approximately 25 teneral females each. Each day, we played recordings of male song to each of the day-cohorts and watched for wing-flick signals. To demonstrate that mated females do not wing-flick, in the 1997 study, we allowed 22 individually-marked (Chapter 2) female M. septendecim to mate once and divided them between two cages along with 22 marked, unmated females of the same age. In each of four consecutive days we played recorded M. septendecim calling songs to the females for two minutes and recorded the number of mated and unmated females responding.

Results. Females first signaled 5-8 days after emerging (Table 5.4). This period of nonresponsiveness is consistent with Magicicada “teneral periods” reported elsewhere (Maier 1982; see also Karban 1981, Young and Josephson 1983) and with similar periods in nine species of phaneropterine katydids (Spooner 1968). Variable weather during maturation probably causes the inexactness of the relationship between age and maturity. Since sclerotization involves chemical processes such as protein cross-linking, it is affected less by time than by temperature and humidity (Chapman 1971 pp. 443-446). In the experiment comparing mated and unmated females, none of the mated females responded to playbacks, while at least half of the unmated females responded each day with wing-flick signals (Table 5.5). This difference was significant in each of the four days (P £ 0.001; Fisher’s Exact Two-Tailed Test).

Response of M. septendecim females to playbacks of heterospecific song

Magicicada aggregations contain as many as four different species (Alexander and Moore 1962; Chapter 6), yet interspecific hybrids are rarely observed (Alexander and

103 Moore 1962; Dybas and Lloyd 1962, White 1973). To demonstrate the species-specific nature of the wing-flick response, in 1996 we played recorded M. septendecim calling phrases alternating with M. cassini calling phrases to 25 caged, unmated M. septendecim females and noted their responses. We also included three unmated M. cassini females as controls; we could not collect enough to include more. We tested the females in groups of five, playing a series of 15 alternating M. septendecim and M. cassini calls (30 calls total) and recording female responses.

Results. M. septendecim females routinely responded to the M. septendecim call phrases in each trial. Only one of 25 M. septendecim females ever responded to a heterospecific playback (Table 5.6). The three M. cassini females in the experiment responded to conspecific calls and never responded to heterospecific playbacks.

Male responses to female signals

Effects of the signal on male behavior in M. septendecim

In 1996 and 1997, to document the effects of the female signal on male behavior, we produced sounds and movements imitating female wing-flick signals for chorusing male M. septendecim. Before we identified a suitable device for producing simulated wing-flicks, we used a strip of paper that we flicked with our fingers. Later, we discovered that toggling an ordinary household electric light switch was more convenient for the experimenter. Because male responses to both artificial stimuli appear indistinguishable, we combined trials with both methods in our statistical analyses.

A single M. -decim call phrase consists of a 1.5-4.0 s buzz (the “main element”) of constant and nearly pure pitch followed by a brief (ca. 0.35 s) “terminal downslur” ending about 500 Hz lower than the main element pitch. To document the importance of correct timing in eliciting male M. septendecim courtship, we produced simulated wing-flicks in response to the calls of males that had landed and begun calling on nearby vegetation,

104 timing our signals either (1) during the main element, (2) during the downslur, or (3) after the downslur. We placed the flicking device within 25 cm of the male along the branch on which he had just landed. We included control trials in which the experimenter approached in the same fashion but did not produce simulated wing-flicks. We scored a male as responding positively if he moved toward the stimulus and began late-stage courtship behaviors such as CII or CIII calling, foreleg-vibrate, or mounting behavior. Each trial ended when the male flew or walked away from the stimulus, or when the male remained motionless longer than 20 seconds.

The above experiment simulated a scenario in which the male alights near a signaling female. To examine male responses in a scenario involving weaker and/or inconsistent female responses, we conducted trials in which we presented individual chorusing males with either a single nearby or repeated distant simulated wing-flicks. We produced single, nearby simulated wing-flicks in response to the first call following a flight at a distance of approximately 25 cm from the male. For multiple distant simulated wing- flicks, we produced the signals after each of the male’s calls from a distance of 1.3 m. We again included control trials in which the experimenter approached in the same manner but did not produce simulated wing-flick signals. In each trial we recorded the number of calls the male made in his current bout, the nature of his next action (sit, walk, fly), and the direction and distance of movement. We stopped monitoring males that paused for longer than 20 seconds.

Results. Only wing-flicks produced after the downslur caused males to respond positively. Males usually responded to such stimuli by walking toward the stimulus while calling. In this behavior, termed “call-walking,” males stopped walking for ca. 1 second immediately following each downslur. This pattern is distinct from chorusing behavior which involves bouts of calling alternated with flights or silent walks. Males were equally

105 unresponsive to the control treatment and to simulated wing-flicks produced during the main element or during the downslur (Table 5.7).

Males responded to single nearby and multiple distant simulated wing-flick signals in a manner suggesting an attempt to localize the stimulus (Table 5.8). Both kinds of stimuli caused males to increase the number of calls in the current calling bout compared to control males; most males then flew to a new calling perch instead of call-walking toward the stimulus. Whether walking or flying after the calling bout, males presented with the single nearby or multiple distant simulated wing-flicks were more likely to move in the direction of the stimulus than control males were. In control trials, males were more likely to move away from the stimulus than toward it, suggesting that the presence of the experimenter probably disturbs the cicada.

Components of the female signal

Because female wing-flick signals are often audible, we assumed early in our investigation that males locate receptive females primarily by listening for wing-flick signals. However, the importance of investigating separately the visual and acoustical components of the female signal became apparent when we substituted an audiotaped recording of a female wing-flick for the mechanically produced simulated flicks of our earlier experiments. Male M. -decim did not give consistent or strong responses to these playbacks, suggesting that they must perceive the visual component of the female signal (actual signals and signals simulated with paper strips and electrical switches involve a synchronized sound and quick movement, while the speaker produces only a sound). Spectral analysis of wing-flick sounds indicates that they are broad-frequency sounds with most of the energy above 4 kHz. It is possible that the different Magicicada species perceive different aspects of female signals; although such high-frequency sounds are within the range of maximal hearing sensitivity of M. -cassini, female wing-flicks may be above the range of maximal hearing sensitivity for male M. –decim (Simmons et al. 1971;

106 Huber et al. 1980). We conducted experiments addressing the effects on male M. –cassini and M. –decim of visual signals, acoustical signals, and signals containing both acoustical and visual information on male chorusing behavior and male courtship behavior.

To examine the effects of timed flick sounds alone on male chorusing behavior, we constructed a clicking device by attaching a 12 volt relay to the end of a 1-meter wooden pole, covering all wires and dark (cicada-colored) parts of the relay with white masking tape. In test trials, we identified an actively chorusing male M. septendecim or M. cassini, placed the relay within 15 cm of him along the same branch, and clicked the relay in time to his calls, following him as he moved and taking care not to make any timed movements observable by the male. For controls, we placed the device in the same manner, but did not click the relay. For each male, we measured the number of calls in each of two calling bouts and the distance moved between the two calling bouts.

To investigate the effects of timed visual signals alone and the significance of other visual cues such as the presence of a cicada-colored object, we approached chorusing male M. septendecim with a model consisting of a black ballpoint pen cap or another model made from an identical cap painted white. Without making sounds, we moved the model once rapidly back and forth about 2 cm at the appropriate time for a female signal, or we held it still. The black pen cap was similar in size and color to a real cicada, while the white cap lacked only the appropriate color stimulus.

To determine whether combined visual and acoustical stimuli are more potent than either alone, we used the following experimental design: From the surrounding chorus, we selected 48 male M. septendecim and 48 male M. cassini one at a time and placed each in a 22x24x22 cm test chamber. Two walls of the chamber consisted of three layers of dark, opaque cloth; the remaining walls consisted of fiberglass window screen. In one treatment, we suspended a motionless model cicada inside the test chamber and responded to the male’s calls with clicks produced behind the opaque chamber sides. In the second

107 treatment, the experimenter held the model inside the chamber and responded to the calling male by moving the model slightly with the appropriate timing, without making clicks. In the third treatment, the experimenter responded by producing clicks and by moving the model. After the start of each trial, we recorded the male’s behaviors for the next two minutes. Male M. septendecim and M. cassini were scored as responding positively to the model if they exhibited any of the following behaviors during the 5 minute trial: CII or CIII call, extrude genitalia, climb onto the model, foreleg vibrate, or attempt to copulate.

Results. Timed click sounds did not affect male M. septendecim behavior, but male M. cassini given click sounds flew significantly shorter distances between calls (Table 5.9). Male M. septendecim courted plastic pen caps that moved in time with their calls, but were less likely to engage in late-stage courtship with white colored caps than with black colored caps (Fig. 5.5), indicating that visual stimuli alone are sufficient to provoke male responses and that the stimulus includes the presence of an appropriately colored cicada- like object. The model that moved and clicked simultaneously was most attractive to M. septendecim and M. cassini (Fig. 5.6). These results suggest that even though sounds or movements may provoke courtship under certain circumstances, in both M. septendecim and M. cassini, synchronized sounds and movements are most likely to be perceived as female signals.

Behavioral bisexuality of males infected by the fungal pathogen Massospora cicadina

Magicicada are sometimes infected by the fungus Massospora cicadina Peck (Speare 1921, Soper 1974, Soper et al. 1976, Lloyd et al. 1982, White and Lloyd 1983). Stage I (conidial phase) of this fungus is asexual; individual Magicicada so infected lose their abdominal terminalia, exposing a chalky mass of conidiospores. Infected cicadas tend to walk around on vegetation, spreading the spores by rubbing the exposed mass on surfaces (Alexander et al. ms.). Individuals contacting these infective conidiospores in the

108 same emergence develop a Stage II (resting spore) infection. These individuals also lose their abdominal terminalia, but they tend to make long flights, during which the friable resting spores spill from the abdomen (Alexander et al. ms.). These resting spores are believed to infect cicada nymphs of the next generation (Soper 1974, Soper et al. 1976).

In 1996, while observing free-flying Magicicada, we observed a Stage I Massospora-infected male M. septendecim respond to a courting male twice with wing- flick signals (Fig. 5.7). The courting male responded as expected by attempting to copulate with the infected, signaling male, who rebuffed the courter by flapping his wings. These observations were striking because, in hundreds of hours spent to that point observing Magicicada in natural and controlled circumstances, we had never observed normal, uninfected males making wing-flick signals in response to other males’ calls. The observation of signaling by infected males led us to use playback experiments to compare the infected males’ responsiveness to that of unmated adult females.

We placed M. septendecim four at a time in a cage and played 15 calling song phrases to each group of cicadas. Trials began with a playback of male calling song at normal pitch (1.39 kHz); then we first increased, then decreased the pitch in steps, and trials ended with a playback of the calling song at normal pitch. We scored cicadas as giving a positive response to a calling phrase if they responded with wing-flick signals to a minimum of two phrases. We played normal and frequency-altered M. septendecim calling songs to 8 wild-caught uninfected males, 6 Stage I Massospora-infected males, and 32 unmated females. We later tested the responsiveness of 5 Stage II Massospora-infected male M. septendecim to playbacks of unaltered M. septendecim calling song.

Results. Stage I fungus-infected M. septendecim responded to playbacks with appropriately timed wing-flicks, while no normal male or Stage II infected male ever responded. Stage I infected males were generally less responsive than mature females (Table 5.10). We have also observed Stage I infected M. tredecassini responding with

109 wing-flick signals to other males’ calls (Fig. 5.8). We expect that fungus-induced bisexuality will be found in all Magicicada species.

Components of male calls

Alexander (1975) noted that “failure to find evidence of nonsinging males searching through cicada choruses implies that the male’s song remains an essential part of his ability to acquire females.” The discovery of the female sexual signal and its timing in relation to the male’s call confirms this prediction and suggests a more detailed functional hypothesis regarding the structural elements of M. –decim and M. -cassini calls, which we have developed and tested primarily in M. septendecim.

The call phrases of M. -decim males are 1.5-4.0 seconds in length, ending with a frequency downslur (Fig. 5.2). The nearly pure-tone main elements of M. -decim calling songs overlap to produce a uniform chorus drone; as male density and activity increases, it becomes difficult to perceive individual calls of individual males. In such situations the downslur is the only portion of the male’s call that stands out from the background chorus. Because females must be able to respond with a signal timed from the end of a male’s call, we suspected that the downslur might function to enhance the end of the male’s call and improve a calling male’s odds of being perceived by a nearby receptive female. Because males have never been observed producing downslurs only, producing a “main element” must also improve a male’s likelihood of provoking a female response. This hypothesis leads to several predictions. First, if the main element alone is ever sufficient to elicit a signal, its effectiveness should decrease as the background chorus increases in intensity. Second, downslurs alone should not be sufficient to elicit wing-flick signals, unless the background chorus is loud enough to sufficiently stimulate the female. Finally, neither the downslur nor the main element alone should be as effective as the intact call in eliciting the female response at any background level.

110 To determine which components of male M. septendecim calls most effectively elicit a female response under different chorus conditions, we constructed a pure-tone artificial calling phrase, a pure-tone phrase lacking a downslur, a pure-tone downslur alone, and a simulated pure-tone background chorus. In previous experiments, we found that females respond similarly to playbacks of recorded and pure-tone artificial calls. We noted the responses of individually marked females in 19 groups of females during 146 playbacks of complete calls, calls lacking slurs, and slurs only at an intensity of 72-79 db (intensity varied within the test chamber) over a background chorus of four different intensities (0 db, 58-62 db, 63-77 db, 65-80 db).

Results. Although call fragments did elicit female responses under certain conditions, females were more likely to respond to whole calls than to partial calls at all background chorus intensities (Table 5.11). As the background chorus intensity increased, female responsiveness to whole calls and main elements of whole calls declined, while females became more responsive to slurs (Table 5.12; Fig. 5.9), such that at the highest intensity, females were more responsive to slurs alone than to main elements alone.

Inter-male acoustical interference behavior

In 1996-1998 we observed a previously undescribed male sound in M. -decim and M. -cassini species; this sound is composed of a short (ca. 0.25 s) “buzz” with a frequency spectrum similar to that of the main element of the calling song but somewhat less pure- toned. This sound is always produced during the downslur of another male’s call (Fig. 5.10). In part because the downslur is important in eliciting female wing-flick signals, we suspected that a male engaged in close-range courtship uses this sound to obscure a newly- arrived competitor’s downslurs, thereby decreasing the likelihood that the courted female will reveal her presence with a wing-flick signal before the competitor completes his short calling bout and departs.

111 Alternatively, the sounds could be aborted calls caused by interruption or when the stimuli inducing the call are insufficient to stimulate a complete call phrase. This hypothesis does not predict that the buzzes should be consistent in length or timing in relation to another male’s call, nor does it predict that the buzz should be produced under limited circumstances. Another explanation for this behavior is that males might use this sound to signal their sex to other males; males often court other males, so the sound may discourage misdirected sexual attention. A similar explanation has been proposed for a “flick-tick” signal produced by M. cassini males (Dunning et al. 1979). This hypothesis predicts that the signal should be observed most often when males are crowded and encounter one another commonly in the chorus.

We conducted three experiments to evaluate these hypotheses in M. septendecim. In the first, designed to determine if crowded chorus conditions alone can induce buzzes, we placed 30 chorusing males together in a 1x1x1 m cage, watched them interact, and listened for the male buzzes. After 20 minutes of watching, we produced simulated wing- flick signals in response to males on one side of the cage for one minute and then observed the males for another five minutes. We repeated this experiment twice.

In the second experiment, intended to simulate the events involved in the appearance of a potential interloper during courtship, we presented 22 individual males with the following series of stimuli: First, we confined each male in a small screen test chamber. We then played one to five minutes of recorded calling song at an intensity designed to simulate a male calling at a distance of approximately 10 cm. Males often began to call during this treatment, but if the playbacks were not sufficient to stimulate calling, we produced simulated wing-flick signals in response to the playbacks using an electrical switch held outside the chamber within view of the male; signals from the switch always induced call-walking from the male. Once the male had begun calling, we stopped the playbacks and responded to the male’s calls for 10-60 seconds, after which we ceased

112 responding. Once the male stopped his calling or courtship songs (some males had begun CII or CIII calling during this duet), we resumed playbacks of calling song from the speaker at the same volume and distance. We noted the context of any interference buzzes produced by the male during the trial.

In the third test, we examined the effects of artificial pure-tone buzzes on female responsiveness by creating a model call and buzz such that the call could be played with or without the buzz. We confined four unmated female M. septendecim in the test chamber and played a sequence of 30 call pairs; each pair consisted of (1) a normal call played from one speaker followed by (2) a call played from the same speaker with a buzz played from a second speaker and superimposed over the slur. We recorded the number of females responding to each call and repeated the experiment 6 times, using different females. We compared the number of females responding positively to each call pair using a Friedman Two Way analysis of variance.

Results. In trial 1 of the first test, no males produced buzzes prior to the production of simulated wing-flick signals, but at least 2 buzzes were heard once we began to respond to the males with simulated wing-flick signals. In trial 2, one buzz was heard during the first 20 minutes, but over 20 were heard once we began responding to the males. Males called often and frequently landed within centimeters of each other during the trials, and although some males harassed or attempted to court other males, these interactions did not lead to male buzzes.

In the second test, males never produced buzzes in response to the initial series of playbacks, never produced buzzes during artificial duets between the playbacks and the experimenter, and never produced buzzes while duetting with the experimenter. However, in 16 of the 22 trials, males began producing buzzes once the playbacks had resumed following the termination of the male’s duet, usually in response to the first or second playback call, but sometimes not until 3 or 4 calls had played. Males producing buzzes did

113 so only during the downslur of the recorded calling song phrases. In five of the trials, the experimenter again began producing simulated wing-flick signals to the playback while the male buzzed; in four of these five cases this caused the male to cease buzzing and begin call-walking near the simulated wing-flick stimulus. In the fifth case the male walked while buzzing after each playback call.

In the third test, the presence of artificial buzzes significantly decreased the likelihood that a female would respond with wing-flick signals to an artificial calling song (Table 5.13). Under the conditions of these experiments, buzzes halved the likelihood of a female response.

Discussion

Wing-flick signals in Cicadidae

Communicative wing-flicking (sometimes called wing-tapping, -banging, - clapping, -clacking, or -clicking) appears to be widespread in cicadas. We use the term “wing-flick” for Magicicada because it connotes movement and sound, both of which are perceived by males; other terms emphasize only the acoustic component of the signal. Male wing-flicking during close-range courtship interactions with females has been reported in Australian and New Zealand Kikihia, and Amphipsalta, (Dugdale and Fleming 1969, Lane 1995) North American Okanagana (Davis 1919, Alexander 1957b, Chapter 8), and European Tibicina (Fonseca 1991), while males combine wing-flicks with long-range calling song in Asian Cicadetta (Popov 1981), Australian and New Zealand Amphipsalta (Dugdale and Fleming 1969, Lane 1995), and Western North American Platypediinae (Moore 1968). Female wing flick signaling is known in North American Magicicada and Okanagana, (Davis 1919, Chapter 8), Australian Cystosoma (Doolan 1981), Cicadetta (Gwynne 1987), and Amphipsalta (Dugdale and Fleming 1969), and New Zealand Amphipsalta, Kikihia, Maoricicada, Notopsalta, and Rhodopsalta (Lane 1995, Dugdale

114 and Fleming 1969), with the most detailed published reports of female wing-flick signaling from Kikihia spp. (Lane 1995), Amphipsalta cingulata (Lane 1995), Cystosoma saundersii (Doolan 1981), and Cicadetta quadricinctata (Gwynne 1987). In each species studied in detail, female wing-flick signals elicit male courtship behavior and appear to have a specific temporal relationship to the male’s song. In a few cases female wing-flick signals, while present, are apparently not always prerequisites for mating: Although sometimes both sexes of Okanagana canadensis and O. rimosa appear to use wing-flicks to signal their presence, females most often signal receptivity simply by approaching stationary calling males (Chapter 8).

Wing-flick signals and the Magicicada courtship sequence

In Magicicada, female wing-flick signals in response to male calling song cause males to cease chorusing and begin courtship. A female Magicicada wing-flick signal consists of a quick, timed movement of the wings that makes a broad-frequency sound varying from a faint rustle to a loud snap; the signal’s timing in relation to the male’s call is species-specific. A chorusing male perceiving a single wing-flick signal or repeated, distant wing-flick signals increases his number of calls in the current bout and then resumes chorusing behavior while moving in the direction of the stimulus, apparently to elicit further responses and obtain better information about the female’s location. If the male receives repeated nearby responses that allow him to localize the female, he ceases chorusing behavior (calls alternated with flights) and call-walks toward the responding female; the male may use short flights to cross gaps in the vegetation, but such short, irregular flights are distinct from chorusing flights. In M. -decim and M. -cassini species, the male may begin CII calling once within approximately 15 cm. of the female. Female M. -decim do not wing-flick during the CII call, which contains no silent gaps; female M. - cassini occasionally respond during males’ CII calling. After contacting the female or while preparing to mount, males of all species begin CIII calling, which they continue until

115 the genitalia are engaged. Receptive females of all species do not wing-flick during CIII and remain still throughout the entire courtship sequence. In nearly every mating that we have observed, the cicadas have performed, in order, each of the behaviors of this courtship sequence; exceptions include rare forced copulations that occur when females are trapped by vegetation or in aggregations of males.

Males respond to both the acoustical and visual components of the signal. We do not know to what extent male Magicicada can localize sound alone without a visual component, but the wing-flick sounds alone do not elicit later courtship stages as rapidly or consistently as do visual cues. The acoustical and visual stimuli sufficient to cause male responses in all Magicicada species must be quite general, because the click and movement of an ordinary electric light switch are highly attractive to courting males. Since male cicadas directed the full repertoire of normal sexual behaviors towards active, signaling models and ignored these same models when no signals were perceptible, male courtship songs and the female wing-flick signals appear to be sufficient for sexual rapprochement in these species.

Male-male acoustic interference competition

The discovery of the rapprochement duet in Magicicada improves our understanding of male-male competitive interactions in choruses. Most significant is the finding that male M. -decim and -cassini appear to use “interference buzzes” to acoustically jam rival males’ signals in competition for mates. Interference buzzes are produced specifically in the context of a close-range male-female duet or a prolonged courtship that has been interrupted by the arrival of a calling, and potentially interloping, male competitor. The courting male emits short buzzes coincident with the downslurs of his rival’s calls; these buzzes obscure the downslur of the rival’s call and reduce the likelihood that the female will perceive and respond to them. If the rival male does not elicit responses from a female, he is likely to continue chorusing and depart. One potential objection to this

116 hypothesis is that the buzz itself could reveal the presence of the nearby receptive female. For the buzz to function to reduce interloping, calling males either must be unable to discern the buzz from the background chorus or they must be unable to hear sufficiently while calling; otherwise, males should have evolved to respond to interference buzzes by intensifying their local search. Male M. septendecim and M. cassini hearing sensitivity is reduced 5-15 dB during calling by the action of muscles that reduce tension on the tympana (Pringle 1954, Simmons et al. 1971); thus calling males may have difficulty perceiving that their signals are being jammed by buzzing males. Also, buzzing males terminate the signal precisely at the end of the interloper’s song, suggesting that buzzing behavior has been selected to be produced only when it is undetectable by the potential interloper.

Other insects are known to make use of specialized male-male competitive signals. For example, some male katydids produce sounds that appear to mimic female responses to their own calling songs, apparently to confuse potential interlopers (Alexander 1975, Grove 1959). A similar function could be served by male wing-clicking during calling in cicadas such as Amphipsalta cingulata, in which males wing-click to their own calling songs with a timing identical to that of female wing-flick responses (Lane 1995). The Magicicada interference buzz appears uniquely complex, however, in that the signal apparently deceives the courted female in order to achieve a competitive advantage over a rival male.

The intense nature of Magicicada male mating competition likely results in few females remaining in the mating pool long after the onset of sexual receptivity; the wing- flick signal is so potent, and searching males so abundant, that newly-receptive females may be detected almost immediately by one of the many chorusing males, in which case wing-flick rapprochement duets may be brief. In addition, unless a teneral female’s readiness to mate is activated all at once, fully formed, a newly adult female may be expected to pass through a period of partial receptivity in which (1) she produces the wing-

117 flick signal very weakly and/or inconsistently and (2) she is more likely to reject males as a result of fluctuations in her mating readiness. If so, males likely have evolved to respond to very weak or intermittent signs of timed wing-flicking in females. Such a period of partial or weak sexual receptivity could be involved in observations of lengthy courtships involving repeated rejections before eventual copulation (Alexander 1968, Dunning et al. 1979) and the apparent coyness of Magicicada females. Subtlety of wing-flick responses in newly-matured females, in addition to rapid detection of signaling females by chorusing males, could explain why the Magicicada wing-flick signal has for so long remained undiscovered.

Acoustic synchrony in M. –cassini .

The ability of M. –cassini males to synchronize their calls may be an adaptation allowing individual males to maximize their probability of perceiving female responses. All available evidence (Simmons et al. 1971; Huber et al. 1980; this study) suggests that male M. –cassini can detect the acoustical component of female wing flick signals. Unlike M. –decim calls, M. –cassini calls are broad frequency, and overlap the frequency range of female wing flick signals (Fig. 5.3). We propose that male M. -cassini have evolved to delay beginning calling phrases until a nearby conspecific begins, so that each male maximizes his chances of hearing nearby female wing flicks without acoustical interference. The net effect of an entire chorus pursuing this strategy is for males to synchronize calling and flying; thus, acoustic detection of female signals may be an explanation for the well-documented phenomenon of synchronized chorusing behavior in M. –cassini (Alexander and Moore 1958).

Wing-flick signals and the parasitic fungus Massospora cicadina

Alexander et al. (in prep.) show that the fungus Massospora cicadina alters Magicicada behavior in ways that likely increase spore transmission, and that the alterations

118 in behavior are different for the two fungal life history stages in ways that are appropriate for the stage-specific modes of transmission. Our observations suggest that the first stage of the fungus has apparently evolved to increase the odds of spore transmission between males by causing infected males to produce wing-flick signals in response to the calls of other males even while retaining normal male courtship behaviors. Thus, while females are in jeopardy from sexual contact primarily with infected males, males are in jeopardy from sexual contact with infected individuals of both sexes. Second stage Massospora-infected males do not produce wing-flick signals, consistent with the theory that spores of this stage do not infect cicadas of the same generation (Soper 1974, Soper et al. 1976). We do not know the physiological mechanism or timing of infection by Stage I fungal spores, but the behavioral bisexuality of infected males strongly supports the argument that spore transmission between individuals of the same generation is crucial for the Stage I fungus.

The evolution of Magicicada chorusing behavior

Alexander (1975) noted that, in Magicicada, it is apparently the sound of the entire chorus that attracts females, rather than the song of an individual male. Therefore, an individual male Magicicada with chorusing behavior that is less effective in long-range attraction but more likely to be detected by stationary receptive females would realize a fitness advantage relative to those males with chorusing behaviors optimal for long-range attraction. For Magicicada, the discovery that receptive females respond to the calls of individual males with timed wing-flicks resolves the question, raised by Alexander (1975), of why males do not silently parasitize the mate-attracting abilities of others by searching without calling: Because male calls provoke female responses, males who search without signaling sacrifice the opportunity to detect females by provoking responses.

Magicicada chorusing behavior lies at one extreme of the range of rapprochement behavior found in cicadas. Males alternate unusually brief bouts of calling (and unusually short call phrases, especially in M. -decim and M. -cassini) with short flights, a

119 rapprochement strategy that is centered around searching for responding females who have moved to the chorus and become stationary. Only one other well-studied cicada species, the Australian Tick-tock cicada (Cicadetta quadricinctata) has a similar rapprochement system, although calling bouts in this species are still three to four times longer in duration than those of Magicicada (Gwynne 1987, Alexander and Moore 1962). Gwynne (1987) noted the similarities of male behavior in Magicicada and Cicadetta quadricincta and all but predicted that female wing-flick signals would be found in Magicicada. At the other extreme are species in which males produce a continuous song (with or without separate phrases) from a single location for relatively long periods, such as in North American Okanagana canadensis and O. rimosa (Chapter 8). This strategy emphasizes stationary male advertisement over searching, and in such species females approach males for mating. A similar range of rapprochement strategies is found in other singing insects such as Phaneropterine katydids (Spooner 1968, Heller and von Helversen 1993). The Magicicada rapprochement system is similar to those of fireflies (see Lloyd 1966, 1979) and substrate- vibrating (Hunt and Nault 1991, Hunt et al. 1992, Claridge 1985) and lacewings (Wells and Henry 1992, Henry 1994).

Understanding the causes of taxonomic variation in rapprochement requires an understanding of the relative costs and benefits of alternative signaling and searching behaviors for males and females in each species under consideration (Alexander et al. 1997). Males, especially in species such as insects with little or no paternal investment, are expected to be more risk-tolerant than females in rapprochement since their reproductive success is most directly influenced by the number of mates obtained (Trivers 1972, Alexander and Borgia 1979, Parker 1979) and hence by the effectiveness of mate-locating activities. Because of the sexes’ unequal risk tolerance, male mate-locating activities are influenced primarily by patterns of female dispersion (Emlen and Oring 1977), while females are influenced by selection for effective parental investment. The relative costs and benefits of alternative rapprochement behaviors for individuals of each sex will be affected

120 by interrelated factors including the kinds and pattern of predation, the spatial distribution of appropriate habitats and resources (Emlen and Oring 1977, Cocroft and Pogue 1996, Heller and von Helversen 1993), the sex-specific costs and benefits of searching vs. remaining stationary, and the current strategies of the opposite sex.

Periodical cicada rapprochement behavior involves a remarkable combination of related unusual features which have likely resulted from adaptation to high population density. Because the phylogenetic relationship of Magicicada to other cicada genera is poorly known, little information is available on the rapprochement behavior of the most recent non-periodical ancestor of Magicicada. However, the uniqueness of the Magicicada rapprochement system suggests that Magicicada must have evolved from ancestors with relatively stationary, advertising males and relatively mobile females. The following general hypothesis explains the unique features of Magicicada; similar selective pressures have likely shaped related attributes of rapprochement systems of other organisms (e.g. Heller and von Helversen 1993).

Upon the evolution of partial or complete periodicity (see Lloyd and Dybas 1966a, b; Cox and Carlton 1991, Long 1993, Yoshimura 1997) high population density arising from lack of ecological control by predators and parasites became a consistent feature of Magicicada ecology. This presumably led to two important effects with dramatic consequences for the Magicicada mating system: First, males and females began to experience reduced risks associated with movement and/or signaling. Second, the greater density of receptive females increased the chances that a male could encounter a receptive female while moving through his environment. These factors together would have improved the potential payoff to individual males of increasing time spent searching at the expense of advertisement; increased male searching would have caused females to benefit most from moving to areas of loud male chorus sound and remaining still once there, in part because increased movement by males would reduce the ability of females to approach

121 individual males before their next flights. This strategy of moving to the chorus and becoming stationary could have resulted in selection favoring a tendency in males to be attracted to the choruses of other males, completing the evolution of the basic elements of the Magicicada rapprochement system.

It is probably true that for any individual of a species in which females approach calling males, unreceptive females can avoid calling males, and, by voluntarily approaching calling males, sexually receptive females are able to mate on demand. For species in which males locate females, however, females can reduce their rapprochement time by making it easier for males to find them. Although in dense Magicicada emergences, it would seem as though males would have little difficulty finding females, any individual female is likely surrounded by a collection of immature cicadas, cicadas of the same sex, and mated cicadas, as well as a small number of sexually receptive members of the opposite sex. Female Magicicada may have evolved to signal their receptivity to ensure timely mating in confusing, dense aggregations. Males would then have been selected to become efficient at detecting these signals and minimizing their chances of mistakenly courting unreceptive cicadas or other inert objects.

It is not difficult to imagine the advent of stationary females resulting in the evolution or refinement of a female wing-flick signal to reduce rapprochement time, although the female signal may have existed from the start in the nonperiodical ancestor. An environment of locally aggregated, stationary females who respond to the call terminus with a wing-flick would select males further for a local search strategy emphasizing short calling bouts, short calling phrases, and short flights. The density of male aggregations would allow the chorus sound itself to assume the female-attraction function of the male’s call, allowing males to modify their calling and searching behaviors to be most effective for local mate searches, possibly resulting in (1) quieter calls, (2) more pure-tone calls in the

122 M. –decim species, and (3) calls with terminal downslurs in the M. –decim and M. cassini species or their common ancestor.

The best tests of the above scenario will likely come from comparative study using data from pair-forming systems in other Cicadidae, but at present few species have been studied in significant detail. One pattern within Magicicada offers initial support for the theory: Two Magicicada species, M. septendecula and M. tredecula, are consistently less abundant than other Magicicada; these species have longer calling bouts and a longer call unit that contains multiple temporal “windows” for female wing-flick responses. Our playback experiments confirm a different aspect of the hypothesis, that male calling song has evolved under selection for distinctiveness in loud choruses. The two-part structure of M. –decim and M. –cassini calling songs may provide a record of the evolution of song distinctiveness from an ancestral call without a downslur: The downslur functions at least in part to mark the end of a call against a background chorus dominated my main elements.

123 Literature Cited Alexander, R. D. 1957a. The taxonomy of the field crickets of the eastern United States (Orthoptera: Gryllidae: Acheta). Ann. Ent. Soc. Am. 50: 584-602. Alexander, R. D. 1957b. Sound production and associated behavior in insects. Ohio Journal of Science 57: 101-113. Alexander, R. D. 1968. Arthropods. Pp. 167-216 in: T. A. Sebeok (ed). Animal Communication. Techniques of Study and Results of Research. (Indiana University Press: Bloomington, IN). Alexander, R. D. 1975. Natural selection and specialized chorusing behavior in chorusing insects. Pp. 35-77 in: D. Pimentel, ed. Insects, Science, and Society . (Academic Press: New York). Alexander, R. D., and T. E. Moore. 1958. Studies on the acoustical behavior of seventeen- year cicadas (Homoptera: Cicadidae: Magicicada). Ohio J. Science. 58: 107-127. Alexander, R. D., and T. E. Moore. 1962. The evolutionary relationships of 17-year and 13-year cicadas, and three new species. (Homoptera: Cicadidae, Magicicada). Misc. Publ. no. 121. Univ. of Michigan Museum of Zoology, Ann Arbor, MI . Alexander, R. D., and G. Borgia. 1979. On the origin and basis of the male-female phenomenon. Pp. 417-440 in: M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects. (Academic Press: New York). Alexander, R. D., D. C. Marshall, and J. R. Cooley. 1997. Evolutionary Perspectives on Insect Mating. Pp. 1-31 in: B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Alexander, R. D., J. R. Cooley, and D. C. Marshall. (ms). A specialized fungal parasite (Massospora cicadina) modifies the behavior of periodical cicadas (Homoptera: Cicadidae: Magicicada). Bigelow, R. S. 1960. Interspecific hybrids and speciation in the genus Acheta (Orthoptera: Gryllidae). Can. J. Zool. 38: 509-524. Bradbury, J. W. 1981. The evolution of leks. Pp. 138-169 in: R. D. Alexander and D. W. Tinkle, eds.. Natural selection and social behavior . (Chiron Press: New York). Bradbury, J. W. 1985. Contrasts between insects and vertebrates in the evolution of male display, female choice, and lek mating. Fortschr. Zool. 31: 273-289. Butlin, R. K., and M. G. Ritchie. 1994. Behaviour and speciation. In: P. J. B. Slater, and T. R. Halliday, eds. Behaviour and Evolution. (Cambridge University Press).

Chapman, R. F. 1971. The Insects: Structure and Function, 2 nd Edition . (Elsevier: New York).

124 Claridge, M. F. 1985. Acoustic signals in the Homoptera: Behavior, Taxonomy, and Evolution. Ann. Rev. Entomol. 30: 297-317. Cocroft, R. B., and M. Pogue. 1996. Social behavior and communication in the neotropical cicada Fidicina mannifera (Fabricius) (Homoptera: Cicadidae). J. Kansas Entomol. Soc . 69(4) suppl.: 85-97. Cox, R. T., and C. E. Carlton. 1991. Evidence of genetic dominance of the 13-year life cycle in periodical cicadas (Homoptera: Cicadidae: Magicicada spp.). Am. Midl. Nat . 125: 63-74. Davis, W.T. 1919. Cicadas of the genera Okanagana, Tibicinoides, and Okanagodes, with descriptions of several new species. J. N. Y. Entomol. Soc. 27: 179-223. Doolan, J. M. 1981. Male spacing and the influence of female courtship behavior in the bladder cicada, Cystosoma saundersii Westwood. Behav. Ecol. Sociobiol. 9:269- 276. Doolan, J. M., and D. Young 1989. Relative importance of song parameters during flight phonotaxis and courtship in the bladder cicada Cystosoma saundersii. J. Exp. Biol. 141 1989: 113-131. Dugdale, J. S., and C. A. Fleming. 1969. Two New Zealand cicadas collected on Cook’s Endeavour voyage, with a description of a new genus. New Zealand J. Sci. 12: 929-957. Dunning, D. C., J. A. Byers, and C. D. Zanger. 1979. Courtship in two species of periodical cicadas, Magicicada septendecim and Magicicada cassini. Anim. Behav. 27: 1073-1090. Dybas, H. S., and M. Lloyd. 1962. Isolation by habitat in two synchronized species of periodical cicadas (Homoptera: Cicadidae: Magicicada). Ecology 43: 444-459. Emlen, S. T., and L. W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science (Wash. D. C.) 197: 215-223. Fonesca, P. J. 1991. Characteristics of the acoustic signals in nine species of cicadas (Homptera: Cicadidae). Bioacoustics 3: 173-182. Fulton, B. B. 1933. Inheritance of song in hybrids of two subspecies of Nemobius fasciatus (Orthoptera). Ann. Ent. Soc. Amer. 26: 368-376. Grove, D. G. 1959. The natural history of the angular-winged katydid Microcentrum rhombifolium PhD. Dissertation, Cornell University. Gwynne, D. T. 1987. Sex-biased predation and the risky mate-locating behaviour of male tick-tock cicadas (Homoptera: Cicadidae). Anim. Behav. 35: 571-576. Heller, K., and D. von Helversen. 1993. Calling behavior in bushcrickets of the genus Poecilimon with differing communication systems (Orthoptera: Tettigonioidea, Phaneropteridae). J. Insect Behav. 6: 361-377. Henry, C. S. 1985. Sibling species, call differences, and speciation in green lacewings (Neuroptera: Chrysopidae: Chrysoperla ). Evolution 39: 965-984.

125 Henry, C. S. 1994. Singing and cryptic speciation in insects. Trends Ecol. and Evol. 9:388-392. Huber, F., D. W. Wohlers, and T. E. Moore. 1980. Auditory nerve and interneurone responses to natural sounds in several species of cicadas. Physiol. Entomol . 5: 25- 45. Hunt, R. E., and L. R. Nault. 1991. Roles of interplant movement, acoustic communication, and phonotaxix in mate-location behavior of the Graminella nigrifrons. Behav. Ecol. Sociobiol. 28: 315-320. Hunt, R. E., J. P. Fox, and K. P. Haynes. 1992. Behavioral response of Graminella nigrifrons (Homptera: Cicadellidae) to experimentally manipulated vibrational signals. J. Ins. Behav. 5: 1-13. Karban, R. 1981. Effects of local density on fecundity and mating speed for periodical cicadas. Oecologia 51: 260-264. Lane, D. H. 1995. The recognition concept of species applied in an analysis of putative hybridization in New Zealand cicadas of the genus Kikihia (Insecta: Hemiptera: Tibicinidae). Ch. 17 in: D. M. Lambert and H. G. Spencer, eds., Speciation and the Recognition Concept: Theory and Application. (John Hopkins). Lloyd, J. E. 1966. Studies on the flash communication system in Photinus fireflies. Misc. Publ. no. 130. Univ. of Michigan Museum of Zoology, Ann Arbor, MI . Lloyd, J. E. 1979. Sexual selection in luminescent beetles. Pp. 293-342 in M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects . (Academic Press: New York). Lloyd, M., and H. S. Dybas. 1966a. The periodical cicada problem. I. Population ecology. Evolution 20: 133-149. Lloyd, M., and H. S. Dybas. 1966b. The periodical cicada problem. II. Evolution. Evolution 20: 466-505. Lloyd, M., J. White, and N. Stanton. 1982. Dispersal of fungus-infected periodical cicadas to new habitat. Environmental Entomology 11:852-858. Long, C. A. 1993. Evolution in periodical cicadas: A genetical explanation. Evol. Theory 10(4): 209-220. Maier, C. T. 1982. Obervations on the seventeen-year periodical cicada, Magicicada septendecim (Hemiptera: Homptera: Cicadidae). Ann. Ent. Soc. Amer. 75: 14-23. Moore, T. E. 1968. The evolution of cicadas and their acoustical behavior. Thirteenth Internat. Congr. Entomol . Abstracts of Papers: 173 (Moscow). Otte, D. 1974. Effects and functions in the evolution of signaling systems. Ann. Rev. Ecol. Syst. 5: 385-417. Otte, D. 1992. Evolution of cricket songs. J. Orthop. Res. 1: 25-49.

126 Otte, D. 1994. The crickets of Hawaii: Origin, systematics and evolution (The Orthopterists’ Society: Philadelphia, PA). Otte, D., and R. D. Alexander. 1983. The Australian crickets (Orthoptera: Gryllidae). Acad. Nat. Sci. Phil. Monogr . 22: 1-477. Olvido, A. E., and T. A. Mousseau. 1995. Effect of rearing environment on calling-song plasticity in the striped ground cricket. Evolution 49: 1271-1277. Parker, G. A., 1979. Sexual selection and sexual conflict. Pp. 123-167 in: M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects . (Academic Press: New York). Popov, A. V. 1981. Sound production and hearing in the cicada Cicadetta sinuatipennis Osh. (Homptera: Cicadidae). J. comp. Physiol. 142: 271-280. Pringle, J. W. S. 1954. A physiological analysis of cicada song. J. Exp. Biol. 34: 525- 560. Ritchie, M. G. 1996. The shape of female mating preferences. Proc. Nat. Acad. Sci. USA 93: 14628-14631. Shaw, K. L. 1996. Polygenic inheritance of a behavioral phenotype—interspecific genetics of song in the Hawaiian cricket genus Laupala. Evolution 50: 256-266. Simmons, J. A., E. G. Wever, and P. M. Pylka. 1971. Periodical cicadas: Sound production and hearing. Science (Wash. D. C. ) 171: 212-213. Soper, R. 1974. The genus Massospora entomopathogenic for cicadas. Part I. Taxonomy of the genus. Mycotaxon 1: 13-40. Soper, R., A. J. Delyzer, and L. F. R. Smith. 1976. The genus Massospora entomopathogenic for cicadas. Part II. Biology of Massospora levispora and its host Okanagana rimosa, with notes on Massospora cicadina on the periodical cicadas. Ann. Ent. Soc. Amer . 69: 89-95. Speare, A. T. 1921. Massospora cicadina Peck, a fungous parasite of the periodical cicada. Mycologia 13: 72-82. Spooner, J. D. 1968. Pair-forming acoustic systems of Phaneropterine katydids (Orthoptera: Tettigoniidae). Anim. Behav. 16: 197-212. Thornhill, R.A., and J. Alcock, 1983. The evolution of insect mating systems . (Harvard University Press: Cambridge, MA). Trivers, R. L. 1972. Parental investment and sexual selection. Pp. 136-179 in: B. Campbell, ed. Sexual selection and the descent of man, 1871-1971 . (Aldine: Chicago). Walker, T. J. 1962. The taxonomy and calling songs of United States tree crickets (Orthoptera, Gryllidae, Oecanthinae). I. The genus Neoxabea and the niveus and varicornis groups of the genus Oecanthus. Ann. Ent. Soc. Amer . 55: 303-322.

127 Walker, T. J. 1963. The taxonomy and calling songs of United States tree crickets (Orthoptera, Gryllidae, Oecanthinae). II. The nigricornis group of the genus Oecanthus. Ann. Ent. Soc. Amer. 56: 772-789. Walker, T. J. 1964. Cryptic species among sound-producing ensiferan Orthoptera (Grylidae and Tettigoniidae). Q. Rev. Biol . 39:345-355. Wells, M. M., and C. S. Henry. 1992. The role of courtship songs in reproductive isolation among populations of green lacewings of the genus Chrysoperla (Neuroptera: Chrysopidae). Evolution 46: 31-42. White J., 1973. Viable hybrid young from crossmated periodical cicadas. Ecology 54: 573-580. White, J., and M. Lloyd. 1983. A , Massospora cicadina Peck (), in emerging nymphs of periodical cicadas (Homoptera: Cicadidae). Environmental Entomology 12: 1245-1252. Williams, K. S., and C. Simon. 1995. The ecology, behavior, and evolution of periodical cicadas. Annu. Rev. Entomol. 40: 269-295. Young, D. and R. K. Josephson. 1983. Pure-tone songs in cicadas with special reference to the genus Magicicada. Journal of Comparative Physiology 152: 197-207. Yoshimura, J. 1997. The evolutionary origins of periodical cicadas during ice ages. Am. Nat. 149: 112-124.

128 Table 5.1. Traits distinguishing Magicicada species. "Pronotal extensions" are lateral extensions of the pronotum between the eyes and wing bases. For additional descriptions see Alexander and Moore 1962.

Species Life Cycle Abdominal sternites Dominant call Pronotal Length of call Range Habitat Peak chorus (yrs.) pitch (kHz) extension (seconds) activity color

orange with black lateral band § midwestern, M. neotredecim Marshall and Cooley 13 1.35 - 1.90 orange 1.5 - 4 upland morning

129 or center Ozark

§ midwestern, M. tredecim (Walsh and Riley) 13 mostly orange 1.00 - 1.25 orange 1.5 - 4 mixed morning southern

orange with black lateral band § M. septendecim (L.) 17 1.25 - 1.50 orange 1.5 - 4 eastern, plains upland morning or center

black, rarely with weak †† 17 > 3.00 2 - 6 §§ eastern, plains lowland afternoon M. cassini (Fisher) orange lateral band black

black, rarely with weak †† midwestern, 13 > 3.00 2 - 6 §§ lowland afternoon M. tredecassini Alexander and Moore orange lateral band black southern

† M. septendecula Alexander and Moore 17 black with orange lateral band > 3.00 black 7 - 14 eastern, plains upland midday

midwestern, 13 > 3.00 † M. tredecula Alexander and Moore black with orange lateral band black 7 - 14 southern upland midday

§ Roughly pure-tone, musical buzz terminating in a noticeable drop in pitch; no ticks. Usually 2-3 calls between flights.

§§ Rapid series of ticks followed by high-pitched, broad-spectrum buzz that rises and then falls in intensity and pitch. Usually 1-2 calls between flights.

† Repeated, rhythmic high-pitched, broad-spectrum tick-buzz phrases, followed by repeated phrases containing only ticks. Usually 1 call between flights.

†† Orange band, if present, often interrupted medially. Table 5.2. The sexual sequence in Magicicada species. Female Female Female Female Chorus Wing Courtship Wing Wing Wing Species Calling Flick Synchrony CI Calling Flick Behaviors CII Calling Flick Behaviors CIII Calling Flick Behaviors

phrases similar to walking, repeated, shortened phrases of the same Foreleg 2-3 phrases walking repeated those used in calling flying type as in court I, with greater emphasis onAt end vibrate, M. -decim between Yes No Yes between staccato No 130 and separated by between the slur, or terminal portion, and without of CII mount, flights phrases buzzes silent gaps phrases intervening silence copulate

repeated, shortened phrases of the same phrases similar to walking, Foreleg 1 phrase type as in court I, with greater emphasis on walking repeated Under some those used in calling flying vibrate, M. -cassini between Yes Yes the slur, or terminal portion, without Yes between staccato No conditions and separated by between mount, flights intervening silences and with additional ticks phrases buzzes silent gaps phrases copulate after the downslur

phrases similar to walking, Foreleg 1 calling bout repeated those used in calling flying vibrate, M. -decula between Yes No Yes none known - none known staccato No and separated by between mount, flights buzzes silent gaps phrases copulate Table 5.3. Study Sites, 1995- 1998.

Year Brood Life Cycle Location County State Characteristics 1995 I 17 Alum Springs Rockbridge VA Logged site 1996 II 17 Horsepen Lake SWF Buckingham VA Logged site 1997 III 17 Siloam Springs SP Brown, Adams IL Clearing 1998 XIX 13 Harold Alexander WMA Sharp AR Cleared field

131 Table 5.4. Length of teneral period, in days, as measured by first female M. septendecim responses to playback of male songs. Females were divided into groups containing approximately 25 cicadas and placed in single-sex cages on living vegetation. Females were exposed daily to playbacks of male song, and the day on which at least one female in the cage responded with wing flick signals was recorded.

Group First response (days after emergence) A 6 B 7 C 8 D 7 E 6 F 5 Average 6.5

132 Table 5.5. Mated and unmated female M. septendecim responses to 2- minute playbacks of male calls. Females responding positively produced at least one wing flick signal in response to playbacks of male songs.

Day Mating n Response P Status (+) (-) (Fisher Exact Test) 1 Mated 22 0 22 Unmated 17 9 8 £ 0.001

2 Mated 22 0 22 Unmated 16 8 8 £ 0.001

3 Mated 20 0 20 Unmated 15 7 8 £ 0.001

4 Mated 18 0 18 Unmated 14 7 7 £ 0.001

133 Table 5.6. Wing flick (WF) responses of 25 female M. septendecim and 3 female M. cassini to a series of 135 alternating male M. septendecim and M. cassini calls.

Number responding Species n to M. septendecim call to M. cassini call M. septendecim 25 25 1 M. cassini 3 0 3

134 Table 5.7. Male M. septendecim responses to artificial wing flicks. We simulated female wing flick signals by flicking a strip of paper or by clicking a heavy-duty light switch at the appropriate time in relation to male call. We scored males as responding positively to the artificial wing flicking if they moved toward the stimulus and began late- stage courtship behaviors such as CII or CIII call, forleg-vibrate, or mounting behavior. Fisher’s Exact 2- tailed tests are used to compare treatments to 46 controls in which the clicking device was presented to the male, but no click was made; only 6 males responded positively to such controls.

n (+) (-) vs. Timing n (+) (-) P Control 46 6 40 End of call 83 66 17 £ 0.001 During Call 41 3 38 £ 0.498 During Slur 13 0 13 £ 0.326

135 e, rpreta- 0.003 (Fisher’s Exact Test) Exact 0.003 (Fisher’s Test) Exact 0.02 (Fisher’s 0.355 (Wilcoxon) 0.005 (Wilcoxon) 0.221 (Wilcoxon) 0.223 (Fisher’s Exact Test) Exact 0.223 (Fisher’s 0.001 (Wilcoxon) ≤ ≤ ≤ P ≤ ≤ ≤ P ≤ 29.8 (n = 47) 15.14 (n= 15) 1.26 (n= 19) ± ± 2.46 (n = 53) ± ± 26.55 (n= 13) 20.98 19.4 (n = 40) 22.60 0.83 (n= 17) 3.53 ± ± 1.16 (n = 43) 3.70 ± ± movement direction (flights and walks) following artificial wing flick signal or control (no signal) or control (no wing flick signal following artificial and walks) direction (flights movement Control signal With Control signal With M. septendecim Male Toward ObserverToward from ObserverAway 16 3Fly after signalDo not fly after signal 4 7 13 13 ObserverToward from ObserverAway 7 0 5 14 2 4 Fly after signalDo not fly after signal 7 36 15 38 Multiple distant simulated wing-flick signals Multiple distant simulated wing-flick Direction Likelihood of flight Distance of flight (cm)Direction 35.76 Distance of flight (cm) 25.75 Call Number 1.94 Likelihood of flight Table 5.8. Table on a clock fac representing the directions as a value from 1-12 with the numbers of movement was recorded treatment. Direction Single simulated wing-flick signal Single simulated with only the directions 11, 12, and 1 considered “toward stimulus” and 4, 5, and 6 “away from stimulus,” to avoid biased inte 4, 5, and 6 “away from stimulus,” “toward stimulus” and 12, and 1 considered 11, with only the directions were not monitored further. paused for longer than 20 seconds lateral movements. Males that tion of ambiguous Call Number 2.28

136 Table 5.9. Experiments on male chorusing behavior. Clicking device placed near chorusing male and clicked or not clicked (control). Number of calls in first and second call bouts and length of first and second flights after treatment or control were measured. Results analyzed with Kruskal-Wallis one-way analysis of variance.

M. septendecim Click vs. control P First call bout (64/46) £ 0.277 First flight (50/37) £ 0.573 Second call bout (45/34) £ 0.712 Second flight (39/32) £ 0.117

M. cassini Click vs. control P First call bout (24/26) £ 0.001 First flight (23/25) £ 0.000 Second call bout (18/22) £ 0.399 Second flight (15/19) £ 0.001

137 Table 5.10. Numbers of M. septendecim responding with wing flick signals to frequency-modified playbacks of male song. Mature females made most reponses to unmodified playback songs, as did Stage I Massospora cicadina infected males. Normal males and Stage II infected males never responded to playbacks.

Song Mature Stage I Stage II Normal Frequency (kHz) Female (32) Male (6) Male (5) Male (8) 2.35 0 0 - 0 1.91 2 0 - 0 1.65 13 1 - 0 1.48 16 2 - 0

1.39 (unmodified) 22 3 0 0

1.30 20 1 - 0 1.13 18 0 - 0 0.956 8 0 - 0 0.869 1 0 - 0

138 Table 5.11. Contrasts of female wing flick responses to playbacks of complete artificial calling songs and partial artificial calling songs (main portion only, slur portion only) at four background chorus intensities. P- values are for Fisher’s Exact 2- tailed tests. Whole artificial calling song is always more effective in eliciting a female response than either partial call. Background is a continuous 1.39 kHz pure tone; background levels: 0db, 58- 62 db, 63- 77 db, 65- 80 db, where maximum value is approximately equal to acoustic intensity of playback call. Background Contrast 0 db 58- 62 db 63- 77 db 65- 80 db Whole vs. Slur 0.001* 0.001* 0.001* 0.005* Whole vs. Main 0.001* 0.001* 0.001* 0.001* Main vs. Slur 0.001* 0.004* 0.465 0.012*

139 Table 5.12. Effects of background chorus intensity on female wing flick responses to playbacks of complete artificial calling songs and partial artificial calling songs (main portion only, slur portion only) at four background chorus intensities. Increasing background chorus intensity significantly decreased the likelihood that whole calls and the main portions of calls would elicit female responses and generally increased the likelihood that slur alone would elicit responses. Background is a continuous 1.39 kHz pure tone; background levels: 0db, 58- 62 db, 63- 77 db, 65- 80 db, where maximum value is approximately equal to acoustic intensity of playback call. Effect of increasing chorus intensity, Kruskal-Wallis one way analysis of variance: Playback type Z P Whole 420 £ 0.001 Main portion only 64 £ 0.001 Slur portion only 44 £ 0.001

140 Table 5.13. Effects of M. -decim “interference buzz.” We confined four, unmated female M. septendecim in a test chamber and played a sequence of 60 calls, alternating normal calls and calls with buzzes, recording the number of females responding to each call. We repeated the experiment 6 times.

Average number of P (Friedman Females responding Two-Way Replicate Buzz to each call analysis of variance.) A Yes 1.13 ± 0.86 £ 0.001 No 2.23 ± 0.77 B Yes 0.57 ± 0.57 £ 0.00 No 2.43 ± 0.68 C Yes 0.80 ± 1.00 £ 0.001 No 1.73 ± 1.08 D Yes 1.27 ± 1.93 £ 0.006 No 0.82 ± 0.58 E Yes 0.93 ± 0.87 £ 0.045 No 1.53 ± 0.78 F Yes 0.47 ± 0.57 £ 0.00 No 2.07 ± 0.98

141

female female female answers answers answers

male CI male CI male CII male CIII

142 male approaches male wf wf wf female, foreleg "sing-fly" male lands male male approaches vibrates, mounts, behavior and calls calls female copulates

Figure 5.1. Stylized sonogram of male call/female wing flick coutship duet for M. -decim. Male begins with cout I (CI) calls; female answers each call with a wing flick. After several such interactions, male begins court II (CII) calling, which consists of repeated phrases of the same general type as CI calling, but shortened and without intervening silence. After the male ceases CII calling, the female wing flicks in response, and the male begins cour III (C III) calling, which consists of repeated staccato buzzes. M. -cassini and M. -decula courtship sequences are similar, except that M. -decula lack clearly defined CII calls. 10

8

6

143 4

2

kHz 0 0 0.5 1.0 1.5 2.0 2.5 S *

Figure 5.2. Sonogram of male M. -decim call and female wing flick response. Female response (marked with asterisk) is a broad-frequency sound. Female produces signal 0.387 s after end of male call (n= 235 examples, ± 0.110s). Wing flick sound enhanced and extraneous background noise removed for clarity. 10 144 8 6 4 2 kHz 0 0 1 2 3 4 S *

Figure 5.3. Sonogram of male M. -cassini call and female wing flick response. Female response (marked with asterisk) is a broad-frequency sound. Female produces signal 0.705 s after end of male call (n= 16 examples, ± 0.112s). Wing flick sound enhanced and background chorus removed from 2 to 4 seconds for clarity

10

8

6 145

4

2

0 kHz

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 S * * Figure 5.4. Sonogram of male M. -decula call and female wing flick responses. Female responses (marked with asterisks) are broad-frequency sounds. Female produces signals during silent portions of male call. Wing flick sound enhanced and extraneous background noise removed for clarity.

Figure 5.5. Comparison of male responses to a model cicada constructed from a pen cap. Cap was either black or white, and experimenter either moved at the appropriate time for a female wing flick signal, or held it still. Male responses, including call-walk towards, CII call, CIII call, and Foreleg vibrate were noted. Fisher's Exact 2- tailed Test P-values for pairwise comparisons of treatements plaed below a stylized sonogram of courtship sequence in M. -decim.

Artificial Artificial Artificial signal signal signal 146

(male CI) Call-Walk towards CII CIII Foreleg vobrate

Black/move (22) vs. black/still (21) 0.001 0.004 0.108 0.001 White/move (21) vs. white/still (20) 0.001 0.233 1.00 1.00 Black/move (22) vs. white/move (21) 0.129 0.045 0.345 0.001 Black/still (21) vs. white /still (20) 0.233 1.00 1.00 1.00 18

16

14

12

10

8

147 6

4

2

0 Move Click Move and Click Move Click Move and Click M. septendecim M. cassini

Figure 5.6. Male responses to actions of model. Males were scored as responding positively if they produced Court II or Court III courtship songs, or if they attempted to mount and copulate with the model. Positive responses marked with shading, negative responses unshaded. Within M. septendecim, the model that moved and clicked was more effective than the model that clicked only (P ² 0.001, Fisher's Exact 2- tailed Test) or the model that moved only (P ² 0.001, Fisher's Exact 2- tailed Test). For M. cassini, the results were similar; the model that moved and clicked was more effective than the model that clicked only (P ² 0.001, Fisher's Exact 2- tailed Test) or the model that moved only (P ² 0.001, Fisher's Exact 2- tailed Test). Responses to click only and move only treatements did not differ in wither species. 6

4 148

2

kHz 0 0 0.5 1.0 1.5 2.0 S * Figure 5.7. Sonogram of response of M. -decim male with Stage I Massospora cicadina infection to an M. -decim call. Male response (marked with asterisk) is a broad-frequency sound similar in timing and acoustical properties to female wing flick signals. Wing flick sound enhanced and extraneous background noise removed for clarity. 10

5 149

0 kHz 0 0.5 1.0 1.5 2.0 S * Figure 5.8. Sonogram of response of M. -cassini male with Stage I Massospora cicadina infection to an artificial M. -cassini call. Male response (marked with asterisk) is a broad-frequency sound similar in timing and acoustical properties to female wing flick signals. Wing flick sound enhanced and extraneous background noise removed for clarity. 0.45

0.40

0.35

0.30

0.25

0.20 Whole call Main portion only Proportion responding 0.15 Slur portion only 0.10 150 0.05

0.00 0 58- 62 63- 77 65- 80

Background intensit (db)

Figure 5.9. Responses of female M. septendecim to playbacks of artificial, pure tone calls and portions of calls against an artificial background chorus. Females were scored as responding positively if they produced wing flick signals in response to playback. No actual chorus of M. septendecim was audible from the test chamber. Intensity of background chorus increased in increments from 0 db to 80 db. Data presented as number of positive responses for a given call and background condition given the total number of trials with those experimental conditions. Whole calls (square markers) were always more likely to elicit responses than either main portion only (circles) or slur portion only (triangles). As background intensity increased, the effectiveness of main portion only decreased, while the effectiveness of slurs alone increased. 4

3 male interference call buzz 2 151 1

kHz 0 0 1 2 3 4 5 S

Figure 5.10. Sonogram of male M. -decim call interference buzz of nearby male. Note how interference buzz overlaps and acoustically obscures calling male's terminal downslur. CHAPTER VI

REPRODUCTIVE CHARACTER DISPLACEMENT AND SPECIATION IN PERIODICAL CICADAS, AND A NEW 13- YEAR SPECIES, MAGICICADA NEOTREDECIM

Abstract

Acoustic mate-attracting signals of related sympatric, synchronic species differ, but those of related allopatric species sometimes do not, suggesting that these signals may evolve to “reinforce” premating species isolation when similar species become sympatric. This model also predicts divergences restricted to regions of sympatry in overlapping species, but such “reproductive character displacement” has rarely been confirmed. We report such a case in the acoustic signals of a previously undetected, new 13-year periodical cicada species, Magicicada neotredecim, here described. Where M. neotredecim overlaps M. tredecim in the central U.S., the dominant chorus pitch (frequency) of M. neotredecim increases from ca. 1.40 to 1.75 kHz, while that of M. tredecim slightly decreases from ca. 1.15 to 1.08 kHz. Female cicadas from sympatry (both species) and allopatry (only M. neotredecim studied) respond best to song pitches similar to those of conspecific males. M. neotredecim differs from 13-year M. tredecim in abdomen coloration, mtDNA, and song pitch, but does not differ consistently from 17-year M. septendecim; thus, like other

152 Magicicada species, M. neotredecim appears most closely related to a geographically- adjacent counterpart with the alternative life cycle. Speciation in Magicicada may often be initiated by life cycle changes that create temporal isolation, and reinforcement could play a novel role by fostering divergence in premating signals prior to speciation. The geographic and phylogenetic relationships of the M. -decim species, and the pattern of reproductive character displacement, suggest two theories of Magicicada life cycle evolution: “nurse brood facilitation” and “canalization of climate-induced shifts.”

Introduction

The periodical cicada species (Magicicada spp.) of eastern North America have extraordinary life histories and phylogenetic relationships (Marlatt 1923; Alexander and Moore 1962; Lloyd and Dybas 1966a, b; Simon 1988; Williams and Simon 1995). Magicicada live underground juvenile lives of either 13 or 17 years, after which they emerge for a brief adult life of about three weeks. Six Magicicada species have been described: In northern and plains states, three morphologically and behaviorally distinct species coexist and emerge together once every 17 years (Fig. 6.1). These species are reproductively isolated in part by distinctive male acoustic signals and female responses (Alexander and Moore 1958, 1962). In the Midwest and South, three similar 13-year species coexist. In this paper, we describe a fourth 13-year species, M. neotredecim. The 13- and 17-year populations are divided regionally into several temporally isolated “broods” that emerge on different schedules. Each 13-year species is probably most closely related to its 17-year counterpart; some of these species pairs can be distinguished only by life cycle (Table 6.1). This pattern suggests that speciation in Magicicada may involve a combination of geographic isolation and life cycle changes that create temporal, or allochronic, isolation (Alexander and Moore 1962; Lloyd and Dybas 1966b; Lloyd and White 1976). Allochronic speciation has been proposed for other organisms (e.g.

153 Alexander 1968; Alexander and Bigelow 1960; Tauber and Tauber 1977a,b), although one case has recently been challenged (Harrison 1979; Harrison and Bogdanowicz 1995).

The male calling songs of sympatric Magicicada species are strikingly different, while those of parapatric life cycle siblings (e.g. 17-year M. cassini and 13-year M. tredecassini) are similar or indistinguishable (Alexander and Moore 1962). This pattern, common in groups with long-range sexual signals, suggests a process in which costly heterospecific sexual interactions lead to selection reinforcing trait differences that promote premating isolation (Dobzhansky 1940; Blair 1955). The hypothesis of reinforcement also predicts greater signal divergence in sympatry, a pattern termed “reproductive character displacement,” when two species’ distributions partially overlap (Brown and Wilson 1956), but this second pattern is only rarely observed (Alexander 1967; Walker 1974; Howard 1993). For example, only one case is known in the singing Orthoptera and Cicadidae (Otte 1989). Our use of these terms, in which reinforcement is a process and reproductive character displacement a resulting pattern, is similar to that of Loftus-Hills and Littlejohn (1992) and Howard (1993), except that in our usage reinforcement can involve selection against wasteful heterospecific sexual interactions of any kind, not merely heterospecific mating as in Howard (1993). Reinforcement of premating isolation is of interest as a mechanism by which barriers to interspecific gene flow are enhanced as a direct result of species interactions in sympatry rather than as incidental outcomes of evolution in allopatry (Dobzhansky 1940; Blair 1955; Howard 1993).

Here we report that the newly described 13-year periodical cicada species, Magicicada neotredecim, shows reproductive character displacement in male song pitch and female song pitch preferences in the central U.S. where it overlaps the 13-year species M. tredecim (Fig. 6.1). M. neotredecim appears to have been derived by a life cycle change from its 17-year counterpart M. septendecim (see also Martin and Simon 1988, 1990; Simon et al. submitted). We propose new hypotheses of life cycle evolution and speciation

154 in Magicicada suggested in part by the geographic and phylogenetic relationships of M. neotredecim to the remaining Magicicada species.

Methods and Results

Background

The Magicicada mating system involves male acoustical signals, female responses, and complex acoustical courtship (Chapter 5). Periodical cicada males aggregate, forming singing choruses attractive to sexually receptive females. Most Magicicada choruses contain three species, a M. –decim1 species that produces a narrow band of sound frequencies with a single dominant pitch between 1 and 2 kHz, and M. -cassini and M. –decula species that each produce broad-spectrum sounds above 3 kHz (Fig. 6.2). While observing 13-year Magicicada in northern Arkansas in 1998, we found choruses with two peak frequencies in the M. -decim range (ca. 1.1 and 1.7 kHz), suggesting a new M. - decim species (Figs. 6.2, 6.3). The pitch differences, easily detectable by ear, are unlikely to result from temperature variations because the song types coexist in the same choruses, because chorus pitch did not vary with air temperature (Fig. 6.4), and because song pitch is not related to tymbal muscle temperature in the closely-related M. septendecim (Young and Josephson 1983). The nature and location of this discovery suggested that the two sympatric M. -decim might correspond to two forms of M. tredecim identified using mitochondrial DNA (mtDNA) and abdomen coloration (Martin and Simon 1988) and found to overlap in a zone reaching from Arkansas to Indiana (Simon et al. submitted). We conducted acoustical, behavioral, and morphological analyses in this region of overlap to test this possibility and to determine the role of song pitch in reproductive isolation of the

1For convenience we refer to Magicicada sibling species groups using the following shorthand: M-decim: M. septendecim (17), M. tredecim (13), and M. neotredecim (13); M -cassini: M. cassini (17) and M. tredecassini (13); M. -decula: M. septendecula (17) and M. tredecula (13).

155 two species. Analysis of geographic song pitch variation led to the discovery of reproductive character displacement in the new species, M. neotredecim.

Documenting Sympatric M. –decim Species with Songs and Morphology

Martin and Simon (1988) identified two forms of 13-year M. tredecim partly on the basis of the amount of orange coloration on the abdominal sternites. To determine if the sympatric M. -decim song types corresponded to these morphs, we measured the calling songs of a random sample of 151 males from a mixed chorus in Sharp Co., AR, and tested for association of song pitch and abdomen color. We collected the males from privately- owned woods on County Rd. 51 approximately 0.25 mi. S. of County Rd. 62 (Lat. 36.256, Lon. -91.462), at a powerline right-of-way, just outside the northwest boundary of the Harold E. Alexander Wildlife Management Area, Sharp Co., AR; we will refer to this location as the “powerline” site.

We recorded male songs using a Sony Professional Walkman cassette recorder and a Sony microphone mounted in a plastic parabola, or a Sony 8mm videocassette recorder with built-in microphone. Because the songs of an individual male do not vary significantly in dominant pitch (Fig. 6.5), we isolated a single call of each individual for spectral analysis. For each recording we generated a power spectrum (plot of sound intensity vs. frequency) and spectrogram (power spectrum vs. time) using Canary 1.1.1 (Cornell Bioacoustics Laboratory) on a Macintosh computer. Individual male M. –decim songs consist of a 1-3 second steady-pitch and nearly pure-tone (sine wave) “main element” followed by a quieter 0.5 second frequency “downslur” that ends approximately 500 kHz lower than the main element pitch (Fig. 6.3; Alexander and Moore 1958; Weber et al. 1987). Because the comparatively minor sound energy in the downslur is distributed across many frequencies, it contributes little to the spectral profile of the call; thus the main element produces the dominant call pitch.

156 We scored the abdomen color of each individually-recorded male using the method of Martin and Simon (1988), in which each male is assigned a value from 1 (50% black) to 4 (all orange). Because male calling song pitches fell into two distinct categories with just one intermediate male out of 151 sampled, while the morphological traits were more continuously distributed, we tested for association between song pitch and coloration by dividing the male sample through an arbitrary intermediate value of 1.5 kHz and examining abdominal color differences between the resulting groups using a Mann-Whitney U.

Results. Within the Sharp Co., AR, powerline site, M. -decim males producing songs with low dominant pitch (mean 1.11 kHz) had the orange abdomen color (Table 6.2; Appendix F) characteristic of Martin and Simon’s (1988) “mtDNA lineage B,” now recognized to be the previously described Magicicada tredecim (Alexander and Moore 1962). The M. -decim males producing higher-pitched songs (mean 1.71 kHz) had the darker abdomen color of Martin and Simon’s “mtDNA lineage A,” and constitute a new species here named Magicicada neotredecim. The mixed chorus produced little sound at pitches intermediate to those of M. tredecim and M. neotredecim (Figs. 6.2, 6.3, 6.6); this and the strong correlation of song and abdomen color suggest that interbreeding is rare (see also Simon et al. submitted). Among approximately 200 cicadas observed during our study, we found just three unambiguous intermediates: One high-pitch male with a yellow abdomen (category 4), one low-pitch male with a darker abdomen (category 2), and one male with an intermediate calling song (1.43 kHz). These could be hybrids or merely phenotypic outliers.

Magicicada neotredecim , new species: Holotype (male): Collected by Cooley and Marshall on 18 May 1998 in Sharp Co., AR., in a powerline right-of-way on County Rd. 51 approximately 0.25 mi. S. of County Rd. 62 (Lat. 36.256, Lon. -91.462), outside the Harold E. Alexander Wildlife Management Area. Dominant pitch of male song phrase 1.68 kHz. Deposited at University of Michigan Museum of Zoology (UMMZ), Ann

157 Arbor, MI with a recording of the male’s call stored on diskette. Allotype: Collected by Cooley and Marshall on 18 May 1998 at same location as holotype. Deposited at UMMZ. For additional description see Table 6.1.

Measuring Female Song Pitch Preferences in Sympatry

Sexually receptive female Magicicada produce “wing-flick” signals in response to male songs (Chapter 5, Cooley and Marshall in prep.). We used this signal as an assay of female mating receptivity to determine if female preference for song pitch was correlated with abdomen color. Using Sound Edit Pro software(MacroMedia), we produced 14 pure- tone model calling phrases differing only in pitch (1.0 kHz to 2.3 kHz, in 0.1 kHz increments) and lacking temporal structure. We modeled artificial songs after field recordings of M. –decim songs (see example using artificial frog calls in Gerhardt 1992). In previous experiments, we have found that females respond similarly to playbacks of recorded and artificial songs. We estimated song pitch preferences of 91 M. -decim females collected from the Sharp Co., AR, powerline site by playing the model songs to individually-marked caged females in both haphazard and ordered sequences using a Macintosh Powerbook computer connected to an amplified portable speaker. Females that did not respond to any songs were dropped from the analysis. We positioned the speaker 25 cm away from the test cage and played the model songs at a volume simulating the intensity of a calling male, as determined by measuring several males’ call intensities with a sound level meter. The playback experiments were carried out between 11:00 and 16:00 in bright overcast or sunny conditions against an acoustic background of a Magicicada chorus containing M. neotredecim, M. tredecim, M. tredecassini and M. tredecula located in woods ca. 8 meters away. For each female, we recorded whether she responded with wing-flick signals and then calculated the weighted average of her responses to the 14 model songs using the following formula:

158 14 å (kHzcall i )(responses to call i ) i =1 (total responses)

We scored female abdomen color using the method described above for males. We tested for association between weighted average pitch preference and abdomen color by dividing the female sample through the same pitch as the male sample (1.5 kHz) and examining the difference in abdomen color between the resulting groups using a Mann- Whitney U.

Results. The playback experiments showed evidence of two classes of females (Fig. 6.6, Appendix G). At the powerline site, females responding maximally to lower pitch songs (M. tredecim) were significantly more orange-colored than females responding maximally to higher pitch songs (M. neotredecim; Table 6.2). About a third (26/91) of the females did not respond to any call. This response rate is similar to that observed in previous work with 17-year M. septendecim females in Virginia and Illinois (Chapter 5; Cooley and Marshall in prep.), where only one M. –decim species is known.

Because the model songs contained no within-phrase temporal structure, differed only in pitch, and accurately distinguished the species, we conclude that song pitch differences may be an important cause of species-specificity in 13-year M. –decim mate recognition. However, natural calls have a pattern of amplitude-modulation that results from individual tymbal pulses (Young and Josephson 1983; Weber et al. 1987); we have not ruled out a role for this temporal structure. Differences in such temporal patterning are probably not the cause of pitch differences in M. -decim cicadas because variations in tymbal contraction rate do not influence dominant pitch (Young and Josephson 1983). We did not analyze temporal call structure because most of our field recordings contained too much background noise, although this did not affect analysis of dominant pitch. Little is

159 known of the relative roles of temporal patterning and frequency content in cicada song in general, although both function in Australian bladder cicadas (Cystosoma; Doolan and Young 1989). In the singing Orthoptera, species-specificity of calling songs usually results from differences in temporal patterning rather than pitch (but see Toms and Otte 1988 for an example in which pitch may be important), while dominant call pitch is important in many Anuran amphibians (Littlejohn 1977; Gerhardt 1988).

Using chorus recordings to estimate species abundance.

For mixed choruses, we developed a method for using the relative intensities of the two species-specific M. –decim dominant chorus pitches to estimate relative proportions of the species. This approach assumes that (1) M. neotredecim and M. tredecim are uniformly distributed at a given site and (2) both species show the same relationship between male abundance and chorus intensity. To test the first assumption, we recorded a continuous chorus sample along a 200m woodside trail in the Harold E. Alexander Wildlife Management Area, Sharp Co., AR, while pointing the parabola/ microphone assembly into the treetops along one side at a 45° angle. We recorded one side while walking in one direction and then recorded the other side while returning. From portions of this recording taken at seven meter intervals, we generated power spectra and measured the intensities of the M. neotredecim and M. tredecim frequency bands; this yielded 47 samples because of gaps in the forest on one side (350m total). If the M. neotredecim and M. tredecim at the site were not uniformly distributed with respect to one another on a local scale, we predicted that this series of samples would show significant variation among locations in the relative intensities of the two species’ chorus bands. To test the second assumption, we compared the distribution of call pitches of a random sample of males collected from the mixed-species Sharp Co., AR chorus to the distribution of song frequencies in the chorus power spectrum, using a Kolmogorov-Smirnov test to determine if the shapes of the two distributions differed significantly.

160 Results. Although the proportions of the two M. –decim species vary on a scale of miles (e.g. Fig. 6.7 insets), the species do not appear to assort significantly within a location. In the 350m continuous recording, the proportion of M. -decim chorus sound produced by the rarer species (M. neotredecim) remained between 10% and 36% (mean = 19.0%, SD = 6.0, n = 47), and the chorus intensities of the two species were not significantly negatively correlated (Pearson coefficient = -0.229, P= 0.121) as would be expected if the species formed exclusive aggregations. Therefore a recording taken at a given position is likely to give a reasonable estimate of the local chorus sound. This result matches our subjective impressions; we never detected by ear variations in the M. -decim chorus sound within a location. If males of the two species sing with similar consistency, intensity, and call duration, the sound intensity at a given species-specific pitch should be related to the local density of males producing that pitch. The random sample of the Sharp Co., AR powerline population verified this relationship: The standardized histogram of call pitches of individually-recorded males was indistinguishable from the standardized quadratic chorus power spectrum (Kolmogorov-Smirnov test, P > 0.05.; Fig. 6.6). Therefore the power spectrum of a mixed chorus can be used to estimate the relative proportions of M. tredecim and M. neotredecim present.

Although the downslurs of the higher-pitched M. neotredecim songs sometimes overlap the higher-frequency portions of the lower-pitched M. tredecim songs, downslurs from even a dense M. neotredecim chorus do not obscure sparse populations of M. tredecim because M. -decim downslurs are brief and their sound energy is distributed across a broad band of frequencies. M. tredecim at the Sharp Co., AR, powerline site produced a distinct peak in the chorus power spectrum (Fig. 6.6) despite constituting just 8% of the study population. However, extremely rare M. tredecim could be overlooked in a dense M. neotredecim population; we attempted to minimize this possibility by recording several minutes at each location and by sweeping the parabola in different directions. In contrast, rare M. neotredecim are always easily detectable in a M. tredecim chorus because

161 M. tredecim males do not produce sounds at frequencies as high as those of M. neotredecim.

Estimating Species Distributions using Geographic Variation in Songs

Once we had demonstrated that the two 13-year M. -decim forms are species differing in song pitch and that the species’ relative abundance can be estimated from chorus recordings, we analyzed the species’ distributions and geographic variation in song pitch using recordings taken from 80 locations distributed throughout the 1998 Magicicada emergence. Periodical cicadas in different regions emerge in different years, and the cicadas of a region are referred to as a “brood” (Marlatt 1923; see individual brood maps in Simon 1988). The 13-year cicadas emerging in 1998 belonged to the largest 13-year brood, Brood XIX, which reaches from Maryland to Oklahoma. The widespread range of this brood allowed a thorough characterization of 13-year M. -decim calling song biogeography. The recording dates were as follows: Sharp, Fulton and Lawrence Co., AR, 12-25 May; remaining Arkansas sites, 28 May; Missouri sites 1-7 June; Randolph, Monroe, Jersey, Sangamon and Piatt Co., IL 29-30 May; remaining Illinois sites 9-14 June; remaining central and southeastern state sites 31 May - 3 June; Maryland sites 29 May - 1 June.

Results. We found M. neotredecim in Missouri, Illinois, western Kentucky, and northern Arkansas (Fig. 6.7, Appendix H; see also Simon et al. submitted). The southernmost M. neotredecim populations overlap M. tredecim in a zone 50-150 km wide reaching from northern Arkansas into southern Missouri, southern Illinois, and eastern Kentucky. The remainder of Brood XIX contains M. tredecim and not M. neotredecim. Simon et al. (submitted) document a similar pattern of overlap using abdomen coloration and mtDNA and describe additional mixed populations in southern Indiana.

162 Geographic variation in dominant chorus pitch of M. neotredecim shows a pattern of reproductive character displacement, a predicted outcome of reinforcement of premating isolation (Dobzhansky 1940; Blair 1955; Howard 1993). M. neotredecim males have the highest song pitch (ca. 1.7 kHz) where they coexist with M. tredecim. North of the overlap zone, M. neotredecim dominant song pitch decreases to approximately 1.4 kHz in Illinois and 1.5 kHz in Missouri (Fig. 6.8, 6.9). Most of the song change occurs over a short distance immediately north of the zone of sympatry, and little variation occurs elsewhere in allopatry. Only song pitch varies across the range of M. neotredecim; abdomen color and mtDNA haplotype are comparatively uniform (Martin and Simon 1988, 1990).

Song pitch variation in M. tredecim is more subtle, only about 1/4th of that observed in M. neotredecim. M. tredecim populations in deep sympatry with M. neotredecim have very low-pitched songs, and M. tredecim peak song pitch slightly increases south and east of the overlap zone from ca. 1.11 to 1.15 kHz (Figs. 6.9, 6.10). However, the significance of this pattern is not clear because some allopatric M. tredecim choruses in the southeast have songs as low-pitched as those in the overlap zone.

While most of the chorus samples included the songs of thousands of males, many of the populations from Missouri and Alabama were recorded late in the emergence when relatively few males remained. For these locations the chances of overlooking a rare type were greater, although it seems unlikely that rare M. tredecim could have been missed in each of the many samples from central and northern Missouri.

Additional Tests of Reproductive Character Displacement

To determine if female song pitch preferences change with male song pitch in M. neotredecim, we measured weighted average song pitch preferences of 33 Magicicada neotredecim females collected from a woodlot off Rt. 1300E 0.8 miles south of White

163 Heath, IL (Piatt Co.), beyond the northern limit of M. tredecim. We completed the playback experiments at Lodge Park Co. Forest Preserve with a nearby background chorus containing M. neotredecim, M. tredecassini, and M. tredecula. We used a Wilcoxon signed-ranks test to determine if the average pitch preference of the Piatt Co. females differed from that of the Sharp Co., AR females. Because of time constraints, we were unable to study allopatric M. tredecim females.

To test the hypothesis that song pitch variations could be related to geographic variation in body size, we compared the song pitches and body sizes of M. neotredecim and M. tredecim males from sympatry at the powerline site with those of 17 M. neotredecim males collected in Allerton Park, Piatt Co., IL, where no M. tredecim are present. We used three characters to approximate size: right wing length, thorax width between the wing articulations, and first abdominal sternite width between the sutures that join it to the first abdominal tergite. We conducted pairwise comparisons among populations using Mann-Whitney-U tests. For each population, we tested for an association between size and song pitch using linear regressions.

Results. Female M. neotredecim preferences change with male song pitch: In sympatry with M. tredecim, (Sharp Co., AR) female M. neotredecim were most responsive to an average pitch of 1.75 kHz (n = 35), while in allopatry (Piatt Co., IL) female preference averaged 1.31 kHz (n = 13; Wilcoxon signed ranks test, P < 0.002).

Size differences do not explain patterns in calling song variation because (1) M. tredecim and M. neotredecim in sympatry overlapped broadly in size but not in song pitch, (2) M. neotredecim populations from Illinois and Arkansas differed in song pitch but not in size, and (3) we found no significant correlation between song pitch and any measure of body size within species in any population (linear regression; Fig. 6.11).

164 Discussion

Reinforcement and Reproductive Character Displacement in M. neotredecim

Reinforcement is often viewed as a model of species formation (e.g. Butlin 1987, 1995; Liou and Price 1994; Kelly and Noor 1996), the only such model in which prezygotic isolation evolves as a direct outcome of natural selection against hybridization. In keeping with this view, Butlin (1987, 1989) suggests that the term “reinforcement” should apply only to cases in which assortative mating is enhanced despite gene flow and that the term “reproductive character displacement” should be used to refer to the strengthening of prezygotic isolation when hybrids are sterile, because speciation is already completed if gene flow is not possible. However, there are potential problems with this distinction (see also Howard 1993). If the fundamental criterion for species status is evolutionary independence (de Queiroz 1998), or low probability of future amalgamation, then populations that have acquired sufficient postzygotic isolation to undergo reinforcement are species even before they come into contact; the occurrence of reinforcement is itself evidence of evolutionary independence. Furthermore, abundant evidence of natural hybridization (Arnold 1997; Arnold and Emms 1998) shows that species should not be identified, nor the completeness of speciation determined, by degree of hybrid sterility. Practical difficulties with Butlin’s definitions arise because uncertainty about past gene flow prevents the application of either term; for example there is little or no evidence that M. neotredecim and M. tredecim currently hybridize, but this does not necessarily mean that gene flow did not occur in the past. Consequently, we use the term reinforcement to refer to a general process defined without respect to gene flow, speciation, or relatedness of interactants, and we use the term reproductive character displacement to identify the resulting pattern.

165 The process of reinforcement is the elaboration of differences between species’ sexual signals in response to selection against heterospecific sexual interactions (Grant 1972, Waage 1979, Howard 1993). If newly established contact between species involves only slight range overlap, differences resulting from reinforcement may be manifest in the form of clines, in which the two species are most extremely different in sympatry and less distinct in allopatry (Brown and Wilson 1956). The clines reflect the relative contributions of conspecific and heterospecific interactions to local optima of sexually-selected characters. Individuals within the zone of contact face net directional selection for distinctiveness from the other species, while conspecifics away from the zone of contact face net selection for compatibility with their own species, with no benefit, and perhaps some costs, to extreme individuals. Thus, in species undergoing reinforcement of premating isolation, differences in species-specific sexual traits will become less extreme away from the zone of contact where individuals face decreased pressures for distinction from heterospecifics and increased selection for compatibility with conspecifics having no history of contact with the other species involved.

A variety of observations support the hypothesis that the high song pitch of M. neotredecim in sympatry with M. tredecim results from reinforcement with M. tredecim. First, M. neotredecim in or adjacent to the area of overlap all have high-pitch songs (>1.6 kHz), the highest M. -decim dominant chorus pitch observed anywhere, while allopatric M. neotredecim never do. Second, the dominant chorus pitches of M. tredecim and M. neotredecim remain comparatively stable in allopatry, as does that of 17-year M. septendecim, which varies only about 0.15 kHz in archived UMMZ recordings taken throughout its widespread range (Table 6.3). Third, the typical song pitch of 17-year M. septendecim, the likely sister group of M. neotredecim (see below), is indistinguishable from that of allopatric M. neotredecim in Illinois (ca. 1.4 kHz); these species are also indistinguishable in behavior, morphology, and mtDNA (Alexander and Moore 1962; Simon et al. submitted). Fourth, song pitch apparently plays an important role in mate

166 recognition of both species. Fifth, M. neotredecim and M. tredecim songs in sympatry differ just enough for average songs of each species to avoid frequency overlap, with M. neotredecim downslurs ending at about the dominant song pitch of M. tredecim (Fig. 6.3). Finally, the change in song pitch does not appear to be an incidental effect of a latitudinal cline in body size: M. neotredecim populations in Arkansas and Illinois produce different song pitches yet are indistinguishable in size, and song pitch does not correlate significantly with body size within any population.

A potential challenge to the hypothesis of reinforcement in M. neotredecim arises from the existence of some Missouri populations apparently well outside the range of M. tredecim with a partially elevated dominant chorus pitch (ca. 1.5 kHz). This pattern could be explained by the reinforcement model if (1) M. neotredecim colonized Missouri from populations in Illinois that were themselves adjacent to the range of M. tredecim, or (2) undiscovered M. tredecim populations exist in Missouri near the locations we sampled. Future surveys should investigate the latter possibility. Because M. tredecim appears to reach its northern limits on Mississippi River and Wabash River lowlands, it may be found only in the vicinity of rivers elsewhere in the northern part of its range, increasing its chances of being overlooked.

Also of interest is that the dominant chorus pitch of M. neotredecim does not appear to be related to the relative abundance of the two 13-year M. -decim species in mixed populations (linear regression, r2 = 0.053, P £ 0.317, n = 21); under certain conditions the reinforcement model predicts this correlation because the strength of reinforcing selection on one species should depend on the abundance of the other (Howard 1993; Noor 1995). However, there are several possible reasons for this correlation to be absent even under the reinforcement model. First, most M. neotredecim in populations sympatric with M. tredecim have calls that are displaced on average just high enough to avoid frequency overlap, suggesting that reinforcing selection ceases after the M. neotredecim song

167 phenotype reaches that point. If only a minor representation by M. tredecim is necessary to drive this change in M. neotredecim, then a correlation between relative abundance and degree of displacement would be detectable only in populations with very rare M. tredecim; unfortunately we recorded few such populations. Second, the anticipated correlation between relative abundance (in sympatry) and degree of displacement would be weakened if mating signals are usually subject to stabilizing selection when not undergoing reinforcement (e.g. Paterson 1993; Alexander et al. 1997); such selection would maintain song displacement upon later changes in relative species abundance. This hypothesis could be falsified in this case by finding evidence of lowered M. neotredecim song pitch where M. tredecim is very rare and the weakly displaced M. neotredecim are unlikely to be immediately descended from populations with undisplaced songs. Third, rapid, continuous fluctuations in relative species abundance would also tend to weaken any correlation between abundance and displacement.

A notable feature of the pattern of song displacement in Magicicada is its asymmetry, although this phenomenon is not unusual in cases of reproductive character displacement (e.g. Littlejohn 1965; Littlejohn and Loftus-Hills 1968; Fouquette 1975; Waage 1979; Noor 1995). In general, because the strength of selection on each species depends on factors that can differ between the species, symmetrical displacement is probably unlikely (Grant 1972; Howard 1993). The selection each species imposes on the other will be a function of (1) the benefits realizable from hybrid sexual interactions (if hybrid offspring are fertile, the cost of hybridization may differ depending on which species is which parent), and (2) the relative abundance of each species (the likelihood that any individual will ever have a cross-species sexual encounter and face the danger of hybridization; but see above). The selection for divergence imposed by heterospecifics (through interspecific gene flow or the threat of it) should be stronger, and the stabilizing selection imposed by conspecifics (through intraspecific gene flow) should be weaker for the rarer of the species. Possible explanations for asymmetrical displacement in Magicicada

168 include: (1) greater numerical abundance of M. tredecim relative to M. neotredecim during critical stages of the interaction, (2) greater M. tredecim female selectiveness upon initial contact, and (3) greater constraints on evolution of lower song pitch.

The pattern of character displacement in M. neotredecim is similar to an example from Gastrophryne (Anura: Blair 1955; Loftus-Hills and Littlejohn 1992) in east Texas. Like M. neotredecim and M. tredecim, Gastrophryne carolinensis and G. olivacea in sympatry are distinguishable by only average differences in morphological characters and show the greatest nonoverlapping differences in dominant song pitch, a trait relevant to species-specific mate recognition. As in the Magicicada example, evidence of hybridization is weak and allopatric populations of the species show nearly or completely nonoverlapping differences, a key observation suggesting that male calls were distinct, and mating errors rare, even prior to sympatry. Like Loftus-Hills and Littlejohn (1992), we note that reinforcement may occur even in the absence of interspecific mating as long as enhancement of differences in male song or female preference increases the efficiency of advertisement or mate-location.

The rarity of reproductive character displacement is a striking paradox (Alexander 1967; Walker 1974; Howard 1993; Alexander et al. 1997) given the ubiquity of song distinctiveness of sympatric species and given other evidence of reinforcement (Coyne and Orr 1989, 1997). Reproductive character displacement has been found in only one other group of singing insects, crickets of the genus Laupala in Hawaii (Otte 1989). The paucity of convincing examples has been interpreted differently by various authors; some point to a lack of adequately studied cases (Walker 1974; Howard 1993), while others suggest that sexual signal evolution may be driven mainly by within-species processes (West-Eberhard 1983; Paterson 1993). A somewhat neglected explanation for the paradox is that character displacement may be quickly obscured by species range changes that disrupt the association of song divergence with range overlap (Waage 1979): As discussed above, song changes

169 due to reinforcement might remain even after species’ ranges cease to overlap if mate- attracting signals and female preferences for them are usually subject to stabilizing selection (Paterson 1993; Alexander et al. 1997). Patterns of character displacement will also be lost if sympatry becomes complete or if allopatric populations with the pre-contact song types are lost.

Another possible explanation of the paradox is that, when male signals differ prior to the establishment of sympatry, reinforcement may involve changes only in the stringency of female mate acceptance criteria (Waage 1979; Gerhardt 1994; Kelly and Noor 1996) as long as the costs of having a novel, minority receiver phenotype are offset by the advantages of avoiding hybrid sexual interactions (Butlin and Ritchie 1994; see Halliday 1983). Such cases might tend to remain undiscovered because of an overemphasis on male signals. This explanation is based in part on the premise that reinforcing selection should occur initially and/or most strongly on females, who usually have the most to lose from mating errors (Alexander et al. 1997). However, reinforcement of female traits alone should be rare because any change in female behavior will tend to cause concerted evolution in male signals: If selection on male signals derives mainly from efficiency in mate attraction, males benefit from adaptations that increase their likelihood of securing a female response, and male signals would evolve in concert with changes in female preferences. Even if females in sympatry evolve only greater stringency of preferences, with no change in the mean preference, variation among males in sympatry should be reduced (relative to males in allopatry) as a result of stronger stabilizing selection.

Life Cycle Differences and Species Status

Designating M. neotredecim and 17-year M. septendecim as separate species may provoke debate because many populations of M. neotredecim differ from M. septendecim only in life cycle length and in geography (e.g. Lloyd 1984). However, if based upon genetic differences, life cycle differences may provide genetic isolation and facilitate

170 evolutionary independence of population lineages, the fundamental criterion for species status (Alexander and Moore 1962; de Queiroz 1998). The following evidence indicates that Magicicada life cycle differences have a genetic basis: (1) in transplant experiments, 13-year cicadas reared north of 17-year cicadas emerged after a 13-year juvenile period (Lloyd 1987); (2) the border between 13- and 17-year species appears historically stable (but see Lloyd et al. 1983); and (3) the failure of small numbers of cicadas emerging off- schedule to satiate avian predators and establish new populations (Marlatt 1923; Dybas 1969; Williams and Simon 1995) suggests persistent selection favoring additional genes canalizing the majority life cycle phenotype. Life cycle differences substantially restrict opportunities for gene flow, because M. neotredecim and M. septendecim emerge simultaneously just once every 221 years. Moreover, during such co-emergences, opportunities for interbreeding are limited further by the minimal geographic overlap between 17- and 13-year populations and the fragmented nature of Midwestern populations (Alexander et al. in prep.). Although some authors have suggested that 13- and 17-year cicadas are widely sympatric (Marlatt 1923; Cox and Carlton 1991), more recent studies (e.g. Lloyd et al. 1983) have found only limited sympatry. Alexander et al. (in prep.) found that 17-year Brood III and 13-year Brood XIX overlap in Illinois in limited regions, by no more than a few kilometers, and are most often separated by prairie gaps. Such limited opportunities for gene flow are unlikely to prevent divergence of M. neotredecim and M. septendecim, especially if differences in life cycle and geography impose divergent selective pressures (Endler 1977; Rice and Hostert 1993). Absence of known life cycle intermediates where M. neotredecim and M. septendecim meet and the stable position of their contact zone suggest that gene flow is sufficiently restricted to permit divergence.

Simon et al. (submitted) discuss life cycle changes in Magicicada and note that this process could bring M. neotredecim populations back into temporal synchrony and panmixis with M. septendecim. Although life-cycle switching from M. septendecim is a credible explanation for the origin of M. neotredecim, amalgamation of M. neotredecim and

171 M. septendecim by this process appears unlikely because (1) permanent life cycle changes are apparently rare, (2) a random switching event has a low probability of establishing synchrony between M. neotredecim and M. septendecim, and (3) a switching event that synchronizes only part of M. neotredecim with M. septendecim will not affect the remaining M. neotredecim.

The Origin of Magicicada neotredecim and Speciation in Periodical Cicadas

Cladistic analysis of M. –decim phylogeny and determination of ancestral life cycles is not yet possible because the informative traits (morphological and molecular) within the group are not shared with other Magicicada species or with other cicada genera; the same problem obscures relationships among the Magicicada sibling groups. However, because M. neotredecim and M. septendecim are in many locations distinguishable only in life cycle and geography, and share the same mtDNA haplotype (Martin and Simon 1988, 1990), one of these two species probably originated from ancestral populations of the other. Together, these apparent cognates (descendants of an exclusive common ancestor or speciation event) differ from M. tredecim in song pitch, abdomen coloration, and in mtDNA by an estimated 2.6% of nucleotides, suggesting that their common ancestor diverged from M. tredecim 1-2 million years ago (Martin and Simon 1990).

Biogeographic evidence suggests that populations of the M. septendecim lineage gave rise to M. neotredecim, rather than vice versa, probably in the time since the last glacial maximum. First, the comparatively restricted distribution of M. neotredecim, surrounded on the west, north, and east by M. septendecim, suggests that M. neotredecim may be the newer lineage; otherwise rapid and implausible range changes would be required to explain the widespread distribution of M. septendecim. Second, the pattern of M. neotredecim song variation, in which most M. neotredecim have undisplaced songs, probably could not have persisted through range changes associated with Pleistocene glacial cycles. As a species’ range shifts southward during glacial cooling, the populations

172 surviving in refuges are most likely founded from the southern margin of the earlier distribution, while the most northern populations simply go extinct (Hewitt 1996). Had M. neotredecim ever experienced such a range shift, all undisplaced M. neotredecim would likely have been lost, because (1) M. neotredecim is completely bounded on the south by M. tredecim, (2) southern M. neotredecim today all have displaced song pitches as a result of overlap with M. tredecim, and (3) all northern, undisplaced M. neotredecim are found in a region that 18,000 years ago may have contained spruce-fir forest and cooler, drier conditions presently found only north of the range of Magicicada (Webb et al. 1993). Thus, the pattern of reproductive character displacement, together with the similarity of northern M. neotredecim and 17-year M. septendecim calling songs, is consistent with the hypothesis that M. neotredecim originated recently via a permanent life cycle change from a 17-year ancestor (Martin and Simon 1988, 1990; Simon et al. submitted), although the reverse hypothesis (17-year M. -decim evolving from Midwestern 13-year M. -decim) cannot yet be ruled out.

Abdomen color and mtDNA data (Simon et al. submitted) indicate that the other large 13-year brood (Brood XXIII), which inhabits most of the lower Mississippi Valley, contains both M. neotredecim and M. tredecim, with M. neotredecim limited to northern populations and partially overlapping M. tredecim. Archived UMMZ recordings from this brood suggest that M. neotredecim has a displaced song pitch (ca. 1.7 kHz) where the two species overlap; unfortunately we do not have recordings from any of the few populations of pure M. neotredecim in Brood XXIII. Archived tapes further suggest that northern populations of the remaining 13-year brood (Brood XXII), found only in southern Mississippi and Louisiana, contain M. tredecim alone (dominant chorus pitch ca. 1.08 kHz; Warren Co., MS, and Catahoula Parish, LA), although additional sampling will be necessary to confirm the apparent absence of M. neotredecim from the brood. Simon et al. (submitted) note that the presence of M. neotredecim or similar cicadas in Brood XXIII could be explained in three ways: (1) by a second derivation from M. septendecim, (2) by a

173 single origin of Brood XIX M. neotredecim followed by the temporal migration of four- year delayed M. neotredecim stragglers to Brood XXIII, or (3) by the formation of Brood XXIII in part from four year delayed Brood XIX populations containing both M. -decim species. Evidence that the developmental plasticity required by hypotheses #2 and #3 occurs in Magicicada includes observations of off-schedule cicadas emerging in numbers difficult to explain by mutation (e.g. Dybas 1969; White and Lloyd 1979; Kritsky and Simon 1996) and by the existence of temporally isolated broods of the same life cycle, presumably formed as a result of short-lived regional climatic events that caused most cicadas to delay or accelerate their emergence once (Alexander and Moore 1962; but see Moore 1993).

Life cycle changes and speciation may be linked in Magicicada (see also Simon et al. submitted). Each species appears most closely related to a cognate with the alternative life cycle; it is therefore inescapable that 13- and 17-year life cycles have evolved repeatedly in the group (Alexander and Moore 1962). The discovery that 17-year cicadas differ from 13-year cicadas in having a 4-year dormancy period (White and Lloyd 1975, Lloyd and White 1976) increases the plausibility of the argument that Magicicada speciation often involves permanent four year life cycle changes, if such a dormancy period could be induced or deleted by few genetic changes. However, identifying plausible evolutionary mechanisms for such life cycle evolution is a challenge because Magicicada populations apparently must number many thousands per hectare to avoid being destroyed by avian predators (Marlatt 1923; Beamer 1928; Alexander and Moore 1962; Dybas 1969; Williams et al. 1993). In considering possible scenarios for the origin of M. neotredecim, we have developed two models of life cycle evolution that may have general relevance for Magicicada speciation, “nurse brood facilitation” and “canalization of climate-induced plasticity.” These models are not exclusive alternatives; rather, the models are intended to convey different ways in which the unique life-history attributes of Magicicada could be involved in species formation.

174 “Nurse brood” facilitation and life cycle changes: Recent confirmation of microgeographic overlap between 13- and 17-year broods in the Midwest (Alexander et al. unpublished data; Lloyd et al. 1983), along with repeated observations of straggler cicadas, suggests a mechanism in which speciation is initiated by temporal migration of life cycle mutants from one brood to an overlapping brood of a different life cycle (Fig. 6.12). For example, if mutations affecting life cycle length sometimes cause 17-year cicadas to emerge with 13-year life cycles, such mutants could escape predation and establish an incipient species (1) if they emerged in the life cycle overlap zone during an emergence of the sympatric 13-year brood and (2) if the sympatric 13-year brood contained only dissimilar 13-year species in the area of overlap. Absence of a confusingly similar species in the 13- year brood would probably be necessary to avoid loss of the temporal immigrants to interspecific hybridization. The existing brood functions as a “nurse brood,” providing predator protection for rare temporal mutants (Alexander amd Moore 1962); this process has also been termed “induction” of one species into a brood lacking that species (Lloyd and Dybas 1966b). One potential difficulty with this hypothesis is that multiple life cycle mutants must appear in the same generation and location, but this problem may be mitigated by the large size of periodical cicada populations, which can reach 3.7 million per hectare (Dybas 1969), and by the possibility that minor genetic changes may influence life cycle length (see above).

Nurse brood facilitation of life cycle mutants should lead to the formation of a new incipient species whenever a given species in one brood does not already have a counterpart in an overlapping brood of a different life cycle; this suggests an explanation for the surprising tendency of Magicicada to repeatedly evolve the same life cycles (Alexander and Moore 1962). In addition, reinforcement of premating isolation among sympatric species of the same life cycle could interact with the nurse brood mechanism to produce new species pairs. For example, reinforcement in M. neotredecim has incidentally caused some populations of this species to differ from M. septendecim in dominant chorus pitch by at

175 least 300 Hz. If future 17-year life cycle mutants from song-displaced M. neotredecim were to co-emerge with 17-year cicadas in the life cycle overlap zone (such as where 17- year Brood X adjoins 13- year Brood XIX), these preexisting song pitch differences might facilitate assortative mating and establishment of a new 17-year species (Fig. 6.13). The general processes of life cycle change, nurse brood facilitation, and reinforcement could explain patterns in Magicicada speciation including (1) pairs of cognate species differing primarily in life cycle length and (2) initial synchronization of different species of the same life cycle.

The hypothesis of nurse brood facilitation has the potential to explain the origin of M. neotredecim from M. septendecim and its synchronization with the 13-year Brood XIX (Fig. 6.14). However, for this hypothesis to be correct, a large region of Midwestern Brood XIX must have contained only M. tredecassini and/or M. tredecula prior to the arrival of M. neotredecim; otherwise most Midwestern M. neotredecim today would have displaced (high-pitched) calling songs from past ecological interaction with M. tredecim. Such a large region of prior range discordance between the 13-year species seems unlikely because their distributions today are coincident throughout the rest of the 13-year range; however such discordance does exist in 17-year cicadas in Oklahoma where the 17-year brood (IV) contains only M. cassini (Dybas and Lloyd 1974). This hypothesis for the origin of M. neotredecim predicts that Midwestern populations of M. tredecassini and/or M. tredecula, which have not yet been adequately examined, are not more closely related to their 17-year counterparts than to southeastern populations.

Canalization of climate-induced life cycle shift: The nurse brood mechanism can explain only instances of life cycle evolution that occurred after broods of both life cycles were already present. Unless 13- and 17-year life cycles appeared prior to the evolution of periodicity and the need for predator satiation, additional mechanisms must be considered. Without the nurse brood mechanism, life cycle change seems unlikely to be initiated by

176 mutation because life cycle mutants are unlikely to appear simultaneously in numbers sufficient to satiate predators. We suggest a second model of Magicicada life cycle evolution based on the assumption (discussed above) that unusual climatic stimuli can trigger the expression of life cycle plasticity.

Developmental plasticity could facilitate life cycle evolution and speciation in Magicicada if extreme climatic conditions sometimes induce periodical cicadas to switch life cycle in numbers sufficient to satiate predators and if the conditions persist for generations. Persistence of such extreme conditions would promote continued expression of the new cycle, leading to selection that would (1) remove genes tending to cause expression of the old phenotype and (2) favor alleles canalizing the new phenotype. Furthermore, if the climate gradually returned to the initial conditions, canalizing selection could lead to the evolution of cicadas that expressed the new life cycle even under the original conditions (Fig. 6.15; Waddington 1953). The process would succeed only if the climate change was substantial, sudden, and persistent. A sudden shift of lesser magnitude could create overlapping broods or lead to local extinction if no populations remained large enough to satiate predators. Gradual climate change could lead to canalization of the original cycle and reduce the probability of a life cycle switch. A sudden change followed by an equally sudden return to the original conditions would lead only to the formation of a temporally- isolated brood of the original life cycle. A general treatment of the role of phenotypic plasticity in facilitating speciation can be found in West-Eberhard (1989).

If different Magicicada species possess similar life cycle plasticity, a climate shift causing a change in the life cycle of one species will likely change sympatric species in synchrony. In this manner, simultaneous, parallel speciation events could occur in Magicicada (Figs. 6.13, 6.14). The hypothesis that different Magicicada species share similar developmental plasticity is supported by the observation that both species (M.

177 septendecim and M. cassini) present in Brood XIII near Chicago produced millions of four-year premature stragglers in the same year, 1969 (Lloyd 1984).

If M. neotredecim was formed from populations of M. septendecim by canalization of a climate-induced life cycle, we might expect to find that the 13-year M. -cassini and M. -decula sympatric with M. neotredecim are also more closely related to their 17-year counterparts, because their ancestors would have been subjected to the same climate fluctuations. There should also be paleoclimatological or other data consistent with recent, dramatic climate change in Missouri and Illinois. The M. neotredecim / M. tredecim overlap zone is concordant with the hypothesized southern boundary of the Prairie Peninsula (Cox and Carlton 1991); perhaps prior to the climate change the border between 13- and 17-year cicadas was located there, in which case prior discordance between the 13- year species’ ranges is not necessary to explain the undisplaced calling song of northern M. neotredecim. One weakness of the canalization hypothesis is that it offers no explanation other than chance for the synchronization of the new Midwestern 13-year cicadas with the remainder of Brood XIX. Simon et al. (submitted) propose a related hypothesis in which successive life cycle changes occur, each one farther north, and each facilitated by the existence of the adjacent Brood XIX by a mechanism similar to the nurse brood process discussed above.

178 Literature Cited

Alexander, R. D. 1967. Acoustical communication in arthropods. Ann. Rev. Entomol. 12: 495- 526. Alexander, R. D. 1968. Life cycle origins, speciation, and related phenomena in crickets. Q. Rev. Biol. 43: 1-41. Alexander, R. D, and R. S. Bigelow. 1960. Allochronic speciation in field crickets, and a new species, Acheta veletis. Evolution 14: 334- 346. Alexander, R. D., and T. E. Moore. 1958. Studies on the acoustical behavior of seventeen- year cicadas (Homoptera: Cicadidae: Magicicada). Ohio J. Sci . 58: 107- 127. Alexander, R. D., and T. E. Moore. 1962. The evolutionary relationships of 17-year and 13-year cicadas, and three new species. (Homoptera: Cicadidae, Magicicada). Misc. Publ. no. 121. Univ. of Michigan Museum of Zoology, Ann Arbor, MI . Alexander, R. D., D. C. Marshall, and J. R. Cooley. 1997. Evolutionary Perspectives on Insect Mating. Pp. 4-31 in: B. Crespi and J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Alexander, R. D., A. F. Richards, D. C. Marshall, and J. R. Cooley. (In prep). The microdistributional relationships of 17- and 13-year cicadas in Illinois (Magicicada): Broods III, 1997, and XIX, 1998. Arnold, M. 1997. Natural hybridization and evolution . (Oxford University Press: New York). Arnold, M. L., and S. K. Emms. 1998. Paradigm lost: Natural hybridization and evolutionary innovations. Pp. 379-389 in: D. J. Howard and S. H. Berlocher, eds. Endless Forms: Species and Speciation . (Oxford University Press: New York). Beamer, R. H. 1928. Studies on the biology of Kansas Cicadidae. Univ. Kansas Sci. Bull . 18: 155- 263. Blair, W. F. 1955. Mating call and stage of speciation in the Microhyla olivacea- M. carolinensis complex. Evolution 9: 469-4 80. Brown, W. L., and E. O. Wilson. 1956. Character displacement. Syst. Zool. 5: 49- 64. Butlin, R. K. 1987. Speciation by reinforcement. TREE 2: 8- 1 3. Butlin, R. K. 1989. Reinforcement of premating isolation. Pp. 158-179 in D. Otte and J. A. Endler, eds. Speciation and its consequences . (Sinauer: Sunderland, MA). Butlin, R. K. 1995. Reinforcement: an idea evolving. TREE 10: 432- 434.

179 Butlin, R. K., and M. G. Ritchie. 1994. Behaviour and speciation. In: P. J. B. Slater, and T. R. Halliday, eds. Behaviour and Evolution. (Cambridge University Press). Cooley, J. R., and D. C. Marshall. (In prep). Sexual signaling and mating behavior in periodical cicadas (Magicicada spp.). Cox, R. T., and C. E. Carlton. 1991. Evidence of genetic dominance of the 13-year life cycle in periodical cicadas (Homoptera: Cicadidae: Magicicada spp.). Am. Midl. Nat . 125: 63- 74. Coyne, J. A., and H. A. Orr. 1989. Patterns of speciation in Drosophila. Evolution 43: 362- 381. Coyne, J. A., and H. A. Orr. 1997. “Patterns of speciation in Drosophila” revisited. Evolution 51: 295- 303. de Queiroz, K. 1998. The general lineage concept of species, species criteria, and the process of speciation. Pp. 57- 79 in: D. J. Howard and S. H. Berlocher, eds. Endless Forms: Species and Speciation . (Oxford University Press: New York). Dobzhansky, T. 1940. Speciation as a stage in evolutionary divergence. Am. Nat . 74: 312- 321. Doolan, J. M., and D. Young. 1989. Relative importance of song parameters during flight phonotaxis and courtship in the bladder cicada Cystosoma saundersii. J. Exp. Biol. 141: 113- 131. Dybas, H. S. 1969. The 17-year cicada: A four-year mistake? Bull. Field Mus. Nat. Hist. 40: 10- 12. Dybas, H. S., and M. Lloyd. 1974. The habitats of 17-year periodical cicadas (Homoptera: Cicadidae: Magicicada spp.). Ecol. Monogr. 44: 279- 324. Endler, J. A. 1977. Geographic variation, speciation and clines . (Princeton University Press: Princeton, NJ). Fouquette, M. J. 1975. Speciation in chorus frogs. I. Reproductive character displacement in the Pseudacris nigrita complex. Syst. Zool . 24: 16- 22. Gerhardt, H. C. 1988. Acoustic properties used in call recognition by frogs and toads. Pp. 455-483 in B. Fritsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak, eds. The evolution of the amphibian auditory system . (Wiley: New York). Gerhardt, H. C.. 1992. Conducting playback experiments and interpreting their results. Pp. 59-77 in: P. K. McGregor, ed. Playback and studies of animal communication . (Plenum Press: New York). Gerhardt, H. C. 1994. Reproductive character displacement of female mate choice in the grey treefrog, Hyla chrysoscelis. Anim. Behav . 47: 959- 969. Grant, P. 1972. Convergent and divergent character displacement. Biol. J. Linn. Soc. 4: 39- 68.

180 Halliday, T. R. 1983. The study of mate choice. Pp. 3- 50 in: P. Bateson, ed. Mate Choice . (Cambridge University Press). Harrison, R. G. 1979. Speciation in North American field crickets: Evidence from electrophoretic comparisons. Evolution 33: 009- 1023. Harrison, R. G., and S. M. Bogdanowicz. 1995. Mitochondrial DNA phylogeny of North American field crickets: perspectives on the evolution of life cycles, songs, and habitat associations. J. Evol. Biol . 8: 209- 232. Hewitt, G. M. 1996. Some genetic consequences of ice ages, and their role in divergence and speciation. Biol. J. Linn . Soc. 58: 247- 276. Howard, D. J. 1993. Reinforcement: Origin, dynamics, and fate of an evolutionary hypothesis. Pp. 46-69 in: R. G. Harrison, ed. Hybrid zones and the evolutionary process . (Oxford University Press: New York). Kelly, J. K., and M. A. Noor. 1996. Speciation by reinforcement: A model derived from studies of Drosophila. Genetics 143: 1485- 1497. Kritsky, G., and S. Simon. 1996. The unexpected 1995 emergence of periodical cicadas (Homoptera: Cicadidae: Magicicada) in Ohio. Ohio J. Sci. 96: 27- 28. Liou, L. W., and T. D. Price. 1994. Speciation by reinforcement of premating isolation. Evolution 48: 1451- 1459. Littlejohn, M. J. 1965. Premating isolation in the Hyla ewingi complex (Anura: Hylidae). Evolution 19: 234- 243. Littlejohn, M. J., and J. J. Loftus-Hills. 1968. An experimental evaluation of premating isolation in the Hyla ewingi complex (Anura: Hylidae). Evolution 22: 659- 663. Littlejohn, M. L. 1977. Long-range acoustic communication in anurans: An integrated and evolutionary approach. Pp. 263-294 in D. H. Taylor and S. I. Guttman, eds. The reproductive biology of amphibians . (Plenum Press: New York). Lloyd, M. 1984. Periodical cicadas. Antenna 8: 79- 91. Lloyd, M. 1987. A successful rearing of 13-year periodical cicadas beyond their present range and beyond that of 17-year cicadas. Am. Midl. Nat . 117: 362- 368. Lloyd, M., and H. S. Dybas. 1966a. The periodical cicada problem. I. Population ecology. Evolution 20: 133- 149. Lloyd, M., and H. S. Dybas. 1966b. The periodical cicada problem. II. Evolution. Evolution 20: 466- 505. Lloyd, M., and J. White. 1976. Sympatry of periodical cicada broods and the hypothetical four-year acceleration. Evolution 30: 786- 801. Lloyd, M., G. Kritsky, and C. Simon. 1983. A simple Mendelian model for the 13- and 17-year life cycle of periodical cicadas, with historical evidence for hybridization between them. Evolution 37: 1162- 1180.

181 Loftus-Hills, J. J., and M. J. Littlejohn. 1992. Reinforcement and reproductive character displacement in Gastrophryne carolinensis and G. olivacea (Anura: Microhylidae): A reexamination. Evolution 46: 896- 906. Marlatt, C. L. 1923. The periodical cicada. U.S.D.A. Bureau of Entomology Bulletin 71, 183 pp. Martin, A. P., and C. Simon. 1988. Anomalous distribution of nuclear and mitochondrial DNA markers in periodical cicadas. Nature (Loondon) 336: 237- 239. Martin, A. P., and C. Simon. 1990. Differing levels of among-population divergence in the mitochondrial DNA of periodical cicadas related to historical biogeography. Evolution 44: 1066- 1080. Moore, T. E. 1993. Acoustic signals and speciation in cicadas (Insecta: Homoptera: Cicadidae). Pp. 269-284 in: D. R. Lees and D. Edwards, eds. Evolutionary patterns and processes . (Academic Press: London). Noor, M. A. 1995. Speciation driven by sexual selection in Drosophila. Nature (London) 375: 674- 675. Otte, D. 1989. Speciation in Hawaiian crickets. Pp. 482-526 in: D. Otte and J. A. Endler, eds. Speciation and its consequences . (Sinauer : Sunderland, MA). Paterson, H. E. H. 1993. Evolution and the recognition concept of species . (Johns Hopkins University Press: Baltimore, MD). Rice, W. R., and E. E. Hostert. 1993. Laboratory experiments on speciation: What have we learned in 40 years? Evolution 47: 1637- 1653. Simon, C. 1988. Evolution of 13- and 17-year periodical cicadas (Homoptera: Cicadidae: Magicicada ). Bull. Entomol. Soc. Am . 34: 163- 176. Simon, C., J. Tang, S. Dalwadi, G. Staley, J. Deniega, and T. Unnasch. Submitted. Restricted gene flow between 13-year cicadas and coexisting “17-year cicadas with 13-year life cycles.” Tauber, C. A., and M. J. Tauber. 1977a. A genetic model for through habitat diversification and seasonal isolation. Nature (London) 268: 702- 705. Tauber, C. A., and M. J. Tauber. 1977b. Sympatric speciation based on allele changes at three loci: evidence from natural populations in two habitats. Science (Wash. D. C.) 197: 1298- 1299. Toms, R. B., and D. Otte. 1988. New genera, species and records of east and southern African tree crickets (Orthoptera: Gryllidae: Oecanthinae). Ann. Transvaal Museum 34: 469- 533. Waage, J. K. 1979. Reproductive character displacement in Calopteryx (Odonata: Calopterygidae). Evolution 33: 104- 116. Waddington, C. H. 1953. Genetic assimilation of an acquired character. Evolution 7: 118- 126.

182 Walker, T. J. 1974. Character displacement in acoustic insects. Am. Zool. 14: 1137- 1150. Webb T. III, P. J. Bartlein, S. P. Harrison, and K. H. Anderson. 1993. Vegetation, lake levels, and climate in eastern North America for the past 18,000 years. Pp. 415-467 in: H. E. Wright, Jr., J. E. Kutzbach, T. Webb III, W. F. Ruddiman, F. A. Street- Perrott, and P. J. Bartlein, eds. Global climates since the last glacial maximum . (University of Minnessota Press, Minneapolis, MN). Weber, T., T. E. Moore, F. Huber, and U. Klein. 1987. Sound production in periodical cicadas (Homoptera: Cicadidae: Magicicada septendecim, M. cassini). Pp. 329-336 in: Proc. 6th Auchen. Meeting, Turin, Italy, 7-11 Sept. 1987. West-Eberhard, M. J. 1983. Sexual selection, social competition, and speciation. Q. Rev. Biol. 58: 155- 183. West-Eberhard, M. J. 1989. Phenotypic plasticity and the origins of diversity. Ann. Rev. Ecol. Syst. 20: 249- 278. White, J., and M. Lloyd. 1975. Growth rates of 17- and 13-year periodical cicadas. Am. Midl. Nat. 94: 127-1 43. White, J., and M. Lloyd. 1979. 17-year cicadas emerging after 18 years: a new brood? Evolution 33: 1193- 1199. Williams, K. S., and C. Simon. 1995. The ecology, behavior, and evolution of periodical cicadas. Ann. Rev. Entomol . 40: 269- 295. Williams, K. S., K. G. Smith, and F. M. Stephen. 1993. Emergence of 13-yr periodical cicadas (Cicadidae: Magicicada): phenology, mortality, and predator satiation. Ecology 74: 1143- 1152. Young, D., and R. K. Josephson. 1983. Pure-tone songs in cicadas with special reference to the genus Magicicada. J. Comp. Physiol . 152: 197- 207.

183 Table 6.1. Traits distinguishing Magicicada neotredecim and other Magicicada species. "Pronotal extension" is the lateral extension of the pronotum behind the eye. For additional description and color photographs see Alexander and Moore (1962)

Species Life Cycle Abdominal sternite color Dominant call Pronotal Length of call (years) (each) pitch (kHz) extension color (seconds)

M. neotredecim Marshall and Cooley (new 13 orange with black lateral band 1.35 - 1.90 § species) or center orange 1.5 - 4 184 M. tredecim (Walsh and Riley) 13 mostly orange 1.00 - 1.25 orange 1.5 - 4 §

17 orange with black lateral band 1.25 - 1.50 § M. septendecim (L.) or center orange 1.5 - 4

†† 17 black, rarely with weak > 3.00 §§ M. cassini (Fisher) orange lateral band black 2 - 6

†† 13 black, rarely with weak > 3.00 §§ M. tredecassini Alexander and Moore orange lateral band black 2 - 6

M. septendecula Alexander and Moore 17 black with orange lateral band > 3.00 black 7 - 14 †

M. tredecula Alexander and Moore 13 black with orange lateral band > 3.00 black 7 - 14 †

§ Roughly pure-tone, musical buzz terminating in a noticeable drop in pitch; no ticks. Usually 2-3 calls between flights. §§ Rapid series of ticks followed by high-pitched, broad-spectrum buzz that rises and then falls in intensity and pitch. Usually 1-2 calls between flights. † Repeated, rhythmic high-pitched, broad-spectrum tick-buzz phrases, followed by repeated phrases containing only ticks. Usually 1 call between flights. †† Orange band, if present, often interrupted medially. Table 6.2. Relationship of abdomen color to male call pitch and female pitch preference in M. -decim at the Sharp Co. AR “powerline” site. The male and female M. -decim samples were each divided on the basis of calling song pitch and pitch preference (at 1.5 kHz). The abdomen colors of these groups are significantly different.

Males Average male dominant call Abdomen Call pitch in kHz color score Species Category n (mean ± SD) (mean ± SD) designation < 1.5 kHz 27 1.11 (±0.07) 3.67 (±0.55) M. tredecim > 1.5 kHz 124 1.70 (±0.07) 2.00 (±0.44) M. neotredecim Abdomen color differs significantly in each group (Mann-Whitney U= 3348.0; P< 0.001)

Females Average female call pitch Abdomen Preference preference in kHz color score Species Category n (mean ± SD) (mean ± SD) designation < 1.5 kHz 13 1.22 (±0.11) 3.23 (±0.83) M. tredecim > 1.5 kHz 35 1.75 (±0.14) 2.20 (±0.72) M. neotredecim Abdomen color differs significantly in each group (Mann-Whitney U= 366.5; P< 0.001)

185 Table 6.3. Dominant frequencies of Magicicada septendecim chorus recordings in the UMMZ sound library.

Brood State Year Frequency I VA 1995 1.41 II CT 1962 1.39 II VA 1996 1.40 III IL 1997 1.35 IV KS 1998 1.34 VIII PA 1976 1.32 XIII IL 1973 1.30-1.31 XIV OH 1957 1.29-1.32

186 M. septendecim (17) - abdomen black and orange - call pitch 1.25-1.50 kHz - mtDNA lineage A

M. cassini (17) M. septendecula (17)

M. neotredecim (13) M. tredecim (13) - abdomen black and orange - abdomen mostly orange - call pitch 1.25-1.90 kHz - call pitch 1.00-1.25 kHz - mtDNA lineage A - mtDNA lineage B

M. tredecassini (13) M. tredecula (13)

Figure 6.1. Distribution of the seven periodical cicada (Magicicada) species, summarized from county-level maps in Simon (1988) and from 1993-1998 field surveys. The 17-year species are sympatric except in peripheral populations: M. cassini alone inhabits Oklahoma and Texas, while only M. septendecim is found in some northern populations (Dybas and Lloyd 1974). Two 13-year species (M. tredecassini and M. tredecula) are sympatric across the entire 13-year range, while the remaining 13-year species, M. tredecim and the new species M. neotredecim, overlap only in the central US. County-level maps overestimate distribution limits, hence the overlap between the 13- and 17-year populations is probably exaggerated. We have found only limited 13/17 overlap during recent surveys in Illinois (1997-1998 emergences of Broods III and XIX, respectively) and have adjusted this map accordingly. The overlap of M. tredecim and M. neotredecim is plotted from recent field surveys (this paper, Simon et al. submitted). Characters distinguishing M. Ðdecim siblings are noted; the M. -cassini and M. -decula siblings are distinguishable only by life cycle. "Call pitch" is dominant pitch of male call phrase; mtDNA lineage refers to types described in Martin and Simon (1990).

187 M. septendecim M. cassini + M. septendecula A.

Power kJ/Hz

1 2 3 4 5 6 7 Pitch (kHz)

B. M. tredecim M. tredecassini + M. tredecula

M. neotredecim

Power kJ/Hz

1 2 3 4 5 6 7 8 Pitch (kHz)

Figure 6.2. Power spectra of typical 17- year chorus containing M. septendecim, M. cassini, and M. septendecula (A); and 13- year chorus containing M. tredecim, M. neotredecim, M. tredecassini, and M. tredecula (B). Note that M. -cassini and M. -decula frequency characteristics differ little in the two choruses, but that the presence of two M. -decim species in the 13- year chorus is readily apparent. 17 - year chorus recording S. of Fandon, McDonough Co., IL, 7 June 1980 (Brood III); 13- year chorus along US 51 near county line in southern Jackson Co., IL 1 June 1976 (Brood XXIII).

188 2.5

2.0

M. neotredecim Pitch 1.5 (kHz) 189 1.0 M. tredecim

0.5

0.0 2.0 4.0 6.0 Time (s)

Figure 6.3. Spectrogram (power spectrum vs. time) showing a two-banded, mixed- species chorus of male calls with one call of each species standing out against the background chorus. Individual calls end with a downslur. Comparatively faint slurs of background chorus males overlap and are not visible. Intervening time between calls has been removed. 1.8

1.7 M. neotredecim

1.6

1.5

1.4 190

1.3 Chorus pitch (kHz)

1.2 M. tredecim 1.1

1 71.1 79.9 81 81.5 81.5 81.7 82 83 83.3 84.9 85.5 87.1 87.4 Air temperature, ¡F

Fig. 6.4. Dominant pitch of a mixed M. neotredecim/M. tredecim as a function of air temperature. No effect of temperature is evident. 191

2 Pitch (kHz) 1

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Time (seconds)

Figure 6.5. Sonogram of 30-second recording of an individual male M. neotredecim in Sharp County, AR, illustrating consistency of individual male calling song pitch.

0.6 600

0.5 500

0.4 400

0.3 300 Proportion 192

0.2 200 Power (nJ/Hz)

0.1 100

0 0 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 Pitch (kHz)

Figure 6.6. Power spectrum (shaded area) of mixed M. neotredecim and M. tredecim chorus at "powerline" study site, Sharp Co., AR, with accompanying frequency histograms of male song pitches (black bars) and female pitch preferences (white bars) of individuals collected randomly at the site. M. tredecim (low song pitch) constitute only 8% of the population. IL 100 km

MD

MO

KY VA 193

TN NC AR

N SC

AL GA

Figure 6.7. Proportions of M. neotredecim (black) and M. tredecim (white) estimated from chorus recordings of the 1998 emergence of Magicicada 13-year Brood XIX. IL

MO N

AR

Dominant pitch of chorus (kHz) 1.68 - 1.77 1.58 - 1.67 1.48 - 1.57 1.38 - 1.47 100 km

Figure 6.8. Geographic variation in dominant chorus pitch of M. neotredecim, showing reproductive character displacement (higher pitch songs) in and near the region of overlap with the low-pitched M. tredecim. Lighter shaded circles indicate higher pitched songs. Shaded region is approximate M. tredecim range. Weak choruses not plotted.

194

Song pitch (kHz)

1.8

1.7

1.6

1.5

1.4 M. neotredecim

1.3

1.2

M. tredecim 1.1

1.0

N S

Figure 6.9. Schematic representation of calling song displacement in Brood XIX M. tredecim and M. neotredecim. Light portions of pie graphs represent proportion of M. tredecim present in choruses; dark portions represent proportion of M. neotredecim. In the north, choruses contain only M. neotredecim, while in the south, choruses contain only M. tredecim. Song pitch is displaced in the relatively narrow zone of overlap betweeen these species.

195 N IL 100 km MD

MO

KY VA 196

TN NC

AR

Dominant pitch of chorus (kHz) SC 1.07 - 1.09

AL GA 1.10 - 1.11 1.12 - 1.14 1.15 - 1.17

Figure 6.10. Geographic variation in dominant chorus pitch of M. tredecim, suggesting weak reproductive character displacement. Lighter shaded circles indicate more displaced (lower pitched) songs. Shaded area is approximate M. neotredecim range. Note that range of variation is only 1/4th of that shown in Fig. 6.7. Weak choruses not plotted. A. B. 1.9 40 1.8 * 1.7 1.6 * 35 1.5 * * 1.4 *

1.3 Right wing length (mm) Call pitch (kHz) * 30 1.2 *

1.1 * 1.0 0.9 25 Arkansas Illinois Arkansas Illinois

M. tredecim M. neotredecim M. tredecim M. neotredecim

C. D. 12 8.0

7.5 11

7.0 * Thorax width (mm)

10 First sternite width (mm) 6.5 * * 9 6.0 Arkansas Illinois Arkansas Illinois

M. tredecim M. neotredecim M. tredecim M. neotredecim

Figure 6.11. Box plots of male calling song pitch (A), right wing length (B), thorax width (C), and first sternite width (D) of M. tredecim (Sharp Co., AR; n = 26) and M. neotredecim (Sharp Co., AR, n = 61; and Piatt Co., IL, n = 17). Asterisks indicate outliers; open circles indicate extreme outliers. Song pitch of M. neotredecim is significantly different in Arkansas (sympatry with M. tredecim) and Illinois (allopatry with M. tredecim; P < 0.001, Mann-Whitney-U), indicating reproductive character displacement, while size measurements show no significant differences within M. neotredecim. M. tredecim is different from M. neotredecim in all size traits (for each, P < 0.001; Mann-Whitney-U).

197

* Life cycle North South overlap zone

17 A 17 B 17 C

13 A' 13 A

198 13 B 13 C

Figure 6.12. Formation of an incipient Magicicada species (13-year A') by "nurse brood facilitation" of life cycle mutants. Three 17-year species each have a similar 13-year counterpart, but the 13-year counterpart of species 17 A is not present where the life cycle types overlap geographically. In this situation 13-year life cycle mutants from 17 A (curved arrow) can establish a new species (13 A') in the overlap zone if they emerge synchronously with the 13-year "nurse" brood (13 B + 13 C). Success of life cycle mutants from 17 A is facilitated because 13 B and 13 C provide the rare mutants with numerical protection from predators. Rare life cycle mutants from 17 B and 17 C are likely to be lost to interspecific hybridization with the similar and abundant 13 B and 13 C. Degree of prior divergence of 13 A and the parent species 17 A will influence the evolutionary outcome (mixture, hybrid zone, or reinforcement) of later contact between 13 A' and 13 A. I II III IV

17 13 17 13 17 13 17 13

a a

a a A a' A a' A A

A'' a' A'' a' 199

Figure 6.13. A model of Magicicada speciation using nurse brood facilitation and reinforcement of premating isolation to explain pairs of similar life cycle cognates. Vertical dimension reflects degree of character divergence.

I. 17-year cicadas (A) mutate (to a') and join an overlapping, co-emerging 13-year brood II. Reinforcement occurs between a' and a in the 13-year brood III. Individuals of 13-year a' mutate and join the 17-year brood, forming A" IV. Reinforcement occurs between A and A" in the 17-year brood

Our data suggest that steps I and II have occurred. Steps III and IV are plausible by extension of the model. Hypothesis 1 Observed pattern Hypothesis 2

17 17 13 13

13 13 13 17

M. -decim M. -decim 13 C 17

Present 17 M. -decim ? M. -cassini 13 M. -decula 13

M. -cassini and M. -decula M. -cassini and M. -decula

17 17 13

B M. -decim 13

M. -decim 17 M. -cassini 13 M. -decula

M. -cassini and/or M. -decula

C 17 17 13 C 13 A M. -decim M. -decim C 17 M. -cassini M. -decula 13 Past M. -cassini and/or M. -decula

Figure 6.14. Two hypotheses for the origins of Midwestern 13- year cicadas. In the first hypothesis (left column), only M. septendecim underwent a life cycle change. At time (A), a region (marked C ) contained 13- year M. -cassini and/or 13- year M. -decula, but lacked 13- year M. -decim. At time (B), some 17- year M. -decim in the shaded area of slight overlap between 13- and 17- year cicadas underwent a permanent life cycle change and became a new 13- year species, M. neotredecim, synchronous with the existing 13- year brood to the south, and escaping predation by the "nurse brood" mechanism. Under the alternative hypothesis (right column), three species simultanesouly underwent a life cycle change. At time (A), the Midwestern region (marked C ) presently inhabited by M. neotredecim was inhabited by 17- year cicadas. During time (B), large numbers of 17- year cicadas in the shaded area underwent a permanent life cycle change, becoming 13- year cicadas. The process leading to mass changes in life cycle possibly involved environmentally- induced canalization of formerly plastic life cycles. Under both hypotheses, time (C) is the present, in which two 13- year M. -decim species exist, but any new 13- year M. -cassini or M. -decula species, if they exist, are cryptic. See text for details.

200 Proportion of life cycle phenotypes

Climate parameter

A B C D Time

Figure 6.15. A model of Magicicada life cycle evolution via canalization of a climate-induced life cycle shift. Graph shows temporal change in a climate parameter such as temperature. Pie charts indicate proportion of cicadas emerging in 17 years (light) and 13 years (dark). During stage A, all cicadas emerge on a 17-year cycle but are capable of expressing life cycle length plasticity and emerging in 13 years under unusual climatic conditions. During stage B, the climate changes suddenly and dramatically such that the majority of cicadas are induced to express the 13- year cycle. During stage C, climatic conditions slowly ameliorate, imposing canalizing selection for the majority life cycle phenotype of 13- years because 17-year stragglers are never abundant enough to survive predation. By stage D, the population has evolved to express the new 13-year cycle even under the original conditions.

201 CHAPTER VII

ASYMMETRY AND MATING SUCCESS IN A PERIODICAL CICADA, MAGICICADA SEPTENDECIM (HOMOPTERA: CICADIDAE)

Abstract

Phenotypic symmetry has been portrayed as an index of environmental or genetic quality. However, like any trait, symmetry results both from a heritable genetic predisposition and the influence of the developmental environment, and thus for species in which choosy females gain only genetic benefits, symmetry will be of no greater utility as a choice criterion than any other indicator trait. When choosy females receive material benefits, any functional impairment imposed by phenotypic asymmetry may reduce the benefits available to females accepting asymmetrical mates. This paper (1) discusses the possible causes of asymmetry and clarifies the specific situations in which symmetry-based mate choice is likely, and (2) presents research on female mate choice and symmetry in the periodical cicada, Magicicada septendecim (L.). Male mating aggregations and lack of mating gifts or parental care make this species an ideal target for studying female choice for

202 superior mates, and thus for a relationship between male symmetry and male mating success.

Introduction

For bilaterally symmetric organisms, phenotypic symmetry results from heritable precision in ontogenetic canalization. Deviations from symmetry result from (1) genetic quality (because genotypes susceptible to perturbation or with poorly interacting constituents lead to asymmetries), (2) environmental quality (because conditions during development influence production of a symmetrical phenotype), or (3) interactions between these two sources. In any population, zero-centered, normally distributed deviations from perfect bilateral symmetry are termed “fluctuating asymmetry” (FA; Van Valen 1962). Asymmetries that are “handed” (“antisymmetry;” e.g., in humans, one hand is generally dominant; some are right- while others are left- handed) or directional (e.g., all humans have an asymmetrical heart in which the right side is responsible for pulmonary circulation) have different population distributions. The terms “fluctuating asymmetry,” “antisymmetry,” and “directional asymmetry” refer only to population distributions and not the underlying causes of asymmetry.

Symmetry has been used to address diverse topics, from mate choice to environmental degradation (Clarke 1993, Møller and Swaddle 1997). However, the widespread application of symmetry theory is not without controversy; some of the most recent and hotly debated exchanges in symmetry theory (see Møller 1983a, Palmer 1996, Møller and Thornhill 1997a, 1997b, Houle 1997, Leamy 1997, Markow and Clarke 1997, Palmer and Strobeck 1997, Pomiankowski 1997, Swaddle 1997, Whitlock and Fowler 1997, Clarke 1998, Houle 1998) result from an oversimplification of the multiple causes of asymmetry and the resulting measurable patterns. For example, some claims that mate choice favors individuals exhibiting developmental stability are based on correlations between mean population developmental stability (as measured by asymmetry) and mean

203 population fitness (Møller 1997 but see Clarke 1998), yet do not take into consideration whether individuals receive any benefits at all from using symmetry as a mate choice criterion. Previous studies also suffer from weak tests of their hypotheses; they succeed in uncovering evidence consistent with their hypotheses, but offer no strong tests of the alternatives (Møller and Höglund 1991, Møller 1993b, c, Møller and Pomiankowski 1993, 1994, Møller and Zamora-Muñoz 1997), a problem exacerbated by potential observer biases (see Houle 1998).

Causes of asymmetry

Environmental stress. The term “developmental stability” refers to an organism’s ability to develop a normal phenotype in spite of environmental perturbations. For bilaterally symmetric organisms, deviations from symmetry due to flaws, damage, developmental inaccuracies, or an inability to withstand hostile aspects of the environment are indicators of “environmental stress” (see Møller 1993 a, b). In any population, normally-distributed, non-adaptive asymmetries with a mean value of zero fit the criteria of fluctuating asymmetry (FA; Van Valen 1962); thus the pattern of fluctuating asymmetry is one estimate of the processes involved in developmental stability. Although fluctuating asymmetry is a population index, if a population exhibits FA, then the asymmetries of its members are individual manifestations of developmental instability (and thus individual FA even though the term itself was defined as a measure of population characteristics; see Møller 1993d, Møller and Swaddle 1997 p. 11).

Symmetry is not, however, synonymous with developmental stability. Antisymmetries or directional asymmetries are often adaptive or result from the pleiotropic effects of adaptations. Adaptively asymmetrical structures, such as the human heart, should be just as subject to developmental stress as bilaterally symmetric characters (Møller and Swaddle 1997), but such characters are rarely used in studies of developmental stability. Perturbations might have different impacts depending on when in the developmental

204 process they occur or how effectively subsequent development compensates for them. For instance, the closer such an event happens to the ontogenetic origin of a cell stem line, the more serious could be its effects (Emlen et al. 1994; see Polak 1994). Environmentally- caused deviations from a genome’s developmental program are identifiable only given some information about how that genotype has been selected to develop in the absence of perturbation. Such direct information is virtually unobtainable, and in most cases, we have only a population average phenotype, or, for a specific genotype, a reaction norm, either of which might be affected by population- or environment- specific stresses. Studies of developmental stability thus rely on characters for which it is safe to assume that normal development produces symmetrical phenotypes, because deviations from normal development are readily identifiable as fluctuating asymmetry (Møller and Swaddle 1997).

Environmentally-induced asymmetries are not, by definition, heritable; apparent heritabilities of environmentally-induced asymmetries result from “common environment” effects (e.g., if parents and offspring develop in similarly stressful environments). Environmentally-induced asymmetries are in addition to or in spite of genetic endowment, and are deviations from the most symmetrical phenotype that a given genotype is capable of producing. Evidence for environmentally-induced asymmetry consists of decreased symmetry in laboratory animals under elevated environmental stress (Parsons 1961; reviewed in Møller and Pomiankowski 1993), and some evidence of elevated asymmetry in animal populations living in marginal habitats (see Parsons 1994).

Genetic quality. Discussions of asymmetry often invoke genomic effects, including the concept of “balanced genotypes” (Lerner 1954, Waddington 1957) and the impact of “genetic balance” on development. Both are indices of how well the constituent genes of a genome interact, such that observable asymmetries result from the pleiotropic effects of the genome in its entirety. For example, heterozygous individuals sometimes exhibit a high degree of symmetry (Mitton 1994), and hybrids, often especially

205 heterozygous, tend to be highly symmetrical (Parsons 1990). Heterozygosity may lead directly to increased symmetry, or may do so by conferring resistance to parasites or other stresses promoting asymmetry.

Genetic effects on asymmetry also include inheritable predisposition/ resistance to asymmetrical development and adaptations that compensate for environmental perturbations. A number of studies demonstrate that sensitivity to environmental perturbations is inheritable, although some are more thorough than others in ruling out common-environment effects (Palmer et al. 1994, Houle 1997, Leamy 1997, Markow and Clarke 1997, Palmer and Strobeck 1997, Pomiankowski 1997, Swaddle 1997, Whitlock and Fowler 1997). The many possible causes of asymmetry make any broad generalizations about heritability and symmetry questionable (Whitlock 1996, Woods et al. 1998, but see Møller and Thornhill 1997a, 1997b; Møller and Swaddle 1997).

If symmetrical males enjoy mating or survival advantages, the ability to develop symmetrically would be subject to selection, with poor variants at a disadvantage (e.g., Møller and Pomiankowski 1994, Palmer and Strobeck 1997). This selection would cause the elimination of genetic variants promoting asymmetry (specific examples of “bad genes”), or place a premium on the evolution of modifier or masking genes (specific examples of “good genes”) that enforce symmetry regardless of (or in spite of) genomic qualities or environmental effects. Such modifier genes are another possible genetic influence on phenotypic symmetry. Reduced levels of asymmetry in sexual ornaments, compared to levels of asymmetry in other morphological characters, suggest that such genetic effects exist and are favored by selection (Møller 1990, Møller and Höglund 1991, Møller 1993c, d, e, Møller and Pomiankowski 1993, 1994, Møller 1995).

206 Symmetry and Mate Choice

Symmetry has been linked to male dominance in fallow deer (Malyon and Healy 1994), attractiveness in humans (Gangstead et al. 1994, Møller et al. 1995), and mating success in barn swallows (Møller 1993b, 1994) and zebra finches (Swaddle 1996). Symmetrical male Japanese scorpionflies generally tend to win both inter- and intra- specific fights for control of mating gifts (Thornhill 1992a, b), the most symmetrical male dung flies and houseflies are also the most successful breeders (Liggett et al. 1993, Møller 1996), and the most symmtrical dragonflies (Coenagrion puella) have the highest lifetime mating success (Harvey and Walsh 1993). Although not all studies show a relationship between symmetry and mating success (see Oakes and Barnard 1994, Tomkins and Simmons 1998), at least some mate choice mechanisms appear to favor symmetry or its correlates.

Phenotypic symmetry is tempting as a focus of mate choice studies, because symmetry gives the appearance of being a universal indicator of genetic quality: Nearly all phenotypes exhibit some form of symmetry, and symmetry at least partly reflects genomic qualities. As a result, symmetry appears to occupy a special status as an accurate and uncheatable quality indicator (Møller and Swaddle 1997). However, like all indicator traits, phenotypic symmetry is subject to a paradox: (1) As the number of genes involved in affecting the trait increases (in the extreme, if expression of the trait reflects the overall genome), the less accurately the trait identifies mates able to provide inheritable genetic benefits, because meiosis and independent segregation ensure that large, coadapted pieces of the genome are not passed on intact (thus offspring quality will depend on the “goodness” of the match between the genes in the sperm and egg that join to create the zygote); yet (2) The smaller the number of genes with a predominant influence on the trait, the more accurately the phenotype identifies an inheritable genetic basis for the trait, but the greater the likelihood that population-wide variation will tend to be homogenized by

207 successive generations of selection. The solution to this paradox is that indicator traits generally are not stable unless they are costly or condition-dependent, reducing the likelihood that expression of advantageous phenotypes can become fixed by selection (Zahavi 1977, Halliday 1978, Borgia 1979, West-Eberhard 1979, Andersson 1982, Hamilton and Zuk 1982, Parker 1982, Price et al. 1993, Kodric-Brown and Brown 1984). Condition-dependence implies that traits similar to category (1) above are most likely to become the basis for stable mate choice systems, because condition-dependence involves the performance of the entire organism, and thus involves the entire genome.

Considering symmetry as a condition-dependent trait opens up new possibilities for understanding the mechanisms by which symmetrical mates are chosen. Although studies of neural networks suggest that relatively simple mechanisms could detect and favor symmetrical mates (Johnstone 1994, Enquist and Arak 1994), if symmetry is condition- dependent, it is also possible that females have no evolved mechanisms for detecting symmetry and incidentally choose on the basis of overall condition or performance. For example, in Japanese scorpionflies (Panorpa japonica), male pheromonal attractiveness correlates with wing symmetry (Thornhill 1992c). These results could suggest that some males have both more symmetrical wings and more attractive pheromones, and that females have mechanisms for detecting these differences. If so, however, after several generations all males would be highly symmetrical and attractive. Alternatively, since these insects must fly to scavenge dead arthropods for food, and since dietary substances may be included in pheromones, there is a functional link between wing symmetry, foraging success, and sexual attractiveness. Thus, females could choose not on the basis of a single trait or a small suite of traits, but on the basis of the performance of the entire male organism. Unlike choice on the basis of a particular character (which would quickly exhaust that character’s heritable variation), choice on the basis of performance require no complex or special apparatus for assessing male quality. Females exercising performance- based choice give their offspring access to any genetic advantages of their mate; those

208 requiring male post-zygotic investment may receive superior resources as well (See Heywood 1989, Hoelzer 1989, and Price et al. 1993).

Thus, considering phenotypic symmetry per se as an indicator trait may be an oversimplification; symmetry may instead be a correlate of factors that influence mating success. For highly asymmetrical individuals, lack of mating success may stem from impaired performance of some behavior necessary for survival or mating (Liggett et al. 1993, Møller and Pomiankowski 1994, Møller 1997, Møller and Swaddle 1997). For example, the enhanced mating success of symmetrical flying insects or birds (Harvery and Walsh 1993, Møller 1994, MacLachlan and Cant 1995, Møller 1996) could be explained by aerodynamics. Instead of choosing the most symmetrical males, females might be likely to mate with the best fliers, most, but not all, of whom will incidentally be symmetrical; this is a form of “indirect” mate choice (Chapter 1). Symmetry may also contribute to the ease with which sexual ornaments are recognized (Johnstone 1994, Enquist and Arak 1994), so that some sexual ornaments may be compound characters whose effectiveness depends both on the symmetry and on the extremity of the display (see Møller 1990, Møller and Höglund 1991, Møller 1992, Møller 1993c, d, e, Møller and Pomiankowski 1993, 1994). To the extent that symmetry enhances performance, selection for performance will favor mechanisms that enforce symmetry (Clarke 1994, Møller and Swaddle 1997). As long as females make relative comparisons, and choose only on the basis of direct or indirect male competitive interactions, variation in male quality is inexhaustible, and mate choice on the basis of performance is possible because males could use any means at their disposal to enhance their competitive abilities (this hypothesis is similar to some solutions of the “lek paradox;” see Pomiankowsi and Møller 1995, Hill 1994).

Mate choice and symmetry in periodical cicadas

Magicicada mating aggregations and complex courtship suggest a long history of sexual selection. Alexander (1975) proposed that Magicicada male choruses are non-

209 resource based leks, and that females restricting mating activities to the choruses receive some benefit unobtainable by mating elsewhere. Males are not known to supply material benefits or mating gifts, so potential benefits to females choosing mates are limited to (1) a superior genetic complement for offspring (genetic benefits) or (2) reduced costs from choosing mates most able to appear at a time and location congruent with female interests (phenotypic benefits). Thus, whether females choose on the basis of genetic or phenotypic benefits, Magicicada seem plausible candidates for species in which male performance attributes, affected by symmetry, influence mating success. Because it is not yet known what females gain by visiting male aggregations, a study of mate discrimination and fluctuating asymmetry stands the greatest chance of finding any relationship if it makes use of characters that could be important in both direct and indirect mate choice (Chapter 1). The absence of any relationship between symmetry and male mating success would be evidence that females do not choose mates on the basis of symmetry or any of its correlates.

Magicicada are highly mobile insects; males engage in acoustical courtship, producing sound with a pair of ribbed organs called tymbals. Sexually receptive females join male singing aggregations and produce timed visual and acoustical receptivity signals in reponse to the calls of a nearby male (Chapter 5, Cooley and Marshall ms.). A male perceiving such signals will court, approach, mount, and copulate with the signaling female. Courting males face severe competition from other nearby males, who seek out the signaling female, so this mating system puts a premium on males’ ability to fly to choruses, call and court females within the choruses, walk to receptive females, and be more effective at these tasks than their competitors. Thus, wing, tymbal, and leg asymmetries are likely candidates for having effects on male mating success.

210 Methods and Results

I studied Brood II Magicicada in a large open field near Horsepen Lake State Wildlife Management Area, Buckingham County, VA, in May 1996. Vegetation in the field consisted primarily of oak, maple, and tulip tree stump sprouts from logging approximately two years prior; the surrounding forest contained primarily oaks and maples, interspersed with patches of planted pines. The morning after their final molt, newly emerged cicadas (“teneral” cicadas) are readily identifiable by their dull color, soft bodies, and by their low positions in the vegetation. I collected unmated, teneral female cicadas early in the mornings several days before beginning experiments and allowed them to mature in single-sex storage cages made by enclosing living vegetation in 200 liter hardware cloth cages. Immediately prior to starting the experiments, I captured active, chorusing males.

I erected two large, cylindrical, 3000 liter white nylon tulle cages over stump sprouts, trimming vegetation and creating access openings so that all parts of the cages could be reached and observed with minimum perturbation. I placed 10 mature females and 20 recently captured males into each cage and carefully examined them at least once per hour. Because normal matings last longer than two hours (Chapter 4), I am confident that no matings were unobserved. Sexually receptive females signal to calling males (Chapter 5), causing all males in the vicinity to rush to find and mate with the signaling female. Therefore, the experimental design of this study probably tends to encourage male-male scramble competition and accentuate quality differences among males, although the densities in these cages were not unnaturally high (Chapter 2). I removed all mating pairs upon first observation and immediately preserved them in 70% ethanol. When all females had been removed, or at the end of the day, the remaining, unpaired males were preserved in a separately marked jar. The experiment was repeated the following day using different cicadas. I collected a total of 27 mating pairs; three unmated male cicadas were lost, and all

211 other cicadas were accounted for. Each preserved male was given a numerical code that, without a key, provided no information about mating status. I investigated three classes of asymmetries and their relationships to mating success: (1) Left-right character size asymmetries, (2) Asymmetries in tymbal rib number, and (3) The presence or absence of anomalous wing vein characters.

Left-right size asymmetries

Methods and results. A trained technician, unaware of the design or purpose of the study (see Houle 1998), made measurements of the right and left sides of three forewing characters, two hindwing characters, and two leg characters (Fig. 7.1). For measurements, the technician used a calibrated ocular micrometer mounted in the eyepiece of a Zeiss binocular dissecting microscope and clamped each wing or leg flat between two pieces of microscope slide glass using a homemade holding jig (Appendix I). All measurements on each cicada were repeated four times on separate occasions, and I recoded the specimens so that the technician was unaware of remeasuring the same specimens. I instructed the technician to skip measurements on damaged cicadas.

For each side of each character, I excluded the most extreme value of the four repetitions, making the measurements conservative estimates of asymmetry. For each character, individuals for which there were not three remaining measurements were eliminated from consideration. I analyzed the repeated measurements using an ANOVA design to ensure that variation came from actual asymmetry differences, not from measurement repetitions (see Palmer and Strobeck 1986, Swaddle et al. 1994, Palmer 1994). Mixed-model ANOVAs had the following structure: Factors were repetition (1-3) and side (left/right), and the dependent variables were the seven characters. In only one case (hindwing character H40) did repetition explain a significant amount of the variation in the data. Kolmogorov-Smirnov tests confirmed this relationship, so this character was discarded (Table 7.1).

212 Using the average values calculated above, I calculated percent asymmetry for each character in each individual according to the following formula:

Right - Left 2* = Scaled % Asymmetry Right + Left

Because this index of asymmetry is scaled to the size of the individual, it allows comparisons of the relative levels of asymmetry among individuals. I used percent asymmetry values for all subsequent analyses and comparisons discussed below (Appendix J).

I calculated average sizes for the right and left sides of the six remaining characters for each specimen. To determine whether any asymmetry was truly fluctuating asymmetry or whether it represented species-typical right/left morphological asymmetry (e.g., the right side is typically different from the left), I used a Lilliefors’ test of normality of scaled, signed asymmetries and Sokal and Rohlf’s test of skew and kurtosis significance (Table 7.2). Only characters F15 and FFEM pass both these tests; H46 passes the kurtosis/skew test, but not the Lilliefors’ test. These analyses indicate that some of the characters do not exhibit normal distributions typical of fluctuating asymmetry (Table 7.2). I also calculated sizes for each character by averaging all right/left values, to allow evaluation of any relationship between size and mating success. The numbers of specimens used in the analysis for each character are noted in Table 7.3.

I examined correlations among asymmetries of the different characters by constructing a Pearson pairwise correlation matrix with Bonferroni corrections (Table 7.4). Wing character asymmetries tend to be correlated, but correlation patterns do not immediately suggest uniformity of expressed asymmetry across all characters within an individual. Correlations among character sizes are stronger (i.e. there is some indication of

213 allometric relationships among characters; Table 7.5), but there is no relationship between character size and degree of asymmetry (Table 7.6).

For each individual, I examined the relationship between both the symmetry and average size of each independent character and mating status, using Kruskal-Wallis One- Way ANOVAs (Table 7.7). Mated males had significantly more symmetrical fore-femurs (Character FFEM) than unmated males. I then calculated a cumulative asymmetry score for each individual by summing the unsigned (absolute-value) asymmetries of all characters. Any individuals that lacked symmetry measurements for all of the characters were not used in this analysis. I calculated an alternative cumulative asymmetry score by summing only the unsigned asymmetries of the two characters exhibiting fluctuating asymmetry, F15 and FFEM. These cumulative measurements of asymmetry should tend to (1) expose genomic effects by allowing repeated measures of developmental stability; and (2) maximize any overall asymmetry differences among individuals. Thus, they should be sensitive assays for whether asymmetries of any combinations of characters influence mating ability (cf. Møller and Swaddle 1997 p.2). Using Kruskal-Wallis One-Way ANOVAs, I found no relationship between cumulative asymmetry and mating status, although cumulative fluctuating asymmetry scores for mated males were significantly lower than for unmated males (Fig. 7.2), and was probably influenced by character FFEM.

Tymbal rib asymmetries

Methods and results. The number of tymbal ribs in M. septendecim varies from 11 to 13, and rarely, the right and left tymbals have different numbers of ribs. On each numerically coded specimen, the number of ribs on both tymbals was counted twice. Because tymbal ribs sometimes branch and merge, only sclerotizations that remained separate from any other structures for more than 3/4 of their total length, and only structures that formed an acute angle with the nearest sclerotization were counted as ribs.

214 There was no relationship between tymbal rib count asymmetries and mating success. (Table 7.8).

Wing venation asymmetries

Methods and results. For the same reasons that asymmetrical males may be inferior mates, females may benefit by avoiding males with atypical wings. Because wing venation in Magicicada septendecim can be highly variable, the technician scored each cicada for the presence/absence of extra wing veins and whether the anomalies were bilaterally symmetrical. These venation anomalies are distinct from “wing crumpling,” or incomplete expansion of the wings at the final molt, noted in some crowded populations (White et al. 1979). None of the cicadas used in this study had crumpled wings. Sixteen of the 77 males used in this study had unusual or atypical wing venation, either in the form of extra or missing wing veins. Four of the sixteen mated, while twelve did not. To determine whether mated males were underrepresented in the group of abnormal-winged males, I created a computer algorithm that resampled the data 10,000 times. Each iteration of the algorithm randomly selected a 16 male sample from the 27 mated and 50 unmated males in the study and then tabulated the number of mated males in the sample. After 10,000 iterations, the algorithm constructed percentiles, which describe the likelihood of finding only four mated males in any 16- member random sample from the original data. Percentiles are reported in Table 7.9, and the code for the resampling algorithm is included in Appendix K. The number of mated and unmated males among the atypically-veined males was not significantly different from that expected by chance.

Discussion

Magicicada seem ideal for mate choice studies because their mating aggregations, complex courtship, and male-male competitive behaviors suggest a history of sexual selection in which females discriminate against some potential mates. Male Magicicada

215 septendecim likely provide females no resources (other than sperm) necessary for reproduction, because females prevented from mating eventually oviposit normal- appearing, unfertilized eggs (pers. obs.; Graham and Cochran 1954); thus, females’ potential benefits from discriminating among mating partners would necessarily be genetic or involve phenotypic benefits other than donated resources.

Mate rapprochement behaviors in Magicicada require males to fly, walk, and fend off competitors. Although slight wing and hindleg asymmetries should affect general mobility, they evidently do not affect the final outcome of courtship or correlate with mating success. Overall asymmetry was also not a predictor of mating success, ruling out general effects of genotypic or phenotypic quality. However, foreleg symmetry correlates with male mating success, suggesting some special functional significance of this character. Walking cicadas appear to use their forelegs to probe surfaces in front of them before venturing onto them. Cicadas with damaged or missing foreleg claws have reduced mobility because they tend to paw at the surface in front of them and remain motionless (pers. obs.); less extreme damage or deformation could have less noticeable effects. The conditions in these experiments probably accentuated male-male scramble competition: Several males could simultaneously perceive a female’s signals. If slight foreleg asymmetries lead to slight reductions in mobility or maneuverability, then males with the most symmetrical forelegs should outcompete less symmetrical cicadas as they race to make contact with signaling females. If such differences are relevant under natural conditions, then male-female rapprochement duets and male-male competition would lead to indirect mate choice (Chapter 1), such that males with the most symmetrical forelegs would be the most successful. Although this study suggests that “indirect” mate choice mechanisms cause females to be more likely to mate males with symmetrical forelegs, there is no evidence that females have mechanisms specifically evolved for favoring symmetrical males. One way to test whether differential male success in these experiments results from differential male abilities independent of any female actions would be to repeat the

216 experiments, replacing the females with an experimenter making artificial signals (Chapter 5) and noting whether the males first reaching the artificial stimulus have the most symmetrical forelegs.

Studies, such as the present one, relying on naturally occurring variation to evaluate the relationship between symmetry and mating success, assume either (1) a positive correlation between symmetry and mating success; or (2) that the sample of individuals used accurately reflects potential mating partners. The first assumption is not valid for most insects, because it requires females to make use of “best-of-n” (BN) mate choice criteria, in which higher quality mates receive incrementally more matings (see Alexander et al. 1997, Janetos 1980, Chapter 1). Insects’ life history patterns and limited opportunities for learning make “threshold” choice mechanisms, in which all potential mates meeting certain criteria are equally acceptable, a more likely alternative (Alexander et al. 1997, Chapter 1). Females exercising threshold choice discriminate among potential mates only if unacceptable ones are present, calling into question assumption (2). There is no evidence that the present study included a class of unacceptable males, because extreme males were distributed proportionally among those that did and those that did not mate. The males used in this study were collected from those engaged in mate search behaviors, the pool from which unconfined females normally draw mates. Consequently, from the present study, there is no evidence that periodical cicada females have specially- evolved mechanisms for acquiring especially symmetrical mates from the pool available to them.

Just as foreleg asymmetries seem to influence male success within small cages, other gross asymmetries affecting mobility could prevent males from entering choruses and encountering females. In most cicada emergences a small class of males has deformed wings from faulty ecdysis; these males have impaired flight and/or signaling capabilities. Comparing symmetry characters of grossly misshapen and normal males would be difficult at best, because deformed males lack the landmarks necessary for accurate measurements.

217 Such a comparison might also be trivial, because deformed males are unable to join treetop chorusing centers where they are most likely to encounter receptive females and thus undoubtedly receive, on average, fewer matings than normal males. Again, no specifically evolved mate choice mechanism need be invoked to explain such discrimination. Because the present study did not include deformed males and confined males in proximity to females, its conclusions are limited to understanding how symmetry affects male mating success within mating aggregations and it provides no information on the differential abilities of males to locate and enter these aggregations.

Given asymmetry’s multiple causes and the diverse interests of individuals choosing mates, there is little reason to expect symmetry to be a universal mate choice criterion. Although we have no reason to expect that phenotypic symmetry is an automatic indicator of quality, the qualities for which symmetry is a proxy may lead to predictable gains for choosy individuals. These predictions lead to testable hypotheses addressing how females might, actively or incidentally, create or participate in situations favoring symmetrical males: 1. Unless females have specially evolved mechanisms for directly assessing symmetry, females must choose males on the basis of functional correlates of symmetry. 2. For those species without male-donated resources, male parental care, or other phenotypic benefits available from differential mating, females can benefit only from choosing among males whose asymmetry differences have some genetic basis. 3. For those species where males provide access to resources, females can benefit from choosing if there is a genetic basis to symmetry or a functional link between symmetry and male performance (which may have heritable components), measured as ability to gather and contribute resources.

These statements lead to testable hypotheses linking sexual selection and asymmetry theory. These testable hypotheses underscore the importance of explicitly considering the possible selective advantages of choosiness and the applicability of specific patterns of asymmetry to hypothetical evolved mechanisms for mate choice in the basis of symmetry.

218 Literature Cited Alexander, R. D., 1975. Natural selection and specialized chorusing behavior in chorusing insects. Pp. 35-77 in: D. Pimentel, ed. Insects, Science, and Society . (Academic Press: New York). Alexander, R. D., D. C. Marshall, and J. R. Cooley, 1997. Evolutionary Perspectives on Insect Mating. Pp. 1- 31 in B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . Cambridge University Press. Andersson, M. 1982. Sexual selection, natural selection, and quality advertisement. Biol. J. Linn. Soc. 17: 375- 393. Borgia, G. 1979. Sexual selection and the evolution of mating systems. Pp. 19- 80 in: M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects 19-80. (Academic Press: New York). Borror, D. J., C. A. Triplehorn, and N. F. Johnson. 1989. An introduction to the study of insects. (Saunders: Philadelphia). Clarke, G. M. 1993. Fluctuating asymmetry of invertebrate populations as a biological indicator of environmental quality. Environmental Pollution 82: 207-11. Clarke, G. M. 1994. The genetic basis of developmental stability. I. Relationships between stability, heterozygosity, and genomic coadaptation. Pp. 17-25 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Clarke, G. M. 1998. Developmental stability and fitness: The evidence is not quite so clear. Am. Nat. 152: 762- 766. Emlen, J. M., D. C. Freeman, and J. H. Graham. 1994. Nonlinear growth dynamics and the origin of fluctuating asymmetry. Pp. 79-96 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Enquist, M., and A. Arak. 1994. Symmetry, beauty, and evolution. Nature (London) 372: 169-172. Gangstead, S. W., R. Thornhill, and R. A. Yeo. 1994. Facial attractiveness, developmental stability, and fluctuating asymmetry. Ethology and Sociobiology 15: 73-85. Graham, C., and A. B. Cochran. 1954. The periodical cicada in Maryland in 1953. J. Econ. Ent . 47:242-244. Halliday, T. R. 1978. Sexual selection and mate choice. Pp. 180-231 in: J. R. Krebs and N. B. Davies, eds. Behavioral Ecology: An Evolutionary Approach . (Blackwell: Oxford).

219 Hamilton, W. D., and M. Zuk. 1982. Heritable true fitness and bright birds: A role for parasites? Science (Wash. D. C.) 218: 384-387. Harvey, I. F., and K. J. Walsh. 1993. Fluctuating asymmetry and lifetime mating success are correlated in males of the Coenagrion puella (Odonata: Coenagrionidae). Ecological Entomology 18:198-202. Heywood, J. S. 1989. Sexual selection by the handicap mechanism. Evolution 43: 1387- 1397. Hill, G. E. 1994. Trait elaboration via adaptive mate choice: sexual conflict in the evolution of signals of male quality. Ethology, Ecology, and Evolution 6: 351- 370. Hoelzer, G. A. 1989. The good parent process of sexual selection. Anim. Behav. 38: 1067- 1078. Houle, D. 1997. Comment on “A meta-analysis of the heritability of developmental stability” by Møller and Thornhill. J. Evol. Biol. 10: 17-20. Houle, D. 1998. High enthusiasm and low R- squared. Evolution 52: 1872- 1876. Janetos, A. C., 1980. Strategies of female mate choice: A theoretical analysis. Behav. Ecol. Sociobiol. 7: 107-112. Johnstone, R. A. 1994. Female preferences for symmetrical males as a by-product of selection for mate recognition. Nature (London): 372: 172-175. Kodric-Brown, A., and J. H. Brown. 1984. Truth in advertising: The kinds of traits favoured by sexual selection. Am. Nat. 124: 309-323. Leamy, L. 1997. Is developmental stability heritable? J. Evol. Biol. 10: 21-29. Lerner, I. M. 1954. Genetic Homeostasis . (MacMillan: New York). Liggett, A. C., I. F. Harvey, and J. T. Manning. 1993. Fluctuating asymmetry in Scatophaga stercoraria L.: Successful males are more symmetrical. Anim. Behav. 45: 1041- 1043. Malyon, C. and S. Healy. 1994. Fluctuating asymmetry in antlers of fallow deer, Dama dama, indicates dominance. Anim. Behav. 48: 248- 250. Markow, T. A., and G. M. Clarke. 1997. Meta-analysis of the heritability of developmental stability: A giant step backward. J. Evol. Biol. 10: 31-37. McLachlan, A., and M. Cant. 1995. Small males are more symmetrical: Mating success in the midge Chironomous plumosus L. (Diptera: Chironomidae) Anim. Behav. 50: 841- 846. Mitton, J. R. 1994. Enzyme heterozygosity, metabolism, and developmental stability. Pp. 49-67 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Møller, A. P. 1990. Fluctuating asymmetry in male sexual ornaments may reliably reveal male quality. Anim. Behav. 40: 1185-1187.

220 Møller, A. P. 1992. Female swallow preference for symmetrical male sexual ornaments. Nature 357: 238-240. Møller, A. P. 1993a. Fluctuating asymmetry. Nature (London): 363: 217 Møller, A. P. 1993b. Female preference for apparently symmetrical male sexual ornaments in the barn swallow Hirundo rustica. Behav. Ecol. Sociobiol . 32: 371- 376. Møller, A. P. 1993c. Patterns of fluctuating asymmetry in sexual ornaments predict female choice. J. Evol. Biol . 6: 481- 491. Møller, A. P. 1993d. Developmental stability, sexual selection and speciation. J. Evol. Biol . 6: 493- 509. Møller, A. P. 1993e. Developmental stability, sexual selection, and the evolution of secondary sexual characters. Etologia . 3: 199- 208. Møller, A. P. 1994. Sexual selection in the barn swallow (Hirundo rustica ). IV. Patterns of fluctuating asymmetry and selection against asymmetry. Evolution 48: 658- 670. Møller, A. P. 1995. Patterns of fluctuating asymmetry in sexual ornaments of birds from marginal and central populations. Am. Nat. 145: 316- 27. Møller, A. P. 1996. Sexual selection, viability selection, and developmental stability in the domestic fly, Musca domestica. Evolution 50: 746- 752. Møller, A. P. 1997. Developmental stability and fitness: A review. Am. Nat. 149: 916- 932. Møller, A. P., and J. Höglund. 1991. Patterns of fluctuating asymmetry in avian feather ornaments: Implications for models of sexual selection. Proc. R. Soc. Lond. B. : 245: 1-5. Møller, A. P., and A. Pomiankowski. 1993. Fluctuating asymmetry and sexual selection. Genetica 89: 267- 279. Møller, A. P., and A. Pomiankowski. 1994. Fluctuating asymmetry and sexual selection. Pp. 269-281 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Møller, A. P. , M. Soler, and R. Thornhill. 1995. Breast asymmetry, sexual selection, and human reproductive success. Ethology and Sociobiology 16: 207-219. Møller, A. P. and J. P. Swaddle. 1997. Asymmery, Developmental stability, and Evolution . (Oxford University Press). Møller, A. P., and R. Thornhill. 1997a. A meta-analysis of the heritability of developmental stability. J. Evol. Biol. 10: 1-16. Møller, A. P., and R. Thornhill. 1997b. Developmental instability is inheritable. J. Evol. Biol. 10: 1-16.

221 Møller, A. P. and C. Zamora- Muñoz. 1997. Antennal asymmetry and sexual selection in a cerambycid beetle. Anim. Behav. 54: 1509- 1515. Oakes, E. J. and P. Barnard. 1994. Fluctuating asymmetry and mate choice in paradise whydahs, Vidua paradisaea: an experimental manipulation. Anim. Behav. 48: 937- 43 Palmer, A. R. 1994. Fluctuating asymmetry analyses: A primer. Pp. 335-364 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Palmer, A. R. 1996. Waltzing with Asymmetry. BioScience 46: 518- 531. Palmer, A. R., C. Strobeck, and A. K. Chippindale. 1994. Bilateral variation and the evolutionary origin of macroscopic asymmetries. Pp. 203-220 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Palmer, A. R. and C. Strobeck 1986. Fluctuating asymmetry: Measurement, analysis, patterns. Ann. Rev. Ecol. Syst. 17: 391- 421. Palmer, A. R., and C. Strobeck. 1997. Fluctuating asymmetry and developmental stability: heritability of observable variation vs. heritability of inferred cause. J. Evol. Biol. 10: 39-49. Parker, G. A. 1982. Phenotype-limited evolutionarily stable strategies. In: Kings’ College Sociobiology Group, eds., Current Problems in Sociobiology 173- 202. (Cambridge: Cambridge University Press). Parsons, P. A. 1961. Fly size, emergence time, and sternopleural chaeta number in Drosophila. Heredity 16: 455- 473. Parsons, P. A. 1990. Fluctuating asymmetry: An epigenetic measure of stress. Biol. Rev. 65: 131- 145. Parsons, P. A. 1994. Developmental variability and the limits of adaptation: Interactions with stress. Pp. 247-255 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Polak, M. 1994. Parasites increase fluctuating asymmetry of male Drosophila nigrospiracula: Implications for sexual selection. Pp. 257-267 in: T. A. Markow, ed. Developmental Instability: Its origins and Evolutionary Implications . (Kluwer, The Netherlands). Pomiankowski, A. 1997. Genetic variation in fluctuating asymmetry. J. Evol. Biol. 10: 51-55. Pomiankowski, A., and A. P. Møller. 1995. A resolution of the lek paradox. Proc. Roy. Soc. Lond. B. 260: 21-29. Price, T. D., D. Schluter, and N. E. Heckman. 1993. Sexual selection when the female directly benefits. Biol. J. Linn. Soc. 48: 187- 211.

222 Simon, C. 1983. Morphological differentiation in wing venation among broods of 13- and 17- year periodical cicadas. Evolution 37: 104- 115. Swaddle, J. P. 1996. Reproductive success and symmetry in zebra finches. Anim. Behav. 51: 203-210. Swaddle, J. P. 1997. On the heritability of developmental stability. J. Evol. Biol. 10: 57- 61. Swaddle, J. P., M. S. Witter, and I. C. Cuthill. 1994. The analysis of fluctuating asymmetry. Anim. Behav . 48: 986-989. Thornhill, R. 1992a. Fluctuating asymmetry and the mating system of the Japanese scorpionfly, Panorpa japonica. Anim. Behav. 44: 867-879. Thornhill, R. 1992b. Fluctuating asymmetry, interspecific aggression, and male mating tactics in two species of Japanese scorpionflies. Behav. Ecol. Sociobiol. 30: 357- 363. Thornhill. R. 1992c. Female preference for the pheromone of males with low fluctuating asymmetry in the Japanese scorpionfly (Panorpa japonica : Mecoptera) Behavioral Ecology 3: 277- 283. Tomkins, J. L., and L. W. Simmons. 1998. Female choice and manipulations of forceps size and symmetry in the earwig Forficula auricularia L. Anim. Behav. 56: 347- 356. Van Valen, L. 1962. A study of fluctuating asymmetry. Evolution 16: 125-142. Waddington, C. H. 1957. Strategy of the Genes . (Oliver and Boyd: London). West-Eberhard, M. J. 1979. Sexual selection, social competition, and evolution. Proc. Am. Phil. Soc. 123: 222-234. White J. 1973. Viable hybrid young from crossmated periodical cicadas. Ecology 54: 573- 580. White, J., M. Lloyd, and J. H. Zar. 1979. Faulty eclosion in crowded suburban periodical cicadas: Populations out of control. Ecology 60: 305-315. Whitlock, M. 1996. The heritability of fluctuating asymmetry and the genetic control of developmental stability. Proc. Roy. Soc. Lond. B. 263: 849-854. Whitlock, M. C., and K. Fowler. 1997. The instability of studies of instability. J. Evol. Biol. 10: 63-67. Woods, R. E., M. J. Hercus, and A. A. Hoffmann. 1998. Estimating the heritability of fluctuating asymmetry in field Drosophila. Evolution 52: 816-824. Zahavi, A. 1977. The cost of honesty. J. Theor. Biol. 67: 603-605.

223 Table 7.1. P- values for tests of measurement error (significant differences among repetitions of the same measurement) and handedness (significant differences between right and left sides of an individual). For Smirnov test, a character is considered to fail if any of the repetitions was significantly different from any other. ANOVA Smirnov Character Repetition Side Repetition F14 0.312 0.994 Pass F15 0.097 0.651 Pass F27 0.413 0.598 Pass H40 0.010 0.615 Fail H46 0.186 0.186 Pass FFem 0.793 0.133 Pass Hfem 0.496 0.391 Pass

224 Table 7.2. Tests for normality of signed, scaled asymmetries. (Skew) (Kurtosis) Character Lilliefor’s P Sokal G1 Sokal G2 Conclusion F14 0.001 Pass Fail Not FA F15 0.566 Pass Pass FA F27 0.001 Fail Fail Not FA H46 0.010 Pass Pass Possibly FA FFem 0.067 Pass Pass FA Hfem 0.003 Fail Pass Not FA Cumulative 0.000 Fail* Fail* Not FA Cumul. FA 0.402 Fail* Fail* Not FA *Since these are cumulative measures of absolute deviations from symmetry, they have no negative values; data were standardized to allow tests.

225 Table 7.3. Numbers of cicadas used in analysis of symmetry and mating success. Mated Unmated Character males used males used F14 25 43 F15 23 47 F27 22 43 H40 26 40 H46 24 42 FFEM 24 45 HFEM 24 40 Cumulative 22 40

226 0.05. ≤

P F14 F15 F27 (H40) H46 FFEM HFEM Pearson pairwise correlation matrix for character asymmetry. correlation matrix for character Pearson pairwise F14F15F27 1.000 0.411* 0.180 1.000 0.483* 1.000 H46 0.258 -0.027 -0.114 -0.098 1.000 (H40) -0.068 -0.257 -0.121 1.000 FFEM 0.054 0.069 -0.042 -0.121 -0.255 1.000 HFEM 0.016 0.205 0.211 -0.169 0.111 0.073 1.000 * Indicates Bonferroni corrected significance * Indicates Bonferroni corrected Table 7.4. Table

227 0.05. ≤

P F14 F15 F27 (H40) H46 FFEM HFEM Pearson pairwise correlation matrix for character size. correlation matrix for character Pearson pairwise F14F15F27 1.000 0.715 0.600 1.000 0.613 1.000 H46 0.702 0.780 0.636 0.630 1.000 (H40) 0.636 0.586 0.581 1.000 FFEM 0.529 0.680 0.561 0.429 0.511 1.000 HFEM 0.515 0.663 0.557 0.600 0.664 0.752 1.000 *All values significant, Bonferroni corrected significance *All values significant, Bonferroni Table 7.5. Table

228 0.05. ≤

P Symmetry F14 F15 F27 (H40) H46 FFEM HFEM Pearson pairwise correlation matrix with Bonferroni probabilities for character sizes and asymmetry. probabilities for character sizes correlation matrix with Bonferroni Pearson pairwise F14F15F27 0.031* -0.141 -0.008 0.089* 0.073 -0.021* H46 0.003 0.204 -0.064 0.055 -0.116* (H40) -0.069 0.092 0.078 0.045* FFEM -0.111 0.030 0.105 0.204 -0.108 -0.084* HFEM -0.184 0.092 0.011 0.064 -0.045 -0.118 -0.106* * Indicates Bonferroni corrected significance * Indicates Bonferroni corrected Size Table 7.6. Table

229 Table 7.7. Kruskal-Wallis One-Way ANOVA analysis of mating status and signed, scaled asymmetry (left columns) and average character size (right columns). Mated and unmated males differed only in levels of asymmetry for character FFEM. Symmetry Size U P U P

FA characters (Table 7.2): F15 479 £ 0.685 514 £ 0.367 FFEM 880 £ 0.028* 673 £ 0.979

Possibly FA character (Table 7.2): H46 266 £ 0.531 287 £ 0.838

Characters not meeting definition of FA (Table 7.2): F14 613 £ 0.508 747 £ 0.442 F27 733 £ 0.539 674 £ 0.991 HFEM 495 £ 0.112 613 £ 0.789

230 Table 7.8. Results of tymbal rib count asymmetry analysis.

Asymmetrical Symmetrical Wings Wings TOTAL

Did mate 2 25 27 Did not mate 5 45 50 TOTAL 7 70 77

Fisher exact test (two-tail) P < 1.000

231 Table 7.9. Results of resampling algorithm to evaluate whether males with wing asymmetries were as likely to mate as normal males. The algorithm draws a random sample of 16 males from a population of 27 mated and 50 unmated males and determines whether the mating frequency in the random sample matches the mating frequency of the 16 males with asymmetrical wing venation. These results demonstrate that the mating frequency (4/16) in the observed sample of 16 asymmetrical males is indistinguishable from the mating frequency in a sample drawn randomly from the cicadas in this study.

Cumulative Number of frequency “mated” males in 10,000 in sample iterations 0 0.0008 £1 0.01 £2 0.043 £3 0.12 £4* 0.27 £5 0.47 £6 0.67 £7 0.83 £8 0.93 £9 0.98 £10 0.99 £11 1 £12 1 £13 1 £14 1 £15 1 £16 1

*Actual value among males with asymmetrical wing venation.

232 F 27 F 14

F 15

H 40 Forewing

H 46 Hindwing

F FEM

H FEM

Foreleg

Hindleg

Figure 7.1. Characters used in analysis of phenotypic symmetry.

F 14: Forewing character #14 from Simon 1983. F 15: Forewing character #15 from Simon 1983. F 27: Forewing character #27 from Simon 1983. H 40: Hindwing character #40 from Simon 1983. H 46: Hindwing character #46 from Simon 1983. F fem: Fore- femur length. H fem: Hind femur length.

Wing diagram adapted from Simon 1983, leg structure diagrams adapted from White 1973 and Borror et al. 1989.

233 0.07

0.06

0.05

0.04

0.03 234

0.02 Scaled, unsigned asymmetry

0.01

0 Mated Unmated Mated Unmated Cumulative Asymmetry Cumulative FA

Figure 7.2. Comparison of cumulative asymmetry scores for mated and unmated males. Left columns: Cumulative asymmetry (sum of unsigned asymmetries for each character); no significant difference between mated and unmated males (Kruskal-Wallis One-Way ANOVA;U = 198, P ² 0.655). Right columns: Cumulative fluctuating asymmetry (sum of unsigned asymmetries for characters F 15 and Fore-Femur); mated males have significantly smaller cumulative FA scores than unmated males (Kruskal-Wallis One-Way ANOVA;U = 667, P ² 0.022). Bars are standard errors. CHAPTER VIII

THE GEOGRAPHIC DISTRIBUTION AND MATING BEHAVIORS OF TWO CICADA SPECIES (OKANAGANA SPP.) IN NORTHEAST MICHIGAN

Abstract

I present the results of a 3 –year biogeographic survey of Okanagana canadensis and O. rimosa in northeastern Michigan. The abundance of these species varied from year to year, although in all years some individuals of each were present within the survey area. In some specific locations, O. rimosa were present in all years, confirming reported observations that this species is not strictly periodical, even on a microgeographic scale. I also describe both species’ sexual behaviors, which involve stationary, calling males and approaching females. These behaviors are in contrast to those of Magicicada spp., the periodical cicadas of North America, in which males actively search for sexually receptive females. I discuss how behaviors such as those found in Okanagana could have evolved into the unusual mating behaviors typical of Magicicada.

235 Introduction

Okanagana canadensis Provancher, and the closely related O. rimosa (Say) are cicadas common in northern Michigan. Life cycles of these species have been reported as four years (Kirby and McPhee 1963) 8-10 years (Huber et al. 1979), or, based upon studies of nymphal development, 9 years (Soper Delyzer, and Smith 1976). Okanagana species are not periodic, because in localities where they are known to occur, some adults can be found in almost any year, although the species’ abundance varies annually.

This study involves two parts: 1. Biogeographic survey of O. canadensis and O. rimosa emergences in northern Michigan, 1996- 1998 and 2. Behavioral studies of O. canadensis and O. rimosa. The biogeographic survey is part of a long-term effort to provide a permanent record of detailed distributional information and resolve ambiguities about these species’ life cycles. The behavioral studies complement ongoing research on the behaviors of other cicadas, including the periodical cicadas of North America (Magicicada spp.). The sister-group relationship of Okanagana and Magicicada has been inferred on the basis of gross similarity of appearance and habits1 (see Simon 1979), although no molecular data support this relationship (Simon pers. comm.). Deciphering the taxonomic relationships of North American cicadas is made difficult by the uniqueness of Magicicada. The sexual behaviors of Magicicada, (in which males search for signaling females; Chapter 5) appear unique2, while the sexual behaviors of Okanagana, (in which, as reported here, females approach calling males) are similar to those found in other cicada genera, such as Tibicen (Alexander 1960, Alexander and Moore 1962, unpub. observations), or Diceroprocta (Alexander and Moore 1962). Regardless of the taxonomic

1 Both Okanagana and Magicicada lack tymbal covers and oviposit in live twigs, while members of another large North American genus, Tibicen, have covered tymbals and oviposit in dead wood or bark. See Appendix L. 2 However, Alexander (1960) and Alexander and Moore (1962) suggest that some of the chorusing behaviors that seem unique to Magicicada are also found in species of the genera Okanagana, Tibicen, and Diceroprocta in dense emergences.

236 relationships among these genera, the sexual behaviors of present-day Okanagana may be similar to those of ancestral Magicicada, and comparisons and contrasts of the behavior and biology of Okanagana and Magicicada allow inferences about how sexual selection shaped the mating system of each.

Methods and Results

Biogeographic surveys

Because of cicadas’ conspicuous, species-specific calls, mapping species distributions over large geographic areas is possible (Marshall et al. 1996, Alexander et al. in prep.). The distributions of O. canadensis and O. rimosa were determined by driving no more than 35 miles per hour along two lane roads with the car windows open and marking the species present on a high quality map, using methods similar to those used for other species of cicadas (Fig. 8.1; see Marshall et al. 1996, Alexander et al. in prep.).

Results. Surveys of O. canadensis and O. rimosa on July 16 1996 and August 1, 1996 (Fig. 8.2), July 25 1997 (Fig. 8.3), and June 20 1998 (Fig 8.4) illustrate that in some locations, O. canadensis were present in all three years of the study, and in 1998, there was a marked increase in the abundance of both O. canadensis and O. rimosa. The increase in abundance in the 1998 survey is not attributable to the earlier date of the survey because there was a decrease in local abundance of O. canadensis in some specific sampling locations, such as at the intersection of Bridge View Road and US 2. O. canadensis were observed primarily in cedars (occasionally aspen), with singing males located low (usually less than 8 feet off the ground n= 20) and on the central trunk or at the base of a large branch. O. canadensis seemed to prefer rocky, high ground in which northern white cedar was present. In contrast, O. rimosa were often present in sandy areas, associated with quaking aspen, maple, or oak, and seldom found on cedars. The calling habits of O. rimosa are more variable than those of O. canadensis; male O. rimosa

237 call from a variety of locations, from grass stems to treetops, and sing from perches at the ends of branches as well as from the bases of branches or trunks.

Upper Peninsula. On July 16, 1996, a substantial (ca. 20) emergence of O. canadensis was discovered at the interchange of US 2 and Bridge View Road, near the ramp from eastbound US 2 to Southbound I-75, in Mackinac County, Michigan. A search of the area yielded no nymphal skins and no fresh oviposition scars; however, some scars reminiscent of cicada eggnests were noted on small branches of white cedar. O. canadensis were present at this location on July 24, 1996, July 25, 1996, August 1, 1996, July 25, 1997 and June 20, 1998.

On July 16, 1996 O. canadensis were noted at I-75 and M-134, and at I-75 and Bay Road, and in the town of Rudyard, near where H-40 crosses a small river and railroad tracks. They were not noted north of Rudyard on that date.

On July 17, 1996, no cicadas were noted between Rudyard and Rexford, or along M-28 from Rudyard to Munising. However, both O. canadensis and O. rimosa were heard at the Seney National Wildlife Refuge, Schoolcraft County. No further Okanagana were noted along M-77 to US 2, or along US 2 from Blaney Park to Brevort. At Brevort and at the intersection of US 2 and Gudmonson Road, dense populations of O. canadensis were noted, with isolated patches at points in between.

The survey of Okanagana emergences continued on July 25, 1996, and the same route as follows was revisited on August 1, 1996, July 25 1997 and June 20, 1998. East of Rudyard on M-48, many O. canadensis were present in aspens at Munuscong Creek before the intersection of M-48 and Fish Road. No cicadas were heard at this location in 1997 or 1998, but O. canadensis were present in 1997 and 1998 along the fork of Munuscong Creek crossing M-48 just east of its intersection with Fish Road. In 1998, east of the intersection of M-48 and 5- mile road, O. rimosa were almost continuously present

238 in wooded areas along M-48 until the intersection with Blair Road; O rimosa were also present along M-48 just east of the intersection with Pennigton Road, and O. canadensis were present at the curve where M-48 joins Fairview Road. O. rimosa were also present in the wet areas associated with East Branch Creek. No cicadas of any species were heard along this stretch in 1996 or 1997.

In 1996 and 1998, O canadensis were present along M-48 at Stalwart, and on high ground just before the intersection of M-48 and Prentiss Bay Road; O. rimosa were present in these locations in 1998. In 1998, both species were present in small numbers in wooded areas between Prentiss Bay Road and North Caribou Drive; O. rimosa may have been present at the curve in M-48 just before intersection of M-48 and North Caribou Drive in 1996 as well, but the only male heard stopped singing after first being heard, so the record is not reliable.

In 1996- 1998, O. canadensis were almost continuously present along M-134 from high ground midway between Albany Island and Cadogan Point to Whitefish Point, and from McKay Bay though Cedarville and continuing west of Cedarville until Nun’s Creek Road. In 1996, O. rimosa were present high in trees at Steele River. In 1998, O. canadensis were present in great abundance along this stretch, with few forested areas silent. O. rimosa were also occasionally present in 1998 in this location. No individuals of either species were heard in any year from a line extending directly north from the western edge of Search Bay west to Ponchartrain Shores, but both species were heard between Ponchartrain Shores and I-75 in 1996. In 1998, O. canadensis were abundant along this stretch. On August 1, 1996, both O. canadensis and O. rimosa were singing along this stretch, whenever trees were present.

In 1997, O. canadensis were present in small numbers along H 63 between M-48 and St. Ignace and were especially abundant near the Carp River. In 1998, O. canadensis were continuously present along this stretch, and O. rimosa were present near Planz Lake.

239 Lower peninsula. On July 16, 1996, and June 20 1998 O. rimosa were heard singing in treetops at the University of Michigan Biological Station, Cheboygan County, Michigan. In 1996 and 1997, few O. rimosa were present along Riggsville Road between I-75 and Bryant Road, along Bryant Road between Riggsville Road and Douglas Lake Road, or along Douglas Lake Road between Bryant Road and the east branch of the Maple River. In 1998, O. rimosa were present in great abundance (> 1000/mile of road) in these locations, with emergence hole densities approaching 50 per square meter. No O. canadensis were found in any year at these sites.

Behavioral observations

Methods

Okanagana canadensis. I conducted intensive behavioral observations of uncaged cicadas near the interchange of US Highway 2 and Interstate 75, Mackinac County, Michigan on July 16, 24, 25, and August 1, 1996, July 27 1997, and on June 20, 1998. I made most observations on O. canadensis in an overgrown field bounded by US 2 (eastbound), an entrance ramp to Interstate 75 (southbound), and Bridge View Road. Vegetation in this field consisted of shrubs and small northern white cedars (Thuja occidentalis). For some observations, I placed cicadas in mesh cages, made by folding a piece of white nylon tulle approximately 1 x 2 m over small cedar seedlings.

On July 25, 1996, I captured and caged 9 male O. canadensis using a dead female Magicicada septendecim specimen as a lure, and I captured 8 additional males by other means. I captured two females who landed on the male cage and placed them in a separate cage, along with four additional females captured by other means. On August 1, I collected four males and one female cicada using similar methods. I made some observations in the field, by placing female cicadas in the male cage and videotaping resultant sexual interactions for later analysis, but I observed no matings in the field. After initial field

240 observations, I separated cicadas by sex, placed them on recently cut tree branches, and transported them to mesh bag cages placed on an elm sapling bordering a pasture at Richard D. Alexander’s farm near the intersection of Bethel Church Road and Schneider Road, Freedom Township, Washtenaw County, MI. Transportation deprived cicadas of fresh vegetation for a maximum of six hours.

In 1997, using similar methods, I captured and transported to Washtenaw County 4 male and 6 female Okanagana canadensis, except instead of placing the cicadas in mesh cages, I observed them in a large (4m x 4m x 4m) screen cage (a “flight cage”) placed over living vegetation. After stocking the cage, I made observations from a truck parked nearby, because O. canadensis not engaged in sexual behavior are easily startled by the presence of humans. I videotaped all sexual interactions. Videotaping required an observer to be out of the truck, but the vehicle was parked in such a way that, from the location of the cages, the observer could not be silhouetted against the horizon. When cicadas were not being observed or videotaped, they were separated by sex.

Okanagana rimosa. Because of low cicada densities, I was unable to capture female O. rimosa in 1996 and 1997. In 1998, O. rimosa emerged in locally high densities (approaching 50 emergence holes per square meter, comparable to Magicicada); I collected and observed them in an open, sandy field dominated by bigtooth aspen (Populus grandidentata) and other shrubby deciduous vegetation near the University of Michigan Biological Station Streams Laboratory (Fig. 8.5). For most observations, I placed cicadas in a large flight cage erected over a small quaking aspen (Populus tremuloides).

Measurements of mating duration. In 1996- 1998, I measured the copulation durations of 4 O. canadensis and 4 O. rimosa mating pairs by observing cicadas continuously during courtship and mating, recording the time at which the male first engaged his genitalia and the time at which he separated completely. In 1995, during the Brood I Magicicada emergence at Alum Springs, Rockbridge County, Virginia, I measured

241 copulation durations of 43 mating pairs of Magicicada septendecim by placing mature, unmated females and recently caught males in 200 liter mesh cages. To establish starting and stopping times for each observed copulation, I monitored cages continuously or scanned them every 10-15 minutes, or ad lib scanned when I heard male courtship sounds. Mating durations for M. septendecim are thus, in some cases, estimates ± 30 minutes.

Results

Okanagana canadensis. Males typically call one per tree. Male O. canadensis tend to perch no more than 8 feet off the ground, sometimes at ground level, (approx. 20 noted, many directly observed, drive by observations of many others for height and tree species information) on the main trunk of small trees, usually cedar (see Davis 1919). Males call with wings slightly open and abdomen extended, sometimes facing down, or, if on a side branch, towards the main trunk. Male calls are broad-frequency, slightly metallic lisps of variable duration. A sonogram of a portion of a male O. canadensis calling bout is included in Fig. 8.6. In 1997, 22 male O. canadensis calling bouts averaged 2:29 minutes, and 13 periods of silence between calling averaged 2:22 minutes (Table 8.1); however, bout lengths are highly variable, and some of the shorter bouts may be courtship behaviors. Males do not engage in “sing-fly” behavior (Alexander and Moore 1962, see also Gwynne 1987), but rather remain on the same calling perch for long periods. For 5 males observed in the field, an estimate of the minimum time spent at a calling perch is 17± 16 minutes (Table 8.1), but males’ lack of movement, and difficulty in determining arrival times at calling perches suggest this is an underestimate.

With few exceptions (see below), calling male O. canadensis were easily startled and difficult to capture (pers. obs; see Davis 1919). Males emit an alarm call when handled, confined, or harassed. Both sexes flew silently from their calling perches at the slightest provocation; alternatively, some calling males ceased calling and remained immobile and silent on their perches. Immobile males were difficult to find because of their

242 mottled coloring, and they remained immobile even if their calling tree was thrashed and shaken. These behaviors suggest that male O. canadensis can adopt two alternative predator avoidance strategies, escape flight or inconspicuous camouflage.

Okanagana rimosa. Male O. rimosa behavior is generally similar to that of O. canadensis, except males appear less selective about calling perches, perching on deciduous trees (often higher than 20 feet), grass stems, or on the ground. A sonogram of a portion of a male O. rimosa calling bout is included in Fig. 8.7. In 1998, although individual male O. rimosa could be found calling on trees, shrubs, or even on the ground as in previous years, there were also single trees that had dense Magicicada – like choruses, except that males in the trees remained still between calling bouts and did not engage in “call-fly” behavior, and males did not synchronize their calls, as in Magicicada cassini and M. tredecassini (Alexander and Moore 1958). The sound intensity at the base of one such tree was 66db, compared to 80+ db for Magicicada. Males’ lack of movement unless perturbed made estimation of residence time at calling perches difficult. During 102 minutes of videotaped sexual interactions during the dense 1998 emergence, in which on average two males were visible at any given time, only 9 spontaneous, undisturbed flights were observed, suggesting that even at high densities, calling males tend to remain on their perches.

The presence of dense aggregations in one tree, while a seemingly similar tree nearby is devoid of males, is reminiscent of the aggregative choruses of Magicicada. It is likely that O. rimosa males actively flew to such trees as opposed to emerging underneath them; although these trees were appropriate species for oviposition, and emergence holes, shed nymphal skins, and ovipositing females were present at each tree, the density of emergence holes and shed skins underneath one such tree could not account for the total number of adult cicadas in the tree. Furthermore, at the study site, nymphal skins and emerging nymphs were found in the forest bordering the open field, but most calling

243 cicadas were along the forest edge or in trees in the field, suggesting net movement away from the forest center, although differences in emergence phenology between the forest and the field cannot be ruled out. Some trees along the forest edge showed evidence of “flagging,” or twig death due to dense oviposition (Marlatt 1923).

Like O. canadensis males, male O. rimosa emit an alarm call when harassed. However, although female O. rimosa were easily startled, calling male O. rimosa were readily approachable and were more efficiently captured by hand than with a net. Unfortunately, all observations of predator avoidance behavior in O. rimosa were made in 1998, during a locally dense emergence; thus, the apparent differences between O. rimosa and O. canadensis could reflect density effects: When sparse, both species could employ the wary behaviors reported for O. canadensis, while higher densities could prompt behaviors more similar to those reported for O. rimosa. Nevertheless, the association of high densities and “predator-foolhardy” behaviors is similar to the lack of predator- avoidance behaviors in dense Magicicada emergences (Karban 1982, Williams et al. 1993, Williams and Simon 1995). Predators were not absent, however: Local residents reported large gull flocks feeding on cicadas (K. Wehrly, pers. comm.; see also Soper, Delyzer, and Smith 1976), and the sarcophagid fly, Colcondamyia auditrix Soper (Soper, Shewell, and Tyrell 1976) an internal parasite of O. rimosa, may also have been present. Of several hundred cicadas examined, two females had extensive damage to their abdominal tergites, suggesting the emergence of fly larvae.

Among several hundred O. rimosa examined in 1998, two males and two females had visible infections of Massopora levispora Soper, a fungal pathogen. Fungal infections were similar in appearance to Stage I (conidiospore-producing) M. cicadina infections of Magicicada (Soper 1963; Soper, Delyzer, and Smith 1976; White and Lloyd 1983). No cicadas producing resting spores were noted, but unlike Stage II (resting spore producing) M. cicadina infections of Magicicada, second-stage infections of M. levispora do not

244 visibly alter cicadas (Soper 1963; Soper, Delyzer, and Smith 1976). First-stage infections of M. levispora, like those of M. cicadina, may cause male hosts to be behaviorally hermaphroditic (Chapter 5): Male Okanagana mount any stationary object they can locate (see below), and normal males do not tolerate being mounted by others; however, I observed one intriguing case in which a fungus infected male O. rimosa repeatedly appeared to tolerate mounts by different males. This male may, due to fungal infection, have been accepting mounts and acting in the interests of the fungus; he did not appear otherwise lethargic or immobile. Thus, although confirmation awaits additional data, Massospora levispora, like M. cicadina, may increase its chances of spreading by causing males to be sexually attractive to males as well as females. The mechanism by which the fungus accomplishes these changes could involve castration, which the fungus accomplishes by destroying the abdomen and its internal organs, and which may result in feminized behaviors.

An interesting correlate of the absence of “sing-fly” behavior in O. rimosa is the fate of individuals infected with the fungus, Massopora levispora. O. rimosa with spore- producing (Stage II) fungal infections exhibit no outward signs of infection and die more or less in situ (Soper 1963; Soper, Delyzer, and Smith 1976), consistent with a fungal adaptation to deposit spores where cicadas are likely to be found in the next generation, and a correlate of the extreme form of philopatry associated with lack of sing-fly behavior. In contrast, normal Magicicada males engage in sing-fly behaviors, and individuals with Stage II Massospora cicadina infections engage in long, spore-dispersing flights consistent with a fungal adaptation to ensure transmission in such highly mobile species (Alexander et al. in prep.).

Sexual behaviors, both species. Males alternated long bouts of calling with bouts of wing flicking (see Davis 1919), producing from 3-5 flicks per wing flick bout. Male Okanagana did not respond to artificial wing flicks produced either during their call or

245 during periods of wing flicking, but sometimes two wing flicking cicadas seemed to produce click bouts in alternation. Females of both species flew to calling males, which they approached by walking. Except to approach a calling male, females remained still and passive prior to mating. In some cases, males made short (< 20 sec.) calls while being approached by females; these shortened calls could be a form of courtship, although it is doubtful that females require males to make such calls since not all males did so. In some cases, mating occurred after a short (ca. 10 minute) bout of wing flicking in which both male and female were seen flicking, and in which the male never called or made any other signal besides wing flicks. Whether or not the male made short calls or either sex wing flicked, once the female made physical contact with the male, they mated without elaborate courtship (5 O. canadensis, 5 O. rimosa mating sequences observed). For each species, four complete matings were observed; O. canadensis matings lasted an average of 34 minutes, while the O. rimosa matings averaged 19 minutes (Table 8.2). Mating durations in these two species are not significantly different, because the mating duration average in O. canadensis is affected by one extreme outlier (Mann-Whitney; U = 9; P £ 0.773, although small sample sizes make this test of dubious validity).

Males presented with an immobile female or similar object were imperturbably persistent in investigating it and were not startled by the presence or actions of a human observer. In 9/9 cases in which a calling male O. canadensis and in 16/20 cases in which a calling male O. rimosa was presented with a model female, the male mounted, extruded his genitalia, and attempted to copulate with the model (Table 8.3). When calling males were presented with live males of their own species, they courted them as if they were females and attempted to copulate with them. The courted male generally responded with a wing flick or an alarm buzz, which were only moderately effective in deterring the courter. Mating durations in Okanagana averaged 26.6 (± 21.75) minutes (Table 8.2). In contrast, for 43 Magicicada mating pairs, copulation duration averaged 243:24 (± 121:48) minutes.

246 Discussion

The surveys of O. canadensis and O. rimosa distributions demonstrate that these species are not strictly periodical, because adults were often present in the same locations in succeeding years, and thus the cicadas in a given locality are not synchronized in their development. However, these species do vary considerably in their yearly abundance, so it may be appropriate to consider them semi-periodical. Some locations, such as the 1998 study site for O. rimosa, may, with repeated sampling, allow final determination of these insects’ life cycles, although the lack of periodicity and year-classes may continue to make estimation of life cycle length based on emergence records difficult.

Females O. canadensis and O. rimosa signal sexual receptivity by physically approaching calling males, and sexual approach is thus under female control. Males attempt to mount nearby, stationary, cicada-like objects, because the only such objects they will ever normally encounter are sexually receptive females. Males have few opportunities for forcing matings since unreceptive females can avoid calling males.

Wing-flicking in Okanagana may have multiple functions. One function of wing flicking may be related to the presence of parasites. Sarcophagid flies (Colcondamyia auditrix ) are attracted to the calling songs of O. rimosa, although infection of O. canadensis is not known (Soper Shewell and Tyrell 1976). The flies lay eggs on the bodies of cicadas. Singing punctuated by bouts of wing-flicking may be an effective deterrent to parasitism because the wing flicking may disrupt the fly’s oviposition activities. Anecdotal evidence, which suggests that touching a male will provoke a bout of wing flicking, supports this explanation for the function of wing flicking (unpublished data). The occasional use of wing flicks by males to deter courtship by other males supports the view that wing flicks in O. canadensis and O. rimosa are general deterrents to close contact or unsolicited touching, either by parasites or by conspecifics. However, parasite

247 avoidance fails to fully explain why females should wing flick, because they produce no acoustical signals attractive to parasites. It is possible that females wing flick only when in proximity to a calling male, and thus when in jeopardy of being found by parasites attracted to the male’s calls.

Alternatively, male wing flicking may be a close-range, low-cost signal. In special circumstances when males and females find themselves nearby or in the same tree without having undergone the usual rapprochement sequence, wing flicks may allow rapprochement without requiring the male to sing extended, easily localized songs and reveal his presence to predators, parasites, or other males. It may benefit females to signal back, to alert males of their proximity and decrease the likelihood that males will make conspicuous acoustical signals. The wing flick interactions of males and females may thus be a quiet, unobtrusive way to facilitate rapprochement while lessening the danger of parasitism3 (see also Heller and von Helverson 1993). The acoustical component of wing flicking, or both acoustical and visual components of this behavior, as in Magicicada, may constitute the signal itself. Alternatively, since other homopterans are known to use vibrational signals (Hunt and Nault 1991, Hunt et al. 1992), and since the female’s final approach is by walking, she could potentially detect vibrations caused by the male’s wing flicking.

Contrasts between Okanagana and Magicicada behavior.

The male genitalia of O. canadensis and O. rimosa (Fig. 8.8) lack the hooks and claspers of Magicicada genitalia (Fig. 8.9; see also Moore and Alexander 1958). Male Okanagana genitalia have a slender adeagus with a scoop-like structure and a long, folded, threadlike extension. Males have been observed to fully penetrate the female genital tract

3 This hypothesis is not intended to be general for cicadas; see Popov 1981 for an account of Cicadetta sinuatipennis, in which tymbal sounds are hypothesized to function in close-range interactions, and wing flicks for long-range attraction.

248 within ten seconds of starting copulation. The inability of the “scoop” to penetrate the female as deeply as the adeagus makes the function of this device as a scoop questionable. If the scoop actually functions to remove sperm stored within the female spermatheca, then the combination of a scoop and long tube, through which ejaculate presumably passes and is deposited near the eggs, could be interpreted as a male adaptation to guard paternity in the face of polyandry. The lack of male genitalic hooks and the female’s control over sexual approach suggest that any degree of polyandry would most likely be under female control. During observations of mated females, none remated.

Alexander et al. (1997) described the influence of the nature of male-female rapprochement on the complexity of different stages of the sexual sequence. When males lure females, using songs, resources, or other advertisements of quality, and when females control sexual approach, courtship is typically complex and genitalia are simple. When males obtain mating by force or control sexual approach, courtship is minimal and genitalia are complex. In Okanagana, males remain stationary and use their songs to lure females, and their genitalia appear to be simple and lack hooks or claspers (Fig. 8.8). Because males are unlikely ever to encounter unreceptive females or to encounter competitive males during or immediately prior to mating, structures that could be adaptations for forcing matings, such as claspers, hooks, etc., are absent.

In contrast, Magicicada males actively search for females and use their calling songs to assist in search. Magicicada genitalia are a virtual battery of clasping, hooking, and ratcheting devices/appliances (Fig. 8.9). Male Magicicada have several opportunities during the sexual sequence to employ force. First, males use their songs to cause females to reveal their locations with wing flick signals and then approach signaling females (Chapter 5). Searching and approaching males have opportunities to force matings on resistant or marginally receptive females, and may use complex genitalia to do so. Second, complex genitalia may be the result of intrasexual competition. The final stages of

249 Magicicada courtship are easily perceived and localized by other, potentially interloping males. A copulating pair is relatively immobile and easily harassed by other males, who attempt to insert their genitalia as well. Once a Magicicada male begins copulation, it is in his interests to hold on as tightly as possible to avoid displacement. Note that displacement may not necessarily be contrary to the female’s interests, since the most competitive male may well sire the most competitive sons. Finally, copulation duration in Magicicada is much longer than in Okanagana, and appears to be in excess of durations required for completeness (Chapter 4). It is possible that male Magicicada use their genitalia to forcibly prolong copulation durations contrary to female interests, perhaps as a form of mate- guarding (Chapter 3).

Hypotheses concerning evolution of the unique mating system of Magicicada.

Suppose that the non-periodical ancestor of periodical cicadas had mating behaviors more similar to those of present-day Okanagana than to those of modern Magicicada. Such an ancestor might have lacked mass mating aggregations, genitalic hooks and claspers, complex courtship, and in most respects been similar to present-day non-periodical cicadas. Females would have signaled their receptivity only by approaching stationary, calling males.

Although the exact mechanisms by which Magicicada became periodical and year classes (or “broods”) formed are unknown, the synergy between (1) the ability of long- lived insects with prime number life cycles to avoid predators’ numerical responses and (2) increasing population sizes (with “selfish herd” connotations) likely promoted periodicity (See Lloyd and Dybas 1966a, b, Hoppensteadt and Keller 1976, Bulmer 1977, Martin and Simon 1990, Heliövaara et al. 1994, Yoshimura 1997, Cox and Carlton 1998, for discussion of the evolution of periodicity). The changes leading to periodicity also would

250 have caused changes in the sexual environment faced by members of the primordial broods.

For cicadas with mating behaviors similar to Okanagana, the task of finding mates is comparatively simple: Adult Okanagana males and females likely only ever encounter receptive members of the opposite sex, because females fly to widely separated singing males. As Magicicada emergences came to involve greater and greater numbers of cicadas, males may have begun to benefit by searching for females instead of waiting for females to come to them. Increasing densities present in a given year would have made it difficult for unreceptive females to avoid the unsolicited attentions of males and at the same time made it easier for males to locate females without waiting for females to come to them. The intersexual conflict thus engendered may have been one factor leading to the complex courtship of Magicicada. Whereas Okanagana male courtship behaviors seem primarily to facilitate rapprochement, much of Magicicada singing and tactile behavior seems focused on persuading females to copulate once contact has been made.

At some point, not far removed in time from the evolution of periodicity, or perhaps contemporary with it, the tasks faced by female periodical cicadas became more complex as the M. -decim, M. -cassini, and M. -decula species came into existence and stable sympatry. The coexistence of ecologically similar, synchronized species would have forced females to require males to produce some form of species identification, the vestiges of which may remain in the differences in the Magicicada species’ male calling songs, regardless of whether those differences arose incidentally in allopatry/allochrony, or as a result of character displacement in sympatry/synchrony. Such a situation would have promoted the equivalent of intersexual arms races, as females evolved to resist and thwart male forcing behaviors and males evolved both more effective tools of persuasion (such as more distinctive songs) and more effective tools for forcing, such as genitalic claspers, hooks, and ridges. Because increasing densities would also have promoted more intense

251 male-male competition, males may have benefited from developing means of persuading or forcing slightly immature females to copulate, and they may have benefited from adaptations for protecting their paternity, perhaps by using prolonged copulations as a form of mate-guarding.

Paradoxically, as Magicicada emergences became denser and came to contain several species, it may have become more and more difficult for females to find receptive mates; a periodical cicada chorus contains teneral, mated, same-sex, conspecific, and heterospecific individuals, all of which could be inappropriate targets of male courtship. Thus, females, who have little to gain by delaying once ready to mate, may, in the past, have benefited from acting in ways to attract the attention of male conspecifics, such as by signaling in response to their calls. The female receptivity signal of periodical cicadas (Chapter 5) may have arisen under such a scenario. As receptive females began to signal males, the intrasexual competition faced by males would have become more severe, because males would have adopted the tactic of converging on signaling females. Genital hooks, claspers, and other appliances for facilitating rapid mating, preventing usurpation, and prolonging copulation would have been favored.

This account of the evolution of the unique mating system of periodical cicadas is prompted by the contrasts between Okanagana and Magicicada. It is not intended to be definitive, but rather to present some of the themes, such as male-female antagonistic coevolution, male-male competition, and mate search strategies, that may have been important in Magicicada mating system evolution. The next step in evaluating these themes is to search for commonalities between Magicicada and other species in which females signal in response to searching males (such as fireflies, Lloyd 1966, 1979, or another cicada, Tibicen pruinosa, pers. obs).

252 Literature Cited Alexander, R. D. 1960. Sound communication in Orthoptera and Cicadidae. In: W. E. Lanyon, W. N .Tavolga, eds, Animal Sounds and Communication. AIBS symposium series publication 7: 38- 92. Alexander, R. D., and T. E. Moore. 1958. Studies on the acoustical behavior of seventeen- year cicadas (Homoptera: Cicadidae: Magicicada). Ohio J. Sci . 58:107-127. Alexander, R. D., and T. E. Moore. 1962. The evolutionary relationships of 17-year and 13-year cicadas, and three new species. (Homoptera: Cicadidae, Magicicada). Misc. Publ. no. 121. Univ. of Michigan Museum of Zoology, Ann Arbor, MI . Alexander, R. D., D. C. Marshall, and J. R. Cooley, 1997. Evolutionary Perspectives on Insect Mating. Pp. 1- 31 in B. Crespi, J. Choe, eds. The evolution of mating systems in insects and arachnids . (Cambridge University Press). Alexander, R. D., A. F. Richards, D. C. Marshall, and J. R. Cooley (in prep.). The microdistributional relationships of 17- and 13- year cicadas in Illinois (Magicicada): Broods III, 1997, and XIX, 1998. Alexander, R. D., J. R. Cooley, and D. C. Marshall. (in prep.). A specialized fungal parasite (Massospora cicadina) modifies the behavior of periodical cicadas (Homoptera: Cicadidae: Magicicada). Bulmer, M. G. 1977. Periodical Insects. Am. Nat. 111: 1099- 1117. Cox., R. T., and C. E. Carlton. 1998. A commentary on prime numbers and life cycles of periodical cicadas. Am. Nat. 152: 162-164. Davis, W.T. 1919. Cicadas of the genera Okanagana, Tibicinoides, and Okanagodes, with descriptions of several new species. Journal Of The New York Entomological Society 27: 179- 223. Gwynne, D. T. 1987. Sex-biased predation and the risky mate-locating behaviour of male tick-tock cicadas (Homoptera: Cicadidae) Anim. Behav. 35: 571- 576. Heliövaara, K., R. Väisänen, and C. Simon. 1994. Evolutionary Ecology of periodical insects. TREE 9: 475- 480. Heller, K. G., and von Helverson, D. 1993. Calling behavior in bushcrickets of the genus Poecilimon with differing communication systems (Orthoptera: Tettigonioidea, Phaneropteridae). J. Ins. Behav. 6: 361- 377. Hoppensteadt, F. C., and J. B. Keller. 1976. Synchronization of periodical cicada emergences. Science 194: 335- 337. Huber, F., D. W. Wohlers, J. L. D. Williams, and T. E. Moore. 1979. Struktur und Funktion der Hörbahn von Singzilkaden. Verh. Dtsch. Zool. Ges . 1979: 212.

253 Hunt, R. E., and L. R. Nault. 1991. Roles of interplant movement, acoustic communication, ad phototaxix in mate-location behavior of the leafhoper Graminella nigrigfrons. Behav. Ecol. Sociobiol 28: 315-320. Hunt, R. E., J. P. Fox, and K. F. Haynes. 1992. Behavioral response of Graminella nigrifrons (Homoptera: Cicadellidae) to experimentally manipulated vibrational signals. J. Ins. Behav. 5: 1- 13. Karban, R. 1982. Increased reproductive success at high densities and predator satiation for periodical cicadas. Ecology 63: 321-328. Kirby, C. S., and H. G. McPhee. 1963. Observations of the cicada, Okanagana rimosa (Say). Canad. Dept. For Bi-Mon. Prog. Rep . 19 (5): 1 Lloyd, J. E. 1979. Sexual selection in luminescent beetles. Pp. 293- 342 in M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects . (Academic Press: New York). Lloyd, J. E., 1966. Studies on the flash communication system in Photinus fireflies. . Misc. Publ. no. 130. Univ. of Michigan Museum of Zoology, Ann Arbor, MI . Lloyd, M. and H. S. Dybas. 1966a. The periodical cicada problem. I. Population Ecology. Evolution 20: 133- 149. Lloyd, M. and H. S. Dybas. 1966b. The periodical cicada problem. II. Evolution. Evolution 20: 466- 505. Marlatt, C. L. 1923. The periodical cicada. U.S.D.A. Bureau of Entomology Bulletin 71, 183 pp. Marshall, D. C., J. R. Cooley, R. D. Alexander, and T. E. Moore. 1996. New records of Michigan cicadidae (Homoptera) with notes on the use of songs to monitor range changes. Great Lakes Ent. 29:165- 169. Martin, A. and C. Simon. 1990. Temporal variation in insect life cycles. Bioscience 40: 359- 367. Moore, T. E., and R. D. Alexander. 1958. The periodical cicada complex (Homoptera: Cicadidae). Proc. X Internat. Cong. Entomol. 1: 349- 355. Popov, A. V. 1981. Sound production and hearing in the cicada, Cicadetta sinuatipennis Osh. (Homoptera: Cicadidae). J. Comp. Physiol. 142: 271-280. Simon, C. M. 1979. The debut of the 17-year-old cicada. Natural History 88: 38-45. Soper, R. S. 1963. Massospora levispora, a new species of fungus pathogenic to the cicada Okanagana rimosa. Can. J. Bot. 41: 875- 878. Soper, R. S., A. J. Delyzer, and L. F. R. Smith 1976. The genus Massospora entomopathogenic for cicadas. Part II. Biology of Massospora levispora and its host Okanagana rimosa, with notes on Massospora cicadina on the periodical cicadas. Ann. Ent. Soc. Am. 69: 89- 95.

254 Soper, R. S., G. E. Shewell, and D. Tyrell. 1976. Colcondamyia auditrix nov. sp. (Diptera: Sarcophagidae), a parasite which is attracted by the mating song of its host, Okanagana rimosa (Homoptera: Cicadidae). The Canadian Entomologist 108: 61- 68. White, J., and M. Lloyd. 1983. A pathogenic fungus, Massospora cicadina Peck (Entomophthorales) in emerging nymphs of periodical cicadas (Homoptera: Cicadidae). Environmental Entomology 12: 1245- 1252. Williams, K. S., K. G. Smith, and F. M. Stephen. 1993. Emergence of 13-yr periodical cicadas (Cicadidae: Magicicada): phenology, mortality, and predator satiation. Ecology 74: 1143- 1152. Williams, K. S., and C. Simon. 1995. The ecology, behavior, and evolution of periodical cicadas. Ann. Rev. Entomol. 40: 269- 295 Yoshimura, J. 1997. The evolutionary origins of periodical cicadas during ice ages. Am. Nat . 149:112-124.

255 Table 8.1. Lengths (minutes) of Okanagana canadensis calling and silent bouts for 5 males.

Calling Silent At perch 4:40 1:32 4:40 15:09 4:28 17:18 9:45 4:32 12:47 0:45 0:34 6:32 3:20 8:23 43:06 1:05 1:42 0:59 0:23 1:30 2:19 1:02 0:34 2:41 0:42 0:32 4:44 2:33 0:34 1:21 0:23 0:11 0:25 1:09 1:26 0:40 0:51 1:26 2:29 0:45 Number of bouts 22 13 5 Average bout length (min:sec) 2:29 2:22 16:53 Standard deviation 3:30 2:27 15:30

256 Table 8.2. Mating durations in Okanagana and Magicicada septendecim. One O. canadensis mating lasted approximately 4 times as long as the average of the others, so two averages, one with this long mating, and one without, are calculated for this species.

Species Duration (min: sec) O. rimosa 19:00 17:00 22:09 17:04 Average 18:48 ± 2:25

O canadensis 80:00 19:06 15:09 23:18 Average 34:23 ± 30:35 Average w/o outlier 19:11 ± 4:05

Magicicada septendecim 195 315 195 330 295 442 167 297 235 235 235 195 310 500 60 124 305 111 354 295 124 331 440 440 310 64 220 444 263 295 354 360 360 250 263 310 65 526 148 91 310 60 405 Average (n= 43) 243:24 ± 121:48

257 Table 8.3. Responses of male cicadas to presentation of model cicada. Males were scored as responding to model if they directed sexual behaviors, such as extruding genitalia, mounting, or copulating, towards the model. In both species, males were more likely to respond positively to the model (X2 goodness of fit; null hypothesis equal likelihood of positive and negative response). Although, in O. rimosa, the likelihood positive and negative responses do not differ from 50% at the significance level of P £ 0.05, this is probably due to small sample size, since the response rates of the two species do not differ (Fisher’s Exact Test P £ 0.280).

Species Trials Responses Positive Negative X2 P O. canadensis 9 9 0 £ 0.05 O. rimosa 20 16 4 £ 0.10

258 Area included in 1996- 1998 O. rimosa and O. canadensis distribution surveys (Figures 2-4) 259

Location of 1998 O. rimosa study site (Figure 5)

Figure 8.1. Locations of study sites, Okanagana canadensis, O. rimosa 1996- 1998.

l l

l l

l

s s l l l l l l l l l s s l s l s Cedarville l l l l l l l l l l l l l l

l l l l l 260 l l l l l

Lake Huron

St. Ignace l l 0 20 40 Miles l l l l Lake Michigan

Figure 8.2. 1996 distribution map of O. rimosa and O. canadensis. Only roads and hydrologic features relevant to distributional study shown. s : O. rimosa; l : O. canadensis. Continuous sequences of symbols indicate continuous distribution of cicadas in appropriate habitat.

l l l

l l l l

l l l

s s l l l l l l l l l s l s l l s l s l Cedarville l l l l l l l l l l l l l l l 261

Lake Huron

l St. Ignace l l l 0 20 40 Miles l l l l l Lake Michigan

Figure 8.3. 1997 distribution map of O. rimosa and O. canadensis. Only roads and hydrologic features relevant to distributional study shown. s : O. rimosa; l : O. canadensis. Continuous sequences of symbols indicate continuous distribution of cicadas in appropriate habitat.

l l l l s s s s l s l l s s s s s l l l l s l l s l l l l s l l l s l s l l l l l l l l l s l l l s l l l s l l s l l s l l l l l l l l l l l s s s s l l s l s l s l l l s l l s l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l s l l l l l Cedarville l l l l l l l l s l l l l l l l l l l l l l l l l l l l l l l l l 262 l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l s l s l l l l l l l l l l l l l l l l l s l l s l l l l l l l l l

l s l l l l l Lake Huron

St. Ignace l l l 0 20 40 Miles l l l l l l l l l l l l l l l l s l l l l l l l l l l Lake Michigan

Figure 8.4. 1998 distribution map of O. rimosa and O. canadensis. Only roads and hydrologic features relevant to distributional study shown. s : O. rimosa; l : O. canadensis. Continuous sequences of symbols indicate continuous distribution of cicadas in appropriate habitat. Open field near UMBS streams laboratory

Douglas Lake 263 Douglas Lake Road

UMBS

Riggsville Road

3 0 3 6 Miles

Figure 8.5. Location of O. rimosa study site in 1998. 10

8

6

4 264

2 kHz

0.0 2.0 4.0 6.0 7.0 Time (s)

Figure 8.6. Sonogram of male Okanagana canadensis call. 10

8

6 265

4

2

kHz

0.0 2.0 4.0 6.0 8.0 10.0 12.0 Time (s)

Figure 8.7. Sonogram of male Okanagana rimosa call. 266

Figure 8.8. Sketches of lateral view of typical Okanagana canadensis male genitalia. Okanagana rimosa are similar. Sketches not to scale (Otte, Anderson).

Magicicada septendecim 267

Magicicada cassini Magicicada septendecula

Figure 8.9. Sketches of lateral view of typical Magicicada male genitalia. Sketches not to scale. (Otter, Anderson). APPENDICES

268 APPENDIX A

C++ code for mating system simulation. Subroutines separated by asterisks (* * * * * *).

param. h: // Constants for program

#define PopSize 15 //Size of male population #define attributes 5 //Number of male attributes in phenotype

#define maxthresh 100 //Maximum threshold #define maxgen 1000 //Maximum number of generations #define pickvalue 10 //Maximum value for random number

#define yes 1 // #define no 0 // (* * * * * * * *) header.h: // Header file //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

#include "param.h" //Sets some global parameters

//function prototypes for simulation void bubblesort(int arr[], int n); //sorts the array its passed void timeseed(void); //seeds random number generator int picker(int n); //picks a random # inspecified range void wipe (void); //wipes contents or an array void dump (void); //dumps male population to screen void choicer(void); //gives the user a choice of functions void inputs(void); //gets user inputs void output(void); //writes output to disk or screen void summarize(void); //provides summary information void simulate(void); //runs the simulation void iterate(void); //runs one iteration void generation(void); //runs one generation void fillfirstgen(void); //Fills first generation of phenotypes void fillnextgen(void); //Fills next generation of phenotypes void bestofn(void); //Chooses best of n males void threshold(void); //Chooses males exceeding a threshold struct male //male data structure { int sum;

269 int mate; int phenotype[attributes]; };

//data structure to hold user inputs struct userparameters { int threshold; int cutoff; int mutation; int mutvalue; char excel; char file; int iterations; int generations; char screen; };

//data structure to hold parameters at a given time struct timeparameters { int genernum; int crash; int iternum; int thdmated; float thdvalue; int bonmated; float bonvalue;

};

(* * * * * * * *) main.cp: // Main program loop. Initializes global variables, sets contents of variables/ //arrays to zero, and calls functions of the siimulation.

#include #include "header.h" //includes all function prototypes #include "param.h" //includes global constants

//***********************Declare some global variables*******************

//initialize population of males, a global array of structures. male thdmales[PopSize]; //male population undergoing threshold choice male bonmales[PopSize]; //male population undergoing bn choice

//initialize a bookkeeping array int bookeep[PopSize];

//Initialize two data structures to hold user inputs and present // iteration/generation values. userparameters userinputs; timeparameters present;

//Results summaries go in these arrays float thdsum[maxgen]; float bonsum[maxgen];

//Keep track of how many times threshold choice population crashes

270 //int tag; int thdcrash[maxgen]; int main(void) {

//********************Start the main program************************** timeseed(); wipe();

//for the first time through, you don't get a choice of what to do inputs(); //Get input parameters simulate(); // Run the simulation

//After the first time through, you get a choice. choicer(); //Run the loop return 0; //End the program }

(* * * * * * * *) bestofn.cpp: //function definition for evaluating and marking best of n males //In case of ties, all typing males contribute to next generation

#include #include #include "header.h" #include "param.h" extern userparameters userinputs; extern male bonmales[PopSize]; extern int bookeep[PopSize]; extern timeparameters present; void bestofn(void) { present.bonmated = 0; present.bonvalue = 0; for (int a = 0; a < PopSize; a++) //Test male to see whether he passes { bonmales[a].sum = 0; for (int b = 0; b < attributes; b++) { bonmales[a].sum = (bonmales[a].sum + bonmales[a].phenotype[b]); } bonmales[a].mate = no; bookeep[a] = bonmales[a].sum; } bubblesort(bookeep, PopSize);

for (a = 0; a < PopSize; a++) //take the top n { //males and mark if (bonmales[a].sum >= bookeep[(PopSize - userinputs.cutoff)]) //them as mated

271 { for (int n = 0; n < userinputs.cutoff; n++) if (bonmales[a].sum >= bookeep[((PopSize - userinputs.cutoff) + n)]) bonmales[a].mate = (yes + n); //Males are ranked //bonmales[a].mate = yes; //High numbers are high quality present.bonmated = present.bonmated + yes; present.bonvalue = (present.bonvalue + bonmales[a].sum); } } present.bonvalue = (present.bonvalue/present.bonmated); }

(* * * * * * * *) bubblesort.cpp: //function definition for c version of bubblesort //must be passed and array and the size of the array

#include #include #include #include "header.h" #include "param.h" void bubblesort(int arr[], int n)

{ //note that because C arrays start at zero, counter i must go down //from arraysize-1 to 0 instead of from arraysize to 1.

for (int i = (n-1); i >= 0; i = (i-1)) for (int j = 0; j < i; j++) if (arr[j] > arr[j+1]) { int k = arr[j]; arr[j] = arr[j+1]; arr[j+1] = k; }

} (* * * * * * * *) choicer.cpp: // choicer-- runs the loop in which the simulation occurs. Gives the user a choice // of entering new parameters, using already entered parameteres, or stopping. //First simulation occurs in Main outside this loop; then Choicer is called and // you get a choice of whether to continue or not

#include #include // for exit() #include "header.h" #include "param.h" void choicer() { int select;

272 extern userparameters userinputs; //only here to provide "WORKING" display on screen. cout << "\n"; cout << "\n"; cout << "Would you like to "; cout << "\n"; cout << " (1) run a simulation with new parameters? "; cout << "\n"; cout << " (2) run a simulation using previous parameters? "; cout << "\n"; cout << " (3) end the simulations? "; cout << "\n"; cin >> select; cout << "\n"; cout << "\n"; if (select == 1) //Run simulation with new input parameters { inputs(); //Get input parameters if (!(userinputs.screen == 121)) { cout << "\n"; cout << " W O R K I N G "; cout << "\n"; } simulate(); //Run the simulation choicer(); //Choose what to do } if (select == 2) //Run simulation with old parameters { if (!(userinputs.screen == 121)) { cout << "\n"; cout << " W O R K I N G "; cout << "\n"; } simulate(); //Run the simulation choicer(); //Choose what to do } if (select == 3) //Stop the simulations { select = select; //Basically nonsense-- just pop back out to Main. }

}

(* * * * * * * *) fillfirstgen.cpp: //Algorithm for filling first generation with randomly generated //phenotypic attributes. Best of N males and threshold males //are filled identically.

#include #include "header.h" #include "param.h"

273 extern male bonmales[PopSize]; extern male thdmales[PopSize]; void fillfirstgen(void) { for (int a = 0; a < PopSize; a++) //ratchet through each male { for (int b = 0; b < attributes; b++) //ratchet through each attribute { bonmales[a].phenotype[b] = picker(pickvalue); //fill bn males thdmales[a].phenotype[b] = bonmales[a].phenotype[b]; //fill thd males the same } }

}

(* * * * * * * *) fillnextgen.cpp: //Algorithm for using mating status in present generation to determine //phenotypic attributes of next generation

#include #include "header.h" #include "param.h" extern male bonmales[PopSize]; extern male thdmales[PopSize]; extern timeparameters present; extern userparameters userinputs; extern int tag; void fillnextgen(void) { int proportion = 0; //used in setting up bonmales proportions int counter = 0; //Used in setting up bonmales proportions male holder[PopSize]; //Temporary array for holding thd or bon males

//Clear holder array************************************************************ for (int a = 0; a < PopSize; a++) //ratchet through each male { holder[a].sum = 0; //Make sure holder is empty holder[a].mate = 0; for (int b = 0; b < attributes; b++) //ratchet through each attribute { holder[a].phenotype[b] = 0; } }

//Transfer Best of N male data to holder*********************************************** for (a = 0; a < PopSize; a++) //ratchet through each male { holder[a].sum = bonmales[a].sum; //transfer to storage bonmales[a].sum = 0; //wipe to zero

274 holder[a].mate = bonmales[a].mate; //transfer to storage bonmales[a].mate = no; //wipe to zero for (int b = 0; b < attributes; b++) //ratchet through each attribute { holder[a].phenotype[b] = bonmales[a].phenotype[b]; //transfer to storage bonmales[a].phenotype[b] = 0; //wipe to zero } }

//Transfer successful Best of N male data back into population array****************** if (present.bonmated > 0) { for (a = 0; a < PopSize; a++) proportion = proportion + holder[a].mate;

proportion = (PopSize/proportion); counter = 0;

for (a = 0; a < PopSize; a++) //Fill next generation with attributes { if (holder[a].mate >= yes) //of successful males { for (int c = (counter* proportion * holder[a].mate); c < PopSize; c++) { for (int b = 0; b < attributes; b++) //ratchet through each attribute { bonmales[c].phenotype[b] = holder[a].phenotype[b]; //transfer successful attributes } counter = (counter +1); } } } }

//Clear holder array************************************************************ for (a = 0; a < PopSize; a++) //ratchet through each male { holder[a].sum = 0; //Make sure holder is empty holder[a].mate = 0; for (int b = 0; b < attributes; b++) //ratchet through each attribute { holder[a].phenotype[b] = 0; } }

//Transfer threshold Choice Male data into holder************************************** for (a = 0; a < PopSize; a++) //ratchet through each male { holder[a].sum = thdmales[a].sum; //transfer to storage thdmales[a].sum = 0; //wipe to zero holder[a].mate = thdmales[a].mate; //transfer to storage

275 thdmales[a].mate = no; //wipe to zero for (int b = 0; b < attributes; b++) //ratchet through each attribute { holder[a].phenotype[b] = thdmales[a].phenotype[b]; //transfer to storage thdmales[a].phenotype[b] = 0; //wipe to zero } }

//Transfer successful threshold choice male data back into population********************* if (present.thdmated > 0) { proportion = (PopSize/present.thdmated); counter = 0;

for (a = 0; a < PopSize; a++) //Fill next generation with attributes { if (holder[a].mate == yes) //of successful males { for (int c = (counter* proportion); c < PopSize; c++) { for (int b = 0; b < attributes; b++) //ratchet through each attribute { thdmales[c].phenotype[b] = holder[a].phenotype[b]; //transfer successful attributes } counter = (counter +1); } } } }

//Handle Mutation Events ************************************************************ //Note: Attributes are sampled with replacement, so the number of mutations //will be (1 < mutations < userinputs.mutations) for (a = 0; a < PopSize; a++) //ratchet through each male { for (int b = 0; b < userinputs.mutation; b++) //number of allowable mutations { //For each male, pick a random attribute and mutate it //to a random integer bonmales[a].phenotype[picker(attributes)] = picker(userinputs.mutvalue); if (present.thdmated > 0) thdmales[a].phenotype[picker(attributes)] = picker(userinputs.mutvalue); } }

} (* * * * * * * *) generation.cpp: //One generation

#include

276 #include "header.h" #include "param.h" extern float thdsum[maxgen]; extern float bonsum[maxgen]; extern timeparameters present; extern userparameters userinputs; void generation(void) { bestofn(); threshold();

//commented out lines are for old way of calculating cumulative totals. //thdsum[present.genernum] = (thdsum[present.genernum] + present.thdvalue); //bonsum[present.genernum] = (bonsum[present.genernum] + present.bonvalue); thdsum[present.genernum] = (thdsum[present.genernum] + (present.thdvalue/userinputs.iterations)); bonsum[present.genernum] = (bonsum[present.genernum] + (present.bonvalue/userinputs.iterations));

//dump(); } (* * * * * * * *) inputs.cpp: //This function provides prompts for user inputs and writes summary information //to a text file, if desired. Most inputs are checked upon entry and set //to a default value if user input exceeds acceptable value.

#include #include #include // for exit() #include "header.h" #include "param.h" const char * file3 = "data.txt"; const char * file4 = "data.xl"; const int Len = 40; extern userparameters userinputs; extern timeparameters present; extern male bonmales[PopSize]; extern male thdmales[PopSize]; extern int bookeep[PopSize]; void inputs(void) { //*****************make sure all variables are empty!************* for (int a = 0; a < PopSize; a++) { bookeep[a] = 0; thdmales[a].sum = 0; bonmales[a].sum = 0; thdmales[a].mate = 0; bonmales[a].mate = 0;

277 for (int b = 0; b < attributes; b++) { thdmales[a].phenotype[b] = 0; bonmales[a].phenotype[b] = 0; } } present.genernum = 0; present.iternum = 0; present.thdmated = 0; present.thdvalue = 0; present.bonmated = 0; present.bonvalue = 0; userinputs.threshold = 0; userinputs.cutoff = 0; userinputs.excel = 0; userinputs.file = 0; userinputs.iterations = 0; userinputs.generations = 0; userinputs.screen = 0;

//User provides information about how many Iterations cout << "\n"; cout << " SIMULATION PARAMETERS "; cout << "\n"; cout << "Please enter the number of iterations"; cout << "\n"; cin >> userinputs.iterations; cout << "\n"; cout << "\n"; cout << "Please enter the number of generations in each iteration"; cout << "\n"; cout << "(must be between 0 and "; cout << maxgen; cout << " )"; cout << "\n"; cin >> userinputs.generations; if (userinputs.generations < 0) userinputs.generations = 0; if (userinputs.generations > maxgen) (userinputs.generations = maxgen); cout << "\n";

cout << "\n"; cout << " THRESHOLD CHOICE PARAMETERS "; cout << "\n"; cout << "Please enter a number between 0 and "; cout << maxthresh; cout << "\n"; cout << " to represent the threshold below which males are unacceptable:"; cout << "\n"; cin >> userinputs.threshold; if (userinputs.threshold < 0) userinputs.threshold = 0; if (userinputs.threshold > maxthresh) (userinputs.threshold = maxthresh);

278 cout << "\n"; cout << "\n"; cout << " BEST-OF-N CHOICE PARAMETERS "; cout << "\n"; cout << "Please enter the number of males who will be allowed to mate"; cout << "\n"; cout << "(must be between 0 and "; cout << PopSize; cout << " )"; cout << "\n"; cin >> userinputs.cutoff; if (userinputs.cutoff < 0) userinputs.cutoff = 0; if (userinputs.cutoff > PopSize) userinputs.cutoff = PopSize; cout << "\n";

cout << "\n"; cout << " MUTATION PARAMETERS "; cout << "\n"; cout << "There are "; cout << attributes; cout << " inheritable attributes."; cout << "\n"; cout << " Enter maximum number of attributes that will mutate each generation: "; cin >> userinputs.mutation; if (userinputs.mutation < 0) userinputs.mutation = 0; if (userinputs.mutation > attributes) userinputs.mutation = attributes; cout << "\n"; cout << "At the beginning of the simulation, phenotype attribute values "; cout << "\n"; cout << " are randomly set to be 0 <= attribute < "; cout << pickvalue; cout << "\n"; cout << "Please enter a value between 0 and "; cout << maxthresh; cout << "\n"; cout << " so that mutations will be 0 <= mutation < value: "; cin >> userinputs.mutvalue; if (userinputs.mutvalue < 1) userinputs.mutvalue = 1; if (userinputs.mutvalue > maxthresh) userinputs.mutvalue = maxthresh; cout << "\n";

//Give the user a choice to write output to a file or screen. cout << "\n"; cout << " OUTPUT OPTIONS "; cout << "\n"; cout << ("Would you like to write the output to a text file \n"); cout << ("called 'data.txt' (y/n) ? "); cin >> userinputs.file; cout << "\n";

279 cout << ("Would you like to write the output to \n"); cout<<("an Excel compatible, tab-delimited text file \n"); cout<<("called 'data.xl' (y/n) ? "); cin >> userinputs.excel; cout << "\n"; cout << ("Would you like to write the output to the screen \n"); cout << ("(this will slow down the simulation) (y/n) ? "); cin >> userinputs.screen; cout << "\n";

//If file output is chosen, // Write summary data to output data.txt, the text file if (userinputs.file == 121) { ofstream fout3(file3, ios::out | ios::app); if (!fout3.good()) // or if (!fin) { cerr << "Can't open " << file3 << " file for output.\n"; exit(1); }

fout3 << "\n"; fout3 << "Iterations in this simulation: "; fout3 << userinputs.iterations;

fout3 << "\n"; fout3 << "Generations in each iteration: "; fout3 << userinputs.generations;

fout3 << "\n"; fout3 << "There are "; fout3 << PopSize; fout3 << " males in each population.";

fout3 << "\n"; fout3 << "In threshold choice, males below this cutoff are rejected: "; fout3 << userinputs.threshold;

fout3 << "\n"; fout3 << "Maximum threshold is: "; fout3 << maxthresh;

fout3 << "\n"; fout3 << "In Best-of-N choice, the top "; fout3 << userinputs.cutoff; fout3 << " males are allowed to mate.";

fout3 << "\n"; fout3 << "Number of attributes in each phenotype: "; fout3 << attributes;

fout3 << "\n"; fout3 << "Number of phenotypic attributes allowed to mutate in each generation: "; fout3 << userinputs.mutation;

fout3 << "\n"; fout3 << "Starting value for phenotypic attributes 0 <= attribute < "; fout3 << pickvalue;

fout3 << "\n";

280 fout3 << "Mutations to phenotypic attributes 0 <= mutation < "; fout3 << userinputs.mutvalue;

fout3 << "\n";

fout3.close(); }

}

(* * * * * * * *) iteration.cpp: //One iteration

#include #include "header.h" #include "param.h" extern userparameters userinputs; extern timeparameters present; //extern int tag; void iterate(void) { wipe(); fillfirstgen(); present.genernum = 0; //tag = 0; for (int b = 0; b < userinputs.generations; b++) { generation(); //output(); fillnextgen(); present.genernum = (present.genernum +1); } }

(* * * * * * * *) picker.cpp: //function definition for picker, which picks a random integer between -1 and n //Uses modulus operator to scale the random number.

#include #include #include "header.h" #include "param.h" int picker(int n) { return rand() % n; }

(* * * * * * * *) simulate.cpp: //Runs the simulation

281 #include #include "header.h" #include "param.h" extern userparameters userinputs; extern timeparameters present; extern int thdcrash[maxgen]; extern thdsum[maxgen]; extern bonsum[maxgen]; void simulate(void) { for (int aa = 0; aa < maxgen; aa++) //Clear out summary information { thdsum[aa] = 0; bonsum[aa] = 0; thdcrash[aa] = 0; } present.iternum = 0; for (int a = 0; a < userinputs.iterations; a++) { iterate(); present.iternum = (present.iternum +1); } summarize(); }

(* * * * * * * *) summarize.cpp: //Provide runtime screen and/or file output. Screen output slows the simulation; //if screen output is not chosen, the procedure Simulate reports only //the number of generations until fixation. Two file outputs are possible; one //is a text file similar to screen output but containing input fields as well; the //other is a tab-delimited text file that can be read by spreadsheet, graphics, //or statistics packages. In Macintosh, since output files have no creator //type, output files can only be read for the first time by a text editor such as BBedit.

#include #include #include // for exit() #include "header.h" #include "param.h" #include const char * afile3 = "data.txt"; const char * afile4 = "data.xl"; const int Len = 40;

extern timeparameters present; extern int thdcrash[maxgen]; extern float thdsum[maxgen]; extern float bonsum[maxgen];

282 extern userparameters userinputs; void summarize() {

//SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS if (userinputs.screen == 121) {

cout << "\n"; cout.setf(ios::fixed, ios::floatfield); //force fixed point display cout << "\n";

for (int a = 0; a < userinputs.generations; a++) { cout << "Average value (thdcrashes thd bn) for generation "; cout << (a + 1); cout << ": "; cout << thdcrash[a]; cout << " "; if (a == 0) cout << (thdsum[a]); if (a > 0) //must adjust numbers for times population has crashed cout << ((thdsum[a]* userinputs.iterations)/ (userinputs.iterations - thdcrash[(a-1)])); cout << " "; cout << (bonsum[a]); cout << "\n"; } } cout << "\n"; cout << "\n";

//SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS

// Write data to output afile1, the text file if (userinputs.file == 121) { ofstream fout3(afile3, ios::out | ios::app); if (!fout3.good()) // or if (!fin) { cerr << "Can't open " << afile3 << " file for output.\n"; exit(1); } //FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF //Code between these F's is written to file fout3.setf(ios::fixed, ios::floatfield); //force fixed point display fout3 << "\n"; for (int b = 0; b < userinputs.generations; b++) { fout3 << "Average value (thdcrashes thd bn) for generation "; fout3 << (b + 1); fout3 << ": ";

283 fout3 << thdcrash[b]; fout3 << " "; if (b == 0) fout3 << (thdsum[b]); if (b > 0) //must adjust numbers for times population has crashed fout3 << ((thdsum[b]* userinputs.iterations)/ (userinputs.iterations - thdcrash[(b-1)])); fout3 << " "; fout3 << bonsum[b]; fout3 << "\n";

} fout3.close(); }

// Write data to output afile2, the Excel file if (userinputs.excel == 121) { ofstream fout4(afile4, ios::out | ios::app); if (!fout4.good()) // or if (!fin) { cerr << "Can't open " << afile4 << " file for output.\n"; exit(1); } fout4.setf(ios::fixed, ios::floatfield); //force fixed point display fout4 << "\n"; for (int c = 0; c < userinputs.generations; c++) { fout4 << (c + 1); fout4 << "\t"; fout4 << thdcrash[c]; fout4 << "\t"; if (c == 0) fout4 << (thdsum[c]); if (c > 0) //must adjust numbers for times population has crashed fout4 << ((thdsum[c]* userinputs.iterations)/ (userinputs.iterations - thdcrash[(c-1)])); fout4 << "\t"; fout4 << bonsum[c]; fout4 << "\n"; } int cumulative = no; //Set to "yes" to have cumulative counts of thd crashes. if (cumulative == no) for (int d = (maxgen - 1); d > 0; d = (d - 1)) thdcrash[d] = (thdcrash[d] - thdcrash[d-1]); for (c = 0; c < userinputs.generations; c++) { fout4 << thdcrash[c]; fout4 << "\n"; } fout4.close(); } //FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF

}

284 (* * * * * * * *) threshold.cpp: //function definition for evaluating whether males pass threshold and //marking ones that do.

#include #include #include "header.h" #include "param.h" extern userparameters userinputs; extern male thdmales[PopSize]; extern timeparameters present; extern int thdcrash[maxgen]; //extern int tag; void threshold(void) { present.thdmated = 0; //Set mating information to zero present.thdvalue = 0; //Set phenotypes of mated males to zero

for (int a = 0; a < PopSize; a++) //Calculate male phenotype { for (int b = 0; b < attributes; b++) { thdmales[a].sum = (thdmales[a].sum + thdmales[a].phenotype[b]); } thdmales[a].mate = no; }

for (a = 0; a < PopSize; a++) { if (thdmales[a].sum >= userinputs.threshold) //test each male { thdmales[a].mate = yes; (present.thdmated = present.thdmated + yes); present.thdvalue = (present.thdvalue + thdmales[a].sum); } }

if (present.thdmated == 0) { //tag = yes; thdcrash[present.genernum] = (thdcrash[present.genernum] + 1); }

if (present.thdmated == 0) for (a = 0; a < PopSize; a++) { thdmales[a].sum = 0; thdmales[a].mate = 0; for (int b = 0; b < attributes; b++) thdmales[a].phenotype[b] = 0; }

if (present.thdmated > 0) present.thdvalue = (present.thdvalue/present.thdmated);

285 }

(* * * * * * * *) timeseed.cpp: //Seeds random number generator using internal clock

#include #include #include "header.h" //includes all function prototypes #include "param.h" //includes global constants void timeseed(void) { time_t t; srand((unsigned) time(&t)); //Seed Random number generator //srand(time(0)); //Alternative seed Random number generator

}

(* * * * * * * *) wipe.cpp: // Wipes variables clean

#include #include "header.h" //includes all function prototypes #include "param.h" //includes global constants extern userparameters userinputs; extern timeparameters present; extern male bonmales[PopSize]; extern male thdmales[PopSize]; extern int bookeep[PopSize]; void wipe(void) { present.genernum = 0; present.thdmated = 0; present.thdvalue = 0; present.bonmated = 0; present.bonvalue = 0; present.crash = 0; for (int a = 0; a < PopSize; a++) { bookeep[a] = 0; thdmales[a].sum = 0; bonmales[a].sum = 0; thdmales[a].mate = 0; bonmales[a].mate = 0; for (int b = 0; b < attributes; b++) { thdmales[a].phenotype[b] = 0; bonmales[a].phenotype[b] = 0; } } }

286 APPENDIX B

Pascal program code for comparing experimental cage densities to densities of free cicadas observed on cage-sized branches. (Think Lightspeed Pascal for the Macintosh)

program Cage; const ObservationNumber = 39; var FarmScan: array[0..39] of integer; greaterdiff, iterations, randomnumber, TargetDensity: integer; Percgreaterdiff: real; {* Seeds Random Number Generator} procedure TIMESEED; begin GetDateTime(randseed); end; procedure Welcome; begin writeln('Welcome to Resampler. You can run up to 10,000 resamplings'); writeln(' of female cicada branch scan data. '); end; {*Picks a random number, scales it to the appropriate size} procedure PICKRAND; begin randomnumber := (Trunc(((ABS(Random)) / 32768) * ObservationNumber)); {writeln('Picked slot ', randomnumber);} end; procedure HardData; begin FarmScan[0] := 7; FarmScan[1] := 12; FarmScan[2] := 14; FarmScan[3] := 7; FarmScan[4] := 7; FarmScan[5] := 11; FarmScan[6] := 16; FarmScan[7] := 22; FarmScan[8] := 17; FarmScan[9] := 22; FarmScan[10] := 6; FarmScan[11] := 9; FarmScan[12] := 12; FarmScan[13] := 13; FarmScan[14] := 12; FarmScan[15] := 27; FarmScan[16] := 3; FarmScan[17] := 11; FarmScan[18] := 10; FarmScan[19] := 13; FarmScan[20] := 17;

287 FarmScan[21] := 26; FarmScan[22] := 7; FarmScan[23] := 15; FarmScan[24] := 16; FarmScan[25] := 17; FarmScan[26] := 4; FarmScan[27] := 11; FarmScan[28] := 9; FarmScan[29] := 16; FarmScan[30] := 24; FarmScan[31] := 15; FarmScan[32] := 5; FarmScan[33] := 13; FarmScan[34] := 10; FarmScan[35] := 8; FarmScan[36] := 16; FarmScan[37] := 11; FarmScan[38] := 7; FarmScan[39] := 12; end; {Gets Input parameters} procedure INPUTS; begin writeln('Please provide the target density of cicadas'); read(TargetDensity); writeln('Please provide the number of iterations'); writeln('of the simulation , maximum 10000 : '); read(iterations); end; {Checks Input parameters} procedure CHECKVAR; begin if (iterations > 10000) or (iterations < 1) then begin writeln('Sorry, but iterations must be an integer less than 10000'); writeln('program has been terminated'); HALT; end; end; procedure Resample; var counter: integer; begin pickrand; {writeln('Picked number ', farmscan[randomnumber]);} if (ABS((12.6 - targetdensity)) < ABS((farmscan[randomnumber] - 12.6))) then greaterdiff := (greaterdiff + 1); end;

procedure iterate; var counter: integer; begin greaterdiff := 0; for counter := 1 to iterations do begin resample; end;

288 end; procedure MakeDistribution; begin Percgreaterdiff := greaterdiff / iterations; writeln(' '); writeln('percent of ', iterations, ' iterations'); writeln('in which picked number was farther from mean of 12.6 than'); writeln('target value = ', targetdensity); writeln(' '); writeln(Percgreaterdiff); end; {The shell of the program} procedure MAKEDECISION; var choicer: integer; begin writeln(' '); writeln('Would you like to:'); writeln(' (1) Resample ?'); writeln(' (2) End this session?'); writeln(' '); writeln('Choose (1) or (2) please'); read(choicer); if choicer = 1 then begin Timeseed; Welcome; inputs; checkvar; HardData; iterate; MakeDistribution; Makedecision; end; if choicer = 2 then HALT; if (choicer < 1) or (choicer > 2) then begin writeln('Hey there are only two choices here!'); Makedecision; end; end; begin MakeDecision; end.

289 APPENDIX C

Of the five bag cage experiments performed in 1995, in only two were females allowed to remate. Both were male silencing experiments designed to assess the influence of male courtship song on mating success. Unmated females were given equal numbers of normal and silenced males with which to mate.

Silencing experiment I 8 unmated females in each of 3 cages. 11 Days Also in each cage 8 normal males, 8 silenced males. 15/28 first matings by silenced males, 13/28 by unmanipulated males. 28 total matings, 6 rematings. 2 cages with 2 rematings each, 1 cage with 4 rematings. 1 female no matings.

Silencing experiment II 4 unmated females into each of 4 cages. 11 Days Also in each cage 4 normal and 4 silenced males 7/14 first matings by normal males 7/14 by silenced males. 14 total matings, 3 rematings. 3 cages with 1 remating each. 2 females no matings.

Male quality, at least as affected by the silencing treatment, is irrelevant because the treatment had no effect on male mating success. For the purposes of the resampling analysis, these experiments can be treated as bag cages in which females had the opportunity to mate multiply. Female remating frequencies in these cages can be compared to remating frequencies in the large flight cage. Although the flight cage data covers a period of 12 days, and the bag cage experiments ran for 11 days, these time periods are comparable, since the number of days is less important than day quality (in terms of weather, etc.). In any case, all but one mating in the flight cage occurred before the cicada’s tenth day from tenerality, well before the end of the flight and bag cage experiments.

290 APPENDIX D

C++ code to evaluate remating frequencies in large and small cages. Subroutines separated by asterisks (************).

****************************** header.h

#include "param.h" //Sets some global parameters

//function prototypes for simulation void choicer(void); //gives the user a choice of functions void simulate(void); //The simulation void inputs(void); //gets user inputs void iterate(void); //iterates the simulation int picker(int n); // pick a random integer between 0 and n void clearfemales(void); //clears female data array void fillfemales(void); //fills female data array with plug info. void output(void); //writes output to disk or screen

//data structure to hold user inputs struct userparameters { char full; char excel; int iterations; char file; char screen;

};

***************************************** param.h

// Constants for program, based on 1995 cage density experiments

#define Maxfemales 78 //Size of female population #define Totalremate 13 //Number of females who remated

************************************** choicer.cp

// choicer-- runs the loop in which the simulation occurs. Gives the //user a choice of entering new parameters, using already entered //parameters, or stopping. First simulation occurs in Main outside this //loop; then Choicer is called and you get a choice of whether to //continue or not

#include #include // for exit()

291 #include "header.h" #include "param.h" extern userparameters userinputs; extern int tally; extern int ptile; extern int iternum; void choicer() { int select;

cout << "\n"; cout << "\n"; cout << "Would you like to "; cout << "\n"; cout << " (1) run a simulation with new parameters? "; cout << "\n"; cout << " (2) run a simulation using previous parameters? "; cout << "\n"; cout << " (3) end the simulations? "; cout << "\n"; cin >> select; cout << "\n"; cout << "\n"; if (select == 1) //Run simulation with new input parameters { ptile = 0; tally = 0; iternum =1; inputs(); //Get input parameters iterate(); //Run the simulation choicer(); //Choose what to do } if (select == 2) //Run simulation with old parameters { ptile = 0; tally = 0; iternum =1; iterate(); //Run the simulation choicer(); //Choose what to do } if (select == 3) //Stop the simulations { select = select; //Basically nonsense-- just pop back out }

}

*********************** clearfemales.cp

#include #include "header.h"

292 #include "param.h" extern int females[Maxfemales]; void clearfemales(void) { for (int i = 0; i < Maxfemales; i++) { females[i] = 0; } }

************************** fillfemales.cp

#include #include "header.h" #include "param.h" extern int females[Maxfemales]; void fillfemales(void) { for (int i = 0; i < Totalremate; i++) { females[i] = 1; } }

************************** inputs.cpp

//This function provides prompts for user inputs and writes summary //information to a text file, if desired. Most inputs are checked upon //entry and set to a default value if user input exceeds acceptable //value.

#include #include #include // for exit() #include "header.h" #include "param.h" const char * file1 = "output1.txt"; const char * file2 = "outXl.txt"; const int Len = 40; extern int females[Maxfemales]; extern userparameters userinputs; extern int TargetValue; extern int CageSize; void inputs(void) { cout << "\n"; cout << " Cage Size Effect Resampler "; cout << "\n"; cout << "\n"; cout << "Cage size effect simulation";

293 cout << "\n"; cout << "Based on 1995 Data"; cout << "\n";

//Tell user how many total females may be in the simulation-- Max //number is set in the param.h file. cout << "\n"; cout << " FEMALE POPULATION CHARACTERISTICS "; cout << "\n"; cout<< "There are "; cout << Maxfemales; cout << " females in the flight cage"; cout << "\n";

cout<< "A total of "; cout << Totalremate; cout << " females in the flight cage remated"; cout << "\n";

cout << "\n";

//User provides information about how many Iterations cout << "\n"; cout << " SIMULATION PARAMETERS "; cout << "\n"; cout << "Please enter the number of iterations"; cout << "\n"; cin >> userinputs.iterations; if (userinputs.iterations < 1) userinputs.iterations = 0; cout << "\n";

//User provides information about Small Cages cout << "\n"; cout << " Characteristics of Small Cages "; cout << "\n"; cout << "Please enter the number of females in a small cage"; cout << "\n"; cin >> CageSize; if (CageSize < 1) CageSize = 0; if (CageSize > Maxfemales) CageSize = Maxfemales; cout << "\n"; cout << "Please enter the observed number of rematings in the small cage"; cout << "\n"; cin >> TargetValue; if (TargetValue < 1) TargetValue = 0; if (TargetValue > CageSize) TargetValue = CageSize; cout << "\n";

294 //Give the user a choice to write output to a file or screen. cout << "\n"; cout << " OUTPUT OPTIONS "; cout << "\n"; cout << ("Would you like to write full output? \n"); cout << ("(instead of abbreviated output) (y/n) ? "); cin >> userinputs.full; cout << "\n"; cout << ("Would you like to write the output to a text file \n"); cout << ("called 'output1.txt' (y/n) ? "); cin >> userinputs.file; cout << "\n"; if (userinputs.full == 121) { cout << ("Would you like to write the output to \n"); cout<<("an Excel compatible, tab-delimited text file \n"); cout<<("called 'outXl.txt' (y/n) ? "); cin >> userinputs.excel; cout << "\n"; } cout << ("Would you like to write the output to the screen \n"); cout << ("(this will slow down the simulation) (y/n) ? "); cin >> userinputs.screen; cout << "\n";

//If file output is chosen, // Write summary data to output file1, the text file if (userinputs.file == 121) { ofstream fout(file1, ios::out | ios::app); if (!fout.good()) // or if (!fin) { cerr << "Can't open " << file1 << " file for output.\n"; exit(1); }

fout << "\n"; fout << "Iterations in this simulation: "; fout << userinputs.iterations; fout << "\n";

fout << "\n"; fout << "\n";

fout<< "There are "; fout << Maxfemales; fout << " females in the flight cage"; fout << "\n";

fout<< "A total of "; fout << Totalremate; fout << " females in the flight cage remated"; fout << "\n";

fout<< "There are ";

295 fout << CageSize; fout << " females in the small cage"; fout << "\n";

fout<< "A total of "; fout << TargetValue; fout << " females in the small cage remated"; fout << "\n";

fout << "\n"; fout << "\n";

fout.close(); }

} ******************************************* iterate.cp

#include #include #include "header.h" //includes all function prototypes #include "param.h" //includes global constants

extern userparameters userinputs; extern int iternum; void iterate() { for (int i = 0; i < userinputs.iterations; i++) { simulate(); iternum = iternum +1; }

} ********************************************* main.cp

//Simulation to analyze cage size effect data #include #include #include "header.h" //includes all function prototypes #include "param.h" //includes global constants

//initialize an integer to tally number of plugged females picked in each turn. int tally;

//initialize an integer to generate percentile int ptile;

//initialize an integer to hold size of small cage we're looking at int CageSize;

296 //initialize an integer to look at target value of matings in small cage int TargetValue;

//initialize array for females int females[Maxfemales];

//Initialize an integer to keep track of which iteration we're on //so that output can display the results properly int iternum =1;

//Declare two characters to control file output. char write1; char write2;

//Initialize two data structures to hold user inputs and present // iteration/generation values. userparameters userinputs; int main(void) {

//********************Start the main program**************************

//Clear out userparameters clearfemales(); fillfemales(); tally = 0; userinputs.iterations = 0; userinputs.excel = 0; userinputs.screen = 0; userinputs. file = 0; ptile= 0;

srand(time(0)); //Seed Random number generator

//for the first time through, you don't get a choice of what to do inputs(); //Get input parameters iterate(); //Run the simulation

//After the first time through, you get a choice. choicer(); //Run the loop return 0; //End the program

} *************************************************** output.cp

//Provide runtime screen and/or file output. Screen output slows the //simulation;if screen output is not chosen, the procedure Simulate //reports only the number of generations until fixation. Two file //outputs are possible; one is a text file similar to screen output but //containing input fields as well; the other is a tab-delimited text //file that can be read by spreadsheet, graphics,or statistics packages. //In Macintosh, since output files have no creator type, output files //can only be read for the first time by a text editor such as BBedit.

297 #include #include #include // for exit() #include "header.h" #include "param.h" #include const char * afile1 = "output1.txt"; const char * afile2 = "outXl.txt"; const int Len = 40; extern int females[Maxfemales]; extern userparameters userinputs; extern int tally; extern int ptile; extern int TargetValue; extern int iternum; void output() { //SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS float percentile; cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint); cout.precision(3); if (userinputs.screen == 121) { if (userinputs.full == 121) { cout << "\n"; cout << "\n"; cout << "ITERATION "; cout << (iternum); cout << "\n"; cout << "\n"; cout << "Number of rematings is "; cout << tally; cout << "\n"; } }

if (iternum == userinputs.iterations) { cout << "\n"; cout << "\n"; cout << "Percentage of iterations in which "; cout << TargetValue; cout << "\n"; cout << "or more rematings occurred: ";

percentile = ptile; //Tricky way this has to be //calculated to percentile = percentile/userinputs.iterations; //avoid implicit conversion

cout << percentile;

298 cout << "\n"; }

//SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS

// Write data to output afile1, the text file if (userinputs.file == 121) { ofstream fout(afile1, ios::out | ios::app); if (!fout.good()) // or if (!fin) { cerr << "Can't open " << afile1 << " file for output.\n"; exit(1); } fout.setf(ios::fixed, ios::floatfield); //force fixed point display //FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF //Code between these F's is written to file if (userinputs.full == 121) { fout << "\n"; fout << "\n"; fout << "ITERATION "; fout << (iternum); fout << "\n"; fout << "\n"; fout << "Number of rematings is "; fout << tally; fout << "\n";

} if (iternum == userinputs.iterations) { fout << "\n"; fout << "\n"; fout << "Percentage of iterations in which "; fout << TargetValue; fout << "\n"; fout << "or more rematings occurred: "; fout << percentile; fout << "\n"; }

fout.close();

} if (userinputs.full == 121) { // Write data to output afile2, the Excel file if (userinputs.excel == 121) { ofstream fout2(afile2, ios::out | ios::app); if (!fout2.good()) // or if (!fin) {

299 cerr << "Can't open " << afile2 << " file for output.\n"; exit(1); }

fout2 << "\n"; fout2 << (tally); fout2.close(); }

//FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF

}

}

****************************************** picker.cp

//function definition for picker, which picks a random integer between //-1 and n //Uses modulus operator to scale the random number.

#include #include #include "header.h" #include "param.h"

int picker(int n) { return rand() % n; }

********************************************* simulate.cp

#include #include #include "header.h" //includes all function prototypes #include "param.h" //includes global constants extern userparameters userinputs; extern int females[Maxfemales]; extern int tally; extern int ptile; extern int TargetValue; extern int CageSize; void simulate() { for (int i = 0; i < CageSize; i++)

300 { int pick = 0; pick = picker(Maxfemales); //Pick a random female tally = (tally + females[pick]); //Tally up number of rematings in this iteration }

if (tally >= TargetValue) //test number of rematings in this iteration ptile = (ptile +1); //prepare to set up percentiles output(); tally = 0;

}

301 APPENDIX E

C++ code for seminal plug resampling algorithm. Subroutines separated by asterisks (* * * * * * * *).

param.h: // Constants for program, based on 1996 Farmville data, //mating interruption experiment, //females interrupted after 1 or 2 hours, assayed for plugs, //and allowed to remate.

#define Maxfemales 36 //Size of female population #define Pluggedfemales 11 //Number of plugged females #define Totalremate 20 //Number of females who remated #define Pluggedremate 4 //Number of plugged females who remated

(* * * * * * * *)

header.h: #include "param.h" //Sets some global parameters

//function prototypes for simulation

void choicer(void); //gives the user a choice of functions void simulate(void); //The simulation void inputs(void); //gets user inputs void iterate(void); //iterates the simulation int picker(int n); // pick a random integer between 0 and n void clearfemales(void); //clears female data array void fillfemales(void); //fills female data array with plug info. void output(void); //writes output to disk or screen

//data structure to hold user inputs struct userparameters { char full; char excel; int iterations; char file; char screen;

}; (* * * * * * * *)

main.cp: //Simulation to analyze 1996 plugged female remating data. Based on //male polygyny simulation.

#include #include #include "header.h" //includes all function prototypes

302 #include "param.h" //includes global constants

//initialize an integer to tally number of plugged females picked in each turn. int tally;

//initialize an integer to generate percentile int ptile;

//initialize array for females int females[Maxfemales];

//Initialize an integer to keep track of which iteration we're on //so that output can display the results properly int iternum =1;

//Declare two characters to control file output. char write1; char write2;

//Initialize two data structures to hold user inputs and present // iteration/generation values. userparameters userinputs; int main(void) {

//********************Start the main program**************************

//Clear out userparameters clearfemales(); fillfemales(); tally = 0; userinputs.iterations = 0; userinputs.excel = 0; userinputs.screen = 0; userinputs. file = 0; ptile= 0;

srand(time(0)); //Seed Random number generator

//for the first time through, you don't get a choice of what to do inputs(); //Get input parameters iterate(); //Run the simulation

//After the first time through, you get a choice. choicer(); //Run the loop return 0; //End the program

}

(* * * * * * * *) inputs.cp: //This function provides prompts for user inputs and writes summary information //to a text file, if desired. Most inputs are checked upon entry and set

303 //to a default value if user input exceeds acceptable value.

#include #include #include // for exit() #include "header.h" #include "param.h" const char * file1 = "output1.txt"; const char * file2 = "outXl.txt"; const int Len = 40; extern int females[Maxfemales]; extern userparameters userinputs; void inputs(void) { cout << "\n"; cout << " PLUGGED FEMALE REMATING SIMULATION "; cout << "\n"; cout << "\n"; cout << "Plugged female remating simulation in which females"; cout << "\n"; cout << "whose first mating is interrupted are assayed for a"; cout << "\n"; cout << "sperm plug and allowed to remate."; cout << "\n"; cout << "Based on 1996 Farmville Data"; cout << "\n";

//Tell user how many total females may be in the simulation-- Max //number is set in the param.h file. cout << "\n"; cout << " FEMALE POPULATION CHARACTERISTICS "; cout << "\n"; cout<< "There are "; cout << Maxfemales; cout << " females in the population"; cout << "\n"; cout<< "There are "; cout << Pluggedfemales; cout << " plugged females in the population"; cout << "\n"; cout<< "A total of "; cout << Totalremate; cout << " females in the population will remate"; cout << "\n";

cout << "\n";

304 //User provides information about how many Iterations cout << "\n"; cout << " SIMULATION PARAMETERS "; cout << "\n"; cout << "Please enter the number of iterations"; cout << "\n"; cin >> userinputs.iterations; if (userinputs.iterations < 1) userinputs.iterations = 0; cout << "\n";

//Give the user a choice to write output to a file or screen. cout << "\n"; cout << " OUTPUT OPTIONS "; cout << "\n"; cout << ("Would you like to write full output? \n"); cout << ("(instead of abbreviated output) (y/n) ? "); cin >> userinputs.full; cout << "\n"; cout << ("Would you like to write the output to a text file \n"); cout << ("called 'output1.txt' (y/n) ? "); cin >> userinputs.file; cout << "\n"; if (userinputs.full == 121) { cout << ("Would you like to write the output to \n"); cout<<("an Excel compatible, tab-delimited text file \n"); cout<<("called 'outXl.txt' (y/n) ? "); cin >> userinputs.excel; cout << "\n"; } cout << ("Would you like to write the output to the screen \n"); cout << ("(this will slow down the simulation) (y/n) ? "); cin >> userinputs.screen; cout << "\n";

//If file output is chosen, // Write summary data to output file1, the text file if (userinputs.file == 121) { ofstream fout(file1, ios::out | ios::app); if (!fout.good()) // or if (!fin) { cerr << "Can't open " << file1 << " file for output.\n"; exit(1); }

fout << "\n"; fout << "Iterations in this simulation: "; fout << userinputs.iterations; fout << "\n";

fout << "\n"; fout << "\n";

305 fout<< "There are "; fout << Maxfemales; fout << " females in the population"; fout << "\n";

fout<< "There are "; fout << Pluggedfemales; fout << " plugged females in the population"; fout << "\n";

fout<< "A total of "; fout << Totalremate; fout << " females in the population will remate"; fout << "\n";

fout<< "A total of "; fout << Pluggedremate; fout << " plugged females in the 1996 population remated"; fout << "\n";

fout << "\n"; fout << "\n";

fout.close(); }

}

(* * * * * * * *) iterate.cp #include #include #include "header.h" //includes all function prototypes #include "param.h" //includes global constants

extern userparameters userinputs; extern int iternum; void iterate() { for (int i = 0; i < userinputs.iterations; i++) { simulate(); iternum = iternum +1; }

}

(* * * * * * * *) simulate.cp: #include #include #include "header.h" //includes all function prototypes

306 #include "param.h" //includes global constants extern userparameters userinputs; extern int females[Maxfemales]; extern int tally; extern int ptile; void simulate() { for (int i = 0; i < Totalremate; i++) { int pick = 0; pick = picker(Maxfemales); //Pick a random female tally = (tally + females[pick]); //Tally up number of matings in this iteration }

if (tally <= Pluggedremate) //test number of plugged matings in this iteration ptile = (ptile +1); //prepare to set up percentiles output(); tally = 0;

}

(* * * * * * * *) fillfemales.cp:

#include #include "header.h" #include "param.h" extern int females[Maxfemales]; void fillfemales(void) { for (int i = 0; i < Pluggedfemales; i++) { females[i] = 1; } }

(* * * * * * * *) clearfemales.cp:

#include #include "header.h" #include "param.h" extern int females[Maxfemales]; void clearfemales(void) {

307 for (int i = 0; i < Maxfemales; i++) { females[i] = 0; } }

(* * * * * * * *) picker.cp: //function definition for picker, which picks a random integer between -1 and n //Uses modulus operator to scale the random number.

#include #include #include "header.h" #include "param.h"

int picker(int n) { return rand() % n; }

(* * * * * * * *) choicer.cp: // choicer-- runs the loop in which the simulation occurs. //Gives the user a choice // of entering new parameters, using already entered parameters, or stopping. //First simulation occurs in Main outside this loop; then Choicer is called and // you get a choice of whether to continue or not

#include #include // for exit() #include "header.h" #include "param.h" extern userparameters userinputs; extern int tally; extern int ptile; extern int iternum; void choicer() { int select;

cout << "\n"; cout << "\n"; cout << "Would you like to "; cout << "\n"; cout << " (1) run a simulation with new parameters? "; cout << "\n"; cout << " (2) run a simulation using previous parameters? "; cout << "\n"; cout << " (3) end the simulations? "; cout << "\n";

308 cin >> select; cout << "\n"; cout << "\n"; if (select == 1) //Run simulation with new input parameters { ptile = 0; tally = 0; iternum =1; inputs(); //Get input parameters iterate(); //Run the simulation choicer(); //Choose what to do } if (select == 2) //Run simulation with old parameters { ptile = 0; tally = 0; iternum =1; iterate(); //Run the simulation choicer(); //Choose what to do } if (select == 3) //Stop the simulations { select = select; //Basically nonsense-- just pop back out to Main. }

}

(* * * * * * * *) output.cp: //Provide runtime screen and/or file output. Screen output slows the simulation; //if screen output is not chosen, the procedure Simulate reports only //the number of generations until fixation. Two file outputs are possible; one //is a text file similar to screen output but containing input fields as well; the //other is a tab-delimited text file that can be read by spreadsheet, graphics, //or statistics packages. In Macintosh, since output files have no creator //type, output files can only be read for the first time by a text editor such as //BBedit.

#include #include #include // for exit() #include "header.h" #include "param.h" #include const char * afile1 = "output1.txt"; const char * afile2 = "outXl.txt"; const int Len = 40; extern int females[Maxfemales]; extern userparameters userinputs; extern int tally; extern int ptile; extern int iternum;

309 void output() { //SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS float percentile; cout.setf(ios::fixed, ios::floatfield); cout.setf(ios::showpoint); cout.precision(3); if (userinputs.screen == 121) { if (userinputs.full == 121) { cout << "\n"; cout << "\n"; cout << "ITERATION "; cout << (iternum); cout << "\n"; cout << "\n"; cout << "Number of matings by plugged females is "; cout << tally; cout << "\n"; } }

if (iternum == userinputs.iterations) { cout << "\n"; cout << "\n"; cout << "Percentage of iterations in which "; cout << Pluggedremate; cout << "\n"; cout << "or fewer plugged females remated: ";

percentile = ptile; //Tricky way this has to be calculated to percentile = percentile/userinputs.iterations; //avoid implicit conversion

cout << percentile; cout << "\n"; }

//SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS

// Write data to output afile1, the text file if (userinputs.file == 121) { ofstream fout(afile1, ios::out | ios::app); if (!fout.good()) // or if (!fin) { cerr << "Can't open " << afile1 << " file for output.\n"; exit(1); } fout.setf(ios::fixed, ios::floatfield); //force fixed point display //FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF

310 //Code between these F's is written to file if (userinputs.full == 121) { fout << "\n"; fout << "\n"; fout << "ITERATION "; fout << (iternum); fout << "\n"; fout << "\n"; fout << "Number of matings by plugged females is "; fout << tally; fout << "\n";

} if (iternum == userinputs.iterations) { fout << "\n"; fout << "\n"; fout << "Percentage of iterations in which "; fout << Pluggedremate; fout << "\n"; fout << "or fewer plugged females remated: "; fout << percentile; fout << "\n"; }

fout.close();

} if (userinputs.full == 121) { // Write data to output afile2, the Excel file if (userinputs.excel == 121) { ofstream fout2(afile2, ios::out | ios::app); if (!fout2.good()) // or if (!fin) { cerr << "Can't open " << afile2 << " file for output.\n"; exit(1); }

fout2 << "\n"; //fout2 << (iternum); //fout2 << "\t"; //fout2 << (Maxfemales); //fout2 << "\t"; //fout2 << (Totalremate); //fout2 << "\t"; //fout2 << (Pluggedfemales); //fout2 << "\t"; fout2 << (tally); fout2.close(); }

311 //FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF

}

}

312 Appendix F. Male frequency collection, 1998 Brood XIX. Series Tape Side Counter Cicada (FC) Main Abdomen Frequency kHz

DJ98V 1 0:13:02 1 1.09 4 DJ98V 1 1:33:52 2 1.55 1 DJ98V 1 0:01:22 3 1.76 2 DJ98V 1 0:10:46 4 1.78 2 DJ98V 1 0:48:02 5 1.70 2 DJ98V 1 0:56:54 6 1.04 4 DJ98V 1 1:03:42 8 1.07 3 DJ98V 1 0:55:02 9 1.70 2 DJ98V 1 0:29:52 10 1.70 3 DJ98V 1 1:19:52 11 1.19 4 DJ98V 1 0:36:23 13 1.70 2 DJ98V 1 1:05:58 14 1.61 2 DJ98V 1 0:58:11 15 1.73 4 DJ98V 1 0:17:37 17 1.71 2 DJ98V 1 0:22:57 18 1.76 2 DJ98V 1 0:06:00 19 1.64 2 DJ98 1 B 96 20 1.75 2 DJ98 B 254 21 1.70 2 DJ98V 1 0:39:24 22 1.77 2 DJ98 2 A 199 23 1.69 2 DJ98 2 A 89 24 1.75 2 DJ98 2 A 183 25 1.64 2 DJ98 2 A 275 27 1.66 2 DJ98 B 111 28 1.70 4 DJ98 1 B 41 30 1.68 2 DJ98 B 211 31 1.69 2 DJ98 B 243 32 1.69 2 DJ98 2 A 288 32 1.74 2 DJ98 2 A 281 33 1.89 1 DJ98 B 222 34 1.78 2 DJ98 B 157 37 1.10 4 DJ98 B 197 38 1.89 3 DJ98 B 231 39 1.72 1 DJ98 B 180 40 1.72 2 DJ98 B 170 42 1.72 2 DJ98 1 B 62 43 1.64 2 DJ98 2 A 217 46 1.70 2 DJ98 B 140 49 1.70 2 DJ98 B 146 51 1.88 3 DJ98 2 A 243 52 1.68 2 DJ98V 1 0:31:56 53 1.73 2 DJ98 54 1.73 2 DJ98 B 126 55 1.65 2 DJ98V 1 0:25:50 59 1.71 1 DJ98 2 A 295 60 1.78 1 DJ98 2 B 185 62 1.70 2 DJ98 2 B 200 63 1.69 2 DJ98 2 B 174 64 1.65 2 DJ98 2 A 267 65 1.55 2

313 DJ98 2 B 72 66 1.73 2 DJ98 2 B 108 67 1.75 2 DJ98V 1 1:17:34 68 1.71 2 DJ98V 1 1:10:40 69 1.10 4 DJ98V 1 0:49:20 70 1.16 3 DJ98V 1 1:16:20 71 1.11 4 DJ98V 1 1:00:37 72 1.11 3 DJ98V 1 0:52:45 73 1.69 2 DJ98V 1 1:12:35 74 1.81 2 DJ98V 1 0:46:55 75 1.66 2 DJ98V 1 1:35:22 76 1.15 4 DJ98V 1 1:56:50 78 1.69 2 DJ98 3 A 195 81 DJ98 2 A 125 83 1.72 2 DJ98 2 A 209 84 1.76 2 DJ98 2 A 44 85 1.08 4 DJ98 2 A 77 86 1.71 2 DJ98 2 A 14 87 1.60 2 DJ98 88 4 DJ98 2 A 110 89 1.70 2 DJ98 2 B 226 90 1.65 2 DJ98 2 A 65 92 1.81 2 DJ98 2 A 249 93 1.68 2 DJ98 2 A 54 94 1.43 3 DJ98 2 A 190 95 1.79 2 DJ98 2 A 236 96 1.58 2 DJ98 2 A 229 97 1.69 2 DJ98 2 A 100 98 1.72 2 DJ98 2 A 22 99 1.71 2 DJ98 2 A 34 100 1.70 2 DJ98V 1 1:47:28 104 1.62 2 DJ98V 2 0:32:52 108 1.63 3 DJ98V 2 0:37:51 111 1.00 3 DJ98V 1 1:59:44 112 1.50 2 DJ98V 1 1:21:39 114 1.06 4 DJ98V 1 1:39:40 115 1.04 4 DJ98V 1 1:31:09 116 1.71 2 DJ98 2 B 66 117 1.72 2 DJ98 2 B 208 118 1.75 1 DJ98V 1 1:48:33 119 1.73 2 DJ98V 1 1:38:39 120 1.69 2 DJ98 2 B 168 121 1.76 2 DJ98V 1 1:57:48 122 1.72 2 DJ98 2 B 158 123 1.70 1 DJ98 2 B 232 124 1.72 2 DJ98V 1 1:55:32 125 1.70 2 DJ98 2 B 115 126 1.13 2 DJ98 2 B 84 127 1.77 2 DJ98V 1 1:28:47 128 1.78 2 DJ98V 1 1:44:12 129 1.61 2 DJ98 2 B 222 130 1.67 3 DJ98 2 B 100 131 1.71 2 DJ98 2 B 196 132 1.73 2 DJ98 2 B 191 133 1.76 2 DJ98 2 B 151 134 1.65 2

314 DJ98V 1 1:52:09 135 1.08 4 DJ98V 1 1:49:51 136 1.67 2 DJ98V 1 1:54:14 137 1.11 3 DJ98V 1 2:00:27 138 1.70 1 DJ98 2 B 90 139 1.10 4 DJ98 2 B 125 140 1.83 2 DJ98 3 A 197 200 1.66 3 DJ98 3 A 364 201 1.69 2 DJ98 3 A 235 202 1.77 2 DJ98 3 A 275 203 1.69 2 DJ98 3 A 265 204 1.65 2 DJ98V 2 1:15:35 205 1.72 2 DJ98V 2 0:51.22 206 1.62 2 DJ98 3 A 270 207 1.61 2 DJ98 3 A 347 209 1.03 4 DJ98 3 A 285 210 1.80 1 DJ98 3 A 241 212 1.63 2 DJ98V 2 1:12:30 213 1.69 2 DJ98V 214 2 DJ98V 2 0:48:02 215 1.77 2 DJ98 3 A 206 217 1.63 2 DJ98 3 A 371 218 1.61 2 DJ98 3 A 281 219 1.64 2 DJ98 3 A 360 220 1.60 2 DJ98 3 A 259 222 2 DJ98V 2 1:05:19 223 1.66 2 DJ98V 2 1:07:00 225 1.64 1 DJ98 3 A 220 227 1.74 2 DJ98V 2 0:43:37 228 1.10 4 DJ98V 2 1:16:02 229 1.72 2 DJ98V 2 0:41:27 230 1.70 3 DJ98V 2 1:14:16 231 1.14 4 DJ98 3 A 321 232 1.71 2 DJ98 3 A 253 233 1.78 2 DJ98 3 A 343 234 1.70 2 DJ98V 2 1:02:27 235 1.69 2 DJ98 3 A 305 236 1.84 3 DJ98 3 A 375 237 1.70 2 DJ98 3 A 212 238 1.83 1 DJ98V 2 0:45:37 239 1.13 4 DJ98 3 A 340 240 1.06 4 DJ98V 2 0:49:26 240 1.10 4 DJ98V 2 1:17:38 241 1.12 3 DJ98 3 A 231 242 1.69 2 DJ98V 2 0:40:12 243 1.75 2 DJ98V 2 0:51:55 244 1.79 2 DJ98 3 A 247 245 1.15 4 DJ98 3 A 328 246 1.72 2 DJ98V 2 1:10:12 247 1.63 2 DJ98 3 A 332 248 1.64 2 DJ98 3 A 356 250 1.84 2

315 Appendix G. Playback Experiments to females, 1998 Brood XIX. Teneral Date Cicada 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 Weighted Average Min Max Abdomen 05/13/98 n 1 2 0 0 0 0 0 0 0 0 0 0 0 0 1.067 1 1.1 3 05/15/98 m 4 3 4 4 4 1 1 0 0 0 0 0 0 0 1.238 1 1.1 05/15/98 o 0 0 0 0 0 0 0 0 0 0 1 0 0 0 2.000 2.0 1.6 05/15/98 p 0 0 0 2 4 2 3 3 4 4 2 2 1 1 1.736 1.3 2.3 05/16/98 1 0 0 0 0 0 0 0 1 2 1 0 0 0 0 1.800 1.7 1.9 1 05/16/98 2 0 0 0 0 0 2 1 3 3 3 2 1 0 0 1.793 1.5 2.1 3 05/16/98 3 4 4 5 3 2 1 0 0 0 0 0 0 0 0 1.189 1.0 1.5 3 05/16/98 4 0 0 1 1 1 1 1 2 0 1 0 0 0 0 1.538 1.2 1.9 1 05/16/98 5 0 0 0 0 1 1 0 1 0 0 1 0 0 0 1.650 1.4 2.0 3 05/16/98 6 0 0 0 1 2 2 2 2 4 2 1 0 1 0 1.700 1.3 2.2 2 05/16/98 7 1 4 5 7 8 8 7 5 5 4 2 1 1 0 1.522 1.0 2.2 2 05/16/98 8 0 0 0 3 4 3 7 3 6 4 4 2 0 0 1.692 1.3 2.1 4 05/16/98 9 0 0 0 0 0 2 1 3 7 7 4 4 0 0 1.857 1.5 2.1 3 05/16/98 10 0 0 0 0 1 2 3 3 2 1 1 0 0 0 1.677 1.4 2.0 2 05/16/98 11 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1.400 1.4 1.4 3 05/16/98 12 0 0 0 1 3 3 4 6 6 5 4 4 4 3 1.837 1.3 2.3 2 05/16/98 13 0 1 1 1 1 1 0 1 0 1 1 1 1 0 1.640 1.1 2.2 2 05/16/98 14 1 0 0 0 0 1 3 3 3 3 2 0 0 0 1.719 1.0 2.0 3 05/16/98 15 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1.500 1.5 1.5 1 05/16/98 16 3 4 4 4 2 1 0 0 0 0 0 0 0 0 1.206 1.0 1.5 4 05/18/98 3a 0 0 0 0 0 1 2 1 2 1 1 1 0 0 1.778 1.5 2.1 1 05/18/98 5a 0 0 0 0 0 0 0 0 1 1 2 3 3 1 2.082 1.8 2.3 3 05/18/98 8a 4 4 3 3 1 0 0 0 0 0 0 0 0 0 1.153 1.0 1.4 4 05/18/98 a 4 6 4 2 0 0 0 0 0 0 0 0 0 0 1.125 1.0 1.3 2 05/18/98 b 0 0 0 0 0 1 0 1 1 0 0 0 0 0 1.667 1.5 1.8 3 05/18/98 c 0 1 1 1 3 5 6 6 5 4 5 3 2 1 1.737 1.1 2.3 2 05/18/98 d 0 1 4 6 8 7 8 8 7 6 6 5 4 0 1.673 1.1 2.2 2 05/18/98 e 5 6 4 3 3 0 0 0 0 0 0 0 0 0 1.167 1.0 1.4 2 05/18/98 f 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1.000 1.0 1.0 05/18/98 g 8 8 6 7 7 5 2 2 1 2 1 1 0 0 1.334 1.0 2.1 3 05/18/98 m 1 2 4 4 3 0 0 0 0 0 0 0 0 0 1.243 1.0 1.4 4 05/18/98 o 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2.100 2.1 2.1 2 05/18/98 p 0 0 1 3 3 4 4 4 4 4 4 4 2 0 1.732 1.2 2.2 2 05/19/98 0 7 8 7 7 5 2 0 0 0 0 0 0 0 0 1.203 1.0 1.5 4 05/19/98 1 1 1 1 2 2 4 4 6 6 6 4 0 0 2 1.695 1.0 2.3 2 05/19/98 2 0 0 0 1 1 1 1 2 0 0 0 0 0 0 1.533 1.3 1.7 3 05/19/98 3 0 0 0 0 0 1 3 4 5 4 3 0 0 0 1.785 1.5 2.0 2 05/19/98 4 0 0 0 1 1 0 1 1 2 3 2 4 0 0 1.847 1.3 2.1 1 05/19/98 5 2 2 1 0 0 0 0 0 0 0 0 0 0 0 1.080 1.3 1.2 05/19/98 6 0 0 0 0 2 3 4 3 2 0 0 1 1 0 1.669 1.4 2.2 2 05/19/98 7 0 0 0 2 0 4 5 8 8 7 6 5 3 4 1.850 1.3 2.3 2 05/19/98 9 0 0 0 0 0 1 3 2 3 2 2 0 0 0 1.762 1.5 2.0 3 05/19/98 10 2 3 5 4 1 0 0 0 0 0 0 0 0 0 1.193 1.0 1.4 4 05/19/98 11 5 6 3 3 0 0 0 0 0 0 0 0 0 0 1.124 1.0 1.3 4 05/19/98 13 0 1 0 0 0 1 0 3 3 3 2 3 2 2 1.905 1.1 2.3 2 05/19/98 14 0 0 0 0 0 0 3 3 5 3 0 1 0 0 1.780 1.6 2.1 2 05/19/98 15 1 2 2 5 7 7 6 7 7 6 4 1 2 2 1.644 1.0 2.3 3 05/19/98 16 0 0 0 0 0 1 1 1 4 5 2 4 1 0 1.900 3 05/19/98 17 0 0 0 0 1 0 0 1 1 1 0 0 0 0 1.700 1.4 1.9 2 05/19/98 18 0 0 0 0 0 0 0 1 2 1 1 0 0 1 1.917 1.7 2.3 2 05/19/98 21 0 0 1 1 4 4 2 1 0 0 0 0 0 0 1.462 1.2 1.7 2 05/19/98 23 0 0 0 0 0 0 3 4 6 6 4 2 2 0 1.867 1.6 2.2 2 05/19/98 24 0 0 0 0 1 2 2 1 2 2 0 0 0 0 1.670 1.4 1.9 2 05/19/98 4 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1.750 05/19/98 4 0 0 0 0 0 1 0 1 1 1 0 0 0 0 1.725 05/19/98 4 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1.750 05/21/98 12 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1.400 05/21/98 3 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1.250 05/23/98 4 0 0 1 0 0 0 0 0 0 0 0 1 0 0 1.650 05/23/98 8 0 0 0 0 0 0 0 0 1 1 0 0 1 0 1.967 05/21/98 3 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1.400 05/23/98 4 0 0 0 0 0 0 0 0 0 1 1 0 0 0 1.950 05/23/98 4 0 0 0 0 0 0 0 0 0 1 1 0 0 0 1.950 05/23/98 8 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1.750 05/23/98 4 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1.750 05/31/98 A 2 0 4 2 2 1 0 0 0 0 0 0 0 0 1.245 1.0 1.5 05/31/98 C 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1.350 1.2 1.5 05/31/98 DD 1 1 2 2 2 1 0 0 0 0 0 0 0 0 1.267 1.0 1.5 05/31/98 F 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1.250 1.2 1.3 05/31/98 FF 0 1 2 2 1 0 0 0 0 0 0 0 0 0 1.250 1.1 1.4 05/31/98 JJ 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1.350 1.3 1.4 05/31/98 M 2 3 3 4 4 3 3 3 3 3 1 0 0 0 1.475 1.0 2.0 05/31/98 MM 1 1 1 1 2 1 1 2 1 1 0 0 0 0 1.467 1.0 1.9 05/31/98 NN 0 1 0 2 0 0 0 0 0 0 0 0 0 0 1.233 1.1 1.3 05/31/98 RR 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1.400 1.4 1.4 05/31/98 W 0 1 2 1 2 0 0 0 0 0 0 0 0 0 1.267 1.1 1.4 05/31/98 X 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1.200 1.2 1.2

316 Appendix H. Locality and chorus pitch data, 1998 Brood XIX. Series Tape Side Counter Location State County Lat Lon Date Hi kHz Hi XJ/Hz X Lo kHz Lo XJ/Hz X DD98 1 B 447 S. of Montgomery, Rt 24 1.5 m E. of US31 at river AL Montgomery 32.176 86.345 06/03/98 1.12 DD98 1 B S. of Montgomery, Rt 24 1.8 mi E of Montgomery Co border at Federal Rd AL Montgomery 32.177 86.386 06/03/98 1.09 1.36 u DD98 1 B 354 #6 at 19.5m N. of Double Springs on Rt 33, Black Warrior WMA at Lawrence Co Rt 90 jnc. AL Lawrence 34.393 87.315 06/03/98 1.12 6.32 u DD98 1 B 354 #5 at 19.5m N. of Double Springs on Rt 33, Black Warrior WMA at Lawrence Co Rt 90 jnc. AL Lawrence 34.393 87.315 06/03/98 1.21 4.42 u DD98 1 B 354 #4 at 19.5m N. of Double Springs on Rt 33, Black Warrior WMA at Lawrence Co Rt 90 jnc. AL Lawrence 34.393 87.315 06/03/98 1.15 2.7 u DD98 1 B 354 #3 at 19.5m N. of Double Springs on Rt 33, Black Warrior WMA at Lawrence Co Rt 90 jnc. AL Lawrence 34.393 87.315 06/03/98 1.11 15.1 u DD98 1 B 354 #2 at 19.5m N. of Double Springs on Rt 33, Black Warrior WMA at Lawrence Co Rt 90 jnc. AL Lawrence 34.393 87.315 06/03/98 1.14 6.1 u DD98 1 B 354 #1 at 19.5m N. of Double Springs on Rt 33, Black Warrior WMA at Lawrence Co Rt 90 jnc. AL Lawrence 34.393 87.315 06/03/98 1.11 2.05 u DD98 S. of Marvin on Rt. 51 2.6 m South of US80 AL 32.404 85.363 1.09 1.39 u DD98 S. of Marvin on Rt. 51 3.8 m S of US80 AL 32.38 85.379 1.08 896 n DJ98 1 B 336 Hardy-Weaver AR Sharp 36.315 91.477 05/19/98 1.69 1.93 n DJ98 3 A 8 Black Rock AR Lawrence 36.122 91.151 05/22/98 1.15 91.3 n DJ98 3 A 45 N of Hardy on 63 AR Fulton 36.337 91.492 05/22/98 1.74 617 n 1.13 117 n DJ98 3 A 70 Mammoth Springs AR Fulton 36.474 91.52 05/22/98 1.65 4740 n 1.05 457 n DJ98 3 A 133 1/3 way south from JNC 342 x 63 to Hardy on 63 AR Fulton 36.442 91.493 05/22/98 1.74 59.6 n 1.11 146 n DJ98 2 B 327 Strawberry AR Sharp 36.111 91.452 05/22/98 1.76 99.8 n 1.17 1190 n DJ98 2 B 345 58 x 115 AR Sharp 35.999 91.487 05/22/98 1.76 14.9 n 1.15 572 n DJ98 2 B 359 115 x Strawberry AR Lawrence 36.028 91.329 05/22/98 1.69 0.041 n 1.13 6.68 n DJ98 2 B 378 Shirey AR Lawrence 36.022 91.235 05/22/98 1.68 4.88 n 1.15 118 n DJ98 2 B 397 310 x 309 AR Lawrence 36.019 91.192 05/22/98 1.15 987 n DJ98 2 B 463 E. of Lawrence Co. 303 AR Lawrence 36.061 91.145 05/22/98 1.8 18.1 n 1.17 159 n DJ98 2 B 487 Lake Charles B AR Lawrence 36.081 91.127 05/22/98 1.74 6.66 n 1.16 776 n DJ98 4 A 135 27 x 14 AR Searcy 35.923 92.617 05/28/98 1.11 67.6 n DJ98 4 A 290 27 x 65 AR Searcy 35.95 92.727 05/28/98 1.14 34.7 n

317 DJ98 4 A 338 235 x 418 AR Searcy 36.106 92.837 05/28/98 1.64 5.05 n DJ98 4 B 451 14 X 281 AR Boone 36.449 93.074 06/02/98 1.64 201 p DD98 1 B 343 #6 at Daisy State Park AR Pike 34.233 93.742 06/02/98 1.17 1.52 u DD98 1 B 342 #5 at Daisy State Park AR Pike 34.233 93.742 06/02/98 1.15 910 n DD98 1 B 316 #4 at Daisy State Park AR Pike 34.233 93.742 06/02/98 1.12 894 n DD98 1 B 240 #3 at Daisy State Park AR Pike 34.233 93.742 06/02/98 1.12 1160 n DD98 1 B 240 #2 at Daisy State Park AR Pike 34.233 93.742 06/02/98 1.14 981 n DD98 1 B 240 #1 at Daisy State Park AR Pike 34.233 93.742 06/02/98 1.11 1.15 u DD98 1 B 200 Ca. 8m W. of Arkadelphia on Rt. 8 AR Clark 34.19 93.222 06/02/98 1.16 766 n DJ98 1 A 28 Harold Alexander WMA Taillight AR Sharp 36.208 91.447 5/12-24/1998 1.73 32 1.14 133 DJ98 1 B 352 Harold Alexander WMA, Power Line AR Sharp 36.256 91.462 5/19-25/1998 1.702167325 442.85 1.12 114.5625 DJ98 4 A 400 Prairie De Rocher IL Randolph 38.088 90.09 05/29/98 1.7 16.7 n 1.1 624 n DJ98 4 A 465 900N IL Monroe 38.156 90.052 05/29/98 1.63 100 n 1.07 500 n DJ98 4 B 64 Pere Marquette IL Jersey 38.972 90.543 05/29/98 1.48 27.6 n DJ98 4 B 214 Sugar Creek IL Sangamon 05/30/98 1.41 2.236 a DJ98 4 B 225 Allerton IL Piatt 40.008 88.644 05/30/98 1.4 346 n DD98 1 A 38 Sam Parr SP nr. Newton IL Jasper 39.016 88.125 05/31/98 1.41 80 n DD98 1 A 18 Fox Ridge SP, SE of Matoon IL Coles 39.403 88.143 05/31/98 1.37 1.3 u DD98 1 A 451 Dixon Spr. SP on Rt 146 #1 IL Pope 37.386 88.671 06/01/98 1.83 0.515 u 1.12 11.8 u DD98 1 A 451 Dixon Spr SP on Rt 146 #2 IL Pope 37.386 88.671 06/01/98 1.12 3.89 u DD98 1 A 451 Dixon Spr SP on Rt 146 #3 IL Pope 37.386 88.671 06/01/98 1.64 0.2 u 1.09 4.9 u DD98 1 A 423 Golconda Marina #1 IL Pope 37.386 88.48 06/01/98 1.67 0.36 u 1.1 9.22 u DD98 1 A 384 Ohio River Rec Area (Golconda) #2 IL Pope 37.377 88.485 06/01/98 1.09 4.19 u DD98 1 A 384 Ohio River Rec Area (Golconda) #1 IL Pope 37.377 88.485 06/01/98 1.09 4.19 u DD98 1 A 355 Rt. 5 6.8 mi. S. of Rt. 145, NW of Golconda Sample 1 IL Pope 37.424 88.514 06/01/98 1.69 0.386 u 1.1 4.2 u DD98 1 A 355 Rt. 5 6.8 mi. S. of Rt. 145, NW of Golconda Sample 2 IL Pope 37.424 88.514 06/01/98 1.65 2 u 1.12 15.1 u DD98 1 A 310 Rt. 145 2.5 mi. S. of Forest Rd. Rt 402 ("Rt 145 Collection Site"), 2.5 mi. S. of Delwood IL Pope 37.546 88.582 06/01/98 1.76 0.903 u 1.12 3.55 u DD98 1 A 275 Rt. 402 0.9m W of Rt. 145 (Delwood), "Site 3" IL Pope 37.58 88.588 06/01/98 1.78 0.972 u 1.1 8.3 u DD98 1 A 250 Rt. 402 0.5m W of Rt. 145 (Delwood), "Site 2" Sample 1 IL Pope 37.58 88.581 06/01/98 1.78 1.41 u 1.12 4.27 u DD98 1 A 250 Rt. 402 0.5m W of Rt. 145 (Delwood), "Site 2" Sample 2 IL Pope 37.58 88.581 06/01/98 1.76 1.5 u 1.12 6.24 u DD98 1 A Rt. 402 0.4m W of Rt. 145 (Delwood), "Site 1" Sample 1 IL Pope 37.58 88.578 06/01/98 1.78 3.15 u 1.12 6.98 u DD98 1 A Rt. 402 0.4m W of Rt. 145 (Delwood), "Site 1" Sample 2 IL Pope 37.58 88.578 06/01/98 1.78 2.35 u 1.09 3.97 u DD98 1 A 163 0.1m off Rt. 145 on Battle Ford Rd. (near Mitchellsville) Sample 3 IL Saline 37.633 88.532 06/01/98 1.77 0.25 u 1.11 10 u DJ98 Nole Alexander Farm IL Piatt 40.071 88.515 06/09/98 1.37 100 u JRC98 2 A 238 IL 37 X I57 IL Effingham 38.993 88.59 06/10/98 1.46 6.58 n JRC98 2 A 275 Forbes 1 IL Marion 38.73 88.777 06/10/98 1.58 41.3 a JRC98 2 A 329 IL 161 X I 57 IL Marion 38.516 88.959 06/10/98 1.57 992 a JRC98 2 A 420 Muddy Creek IL Jefferson 38.31 88.988 06/11/98 1.09 30.7 n JRC98 2 A 451 Rayse Creek IL Jefferson 38.328 89.107 06/11/98 1.65 17.1 n 1.13 17.5 n JRC98 2 B 8 1425 X 400 IL Jefferson 38.332 89.071 06/11/98 1.66 2.25 n 1.1 12.2 n JRC98 2 B 52 500E IL Jefferson 38.402 89.052 06/11/98 1.66 14.6 n 1.08 9.9 n JRC98 2 B 159 2000N IL Jefferson 38.417 88.9 06/11/98 1.62 833 p 1.09 1 p JRC98 2 B 183 Lake Jaycee B IL Jefferson 38.37 88.887 06/11/98 1.65 16 n JRC98 2 B 207 Sam Dale A IL Wayne 38.55 88.586 06/12/98 1.65 1 p 1.09 939 p JRC98 2 B 325 Rend Lake IL Jefferson 38.216 88.986 06/12/98 1.11 130 n JRC98 2 B 464 Lonely Oak IL Washington 38.281 89.357 06/12/98 1.72 1.52 a JRC98 3 A 42 100E X Berryville IL Lawrence 38.6 87.891 06/13/98 1.13 2.48 n JRC98 3 A 105 Middle Fork IL Champaign 40.385 87.959 06/14/98 1.37 4.16 n JRC98 3 A 178 1100N x 2300E IL Livingston 40.787 88.489 06/14/98 1.36 286 p DD98 1 B 184 Lake Barkley Collection Site Sample #1 KY Trigg 36.849 87.914 06/01/98 1.1 50 n DD98 1 B 184 Lake Barkley Collection Site Sample #2 KY Trigg 36.849 87.914 06/01/98 1.1 1 u DD98 1 B 50 Lake Barkley near powerlines KY Trigg 36.849 87.914 06/01/98 1.09 16.6 n DD98 1 B 50 Lake Barkley near powerlines KY Trigg 36.849 87.914 06/01/98 1.12 DD98 1 B 20 Rt. 62 2m. E. of Princeton KY Caldwell 37.107 87.941 06/01/98 1.15 605 n ZV98 Jct Rt 235/Diamond Lane (S. of Dameron) MD St. Mary's 38.145 76.365 05/25/98 1.08 ZV98 Bishop House end of Demko Rd MD St. Mary's 38.169 76.371 05/27/98 1.05

318 ZV98 End of Three-second lane, N. of Dameron MD St. Mary's 38.16 76.369 05/28/98 1.1 ZV98 End of Kessler Way MD St. Mary's 0 06/01/98 1.12 DJ98 4 B 373 I44 X US 50 MO St. Louis 38.536 90.503 06/01/98 1.48 15.6 n DJ98 4 B 406 I44 ex 2301 Rol Hil MO Franklin 38.279 91.124 06/01/98 1.46 736 p JRC98 1 A 4 Rte 112 MTNF MO Barry 36.543 93.859 06/03/98 1.67 194 p JRC98 1 A 54 Rte. 112 pullout MO Barry 36.574 93.836 06/03/98 1.59 1.43 n JRC98 1 A 72 Roaring River MO Barry 36.586 93.837 06/03/98 1.59 463 p JRC98 1 A 82 FR 1200 MO Barry 36.602 93.802 06/03/98 1.59 40.05 a JRC98 1 A 102 Hwy 173 MO Stone 36.73 93.511 06/03/98 1.59 87.15 a JRC98 1 A 111 Eag Cr 1 M N of Lake MO Stone 36.735 93.489 06/03/98 1.55 7.45 n JRC98 1 A 164 AA914 x Grace Rd. MO La Clede 37.7 92.709 06/03/98 1.48 48.5 n JRC98 1 A 199 Bennet Nature MO La Clede 37.726 92.856 06/03/98 1.48 4.59 n JRC98 1 A 452 14 x AP MO Howell 36.811 92.054 06/06/98 1.6 98.6 a JRC98 1 A 475 14 x Co 5130 MO Howell 36.8 91.991 06/06/98 1.48 69 p JRC98 1 B 8 1-2960 x WW MO Howell 36.961 91.714 06/06/98 1.62 64.2 p JRC98 1 B 29 1.1- 2960 x WW MO Howell 36.962 91.709 06/06/98 1.6 49.3 p JRC98 1 B 60 60-571 MO Shannon 36.999 91.31 06/06/98 1.67 140 a JRC98 1 B 248 Possum Creek MO Wayne 36.968 90.297 06/06/98 1.19 687 p JRC98 1 B 285 Hilltop MO Wayne 37.042 90.296 06/06/98 1.77 351 p JRC98 1 B 311 1.0C X 65 MO Dallas 37.683 93.124 06/07/98 1.48 3.41 n JRC98 1 B 333 1.5C X 65 MO Dallas 37.682 93.126 06/07/98 1.49 13.8 n JRC98 1 B 350 0.5 140 X 399 MO Hickory 37.962 93.197 06/07/98 1.49 2.3 n JRC98 1 B 365 0.7 140 X 399 MO Hickory 37.962 93.197 06/07/98 1.5 12.6 n JRC98 1 B 383 1.5 dn 391 MO Benton 38.17 93.138 06/07/98 1.5 7.56 n JRC98 1 B 399 Mt. Olivet MO Benton 38.451 93.275 06/07/98 1.46 7.37 n JRC98 1 B 424 Fristo X Wasson MO Pettis 38.786 93.2 06/07/98 1.45 272 a JRC98 2 A 62 439 X US24 MO Chariton 39.426 92.754 06/07/98 1.48 151 p JRC98 2 A 159 Karma 1 MO Randolph 39.608 92.455 06/07/98 1.5 110 p DD98 0.8 mi S from Rt 49 on Tuckertown Rd, N. of Tuckertown NC 35.506 80.181 1.08 DD98 Vista Pt SRA E. of Pittsboro NC 35.703 79.05 1.15 5.05 u DD98 Rt. 87 N. of Pittsboro, 1.0 m S of SR1551 NC 35.818 79.254 1.1 2.14 u DD98 Nashville, Percy Warner SP TN 36.069 86.886 1.06 935 APPENDIX I

Holding jig for symmetry measurements.

An X-acto “X-tra Hands” tool (Hunt Mfg. Co., Speedball Road, Statesville, NC 28677) was modified in the following manner to make a holding jig.

1. Alligator clips were removed and discarded.

2. Copper alligator clips were filed so that the gripping surfaces were flat when gripping two microscope slides and two sheets of brass.

3. Small rectangles of brass were soldered into the filed jaws of the copper alligator clips.

4. The brass rectangles were glued to two pieces of microscope glass using “liquid nails” adhesive.

5. All exposed areas of adhesive were coated with red GLPT insulating varnish. This step is critical since the insulating varnish forms a barrier that protects the adhesive from exposure to alcohol or other preserving fluids.

6. The clip/slide assembly was installed in the X-tra hands device.

319 Appendix J. Symmetry and suze measurements for mated and unmated cicadas. Mate F14 F 14 F 15 F 15 F 27 F 27 H 40 H 40 H 46 H 46 Fore Femur Fore Femur Hind Femur Hind Femur Tymbal Rib Cumulative Cumulative Status Cicada Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry (FA Char.) (All) No 1 0.00991736 11.4233842 -0.0096 11.8010167 0.00291121 12.9716776 0.0221169 11.9520697 0.00297177 12.7073348 0.029752066 4.459616546 0.003724395 4.91776455 0 0.039352066 0.058876793 No 2 0.01324503 11.4045025 0 0.00578035 13.0660857 0 11.328976 0.00561798 13.4437182 -0.013029316 4.525957949 -0.014234875 5.146710758 0 No 3 0 11.2912128 0.00310078 12.1786492 -0.0173913 13.0283224 -0.02955665 11.4989107 0 12.1220044 -0.00625 4.717610892 0 5.256604938 0 0.009350775 0.02674208 No 4 -0.02040816 11.1023965 -0.02764977 12.291939 -0.0149925 12.594045 -0.01011804 11.1968046 0.01201201 12.5751634 0.00619195 4.761838494 -0.017761989 5.155868607 0 0.03384172 0.099016388 No 5 -0.0033389 11.3100944 0 0 12.4618736 0.01658375 11.3856209 0.01447178 13.0472041 0.006557377 4.496472881 0.010810811 5.08260582 0 No 6 0.01333333 11.328976 0.00304414 12.4052288 -0.01801802 12.5751634 0.01360544 11.1023965 0.00298954 12.6318083 0.006493506 4.540700483 -0.003853565 4.75292328 0 0.009537647 0.047732099 No 7 0 11.8198983 -0.00930233 12.1786492 0.01146132 13.1793755 -0.0033389 11.3100944 0.00300752 12.5562818 0.016638935 4.430131478 -0.019267823 4.75292328 0 0.025941261 0.05967792 No 8 0 13.5947712 0.00283688 13.3115468 0.0137931 13.6891794 0 12.8395062 0 0.00317965 4.636526954 -0.031088083 5.30239418 0 0.00601653 No 16 -0.00317965 11.8765432 -0.01791045 12.6506899 -0.0115942 13.0283224 0 11.6310821 0 -0.006493506 4.540700483 0.010810811 5.08260582 0 0.024403954 No 17 0.00305344 12.3674655 0.00860832 13.1604938 0.01142857 13.2171387 0.01536098 12.291939 0 0 4.599670619 0 5.494708995 1 0.008608321 No 18 -0.01769912 12.8017429 0.00291971 12.9339143 0.01671309 13.557008 0.03389831 12.2541757 0 0.012269939 4.806066096 0.007017544 5.219973545 0 0.015189647 No 19 -0.0356564 11.6499637 0 12.4241104 -0.00302572 12.4807553 -0.01011804 11.1968046 0 13.0283224 0.019480519 4.540700483 -0.007326007 5.000185185 0 0.019480519 0.065488647 No 20 0.01550388 12.1786492 -0.00902256 12.5562818 0 13.0283224 0.00638978 11.8198983 0.01438849 13.1227306 0.009508716 4.651269488 -0.007042254 5.201657848 0 0.018531273 0.055465891 No 21 -0.02843602 11.9520697 0.02067947 12.7828613 -0.0057971 13.0283224 0 0 -0.026578073 4.437502745 -0.003565062 5.13755291 0 0.047257541 No 22 0.05405405 12.5751634 0 0.00581395 12.9905592 0.01282051 11.7821351 0 0.009538951 4.636526954 0 5.091763668 0 No 23 -0.02346041 12.8772694 -0.03156385 13.1604938 -0.04255319 13.3115468 0.00956938 11.83878 -0.00277393 13.6136529 0.003025719 4.872407499 -0.016977929 5.393972663 0 0.034589564 0.12035502 No 24 0.02337229 11.3100944 -0.01246106 12.1220044 0.00569801 13.254902 0.01960784 11.5555556 0 0.009756098 4.533329216 0.003656307 5.009343034 0 0.022217157 0.054943757 No 25 -0.0030349 12.442992 0 12.4618736 0.00302572 12.4807553 0.00636943 11.8576616 0 13.5947712 0.00317965 4.636526954 0 5.219973545 1 0.00317965 0.00924027 No 26 -0.01298701 11.6310821 -0.00593472 12.7262164 0 13.254902 0.0294599 11.5366739 -0.00289436 13.0472041 0.006644518 4.437502745 0.003696858 4.954395944 0 0.012579236 0.032157463

320 No 27 0 12.4618736 0.01503759 12.5562818 0.01694915 13.3681917 3.0446E-16 11.6688453 -0.00284495 13.2737836 0.003305785 4.459616546 0.010810811 5.08260582 0 0.018343379 0.048948293 No 28 -0.00634921 11.8954248 -0.00291121 12.9716776 -0.00554017 13.6325345 -0.03647416 12.4241104 -0.01915185 13.8024691 0 4.614413153 0 0 0.002911208 No 29 -0.04334365 12.1975309 0.03003003 12.5751634 -0.02253521 13.405955 0.01875 12.0842411 0.01111111 13.5947712 0.01610306 4.577556818 0.014184397 5.165026455 0 0.04613309 0.137307462 No 30 0.01948052 11.6310821 -0.01230769 12.2730574 0.02105263 12.5562818 0.00327332 11.5366739 0.01204819 12.5374001 -0.010050251 4.40064641 -0.011385199 4.826186067 0 0.022357944 0.086324487 No 31 0.01360544 11.1023965 0.00638978 11.8198983 -0.00623053 12.1220044 0.01727116 10.9324619 0.01494768 12.6318083 0.003327787 4.430131478 -0.003710575 4.936080247 0 0.009717563 0.048211793 No 39 0 11.328976 -0.00913242 12.4052288 0 12.4996369 0 10.7625272 -0.00886263 12.7828613 0.022764228 4.533329216 -0.010544815 5.210815697 0 0.031896648 0.051304092 No 40 0.01215805 12.4241104 -0.00615385 12.2730574 0.00297177 12.7073348 0.00953895 11.8765432 0 0 4.422760211 0 5.018500882 0 0.006153846 No 49 0.00916031 12.3674655 0 0.00272851 13.8402324 0 0 0.009630819 4.592299352 0 5.165026455 0 No 50 -0.0033389 11.3100944 -0.00607903 12.4241104 0 13.0283224 0.01886792 12.0087146 -2.6803E-16 13.254902 -0.013157895 4.481730347 -0.003629764 5.045974427 0 0.019236922 0.026205584 No 51 -0.02033898 11.1401598 0 -0.01526718 12.3674655 0.022187 11.9143065 0.01898734 11.9331881 0.009569378 4.62178442 0.003514938 5.210815697 1 No 52 0.01269841 11.8954248 -0.01183432 12.7639797 -0.00297177 12.7073348 0.00329489 11.4611474 0.02279202 13.254902 -0.01610306 4.577556818 -0.010657194 5.155868607 0 0.027937379 0.077056776 No 53 0 11.8954248 -0.02143951 12.3297023 -0.01204819 12.5374001 0 12.2730574 0.00285307 13.2360203 0 4.422760211 -0.003565062 5.13755291 0 0.02143951 0.039905832 No 54 -0.00933126 12.140886 -1.4E-16 12.6884532 0.00561798 13.4437182 0 11.1401598 0 -0.016051364 4.592299352 0.021505376 5.110079365 0 0.016051364 No 55 -0.0125 12.0842411 -0.01176471 12.8395062 -0.01430615 13.1982571 -0.01526718 12.3674655 0.00272851 13.8402324 0.012195122 4.835551164 -0.020408163 5.384814815 0 0.023959828 0.073902656 No 56 -0.00350263 10.7814089 -0.0096 11.8010167 -0.00323102 11.6877269 -0.03231598 10.5170661 -0.00636943 11.8576616 -0.006688963 4.408017677 -0.03364486 4.899448854 1 0.016288963 0.063036895 No 57 0.00630915 11.9709513 -0.00308166 12.2541757 0 13.5947712 0.02782071 12.2164125 0 13.2926652 0.012779553 4.614413153 0.003514938 5.210815697 0 0.015861217 0.025685304 No 58 -0.02643172 12.8583878 -0.02043796 12.9339143 0.01426534 13.2360203 0.00636943 11.8576616 -0.00557103 13.557008 0.012779553 4.614413153 -0.007092199 5.165026455 0 0.033217509 0.086577791 No 59 0.00649351 11.6310821 0.00584795 12.9150327 -0.0027894 13.5381264 0 0 0 4.673383289 0 5.201657848 0 0.005847953 0.01513086 No 60 0.01960784 11.5555556 -0.00944882 11.989833 0.00578035 13.0660857 -0.00638978 11.8198983 0.02890173 13.0660857 0.003327787 4.430131478 -0.003577818 5.119237213 0 0.012776606 0.070644348 No 61 -0.02368866 11.1590414 0.00316957 11.9143065 0.00283688 13.3115468 0.00680272 11.1023965 0.02157165 12.2541757 -0.01320132 4.466987813 -0.007434944 4.926922399 0 0.016370892 0.071903028 No 62 0.02150538 12.291939 0.02756508 12.3297023 0.02008608 13.1604938 -0.01615509 11.6877269 0 13.3304285 0.02907916 4.562814284 0.010810811 5.08260582 0 0.056644244 0.109046515 No 63 0 11.328976 -0.00581395 12.9905592 -0.00285307 13.2360203 0 11.8954248 -0.00278164 13.5758896 -0.009360374 4.724982159 0.003384095 5.41228836 0 0.015174328 0.024193131 No 64 0.0030349 12.442992 0.02002861 13.1982571 0.00266312 14.1801017 -0.00626959 12.0464779 0 13.8213508 0.00312989 4.710239625 0 5.494708995 1 0.023158503 0.02885652 No 65 0.00351494 10.7436456 0.00328407 11.4989107 0.00625 12.0842411 -0.01769912 10.6681191 0.00316957 11.9143065 0.017035775 4.32693374 0.003898635 4.69797619 0 0.020319847 0.037152993 No 66 -0.00321027 11.7632534 0.00609756 12.3863471 0.00301659 12.5185185 -0.02547771 11.8576616 0 13.2926652 0.029268293 4.533329216 0 5.036816578 0 0.035365854 0.041592718 No 67 0.00343053 12.3674655 -0.00617284 13.1604938 0 13.2171387 -0.0034904 12.291939 0 0.02027027 4.599670619 -0.011173184 5.494708995 0 0.02644311 No 68 0.0060241 12.8017429 0.00865801 12.9339143 0.00557103 13.557008 -0.01846154 12.2541757 0 -0.003149606 4.806066096 0 5.219973545 0 0.011807615 No 69 0.00628931 11.6499637 0.0140647 12.4241104 -0.01117318 12.4807553 0.00623053 11.1968046 0.00921659 4.540700483 0.003350084 5.000185185 0 0.023281287 0.044093864 No 70 -0.02492212 12.1220044 -0.01173021 12.8772694 -0.01149425 13.1416122 0 12.1597676 -0.00834492 13.5758896 0.009884679 4.47435908 0.007092199 5.165026455 0 0.021614884 0.073468377 No 71 0.02298851 11.4989107 0.00621118 12.1597676 -0.00297177 12.7073348 0 11.7066086 0 0.0224 4.607041886 -0.003514938 5.210815697 0 0.02861118 No 72 -0.00331675 11.3856209 -0.00309119 12.2164125 -0.00607903 12.4241104 0 11.6310821 0.02014388 13.1227306 0.009884679 4.47435908 -0.021428571 5.128395062 0 0.012975869 0.063944102 Yes 9 -0.00327332 11.5366739 0.01490313 12.6695715 0.00873362 12.9716776 -0.00953895 11.8765432 0 0.012383901 4.761838494 0.013745704 5.329867725 1 0.027287031 Yes 10 -0.01584786 11.9143065 0 0.00289436 13.0472041 0.00956938 11.83878 -0.00550964 13.708061 0.003159558 4.666012022 0.017953321 5.100921517 0 0.003159558 0.045364737 Mate F14 F 14 F 15 F 15 F 27 F 27 H 40 H 40 H 46 H 46 Fore Femur Fore Femur Hind Femur Hind Femur Tymbal Rib Cumulative Cumulative Status Cicada Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry Size Symmetry (FA Char.) (All) Yes 11 -0.02994012 12.6129267 -0.02086438 12.6695715 -0.01101928 13.708061 0 0 -0.012820513 4.599670619 -0.025316456 5.064290123 0 0.033684894 Yes 12 0.02446483 12.3485839 0.01173021 12.8772694 -0.00557103 13.557008 -0.01550388 12.1786492 0.01680672 13.4814815 -0.003189793 4.62178442 -0.003577818 5.119237213 0 0.014919998 0.065340401 Yes 13 0.03389831 11.1401598 0.01941748 11.6688453 0.00571429 13.2171387 0.02261712 11.6877269 0 0.003327787 4.430131478 0.007092199 5.165026455 0 0.022745263 Yes 14 0 11.4045025 0.02439024 12.3863471 0.00292826 12.8961511 0.01269841 11.8954248 0 -0.003231018 4.562814284 0.017761989 5.155868607 0 0.027621262 Yes 15 0.00318979 11.83878 0 0.00578035 13.0660857 0.04388715 12.0464779 -0.00825309 13.7269426 -0.013245033 4.452245279 -0.011378556 5.231420855 0 0.013245033 0.041846823 Yes 32 0.00657895 11.480029 0.00641026 11.7821351 0.0058651 12.8772694 0.01324503 11.4045025 0.00584795 12.9150327 0.013157895 4.481730347 0.003552398 5.155868607 0 0.019568151 0.041412552 Yes 33 -0.04778157 11.0646333 -0.02461538 12.2730574 0 12.4618736 0.00664452 11.3667393 0.01173021 12.8772694 0.003327787 4.430131478 -0.007434944 4.926922399 0 0.027943172 0.094889891

321 Yes 34 0.0126183 11.9709513 -0.00636943 11.8576616 0.01574803 11.989833 0.02020202 11.2156863 0.00302572 12.4807553 0.003395586 4.341676274 0 4.945238095 0 0.009765012 0.041157059 Yes 35 0.00941915 12.0275962 0 0 13.4437182 0.02531646 11.9331881 0.0056338 13.405955 0.003210273 4.592299352 -0.007067138 5.183342152 0 Yes 36 -0.00317965 11.8765432 0 12.5374001 0.00289436 13.0472041 -0.02702703 11.177923 0 13.0283224 0.012944984 4.555443017 -0.01814882 5.045974427 0 0.012944984 0.03716781 Yes 37 0.04636785 12.2164125 0.01230769 12.2730574 0.01485884 12.7073348 0.00638978 11.8198983 0.0058309 12.9527959 0 4.422760211 0 0 0.012307692 Yes 38 0.00322061 11.7254902 -0.00606061 12.4618736 0.00878477 12.8961511 0.02329451 11.3478577 0.00557103 13.557008 0.006451613 4.570185551 0.003539823 5.174184303 0 0.012512219 0.033628458 Yes 41 -0.00668896 11.2912128 0 -0.01569859 12.0275962 0.00981997 11.5366739 -0.00611621 12.3485839 0.013071895 4.511215415 0.018416206 4.97271164 1 Yes 42 0.00913242 12.4052288 0 12.4618736 0.00846262 13.3870733 0.02870813 11.83878 0.01958042 13.5003631 -0.003159558 4.666012022 0.003490401 5.24744709 0 0.003159558 0.043825422 Yes 43 0.01043478 10.8569354 0 0.0143472 13.1604938 0.00315956 11.9520697 0.01204819 12.5374001 0.009917355 4.459616546 0.02166065 5.073447972 0 Yes 44 0 12.6506899 -0.00281294 13.4248366 0 13.5947712 0.02143951 12.3297023 0 0 4.865036232 0 5.219973545 0 0.00281294 Yes 45 0.01540832 12.2541757 0.0089955 12.594045 0.002886 13.0849673 0.03389831 12.2541757 0 0.009360374 4.724982159 0.006968641 5.256604938 0 0.018355877 Yes 46 -0.01476015 10.2338417 -0.00664452 11.3667393 0.00662252 11.4045025 0 0 0 3.950999122 0.007936508 4.615555556 0 0.006644518 Yes 47 0.01574803 11.989833 0.00588235 12.8395062 0 13.0283224 0.02276423 11.6122004 0.01117318 13.5192447 0.003025719 4.872407499 0.017152659 5.339025573 0 0.008908072 0.052981946 Yes 48 0.00607903 12.4241104 0 -0.00863309 13.1227306 0 12.3863471 0 13.0283224 0.022544283 4.577556818 0.014234875 5.146710758 0 Yes 73 0.00630915 11.9709513 0 -0.01117318 13.5192447 -0.03870968 11.7066086 0.00273598 13.8024691 0.003072197 4.798694829 0.013605442 5.384814815 0 Yes 74 -0.01647446 11.4611474 -2.9584E-16 12.0087146 0 13.5947712 -0.03355705 11.2534495 0 -0.00317965 4.636526954 -0.010544815 5.210815697 0 0.00317965 Yes 75 0.00332779 11.3478577 0 12.9150327 -0.00598802 12.6129267 -0.01642036 11.4989107 0 0 4.658640755 0 5.201657848 0 0 Yes 76 0.02711864 11.1401598 0.01197605 12.6129267 0.00881057 12.8583878 0.00699301 10.8002905 0.01204819 12.5374001 0.006802721 4.334305007 -0.041745731 4.826186067 0 0.018778769 0.108501909 Yes 77 0.01733102 10.8946986 0.00316957 11.9143065 0.002886 13.0849673 0.00966184 11.7254902 0 -0.006644518 4.437502745 0.003642987 5.02765873 0 0.00981409 APPENDIX K

Code for resampling algorithm to evaluate whether males with asymmetrical wing venation were over- or under-represented among males that did not mate in mating experiments. (Think Lightspeed Pascal for the Macintosh)

program Extra;

const TargetMatings = 2; {1= not mated; 2= mated} malenumber = 16;

var Tabarray, NewData, HardData: array[0..77] of integer; Distribution: array[0..10000] of integer; NumWTargMatings, iterations, randomnumber, FemaleNumber: integer;

{* Seeds Random Number Generator} procedure TIMESEED; begin GetDateTime(randseed); end;

procedure Welcome; begin writeln('Welcome to Resampler. You can run up to 10,000 resamplings'); writeln(' of abnormal wing veined male mating data. '); end;

{*Picks a random number, scales it to the appropriate size} procedure PICKRAND; begin randomnumber := (Trunc(((ABS(Random)) / 32768) * 77)); {writeln('Picked number ', randomnumber);} end;

procedure FillHardSlot; begin pickrand; {writeln(randomnumber);} if HardData[randomnumber] = 2 then begin FillHardSlot; {writeln('repeat picked!');} end; HardData[randomnumber] := 2; end;

322 procedure HardDataMaker; var counter: integer; begin for counter := 1 to 77 do HardData[counter] := 1; for counter := 1 to 27 do begin FillHardSlot; end; for counter := 1 to 77 do {Writeln(HardData[counter]);} end;

{Gets Input parameters} procedure INPUTS; begin writeln('Please provide the number of iterations'); writeln('of the simulation , maximum 10000 : '); read(iterations); end;

{Checks Input parameters} procedure CHECKVAR; begin if (iterations > 10000) or (iterations < 1) then begin writeln('Sorry, but iterations must be an integer less than 10000'); writeln('program has been terminated'); HALT; end; end;

procedure Resample; var counter: integer; begin for counter := 1 to malenumber do begin pickrand; if randomnumber = 0 then pickrand; NewData[counter] := HardData[randomnumber]; {Writeln(NewData[counter]);} end; end;

{clears this array} procedure ClearDistribution; var counter: integer; begin for counter := 0 to 10000 do Distribution[counter] := 0;

323 end;

{creates a variable containing number of females that got TargetMatings matings} procedure TABULATEmATINGS; var counter: integer; begin NumWTargMatings := 0; for counter := 1 to malenumber do begin if (NewData[counter] = TargetMatings) then NumWTargMatings := NumWTargMatings + 1; end; end;

procedure iterate; var counter: integer; begin for counter := 1 to iterations do begin resample; TabulateMatings; Distribution[counter] := NumWTargMatings; {Writeln(counter, 'Distribution at counter', distribution[counter]);} {Writeln(NumWTargMatings, 'NumWTargMatings');} {Writeln(' ');} end; end;

{Sorts distribution array to prepare for generating percentiles} procedure bubblesort; var i, j, k: integer; begin for i := iterations downto 2 do for j := 1 to i - 1 do if (distribution[j] > distribution[j + 1]) then begin k := distribution[j]; distribution[j] := distribution[j + 1]; distribution[j + 1] := k; end; end;

procedure MakeDistribution; var DistributionTab, count, cumucount, countera, counterb, counterc: integer; counter: integer; begin bubblesort; {for counterc := 0 to iterations do} {begin} {writeln('post-sort');} {writeln(distribution[counterc]);} {end;} cumucount := 0; {ratchets thru loking for each value; set some kind of a tab ratchet on it each}

324 {time it gets a "hit"} for DistributionTab := 0 to malenumber do begin count := 0; for counterb := 1 to iterations do if (Distribution[counterb] = DistributionTab) or (Distribution[counterb] < DistributionTab) then count := count + 1; {cumucount := cumucount + count;} Writeln('Frequency of trials in which', DistributionTab, ' or fewer'); Writeln('deformed males achieved matings'); {Writeln(cumucount / iterations);} Writeln(count / iterations); {Writeln(count);} Writeln(' '); end; writeln(' '); {for counter := 1 to iterations do} {Writeln(distribution[counter]);} {writeln(' ');} end;

{The shell of the program} procedure MAKEDECISION; var choicer: integer; begin writeln(' '); writeln('Would you like to:'); writeln(' (1) Resample ?'); writeln(' (2) End this session?'); writeln(' '); writeln('Choose (1) or (2) please'); read(choicer); if choicer = 1 then begin Timeseed; Welcome; inputs; checkvar; ClearDistribution; HardDataMaker; iterate; MakeDistribution; Makedecision; end; if choicer = 2 then HALT; if (choicer < 1) or (choicer > 2) then begin writeln('Hey there are only two choices here!'); Makedecision; end; end;

begin MakeDecision; end.

325 APPENDIX L

Miscellaneous field notes on Okanagana

1. The calling behaviors and morphology of O. canadensis and O. rimosa are superficially similar to those of O. bella found in the western United States. The mating system of this species may be similar to that of the eastern cicadas. 2. Nymphs and eclosing nymphs were found only during the dense 1998 emergence of O. rimosa. Nymphal skins of O. rimosa and O. bella are similar, with prominent dark, dorsal abdominal bands. Emerging O. rimosa nymphs are a pale green, and can be found between approximately 10 AM and 2 PM, not in the evening as with Tibicen or the evening and morning as with Magicicada. Nymphs darken within an hour of emergence, and, like newly emerged Tibicen nymphs, are capable of weak flight within this period. 3. Okanagana oviposit in live twigs. Females excavate shallow, unbifurcated eggnests.

326 Errata

Sexual behavior in North American cicadas of the genera Magicicada and Okanagana

by

John Richard Cooley

Page 78: First line of results should read: ". . .matings lasted an average of 270.4 minutes. . ."

Table 4.1 (Page 90)

Average for 43 cases should read "270.42 ± 122.07"

Page 103: Line 4 should read: ". . . we made six daily collections. . ."

Page 118: Two lines omitted at end of page.

Last paragraph should read: ". . .and that the alterations in behavior are different for the two fungal life history stages in ways that are appropriate for the stage-specific modes of transmission. Our observations suggest that the first stage of the fungus has apparently evolved. . ."

Table 5.9 (Page 137)

P- value for M. cassini "click vs. control, first call bout: should read "P< 0.980."

Table 5.13 (Page 141)

Replicate "D" "no buzz" treatment: 1.93 ± 0.58 females responded to each call.

Figure 5.1 (Page 142)

Caption line 1 should read ". . . male begins court I (CI) calls. . ."

Caption line 4 should read " . . .male begins court III (CIII) calling. . ."

Figure 5.5 (Page 146)

Caption line 4 should read ". . . pairwise comparisons of treatments placed below. . . "

Table line 1 should read "Foreleg vibrate."

Figure 5.6 (Page 147)

Y axis label should read "Number of males"

"Less than or equal to" signs missing from caption

Caption line 7 should read ". . . treatments did not differ in either species."

Figure 5.9 (Page 150) X axis should read "Background intensity"

Page 247

First line, second paragraph should read: "Female O. canadensis and O. rimosa signal. . . "