INFORMATION TO USERS

This manuscript has been rqiroduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter 6ce, \ ^ e others may be from aity type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely afreet reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, b^inning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 Noith Zed} Road, Ann Aibor MI 48106-1346 USA 313/761-4700 800/521-0600

BEHAVIORAL, NEUROPHYSIOLOGICAL, AND BIOPHYSICAL STUDIES ON COMMUNICATION IN “EARLESS” FROGS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

By

Erik D. Lindquist, B.S., M.S.

The Ohio State University 1997

Dissertation Committee: Approved by Dr. Thomas E. Hetherington, Adviser

Dr. W. Mitchell Masters Adviser Ms. Sandra L. L. Gaunt, M.S. Zoology Department

Dr. Brian H. Smith UMI Number: 9801734

UMI Microform 9801734 Copyright 1997, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 ABSTRACT

Information on developmental stages (larvae, recent metamorphs and subadults) and ecology of the Panamanian golden frog [Atelopus zeteki) is presented. Tadpoles were found in stream pools away from swift torrents and recent metamorphs were observed on gravel bars adjacent to these pools.

Water quality of A. zeteki streams was found to have high dissolved oxygen content, neutral pH, low turbidity, and low levels of sulfates, nitrates, and reactive phosphorous.

As opposed to the aposematically colored adults, metamorphs and subadults of the two populations studied were cryptically colored, suggesting the absence of significant skin toxicity in juveniles. Subadult juveniles often were seen in close proximity to adult males. This is surprising because outside of courting gravid females, adult males are aggressive toward other conspecifics within sight.

Field studies were done on the visual and acoustic behavior of Atelopus zeteki, a species lacking a tympanic middle ear. Males displayed three stereotyped behaviors in response to playbacks of male pulsed vocalizations: foot signaling (semaphoring), vocalization, and repositioning orientation.

Frequencies of semaphoring and orientation responses were significantly increased by presentation of playback calls. Vocal responses also increased

(non-significantly) during the playback period and continued to increase during post-playback silence. These results provide the first experimental evidence that an "earless" anuran species displays behavioral responses to sound, and that vocalization mav play an important role in communication.

Orientation by the frogs toward the playback speaker suggests that this

"earless" species is capable of localizing a sound source.

Social communication in anuran amphibians is mediated predominantly by acoustic signals. The semaphoring observed in this species appears to represent visual signals used intraspecifically in male agonistic behavior. The use of visual signaling by A. zeteki may be correlated with their noisy, montane stream habitat. The communicative nature of semaphoring was supported by experimental manipulations using mirrored self-image presentations and nonresident introductions. Male frogs semaphored significantly more when presented with a mirrored self-image than with a nonreflective control. Novel encounters between resident males and nonresident frogs showed that semaphores were used directionally and were displayed toward target

m individuals. Females semaphored frequently and this observation represents a

rare case of signaling by females in a typically male-biased communicative

regime among anurans. Semaphore actions were clearly linked to a locomotory

gait pattern and appear to have originated as an elaboration of a standard

stepping motion. Laboratory studies were conducted on the acoustic

responsiveness of various body surface tissues and auditory midbrain

sensitivities to sound in species of Atelopus that either possess or lack a

tympanic middle ear. Acoustic responsiveness of body tissues was measured

with a laser Doppler vibrometer, and the body wall directly overlying the lung

was found to be highly responsive to sound in all species tested. The otic

(lateral head) region showed little responsiveness in earless species. The otic

region of the eared A.flavescens showed significant responses to high frequency

sound, but body wall responsiveness was higher than otic responses over the entire frequency range. Peak body wall responsiveness occurred within the dominant frequency range of each species’ advertisement vocalization. Peak responsiveness of the otic region of the eared A. flavesceiis also occurred within the dominant frequency range for the species advertisement vocalization.

Multi-unit neurophysiological recordings of the auditory midbrain {toms semicircularis) were obtained for one eared and two earless species of Atelopus.

Auditory sensitivity curves were constructed and showed three distinct regions

IV of sensitivity at low, middle, and high frequency ranges, the latter sensitivity falling within the dominant frequency range of each species’ advertisement vocalization.

Auditory sensitivity curves showed a close match with the acoustic responsiveness of the body wall of all species and with that of the otic region of the eared species. This correlation suggests that the body wall/lungs serves as the route of sound transfer to the inner ear in earless species and possibly in eared species as well.

The mechanisms of hearing in the earless frog Bombina orientalis were studied using laser Doppler vibrometric and neurophysiological techniques.

Laser vibrometry demonstrated that the body wall overlying the lung was highy responsive to sound compared to the lateral head surface overlying the inner ear. Detemtination of auditory midbrain thresholds found that covering the body surface between the forelimbs and hindlimbs dramatically decreased auditory sensitivity at all tested frequencies. Filling the lungs with oxygenated saline also drastically decreased hearing sensitivity. Tests showed that suturing the glottis closed had a minimal effect on midbrain responses to sound, suggesting that sound transfer from the lungs to the inner ear via the buccal cavity is relatively unimportant. The precise route of lung-based sound transfer remains unclear. The lung-based hearing system of the earless Bombina ürimtalis may represent the retention of a system used by primitive aquatic vertebrates that possessed a lung. Results of this study demonstrate that such a lung-based system can be used for hearing airborne sound as well.

VI Dedicated to mv wife Mollv

Vll ACKNOWLEDGMENTS

I wish to thank Thomas Hetherington for the constant support that he

provided me in the field and in the laboratory. I am grateful for his technical

and intellectual assistance during the experimentation, observation, analysis,

and dissemination of this dissertation research.

I thank A. Stanley Rand, Roberto Ibanez, and Neal Smith for their

technical support while working in Panama. Also I am indebted to the

logistical staff of the Smithsonian Tropical Research Institute for their help

and insight: namely Georgina de Alba, Maria Morello, and Gloria Maggiori.

I thank Marty Crump, Roy McDiarmid, Alan Jaslow, Roberto Ibanez,

Carlos Navas, and Robin Saunders for information and insight on the genus

Atelopus.

I thank Susan Volman, W. Mitch Masters, Kris Schuett, Andrew

Hocknian, and Jim Fox for their technical support related to laboratory experimentation.

I am grateful for the field assistance by Tom Hetherington, Alberto

Castillo, Dan Badgley, David Swihart, Molly Lindquist, David Dennis, and

v u I Vasco Morales. I also wish to recognize the superb illustrative and

photographic work provided by David Dennis.

I thank Brian Smith, Doug Nelson, and Paul Doherty for their assistance in

the statistical analysis of this research.

I am grateful to Sandra Gaunt, Doug Nelson and the Horror Laboratory of

Bioacoustics for use of recording, playback, and analysis equipment.

I thank the following people for their numerous contributions and

suggestions they made in revising the published portions of this dissertation:

A. Stanley Rand, Marty Crump, Roy McDiarmid, Tom Grubb, Jerry

Downhower, Sandy Gaunt, Susan Volman, and Alison Snow.

1 appreciate the support given by IN.RE.NA.RE. through the granting of

scientific permits along with their enthusiasm to educate the public on

threatened native amphibians. I thank José Cheng, and the national park staff

at Cerro Campana and El Cope for assisting me with endangered populations.

Funding for this research was provided by short-term graduate research

fellowships through the Smithsonian Tropical Research Institute, a Grant-in-

Aid of Research through Sigma Xi: The Scientific Research Society, Columbus

Zoo research grants, and a Graduate School Alumni Research Award from The

Ohio State Universitv.

IX Lastly and most importantly, I am indebted to my wife, Molly and our children for showing patience and flexibility in accommodating the necessities of graduate student life. Without their support and prayers, this research could not have been accomplished. 1 likewise am grateful for the support of my parents, parents-in-Iavv, family, and friends. VITA

December 30, 1967 Bom - Des Plaines, IL, USA

1991 B.S. Zoology, Southern Illinois University

1995 M.S. Zoology, The Ohio State University

1994/1996 Smithsonian Institution Short-term Graduate Fellow, Smithsonian Tropical Research Institute

1992 - present Graduate Teaching, Research, and Administrative Associate, The Ohio State Universitv

PUBLICATIONS

Research Publication

1. Lindquist, E. D. and T. E. Hetherington. In Press. Tadpoles and juveniles of the Panamanian golden frog, Atelopus zeteki (Bufonidae) with notes on ontogenetic shift in coloration and patterning. Herpetologica.

2. Lindquist, E. and D. W. Swihart. In Press. Atelopus chhiciuiejisis (Chiriquf Harlequin Frog). Mating behavior and egglaying. Herpetol. Rev.

3. Lindquist, E. D. and T. E. Hetherington. 1996. Field studies on visual and acoustic signaling in the "earless" Panamanian golden frog, Atelopus zeteki. J. Herpetol. 30:347-354.

XI 4. Lindquist, E. 1995. Atelopus zeteki (Panamanian Golden Frog). Pure tonal vocalization. Herpetol. Rev. 26(4):200-201.

FIELDS OF STUDY

Major Field: Zoology

XU TABLE OF CONTENTS

Page A bstract ...... ii

D edication ...... vii

Acknowledgments ...... viii

V ita ...... xi

List of T ab les ...... xvii

List of Figures ...... xviii

Chapters:

1. Introduction...... 1

Overview ...... 1 Communication in animals ...... 2 Acoustic communication in anuran amphibians ...... 3 Hearing in anuran amphibians ...... 4 Hearing in earless frogs ...... 5 Visual communication in anuran amphibians ...... 7 Research Plan ...... 9 Study subjects ...... 9 Specific a im s ...... 11

2. Tadpoles and juveniles of the Panamanian golden irog, Atelopus zeteki (Bufonidae) with infonnation on development of coloration and patterning ...... 13

A bstract ...... 13 Introduction ...... 14

XllI M aterials...... 16 Larval specim ens ...... 16 Water analysis ...... 16 Juvenile specimens ...... 17 Results and discussion ...... 17 Larvae ...... 17 Description ...... 17 Coloration in preservative ...... 20 Coloration in life ...... 2 I Ecology ...... 21 Water analysis ...... 2 1 Juveniles ...... 23 Recently metamorphosed juveniles ...... 23 Subaduits ...... 25 Ecology ...... 26 Implications of color variation ...... 26 Resum en ...... 32 Acknowledgements ...... 33 Bibliography ...... 34

3. Field studies on visual and acoustic signaling in the “earless” Panamanian golden frog, Atelopus zeteki ...... 37

A bstract ...... 37 Introduction ...... 38 Materials and methods ...... 41 Study sites ...... 41 Experimental design ...... 4 1 Results...... 43 Types of behavioral responses ...... 43 Correlations of behavior to playback vocalizations ...... 45 Discussion ...... 6 1 Pulsed vocalization ...... 62 Visual signaling ...... 62 Orienting behavior ...... 63 Interaction of visual and acoustic signaling ...... 64 Correlation of visual signaling and earlessness ...... 66 Resum en ...... 68 Acknowledgments ...... 69 Bibliography ...... 70

XIV 4. Semaphoring in an earless frog: The origin of a novel visual signal ...... 73

A bstract ...... 73 Introduction ...... 74 M ethods ...... 78 Study subjects and sites ...... 78 Mirrored self-image presentations ...... 78 Conspecific nonresident introductions ...... 79 R esults...... 80 Mirrored self-image presentation ...... 80 Conspecific nonresident introduction ...... 80 Directional use of semaphores ...... 83 Sex bias in semaphore u se ...... 87 Neural basis of semaphoring ...... 87 Discussion ...... 88 Semaphores as visual signals ...... 88 Sex bias in semaphore u se ...... 90 Sexual selection and semaphoring ...... 91 Acknowledgments ...... 92 Bibliography ...... 92

5. Biophysical and neurophysiological studies on audition in eared and earless anurans of the genus Atelopus ...... 95

A bstract ...... 95 Introduction ...... 96 M ethods ...... 99 A nim als ...... 99 Laser Vibrometry ...... 100 Neurophysiology ...... 101 R esults...... 104 Laser Vibrometry ...... 104 Neurophysiology ...... 111 D iscussion ...... 111 Comparative TS sensitivities ...... 111 Correlations between LDV measurements and TS sensitivity ...... 121 The structural basis of hearing in eared and earless species ...... 122 The functional basis of middle ear loss in frogs ...... 132

XV Acknowledgment ...... 133 Bibliography ...... 134

6. Lung-based hearing in an “earless” frog ...... 137

A bstract ...... 137 Introduction ...... 138 Materials and Methods ...... 141 Laser Doppler vibrometry ...... 141 Neurophysiological measurements of auditory sensitivity ...... 142 Attenuation of lung sound reception ...... 143 Obstructing sound transfer through the glottis ...... 145 R esults...... 145 Laser Doppler vibrometry ...... 145 Effects of covering the lateral body wall on hearing sensitivity ...... 145 Effects of lung filling on auditory sensitivity ...... 150 Effects of glottal obstruction on hearing sensitivity ...... 155 Discussion ...... 155 Acknowledgements ...... 163 Bibliography ...... 163

Complete Bibliography ...... 166

XVI LIST OF TABLES

Table Page

2.1 Water sample analyses from El Cope and Campana Heights. FTU = Foramizin Turbidity Units ...... 22

3.1 Descriptive statistics for each five minute period of experimentation (SI, S2, PI, P2, and PP) for behaviorally responsive male frogs (n= 1 T, n=7; n= 10); and paired t-test values comparing response means between the 10 minutes of silence (S=SI+S2) and the 10 minutes of playback (P=P1 + P2) for all frogs tested (Minitab rel. 10 for Windows). Total number of tested individuals; N= 15. A. Semaphoring Responses; 8. Vocal Responses; C. Orienting Responses ...... 59

3.2 Taxonomic distribution, morphological and ecological traits associated with foot signaling anurans. HABITAT: C=Cloudforesi, R= Rainforest, D=Tropical Dry Forest; M icrohabitat : C = Cascade, F=Forest, P=Pond, S=Stream, Sw=Swamp; Activity PATTERN; D=Diurnal, N=Nocturnal, ' Foot signals during 1st hours of daylight, ■ during moonlit nights;M iddle EAR: A=Absent, P = P resent ...... 67

xvii LIST OF FIGURES

Figure Page

2.1 Atelopus zeteki tadpole at stage 36 (Gosner, I960) from Parque General Omar Torrijos - El Cope National Park deposited in the herpetological museum collection at the Smithsonian Tropical Research Institute, Balboa, Panama as STRI-0311. Illustrations by David Dennis. Top, Dorsal view; Center, Lateral view; Bottom, Ventral view. Bar represents 1 m m ...... 18

2.2 Ontogenetic color and patterning change in A. zeteki. Photography by David M. Dennis ...... 27 A. Immaculate yellow adult male from El C ope ...... 28 B. Patterned adult male from Campana Heights ...... 29 C. Metamorphic juvenile from El Cope ...... 30 D. Subadult juvenile from Campana Heights ...... 31

3.1 Sequential illustration of fore-foot semaphoring by a male A. zeteki (see text for further d etails) ...... 44

3.2 Number of semaphores of all 15 frogs per minute of experimentation. The pre-playback period of silence is represented by minutes 1-10, the playback period (boxed area) is represented by minutes 11-20, and the post­ playback period of silence is represented by minutes 21 - 25. Each circle represents one semaphore. Frogs numbered 1 -8 represent animals from El Cope and those numbered 9-15 represent animals from Cerro Campana ...... 46

3.3 Number of vocalizations of all 15 frogs per minute of experimentation. The pre-playback period of silence is represented by minutes 1-10, the playback period (boxed area) is represented by minutes 11-20, and the post-

XVllI playback period of silence is represented by minutes 21 - 25. Each circle represents one call. Frogs numbered 1-8 represent animals from El Cope and those numbered 9-15 represent animals from Cerro Campana ...... 48

3.4 Number of turns of all 15 frogs per minute of experimentation. The pre-playback period of silence is represented by minutes 1-10, the playback period (boxed area) is represented by minutes 11-20, and the post­ playback period of silence is represented by minutes 21 - 25. Each circle represents one turn. Frogs numbered 1-8 represent animals from El Cope and those numbered 9-15 represent animals from Cerro Campana ...... 50

3.5 Total number of semaphores and semaphoring individuals (out of 15 frogs tested) per experimental period. The pre-playback period of silence and the playback period is divided into two, five minute periods each (SI and 32 and PI and P2, respectively), followed by a five minute post-playback period of silence (PP). Solid circles represent the total number of semaphores and the open circles represent the number of semaphoring individuals ...... 53

3.6 Total number of calls and calling individuals (out of 15 frogs tested) per experimental period. The pre-playback period of silence and the playback period is divided into two, five minute periods each (S1 and S2 and PI and P2, respectively), followed by a five minute post-playback period of silence (PP). Solid circles represent the total number of calls and open circles represent the number of semaphoring individuals...... 55

3.7 Total number of turns and turning individuals (out of 15 frogs tested) per experimental period. The pre-playback period of silence and the playback period is divided into two, five minute periods each (S1 and S2 and PI and P2, respectively), followed by a five minute post-playback period of silence (PP). Solid circles represent line

XIX represents the total number of turns and open circles represent the number of turning individuals ...... 57

4.1 Visually signaling golden frog showing the conspicuous forelimb rotation, referred to as semaphoring. Illustration by David Dennis ...... 75

4.2 Histogram illustrating the differential semaphore use between mirror and nonreflective control presentations. Each column represents one individual. Bars above and below the axis represents summed responses to mirrored self image and nonreflective control, respectively ...... 81

4.3 Illustration of 90° lateral fields of view in A. zeteki...... 84

4.4 Histogram illustrating the directional use of semaphores in relation to target animal position. Each column represents one individual. Bars above and below the axis represents summed semaphores scored as "onside" and "offside," respectively ...... 85

5.1 Mean acoustic responsiveness of tissue surfaces: solid line, body wall region (lung) with solid standard error bars; long dashed line, otic region with standard error bars; short dashed line, rostral region. Bar represents the dominant frequency range of the species advertisement call (Lescure 1981). AtelopusJlcivescens (n = 5 ) ...... 105

5.2 Mean acoustic responsiveness of tissue surfaces: solid line, body wall region (lung) with solid standard error bars; long dashed line, otic region; short dashed line, rostral region. Bar represents the dominant frequency range of the species advertisement call (C. Navas, pers. com.). Atelopus sp. (Chingaza, n = 4) ...... 107

5.3 Mean acoustic responsiveness of tissue surfaces: solid line, body wall region (lung) with solid standard error bars; long dashed line, otic region; short dashed line, rostral region. Bar represents the dominant frequency range of

XX the species advertisement call (R. Ibanez, pers. com.). Atelopus sp. (Nusagandi, n = 4) ...... 109

5.4 Multiunit auditory sensitivity threshold curves recorded from torus sanicircularis: upper solid line, mean auditory threshold with standard error bars; lower solid line, minimum auditory sensitivity for species. Bar represents the dominant frequency range of the species’ advertisement vocalization (Lescure, 198 1). Atelopus flavesceiis (n = 2 ) ...... 112

5.5 Multiunit auditory sensitivity threshold curves recorded from torus sanicircularis: upper solid line, mean auditory threshold with standard error bars; lower solid line, minimum auditory sensitivity for species. Bar represents the dominant frequency range of the species’ advertisement vocalization (C. Navas, pers. com.). Atelopus sp. (Chingaza, n=5) ...... 114

5.6 Multiunit auditory sensitivity threshold curves recorded from toms sanicirailaris: upper solid line, mean auditory threshold with standard error bars; lower solid line, minimum auditory sensitivity for species. Bar represents the dominant frequency range of the species’ advertisement vocalization (R. Ibanez, pers. com.). Atelopus sp. (Nusagandi, n=7) ...... 116

5.7 Mean TS auditory sensitivity for eared and earless species compared. ¥= A. flavesceiis; C= Atelopus sp. (Chingaza); N = Atelopus sp. (Nusagandi) ...... 118

5.8 Correlation between mean TS auditory threshold and surface responsiveness of the body wall/lung region: upper solid line, mean TS auditory threshold; lower solid line, mean acoustic responsiveness of body wall/lung surface; dashed line, mean acoustic responsiveness of tympanic region surface. Bar represents the dominant frequencv range of the species advertisement vocalization (Lescure, 1981). Atelopus flavescais...... 123

XXI 5.9 Correlation between mean TS auditory threshold and surface responsiveness of the body wall/Iung region: upper solid line, mean TS auditory threshold; lower solid line, mean acoustic responsiveness of body wall/lung surface. Bar represents the dominant frequency range of the species advertisement vocalization (C. Navas, pers. com.). Atelopus sp. (Chingaza) ...... 125

5.10 Correlation between mean TS auditory threshold and surface responsiveness of the body wall/lung region: upper solid line, mean TS auditory threshold; lower solid line, mean acoustic responsiveness of body wall/lung surface. Bar represents the dominant frequency range of the species advertisement vocalization (R. Ibanez, pers. com.). Atelopus sp. (Nusagandi) ...... 127

6.1 Laser Doppler vibrometric measurements of the acoustic responsiveness of the lateral head tissues (dashed lines) and body wall overlying the lung (solid lines) for three Bombina orientalis. Velocity is represented on a relative dB scale. The body wall over the lung is much more responsive than the lateral head surface and shows a low and high frequency peak in m otion ...... 146

6.2 Comparison of mean midbrain hearing thresholds when the body walls are covered (dashed line) and uncovered (solid line) by a layer of silicone grease. The lines represent the average of responses from five individuals, and the vertical bars represent the standard error. Covering the body wall dramatically increased auditory thresholds ...... 148

6.3 Comparison of midbrain hearing thresholds when the lungs are filled with air (solid lines) or with saline (dashed lines). Data from three individuals in which thresholds were measured first with the lungs filled with air and subsequently with the lungs filled with saline. Filling the lungs with saline dramatically elevated hearing thresholds. No midbrain responses could be recorded above 300 Hz in frogs with lungs filled with saline ...... 151

X.X11 6.4 Comparison of midbrain hearing thresholds when the lungs are filled with air (solid lines) or with saline (dashed lines). Data from two individuals in which thresholds were measured first with the lungs filled with saline and subsequently with the lungs emptied of saline. Hearing thresholds were raised in animals with lungs filled with saline. The effect of saline-filled lungs on hearing sensitivity in these two animals was not as great as that seen in Figure 6.3, probably because the lungs were not completely emptied of saline (see text) ...... 153

6.5 Comparison of mean midbrain hearing thresholds when the glottis is open (solid line) and sutured shut (dashed line). Data from five individuals. Suturing the glottis produced no significant effect on hearing thresholds...... 156

XXlll CHAPTER 1

INTRODUCTION

Overvieiv

This dissertation describes research that investigated communication and

the perception of acoustic signals in frogs of the genus Atelopiis that lack a

tympanic middle ear. These frogs raise interesting questions about the

mechanisms of perceiving airborne sound cues. In addition, because they also

employ visual signals for communication, and because frogs as a group relv

mostly on acoustic communication, species of Atelopus provide an opportunitv

to analyze the factors that underlie the evolution and use of a novel mode of

communication. Frogs of the genus Atelopus therefore allow one to investigate

the relative importance and environmental context of different signaling modalities, as well as providing an opportunity to more fully explore the functional basis of hearing in vertebrates. Communication in Animals .—The use of signals to convey information is ubiquitous among animals (Hasson, 1997; Hauser, 1996; see Complete

Bibliography for Introduction references). In order for a signaling system to be reliable, signals need to honestly convey important biological information such as mate quality, aposematism, aggression, etc (Hasson, 1997; Godfray, 1995;

Ryan, 1988). Signaling systems are dynamic in that numerous environmental factors influence the method of transmission and content of the messages animals seek to convey (Marier, 1977). The use of a particular signal modality is essential for optimizing reliability in communication (Smith, 1977; Hauser,

1996). For example, compared to other cues such as visual signals, acoustic signaling can be effective over relatively longer distances, in obstructed environments, and in extreme darkness. Nocturnal chorusing male frogs attract females from afar to a site of courtship and oviposition. On the other hand, when environments are acoustically saturated, as in areas of high wind or water noise, or by the acoustic signals of other animals, visual signals may provide greater reliabilitv. When the functional demands of signaling are consolidated with ecological, morphological, and physiological constraints, a complex communicative repertoire can arise (Smith, 1977; Rand, 1988). A complex set of factors may therefore determine the particular pattern of communication observed within a given taxon. Acoustic Communication in Anuran Amphibians .—Vocalizations plav an

important role in the behavioral repertoire of most anuran amphibians (Wells,

1977a, b; Gerhardt, 1982; Ryan, 1985; Rand, 1988). In frogs, this mode of communication predominantly serves a function in mate attraction, territoriality, courtship, aggression, release, and distress (Rand, 1988; Wells,

1977a, b). As an advertisement, calling is performed by males to attract distant females to the male’s location where courtship and perhaps amplexus and egg fertilization may ensue. Advertisement calling also serves to declare occupancy of a territory to males of the same or different species (Wells,

1977a, b). Some species modifv their advertisement calls into courtship calls as a female approaches the male in close range, while females of some species may indicate interest by calling as well (Wells, 1977b). In aggression, males and even females may use vocalization in defending nests or other vital resources, or in combative wrestling (Drewry, 1970; Jaslow, 1979; Townsend et al., 1984). Release calls are produced by uninterested individuals that resist amplexus. Finally, calling may be used in situations of distress such as seizure by a predator, evasion of a predator, or loss of a wrestling bout (Jaslow, 1979;

Duellman and Trueb, 1985; Hodl and Gollman, 1986). Not only does calling convey behavioral information from sender to receiver, but also may provide information about position, identity, sex, size, fitness, and physiological condition of the sender (Rand, 1988). Morphologically, males have well

developed laryngeal apparatuses that allow them to produce a variety of

vocalizations. These structures are developmentally rudimentary in females

and function in only a few species (Wells, 1988; Stewart and Rand, 1991; Roy

et al., 1995; Emerson, pers. com.). The usual anuran regime of intraspecific

acoustic communication is sex biased in that males may be both senders or

receivers, whereas females are solely receivers.

Hearing in Amiran Amphibians .—In most anuran amphibians, reception of airborne sounds is accomplished through a tympanic middle ear that consists of an external tympanum, middle ear cavity, and auditory ossicle (columella or stapes), and can transfer sound energy to the inner ear. Sound pressure waves striking the tympanum produce columellar motion that displaces endolymphatic fluid at the oval window of the inner ear. This fluid displacement drives resonation of the auditory papillae, where auditory hair cells transduce acoustic stimuli into action potentials that are sent to the higher auditory system in the brain. Typically three auditory papillae are present and they respond best to discrete ranges of acoustic frequencies. The saccule, amphibian, and basilar papillae are tuned to low, middle and high frequency ranges respectively (Feng et ai., 1975: Moffat and Capranica, 1976;

Wilczynski and Capranica, 1984; Zakon and Wiiczynski, 1988).

Although acoustic signals are utilized as the dominant mode of

communication, several vocalizing genera completly lack a tympanic middle

ear (Jaslow et al., 1988). This condition is often referred to as “earless,”

although standard inner ears, containing well-developed auditory papillae, are

present (Wever, 1985; Jaslow et al., 1988).

Hearing in Earless Frogs.—’’Earless” frogs that vocalize raise interesting

questions concerning their relative auditory sensitivity and mechanisms of

sound reception. Wever (1985) was able to measure cochlear microphonie

responses to airborne sound in several species of earless Atelopus, and studies of

midbrain responses by Jaslow and Lombard (1996) found that A. cliiiiquiensis is

surprisingly sensitive to sound. The level of auditory sensitivity observed

suggests that effective nontyntpanic pathways of sound transfer are present in

frogs.

Recent research has demonstrated that some frogs may use the lateral body wall and lungs as an additional route for sound reception (Narins et al.,

1988; Ehret et al. 1990; Hetherington, 1992; Ehret et al., 1994).

Measurements of the acoustic responsiveness of body surfaces using a laser Doppler vibrometer have demonstrated that the lateral body wall immediately

over the lungs is highly responsive to sound (Narins et al., 1988; Jorgensen,

1991; Hetherington, 1992; Ehret et al., 1994). In eared species, pressure waves

within the lungs can pass forward through the glottis, mouth cavity, and

Eustachian tubes to the middle ear cavity and affect tympanic motion (Ehret

et al., 1990; Jorgensen et al., 1991). This additional route of sound input to

the tympanum may have important consequences for directional characteristics

of the tympanic middle ear (Narins et al., 1988). Ehret et al. (1994) also

demonstrated that attenuation of high frequency sound along the lung- eardrum pathway is pronounced, so that the contribution of the lung to tympanic responsiveness is restricted to relatively low frequencies. It has been hypothesized that the body wall and lungs can transfer sound energy to the inner ear of earless species of frogs (Narins et al., 1988; Hetherington, 1992), although there has been no experimental demonstration of the precise route.

Earless frogs also test assumptions of the structural basis of directional hearing in frogs. All hypothetical models of sound localization (phonotaxis) in frogs rely upon the presence of a tympanic middle ear to behave as a directionally sensitive pressure-gradient receiver (Eggemtont, 1988). Acoustic signal detection without the perception of spatial infomiation may be of little selective advantage, so it seems conceivable that earless frogs that vocalize are capable of localizing a sound source. There has yet to be a hypothesis of how phonotaxis may be achieved by anurans that do not have a standard tympanic middle ear. If these “earless” frogs can detect and localize an acoustic source, current perspectives on the mechanisms of hearing and sound localization will need to be re-evaluated. Therefore, understanding the alternative auditory pathways utilized by earless anurans may significantly change our understanding of how auditory systems function.

Visual Communication in Anuran Amphibians. — Frogs in the genus Atelopus also appear to use visual signals in the form of foot waving (semaphoring) to communicate. The use of foot signals for visual communication in anurans is rare and has onlv recently been documented in six genera other than Atelopus.

Crump (1988) described instances where foot displays accompanied vocalizations in A. varius and also documented an association of foot signaling with intraspecific male-male aggression in this species. All of the species that have been reported to use visual signaling are diurnal except some species in the genus Litoria, which perform visual signals in moonlight (Richards and

James, 1992). All but one of these visually signaling species occur in cascading stream environments where acoustic signals may be masked by high levels of ambient noise (Heyer et al., 1990). The occurrence of this potential signaling system in animals that appear to rely heavily upon acoustic communication raises interesting questions regarding how animal systems may resolve problems of signal reliability and detection in habitats of high ambient noise and how they may adaptively switch between communication modalities. Foot displays have been observed during courtship and male-male aggression in

Staiirois (Harding, 1982; Davison, 1984), and during male aggression in Litoria and Brachycephaliis (Richards and James, 1992; Pombal et al., 1994), but the context of visual display behavior remains unclear in other species in which it has been observed.

The circumstances under which visual signaling is used in Atelopus (or in any frog) have not been experimentally examined, but many hypotheses may be put forth to explain their use of visual signals. These frogs are brilliantly colored and highly toxic (Brown et al., 1977) and foot signaling mav draw attention to their aposematic (warning) coloration. Such signaling may provide interspecific warning of unpalatability and toxicity, but this function would not explain intraspecific use. Alternatively, visual signaling may have arisen in this genus as a response to the presence of high background noise and/or the loss of the tympanic middle ear. The latter appears to be less likely as all other visually signaling anurans have complete tympanic middle ears.

The hypothesis put forth by Heyer et al. (1990), that visual signaling has

8 evolved in species living along noisy cascade habitats where acoustic signals are

masked and less reliable than visual signals, appears more likely to be

applicable. The relative importance or preferential use of acoustic and visual

signals, and the degree of interaction between visual and acoustic sensory

modalities, has yet to be empirically examined in anurans that display both

types of communication. The biological significance of visual displays also

needs to be examined in the context of epiphenomenal signaling. This

dissertation research provides the first of such investigations into visual

signaling in anuran amphibians.

Research Plan

Study Subjects .—The primary study subjects in this dissertation studv were

harlequin frogs of the neotropical genus Atelopus, which are taxonomicallv

placed in the toad family Bufonidae. This genus, with few exceptions, is

restricted to montane wet forests and paramo (treeless highlands), and life history observations are less numerous than for other anuran groups (Lotters,

1996; McDiamiid, 1971 ). Most species in the harlequin frog genus Atelopus, such as the Panamanian golden frog, A. zeteki (in the A varius group) lack the tympanic middle ear pathway and yet vocalize (McDiamiid, 1971 ; Jaslow,

1979; Cocroft et al., 1990). There are, however, also species in this genus. such as A. flavescens (in the A. flavescens group) that have a nearly complete tympanic middle ear (McDiarmid, 1971; Lescure, 1981a). A middle ear cavity and columella is present, but the extracolumella attaches only to unspecialized skin rather than to a more specialized tympanum. The presence of eared and earless species in the same genus is useful for comparative study of auditory function in closely related and very similar animals with very dissimilar auditory morphologies. In fact, Atelopus appears to be the only vertebrate genus in which a direct comparative study of sound reception strategies can be conducted on congeners with and without tympanic middle ears.

Male Atelopus produce a variety of vocalizations, although the functional significance of their calls has not been elucidated. Across the genus, three general vocalization patterns have been described: the pulsed call, used primarily in advertisement and aggression; the pure tonal whistle, likely used as a distress signal by males during and after losing agonistic wrestling bouts; and the short call, produced when seized, disturbed, put in close containment with conspecifics, or while wrestling another male; (Jaslow, 1979; Lescure,

1981a; Crump, 1988; Cocroft et al., 1990; Lindquist, 1995). Although vocalizations in Atelopus have been documented primarily in cases of male-male aggression and disturbance (Jaslow, 1979; Crump, 1988; Cocroft et al., 1990), males of two South American species, A. pulcher (= A. spumarius) and A.

10 flavescms, appear to use pulsed calls in mate attraction (Jaslow, 1979; and pers. obs). Females are not reported to vocalize in this genus. All known vocalizations in Atelopus have dominant carrier frequencies above 1500 Hz.

Neurophysiological experiments were conducted using another vocalizing earless species, Bombina orientalis, which is more readily available. This unrelated Asian species is semi-aquatic, lacks a tympanum and middle ear cavity, and has only a reduced columella not well connected to the oval window of the inner ear (Wever, 1985; Jaslow et al., 1988). Although

Eustachian tubes are present, there is no middle ear cavitv to which they normally would connect (Jaslow et al., 1988). Therefore the species can be considered to be functionally “earless.”

Vocalization in B. orientalis is similar to Atelopus in that several call types are present (pers. obs.). These calls serve a variety of functions such as advertisement for courtship, territory, and release (Walkowiak, 1980). The dominant frequency of these vocalizations is considerably lower (500-1000

Hz) than that of Atelopus.

Specific Aims.—The research presented here investigated hearing and the use of both acoustic and visual communication in the genus Atelopus through behavioral, neurophysiological, and biomechanical tests. Behavioral field

11 research on the earless A. zeteki focused on ( I ) examining the effects of

intraspecific vocalizations on behavior, (2) determining the ability of the frogs

to localize sound, (3) determining whether semaphoring is a fomi of visual

communication, and (4) investigating the interaction between visual and

acoustic communication. Neurophysiological and functional studies on

Atelopus ( 1 ) compared the acoustic sensitivity of the auditory midbrain in eared

and earless species, (2) investigated the acoustic responsiveness of various bodv

tissue surfaces in eared and earless species, and (3) correlated

neurophysiological sensitivities and tissue responsiveness to suggest potential

pathways of sound transfer in earless species. Additional neurophysiological experiments on Bombina tested the effectiveness of different potential pathways of sound transfer. Chapters two through six of this dissertation is comprised of

five papers that have been submitted for publication in scientific journals.

Chapter two considers the importance of color in indirect visual signaling, a peripheral, yet integral topic to the total communication system \n Atelopus.

12 CHAPTER 2

TADPOLES AND JUVENILES OF THE PANAMANIAN GOLDEN FROG. ATELOPUS ZETEKI (BUFONIDAE) WITH INFORMATION ON DEVELOPMENT OF COLORATION AND PATTERNING

Abstract.—Published information regarding the life history and ecology of

the endangered Panamanian golden frog, Atelopus zeteki is completely absent.

Information on developmental stages (larvae, recent metamorphs and

subadults) and ecology of the A. zeteki is presented. Tadpoles were found in

stream pools away from swift torrents and recent metamorphs were observed

on gravel bars adjacent to these pools. Water quality of A. zeteki streams was

found to have high dissolved oxygen content, neutral pH, low turbidity, and

low levels of sulfates, nitrates, and reactive phosphorous.

As opposed to the aposematically colored adults, metamorphs and subadults of the two populations studied were cr\'pticallv colored, suggesting the absence of significant skin toxicity in the latter. Subadult juveniles often were seen in close proximity to adult males. This is unexpected because

13 outside of courting gravid females, adult males are aggressive toward other

conspecifics within sight.

INTRODUCTION

Information on the life history and reproductive biology of the neotropical

anuran genus Atelopus is limited (McDiarmid, 1971). This genus, with few

exceptions, is restricted to montane wet forests and paramo, and life historv

observations are less numerous than for other anuran groups (Lotters, 1996;

McDiamiid, 1971). A. zeteki Dunn (1933), inhabits streams along the

montane slopes of the Central Cordilleran cloud forests of western-central

Panama (Cocroft et al., 1990; Miller, 1987; Savage, 1972; pers. obs.). Even

less infomiation regarding the life history and ecology of this endangered

species is available. This report describes newly discovered developmental

fomis and provides ecological information on larval and juvenile A. zeteki

collected during a study conducted over the initial three months of the wet

season (April - July 1996) in populations found at the westem and eastern

extent of the species’ known range.

Several tadpoles, newly transfomied juveniles, subadult juveniles, and adults were seen at streams in Parque General de Division Omar Torrijos-El Cope (El

14 Copé), Provincia de Codé, and Parque Nadonal Altos de Campana (Campana

Heights), Provincia de Panama. Exact localities are withheld due to the continued threat of illegal poaching (R. Arias de Para, pers. com.; pers. obs.).

Frogs were found at streams in both primary and secondary growth low elevation cloud forests (Myers, 1969), although on a few occasions some individuals were encountered along streams passing through narrow patches of disturbed forest bordered by cattle pasture (pers. obs.). Although all adults and juveniles were found within three meters of montane streams, pulsed calls given by males deep in the forest far away from water (>50 m) were heard occasionally. These two observations are consistent with reports bv Dunn

(1933) and others (Angher, G. and Eisenmann, D., pers. com.; Dole and

Durant, 1974).

Tadpoles and recent metamorphs have not been described for A. zeteki.

Tadpoles have been described for A. bcilios (Coloma and Lotters, 1996), A.

(Mebs, 1980), A. flavesceiis {Lescure, 1981), A. (Grav and

Canatella, 1985), A. pulcher {Gascon, 1989), A. suboniatus {Lynch, 1986), A. varius (Starrett, 1967), A. spumarius, A. certus,A. iguescens (Duellman and

Lynch, 1969), and A. friivfor (Lavilla et al., 1997). Lotters (1996) has compiled a comprehensive summary of morphological features for most described tadpoles of the genus. Metamorphic juveniles have been described

15 only for A balios (Coloma and Lotters, 1996). Dunn ( 1933) described the

appearance of juveniles of A. zeteki as having “the dark predominating,” but

provided no detailed information as to their size, coloration, and patterning.

MATERIALS AND METHODS

Larval Specimens.—Due to the endangered status of A. zeteki, the following tadpole descriptions and measurements (mm) are based on only four individuals between stages 25-45 (Gosner, I960). Specimens have been deposited in the herpetological museum collection at the Smithsonian Tropical

Research Institute, Balboa, Panama as STRI-0311. All tadpole specimens were collected from a stream in the park at El Copé at an altitude of 450 m above sea level. Tadpoles were fixed in 10% fomtalin.

Water analysis.—Water samples from El Copé were obtained on 5/7/96 and tested on 5/10/96. Water samples from Campana were obtained on 5/14/96 and tested on 5/15/96. Water was taken from the center of pools in a 250 ml vacuum sealed plastic bottle. W ater quality of pool samples was assessed using a HACH DR/2000 Spectrophotometer at 23 C at the STRI aquatic laboratorv'^ at Barro Colorado Island, Panama. Water analysis values using the HACH

16 DR/2000 were averaged between two independently assessed readings from samples. The pH was assessed once at the field site and is presented in Table

2.1 (p. 22) along with other measured physical parameters.

Juvenile Spedmefis .—Recent metamorphs and subadult juveniles were caught and measured alive and released at El Copé and Campana Heights. Newly metamorphosed specimens with tail vestiges were obtained at El Copé. The terminology of larv^al features follows Coloma and Lotters (1996), Duellman and Lynch (1969), and Altig and Johnston (1989) and the developmental stages correspond to those of Gosner (1960). Measurements and weights of specimens were made to the nearest 0.1 mm with a Manostat dial caliper and to the nearest 0.1 g with an A\inet ( 10 g) scale.

RESULTS AND DISCUSSION

Larvae

Description .—Measurements (mm) of four developmental stages of a series of four tadpoles from El Copé are given in the following description. This description and Fig. 2.1 is based on a single indhidual at developmental stage

36. Atelopus zeteki tadpole is characterized as gastromyzophorous according to

17 4^ ^ ■

Figure 2.1. Atelopus zeteki tadpole at stage 36 (Gosner, I960) from Parque General Omar Torrijos - El Copé National Park deposited in the herpetological museum collection at the Smithsonian Tropical Research Institute, Balboa, Panama as STRI-0311. Illustrations by David Dennis. Top, Dorsal view; Center, Lateral view; Bottom, Ventral view. Bar represents I mm.

18 Altig and Johnston’s ecological guild ( 1989) and is a Type-IV larva as

categorized by Orton (1953). Body length (snout apex - vent tube base) 5.8

mm, body width 4.3 mm, total length 12.2 mm. Body ovoid, depressed,

flattened ventrally, about 3/5 as high as wide; greatest width at about one half

of body length. In dorsal view snout broadly rounded; dorsal contour

gradually curv'ed anteroventrally to tip of snout; bodv slightlv expanded just

posterior to eyes; nostrils small, closer to eyes than to tip of snout; eyes dorsal,

directed anterodorsolaterally; moderately large, non-bulging, separated bv a

distance 1.5 times the diameter of eye. Spiracle sinistral, elongate, ventral to

the horizontal body axis, directed posteriorly originating about one half of

body length; vent tube moderate in length, medial. Caudal musculature well pronounced on anterior half of tail; musculature deep on anterior half, narrowing abruptly at midlength of tail, extending nearly to tip of moderately rounded tail; caudal fins tallest at about three-fifths of length of tail; dorsal fin arched throughout posterior third; ventral fin not varying in height except for the posterior fifth; dorsal fin temiinating anteriorly to juncture of tail and body; ventral fin temiinating posterior to body.

Mouth large, ventral, surrounded by labia fomiing an unbroken oral disc

3.6 mm wide; posterior lip with no papillae; other lips bearing a single row of small blunt papillae. Labial tooth fomiula is 2/3 as is typical for most Atelopus,

19 rows complete, roughly equal in length, first upper row angulate; beaks thin, smooth; upper beak parallel to arc of snout (from ventral aspect), about four-

fifths length of broadly U-shaped lower beak. Large ventral suctorial disc extends and broadens from posterior labium, widest anteriorly, extending two- thirds length of body, forming a completely round sucker; suctorial disc lacks papillae.

Developmental stage 25 has a body length of 2.7 mm and a total length of

7.1 mm, and has its greatest width at 2.1 mm. Developmental stage 27 has a body length of 4.6 mm and a total length of 9.7 mm, and has its greatest width at 2.8 mm. Developmental stage 45 has a body length of 7.6 mm and a total length of 9.4 mm, and has its greatest width at 2.3 mm.

Coloration in presenmtive .—Dorsum dark brownish gray and is flanked with scattered, lighter gray areas, nearly unpigmented ventrally; venter translucent; suctorial disc lacking pigment except for a small central spot; gut clearly visibly; spiracle unpigmented; hind limbs covered laterally with minuscule dark marks; toe tips pigmented; tail pigmented dorsally and laterally; posterior two- thirds of tail unpigmented.

20 Coloration in life .— Dark brown to black dorsally with metallic gold flecks;

venter pale gray to translucent. A metamorphosing tadpole at stage 45 had

lost the gold flecking, had small dark green markings against the dark brown

dorsum, and overall was slightly darker than fully metamorphic juveniles found

at the same site (see description below).

Ecology .—Tadpoles of A. zeteki were commonly found resting on the top of

stones and stream gravel at the edges of shallow pools below cascades, similar

to reports for A. certiis by Duellman and Lynch (1969). They were found

anywhere along the stream where water had the chance to pool, as long as the pool was directly connected to flowing channels. Tadpoles were not observed in flowing channels and were found in water depths between 5 and 35 cm.

Examination of deeper levels revealed no tadpoles.

Water analysis.—W ater samples were drawn at two areas at each study site; one at a locality signifying the upstream extent of where individuals were found, and another in the approximate center of each study site (see Table

2.1). Temperatures taken throughout the duration of study at both sites averaged 20.9 C (range 20.4-21.3 C; n = 11 ) for water and averaged 24.2 C

(21.5-28.4 C; n = 7) for air. Atelopus zeteki streams showed comparable pH

21 (7.04-7.76) values to those reported for streams inhabited by Æ balios (pH

7.7; Coloma and Lotters, 1996). Dissolved oxygen levels for A. zeteki

streams (x = 6.99 m ^, n = 4) were intemtediate to values reported for

streams that A. balios (4.0-5.0 mg/1) SindA. Jlavescais (7.5 m ^) inhabit

(Coloma and Lotters, 1996; Lescure, 1981).

El Copé Campana Heights Parameter [±SE] Upstream Midstream Upstream Midstream pH 7.04 7.49 7.76 7.6

Dissolved [ ±0.201 mg/L 6.7 6.75 7.3 7.2

Nitrates (NO^ ) [±o.o3| 0.3 0.3 0.3 0.3

Sulfates (SO / ) [± 2 .2 ]mg/L 0.0 0.5 1.0 1.0

Phosphates (PO/^) [±0 .0 1 1 mg/L 0.015 0.080 0.075 0.025

Turbidity [±2.0] FTU 1.5 0.5 0.0 1.0

Elevation m above sea level 600 350 720 500

Table 2.1. Water sample analyses from El Copé and Campana Height. FTU = Fomiazin Turbidity Units.

22 Juveniles

Recently metamorphosed juveniles .—Metamorphic juveniles were observed at both field sites and were representative of stage 46+ (Gosner, 1960). During the course of study, El Copé metamorphic juveniles ranged between 8.4 and 14 mm SVL (snout-vent length) (x = 10.20 ± 2.09 s.d.; n = 9). Campana metamorphic juveniles were slightly larger, measuring between 14.5 and 17.1 mm SVL (x = 15.73 ±1.31 s.d.; n = 3). Metamorphic and subadult juvenile frogs were too small to obtain accurate weights in the field. The size difference may be explained by earlier breeding at Campana and it seems likely that the

El Copé animals are more representative of recently metamorphosed individuals.

Dorsal ground coloration in life of recent metamorphs at El Copé was a deep, vivid green. This color generally matched the green of mosses growing on stones in and around the streams. Dark brown to black dorsal markings were invariably present and always included: a rostral spot or transverse band, a distinct interorbital “X” reaching from the crest of the eyes posterior to the suprascapular region, a single mid-dorsal (lumbar) chevron-like band opening posteriorly and terminating at the flank, a tranverse postsacral band, and 1 -3 complete lateral bands surrounding amis and legs. Some individuals bore small dark markings on digits. Numerous brick-red warts were scattered

23 within the dark dorsal markings. The ground color of the venter was either white or goldenrod yellow (6 white individuals and 3 yellow individuals). Dark spotting was found on the venter of some individuals and was not correlated with ground color. Individuals possessing a white venter had a lemon to goldenrod yellow anal patch. Palmar and plantar surfaces were goldenrod yellow.

The metamorphosed juveniles found at Campana Heights resembled those at El Copé except that the dorsal ground color was a lemon to goldenrod yellow rather than green. The dark dorsal markings were much wider, so the extent of the ground color was more limited. The venters and palmar and plantar surfaces of the Campana Heights juveniles were goldenrod. The occurrence of two distinctly colored venters in El Copé metamorphs (a white venter with lemon to goldenrod yellow anal patch, or a completely yellow to goldenrod venter) is interesting. Similar differences in ventral coloration related to sexual dimorphism are seen in adult A. limosus and Atelopus sp. from

Nusagandi, Kuna Yala, Panama (Ibanez et al., 1995; pers. obs.). However, this juvenile character was only seen in El Copé metamorphs and is completely absent in adults, so the significance of the two color patterns is unclear.

24 Subadults .—I saw several subadults at Campana Heights and one subadult

at El Cope during the month of April 1996. The El Cope individual measured

28.3 mm and weighed 1-1 g and more closely resembled in color (greenish)

metamorphic juveniles than the typical goldenrod to yellow adults. In

addition to this individual, I had observed a yellow and black subadult of

comparable size in July of 1994, but lack data on its length and weight. The

SVL of Campana Heights subadults averaged 24.9 mm and weighed 1.1 g (n =

8). Their coloration was similar to that of adults, but the dark dorsal markings

were distinctly wider in proportion to the yellow, although narrower than

those of smaller juveniles (see above). Adult A. zeteki may be either

immaculate yellow or yellow with black markings. Patterned adults have

narrower black markings than both metamorphic and subadult juveniles (see

Fig. 2.2, pp. 27-29). Measurement of the mean percentages of dark coloration

along the vertebral axis (percentage of SVL along dorsal midline) found that

subadults (x = 60.00%, s.d. = 6.29; n = 8) had significantly more dark

patterning than adults (x = 28.57%, s.d. = 16.79; n = 8; P = 0.0022, paired t-test) in the Campana Heights population. Although not many subadults were seen at El Cope, it is likely that the same phenomenon exists there as well, as El Cope adults are frequently entirely goldenrod.

25 Ecology. —Observ'ed juveniles were always within 2 m of the stream. At El

Cope, recently metamorphosed juveniles usually were next to stream pools

teeming with tadpoles. Atelopus zeteki subadults seen during April 1996 at

Campana Heights often were seen in the immediate vicinity of, and sometimes

even touching, an adult male. This observation was surprising in that adult

male A. chiriquieiisis, A. varius, and A. zeteki are normally aggressive and

typically do not allow conspecifics other than gravid females in close proximity

(Crump, 1988; Jaslow, 1979; pers. obs.). At the onset of consistent and heav\'

rains, all juveniles disappeared from open streamside areas frequented by adult

males. It may be that territorial behavior in adult males is initiated bv these

rains.

Implications of color variation.—Daly et al. (1994) discovered a dietary

uptake system for alkaloid toxin deposition in the skin of several dendrobatid

frogs. The potent tetrodotoxin also appears to be acquired by dietarv or

symbiotic bacterial uptake in A. varius from Fortuna, Panama (Daly et al., in

press). Adult A. zeteki are very poisonous, with at least two known

atelopidtoxins (zetekitoxin AB and zetekitoxin C), tetrodotoxin, and

bufadienolides present in the skin (Brown et al., 1977; Fuhrman et al., 1969).

The cryptic coloration of metamorphs and young juveniles ( Fig. 2.2 C & D)

26 Figure 2.2. Ontogenetic color and patterning change in A. zeteki. Photography by David M. Dennis

A. Immaculate yellow adult male from El Cope B. Patterned adult male from Campana Heights C. Metamorphic juvenile from El Cope D. Subadult juvenile from Campana Heights

27 Figure 2.2 A.

2 8 Figure 2.2 B.

29 Figure 2.2 C.

30 Figure 2.2 D.

31 may be related to a lack of significant skin toxicity and/or dietary uptake of

carotenoids (Goodwin, 1984). I hypothesize that skin toxins accumulate with

age and that aposematic (yellow) coloration develops while the proportion of

dark dorsal patterning decreases (see Fig. 2.2.). Although direct testing of this

hypothesis on the endangered A. zeteki is unlikely, examination of the toxicity

of juveniles of other species of Atelopus could resolve the question. The developmental shift in color of A. zeteki from cryptic to aposematic may be comparable to that of Phyllobates terribilis, whose juveniles are more cryptic and less toxic than the adults (Myers et al., 1978). Behavior of A. zeteki may shift developmentally in relation to skin toxicity and aposematic coloration. This hypothesis may explain the lack of observations of juveniles; metamorphs and young juveniles are secretive and typically only the adults move about conspicuously in the open.

RESUMEN

Informacidn sobre las etapas de desarrollo (larvas, juveniles recién transfomiados y subadultos) y ecologia de la rana dorada de Panama {Atelopus zeteki) es presentada. Los renacuajos fueron encontrados en las piletas de las quebradas, lejos de corrientes fuertes, y los juveniles recién transfomiados

32 fueron observados en los bancos de grava a los lados de estas piletas. La

calidad de agua de las quebradas fue determinada, esta tema un contenido de

oxigeno disuelto alto, pH neutro y niveles bajos de sulfatos, nitratos, fôsforo

reactivo y turbidez.

A diferencia de los adultos, que tienen una coloraciôn aposemâtica, los metamorfos y subadultos de las dos poblaciones estudiadas tenian una coloraciôn criptica, lo que sugire la ausencia de toxicidad significativa en la piel de éstos ûltimos. Los juveniles subadultos fueron vistos frecuentemente creca de machos adultos. Esta observadôn es inusual, porque exceptuando a las hembras gravidas que estân siendo cortejadas, los machos son agresivos hacia cualquier otro individuo coespecifico dentro de su campo visual.

Acknoivledgttients .—1 thank A. Stanley Rand and Roberto Ibanez and the logistical staff of the Smithsonian Tropical Research Institute (STRI), Balboa,

Panama for their help and insight. I am grateful for the field assistance by

Tom Hetherington, Alberto Castillo, Dan Badgley, and David Dennis. We also wish to recognize the superb illustrative and photographic work provided by David Dennis. Funding for this research was provided by a short-term graduate research fellowship through STRI, a Grant-in-Aid of Research through Sigma Xi, The Scientific Research Society, and a Graduate School

33 Alumni Research Award from Ohio State University. I particularly appreciate

the support given by IN.RE.NA.RE. through the granting of scientific permits along with their enthusiasm to educate the public on threatened native amphibians.

BIBLIOGRAPHY

Altig, R. and G. F. Johnston. 1989. Guilds of anuran larvae: Relationships among developmental modes, morphologies, and habitats. Herpetol. Monogr. 3:81-109.

Brown, G. B., Y. H. Kim, H. Küntzel, and H. S. Mosher. 1977. Chemistry and pharmacology of skin toxins from the kog Atelopus zeteki (Atelopidtoxin: Zetekitoxin). Toxicon 15:115-118.

Cocroft, R. B., R. W. Mcdiarmid, A. P. Jaslow, and P. M. Ruiz-Carranza. 1990. Vocalizations of eight species o( Atelopus (Anura: Bufonidae) with comments on communication in the genus. Copeia 1990:631-643.

Coloma, L. A., and S. Letters. 1996. The tadpole o(Atelopus balios (Anura: Bufonidae) from the Pacific lowlands of Ecuador. Herpetologica 52:66- 70.

Crump, M. L. 1988. Aggression in harlequin frogs: Male-male competition and a possible conflict of interest between the sexes. Anim. Behav. 36:1064- 1077.

Daly, J. W., S. 1. Secunda, H. M. Garraffo, T. P. Spande, A. Wisnieski, and J. F. Cover Jr. 1994. An uptake system for dietary alkaloids in poison frogs (Dendrobatidae). Toxicon 32:657-663.

Daly, J. W., W. L. Padgett, R. L Saunders, and J. F. Cover Jr. Absence of tetrodotoxins in a captive-raised riparian frog, Atelopus varius. Toxicon: In press.

34 Dole, J. W„ and P. Durant. 1974. Movements and seasonal activity of Atelopus oxyrhynchus_{Anura: Atelopodidae) in a Venezuelan cloud forest. Copeia 1974: 230-235.

Duellman, W. E., and J. D. Lynch. 1969. Descriptions o f Atelopus tadpoles and their relevance to atelopodid classification. Herpetologica: 25 231 -240.

Dunn, E. R. 1933. Amphibians and reptiles from El Valle de Anton, Panama. Occas. Pap. Boston Soc. Nat. Hist. 8:65-79.

Fuhrman, F. A., G. Fuhman, and H. S. Mosher. 1969. Toxin from the skins of frogs of the genus Atelopus: Differentiation from dendrobatid toxins. Science 165:1376-1377.

Gascon, C. 1989. The tadpole o f Atelopus pulcher Boulenger (Anura, Bufonidae) from Manaus, Amazonas. Revta. Bras. Zool. 6:235-239.

Gosner, K. L. 1960. A simplified table for staging anuran embry^os and larv'ae with notes on identification. Herpetologica 16:183-190.

Goodwin, T. W. 1984. The Biochemistry of the Carotenoids 2"^^ ed.. Chapman and Hall, New York, New York.

Gray, P. And D. C. Cannatella. 1985. A new species o f Atelopus (Anura: Bufonidae) from the Andes of nonhem Peru. Copeia 1985:910-917.

Ibanez, R. D., C. A. Jamarillo, and F. A. Solis. 1995. Una especie nueva de Atelopus (Amphibia: Bufonidae) de Panama. Carrib. J. Sci. 31:57-64.

Jaslow, A. P. 1979. Vocalization and aggression m Atelopus chiriquiemis (Amphibia, Anura, Bufonidae). J. Herpetol. 13:141-145.

Lavilla, E. O., R. O. de Sa, and I. De la Riva. 1997. Description of the tadpole of Atelopus tricolor. J. Herpetol. 31:121-124.

Lescure, J. 1981. Contribution à l’étude des Amphibiens de Guyane Française IX. Le têtard gastromyzophore 6'Atelopusflavescens Duméril et Bibron (Anura, Bufonidae). Àmphibia-Reptilia 2: 209-215.

35 Letters, S. 1996. The Neotropical Toad Genus Atelopus: Checklist-Biolog)^- Distribution. M. Vences and F. Claw Verlags. Koln, Germany.

Lynch, J. D. 1986. Notes on the reproductive biology o i Atelopus siiboniatus. J. Herpetol. 20:126-129.

McDiarmid, R. W. 1971. Comparative morphology and evolution of frogs of the neotropical genera Atelopus, Dmdrophrynisciis, Melanophiynisais, and Oreophrynella. Bull. Los Angeles Mus. Nat. Hist. Sci. 12: 1-66.

Mebs, D. 1980. Zur fortpflanzung von m/cÿer (Amphibia: Salientia: Bufonidae. Salamandra 16:65-81.

Miller, T. 1987. Notes on Central hm tncA n Atelopus. The Herpetoculturist 1:25-28.

Myers, C. W. 1969. The ecological geography of cloud forest in Panama. Amer. Mus. Nov. 2396:1-52.

Myers, C. W., J. W. Daly, and B. Malkin. 1978. A dangerously toxic new frog (Phyllobates) used by Emberâ indians of western Colombia, with discussion of blowgun fabrication and dart poisoning. Bull. Am. Mus. Nat. Hist. 161(2):309-365.

Orton, G.L. 1953. The systematics of vertebrate larvae. Syst. Zool. 2:63-75.

Savage, J. M. 1972. The harlequin frogs, genus Atelopus, of Costa Rica and western Panama. Herpetologica 28:77-94.

Starrett, P. 1967. Observations on the life history of frogs of the family Atelopodidae. Herpetologica 23:195-204.

36 CHAPTERS

FIELD STUDIES ON VISUAL AND ACOUSTIC SIGNALING IN THE "EARLESS" PANAMANIAN GOLDEN FROG, ATELOPUS ZETEKI

Abstract.—Field studies were done on the visual and acoustic behavior of the Panamanian golden frog, Atelopus zeteki, a species lacking a t\Tnpanic middle ear. Males displaved three stereotyped behaviors in response to playbacks of male pulsed vocalizations: foot signaling, vocalization, and repositioning orientation. Frequencies of foot signaling and orientation responses were significantlv increased bv presentation of plavback vocalizations. Vocal responses also increased (non-significantlv) during the playback period and continued to increase during post-playback silence. These results provide the first experimental evidence that an "earless" anuran species displays behavioral responses to sound, and that vocalization may play an important role in communication. The fore-foot waving, and possibly hind- foot raising, observed in this species appear to represent visual signals used intraspecifically in male agonistic behavior. Male frogs appear to relv

37 preferentially on visual signaling (foot signaling) as compared to acoustic

signaling (vocalizing). The existence and preferential use of visual signaling in

this species seems to be correlated with their noisy, montane stream habitat.

Orientation by the frogs toward the playback speaker suggests that this

"earless" species of anuran is capable of localizing a sound source.

INTRODUCTION

Vocalizations play an important role in the mating and territorial

behavior of most anuran amphibians (Wells, 1977; Gerhardt, 1982; Ryan,

1985). Although acoustic signals are used as the dominant mode of communication, several vocalizing genera lack conventional middle ear structures such as the external tympanum, middle ear cavity, and auditory ossicle. This condition is often referred to as "earless," although standard inner ears, containing well-developed auditory end organs, are present (Wever, 1985;

Jaslow et al., 1988). Most species within the genus Atelopus lack a tympanic middle ear yet produce a variety of vocalizations (McDiarmid, 1971; Cocroft et al., 1990), and, therefore, provide an excellent opportunity to study acoustic behavior in an earless anuran. The Panamanian golden frog, Atelopus zeteki, is an endangered species, endemic to the humid foothills of west-central Panama.

38 In spite of low population numbers, I had the opportunity to make several

observations and perform experiments to test their acoustic behavior.

Although male Atelopus produce vocalizations, the functional significance

of their calls is not fully understood. Across the genus Atelopus, three general

call patterns have been described: the pulsed call, used primarily in

advertisement and confrontation; the pure tonal whistle, likely used as a

surrender signal by males after losing agonistic wrestling bouts; and the short

call, produced when disturbed, in close containment, or while wrestling

another male; (Jaslow, 1979; Lescure, 1981a; Crump, 1988; Cocroft et al.,

1990; Lindquist, 1995). Although vocalizations have been documented

primarily in cases of male-male aggression and disturbance (Jaslow, 1979;

Crump, 1988; Cocroft et al., 1990), males of two South American species, A. pulcher (= A. spumarius) and A. Jlavescens, appear to use pulsed calls in mate attraction (Jaslow, 1979; and pers. obs). In A. zeteki, all three call types have been documented (Cocroft et al., 1990; Lindquist, 1995).

Social communication in Atelopus may not be restricted to vocalization.

Apparent visual communication in the form of foot signaling (semaphoring) has been observed in a few species oi Atelopus (Crump, 1988; Ibanez et al.,

1995; R. Saunders, pers. com.). Cmmp (1988) described instances where foot displays were accompanied with vocalizations in A. varius and also documented

39 an association of semaphoring with intraspecific male-male aggression in this species. Atelopus chiricjuiensis males have been observed to semaphore in response to speaker playbacks of male vocalizations in the field, and a captive male used semaphores while attempting to amplex a female (pers. obs.).

The use of foot signals for visual communication in anurans is rare, but nonlocomotory limb movements have been documented in six other genera:

Staurois parvus and 5. latopalmatiis (Ranidae; Harding, 1982; Davison, 1984),

Hylodes aspenis (Leptodactylidae; Heyer et al. 1990), Taudactyliis eungelleiisis

(Myobatrachidae; Winter &. McDonald, 1986), Dendrobates parvulus

(Dendrobatidae: Wevers, 1988), Bracfiycephalus ephippiwn (Brachycephalidae;

Pombal et al., 1994), and Litoria fallax, L. genimaadata, L. nannotis, and L. rheocola (Hylidae; Richards and James, 1992). All these species are diurnal except those of the genus Litoria, which perform visual signals in moonlight

(Richards &. James, 1992). Hind foot semaphoric displays have been observed during courtship and male-male aggression in Staurois (Harding, 1982;

Davison, 1984), and during male aggression in Litoria and Bracliycephalus

(Richards &. James, 1992; Pombal et al., 1994), but the context of visual display behavior remains unclear in other species in which it has been observed.

40 The aim of this study was threefold; ( I ) to examine the effects of

intraspecific vocalizations on the behavior of an "earless" anuran species, A.

zeteki; (2) to determine if this species is capable of localizing a sound source;

(3) to investigate any interaction of visual and acoustic communication in this

species.

MATERIALS AND METHODS

Study Sites .—Field observations were carried out on populations of A.

zeteki from Parque General de Division Omar Torrijos-El Cope, Provincia de

Codé, and Parque Nadonal Los Altos de Campana, Provincia de Panama

during July 1994.

Experimental Design.—A series of call playback experiments was performed

to test behavioral responses of resident males along their rocky stream habitat.

After male frogs were located, a full range Audix PH-3 speaker was placed between 1 m and 2.5 m from an individual. Speaker position and distance varied due to cascade topography and location of each animal. Observers sat at least 3 m from the frog and speaker and remained quiet and motionless throughout the experimentation period. Tests were conducted in the following

41 manner: 10 min pre-playback period allowed the test subjects to acclimate to

the presence of a speaker and the observer(s), this was followed bv a 10 min

playback period during which a continuous series of pulsed vocalizations on an

endless cassette loop were played through a Sony TCD-D5 cassette recorder,

following the 10 min playback period, there was a 5 min post-plavback period

of silence.

Pulsed call recordings (obtained from A. jaslow) lasted about 400 msec

and were played even^ 3.4 sec. Only one good quality recording with minimal

background noise was available, therefore, onlv one exemplar of the pulsed call could be created for the plavback tape. I recognize that the standard protocol for playback experiments dictates that several exemplars should be used to avoid pseudo-replication error (Kroodsma, 1989 &. 1990). However, considering the fundamental questions addressed by this study (of hearing capabilities), pseudoreplication error should have minimal impact on the interpretation of the experimental results. Sound amplitude was set at 86 dB

SPL at 1 m from the speaker (linearly calibrated with a Quest 215 sound level meter: flat; peak). This sound pressure is comparable to that produced bv a male calling at 1 m (pers. obs.).

All behavioral responses during each period of experimentation were manually recorded, and the responses of some of the tested animals were

42 videotaped. A total of 15 male frogs were tested, eight at the El Cope site and seven at the Cerro Campana site. Sound recordings made during experiments were deposited in the archive of the Borror Laboratory of Bioacoustics at The

Ohio State Universitv.

RESULTS

Types of Behavioral Responses.—Atelopus zeteki males displayed three conspicuous behaviors during the playback experiments: semaphoring, vocalization, and change in body orientation. The semaphoric fore-foot displays were observed in 11 of 15 males and were similar to those reported by

Crump (1988) and Ibanez, et al. (1995). Semaphoric displays typically consisted of one to three rotational movements while the body was braced by the opposing front leg or while the animal was walking (Fig. 3.1; Crump, 1988; pers. obs.). Periodically, the digits on each hand were drawn toward the palm sequentially from digits 4 to 1 (thumb) as the fore-limb was being set down after a wave. In addition to fore-foot waving, three males perfomied raised hind-limb displays, a behavior not previously reported in Atelopus. This foot- raising in A. zeteki is different from the hind limb extension ("flagging")

43 Fig. 3.1 Sequential illustrations of fore-foot semaphoring bv a male A. zeteki (see text for further details). Sequence duration may vary from 300 - 800 msec.

observed in other visually signaling genera (see Chapter Introduction) and was

a subtle, but apparently deliberate action. Due to the rarity and subtle nature

of foot raising behavior, it was decided not to score it as a type of visual signal

in the statistical and graphical analyses that follow. During foot-raising, the

leg was never extended, but always tucked in toward the body, lifted dorsally,

and reset back on the ground. Occasionally, this motion was repeated a few

times before the limb was set back on the ground, and in some cases the digits

undulated or twitched spasmodically.

During the experiments, seven males responded by producing vocalizations. The majority of these were of the pulsed variety, but in one trial, a series of pure tonal whistles with no frequency modulation was added

(Lindquist, 1995). Ten of the frogs also showed changes in body orientation

44 during the experiments. The frogs began turning, usually in place, in response

to speaker playbacks. Of the 10 test males that responded by turning, all but

one made their initial turn toward the playback speaker. The males which

accurately turned were initially positioned in the same horizontal plane as that

of the speaker, and by turning could directly face the speaker. The one animal

that did not turn toward the speaker was clinging to a rock surface perpendicular to the plane of the playback speaker, and therefore could not make an accurate turn toward it. After an initial turn, 8 of 10 frogs began to walk, then started another series of turns. Following the initial turn, subsequent body movements and turns typically varied in orientation to the speaker.

Correlations of Behavior with Playback Vocalizations.—Males showed clear changes in behavior in response to playback vocalizations. The behavioral responses of all 15 males per minute of each experimental period are shown in

Figs. 3.2 -3.4. To further analyze the relationships between observed behaviors and playback vocalizations, each entire experimental trial was divided into five segments. The pre-playback period of silence and the playback period each were divided into two, five minute periods (SI & 52 and

PI &. P2, respectively), followed by a five minute post-playback period of

45 Figure 3.2 Number of semaphores of all 15 frogs per minute of experimentation. The pre-playback period of silence is represented by minutes I-10, the playback period (boxed region) is represented by minutes 11-20, and the post-playback period of silence is represented by minutes 21-25. Each circle represents one semaphore. Frogs numbered 1-8 represent animals from El Cope and those numbered 9-15 represent animals from Cerro Campana.

46 Figure 3.2 Semaphoring Responses of Individual Frogs

Individual Minmc 8 9 10 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 11 .... 12 . .. V 13 ... L 14 .. A 15 . ## . Y 16 . H 17 . A 18 .. c: 19 . K 20 . ## 21 . 22 . 23 24 25 Figure 3.3 Number of vocalizations of all 15 frogs per minute of experimentation. The pre-playback period of silence is represented by minutes I-10, the playback period (boxed region) is represented by minutes 11-20, and the post-playback period of silence is represented by minutes 21-25. Each circle represents one call. Frogs num bered 1-8 represent animals from El Cope and those numbered 9-15 represent animals from Cerro Campana.

48 Figure 3.3 Vocalization Responses of Individual Frogs

Individual Minute 8 9 10 11 12 13 14 15 1 2 3 4

() 7 8 9 10

12 V 13 L 14 A 15 Y 16 H 17 A 18 c: 19 K 20 21 22 23 24 25 Figure 3.4 Number of turns of all 15 frogs per minute of experimentation. The pre-playback period of silence is represented by minutes 1-10, the playback period (boxed region) is represented by minutes 11-20, and the post-playback period of silence is represented by minutes 21-25. Each circle represents one turn. Frogs numbered 1 -8 represent animals from El Cope and those numbered 9-15 represent animals from Cerro Campana.

50 Figure 3.4 Orientation Response of Individual Frogs

Individual Minute 8 9 10 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 11 . m m m 12 m m P 13 m , L Ln 14 . A 15 .. . m Y If) mm mm K 17 • • m m A 18 . m • • c; 19 mm m m K 20 m m 21 22 23 24 25 silence (PP). Certain behavioral trends become evident when the data are presented in this manner.

Foot signaling responses, across the 25 minutes of observation, typically were either initiated or increased in frequency during the presence of call playbacks. After the playbacks were terminated, the frequency of responses dropped. This clustering of responses within the playback portion of experimentation is apparent in Fig. 3.2. The distribution of total foot wave responses peaks in the latter 5 minutes of playback (P2), with a moderate drop during post-playback (PP, Fig. 3.5: solid circle). The distribution of foot signaling individuals follows a similar pattern (Fig. 3.5: open circle). If data for only responsive frogs (those that displayed semaphoring) are considered, the response trend becomes more apparent. The means and standard deviations of foot signaling responses per responsive individual for each of the five periods are indicated in Table 3.1 A. The largest mean foot signaling rate occurred during P2. The number of foot signals increased significantly during the 10 minutes of playback (P1&P2) compared to the ten minutes of silence (S1&S2)

(Table 3.1 A; paired t-test: P<0.0120, df=14).

Vocal responses increased across the 25 minutes of observation, particularly during the final ten minutes of experimentation (Fig. 3.3). The number of total vocal responses increased dramatically during the latter part of

52 Figure 3.5 Total number of semaphores and semaphoring individuals (out of the 15 frogs tested) per experimental period. The pre-playback period of silence and the playback period is divided into two, five min periods each (SI & S2 and P1 &. P2, respectively), followed by a five min post-playback period of silence (PP). The solid line represents the total number of semaphores and the dashed line represents the number of individuals semaphoring.

53 Figure 3.5

Semaphores 120 12

100 10 (U CD I 80 8 :g C > 0 • T3 1 60 6 £ I , I et -M H 20

0 0 SI S2 PI P2 PP Experimental Period

54 Figure 3.6 Total number of vocalizations and vocalizing individuals (out of the 15 frogs tested) per experimental period. The pre-playback period of silence and the playback period is divided into two, five min periods each (SI & S2 and PI &. P2, respectively), followed by a five min post-playback period of silence (PP). The solid line represents the total number of vocalizations and the dashed line represents the number of individuals vocalizing.

55 Figure 3.6

Vocalizations 120 12

100 • 10

C/3 I t g u 80 8 -o o > -gË 60 6 c C

É^ 40 . 4 ^ £ H Z 20 2

0 • 0 SI S2 PI P2 PP Experimental Period

56 Figure 3.7 Total number of turns and turning individuals (out of the 15 frogs tested) per experimental period. The pre-playback period of silence and the playback period is divided into two, five min periods each (SI & 52 and PI P2, respectively), followed by a five min post-playback period of silence (PP). The solid line represents the total number of turns and the dashed line represents the number of individuals turning.

57 Total Number of Turns ro (jo ON o o o o o T1

R ^ • w Lj Is. O 3 n>

3 I— go Ul a 00 o n> 3 • g o CL ""0

to 4L ON 00 ^ ^ O to Number of Individuals Period ktd. Respttnses X Sid. Dev, S.12. Mean iüdlv^MM. Paired t-tesi (M„: S=P) A. Semaphoring S1 3 15 1.364 2.618 0.789 8 df=14 (n„„=15, i\= l 1) S2 2 H 0.727 1.618 0.488 4 T=2.880 PI 9 35 3.180 3.600 1.090 10 P=0.0120 P2 10 70 6.360 5.370 1.620 16 X diff. = 5.400 PP 7 21 1.909 2.300 0.694 6 std. dev. diff. = 7.270 Total 1 1 149 13.550 1 1.180 3.370 32 sill. err. x diff.= 1.88

B. Vocalizing SI 1 1 1 1.571 4.158 1.570 11 df= 14 (n„„=15, n, = 7) S2 1 7 1.000 2.646 1.000 7 1 = 1.400 PI 1 2 0.286 0.756 0.286 2 P=0.1800

P2 5 45 6.429 4.791 1.810 13 X diff.= 1.933 PP 6 101 14.429 9.744 3.680 28 std. dev. diff. = 5.351 Ttnal 7 166 23.710 1 1.370 4.300 38 std. err. x diff.= 1.38 Ul vO C. Orienting SI 1 1 0.100 0.316 0.100 1 df= 14 (n„„=15, n„=10) S2 0 0 0.000 0.000 0.000 0 T=3.270 PI 9 35 3.500 3.408 1.080 6 P=0.0056

P2 9 31 3.100 2.283 0.722 8 X diff.=4.330 PP 4 6 0.600 0.843 0.267 2 std. dev. diff.=5.140 Total 10 73 7.300 5.120 1.620 15 std. err. % diff.= 1.33

fable 3.1. Descriptive statistics for each five iitinuie period of experiniematioM (S1, S2, P I, P2, and PP) for behaviorally responsive male frogs (n= 1 1 ; n=7; n= 10); and paired t-test values comparing respoitse meaits between the 10 minutes of silence (S = SI +S2) and the 10 minutes of playback (P=PI +P2) for all frogs tested (Minitab rel. 10 for Windows). Total number of tested individuals; N= 15. A. Semaphoring Response; B. Vocal Responses; C. Orienting Responses. playback (P2) and peaked during post-playback (PP, Fig. 3.6: solid circle).

The number of responsive individuals (those that vocalized) followed a similar

pattern (Fig. 3.6: open circle). Again, the pattern is apparent only when data

for responsive individuals are considered (Table 3.1 B). There was a non­

significant trend of increase in the number of vocalizations during the ten

minutes of playback (P1&P2) compared to the ten minutes of silence (SI&S2,

Table 3.1 B; paired t-test: P= 0.1800, df= 14). This lack of significance may be

due to the different response distributions and to the lower number of

responsive individuals.

Orienting responses, across the 25 minutes of observ ation, were initiated during the call playbacks bv all but one individual. That individual turned

toward the observer in the pre-playback. After the playbacks ceased, the number and frequency' of responses greatly diminished. This clustering of turning responses during the playback segment is apparent in Fig. 3.4. The number of total orienting responses peaked in the initial 5 minutes of playback

(PI ), nearly plateaued in the latter (P2), and steeply dropped during post­ playback (PP, Fig. 3.7: solid circle). Again, the distribution of the number of individuals has a similar shape (Figure 3.7: open circle). If data for only responsive frogs (those that displayed orientational movements) are considered, the response trend is more apparent (Table 3.1 C).

60 The number of turns increased significantly during the ten minutes of playback (PI&P2) compared to the ten minutes of silence (S1&S2, Table 3.1

C; paired t-test: P=0.0056, df=14).

DISCUSSION

This study provides clear evidence that A. zeteki, a species lacking a tvmpanic middle ear, responds to sound. This study complements physiological studies on some earless frogs that have demonstrated significant sensitivity to sound at certain frequencies (Jaslovv & Lombard, 1996; Wever 1985), and clearlv establishes the importance of vocalization in this earless species. Foot waving and change in body orientation significantly increased when male vocalizations were played to male frogs, and both subsequently decreased after playbacks ceased. Vocalizations of experimental males also showed a moderate trend of increase during playbacks and continued to increase after playbacks ended.

These results indicate that vocalizations do serve a function in an "earless" anuran species. The results of this study support physiological evidence that

''taxless" Atelopus are sensitive to sound (see Chapter 5).

61 Pulsed Vocalization.—The pulsed call, when played to males of A. zeteki, evoked several agonistic behaviors and supports the suggestion of Crump

( 1988) that the pulsed call advertizes male position and is used during aggressive encounters. M. Crump (1986) described many territorial behaviors in A. varias, such as homing, site fidelity, and active defense. This study on A. zeteki suggests that males of this species, as in A. varias, also are territorial and aggressive when another male is perceived to be nearbv.

Visual Signaling. —Visual signaling in the form of foot waving and raising likely is used in intraspecific, male-male agonistic displays, although visual signaling by females or between males and females cannot be ruled out. Foot waving in A. zeteki was at times elicited solely by the presence of a human observer (pers. obs.). This observation may suggest that these visual signals function as interspecific displays toward a potential predator. These frogs are brilliantly colored and highly toxic as adults (Brown et al., 1977) and foot waving and raising may draw attention to their aposematic coloration. Of the numerous cryptic juveniles observed, none were seen semaphoring. These limb movements conceivably could be triggered by any agitating situation.

Nonetheless, conspecific acoustic signals trigger this response pattern quite markedly, and this finding strongly suggests a role in intraspecific

62 communication. Foot waving behavior toward a human observer may simply

represent extreme interspecific territorial behavior or aposematic signaling. For

example, during a preliminary observation, one male appeared to be disturbed

by the close proximity of the observer. The frog repeatedly foot waved and

vocalized as it moved from 2.5 m to within 5-10 cm of the observer. This was

the most dramatic case of semaphoring directed toward a human observer, but

other males under similar circumstances also showed some degree of agitation

by a display of foot waving. M. Crump has seen identical behavior directed toward her while observing A. variiis (pers. com.).

This study also described hind-foot raising behavior in some male A. zeteki. Because it was a rare and subtle behavior, it was not treated as a visual signal in this study, and research presented in Chapter 4 establishes that it does not function in visual communication.

Orienting Behavior .—All nine males that showed turning behavior, and were capable of orienting toward the speaker, directed their initial turn toward the playback speaker. In two trials, the frogs walked and jumped toward the speaker, at times moving over rocks and along branches. These are the first experimental data to suggest that an "earless" anuran is capable of directional hearing. These data are especially interesting because hypotheses explaining

63 directional hearing in anurans have focused on the role of a tympanic middle

ear as an inherently directional pressure gradient receiver (Eggermont, 1988).

Considering the small size of A. zeteki, it is difficult to envision how a non-

tympanic system of hearing may provide spatial cues in this species.

Nonetheless, nontympanic directional hearing appears to be possible and the

question merits further experimental inquiry.

Interaction of Visual and Acoustic Signaling .—Male A. zeteki showed some

consistent patterns in acoustic and visual signaling behavior. Most males

displayed a tvq)ical series of responses after initiation of call playbacks. At the onset, the individual sat in place for a few seconds to a few minutes. The frog

then repositioned himself toward the speaker and/or produced a series of foot waves or raises. Frequency of foot waving and raising increased up to the point of post-playback, after which the frequency decreased considerably (Figs. 3.2

&. 3.5). Except for one orientation to human presence by one individual in a pre-playback period, all turning responses exhibited by males occurred during the playback period, as if these males were actively searching for a perceived male (Figs. 3.4 3.7). Approximately five minutes into the playback period, many individuals started to call in response. In some cases, calls were given in between calls of the playback, or accompanied by foot waving or raising, or

64 walking. As playback progressed into post-playback, frogs continued to

increase the frequency of calling (Figs. 3.3 &. 3.6). These observations suggest

that visual signaling (foot waving) and visual confirmation (orientation) are

used as first responses to calls, perhaps because of the greater signal reliability

of visual cues compared to auditory cues in habitats with high background

noise. However, because visual communication is limited by the visual

capability of the receiver and the position of the transmitter, the mode breaks down in cases of obstruction. Acoustic communication, although perhaps less

reliable in noisy habitats, is more reliable in transmitting information around obstructions.

1 suggest that the brightlv colored A. zeteki prefers visual over acoustic communication. In situations where a male perceives a rival acousticallv, but there is no visual contact, acoustic signaling is initiated. This point will be directly tested in the future by conducting experiments that employ visual playback images with and without sound. A residual effect results in an increased frequency of calling after the cessation of playbacks. This latter phenomenon may be related to the animal's agitated state and/or the reliabilltv of acoustic signals in communicating information to an individual out of view.

65 Correlation of Visual Siptaling and Earlessness.—Several families of anurans lack a tympanic middle ear, and while many "earless" species inhabit torrential stream and noisy cascade environments, there is no clear, single factor correlated with this morphological condition (Jaslow et al„ 1988). A compelling argument can be made that environments such as these facilitate the evolutionary reduction or loss of the tympanic middle ear and the use of visual signals for intraspecific communication. Observations on Atelopus sp. nov. from Parque Nacional Chingaza, Colombia, support the hypothesis that visual signaling is correlated with noisy habitats. The males of this species vocalize and amplex females far from water in areas of relatively low background noise, and show no evidence of foot signaling behavior (C. Navas, pers. com.).

Table 3.2 summarizes some traits associated with foot signaling anurans.

Although visual communication in anurans may be correlated with high noise environments, reduction and loss of a tympanic middle ear may not. For example, all visually signaling species other than Atelopus possess a complete tympanic middle ear. Even A. spumariiis, which possesses a nearly complete tympanic middle ear, has been reported to use foot waving displays between males (Saunders pers. com.). The common factor that all visually signaling anuran species (except Dendrobates panndus and Litoria falla.x) share is general

6 6 Family Species Habitat Microhabitat Activity Ear References Brachycephalidae Bracliyccpluilus epiiippium R S DP Pombal et al., 1994 Dendrobatidae Dmdrohatcs pmyulus R F DP Wevers, 1988; Duellman, Leptodactylidae Hylodcs aspmis R S,C DP Heyer et al., 1990 Myobatrachidae Taudactylus cungclkmis R S,C DP Winter &. McDonald, 1986 Ranidae Staurois panms R S,C . N ' . P Harding, 1982 Ranidae Staurois latopalinntus RS,C N’ P Davison, 1984 Hylidae Litoria gcnimaculata RS NP Richards & lames, 1992 Hylidae Litoria nannotis R s.c NP Richards & James, 1992 Hylidae Litoria rhcocola R s,c NP Richards & James, 1992 Hylidae Litoria fallax D P.Sw N' P Richards & James, 1992 Bufonidae Atelopus limosus R s DA Ibanez et al., 1995 O' Bufonidae A telopus spumarius R s D P Saunders, pers. com. Bufonidae Atelopus varius R,C s,c D A Crump, 1988 Bufonidae A telopus cli iritju iemis C s,c D A pers. obs. Bufonidae Atelopus zeteki R,C s,c D A pers. obs.

Table 3.2. Taxonomic distribution, morphological and ecological traits associated with foot signaling anurans. HAlirrA'l’: C=Cloud forest, R=Rainforest, D=Tropical Dry Forest; MicrohabitaT; C=Cascade, F= Forest, P=Pond, S=Stream, Sw=Swamp; ACTIVITY P attern: D=Diurnal, N = Nocturnal, * Foot signals during 1st hours of daylight, ' during moonlit nights; MIDDLE EAR: A=Absent, P=Present. habitat type (Duellman, 1978; Richards and James, 1992). Therefore, visual

signaling in anurans appears to be correlated with montane stream habitats with high water noise and not with the loss or reduction of the tympanic

middle ear as previously hypothesized by Heyer et al. (1990). Additional work on visual and acoustic communication in anurans is needed for a better understanding of their functional interrelationship and roles in the behavioral repertoires of bimodally signaling species, such as A. zeteki.

R£SUMEN

Se llevaron a cabo estudios de campo sobre el comportamiento visual y acustico de la rana dorada de Panama, Atelopus zeteki, una especie que carece del oido medio timpanico (“sorda”). Se observé, en los machos, très tipos de comportamiento estereotipado como respuesta a grabaciones de vocalizaciones de individuos del mismo sexo: movimientos de las patas, vocalizaciones y cambios de orientaciôn. La frequencia en el movimiento de las patas y en los cambios de orientaciôn aumentô en forma significativa cuando se usé las grabaciones. Las vocalizaciones también aumentaron (pero no significamente) durante y después de las grabaciones. Estos resultados constituyen la primera evidencia experimental de que las especies de anuros “sordos” son capaces de

68 responder a sonidos y de que las vocalizaciones pueden jugar un papel

importante en la comunicaciôn. El movimiento circular de las patas

delanteras, y posiblemente el levantamiento de las patas traseras parecen ser

senales visuales usadas intraespecificamente en las relaciones agonisticas. Los

machos parecen depender mas de senales visuales (movimientos de las patas) que de senales acûsticas (vocalizaciones). La existencia y el uso preferencial de senales visuales en esta especie pueden ser relacionados al nivel de ruido existente en su habitat de rfos y riachuelos montanosos. Los cambios de orientaciôn en direcciôn a las grabaciones empleadas sugieren que esta especie

“sorda” de anuros es capaz de localizar eficientemente una fuente de sonido.

Acknowledgments .—This field study was funded by a short-term graduate fellowship through the Smithsonian Tropical Research Institute. 1 am deeplv grateful to A.S. Rand, R. Ibanez, and N. Smith for their availability and assistance throughout the study. 1 am indebted to T. Hetherington, D.

Swihart, A. Castillo (A.C.E.P.E.), V. Morales, and M. Lindquist for their field assistance and companionship, and to M. Crump, R.W. McDiarmid, A. Jaslovv,

R. Saunders, T. Barker, and M. Bernal for providing information on behavior and localities of A. zeteki. I thank IN. RE. NA. RE, J. Cheng, and the national park staff at Cerro Campana and El Cope for permitting me to work on the

69 endangered golden frog populations, and are indebted to the staff at STRJ: G.

de Alba, M. Morello, G. Maggiori, I. Ivandc, and A. Armùelles. I thank A.

Jaslow for use of his audio recordings, and D. Stetson and D. Dennis for their

assistance with the illustrations. I am thankful to N. Arguedas and D. Slack

for their grammatical construction and review of the Spanish abstract. I thank

B. Smith for assistance in statistical analysis and the Borror Laboratory of

Bioacoustics for tape production and use of their recording and analysis equipment. Finally 1 am grateful for the suggestions made by A. S. Rand, M.

Crump, and many others who carefully reviewed a preliminary version of this manuscript.

BIBLIOGRAPHY

Brown, G. B., Y. H. Kim, H. Kiintzel, H. S. Mosher, G. J. Fuhmian, and F. A. Fuhmian. 1977. Chemistry and pharmacology of skin toxins from the frog Atelopus zeteki (Atelopidtoxin: zetekitoxin). Toxicon 15:115-128.

Cocroft, R. B., R. W. Mcdiarmid, A. P. Jaslow, and P. M. Ruiz-Carranza. 1990. Vocalizations of eight species o( Atelopus (Anura: Bufonidae) with comments on communication in the genus. Copeia 1990: 631-643.

Crump, M. L. 1986. Homing and site fidelity in a neotropical frog, Atelopus varius (Bufonidae). Anim. Behav. 34: 438-444.

Crump, M. L. 1988. Aggression in harlequin frogs: male-male competition and a possible conflict of interest between the sexes. Anim. Behav. 36: 1064- 1077.

70 Davison, G. 1984. Foot-flagging displays in Bornean frogs. Sarawak Mus. J. 33(57):177-178.

Duellman, W. 1978. The biology of an equatorial herpetofauna in Amazonian Ecuador. Misc. Publ. Mus. Nat. Hist. Kans. 65: 1-352.

Eggermont, J. J. 1988. Mechanisms of sound localization in anurans. In: B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak (eds.). The Evolution of the Amphibian Auditory System, pp. 307-336. New York: John Wiley &. Sons.

Gerhardt, H. C. 1982. Sound pattern recognition in some North Anierican tree frogs (Anura: Hvlidae): implications for mate choice. Am. Zool. 22: 581- 595.

Harding, K. A. 1982. Courtship displav in a Bornean frog. Proc. Biol. Soc. of Wash., 95: 621-624.

Heyer, W. R., A. S. Rand, C. A. G. Da Cruz, O. L. Peixoto, and C. E. Nelson 1990. Frogs of Boracea. Arq. Zool. 31: 231-410.

Ibanez, R. D., C. A. Jamarillo, y F. A. Solis. 1995. Una especies nueva de Atelopus (Amphibia: Bufonidae) de Panama. Carrib. J. Sci. 31: 57-64.

Jaslow, A. P. 1979. Vocalization and aggression in Atelopus chiriquiensis (Amphibia, Anura, Bufonidae). J. Herpetol. 13: 141-145.

Jaslow, A. P., and R. E. Lombard. 1996. Hearing in the Neotropical Frog, Atelopus chiriquiensis. Copeia 1996:428-432.

Jaslow, A. P., T.E. Hetherington, and R.E. Lombard. 1988. Structure and function of the amphibian middle ear. In: B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak (eds.). The Evolution of the Amphibian Auditory System, pp. 69-92. New York: John Wiley & Sons.

Kroodsma D.E. 1989. Suggested experimental designs for song playbacks. Anim. Behav. 37:600-690.

71 Kroodsma D.E. 1990. Using appropriate experimental designs for intended hypotheses in song" playbacks, with examples for testing effects of song repertoire sizes. Anim. Behav. 40: 1138-1150.

Lescure, J. 1973. Contribtion al'etude des Amphibiens de Guyane française. I. Notes sur Atelopus Jlavesceiis Dumeril et Bibron et description d'une nouvelle espece. Vie Milieu 23:125-141.

Lindquist, E D. \995. Atelopus zeteki-. Pure tonal vocalization. Herpetol. Rev. 26(4): 200-201

McDiamtid, R.W. 1971. Comparative morphology and evolution of frogs of the genus Atelopus, Dendrophiyniscus, Melnnopluyuisais, and Oreophiynella. Bull. Los Ang. Co. Mus. Nat. Hist. 12:1-66.

Pombal, J.P. Jr., I. Sazima, and C.F.B. Haddad. 1994. Breeding behavior of the Pumpkin Toadlet, Brachycephalus epiiippium (Brachycephalidae). J. Herpetol. 28:516-519.

Richards, S.J. and C. James. 1992. Foot-flagging displays of some Australian frogs. Mem. Qd. Mus. 32:302.

Ryan, M.J. 1985. The Tungara Frog: a Study of Sexual Selection and Communication. Chicago: University of Chicago Press, pp. ix-230.

Wells, K.D. 1977. The courtship of frogs, In D.H. Taylor and S.I. Guttman (eds.). The Reproductive Biology of Amphibians. New York: Plenum Press, pp. 233-262.

Wever, E.G. 1985. The Amphibian Ear. Princeton: Princeton Univ. Press.

Wevers S. E. 1988. Enige opmerkingen over de pijlgifkikker Dendrobates parvulus. Lacerta 46:51-53.

Winter, J. and K.R. McDonald. 1986. Eungella: The land of the cloud. Aust. Nat. Hist. 22:39-43.

72 CHAPTER 4

SEMAPHORING IN AN EARLESS FROG: THE ORIGIN OF A NOVEL VISUAL SIGNAL

Abstract.—Social communication in anuran amphibians (frogs and toads) is mediated predominantly by acoustic signals. Unlike most anurans, the

Panamanian golden frog, Atelopus zeteki, lacks a standard tympanic middle ear and appears to have augmented its communicative repertoire to include rotational limb motions as visual signals, referred to here as semaphores. The communicative nature of semaphoring was suggested by experimental manipulations using mirrored self-image presentations and nonresident introductions. Male frogs semaphored significantly more when presented with a mirrored self-image than with a nonreflective control. Novel encoimters between resident males and nonresident frogs demonstrated that semaphores were used directionally and were displayed toward target individuals. Females semaphored frequently and this observation represents a rare case of signaling by females in a typically male-biased communicative regime. Semaphore

73 actions were dearly linked to a locomotory gait pattern and appear to have

originated as an elaboration of a standard stepping motion.

INTRODUCTION

Animals employ many sensory channels (acoustic, visual, chemical, etc.) for

communication (W. Smith, 1977). Visual signaling, in the form of discrete

limb motions, is widely used among vertebrate and invertebrate taxa and

indudes a diversity of movements from daw- and leg-waving in fiddler crabs of

the genus Uca (Zucker, 1983; Hyatt, 1977) to wing-flick displays in parid birds

(S. Smith, 1996; Clemmons and Lambrechts, 1992). Although sodal

communication in anurans appears to rely mostly on acoustic signals

(vocalizations), a few genera apparently have incorporated visual signals in the

form of conspicuous limb motions. The neotropical germs Atelopus

(Bufonidae), commonly known as harlequin or down frogs, is one such group

in which the use of visual signals appears to be a significant mode of

communication for agonistic displays between males (Lindquist and

Hetherington, 1996; Ibanez et al., 1995; Crump, 1988). These visual signals in

Atelopus involve semaphoric movements of the forelimbs (Fig. 4.1). Crump

(1988) described instances in which foot displays were accompanied by

74 vocalizations in A. varius and also documented an association of semaphoring with intraspecific male-male aggression in this species. Atelopus chiriquiensis males also have been observed to semaphore in response to speaker playbacks of male vocalizations in the field (pers. obs.).

The use of visual semaphores for communication in anurans is rare and has been documented in six genera other than Atelopus: Staurois parvus and S. latopalmatus (Ranidae: Harding, 1982; Davison, 1984), Hylodes asperus

(Leptodactylidae: Heyer et al. 1990), Taudactylus eungellensis (Myobatrachidae:

Winter and McDonald, 1986), Dendrobates parvulus (Dendrobatidae: Wevers,

Figure 4.1. Visually signaling golden frog showing the conspicuous forelimb rotation, referred to as semaphoring. Illustration by David Dennis.

75 1988), Brachycephalus ep/iippiMm(Brachycephalidae; Pombal et al., 1994), and

Litoria fallax, L. genimaculata, L nannotis, and L. rheocola (Hylidae: Richards and

James, 1992). All of these species are diurnal except those of the genus Litoria, which visually signal in moonlight (Richards and James, 1992). Hind foot semaphoric displays have been observed during courtship and male-male aggression in Staurois (Harding, 1982; Davison, 1984), and during male aggression in Litoria and Brachycephalus (Richards and James, 1992; Pombal et al., 1994). The function of visual displays remains unclear in the other species in which it has been observed, and empirical studies testing the communicative nature of these displays are lacking.

The apparent use of visual signaling in the genus Atelopus is especially intriguing because, although these frogs readily vocalize, most species lack typical middle ear structures (a tympanum, an air-filled middle ear cavity, and auditory ossicle). This anatomical condition is often referred to as earlessness, although standard inner ears with well-developed auditory sensory organs, are present (Wever, 1985; Jaslowet al., 1988). Neurophysiological studies have established that earless species oï Atelopus are indeed less sensitive to sound than fully eared species (Jaslow and Lombard, 1996; Wever, 1985; pers. obs.) but retain significant sensitivity despite the lack of a standard tympanic middle ear. Field studies of the earless A. zeteki also have demonstrated that

76 conspedfic vocalizations affect male behavior and seem important in agonistic encounters (Lindquist and Hetherington, 1996). Earless frogs may use nontympanic auditory pathways, such as the lateral body wall and lungs, as transmission routes for sound energy to the irmer ear (Ehret et al., 1990;

Hetherington 1992; pers. obs.).

Almost all of the spedes that use visual signals inhabit montane, cascade stream environments with high ambient noise levels (Lindquist and

Hetherington, 1996). These environmental factors have been suggested to explain both the origin of semaphoring and earlessness in frogs (Lindquist and

Hetherington, 1996), and Heyer et al. (1990) hypothesized that visual communication should be more reliable than acoustic communication in such environments because of the masking effect of ambient noise. McDiarmid

(1971) concluded that the loss of the tympanic middle ear was a derived condition in Atelopus and the characteristically high ambient noise of the habitat of most spedes of Atelopus m ay explain its loss.

Here I present data suggesting that the forelimb movements of male and femsàe Atelopus zeteki are true visual signals (semaphores). I provide evidence that the semaphoring action represents an elaboration of a normal stepping movement derived from a locomotory neural program.

77 METHODS

Study Subjects and Sites. —The Panamanian golden frog, Atelopus zeteki Dunn

1933, inhabits streams along the montane slopes of the Central Cordilleran cloud forests of west-central Panama. I studied A. zeteki populations from streams within Campana Heights and El Cope National Parks, at the eastern and western extent of its range, respectively. Resident Atelopus are often aggressive toward conspedfrcs entering their territory and often produce vocalizations and sometimes semaphores in such encounters (Crump, 1988; pers. obs). I used mirror presentations to test whether a self image would elicit semaphores, and introductions of non-resident frogs into the territory of a resident male to examine the use of semaphores in a potentially agonistic encounter. Field observations and behavioral experiments were performed between 0700 hr and 1700 hr, from April 2 to June 30, 1996.

Mirrored Self-image Presentations.—Mirrored self-image presentation experiments used the presence of a 10 cm x 10 cm mirror (self image) and a control (opaque side of mirror) placed 4-8 cm from a resident male frog. The mirror/control was frxed to the end of a 2 m aluminum pole so as to minimize observer interference. Trials lasted 15 min, beginning with a 5 min pre­

78 treatment observation period (as a control for the presence of a human

observer) followed by two, 5 min treatment periods (mirror and control).

Mirror presentation experiments were performed on 23 resident males. The

presentation order of experimental treatments was randomly alternated in each

trial to control for order effects (12 with control first presentation and 11 with

mirror first presentation). The observer sat motionless 2 m from the subject.

Behavioral responses were noted and many were videotaped. Tested frogs were

by necessity male as female frogs were less commonly encountered.

Conspedfic Nonresident Introductions .—Introduction experiments tested the

behavioral responses of frogs to encounters with novel conspedfrcs. A total of

33 conspedfic introduction experiments were performed, involving 33 males

and six females. Residents always were male (n=20). In eight cases the

intruders were female (two females were tested in gravid and non-gravid

condition), and in 25 cases the intruders were male (some males were used

both as an intruder and as a resident). Individual non-resident males and

females were placed approximately 50 cm away from a resident male, and behavioral responses were videotaped on a Nikon VN360 for later analysis.

Introduction trials ran as long as individuals interacted and were halted if

79 neither frog responded for ten min or ten min after the last response of either individual.

RESULTS

Mirrored Self-image Presentations.—Of the 23 animals tested, 15 did not respond, eight responded with semaphoring and none responded by vocalizing.

In the trials with responsive individuals, 101 semaphores were used: none were observed during the pretreatment period, a mean of 10.38 (±10.45 sd) occurred during mirror presentation, and 2.25 (±3.96 sd) during control presentation. These data demonstrate that significantly more semaphores were produced during the presentation of a mirrored self image compared to that of the control (P=0.0423, Wilcoxon Matched-Pairs Signed-Rank Test [NoruAis,

1993], two-tailed; n=8. Fig. 4.2). Whole-body movement of the mirrored self- image seemed to be important in evoking semaphores, as the majority of semaphoring frogs moved before responding (i.e on seeing their reflection move), and frogs that did not move did not semaphore had not moved.

Conspedfic Nonresident Introductions .—Semaphores were given by a total of 12 males and six females. Interactions between individuals were often complex,

80 Figure 4.2 Histogram illustrating the differential semaphore use between mirror and nonreflective control presentations. Each column represents one individual. Bars above and below the axis represent summed responses to mirrored self-image and nonreflective control, respectively.

81 Figure 4.2

Semaphoring Individuals 30 Mirror Presentation

Control Presentation Wilcoxon Matched-Pairs Signed-Rank, P= 0.0423

82 and occasionally escalated to chases and wrestling bouts. Here we emphasize analysis of precontact behaviors. There was a clear asymmetry in behavior of the introduced frogs and resident male frogs. Resident males primarily vocalized (x=52.27 ± 68.42, n=15) and occasionally semaphored (x=9.86 ±

7.88, n=7). In contrast, introduced frogs were silent but frequently semaphored (x=16.17 ± 23.10, n=12).

Directional Use of Semaphores .—Data from mirrored self-image presentations and introduction experiments were included in an analysis of the directional use of semaphores. Of 558 recorded cases involving individuals, the handedness of semaphoring is completely random within individuals and overall (right, 277; left, 281). We were able to discern the position of the target individual clearly in 344 of the 558 recorded semaphores. To ascertain the directionality of semaphore use, videotaped trials were analyzed by scoring the semaphores in either of two lateral 90 degree fields of view (Fig. 4.3).

In this analysis, if the target frog was in either lateral 90° field, a semaphore by the forelimb on the same side as the target animal was scored as an

“onside” semaphore, and a semaphore by the contralateral forelimb was scored as “offside.” This approach was used because target frogs in the lateral fields of view were visible only to either the left or the right eye of the semaphoring

83 animal. It is likely that there is a stereoscopic overlap in the forward field, as morphological observations suggest that the field of view of an eye extends approximately 26 degrees across the midline (pers. obs., n=5)].

Onside semaphores per individual frog (x=8.43 ± 11.77 s.d.) were used significantly more often than offside semaphores (x=3.00 ± 5.67 s.d.)

[P=0.0167, Wilcoxon Matched-Pairs Signed-Rank Test (NoruSis, 1993), two tailed; n= 14. Fig. 4.4].

Figure 4.3. Illustration of 90° lateral fields of view in A zeteki.

84 Figure 4.4 Histogram illustrating the directional use of semaphores in relation to target animal position. Each column represents one individual. Bars above and below the axis represent summed semaphores scored as “onside” and “offside,” respectively.

85 Figure 4.4

Semaphoring Individuals

40

u_ O k. (U XI £ 3 z Offside Semaphores Wilcoxon Matched-Pairs Signed-Rank, P = 0.0 I6 7

8 6 Sex Bias in Semaphore Use.—Semaphores were displayed more readily by female A. zeteki (six of six tested) than males (11 of 33 tested; P=0.0038,

Fisher’s Exact Test [NoniSis, 1993]). The majority of female semaphoring was observed in the context of courtship (i.e., the interaction resulted in amplexus).

In contrast, males used only vocalizations during courtship and never semaphored

Neural Basis of Semaphoring .—Following a semaphore, we often observed what appeared to be intention movements (twitching or motion) of the contralateral (opposite) hindlimb. Amphibian gaits typically follow a diagonal sequence in which a forelimb step is followed by a step of the contralateral hindlimb. The motions of the contralateral hindlimb following a semaphore suggested that the semaphore action is derived from a locomotory neural program. To test this hypothesis, we recorded the first limb movement subsequent to a forelimb semaphore. Of 558 semaphores recorded in 19 frogs,

222 were followed immediately by some type of limb motion. In 206 of the

222 cases, the motion involved a perceptible twitch or actual step of the contralateral hindlimb. The difference between motions of the contralateral hindlimb (x= 10.79 ± 17.08 sd) and those of other limbs (x=0.84 ± 1.30 sd) per individual frog is highly significant (P=0.0002, Wilcoxon Matched-Pairs

87 Signed-Rank Test [NoruSis, 1993], two tailed). This pattern suggests that a semaphore represents an elaboration of a standard stepping movement of the normal locomotory program. The motion of a forelimb semaphore similarly begins as does forelimb step, but is subsequently drawn upward instead of being set down as in locomotion. The observation of hindlimb motion subsequent to semaphoring probably explains the hind-foot raising behavior reported in some male A. zeteki in a previous study (Lindquist and

Hetherington, 1996).

DISCUSSION

Semaphores as Visual Signals.—The results of this study suggest that semaphoring represents visual communication in A. zeteki. The preferential elicitation of semaphores by a mirror self-image demonstrated that the response is elicited by specific cues most likely associated with another intraspecific individual. The directional use of semaphores suggest that the action is directed toward a specific target. Both findings are supportive of a signaling function of semaphore movements.

Demonstration of changes in the behavior of target frogs in response to semaphores was problematic in this study. Behavioral interactions during

88 introduction experiments were complicated and, in addition to semaphoring,

frogs also produced calls, chased each other, and wrestled. In most cases it was

not possible to accurately determine if the actions of any given animal were in

response to a semaphore or another behavior exhibited by a nearby conspecific.

However, some pertinent observations can be detailed from eight introduction

trials in which a semaphore was the first observable action. These semaphores were performed only by intruders, and were the only observable actions by

these frogs. In six of the eight cases in which intruders semaphored, the

resident frogs quickly responded (within 24 sec) by calling (four cases), chasing

(one case), and semaphoring (one case). In the other two cases, the resident likely could not see the semaphoring intruder, and therefore did not respond to the semaphores. Therefore, in all cases in which a resident male could see semaphores, the semaphores elicited responses. These data suggest that semaphores do elicit behavioral responses in targeted individuals, but additional experiments are needed to establish the functional effects of semaphores. Fore example, the result of a pilot study with a yellow flag moved to mimic a semaphore showed that some target frogs responded with semaphores, calls, and even attacks on the rotating flag. More precise evidence of the role semaphoring plays in communication may be provided by additional tests with artificial semaphores and/or semaphoring models.

89 Visual signaling by semaphoring clearly is used intraspedfically. There are intrasexual, male-male agonistic displays, and intersexual female-male displays.

Semaphoring inÆ zeteki was at times elicited solely by the presence of a human observer (pers. obs.). Movement of a brightly colored limb may function to draw attention to the aposematic coloration of these toxic animals.

Also, semaphoring toward a human observer may represent extreme interspecific territorial behavior. For example some males appeared to be disturbed by the close proximity of the observer. Frogs on a few occasions approached the observer while repeatedly semaphoring and vocalizing. In two cases males actually climbed up the observer’s leg while displaying. These were the most dramatic cases of semaphoring directed toward a human observer.

M. Crump has seen similar behavior directed toward her while observing A varius (pers. com.).

Sex Bias in Semaphore Use.—Communication in anuran amphibians is generally sex-biased in that males of almost all species act as the sole sender

(Roy et al.,1995; Wells, 1977; Wells, 1988). However, this study foimd that female A zeteki displayed more semaphores than males. Female semaphoring occurred during courtship, whereas males only vocalized and never semaphored in encounters with gravid females. This differential use of

90 semaphores by females and males is intriguing and suggests that semaphoring

may be used in multiple sodal contexts. This differential semaphoring may

reflect an origin other than that held by the ambient noise hypothesis.

Sexual Selection and Semaphoring .—While the context of visual signaling in A. zeteki is only partly understood, a number of hypotheses may be proposed with

regards to sexual selection and signaling modality. From the standpoint of

intrasexual selection, semaphoring between males was commonly performed by

intruders. These semaphores may represent an appeasement display given to a resident male. However, resident males also semaphored toward intruding males, and chases by resident males often included semaphoring, both of which would appear to conflict with an appeasement function. It seems more likely that semaphores are used as a measure of territorial vigilance by either the resident or challenging intmder. Numerous semaphores given by either male may signal a high degree of aggressiveness and may serve to preempt exhaustive bouts of wrestling. Why intmding males are less likely to vocalize remains unclear. We did not observe female - female interactions in this study, but Crump (1988) noted that A. varius females readily semaphored to each other when one individual was introduced into the territory of another.

Crump’s observations found that females did not wrestle as did males. This

91 may simply reflect a tendency for males to generally be more aggressive and

defend sites used for egglaying, in addition to defending nutritive resources.

With regards to intersexual selection, gravid females may display

semaphores to evaluate vigilance in males before amplexus. In addition, the

conspicuous semaphoring behavior also may draw the attention of males. It

may then be predicted that gravid females that semaphore (or semaphore more

frequently/conspicuously) should be more effective in courtship and amplexus.

Semaphoring between nongravid females and males simply may be comparable

to male - male semaphoring in that both may gauge territorial vigilance.

Acknowledgements .—I thank T. Hetherington, A. S. Rand, D. Nelson, T.

Grubb, J. Downhower, S. Gaunt, and A. Snow for comments, review, and discussion and A. Castillo for field assistance. This work was financially supported by a Smithsonian Tropical Research Institute Short-term Graduate

Fellowship, a Grant-In-Aid of Research from Sigma Xi, The Scientific Research

Society, and a Graduate School Alumni Research Award from The Ohio State

University.

92 BIBUOGRAPHY

Clemmons, J. R. and M. M. Lambrechts. 1992. The waving display and other nest site anti-predator behavior of the Black-capped Chickadee. Wilson Bull. 104, 749-756.

Crump, M. 1988. Aggression in harlequin frogs: male-male competition and a possible conflict of interest between the sexes. Anim. Behav. 36,1064-1077.

Davison, G. 1984. Foot-flagging displays in Bornean frogs. Sarawak Mus. J. 33, 177-178.

Ehret, G., J. Tautz, and B. Schmitz 1990. Hearing through the lungs: lung- eardrum transmission of sound in the frog Eleutherodactylus coqui. Naturwissenschaften 77, 192-194.

Harding, K. A. 1982. Courtship display in a Bornean frog. Proc. Biol. Soc. of Wash., 95, 621-624.

Hetherington, T. E. 1992. The effects of body size on functional properties of middle ear systems in anuran amphibians. Brain Behav. Evol. 39, 133-142.

Heyer, W. R., A. S. Rand, C. A. G. Da Cruz, O. L. Peixoto, and C. E. Nelson 1990. Frogs of Boracea. Arq. Zool. 31, 231-410.

Hyatt, G. W. 1977. Field studies of size-dependent changes in waving display and other behavior in the fiddler crab, Uca pugilator (Brachyura, Ocypodidae) Mar. Behav. Physiol. 4, 283-292.

Ibanez, R. D., C. A. Jamarillo, y F. A. Solis. 1995. Una especies nueva de Atelopus (Amphibia: Bufonidae) de Panama. Carrib. J. Sd. 31, 57-64.

Jaslow, A. P., T. E. Hetherington, and R. E. Lombard. 1988. Structure and function of the amphibian middle ear. In: The Evolution of the Amphibian Auditory System (Ed. by B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak), pp. 69-92. New York: John Wiley and Sons.

Jaslow, A. P. and R. E. Lombard. 1996. Hearing in the Neotropical Frog, Atelopus chiriquiensis. Copeia 1996, 428-432.

93 Lindquist, E. D. and T. E. Hetherington. 1996. Field studies on visual and acoustic signaling in the “earless Panamanian golden frog, Atelopus zeteki. J. Herpetol. 30, 347-354.

McDiarmid, R.W. 1971. Comparative morphology and evolution of frogs of the genus Atelopus, Dendrophiyniscus, Melanophryniscus, and Oreophrynella. Bull. Los Ang. Co. Mus. Nat. Hist. 12:1-66.

NoruSis, M. J. 1993. SPSS for Windows™, Rel. 6.0. NoruSis/SPSS Inc.

Pombal, J. P. Jr., I. Sazima, and C. F. B. Haddad. 1994. Breeding behavior of the Pumpkin Toadlet, Brachycephalus ephippium (Brachycephalidae). J. Herpetol. 28, 516-519.

Richards, S. J. and C. James. 1992. Foot-flagging displays of some Australian frogs. Mem. Qd. Mus. 32, 302.

Roy, D., B. Bijoylakshmi, and A. Sarma. 1995. Analysis and significance of female reciprocal call in frogs. Curr. Sd. 69, 265-270.

Smith, S. M. 1996. The single wing-flick display of the Black-capped Chickadee. Condor 98, 885-887.

Smith, W. J. 1977. The Behavior of Communicating: An Ethological Approach, Cambridge: Harvard University Press.

Wells, K. D. 1977. The courtship of frogs. In: The Reproductive Biology of Amphibians (Ed. by D. H. Taylor and S. I. Guttman). pp. 233-262. New York: Plenum Press.

Wells, IC D. 1988. The effect of sodal interactions on anuran vocal behavior. In: The Evolution of the Amphibian Auditory System (Ed. by B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak), pp. 433-454. New York: John Wiley and Sons.

Wever, E.G. 1985. The Amphibian Ear. Princeton: Princeton Univ. Press.

Zucker, N. 1983. Courtship variation in the Neo-tropical fiddler crab, Uca deichmanni: Another example of female indtation to male competition? Mar. Behav. Physiol. 10, 57-79.

94 CHAPTERS

BIOPHYSICAL AND NEUROPHYSIOLOGICAL STUDIES ON AUDITION IN EARED AND EARLESS ANURANS OF THE GENUS ATELOPUS.

Abstract.—The acoustic responsiveness of different body surface tissues and

auditory midbrain sensitivities to sound were measured in species of Atelopus

that either possess or lack a tympanic middle ear (“eared” and “earless”,

respectively). Acoustic responsiveness of bodv tissues was measured with a

laser Doppler vibrometer, and the body wall directly overlying the lung was

found to be very responsive to sound in all species tested. The otic (lateral

head) region showed little responsiveness in earless species. The otic region of

the eared A. flavescais showed significant responses to high frequency sound,

but body wall responsiveness was higher than otic responses over the entire

frequency range. Peak body wall responsiveness occurred within the dominant

frequency range of each species’ advertisement vocalization. Peak

responsiveness of the otic region of the eared A. flavesceiis also occurred within the dominant frequency range for the species advertisement vocalization.

95 Multi-unit neurophysiological recordings of the auditory midbrain {toms semicircularis) were obtained for one eared and two earless species of Atelopus.

Auditory sensitivity curves were constructed and showed three distinct regions of sensitivity at low, middle, and high frequency ranges, the latter sensitivity falling within the dominant frequency range of each species’ advertisement vocalization.

Auditory sensitivity cur\^es showed a close match with the acoustic responsiveness of the body wall of all species and with that of the otic region of the eared species. This correlation suggests that the body wall/lungs serve as the route of sound transfer to the inner ear in earless species and possibly in eared species as well.

INTRODUCTION

Vocalizations play an important role in the mating and territorial behavior of most anuran amphibians (Wells, 1977), and most species have well- developed tympanic middle ears, consisting of an external tympanum, air-filled middle ear cavity, Eustachian tubes, and auditory ossicle, for sound reception

(Wever, 1985; Jaslowet al., 1988). However, several species of anurans entirely lack a tympanic middle ear, an anatomical condition often referred to

96 as “earless,” although they have standard inner ears containing well-developed auditory sensory organs (Wever, 1985; Jaslowet al., 1988). Despite the lack of a tympanic middle ear, many earless species of frogs produce a variety of vocalizations, suggesting that they are using alternative pathways of sound reception for detecting their calls (Jaslow et al., 1988; Cocroft et al., 1990).

Most species in the genus Atelopus entirely lack tympanic middle ears ( = earless), but some species possess reduced tympanic middle ears (= eared).

The latter middle ear configuration includes all the standard components except a specialized tympanum. McDiarmid ( 1971 ) concluded that the loss of the tympanic middle ear was a derived condition in Atelopus. In the eared species of Atelopus the extracolumella attaches to unspecialized skin bordering the lateral edge of the middle ear cavity. Both eared and earless species of

Atelopus produce a variety of calls (Cocroft et al., 1990), and Lindquist and

Hetherington (1996) demonstrated that the earless A. zeteki displays behavioral responses to conspecific vocalizations in the field. Wever (1985) was able to measure cochlear microphonie responses to airborne sound in several species of earless Atelopus, and studies of midbrain responses by Jaslow and Lombard

(1996) found that A. cliiriquieusis is surprisingly sensitive to sound. The level of auditory sensitivity observed suggests that effective nontvmpanic path wavs of sound transfer are present in these species.

97 There is evidence that some frogs may use the lateral body wall and lungs as a nontympanic route for sound reception (Narins et ai., 1988; Ehret et al.,

1990; Hetherington, 1992; Ehret et al., 1994). Measurements of the acoustic responsiveness of body surfaces using a laser Doppler vibrometer have demonstrated that the lateral body wall directly overlying the lungs is verv responsive to sound (Narins et al., 1988; Jorgensen, 1991; Hetherington,

1992; Ehret et al., 1994). In completely eared species, pressure waves within the lungs can pass forward through the glottis, mouth cavity, and Eustachian tubes to the middle ear cavity and affect tympanic motion (Ehret et al., 1990; lorgensen et al., 1991). Ehret et al. (1994) demonstrated that attenuation of high frequency sound along the lung-eardrum pathway is pronounced, so that the contribution of the lung to tympanic responsiveness is restricted to relatively low frequencies. It has been hypothesized that the body wall and lungs can directly transfer sound energy to the inner ear of earless species of frogs (Narins et al., 1988; Hetherington, 1992), although there has been no experimental demonstration of the precise route.

The presence of both eared and earless species within the genus Atelopus provides a unique opportunity to study different strategies of sound reception

(tympanic vs. nontympanic types) in closely related and otherwise anatomicallv similar animals. This studv analvzed both the auditorv midbrain

98 auditory sensitivity and the acoustic responsiveness of different body surfaces in eared and earless species of Atelopus. Potential peripheral sound pathways were analyzed by directly correlating the acoustic properties of the different bodv surfaces with auditorv midbrain sensitivities.

METHODS

Animals.—The research reported here was performed under guidelines established by the National Institutes of Health (USA). Laser vibrometric and neurophysiological tests were conducted on Atelopus Jlavescens, an “eared” species, and on two undescribed “earless” species, Atelopus sp. (Chingaza) from

Parque Natural Nacional Chingaza, Colombia and Atelopus sp. (Nusagandi), from Parque Nacional Nusagandi, Kuna Yala, Panama. Measurement ranges for each species/sex are as follows: A. Jlavescens o"=28mm/1.3g and ? =29-

33m m /1.8-1.9g, AWopws sp. (Chingaza) o"=27-33mm/1.4-1.8g and 9=32-

39mm/1.5-1.3.Ig, Atelopus sp. (Nusagandi) o"=27-28mm/l.l-1.5g and

? =31 mm/1.7g Histological examination has demonstrated that A. jlavescens lacks a specialized tympanum but possesses a well-developed middle ear cavity and auditory ossicle (McDiamiid, 1971; pers. obs.). All frogs were wild caught adults and were obtained legally from the country of origin. Chingaza frogs

99 are a high elevation species (4500 m above sea level) and were kept in a

standard 40 gallon terrarium in a refrigerated room (7-15 C). Atelopus sp.

(Nusagandi) SLnd A. flavescens are low elevation species (500 -1700 m above sea

level) and were kept in standard 10 gallon and custom made 150 gallon

terraria at room temperature (23 C) at the Ohio State University, Columbus,

Ohio, USA. The animals were maintained on a 12:12 light/dark cycle.

Laser Vibrometry.—Laser Doppler vibrometric (LDV) experiments were

perfomied on animals lightly anesthetized by immersion in a 1 % tricaine

methanosulfonate salt (MS-222) solution so that they sat motionless during

experimentation. Animals were then placed in a sound attenuating box where

a neon-helium laser beam from a laser Doppler vibrometer (Polytec OFV

1000) was focused on a small metallic reflector (0.3 x 0.3 mm) placed on certain body surfaces. LDV measurements were made at three body surface points: the lateral body wall directly over the lung, the otic region over the inner region, and rostrum (snout midway between the external nares and eye).

The latter represented a control body surface. Animals were moistened frequently to keep the skin surfaces from drying. Adult frogs (A. Jlavesccns

[Io":4?]; Atelopus sp. {Chingaza} [Icf-.l?]; Atelopus sp. {Nusagandi ) [3cf:l?]) were used in all LDV experiments.

1 0 0 Computer generated sinusoidal acoustic frequencies were amplified and

delivered to frogs through a full range 25 cm speaker using Wave SE acoustic

software. One third octave frequencies ranging from 160 Hz to 6300 kHz

were played at 90 dB SPL (RMS - slow) from a distance of 30 cm from the test

subject. LDV measurements were made on body surfaces ipsilateral to the

speaker. Sound pressure level (SPL) was monitored with a 1 cm condenser

microphone (Briiel &. Kjaer Type 4155) positioned above the animals, aimed

toward the speaker, and connected to a sound level meter (Briiel &. Kjaer Type

2230) and third octave/octave filter set (Briiel &. Kjaer Type 1625). LDV output voltage for each tested frequency was measured on a wave analyzer

(Hewlett Packard 3581 A), averaged across individuals of each species, and converted to relative velocity (dB=20 log V/V^; V„= 1 /im/sec). Individual animals were tested once and values were obtained for each anatomical region at every frequency. The acoustic responsiveness of each anatomical region for each individual was averaged at each frequency point to vield an average tissue surface response curve for each species.

Neurophysiology .—Surgeries for neurophvsiological experiments were performed on animals anesthetized by immersion in a 1% tricaine methanosulfonate salt solution. The dorsal surface of the midbrain was

101 exposed through an opening in the parietal bones of the skull. After surgery, animals were immobilized for multi-unit midbrain recordings with an intramuscular injection of 15 fxVg body weight of 9 m^ml d-tubocurarine chloride. The dosage required for Atelopus is unusually high for anurans.

Individual frogs were placed on an experimental stage located in a sound and vibration attenuating booth. Animals used in the neurophvsiological experiments had to be completely immobilized and were positioned in a resting posture less upright than in the LDV experiments. A tungsten electrode (FHC Inc, 11-13 MÛ) was inserted 220 to 775 fim deep into the torus seinidradaris (TS, auditory midbrain) by use of a stepping-motor microdriver (M. Walsh Engineering, UD - 200). Computer generated sinusoidal tone pips were amplified and delivered to frogs through a full range speaker using custom acoustic software (as described in Volman, 1996).

Auditory stimuli lasted 250 msec and had 5 msec rise and fall times and were repeated every 3.5 sec. Multicellular responses were amplified with an AC amplifier (A-M Systems 1800) and played on an external speaker. Hearing thresholds for each species were determined by auditory monitoring of multiunit responses as the amplitude of tone pips was manuallv attenuated bv

1 dB increments. TS threshold sensitivities were obtained at frequencies

102 ranging from 100 Hz to 5000 kHz at 100 Hz intervals. Presentation order of stimulus frequencies was randomized.

Atelopus species are small and have been found to be relatively fragile during surgery and neurophysiological experiments compared to other anurans; hence, recordings were made from the TS rather than the eighth nerve. The midbrain is quite small, and in some individuals only one complete electrode pass was possible before TS responses decreased. Midbrain responses were measured from two A. flavescais (two complete curves from one animal and one complete and one partial curve from another individual), six male and one female Atelopus sp. (Nusagandi) (three complete curves from one individual, two curves from each of two other individuals, and a single curve from each of the remaining four individuals), and five male Atelopus sp.

(Chingaza) (three complete curves from one individual, two curves from each of two other individuals, and a single curve each from another two individuals). The minimum audible thresholds from each individual were averaged to yield a mean audibility curve for each species. A minimum threshold curve for each species also was plotted by incorporating minimum values among all individuals of that species at all sound frequencies.

103 RESULTS

Laser Vibrometiy.—The acoustic responsiveness of tissue surfaces varied dramatically among the anatomical regions tested. Acoustic responses of the body wall directly over the lung yielded the highest vibration velocities for all species (Figs. 5.1 - 5.3). The body wall showed high responsiveness at low frequencies below about 400 Hz and even greater responsiveness at higher frequencies centered around 2500 Hz (Figs. 5.1 - 5.3). The high frequency range of acoustic responsiveness falls within the dominant frequency range of the species’ advertisement vocalization for each species (shown as a bar in Figs.

5.1 - 5.3). The rostral (control) region of the head provided consistently low vibration velocities at all test frequencies for each species (Figs. 5.1 - 5.3). The mean acoustic sensitivity across all test frequencies of bodv wall tissues was significantly greater than that of the rostrum (P <0.003 for all species; paired t-test). The acoustic responsiveness of the otic region was low for the two earless Atelopus sp. over the entire range of frequencies (Figs. 5.2 - 5.3).

Responsiveness of the otic region of the eared A. flavescens was similarly low at most frequencies, but markedly greater around 2500 Hz, thereby generallv matching the dominant frequency' range of the species advertisement vocalization (Fig. 5.1).

104 Figure 5.1 Mean acoustic responsiveness of tissue surfaces: solid line, body wall region (lung) with solid standard error bars; long dashed line, otic region with standard error bars; short dashed line, rostral region. Bar represents the dominant frequency range of the species’ advertisement vocalization (Lescure, 1981). A. flavescens (n=5)

105 Figure 5.1

Atelopus flavescens

CO T 3 \ \ 0o > > 1 \ « 1 /

10 dB

100 1000 10000 Frequency (Hz)

106 Figure 5.2 Mean acoustic responsiveness of tissue surfaces: solid line, body wall region (lung) with solid standard error bars; long dashed line, otic region; short dashed line, rostral region. Bar represents the dominant frequency range of the species’ advertisement vocalization (C. Navas, pers. com.). Atelopus sp. (Chingaza, n=4)

107 Figure 5.2

Atelopus sp. (Chingaza)

-oSQ

U \

.Ë V

'Z

10 dB

100 1000 10000 Frequency (Hz)

108 Figure 5.3 Mean acoustic responsiveness of tissue surfaces: solid line, body wall region (lung) with solid standard error bars; long dashed line, otic region; short dashed line, rostral region. Bar represents the dominant frequency range of the species' advertisement vocalization (R. Ibanez, pers. com.) Atelopus sp. (Nusagandi, n=4)

109 Figure 5.3

Atelopus sp. (Nusagandi)

ca •a

o .Ë « u

10 dB

ICO 1000 10000 Frequency (Hz)

1 1 0 Neurophysiology. —TS threshold curves showed three distinct frequency"

regions of sensitivity in both the eared and earless species of Atelopus: a low

frequency (below 400 Hz) region, a middle frequency (centered around 1500

Hz) region, and a high frequency (> 2,000 Hz) region (Figs. 5.4 - 5.6). The high frequency range of auditory sensitivity falls within the dominant frequency range of the species advertisement vocalization for each species

(Figs. 5.4 - 5.6).

Figure 5.7 shows a comparison of the average auditory threshold curves for eared and earless species of Atelopus. The auditory sensitivity measured for the two specimens oi A. flavescens demonstrate that this eared species is slightly (8-

13 dB SPL) more sensitive to high frequency sound than both earless species.

Earless Atelopus on the other hand appear to have greater sensitivity to low frequency (below about 400 Hz) sound than does A. flavescens.

DISCUSSION

Comparative TS Sensitivities.—Overall, the shape of TS sensitivity curves in eared and earless Atelopus was similar, although differences in absolute sensitivity were apparent at certain sound frequencies. The eared A. flavescens was roughly 8 to 13 dB more sensitive at high frequencies (>2.5 kHz)

111 Figure 5.4 Multiunit auditory sensitivity threshold curves recorded from toms sanidrculciris: upper solid line, mean auditory threshold with standard error bars; lower solid line, minimum auditory sensitivity for species. Bar represents the dominant frequency range of the species’ advertisement vocalization (Lescure, 1981). A. flavescens (n=2)

112 Figure 5.4

no Atelopus flavescens

ICO

90 ^ C? Ch ce 80 /A 2 / c I ™ _c H 60

50 ■

40 100 1000 10000 Frequency (Hz)

113 Figure 5.5 Multiunit auditory sensitivity threshold curves recorded from toms seniicircularis: upper solid line, mean auditory threshold with standard error bars; lower solid line, minimum auditory sensitivity for species. Bar represents the dominant frequency range of the species’ advertisement vocalization (C. Navas, pers. com.). Atelopus sp. (Chingaza, n=5)

114 Figure 5.5

no Atelopus sp. (Chingaza) JV ICO A i- 90 - / \ Ch OO \ / > , A" ca 80 2- 2 o 70 CJC/5

60

50

40 100 1000 10000 Frequency (Hz)

115 Figure 5.6 Multiunit auditory sensitivity threshold curves recorded from toms smicirculciris: upper solid line, mean auditory threshold with standard error bars; lower solid line, minimum auditory sensitivitv for species. Bar represents the dominant frequency range of the species’ advertisement vocalization (R. Ibanez, pers. com.). Atelopus sp. (Nusagandi, n=7)

116 Figure 5.6

110 Atelopus sp. (Nusagandi)

100 S 90 - I' , & c/2 J CQ 80 - 3 v,!r 2 Il I o ■ü 70 - ï c-, 60 -

50

40 100 1000 10000 Frequency (Hz)

117 Figure 5.7 Mean TS auditory sensitivity for eared and earless species compared. F = A. flavt'scais; C = Atelopus sp. (Chingaza); and N = Atelopus sp. (Nusagandi).

1 1 8 Figure 5.7

no Comparative TS Sensitivity inAtelopus

100

90 N

c/2 ' F ffl 80 3 20 1 ™ c_JZ 60 -

50

40 100 1000 10000 Frequency (Hz)

119 compared to the earless Nusagandi and Chinga-za. Atelopus sp. (see Fig. 5.7).

However, earless species showed higher auditory midbrain sensitivity (by as

much 24 dB) at low frequencies below 500 Hz (see Fig. 5.7).

The general shape of the midbrain tuning curves was similar for both eared

and earless species. All specimens showed good sensitivity to low frequency sounds below about 400 Hz, to middle frequency sounds around 1500 Hz, and to higher frequency sounds from about 2500-3000 Hz. These three most sensitive regions of the tuning curves appear to represent responses from different auditory hair cell populations. The low, middle, and high frequency sensitivities likely are associated with the saccule and/or amphibian papilla, the amphibian papilla, and basilar papilla respectively. This pattern is reflected in many other anuran species (Feng et al., 1975; Moffat and Capranica, 1976;

Wilczynski and Capranica, 1984; Zakon and Wilczynski, 1988). The high frequency TS sensitivities of both the eared and earless Atelopus are tuned to the range of dominant frequencies of the species’ advertisement vocalization

(Figs. 5.4 - 5.6). The band of auditory sensitivity centered around 1,500 Hz found in all three species examined in this study (Figs. 5.4 - 5.6) was not described by Jaslow and Lombard (1996) for A. chriricpdensis. The difference in techniques used in the two studies, specifically the use of chronic electrode implants on three individuals and pulsed sound signals in the Jaslow and

1 2 0 Lombard study, may be responsible for the differences in the shape of the

sensithdty curves.

Correlations Betiveai L D V Measuremaits and TS Sensitivity.—This study demonstrated that the lateral body wall overKnng the lungs is extremely responsive to sound in both eared and earless species of Atelopus. The bodv wall showed a low frequency range of responsiveness below about 400 Hz and a higher frequency range of responsiveness centered around 2500 Hz. The latter zone of acoustic responsiveness corresponds well to the range of dominant frequencies in the species’ advertisement vocalization (Figs. 5.1 -

5.3). The body wall was much more responsive to sound than the rostral region of all species and the otic (lateral head) region of the earless species across the full range of frequencies. The body wall area of the eared A. flavescens also was more responsive than the otic region of this species across the full range of frequencies, although the otic area did show a high frequency peak in responsiveness at about 2500 Hz. The latter peak in otic responsiveness was nonetheless significantly lower than the same high frequency peak in body wall sensitivity. Therefore, both the otic region and body wall of the eared A. flavescens displayed peaks in responsiveness near the dominant frequencies of advertisement vocalizations, although the body wall

121 showed higher amplitude responses. The close match in peak responsiveness in otic and body wall surfaces may reflect only that lung resonance is driving the otic response. Figures 5.8 - 5.10 show the correspondence between body wall responsiveness (and otic responsiveness (or A. Jlavescens)sinà. the high frequency region of sensitivity of the TS in Atelopus.

The Structural Basis of Hearing in Eared and Earless Species.—The surprising similarity in auditory midbrain sensitivities between eared and earless species of Atelopus suggests that nontympanic pathways of sound reception can be as effective as a standard tyntpanic middle ear. The presence of a tympanic middle ear typically would be expected to impart better sensitivity to high frequency sound, and this was indeed observed, although the increase in sensitivity (8-13 dB) was not dramatic. A. flavescens lacks a specialized tympanum, however, and comparison of the high frequency sensitivity of the earless species to that of a species with a fully developed tympanum may provide more pronounced differences. The body wall of all of the species examined showed patterns of acoustic responsiveness that matched the high frequency' areas of peak TS sensitKaty. The middle frequency peak in TS sensitivity was not reflected by pronounced responsiveness of the body wall.

Perhaps at these middle frequencies, likely associated with responses from the

122 Figure 5.8 Correlation between mean TS auditory threshold and surface responsiveness of the body wall/lung region: upper solid line, mean TS auditory threshold; lower solid line, mean acoustic responsiveness of body wall/lung surface; dashed line, mean acoustic responsiveness of otic surface. Bar represents the dominant frequency range of the species’ advertisement vocalization (Lescure, 1981). Atelopus flavescens

123 Figure 5.8

110 Atelopus flavescens 100

90

80

CQ É. 70 -a

® 60 - _oO O _o 50 - > g -c 40, •

30

20 10 dB

10

0 100 1000 10000 Frequency (Hz)

124 Figure 5.9 Correlation between mean TS auditory threshold and surface responsiveness of the body wall/Iung region: upper solid line, mean TS auditory threshold; lower solid line, mean acoustic responsiveness of body wall/lung surface. Bar represents the dominant frequency range of the species’ advertisement vocalization (C. Navas, pers. com.). Atelopus sp. (Chingaza)

125 Figure 5.9

110 . Atelopus sp. (Chingaza) ICO -

es "O

UO g .Ë ce i 30

20 10 dB

100 1000 10000 Frequency (Hz)

126 Figure 5.10 Correlation between mean TS auditory threshold and surface responsiveness of the body wall/Iung region: upper solid line, mean TS auditory threshold: lower solid line, mean acoustic responsiveness of body wall/Iung surface. Bar represents the dominant frequency range of the species’ advertisement vocalization (R. Ibanez, pers. com.). Atelopus sp. (Nusagandi)

127 Figure 5.10

no . Atebpus sp. (Nusagandi) 100

90

80 -

S 70 -

3 60 OU 2 g o 50 O "w > E f . 40 H 30 -

20 - 10 dB 10 -

0 - 100 1000 10000 Frequency (Hz)

128 amphibian papilla (Feng et al., 1975; Wilczynski and Capranica, 1984; Zakon

and Wilczynski, 1988), other nontympanic pathways of sound reception are operating.

Most studies of body wall/lung responses to sound have emphasized how lung-bome sound can be transferred through the glottis and Eustachian tubes and directly modify tympanic motion (Narins et al., 1988, Ehret et al., 1990;

Ehret et al., 1994). Given that earless species ofAtdopus entirely lack a tympanic middle ear, including Eustachian tubes, sound energy penetrating the lateral body wall and entering the lung must be transmitted to the inner ear by another route. Narins et al. (1988) proposed one potential route from the lung to fluid spaces surrounding the spinal column, and from there to the endolymphatic sac and into the inner ear. Several possible pathways exist, and additional research is needed to isolate the routes of sound transfer.

Ehret et al. (1994) found that high frequency signals passing from the lungs to the middle ear cavity were greatly attenuated, and that this route was more effective for transmission of low frequency signals. This study demonstrates that earless species are on average 8-13 dB less sensitive to high frequencies than a species with a slightly reduced tympanic middle ear, and anurans with a specialized tympanum may be even more sensitive to high frequencies.

Although the lung-inner ear pathway may be less effective at high frequencies

129 than a tympanic middle ear, this study suggests that high frequency signals

nonetheless can be transferred from the Itmgs to the inner ear. The alternative

would be that high frequency sound might directly penetrate the lateral head

tissues and pass to the inner ear, but the very low acoustic responsiveness of

head tissues to high frequencies observed in this study suggests that this is not

the case.

It is unclear how much of the auditory sensitivity of the tAxtd A. Jlavescens

depends on lung sound reception. As mentioned above, it has been

demonstrated that sound acting on the bodv wall and lungs of frogs with

tympanic middle ears can pass through the glottis and Eustachian and directly

produce tympanic motion (Narins et al., 1988; Ehret et al., 1990; Ehret et al.,

1994). However, nontympanic routes of sound transfer from the lungs to the

inner ear that operate in earless frogs may function in eared species as well.

VVilcz\Tiski et al. (1987) demonstrated that in Rana pipiens auditory nerve

responses to sounds below about 500 Hz could be produced equallv well by

sound transfer along tympanic or nontympanic pathways. However, aspects of nontympanic sound reception may be body-size dependent, as the bodv wall and lungs of small anurans may be more responsive to high frequency sound

(Hetherington, 1992). As the species o f Atelopus used in this studv are smaller than Rana pipiens, the effectiveness of their nontympanic sound transmission

130 may extend to higher frequencies. Therefore, it is possible that body wall reception of sound can contribute to auditory responses over a broad range of frequencies in the eared A. flavescens.

The greater auditory sensitivity o î A. flavescens to high frequency sounds compared to its earless congeners may result from two potential mechanisms.

First, such sensitivity may be due to direct reception of these signals by the otic area and extracolumella/columella. Although the body wall actually is more responsive than the otic region to these high frequency sounds, there may be pronounced attenuation of sound energy, especially at high frequencies, during transfer from the lungs. Second, sound transferred from the body wall and lung via the glottis and Eustachian tubes may increase otic area/columellar responses to sound at these frequencies. Overall, however, the contribution of the body wall/ lung to hearing in the &2iTté. A. flavescens probably is important at low frequencies. Because the otic region of this species showed little responsiveness to low frequency sound, direct transfer of sound energ): from the lungs to the inner ear seems the most likely explanation.

The poorer auditory sensitivity of the eared A. flavescens to low frequencies compared to the earless species may be due to phase cancellation between sound concurrently transmitted from the body wall and otic region, but can

131 not precisely be explained by this study. Possibly the earless species possess

specializations of either their nontympanic pathways or inner ear that enhance

sensitivity to lower frequencies.

The Functional Basis of Middle Ear Loss in Frogs .—The adaptive significance of

the differences in middle ear morphology in the neotropical genus Atelopus

remains speculative. Both eared and earless species commonly are found along streams, but the eared species tend to live in lowland areas and the earless species inhabit both lowland and montane regions. It has been proposed that the montane stream habitats are noisier, thereby making acoustic communication more difficult (Heyer et al., 1990) and minimizing the selective advantage of a tympanic middle ear. However, results from this and other studies demonstrate that earless Atelopus hear fairly well, and field studies have established that they respond behaviorally to conspecific vocalizations (Lindquist and Hetherington, 1996). It is possible that the major selective advantage of a tympanic middle ear concerns its function in sound localization rather than any enhancement of hearing sensitivity. The anuran tympanic middle ear has been proposed to function as a pressure gradient receiver (Feng and Capranica, 1976; Feng, 1981; Feng and Shofner,

1981), and sound transfer from the lungs via the glottis has been proposed to

132 improve the directional sensitivity of the tympanum to low frequency sound

(Ehret et al., 1990; forgensen et al., 1991; Ehret et al., 1994). Although

behavioral tests have demonstrated that earless Atelopus can locate sound

sources (Lindquist and Hetherington, 1996), the relative effectiveness of a

lung-based system of sound reception not involving a tympanic middle ear for

sound localization remains unclear.

Additional insights into the functional and evolutionary basis of different

middle ear configurations in anurans may be obtained by direct comparisons of

audition in closely related species with fullv-developed, reduced, and absent

tympanic middle ears. Although no species of Atelopus possess a complete

tympanic middle ear, there is a closely related species {Frostius pemambucensisi

fomxtrly Atelopus peniambiicensis) that does possess one (Cannatella, 1986).

Studies of this species could provide a more comprehensive comparison of the

effectiveness of sound reception by different middle ear configurations.

Acbwwledgtnents. —I wish to thank Carlos Navas and Roberto Ibanez Diaz

for information on dominant call frequencies of Atelopus sp. (Chingaza) and

(Nusagandi) respectively. I am grateful to Susan Volman for the use of her neurophysiological equipment and laboratory. A1 Feng provided useful advice on surgical procedures. I also am grateful for the technical assistance of Susan

133 Volman, Kris Schuett, Mitch Masters, and Jim Fox. Walt Wilczynski

provided useful comments on an early draft of this chapter. This work was

supported in part by NIH grant #MH47330. These experiments comply with

the “Principles of animal care,” publication No. 86-23, revised 1985 of the

National Institutes of Health and also with the current laws of the State of

Ohio, USA.

BIBLIOGRAPHY

Cannatella, D. C. 1986. A new genus of bufonid (Anura) from South America, and phylogenetic relationships of the neotropical genera. Herpetologica 42:197-205.

Cocroft, R. B., R. W. McDiamtid, A. P. Jaslow, and P. M. Ruiz-Carranza. 1990. Vocalizations of eight species oî Atelopus (Anura: Bufonidae) with comments on communication in the genus. Copeia 1990:631-643.

Ehret, G., J. Tautz, and B. Schmitz. 1990. Hearing through the lungs: lung- eardrum transmission of sound in the frog Eleutherodactyliis cocpii. Naturvvissenschaften 77:192-194.

Ehret, G., E. Keilwerth, and T. Kamada. 1994. The lung-eardrum pathway in three treefrog and four dendrobatid frog species: some properties of sound transmission. J. Exp. Biol. 195:329-343.

Feng, A. S. .1981. Directional response characteristics of single neurons in the torus semicircularis of the Leopard frog [Ram pipiens). J. Comp. Phvsiol. 144:419-428. Feng, A., P. M. Narins, and R. R. Capranica. 1975. Three populations of primary auditor)^ fibers in the bullfrog {Rana catesbiana): Their peripheral origins and frequency sensitivities. J. Comp. Physiol. 100:221-229.

134 Feng, A. S. and R. R. Capranica. 1976. Sound localization in anurans 1. Evidence of binaural interaction in the dorsal medullary nucleus of bullfrogs {Rana catesbiana). J. Neurophysiol. 39:871-881.

Feng, A. S. and W. P. Shofner 1981. Peripheral basis of sound localization in anurans, acoustic properties of the frog’s ear. Hearing Res. 5:201-216.

Hetherington, T. E. 1991. The effects of body size on the evolution of the amphibian middle ear. In A. Popper, D. Webster, R. Fay (eds). The evolutionary biology of hearing. Springer-Verlag, New York pp. 421-437.

Hetherington, T. E. 1992. The effects of body size on functional properties of middle ear systems in anuran amphibians. Brain Behav. Evol. 39:133-142

Jaslow, A. P., T. E. Hetherington, and R. E. Lombard. 1988. Structure and function of the amphibian middle ear. In B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, W. Walkowiak (eds). The evolution of the amphibian auditory system. John Wiley and Sons, New York pp. 69-92.

Jaslow, A. P. and R. E. Lombard 1996. Hearing in the Neotropical Frog, Atelopusshiriqiiiensis. Copeia 1996:428-432.

Jorgensen, M. B. 1991. Comparative studies of the biophysics of directional hearing in anurans. J. Comp. Physiol. A 169:591-598.

Jorgensen, M. B., B. Schmitz, and J. Christensen-Dalsgaard 1991. Biophysics of directional hearing in the frog Eleutherodactyliis coqui. J. Comp. Phvsiol. A 168:223-232.

Lescure, J. 1973. Contribution a l’étude des Amphibiens de Guyane française. I. Notes SUT Atelopus flavescens Duméril et Bibron et description d’une nouvelle espèce. Vie Milieu 23:125-141.

Lescure, J. 1981. Contribution a l’étude des Amphibiens de Guyane française. Vlll. Validation dAtelopus spumarius Cope, 1871, et désignation d’un néotype. Description dAtelopus spumarius barbotini nov. ssp. Données étho- écologiques et biogéographiques sur les Atelopus du groupe flavescens (Anoures, Bufonidés). Bull. Mus. Natn. Hist. Nat. Paris 3:893-910.

135 Lindquist, E. D. and T. E. Hetherington. 1996. Field studies on visual and acoustic signaling in the "earless" Panamanian golden frog, Atelopus zeteki. J. Herpetol. 30:347-354.

McDiamtid, R. W. 1971. Comparative morphology and evolution of frogs of the genus Atelopus, Dendrophrynisais, Melanophrynisais, and Oreophtynellci. Bull. Los Ang. Co. Mus. Nat. Hist. 12:1-66.

Moffat, A. J. M. and R. R Capranica. 1976. Auditory sensitivity of the saccule in the American toad {Bufo americams). J. Comp. Physiol. A 105:1-8.

Narins, P. M., G. Ehret, and J. Tautz. 1988. Accessory pathway for sound transfer in a neotropical frog. Proc. Nat. Acad. Sci. USA. 85:1508-1512.

Volman, S. 1996. Qualitative assessment of song-selectivity in the zebra finch “high vocal center.” }. Comp. Physiol. A 178:849-862.

Wells, K. D. 1977. The courtship of frogs. In D. H. Taylor and S. I. Guttman (eds) The reproductive biology of amphibians. Plenum Press, New York, pp. 233-262.

Wever, E. G. 1985. The amphibian ear. Princeton University Press, Princeton.

Wilczynski, W. and R. R. Capranica. 1984. The auditory system of anuran amphibians. Progress. Neurobiol. 22:1-38.

Wilczynski, W., C. Res 1er, and R. R. Capranica. 1987. Tympanic and extratvmpanic sound transmission to the leopard frog. J. Comp. Phvsiol. 161:659-669.

Zakon, H. H. and W. Wilczynski. 1988. The physiology of the anuran eighth nerve. In B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak (eds). The evolution of the amphibian auditory system. John Wiley and Sons, New York pp. 125-155.

136 CHAPTER 6

LUNG-BASED HEARING IN AN “EARLESS” FROG

Abstract.—The mechanisms of hearing in the “earless” hog Bombina orientalis were studied using laser Doppler vibrometric and neurophysiological techniques. Laser vibrometry demonstrated that the body wall overlying the lung was very responsive to sound compared to the lateral head surface overlying the inner ear. Auditory midbrain thresholds were dramatically decreased at all frequencies tested after covering the body surface between the forelimbs and hindlimbs with silicone grease. Filling the lungs with oxygenated saline also drastically decreased hearing sensitivity. Tests showed that suturing the glottis closed had minimal effect on midbrain responses to sound, suggesting that sound transfer from the lungs to the inner ear via the buccal cavity is relatively unimportant. The precise route of lung-based sound transfer remains unclear. Results of this study demonstrate that such a lung- based system can be used for hearing airborne sound as well.

137 INTRODUCTION

Most anurans possess a tympanic middle ear, comprised of a tympanum,

air-filled middle ear cavity, and ossicular elements (the extracolumella and

columella), that is specialized for reception of air-bome sound (Capranica,

1976; Wilczynski and Capranica, 1984; Jaslow et al., 1988). However, several

species of anurans have reduced tympanic middle ears (having lost a

tympanum) or lack them completely (Jaslow et al., 1988). The latter often are

referred to as “earless” species, although all retain well-developed auditory end

organs within the inner ear (Wever, 1985) and many produce a variety of

vocalizations and respond to sound behaviorally (Jaslow et al., 1988; Lindquist

and Hetherington, 1996). Such earless frogs presumably utilize non-tympanic

pathways of sound reception, and recent research has focused on the lungs as

an alternative route for sound transduction.

The lungs of frogs lie directly below the skin of the lateral body wall, and

the latter surface is very responsive to sound (Narins et al. 1988; Ehret et al.

1990; Jorgensen et al. 1991; Hetherington, 1992). In anurans with standard tympanic middle ears, sound pressure acting on the body wall can generate pressure waves within the lungs that can pass forward through the glottis and

Eustachian tubes and modify tympanic responses to sound (Narins et al.,

138 1988; Ehret et al., 1990). This body wall-lung pathway of sound reception and transfer may improve the directional sensitivity of the tympanic ear, especially at lower frequencies where the response of the body wall is greatest.

Effects of body wall sound reception on directional characteristics of the tympanum have been demonstrated in species with fully-developed tvmpanic middle ears (Jorgensen et al., 1991 ). The lungs have also been found to function as a low frequency channel for hearing underwater sound in the aquatic hogXenopus Icievis (Christensen-Dalsgaard and Elepfandt, 1995).

It has been proposed that sound energy can be transferred from the body wall and lungs directly to the inner ear in earless frogs. Narins et al. ( 1988) suggested that sound energy may be transferred along the endolymphatic sac system that extends from the vertebral canal to the inner ear (Narins et al.

1988). Sound energy also might be transmitted directly along the vertebral column. For example, Schellart and Popper (1992) suggested that sound pressure transduction by the airbladder of eels may produce compression waves through the vertebral column that would be less attenuated than those passing through soft tissues to the inner ear. It also is possible that sound energy from the lungs can pass through the glottis and into the mouth cavity and from there through the round window and into the inner ear. The round window of frogs typically is covered with soft body tissues and directed

139 somewhat posteriorly and ventrally toward the mouth cavity. Pressure waves

in the mouth cavity potentially could pass through soft tissues overlying the

round window and subsequently stimulate the auditory end organs (basilar

and amphibian papillae) and exit at the oval window. Such a "backward"

system apparently operates for underwater hearing in tadpoles. Witschi

(1949) described a bronchial columella in larval Rana tennporaria that presumably transfers motion of the lung wall generated by sound pressure to the round window. The bronchial columella is lost at metamorphosis, but, as mentioned above, sound energy may pass from the lung to round window via the glottis and buccal cavity in metamorphosed, earless anurans.

To date there has been no direct demonstration that earless frogs hear airborne sound via the body wall and lungs. One aim of this study, therefore, was to test whether sound reception via the body wall and lungs of an earless frog can generate responses to sound in the inner ear. Additional experiments then were conducted to test the different hypotheses concerning the precise pathways of sound transmission from the lungs to the inner ear. The fire- bellied toad {Bombina orientalis) was used in these studies. This readily obtainable species is essentially earless, as it lacks a tympanum and middle ear cavity (Wever, 1985; pers. obs.). Avery reduced columella remains, but it is attached only to a process of the hyoid apparatus.

140 MATERIALS AND METHODS

The experimental research reported herein comply with the “Principles of animal care,” publication No. 86-23, revised 1985 of the National Institutes of Health (USA) and with current laws of the state of Ohio (USA). Adult fire- bellied toads (B. orientalis) ranging in size from about 4-6 g were used in this studv.

Laser Doppler Vibrometiy.—Laser Doppler \abrometric (LDV) measurements of the acoustic responsiveness of the body wall and lateral head region were made on three animals lightly anesthetized by immersion in a 1 % tricaine methanosulfonate salt (MS-222) solution. Once anesthetized, the frogs sat motionless during measurements in the dark sound attenuating box used for the tests. A neon-helium laser beam from a laser Doppler vibrometer (Polytec

OFV 1000) was focused on a small metallic reflector (0.3 x 0.3 mm) placed on either the lateral head surface overlying the inner ear or on the lateral body wall directly over the lung. Animals were moistened frequently to keep the skin surfaces from drying. Pure tone sounds (at 100, 150, 200, 300, 400, 500,

600, 700, 800, 900, 1000, 1250, 1600, 2000, and 2500 Hz) were delivered at

90 dB SPL by a full range 25 cm speaker positioned 30 cm from the animals.

141 During experimentation, sound pressure level (SPL) was monitored with a I

cm condenser microphone (Brûel & Kjaer Type 4155) positioned 5 cm above

the animals, directed toward the speaker, and connected to a sound level meter

(Brûel & Kjaer Type 2230) and third-octave/octave filter set (Brûel &. Kjaer

Type 1625). LDV measurements were made on body surfaces ipsilateral to the

speaker. LDV output voltage for each tested frequency was measured on a

wave analyzer (Hewlett Packard 3581A) and converted to relative average

velocity (dB).

Neurophysiological Measuranents of Auditory Sensitivity .—Hearing sensitivity

was measured by determining auditory midbrain thresholds to sound stimuli.

Surgeries for midbrain recording were performed on animals anesthetized by

immersion in a 1 % tricaine methanosulfonate salt solution. The dorsal surface

of the midbrain was exposed through an opening in the parietal bones of the

skull. After surgery, animals were immobilized for multi-unit midbrain

recordings with an intramuscular injection of 1.5 fA/g body weight of 3 mg/ml d-tubocurarine chloride. Individual frogs were placed on an experimental stage located in a sound and vibration attenuating booth. Animals used in the neurophysiological experiments had to be completely immobilized and were positioned in a resting posture less upright than in the LDV experiments. A

142 tungsten electrode (FHC Inc, 11-13 M) was inserted at a depth of 115-675 /j.m into the torus semicircularis (auditory midbrain) by use of a stepping-motor microdriver (M. Walsh Engineering, UD - 200). Computer generated sinusoidal tone pips were amplified and delivered to frogs through a full range speaker using custom acoustic software (as described in Volman, 1996).

Auditory stimuli lasted 250 msec and had 5 msec rise and fall times and were repeated every 4.5 sec. Multicellular responses were amplified with an AC amplifier (A-M Systems 1800) and played on an external speaker. Hearing thresholds were determined by auditory monitoring of multiunit responses as the amplitude of tone pips was manually attenuated by I dB increments.

Midbrain threshold sensitivities were obtained at frequencies ranging from 100

Hz to 1000 Hz at 100 Hz intervals. Higher frequencies were sampled, but sensitivities were low and often not measurable following experimental manipulations. The order of presentation of stimulus frequencies was randomized.

Attenuation of Lung Sound Reception .—Two experimental manipulations were perfomied to test the role of the body wall and lung in sound reception. First, the effect of covering the lateral body wall with silicone grease on the auditory sensitivity of frogs was examined. These tests were conducted on five frogs.

143 and the order of treatment (body wall uncovered or covered) was altered to control for treatment order effect. In two cases, control midbrain thresholds were determined and then the entire body of the frogs between the forelimbs and hindlimbs was covered by an approximately 1.5 cm layer of silicone grease held within a strip of cheesecloth fabric. In three other cases, responses with the body wall covered were recorded first and the body wall subsequently uncovered for a series of control threshold measurements. This alternation of treatments insured that threshold differences were a result of the experimental manipulation and not a result of mortality.

Covering the entire body between the forelimbs and hindlimbs with silicone grease potentially could block nontympanic pathways of sound reception other than the lungs, so another test, in which the lungs were filled with saline, was conducted on five additional frogs. In three animals, control midbrain thresholds were measured and the lungs of the animals then filled with oxygenated saline by injection through the glottis. Midbrain sensitivity was then measured again. In two other animals, the lungs first were filled with oxygenated saline, midbrain responses measured, and then the lungs were emptied of saline by gently applied pressure to the body wall. The lungs were then re-inflated with air with a 10 cc syringe. A series of control midbrain thresholds were subsequently measured.

144 Obstructing Sound Transfer Through the Glottis .—An additional test was

performed to examine whether sound energy is transferred from the lungs to

the inner ear via the glottis and mouth cavity. In five frogs, midbrain

threshold sensitivities were measured, and then the glottis of each animal was

closed by a silk suture. Subsequently, another set of midbrain thresholds were

measured.

RESULTS

Laser Doppler Vibrometry.—The body wall overlying the lungs had greater

acoustic responsiveness than the lateral head tissues over the inner ear (Fig.

6.1 ). The lateral body wall showed peak motion at low frequencies below about 300 Hz and also at higher frequencies close to 1000 Hz. Motion of the lateral head surface was much less responsive through the frequency range, although there was a slight increase in responsiveness around 1000 Hz.

Effects of Covering the Lateral Body Wall on Hearing Sensitivity .—Covering the lateral body wall with a layer of silicone grease dramaticallv decreased midbrain sensitivity across all frequencies tested (Fig. 6.2). The effect was the same regardless of the treatment order (i.e., animals in which the bodv wall

145 Figure 6.1 Laser Doppler vibrometric measurements of the acoustic responsiveness of the lateral head tissues (dashed lines) and body wall overlying the lung (solid lines) for three Bombina orientalis. Velocity is represented on a relative dB scale. The body wall over the lung is much more responsive than the lateral head surface and shows a low and high frequency peak in motion.

146 Figure 6.1

CQ T 3

_o > I I r \ r >

g

5dB

100 1000 10000 Frequency (Hz)

147 Figure 6.2 Comparison of auditory midbrain thresholds when the body wall is covered (dashed line) and uncovered (solid line) by a layer of silicone grease. The lines represent the average of responses from five individuals, and the vertical bars represent the standard error of the mean. Covering the body wall dramatically increased hearing thresholds.

1 4 8 Figure 6.2

100

90

80

£

60 -

50 100 1000 Frequency

149 was blocked first showed a dramatic increase in sensitivity after removal of the

silicone grease layer). On average, thresholds rose by at least 20-25 dB

following covering of the body wall (Fig. 6.2).

Effects of Lung Filling on Auditoiy Sensitivity .—Filling the lungs with saline

also had a dramatic effect on hearing sensitivity (Figs. 6.3 & 6.4). Hearing

sensitivity decreased across the entire range of frequencies tested. There

appeared to be a treatment order effect. The difference between thresholds

with filled and unfilled lungs was most pronounced in those animals in which

control thresholds were measured first and the lungs subsequently filled (Fig.

6.3). Thresholds after filling the lungs were elevated bv about 25-30 dB at each frequency measured. The difference between thresholds with filled and unfilled lungs was less pronounced, but still apparent, when midbrain responses were measured first with filled lungs and then again after the lungs were emptied (Fig. 6.4). Thresholds with the lungs filled were about 10-15 dB higher than those with the lungs empty. Midbrain thresholds taken from animals after the lungs had been emptied of saline (solid lines. Fig. 6.3) were elevated, by as much as about 15 dB at a given frequency, compared to thresholds measured before filling of the lungs (solid lines. Fig. 6.4). The

150 Figure 6.3 Comparison of auditory midbrain thresholds when the lungs are filled with air (solid lines ) or with oxygenated saline (dashed lines). Data from three individuals in which thresholds were measured first with the lungs filled with air and subsequently with the lungs filled with oxygenated saline. Filling the lungs with saline dramatically elevated auditory thresholds. No midbrain responses could be recorded above 300 Hz in frogs with lungs filled with saline.

151 Figure 6.3

100

90

■a 0 1

50 — 1000 100 Frequency (kHz)

152 Figure 6.4 Comparison of midbrain hearing thresholds when the lungs are filled with air (solid lines ) or with oxygenated saline (dashed lines). Data from two individuals in which thresholds were measured first with the lungs filled with oxygenated saline and subsequently with the lungs emptied of saline. Auditory thresholds were elevated in animals with lungs filled with saline. The effect of saline-filled lungs on hearing sensitivity in these two animals was not as great as that seen in Figure 6.3, probably because the lungs could not be completely emptied of saline (see text).

153 Figure 6.4

100

90

0. '■ n CQ 80 2 . / 2 o 70

H 60

50 100 1000 Frequency (Hz)

154 treatment order effect likely is related to a failure to completely empty the

lungs of saline after filling them (see Discussion).

Effects of Glottal Obstruction on Hearing Sensitivity .—Suturing the glottis

had little effect on midbrain thresholds (Fig. 6.5). Hearing sensitivity to low

frequencies below about 300 Hz was slightly higher (by no more than 3 dB)

when the glottis was closed, and hearing sensitivity to higher frequencies above

300 Hz was slightly higher (by no more than 5 dB) when the glottis was open.

DISCUSSION

The experiments described in this paper strongly support the hypothesis

that the lungs are the major route of sound reception in frogs that lack a

tympanic middle ear. Covering the body wall with a layer of silicone grease,

while keeping the head uncovered, dramatically decreased hearing sensitivity

across all sound frequencies. Filling the lungs with saline had the same effect,

and this latter test more precisely pinpoints the lungs, and not the body as a whole, as the important route of sound reception. The tests involving filling the lungs with saline found that the difference in hearing thresholds between

155 Figure 6.5 Comparison of midbrain hearing thresholds when the glottis is open (solid line) and sutured shut (dashed line). Data from five individuals. Suturing the glottis produced no significant effect on hearing thresholds.

156 Figure 6.5

100

90 -

G,

CQ 80 2 o s: 70 H 60

50 100 1000 Frequency (Hz)

157 the filled and unfilled condition was less, but still significant, when midbrain

responses were measured first from frogs with the lungs filled and subsequently

again when the lungs had been emptied. This less pronounced difference may

have been a result of a failure to completely remove saline from the lungs of

the frogs and/or insufficient inflation of the lungs after removal of the saline.

The midbrain thresholds determined after the lungs were emptied were

elevated by as much as 15 dB compared to the thresholds measured in animals

with normal, air-filled lungs (Figs. 6.3 &. 6.4). This suggests that the effect

caused by filling the lungs remained after the attempt to empty them, most

likely because of the two possible factors mentioned above. Regardless, the

results of the tests with air-filled and saline-filled lungs clearly demonstrate that the lung is important in hearing in these frogs.

Laser Doppler vibrometry found that the body wall over the lungs is indeed quite responsive to sound (Fig. 6.1), much more so than the lateral head tissues over the inner ear region. The body wall showed peaks in responsiveness at low frequencies below about 300 Hz and also at higher frequencies near 1000 Hz. Interestingly, the high frequency peak is not reflected in the threshold hearing curve of this species. As can be seen in the control (uncovered body wall) sensitivity curves in Figure 6.2, hearing sensitivity is best at ver\' low frequencies below about 200 Hz and falls off

158 with increasing frequency. There was evidence of a slight increase in

sensitivity in some animals around 1000 Hz, but no conspicuous shift

comparable to the peak in responsiveness found in the vibrometric

measurements. Although the high frequency peak in lung body wall motion

was not well reflected in midbrain sensitivities, the lung was still responsible

for auditory responses at those frequencies as covering the body wall or filling

the lung with saline dramatically decreased hearing sensitivity. Possibly the

resonations of the lung at 1000 Hz or above are not effectively transferred to

the inner ear. This would match the findings of Ehret et al. (1994) that high

frequency signals passing from the lungs of various species of anurans are

greatly attenuated as they are transmitted into the middle ear cavity.

Closure of the glottis had a minimal effect on midbrain responses to

sound, suggesting that sound energy from the lungs passes to the inner ear

along some route other than through the glottis and mouth cavity. At very low

frequencies, thresholds were slightly higher (by no more than 3 dB) when the glottis was sutured closed, and at higher frequencies thresholds were slightly

higher (by no more than 5 dB) when the glottis was open. These small differences probably have minimal functional significance. This finding is especially interesting because other workers studying frogs with well-developed tympanic middle ears have demonstrated that sound signals from the lungs can

159 be transmitted through the glottis and into the middle ear cavity and thereby

modify tympanic motion (Ehret et al. 1990; Jorgensen, 1991;). Also,

Christensen-Dalsgaard and Elepfandt ( 1995) found that closing the glottis of

Xenopus laevis significantly decreased auditory responses to low frequency

underwater sound, presumably preventing sound energy absorbed by the lungs

from reaching and acting on elements of the tympanic middle ear of this

species. Given these previous findings, it seemed reasonable to hypothesize

that sound energy from the lungs could pass through the glottis and enter the

mouth cavity of “earless” frogs, eventually reaching the inner ear through the

soft tissue layers overlying the round window or through the oval window.

However, the results of the tests involving suturing of the glottis refute this

hypothesis. The actual route of sound transfer from the lungs to the inner ear

remains to be determined. As discussed in the Introduction (to this chapter),

other routes have been proposed, such as transfer through the endolymphatic

sac that extends from the vertebral canal to the inner ear (Narins et al., 1988).

However, sound energy also could pass along several soft tissue pathways, as appears to be the case in several species of fishes that hear using an air bladder but have no clear structural connection between the air bladder and inner ear

(Popper and Coombs, 1982; Schellart and Popper, 1991). Determination of

160 the precise route of sound transfer in earless frogs therefore may provide a

challenge for further experimental work.

Many living fishes use air bladders for reception of underwater sound

(Schellart and Popper, 1991). Pulsations of the air bladders are transferred to

the inner ear either by structural elements or along unspecialized pathways. It

seems likely that early lunged aquatic vertebrates used lungs for transducing

underwater sound pressure energy. This study suggests that such a system can

still function for reception of airborne sound, and that some living tetrapods,

such as earless frogs, use their lungs for hearing. The lungs also have been

implicated in hearing in snakes, a group that also lacks a tympanic middle ear

(Hartline, 1971; Shulse et al., in prep). It has been proposed that

nontympanic hearing using the lungs is very size-dependent (Hetherington,

1992). Because of resonant properties related to lung size, the lung pathway of large tetrapods most likely would be effective only at very low frequencies, whereas that of small tetrapods could be effective at much higher frequencies.

This study, however, has found that in the earless B. orieutalis, a fairly small

frog species, sensitive hearing is restricted to low frequencies somewhat below the resonant frequency of the lung. In contrast, other earless species of frogs, such as those in the genus Atelopus, hear higher frequency sounds (around

2500 Hz) quite well (Jaslow and Lombard, 1996; Lindquist et al., in prep).

161 suggesting that the lung can be effective for sound reception at these high

frequencies. It therefore is not possible to categorize the lung pathway as only a low frequency channel in earless frogs.

The use of the lung for hearing air-bome sound raises the question of the adaptive significance of the tympanic middle ear. Lombard and Bolt

(1979) reported that the earliest tetrapods lacked a tympanic middle ear, and it seems reasonable to suggest that they could use the body wall and lung as a hearing pathway for at least low frequency sound. According to Lombard and

Bolt (1979) the tympanic middle ear appears to have evolved independently in amphibians, diapsid reptiles, and the mammalian lineage. It may be proposed that a tympanic middle ear generally would be more effective for reception of high frequency sound, but perhaps more importantly it may be more accurate for purposes of sound localization. It is unclear how earless tetrapods might localize sound, although there is evidence that earless frogs of the genus

Atelopus may be capable of doing so (Lindquist and Hetherington, 1996).

Experimental studies of sound localization in tetrapods with and without tympanic middle ears may provide insight into the selective importance of enhanced directionality via tympanic hearing.

162 Acknowledgments .—I wish to thank Susan Volman for the use of her neurophysiological equipment and laboratory. I also am grateful for the technical assistance of Thomas Hetherington, Susan Volman, Kris Schuett,

Mitch Masters, Andrew Hockman, and Jim Fox. These experiments comply with the “Principles of animal care,” publication No. 86-23, revised 1985 of the National Institutes of Health and also with the current laws of the State of

Ohio, USA.

BIBLIOGRAPHY

Capranica, R. R. 1976. Morphology and physiology of the auditory system. In, Frog Neurobiology. R. Uinas and W. Precht (eds). Springer-Verlag, Berlin, pp. 551-575.

Christensen-Dalsgaard, J. and A. Elepfandt. 1995. Biophysics of underwater hearing in the clawed frog, Xenopus Uievis. J. Comp. Physiol. A 176:317- 324.

Ehret, G., J. Tautz, and B. Schmitz. 1990. Hearing through the lungs: lung- eardrum transmission of sound in the frog Eleutherodactylus cocpd. Natunvissenschaften 77:192-194

Ehret, G., E. Keilwerth, and T. Kamada. 1994. The lung-eardrum pathway in three treefrog and four dendrobatid frog species: some properties of sound transmission. J. Exp. Biol. 195:329-343.

Hartline, P.M. 1971. Mid-brain responses of the auditory and somatic vibration systems in snakes. J. Exp. Biol. 54:373-390.

163 Hetherington, T. E. 1992. The effects of body size on functional properties of middle ear systems in anuran amphibians. Brain Behav. Evol. 39:133-142.

Jaslow, A. P., T. E. Hetherington, R. E. Lombard. 1988. Structure and function of the amphibian middle ear. In, B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, W. Walkowiak (eds). The Evolution of the Amphibian Auditory system. John W iley & Sons, New York. pp. 69- 92.

Jaslow, A. P. and R. E. Lombard. 1996. Hearing in the neotropical frog, Atelopus chiriquieiisis. Copeia 1996:428-432.

Jorgensen, M. B. 1991. Comparatiye studies of the biophysics of directional hearing in anurans. J. Comp. Physiol. A 169:591-598.

Jorgensen, M. B., B. Schmitz, J. Christensen-Dalsgaard. 1991. Biophysics of directional hearing in the frog Eleutherodactylus coipii. J. Comp. Physiol. A 168:223-232.

Lindquist, E. D. and T. E. Hetherington. 1996. Field studies on visual and acoustic signaling in the "earless" Panamanian golden frog, Atelopus zeteki. J. Herpetol. 30:347-354.

Lombard, R.E. and J. Bolt 1979. Evolution of the tetrapod ear: an analysis and reinterpretation. Biol. J. Linn. Soc. 11:19-76.

Narins, P. M., G. Ehret, and J. Tautz. 1988. Accessory pathway for sound transfer in a neotropical frog. Proc. Nat. Acad. Sci. USA. 85:1508-1512.

Popper, A. N. and S. Coombs. 1982. The morphology and evolution of the ear in Actinopterygian fishes. Amer. Zool. 22:311-328.

Schellart, N. A. M. and A. N. Popper. 1991. Functional aspects of the evolution of the auditory system of Actinopterygian fishes. In, The Evolutionary Biology of Hearing. D. B. Webster, R. R. Fay, and A. N. Popper (eds.). Springer-Verlag: New York, pp 295-322.

Volman, S. 1996. Qualitative assessment of song-selectivity in the zebra finch “high vocal center.” J. Comp. Physiol. A 105:1-8.

1 6 4 Wever, E. G. 1985. The Amphibian Ear. Princeton University Press, Princeton.

Wilczynski, W. and R. R. Capranica. 1984. The auditory system of anuran amphibians. Progress. Neurobioi. 22:1-38.

Witschi, E. 1949. The larval ear of the frog and its transformation during metamorphosis. Z. Naturf. Wiesbaden 4:230-242.

165 COMPLETE BIBLIOGRAPHY

Altig, R. & G. F. Johnston. 1989. Guilds of anuran larvae: Relationships among developmental modes, morphologies, and habitats. Herpetol. Monogr. 3 :81-109.

Brown, G. B., Y. H. Kim, H. Kuntzel, H. S. Mosher, G. J. Fuhrman, &. F. A. Fuhrman. 1977. Chemistry and pharmacology of skin toxins from the ïxog Atelopus zeteki (Atelopidtoxin: zetekitoxin). Toxicon 15:115-128.

Cannatella, D. C. 1986. A new genus of bufonid (Anura) from South America, and phvlogenetic relationships of the neotropical genera. Herpetologica 42:197-205.

Capranica, R. R. 1976. Morphology and physiology of the auditory system. In, Frog Neurobiology. R. Llinas and W. Precht (eds). Springer-Verlag, Berlin, pp. 551-575.

Christensen-Dalsgaard, J. and A. Elepfandt. 1995. Biophysics of underwater hearing in the clawed frog, Xaiopus laevis. J. Comp. Physiol. A 176:317- 324.

Clemmons, J. R. and M. M. Lambrechts. 1992. The waving display and other nest site anti-predator behavior of the Black-capped Chickadee. Wilson Bull. 104:749-756.

Cocroft, R. B., R. W. McDiamtid, A. P. Jaslow, &. P. M. Ruiz-Carranza. 1990. Vocalizations of eight species of Atelopus (Anura: Bufonidae) with comments on communication in the genus. Copeia 1990:631-643.

Coloma, L. A. &. S. Letters. The tadpole o( Atelopus balios (Anura: Bufonidae) from the Pacific lowlands of Ecuador. Herpetologica 52:66-70.

166 Crump, M. L. 1986. Homing and site fidelity in a neotropical frog, Atelopus varius (Bufonidae). Anim. Behav. 34:438-444.

Cmmp, M. L. 1988. Aggression in harlequin frogs: male-male competition and a possible conflict of interest between the sexes. Anim. Behav. 36:1064- 1077.

Daly, f. VV., S. I. Secunda, H. M. Garraffo, T. F. Spande, A. Wisnieski, &. J. F. Cover Jr. 1994. An uptake s\'stem for dietarv alkaloids in poison frogs (Dendrobatidae). Toxicon 32: 657-663.

Daly, J. W., W. L. Padgett, R. L Saunders, & J. F. Cover Jr. Absence of tetrodotoxins in a captive-raised riparian frog, Atelopus varius. Toxicon. In press.

Davison, G. 1984. Foot-flagging displavs in Bornean frogs. Sarawak Mus. J. 33(57): 177-178.

Dole, J. W. &. P. Durant. 1974. Movements and seasonal activity oî Atelopus o.vyrhytiehus (Anura: Atelopodidae) in a Venezuelan cloud forest. Copeia 1974:230-235.

DreuTy, G. E. 1970. The role of amphibians in the ecology of the Puerto Rican rainforest. In Clements, R.G., Drewry, G.E., and R.J. La vigne, (eds.). The Rainforest Project Annual Report - 1970.

Duellman, VV. E. 1978. The biology of an equatorial herpetofauna in Amazonian Ecuador. Misc. Publ. Mus. Nat. Hist. Kans. 65:1-352.

Duellman, W. E. and L. Trueb. 1985. Biology of Amphibians. MacGraw-Hill, Inc., New York. pp. ix-670.

Duellman, VV. E. and J. D. Lvmch. 1969. Descriptions o î Atelopus tadpoles and their relevance to atelopodid classification. Herpetologica: 25: 231-240.

Dunn, E. R. 1933. Amphibians and reptiles from El Valle de Anton, Panama. Occas. Pap. Boston Soc. Nat. Hist. 8:65-79.

167 Eggermont, J. J. 1988. Mechanisms of sound localization in anurans. In B. Fritzsch, M.J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkovviak (eds.). The Evolution of the Amphibian Auditory System, pp. 307-336. New York: John Wiley & Sons.

Ehret, G., J. Tautz, and B. Schmitz. 1990. Hearing through the lungs: lung- eardrum transmission of sound in the frog Eleutherodactylus coqui. Naturwissenschaften, 77: 192-194.

Ehret, G., E. Keilwerth, and T. Kamada, 1994. The lung-eardrum pathway in three treefrog and four dendrobatid frog species: some properties of sound transmission. J. Exp. Biol. 195:329-343.

Feng, A. S. .1981. Directional response characteristics of single neurons in the torus semicircularis of the Leopard frog {Rami pipiais). J. Comp. Physiol. 144:419-428.

Feng, A. S., P. M. Narins, and R. R. Capranica. 1975. Three populations of primary auditory fibers in the bullfrog (Rana catesbiana): Their peripheral origins and frequency sensitivities. J. Comp. Phvsiol. 100:221-229.

Feng, A. S. and R. R. Capranica. 1976. Sound localization in anurans I. Evidence of binaural interaction in the dorsal medullary nucleus of bullfrogs {Rana catesbiana). J. Neurophysiol. 39:871-881.

Feng, A. S. and W. P. Shofner 1981. Peripheral basis of sound localization in anurans, acoustic properties of the frog’s ear. Hearing Res. 5:201-216.

Fuhrman, F. A., G. Fuhrman, &. H. S. Mosher. 1969. Toxin from the skins of frogs of the genus Atelopus: Differentiation from dendrobatid toxins. Science 165:1376-1377.

Gascon, C. 1989. The tadpole of Atelopus pulcher Boulenger (Anura, Bufonidae) from Manaus, Amazonas. Revta. Bras. Zool. 6:235-239.

Gerhardt, H. C. 1982. Sound pattern recognition in some North American tree frogs (Anura: Hvlidae): implications for mate choice. Am. Zool. 22:581- 595.

1 6 8 Godfray, H. C. J. 1995. Signaling of need between parents and young: Parent- offspring conflict and sibling rivalry. Am. Nat. 146:1 -24.

Gosner, K. L. 1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16:183-190.

Goodwin, T. W. 1984. The Biochemistry of the Carotenoids (2"*^ ed.). Chapman and Hall, NY. pp. 154-157.

Gray, P. &. D. C. Cannatella. 1985. A new species oi Atelopus (Anura: Bufonidae) from the Andes of northern Peru. Copeia 1985: 910-917.

Harding, K. A. 1982. Courtship display in a Bornean frog. Proc. Biol. Soc. of Wash., 95:621-624.

Hartline, P.H. 1971. Mid-brain responses of the auditory and somatic vibration systems in snakes. J. Exp. Biol. 54:373-390.

Hasson, O. 1997. Towards a general theorv of biological signaling. J. Theor. Biol. 185:139-156.

Hauser, M. D. 1996. The Evolution of Communication. Cambridge: Massachusetts Institute of Technology Press.

Hetherington, T. E. 1991. The effects of body size on the evolution of the amphibian middle ear. In A. Popper, D. Webster, R. Fay (eds). The evolutionarv biology of hearing. Springer-Verlag, New York pp. 421- 437.

Hetherington, T.E. 1992. The effects of body size on functional properties of middle ear svstems of anuran amphibians. Brain Behav. Evol. 39: 133- 142.

Heyer, W. R., A. S. Rand, C. A. G. Da Cruz, O. L. Peixoto, and C. E. Nelson 1990. Frogs of Boracea. Arq. Zool. 31:231-410.

Hodl, W., and Gollman, G. 1986. Distress calls in neotropical frogs. Amph. Rept. 7:11-21.

169 Hyatt, G. W. 1977. Field studies of size-dependent changes in waving display and other behavior in the fiddler crab, Uca pugilator (Brachyura, Ocypodidae) Mar. Behav. Physiol. 4:283-292.

Ibanez, R. D., C. A. Jamarillo, Y. F. A. Solis. 1995. Una especies nueva de Atelopus (Amphibia: Bufonidae) de Panama. Carrib. J. Sci. 31:57-64.

Jaslow, A. P. 1979. Vocalization and aggression in Atelopus cliiriquiensis (Amphibia, Anura, Bufonidae). J. Herpetol. 13:141-145.

Jaslow, A. P., and R. E. Lombard. 1996. Hearing in the Neotropical Frog, Atelopus chiricpiiensis. Copeia 1996:428-432.

Jaslow, A. P., T. E. Hetherington, and R. E. Lombard. 1988. Structure and function of the amphibian middle ear. In B. Fritzsch, M.J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak (eds.). The Evolution of the Amphibian Auditory Svstem, pp. 69-92. New York: John Wiley &. Sons.

Jorgensen, M. B. 1991. Comparative studies of the biophvsics of directional hearing in anurans. J. Comp. Physiol. A 169:591-598.

Jorgensen, M. B., B. Schmitz, J. Christensen-Dalsgaard 1991. Biophysics of directional hearing in the frog Eleutherodactylus coqui. J. Comp. Phvsiol. A 168:223-232

Kroodsma D. E. 1989. Suggested experimental designs for song plavbacks. Anim. Behav. 37:600-690.

Kroodsma D. E. 1990. Using appropriate experimental designs for intended hypotheses in 'song' playbacks, with examples for testing effects of song repertoire sizes. Anim. Behav. 40:1138-1150.

La villa, E. O., R. O. de Sa, and 1. De la Riva. 1997. Description of the tadpole of Atelopus tricolor. J. Herpetol. 31:121-124.

Lescure, J. 1973. Contribtion al'etude des Amphibiens de Guyane française. 1. Notes sur AtelopusJlavescens Dumeril et Bibron et description d'une nouvelle espece. Vie Milieu 23:125-141.

170 Lescure, J. 198 la. Contribution à l’étude des Amphibiens de Guyane française IX. Le têtard gastromyzophore d ’Atelopusflavescens Duméril et Bibron (Anura, Bufonidae). Àmphibia-Reptilia 2: 209-215.

Lescure, J. 1981b. Contribution à l’étude des Amphibiens de Guyane française. VIII. Validation à.'Atelopus spumarius Cope, 1871, et désignation d’un néotype. Description dAtelopus spumarius barbotini nov. sp. Données étho-écologiques et biogéographiques sur les Atelopus du groupe flavescens (Anoures, Bufonidés). Bull. Mus. Natn. Hist. Nat. Paris 3:893-910

Lindquist, E. D. 1995. Atelopus zeteki: Pure tonal vocalization. Herpetol. Rev. 26:200-201.

Lindquist, E. D. and T. E. Hetherington. 1996. Field studies on visual and acoustic signaling in the “earless Panamanian golden frog, Atelopus zeteki. J. Herpetol. 30:347-354.

Lombard, R E. and f. Bolt 1979. Evolution of the tetrapod ear: an analysis and reinterpretation. Biol. J. Linn. Soc. 11:19-76.

Lotters, S. 1996. The Neotropical Toad C tm .s Atelopus: Checklist-Biology- Distribution. M. Vences and F. Claw Verlags. Koln, Gennany.

Lynch, J. D. 1986. Notes on the reproductive hicAo^ o(Atelopus suboniatus. J. Herpetol. 20:126-129.

Marier, P. 1977. The evolution of communication. In T. A. Sebeok (ed). How Animals Communicate, pp. 45-70. Bloomington & London: Indiana Univ. Press.

McDiamtid, R. W. 1971. Comparative morphology and evolution of frogs of the genus Atelopus, Dendrophryuisais, Melanophrynisais, and Oreopluynella. Bull. Los Ang. Co. Mus. Nat. Hist. 12:1-66.

Mebs, D. 1980. Zur fortpflanzung von Atf/opMS cruciger (Amphibia: Salientia: Bufonidae. Salamandra 16: 65-81.

Miller, T. 1987. Notes on Central Amtncsax Atelopus. The Herpetoculturist 1:25-28.

171 Moffat, A. J. M„ and R. R. Capranica 1976. Auditory sensitivity of the saccule in the American toad {Biifo americaniis). J. Comp. Physiol. A 105:1-8.

Myers, C. W. 1969. The ecological geography of cloud forest in Panama. Amer.Mus. Nov. 2396:1-52.

Myers, C. W., J. W. Daly, and B. Malkin. 1978. A dangerously toxic new frog {Phyllobates) used by Embera indians of western Colombia, with discussion of blowgun fabrication and dart poisoning. Bull. Am. Mus. Nat. Hist. 161(2):309-365.

Narins, P. M., G. Ehret, and J. Tautz 1988. Accessory pathway for sound transfer in a neotropical frog. Proc. Nat. Acad. Sci. USÀ. 85:1508-1512.

Norusis, M. J. 1993. SPSS for Windows™, Rel. 6.0. Noru§is/SPSS Inc.

Orton, G.L. 1953. The systematics of vertebrate larvae. Syst. Zool. 2:63-75.

Pombal, J. P. Jr., I. Sazima, &.C. F. B. Haddad. 1994. Breeding behavior of the Pumpkin Toadlet, Bracliyceplialus epiiippium (Brachycephalidae). J. Herpetol. 28:516-519.

Popper, A. N. and S. Coombs. 1982. The morphology and evolution of the ear in Actinopterygian fishes. Amer. Zool. 22:311-328.

Pounds, A. J. 1994. Amphibian declines climate disturbance: the case of the golden toad and the harlequin frog. Cons. Biol. 8( l):72-85.

Rand, A. S. 1988. An overview of anuran acoustic communication. In Fritsch, B., Ryan, M., Wilczynski, W., Hetherington, T.E. and Walkowiak, W., (eds.). The Evolution of the Amphibian Auditory System, pp. 69-91. New York: John Wiley &. Sons.

Richards, S. J. &.C. James. 1992. Foot-flagging displays of some Australian frogs. Mem. Qd. Mus. 32:302.

Roy, D., B. Bijoylakshmi, and A. Sarma. 1995. Analysis and significance of female reciprocal call in frogs. Curr. Sci. 69:265-270.

172 Ryan, M. J. 1985. The Tungara Frog: a Study of Sexual Selection and Communication. Chicago: University of Chicago Press, pp. Lx-230.

Ryan, M. J. 1988. Constraints and patterns in the evolution of anuran acoustic communication. In Fritsch, B., Ryan, M., Wilczynski, W., Hetherington, T.E. and Walkowiak, W., (eds.). The Evolution of the Amphibian Auditory System, pp. 637-677. New York: John Wiley & Sons.

Savage, J. M. 1972. The harlequin frogs, genus Atelopus, of Costa Rica and western Panama, Herpetologica 28:77-94.

Schellart, N. A. M. and A. N. Popper. 1991. Functional aspects of the evolution of the auditory system of Actinopterygian fishes. In, The Evolutionary Biology of Hearing. D. B. Webster, R. R. Fay, and A. N. Popper (eds.). New York: Springer-Verlag. pp 295-322.

Smith, S. M. 1996. The single wing-flick display of the Black-capped Chickadee. Condor 98:885-887.

Smith, W. J. 1977. The Behavior of Communicating: An Ethological Approach, Harvard University Press, Cambridge.

Starrett, P. 1967. Observations on the life history of frogs of the familv Atelopodidae. Herpetologica 23: 195-204.

Stewart, M.M. and A.S. Rand. 1991. Vocalizations and the defense of retreat sites bv male and female frogs, Eleutherodactylus coqui. Copeia 1991(4):1013-1024.

Townsend, D. S., Stewart, M. M., and F. H. Pough. 1984. Male parental care and its adaptive significance in a neotropical frog. Anim. Behav. 32:421- 431.

Volman, S. 1996. Qualitative assessment of song-selectivity in the zebra finch “high vocal center.” J. Comp. Physiol. A 178:849-862.

Walkowiak, W. 1980. The coding of auditory signals in the Torus semicircularis of the Fire-bellied Toad and the Grass Frog: Responses to simple stimuli and to conspecific calls. J. Comp. Phys. A 138:131-148.

173 Wells, K. D. 1977a. The courtship of frogs, In D.H. Taylor and S. I. Guttman (eds.). The Reproductive Biology of Amphibians. New York: Plenum Press, pp. 233-262.

Wells, K. D. 1977b. The social behavior of anuran amphibians. Anim. Behav. 25:666-693.

Wells, K. D. 1988. The effect of social interactions on anuran vocal behavior. In B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak (eds.). The Evolution of the Amphibian Auditory System, pp. 433-454. New York: John Wiley &. Sons.

Wever, E. G. 1985. The Amphibian Ear. Princeton: Princeton Univ. Press.

Wevers, S. E. 1988. Enige opmerkingen over de pijlgifkikker Otr/ir/wZjrtto parvuliis. Lacerta 46:51-53.

Wilczynski, W., and R. R. Capranica. 1984. The auditory system of anuran amphibians. Progress. Neurobioi. 22:1-38.

Wilczynski, W., C. Resler, and R. R. Capranica. 1987. Tympanic and extratvmpanic sound transmission to the leopard frog. J. Comp. Phvsiol. 161:659-669.

Winter, J. &. K. R. Me Donald. 1986. Eungella: The land of the cloud. Aust. Nat. Hist. 22:39-43.

Witschi, E. 1949. The larval ear of the frog and its transfomtation during metamorphosis. Z. Naturf. Wiesbaden 4:230-242.

Zakon, H. H., and W. Wilczynski. 1988. The physiology of the anuran eighth nerve. In B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington, and W. Walkowiak (eds), The evolution of the amphibian auditorv system. New York: John Wiley &. Sons. pp. 125-155.

Zucker, N. 1983. Courtship variation in the Neo-tropical fiddler crab, Uca ddchmanni: Another example of female incitation to male competition? Mar. Behav. Phvsiol. 10:57-79.

174