1 FROM GENES TO COMMUNITIES: INTEGRATIVE BIOLOGY OF ACOUSTIC INTERACTIONS IN NEOTROPICAL SINGING MICE By BRET PASCH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011
2 2011 Bret Pasch
3 To Mary and Eugnie -wise and adventurous grandmothers
4 ACKNOWLEDGMENTS I thank my wife, Kate, for her love, patience, and encouragement; our n ew daughter, Mary, for her vibrance; and my parents and brother for their enduring support. My advisor, Steven Phelps, inspired mechanistic approaches to animal behavior while allowing me the creative space to pursue my interests. His vision and guidance h ave been invaluable. Comfmittee members Jim Austin, Doug Levey, Brian McNab, and David Reed provided important perspective and insight as my dissertation evolved. I am grateful to my fellow lab members, including Dimitri Blondel, Polly Campbell, Ondi Crino Alex Ophir, and Jorge Pino for their constructive feedback and camaraderie I enjoyed working alongside three undergraduate students, Andreas George, Molly Phillips, and Rachel Sanford. Their passion and inquisitiveness stirred my own and enriched the re search experience I thank Ben Bolker and Craig Osenberg for their advice on experimental design and data analysis, Harvey Ramirez and Dania del Castillo for their guidance on surgical technique, and Russ Charif and Dean Hawthorne at the Bioacoustics Resea rch Program, Cornell Lab of Ornithology for their advice on sound analysis. Finally, a warm thanks to F. Michonneau, J. Monaghan, A. Seifert, M.D. Smith, and A. Stier, for their friendship and collaborations. Research was conducted with permits from UF IAC UC, the Costa Rican Ministerio del Ambiente y Energia, and the Panamanian Autoridad Nacional del Ambiente. The research was supported by grants from Sigma Xi, Am erican Society of Mammalogists, American Museum of Natural History, Society for Integrative and Comparative Biology, Organization for Tropical Studies, University of Florida Alumni Fellowship, and N ational S cience F oundation D octoral Dissertation Improvement Grant # 0909769
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF OBJECTS ................................ ................................ ................................ ......... 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCT ION ................................ ................................ ................................ .... 12 2 ANDROGENS MODULATE SONG EFFORT AND AGGRESSION IN NEOTROPICAL SINGING MICE ................................ ................................ ............ 14 Background ................................ ................................ ................................ ............. 14 Materials and M ethods ................................ ................................ ............................ 16 Study Animals ................................ ................................ ................................ .. 16 Surgical Procedure ................................ ................................ ........................... 16 Song Recording and Measurement ................................ ................................ .. 17 Resident Intruder Trial s ................................ ................................ .................... 18 Testosterone Radioimmunoassay ................................ ................................ .... 18 Data Analysis ................................ ................................ ................................ ... 19 Results ................................ ................................ ................................ .................... 20 Plasma Androgen Concentrations ................................ ................................ .... 20 Resident Intruder Trial s ................................ ................................ .................... 20 Singing Behavior ................................ ................................ .............................. 20 Song Parameters ................................ ................................ ............................. 20 Discussion ................................ ................................ ................................ .............. 22 3 ANDROGEN DEPENDENT MALE VOCAL PERFORMANCE I NFLUENCES FEMALE PREFERENCE IN NEOTROPICAL SINGING MICE ............................... 34 Background ................................ ................................ ................................ ............. 34 Materials and Methods ................................ ................................ ............................ 37 Generating the Population Performance Limit ................................ .................. 37 Androgen Manipulation ................................ ................................ .................... 38 Female Phonotaxis ................................ ................................ ........................... 39 Statistical Analyses ................................ ................................ .......................... 41 Results ................................ ................................ ................................ .................... 43 Population Performance Limit ................................ ................................ .......... 43 Androgenic Effects on Performance Scores ................................ ..................... 43
6 Female Phonotaxis ................................ ................................ ........................... 44 Discussion ................................ ................................ ................................ .............. 44 4 INTERSPECIFIC DOMINANCE VIA VOCAL INTERACTIONS MEDIATES ALTITUDINAL ZONATION IN NEOTROPICAL SINGING MICE ............................ 53 Background ................................ ................................ ................................ ............. 53 Materials and Methods ................................ ................................ ............................ 56 Altitudinal Replacement Zones and Temperature Regimes ............................. 56 Interspecific Behavioral Trials ................................ ................................ ........... 57 Reciprocal Removal Ex periment ................................ ................................ ...... 57 Field Playback Experiments ................................ ................................ ............. 58 Laboratory Playback Experiments ................................ ................................ .... 59 Statistical Analyses ................................ ................................ .......................... 60 Results ................................ ................................ ................................ .................... 61 Altitudinal Replacement Zones ................................ ................................ ......... 61 Interspecific Behavioral Trials ................................ ................................ ........... 62 Reciprocal Removal Experiment ................................ ................................ ...... 62 Field Playback Experiments ................................ ................................ ............. 63 Laboratory Playback Experim ents ................................ ................................ .... 63 Discussion ................................ ................................ ................................ .............. 64 Role of Acoustic Communication in Interspecific Interactions .......................... 68 Conclusion ................................ ................................ ................................ ........ 69 5 AGONISTIC CHARACTER DISPLACEMENT IN SOCIAL COGNITION IN A TROPICAL MOUSE ................................ ................................ ................................ 77 Background ................................ ................................ ................................ ............. 77 Materials and Methods ................................ ................................ ............................ 79 Reciprocal Playback Experiments ................................ ................................ .... 79 Tissue Preparation ................................ ................................ ........................... 80 Egr 1 Immunocytochemistry ................................ ................................ ............. 81 Analysis of Egr 1 Immunoreactivity ................................ ................................ .. 81 Statistical Analyses ................................ ................................ .......................... 82 Results ................................ ................................ ................................ .................... 82 Discussion ................................ ................................ ................................ .............. 83 6 CONCLUSIONS ................................ ................................ ................................ ..... 87 LIST OF REFERENCES ................................ ................................ ............................... 92 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 109
7 LIST OF TABLES Table page 2 1 Results of paired t tests of acoustic variables for male singing mice post pre treatment.. ................................ ................................ ................................ .......... 26
8 LIST OF FIGURES Figure page 2 1 S. teguina ). ................................ ............... 27 2 2 Androgenic effects on aggressive behavior. Number of submissive supine displays by resident male singing mice to intru ding male consp ecifics pre and post treatment ................................ ................................ ............................. 28 2 3 Androgenic effects on singing behavio r ................................ .............................. 29 2 4 Mouse song pre and post ca stration ................................ ................................ 30 2 5 Androgenic effects on song parameters ................................ ............................. 31 2 6 Relationship between androgen concentrations and song parameters i n singing mice. ................................ ................................ ................................ ....... 32 3 1 Design of female preference experiment. ................................ ........................... 49 3 2 Vocal performance in male singing mice. ................................ ........................... 50 3 3 Female response to experimentally manipulated male trills. .............................. 51 4 1 Latitudinal and altitudinal distributions of Neotropical singing mice .................... 70 4 2 Altitudinal distributions and temperature regimes of Neotropical singing mice ... 71 4 3 Interspecific aggression in Neotropic al singing mice ................................ .......... 72 4 4 Reciprocal removal experiment on Volcn Iraz, Costa Rica ............................. 73 4 5 Response to playback trials in the fie ld at Volcn Iraz, Costa Rica .................. 74 4 6 Response to playback trials in the laboratory at Volcn Iraz, Costa Rica ......... 75 5 1 Altitudinal distributions of Neotropical singing mice at our study site s in Costa Rica ................................ ................................ ................................ .................... 84 5 2 Response to playback trials in the laboratory at Volcn Iraz (sympatry) and Cerros La Carpintera (allopatry) Costa Rica ................................ ..................... 85 5 3 Neuronal immediate early gene expression in response to playbacks in the laboratory at Volcn Iraz (sympatry) and Cerros La Carpintera (allopatry), Costa Rica ................................ ................................ ................................ .......... 86
9 LIST OF OBJECTS Object page 2 1 ................................ ....................... 33 2 2 ................................ .... 33 2 3 ................................ ........ 33 3 1 singing mouse song ................................ .... 52 4 1 to playback of conspecific song ................................ ................................ ................................ .................... 7 6 4 2 Song of a male Chiriqu singing mouse ................................ .............................. 76
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for th e Degree of Doctor of Philosophy FROM GENES TO COMMUNITIES: INTEGRATIVE BIOLOGY OF ACOUSTIC INTERACTIONS IN NEOTROPICAL SINGING MICE By Bret Pasch December 2011 Chair: Steven M. Phelps Major: Zoology My dis sertation addressed proximate and ultimate fac tors governing acoustic communication in Neotropical singing mice ( Scotinomys ). A major goal of the work was to understand how interspecific communication contributed to the abrupt altitudinal replacement of singing mice in Costa Rica and Panam I integra ted data from biogeographical surveys, long term environmental monitoring, reciprocal removal experiments, staged behavioral trials, reciprocal pla yback experiments hormonal manipulation, and analysis of neuronal immediate early gene expression. I found t hat the production of mouse trills is constrained by a performance trade off between how fast notes are repeated and the frequency bandwidth of each note Males vary in vocal performance, and females prefer males that are better performers. I also found th at vocalizations and associated aggression are modulated by androgens, an important class of steroid hormones that mediate the expression of male secondary sex characters. These data suggest that vocalizations are used to attract mates and repel rivals and provided an important foundation to address the role of song in mediating interspecific interactions. I found that the larger species, S. xerampelinus is behaviorally dominant and excludes S. teguina from higher altitudes. P layback
11 experiments demonstrat ed that S. teguina is silenced and repelled by heterospecific song, whereas S. xerampelinus responded to heterospecifics. Thus, response to song reflects underlying dominance and suggests that interspecific communication mediates altitudinal zonation of ec ologically similar congeners. To investigate the neural mechanisms involved in species recognition, I studied brains of male S teguina from two populations (allopatric and sympatric with S. xerampelinus ) that were broadcast heterospecific vocalizations. I n contrast to sympatric males, allopatric animals responded to heterospecifics Both populations exhibited significant induction of egr 1 in the auditory cortex and medial geniculate in response to conspecifics when compared to control stimuli. However he terospecific songs elicited stronger responses in sympatry than in allopatry, despite the fact that these songs elicit ed less singing behavior in sympatry. I suggest that differences in aggressive experience with heterospecifics can modify auditory tuning to match ecologically important stimuli. The integrative nature of the study elucidates link s between changes in perception at the neural level and changes in distribution at the geographic level.
12 CHAPTER 1 INTRODUCTION Many animals use long distance acoustic signals to advertise their presence t o a network of potential mates and competitors. A rich tradition of studies on acoustic communication in birds, frogs, and insects has led to important insight into disparate disciplines of biology, including ecology and evolution, behavioral endocrinology, and neuroscience (Kroodsma and Miller 1996; Gerhardt and Huber 2002 ). These advances have necessitated integration of proximate and ultimate factors that govern acoustic signaling. In my dissertation, I aim to synthesize the proximate and ultimate factors mediating acoustic communication in Neotropical singing mice. While features of acoustic communication have been well described in laboratory rodents (Brudzynski 2010), this work attempts to place advertisem ent signaling in a broader ecological and evolutionary context. Chapter 2 assesses the role of androgens an important class of steroid hormones that mediate the expression of male secondary sex characters in modulating the production of advertisement voc alizations. I manipulated androgen leve ls in male mice via castration and hormone replacement and measured their aggressive and singing behavior pre and post treatment I discuss the role of androgens in linking gonadal status with decisions about investm ent in reproductive behaviors. Chapter 3 explores the influence of androgens on mediating individual variation in male advertisement vocalizations, and the response of females to that variation. Vocalizations from males captured and recorded at different f ield sites throughout Costa Rica and Panam were analyzed to assess within vs. between individual variation. Vocalizations were then compared to those produced by laboratory animals whose
13 circulating levels of androgens were experimentally manipulated. Th e trills were then manipulated and presented to females to determine their response. I discuss the influence of a ndrogens on modulating the production of physically challenging trills, and how signal variation influences female preference Chapter 4 invest igates the role of acoustic communication in influencing singing mouse geographic limits I use b iogeographical surveys, reciprocal removal experiments, interspecific behavioral trials, and reciprocal playback experiments to determine the biotic and abioti c factors responsible for the abrupt altitudinal replacement of S. teguina by S. xerampelinus I discuss how acoustic communication reflects underlying dominance to mediate altitudinal zonation between ecologically similar congeners. Finally, Chapter 5 inv estigates the neuronal mechanisms involved in perception of auditory signals from an ecological perspective. Playback experiments were performed on male S. teguina sympatric with dominant heterospecifics and in allopatry. I examined song evoked induction of an immediate early gene in two auditory regions of the brain. I discuss the pattern of character displacement and how agonistic experience with heterospecifics can modify auditory tuning to match ecologically important stimuli.
14 CHAPTER 2 ANDROGENS MO DULATE SONG EFFORT AND AGGRESSION IN NEOTROPICAL SINGING MICE 1 Background Androgens are critically involved in the expression and evolution of seconda ry sex characters (Adkins Regan 2005), including vocalizations produced in the context of aggression and mat e attraction (Ketterson et al. 1996; Ball et al. 2003; Moore et al., 2005; Wilczynski et al. 2005). In birds, frogs, and fish, androgens act on neural pathways and peripheral musculature to influence vocalization rate, signal d uration (Bass and Remage H ealey 2008), amplitude ( Connaughton et al. 1997), and d ominant frequency (Beani et al. 1995; Cynx et al. 2005). In turn, males that vocalize often and produce longer, louder signals with lower frequencies are often perceived as greater threats by rival mal es and preferred by females (Klump and Gerhardt 1987; Clayton and Prove 1989; Ryan and Keddy Hector 1992; Reby et al. 2005; Apicella et al. 2007; Mager et al. 2007; Wyman et al. 2008) Androgens thus play a key role in modulating information content of voc al signals, and their manipulation enables exploration into how hormonal mechanisms influence behavior. In rodents, androgens modulate the rate of ultrasonic vocalizations (USV) produced in reproductiv e contexts (Sales 1972; Floody 1981; Nyby et al. 1992; Yamaguc hi and Kelly 2002; James et al. 2006; Hammerschmidt et al. 2009). However, examples of androgenic effects on other attributes of mammalian vocalizations are 18 th century 1 Reprin ted with permission from Hormones and Behavior. Pasch, B., George, A.S., Hamlin, H.J., Guillette, L.J., and S.M. Phelps. 2011b. Androgens modulate song effort and aggr ession in Neotropical singing mice. Hormones and Behavior 59: 90 97
15 (Jenkins 1 998; Nelson 2005) and adult men administered te stosterone therapy (King et al. 2001; Akcam et al. 2004; Hartgens and Kuipers 2004). While dominant frequency and amplitude features are correlated with fighting ability and female mate choice in some mammals (Reby et al. 2005; Wyman et al. 2008), the role of androgens in controlling such vocal attributes is poorly known. Herein, we investigate androgenic effects on vocalizations of Neotropical singing mice (genus Scotinomys ), diurnal insectivorous rodents dis tributed throughout Central American cloud forests ( Wilson and Reeder 2005 ). Singing mice provide a tractable mammalian model because adult males commonly emit loud, frequency modulated trills that span wide bandwidths (10 43 kHz; Hooper and Carleton 1976; Miller and Engstrom 2007; Campbell et al. 2010; Figure 2 1; Object 2 1 ) features characteristic of advertisement signals used to maximize transmission distance and facilitate localization (Bradbury and Vehrencamp 1998). Presentation and subsequent remova l of females increases male singing rate ( Fernandez Vargas et al. 2008), and males countersing in response to broadcast songs of intruding males in the field and laboratory ( Chapter 4). Thus, singing mouse trills appear to function similarly to frog and bi rd trills used in male male aggressi on and mate attraction (Emerson 2001; Collins 2004). In this study, we tested the hypothesis that androgens modulate vocal behavior, song S. teguina ) by experim entally manipulating circulating androgen concentrations. We predicted that males receiving no androgen replacement following castration would sing fewer songs and be less aggressive, whereas androgen replacement would maintain or increase song effort and
16 aggression. Lastly, we examined whether individual differences in androgen titers predicted individual differences in behavior. Materials and Methods Study A nimals Experimental animals used in the study were lab reared offspring (F1) of wild captured mice from Cerro La Carpintera, Costa Rica. Sexually experienced adult males were assigned randomly to treatment groups and housed individually in cages for 1 month prior to experimentation. Animals were given cat chow and water ad libitum and maintained on a 12 L:12D cycle at 20 3 C. The Institutional Animal Care and Use Committee at the University of Florida (UF #200801939) approved all procedures. Surgical P rocedure Mice were anesthetized with isoflurane (SurgiVet Isotec T 3 Classic Isoflurane Vaporizer ) and castrated (Cast) bilaterally through a 1 cm scrotal incision. We placed a 10 mm silastic implant (1.47 mm inner diameter X 1.96 mm outer diameter Dow Corning Corporation, Midland, Michigan) subcutaneously along the dorsal mid line. Implants were filled wi th either 1 mm (15 mg) of testosterone (T; Sigma T1500), 2mm (30 mg) of testosterone, 1mm of dihydrotestosterone (DHT; Sigma A8380), or left empty ( n = 9 per treatment). The DHT treatment group was included to identify whether androgens were sufficient to cause behavioral and acoustic changes in the absenc e of aromatization (Nyby et al. 1992). All implants were sealed with silicon adhesive, sterilized with ethylene oxide, and soaked in 0.1% saline solution at 37 C for 12 hours prior to implantation. Resear chers were blind to animal treatment and implant concentrations approximated physiological doses of plasma T from adult field captured animals (see Results).
17 Song Recording and M easurement Fourteen days pre and post surgery, we placed mouse cages inside an acoustic chamber (42 x 42 x 39 cm) made from expanded PVC and lined with anechoic foam. Between 0700 1000h, we recorded spontaneous song rate for 30 min, broadcast a single conspecific song, and measured the latency to counter sing for 5 min. The broadc ast song was randomly chosen from among 51 laboratory recordings of 12 wild caught males. The stimulus was within 2 standard deviations of the population mean for all s ong parameters (Campbell et al. 2010). We played stimuli at 50 dB SPL at 1 m from a Pion eer TS 250 speaker (flat frequency response 6 40 kHz) via an external soundcard (Edirol FA 66). We used an ACO Pacific microphone and preamplifier (Model 7016 and 4116) connected to a laptop running Raven Pro 1.3 via an external soundcard (Edirol FA 66) to record mouse songs at 30 cm. Microphones were calibrated prior to recording sessions with a Brel and Kjr sound level meter (Type 2219) and calibrator (Type 4230). All songs at were sampled at 96 kHz and 24 bits. To analyze recordings, we used automat ed code written in MATLAB (Phelps, available upon request) to quantify a variety of song parameters including note number, trill rate (note/s), power (dB SPL), frequency bandwidth (kHz), and dominant frequency (kHz) from three spontaneous songs per individ ual pre and post treatment. We then measured a subset of parameters for every note in the trill and used a quadratic function, f ( x ) = A x 2 + B x + C, to summarize how parameters changed over the course of a song. For example, to describe frequency bandwidth over the course of the song, f(x) is defined as the bandwidth at note number x, C represents the bandwidth of the first note, B corresponds to change in bandwidth between the first and second notes, and A describes how rapidly bandwidth changes ov er the c ourse of a song (Figure 2 1 D ). We
18 refer to these parameters as note bandwidth A, B, and C in Table 2 1. Finally, we averaged the coefficients of variation (CV= [s.d./mean] x 100) from each song parameter to assess the degree of song variability pre and po st treatment. Resident Intruder T rials We performed resident intruder trials to test whether androgens mediate aggressive behavior in S. teguina The day following pre surgery song recording, we introduced a random novel male (intruder) to the home cage of the treatment animal (resident) and videotaped their interactions for 5 min. Cages of residents were transferred from isolation to the behavior room 30 min prior to initiation of trials. We recorded number of attacks, latency to attack, and number of subm issive supine displays of treatment animals. We repeated the same procedure fifteen days post treatment, but presented treatment animals with novel intruders so that each intruder was used only once. Testosterone R adioimmunoassay All animals were euthaniz ed 16 days post surgery and trunk blood was collected with heparinized capillary tubes. We centrifuged blood at 2500 rcf for 15 min. We used a solid phase, 96 well plate (Perkin Elmer, Boston, MA, Protein A Flash Plate Plus) radioimmunoassay (RIA) to deter mine plasma T concentrations as previousl y described (Hamlin et al., 2011 ). Briefly, an antibody specific to T (Fitzgerald Industries, Concord, MA, Cat # 20 TR05T) was diluted in phosphate buffered saline gelatin to a concentration of 1:15,000. We incuba ted plates with antibody for two hours at room temperature. 3 H labeled steroid was then added at 12,000 counts/min per 100ul and plates were incubated for three hours at room temperature. Prior to sample analyses, we serially diluted a plasma pool to ens ure parallelism with the standard curve.
19 Interassay variance wells were similarly prepared from a plasma pool of 5 animals. Intra assay variance averaged 3.7%, whereas interassay variance averaged 6.4%. T concentrations of DHT treated males reflected in part the cross reactivity of the assay and are reported below but not included in our analyses. Although the T antibody used in our RIA is reported to have 1% cross reactivity with DHT (Fitzgerald Industries), the "T" levels detected in our DHT treated ca strates suggest androgen cross reactivity is more substantial. It is worth noting, however, that other studies show DHT treated castrates to have substantial levels of T (e.g. Crews et al. 1978). We cannot rule out the possibility that DHT promotes T relea se from non testicular sources. We compared T concentrations of our experimental animals to those of adult males ( n = 12) sampled in the field and processed in an identical manner. Data A nalysis We first used analysis of variance (ANOVA) and Kruskal Wallis tests to assess variation of all pre treatment measures among groups. Paired t tests and Wilcoxon signed rank tests were then used to assess differences within groups pre and post treatment. We corrected for multiple comparisons of song parameters using the Benjamini treatment T concentrations and post pre differences within groups among treatments. We conducted post between circulat ing concentrations of T and post measures on song characteristics, we fit a logarithmic curve to individuals across treatments and used linear regressions to highlight patterns within treatment groups. All statistics were performed in JMP Version 7 (SAS I nstitute Inc. Cary, NC, 1989 2007). We report means standard error (s.e.) throughout the text unless otherwise noted.
20 Results Plasma Androgen C oncentrations T concentrations ( s.d.) varied significantly among groups post treatment (Cast + empty: 0.39 0.1 ng/ml, Cast + DHT: 2.03 3.3 ng/ml, Cast + T (1mm): 5.33 2.8 ng/ml, Cast + T (2mm): 13.28 3.1 ng/ml; P < 0.0001) and approximated the physiological range of adult males sampled in the field (0.76 6.35 ng/ml). Resident Intruder T rials The number of attacks or latency to attack intruding males did not differ pre and post treatment in any group ( P > 0.05). However, empty implanted males engaged in more submissive supine displays post treatment (Pre: 0.1 0.1, Post: 1.1 0.3; W = 14.0, P = 0.01; F igure 2 2). Singing B ehavior Pre treatment measures for all behavioral variables did not differ among groups ( P > 0.05). Empty implanted males decreased spontaneous song rate (Pre, 1. 6 0.3 songs/10 min; Post, 0.48 0.2 songs/10 min; W = 22.5, P = 0. 004; Fig. 2 3 ) and increased latency to sing in response to a conspecific song (Pre, 217.6 35s; Post, 275.3 23s; W = 7.5, P = 0.03; Figure 2 3 ), whereas DHT, T (1mm), and T (2mm) implanted males increased spontaneous song rate ( W = 9.5, 14.0, 17.5 resp ectively, P < 0.05) but exhibited no change in latency to respond ( W = 1.5, 1.5, 7.5 respectively, P > 0.05; Figure 2 3 ). Song P arameters Experimental manipulation had significant effects on a variety of song parameters in empty implanted males, as exempl song pre and post treatment ( Figure 2 4; Object s 2 2 and 2 3 ). Empty implanted males
21 sang songs with fewer notes (Pre, 101.4 6 notes; Post, 54.7 6 notes; P > 0.001) that were lower in power (Pre, 71.6 1 dB SPL; Po st, 62.6 1.9 dB SPL; P > 0.001) with slightly higher dominant frequencies (Pre, 19.9 0.8; Post, 22.5 2 kHz; p = 0.18; Table 2 1; Figure 2 5 ). Such changes were manifest in animals starting songs with longer note durations and internote intervals that quickly decreased in power (Note Duration C, INI C, and RMS amplitude B; Table 2 1 ). Empty implanted males also sang less stereotyped songs after treatment (Pre, 6.8 0.3%; Post, 11.1 1.7%; P > 0.05; Table 2 1; Figure 2 5 D ). Conversely, implants mainta ined the majority of song parameters in DHT and T treated groups, with the exception of T (1mm) treated males exhibiting increased frequency bandwidth (Pre, 25.2 0.6 kHz; Post, 26.2 0.5 kHz; p > 0.01) and T (2mm) treated males showing decreased domin ant frequency post treatment (Pre, 21.2 0.7 kHz; Post, 18.8 0.5 kHz; P > 0.01; Table 2 1; Figure 2 5 ). The relationship between circulating T concentrations and number of notes post treatment showed a highly non linear effect across treatments (y = 10 .5(ln)x + 68, r 2 = 0.52, F 1,23 = 24.73, P < 0.0001 ; Figure 2 6 ), reflected in varying slopes of the relationship within treatment groups. Among empty implanted individuals, for example, low circulating concentrations (below physiological range of adult mal es captured in the field) of androgens showed a steep positive correlation with number of notes ( m = 137.6, r 2 = 0.63, P = 0.02). T (1mm) treated males showed a less steep association ( m = 1.5, r 2 = 0.11, P = 0.4), whereas T (2mm) treated males that were above physiological levels of T showed no relationship ( m = 0.71, r 2 = 0.05, P = 0.56). We found similar non linear effects between circulating T concentrations and power (y = 2.6(ln)x + 65.8, r 2 = 0.57, F 1,23 = 7.53, P < 0.0001), dominant frequency y = 2 1.6(ln)x 1.1, r 2 = 0.26,
22 F 1,23 = 7.55, p = 0.01, and song variability (y = .017(ln)x + 0.08, r 2 = 0.37, F 1,23 = 11.21, P = 0.003; Figure 2 6 B ). Discussion We explored the role of androgens in modulating singing behavior, song characteristics, and assoc found that androgens influenced a suite of measures, highlighting their role in linking gonadal status to decisions about investment in reproductive behaviors. Some effects of castration were profound and conspicuous (e.g. reductions in song rate, length, and power), whereas their influence on aggression was subtler. Administration of T and DHT influenced measures of song rate and response latency similarly. Conversely, empty implanted males sang less and showed longer latencies to respond to conspecific songs. Because DHT cannot be aromatized to estrogen, androgens appear to act directly on androgen receptors to mediate vocal behavior in singing mice. USVs are androgen dependent in other rodents, tho ugh estrogens facilitate the production of vocalizatio ns in rats and gerbils (Floody 1981; Nyby et al. 1992 ). Androgenic effects on song rate and latency reflect motivational state in many taxa including rodents (Nunez et al. 1978) and likely act on neural pathways that govern vocal production. Androgens had a profound influence on the structure of vocalizations. When empty implanted males sang, their trills were shorter, lower in power, higher in frequency, and more variable. Song length is a classic indi cator of motivation in anurans and birds ( Catchpole et al. 1986; Bensch and Hass elquist 1991; Welch et al. 1998 ), and power can convey information on receiver intent and condition ( Gerhardt and Huber
23 2002; Forstmeier et al. 2002; Brumm and Todt 2004; Wyman et al. 2008). The fact that empty males sang at lower frequencies after treatment indicates that the larynx remains androgen responsive after puberty. Androgens may influence lar ynx size or modulate expression of myosin heavy chain isoforms that govern the velocity and force of muscul ar contractions (Fischer et al. 1993; Dabbs and Mallinger 1999), thus providing a potential mechanism for changes in song attributes. The high song v ariability of empty implanted males suggest that androgens are involved in coordinating both the fine neuromotor control and highly motivated state associated with son g stereotypy (Kao and Brainard 2006; Woolley and Doupe 2008). Thus, the extensive effect s of androgens on vocalizations suggests they act on a suite of targets, including brain regions that relay impulses to the larynx and peripheral muscles that influence dominant frequency and amplitude during exhalation (Beani et al. 1995). While androgen s had a profound influence on song, their effects on aggression were significant but subtler. Androgens significantly reduced the probability of submissive supine behavior, but we found relatively little offensive aggression in any treatment group. Moreove r, animals exhibited low levels of offensive aggression prior to castration, indicating that the surgery itself was not causing low aggression levels. We suspect that our testing procedures contributed to low levels of aggression. Blondel (2004) placed mal e residents of S. xerampelinus a sister species, in the testing room the day before testing, performed trials under red light, and found that animals were very aggressive. The fact that we staged encounters 30 min after placing males in the novel testing room may have led subjects to behave as intruders rather than residents.
24 Nevertheless, the ability of castration to promote supine displays indicates androgens modulate aggression as well as vocalization in singing mice. Interestingly, the influence of and rogens on vocal attributes followed a strongly non linear relationship ( Figure 2 6 ). The relationship between number of notes and plasma testosterone titers, for example, is steepest in the range of hormone concentrations exhibited by castrates, shallow in the range of intact males, and flat at supra describes how animals translate variation in circumstance (in this case reproductive status signaled by androgen concentration s) into changes in phenotype (vocal effort). If this relationship is adaptive, the steep increase in vocal effort elicited by T concentrations just below those of healthy adult males may be a means of ensuring that males engage in high vocal effort when te stes are producing gametes, or none at all when conditions are unfavorable for reproduction. The more shallow relationship within the range of adult T titers suggests that natural variation in androgens coordinate subtle differences in vocal effort among m ales. Lastly, the flat relationship at supra physiological doses suggests that increasing vocal effort beyond that plateau will provide no further increase in fitness regardless of reproductive condition or social status. Such non linear curves were consi stent across measures of vocalization, and have long been observed in behavioral endocrinology ( Beach 1948; Adkins Regan 2005; Ball and Balthazart 2008). Reaction norms could provide a useful and general framework for investigating the relationship between hormone concentrations and behavior.
25 The hormonal coordination of multiple dimensions of vocal effort is similar to proposed endocrine control over vocalization rates in anurans ( Marler and Ryan 1996; Emerson 2001; Leary et al. 2004). The model suggests that androgens have graded effects on vocal characteristics and are suppressed by negative feedback from corticosterone associated with energetic constraints of calling. Future studies that incorporate androgenic interactions with other hormones and neuro modulators will provide important insight into the full suite of mechanisms involved in signal production. By introducing a novel mammalian species, we hope to broaden the taxonomic scope of work in acoustic signaling to facilitate integration of hormones behavior, and evolution.
26 Table 2 1 Results of paired t tests of acoustic variables for male singing mice post pre treatment. Bolded values are significant after Benjamini Hochberg procedure ( Q = 0.10). INI = inter note interval, RMS am p = root mean square amplitude, DF = dominant frequency. Change is RMS Amp is linear (values = 0). Acoustic variable Empty DHT (1mm) T (1mm) T (2mm) t 8 p t 8 p t 8 p t 8 p Amplitude measures # notes 6.42 0.0002 0.13 0.89 0.37 0.72 0.58 0.57 Trill rate (notes/s) 1.16 0.27 2.03 0.07 2.01 0.07 0.23 0.82 Power 5.34 0.0007 1.76 0.11 0.78 0.45 0.23 0.82 Note duration A 1.51 0.16 0.88 0.41 2.06 0.07 0.04 0.96 Note duration B 2.51 0.03 0.19 0.85 0.77 0.46 0.67 0.51 Note duration C 1. 37 0.02 0.24 0.81 2.10 0.06 0.15 0.88 INI A 2.02 0.07 1.38 0.20 2.82 0.02 1.13 0.28 INI B 0.66 0.52 1.12 0.29 2.92 0.01 1.21 0.25 INI C 2.59 0.03 1.16 0.27 1.50 0.17 2.03 0.07 Note RMS Amp A Note RMS Amp B 3.32 0.01 1.28 0 .23 0.15 0.88 0.70 0.49 Note RMS Amp C 0.63 0.54 0.55 0.59 0.47 0.65 1.74 0.11 Frequency measures Bandwidth 1.19 0.08 1.05 0.32 3.48 0.008 0.27 0.79 Dominant frequency 1.46 0.18 0.74 0.47 1.29 0.23 3.72 0.005 Note bandwidth A 0.15 0. 88 0.46 0.65 0.28 0.78 0.71 0.49 Note bandwidth B 0.28 0.78 0.32 0.75 0.07 0.94 0.40 0.69 Note bandwidth C 0.07 0.93 0.66 0.52 1.15 0.28 0.21 0.83 Note DF A 0.61 0.55 0.63 0.54 0.06 0.94 0.37 0.71 Note DF B 0.13 0.89 1.30 0.22 0.25 0 .80 1.12 0.29 Note DF C 0.38 0.71 1.40 0.19 0.58 0.57 2.51 0.03 Stereotypy measure Song variability 2.77 0.02 0.50 0.62 1.45 0.18 0.49 0.64
27 Figure 2 S. teguina ). singing mouse ( S. teguina ) song. B) Spectrogram and power spectrum of 3 notes underlined in A). C) Adult male S. teguina singing. (D) Note measurements on change in frequency bandwidth, inter note interval, and note duration over the course of song. DF = dominant frequency.
28 Figure 2 2. Androgenic effects on aggressive behavior. Number of submissive supine displays by resident male singing mice to intruding male conspecifics pre and post treatment. Error bars are + 1 s.e. p < 0.05 for p within groups. Groups not sharing the same letter have significantly different p < 0.05). Cast= castrate, DHT = dihydrotestosterone, T = testosterone.
29 Figure 2 3. Androgenic effects on sin ging behavior A) Spontaneou s song rate (songs/10 min) and B) latency to sing (s) in response to conspecific song of male singing mice pre and post treatment. Error bars are + 1 s.e. p < 0.05 er p < 0.05). Cast = castrate, DHT = dihydrotestosterone, T = testosterone.
30 Figure 2 4. Mouse song pre and post castration. Representative songs of an individual male singing mouse castra ted + empty imp lant pre and post treatment. DF = dominant frequency.
31 Figure 2 5. Androgen ic effects on song parameters. A) Number of notes, B) Power (dB SPL at 30 cm, C) Dominant frequency (kHz), and D) Song variability (CV *100) of male singing mice pre and post treatment. Error bars are + 1 s.e. ** p < 0.001, p < HSD, p < 0.05). Cast= castrate, DHT = dihydrotestosterone, T = testosterone.
32 Figure 2 6. Relationship between andro gen concentrations and so ng parameters in singing mice. A) Testosterone concentrations (ng/ml) of hormone treated male singing mice by number of notes. Black line along x axis indicates the range of physiological concentrations of testosterone from field c aptured mice in Costa Rica and Panam. m = slope, p < 0.10. B) Testosterone concentrations (ng/ml) of hormone treated male S. teguina by power (dB SPL at 30 cm), dominant frequency (kHz), and song variability (CV *100).
33 Object 2 1. Song of intact male A 0.71 M B) Object 2 2. and played back at speed. (.wav file 1.82 MB) Object 2 cas tration. Recorded at 96 kHz and played back at speed. (.wav file 0.91 MB)
34 CHAPTER 3 ANDROGEN DEPENDENT MALE VOCAL PERFORMANCE INFLUENC ES FEMALE PREFERENCE IN NEOTRO PICAL SINGING MICE 1 Background Vocalizations used in aggressive and mating contexts a re among the most diverse and elaborate displays in the animal kingdom. Physical or physiological constraints can result in signals that convey information about signaler condition (Maynard Smith and Harper 1995; Vehrencamp 2000; Fitch and Hauser 2003; Hur d and Enquist 2005; Taylor and Reby 2010). Receiver responses to such signals can generate substantial intra and inter sexual selection (Andersson and Simmons 2006). In birds, anurans, and mammals, conspicuous features of vocal signals (e.g. duration and fundamental frequency) provide accurate indicators of male quality and size that are attractive to females and threatening to rival males ( Clayton and Prove 1989; Welch et al. 1998; Reby et al. 2005). Such displays are thought to transmit honest informati on because only individuals in better condition are able to afford costs associated with signal production. In vertebrates, vocal displays associated with reproduction are often mediated by androgens released from the testes (Floody 1981; Ball et al. 2003 ; Moore et al. 2005; Bass and Remage Healey 2008). Androgens can influence vocalizations through actions on motivational centers and vocal motor pathways in the central nervous system (reviewed in Yamaguchi and Kelly 2002; Bass and Remage Healey 2008) or v ia modulation of peripheral structures involved in signal production. For example, changes 1 Reprinted with per mission from Animal Behaviour. Pasch, B., George, A.S., Campbell, P., and S.M. Phelps 2011 a Androgen dependent male vocal performance influences female pre ference in Neotropical singing mice. Animal Behaviour 82: 177 183.
35 in fundamental frequency result from anabolic effects of androgens on the larynx and 1985; Beani et al. 1995; Cynx et al. 2005; Evans et al. 2008; Pasch et al. 2011; Taylor and Reby 2010). By modulating signal duration, rate, or fundamental frequency, androgens can provide a mechanistic link between male condition and signal form (Remage H ealey and Bass 2007; Charlton et al. 2011 ). While androgens influence conspicuous features of vocal signals, little is known about their effects on vocal attributes that reflect male performance. For example, many avian species emit a series of rapidly re peated notes of similar structure termed trills. Trill production is constrained by a mechanical trade off between how fast notes are repeated (trill rate) and the frequency bandwidth of each note, resulting in a performance limit (Podos 1997) Males with high vocal performance scores are considered greater threats by rival males and are more attractive to females in a variety of species (Ballentine et al. 2004; Illes et al. 2006 ). Vocal performance is also correlated with age, social status, and reproducti ve success ( Janicke et al. 2008; Ballentine 2009; Botero et al. 2009; de Kort et al. 2009). Despite the association between circulating testosterone and these same variables (age: Morton et al. 1990; Vleck and Brown 1999, social status: Wingfield 1985; Win gfield and Moore 1987, reproductive success: Reed et al. 2006) the role of androgens in mediating variation in vocal performance is unclear. In birds, vocal performance trade whereby sound frequencies tr ack beak gape, with wide gapes corresponding to high frequencies and small gapes corresponding to low frequencies (Westneat et al. 1993; Hoese et al 2000). However, other aspects of respiratory and syringeal motor patterns
36 are important in modulating aspec ts of note rate and frequency modulation (Goller and birds take between each note suggest a potential role for respiratory musculature in limiting temporal aspects of n ote production (Calder 1970; Hartley and Suthers 1989; Lambrechts 1996 ). While the influence of androgens on avian respiratory musculature has not been well characterized, the mammalian diaphragm is highly sensitive to androgens that can influence the velo city and force of respiration (Prezant et al. 1997). Similarly, the craniomandibular musculature of rodents is responsive to androgens (Lyons et al. 1986; Eason et al. 2000), which suggests a mechanism by which androgens could influence the rate of gape an d frequency modulation. These data suggest that androgens could provide an important yet underappreciated mechanism modulating vocal performance, and highlight the potential for an appropriate mammalian model to inform our understanding of vocal signals. N eotropical singing mice (genus Scotinomys ) are diurnal insectivorous rodents that inhabit montane cloud forests throughout Central America (Hooper and Carleton 1976; Wilson and Reeder 2005 ). Adult males commonly emit stereotyped trills that are androgen de pendent and used in male male aggression and mate attraction (Pasch et al. 2011; herein). Presentation and subsequent removal of females increases male song rate ( Fernandez Vargas 2006), and males countersing in response to broadcast songs of intruding mal es in the field and laboratory (Pasch et al. 2011; Chapter 4 ). When singing mice trill ( Object 2 1 modulate the fundamental frequency (10 43 kHz) with their mouth gape (Miller and
37 Engstrom 2007; Obje ct 3 1 ) in a manner analogous to some bird song (Suthers et al. 1999). S. teguina ) exhibits a vocal performance trade off between trill rate and frequency bandwidth. We then manipulate androgens in la b reared males to investigate whether they play a role in modulating vocal performance. Finally, we experimentally manipulate trills to determine if females responded to versions resembling those of androgen treated males. We predicted that androgen remova l would negatively impact vocal performance, and that females would prefer male mouse trills with higher performance scores over those with lower scores. Materials and Methods Generating t he P opu lation Performance L imit To assess variation in male trills, we live captured animals in Sherman traps baited with peanut butter and oats in four localities throughout Costa Rica and Panam from 2006 2009 (see Campbell et al. 2010). We measured their mass and hindfoot length prior to transferring males to PVC coated wire mesh cages (28 cm x 28 cm x 28 cm). Cages were placed inside anechoic recording chambers (42 cm x 42 cm x 39 cm) made from expanded PVC. Details on song recording can be found in Campbell et al. (2010). We recorded spontaneous trills from 102 males ( N = 3.73 1.57 trills/male, range 1 12). Animals were uniquely marked and released at the site of capture. All capture and recording procedures in the field were conducted with approval from the Institutional Animal Care and Use Committee (IACUC) at the U niversity of Florida (#E436), the Costa Rican Ministerio del Ambiente y Energia, and the Panamanian Autoridad Nacional del Ambiente.
38 We used an automated code in MATLAB (available upon request) to quantify the average trill rate and frequency bandwidth fo r the entire song. Frequency bandwidths were calculated by a threshold routine that identified frequencies at the onset and offset of notes that were 8 standard deviations above the average background noise (sampling rate: 96 kHz; Fast Fourier Transform si ze: 256; Hann window, time resolution = 50 ms). We s songs (filled circles, Figure 3 2 A ) and calculated vocal performance limits as described in detail elsewhere ( Podos 1997; Ballentine et al. 2004 ). Briefly, we first pooled bins. For each bin, we determined the male with the largest frequency bandwidth (open circles, Figure 3 2 A ) and calculated a linear regression through these maxima (after Blackburn et al. 1992). This statistical tec hnique attempts to describe biological boundaries by defining theoretical limits to bivariate plots. In this context, the upper bound regression operationally defines the population performance limit on trill production, with trills closer to the limit bei ng more difficult to produce (Podos 1997). Androgen M anipulation We experimentally manipulated androgen levels of lab reared offspring (F1) of wild captured mice from Cerros de la Carpintera, Costa Rica. Sexually experienced adult males were assigned rando mly to one of three treatment groups and housed individually in cages for 1 month prior to experimentation. Animals were given cat chow and water ad libitum and maintained on a 12L:12D cycle at 20 3 C. Mice were anesthetized with isoflurane (SurgiVet I sotec T 3 Classic Isoflurane Vaporizer ) and castrated (Cast) bilaterally. We placed a 10 mm silastic implant (1.47 mm i.d. X 1.96 mm o.d., Dow Corning Corporation, Midland, Michigan) subcutaneously along the
39 dorsal midline. Implants were filled with either 1 mm (15 mg) of testosterone (T; Sigma T1500), 1 mm of dihydrotestosterone (DHT; Sigma A8380), or left empty ( N = 9 per treatment). The DHT treatment group was included to identify whether androgens were sufficient to cause acoustic changes in the absence of aromatization (Nyby et al. 1992). Implants were sealed with silicon adhesive, sterilized with ethylene oxide, and soaked in 0.1% saline solution at 37 C for 12 hours prior to implantation. Implant concentrations approximated physiological doses of plas ma T of field captured animals (see Pasch et al. 2011). Fourteen days pre and post surgery, we placed mouse cages inside an anechoic recording chamber to record trills ( N = 3/male) as described above. We then plotted trill rate and frequency bandwidth for all experimental animals pre and post treatment and calculated performance scores as the orthogonal deviation (i.e. vocal deviation in Podos 2001) from the performance limit generated from field captured males. Researchers were blind to animal treatment All laboratory procedures were conducted with permits from IACUC at the University of Florida (#200801939). Female P honotaxis To assess female response to male trills, we live captured females at Cerros de la Carpintera, Costa Rica in February 2010. We used only parous females ( N = 18) as indicated by open pubic symphyses and enlarged nipples, because pregnant and lactating females ( N = 4) showed no phonotaxis towards male trills in preliminary studies. We isolated females in wire mesh cages placed ins ide anechoic chambers for 48 hours prior to experimentation. Females were given cat chow and water ad libitum and were released at the site of capture after our experiments.
40 To synthesize experimentally high performance scores, we randomly selected 6 tril 70% of gaps of silence between each note using bioacoustics software Raven Pro 1.3 ( Cornell Lab of Ornithology, Ithaca, NY) Thus for each Slow stimulus we had a correspondin By increasing trill rate without changing bandwidth, Fast trills were consistently closer to the population performance limit ( Slow: 6.8 0.90; Fast: 1.6 0.92; Paired t test, t 5 = 12.16, P < 0.0001; Figure 3 3 A ). Each Slow Fast stimulus pair was used to test 3 different females. Thus we used a total of 6 stimulus pairs to test 18 females. We tested females in a 92 cm x 8 cm x 23 cm arena made from translucent white acry lic with a central compartment to house the female subject. The arena was a single runway 8 cm wide, 23 cm tall, and 92 cm long, bisected by a central compartment 9 cm x 8 cm for housing the subject female. Each sliding door of the central chamber had a 5 cm diameter hole covered with stainless steel mesh to allow females to hear stimuli. Two speakers (Pioneer TS 250; flat frequency response 6 40 kHz) were mounted at opposite ends ( Figure 3 1 A ). Trills were broadcast from a laptop to each speaker via an external soundcard (Edirol FA 66) with a peak amplitude of 50 dB SPL at 1 m, which approximates natural trill amplitudes. Speakers were calibrated daily with a Brel and Kjr sound level meter (Type 2219) and calibrator (Type 4230). Females were placed i n the central chamber of the choice arena between 1500 1830 two days following capture. We allowed females to acclimate for 17 min prior to the first test stimulus. Slow and Fast stimuli from the same male were broadcast antiphonally from each speaker begi nning 3 min before the doors were lifted. Each stimulus was repeated every 60 s,
41 and the onset of the two stimuli were 30 s out phase. We then ceased playback, opened the doors and recorded female responses for 5 min. After each trial, the arena was cleane d with water and alcohol and the female was placed in the central chamber for a 17 min inter trial acclimation. The same stimuli were used for a second test, but stimulus order and side were reversed. We calculated the latency to approach (300 s if no appr oach) and time spent near (within 30 cm of speaker) each stimulus. Researchers blind to treatments scored all trials. Statistical A nalyses Linear regressions were used to determine the relationship between trill rate and frequency bandwidth, to calculate the performance limit, and to assess how body condition influenced performance scores of male trills recorded in the field. We used model II analysis of variance (ANOVA) to test whether variation in vocal performance was greater among vs. within individua ls. Variance components were then used to estimate repeatability of vocal performance within males ( Lessells and Boag 1987 ). For experimentally manipulated animals in the lab, we used ANOVA to assess variation of pre and post treatment performance scores among groups, and paired t te sts to assess differences within groups pre and post treatment. Significance of post treatment performance scores among groups was assessed using Bonferroni corrected alpha To estimate female response to male vocal performance, we iden tified the speaker that was approached first following playback. The null hypothesis of no preference predicts females will be equally likely to approach either stimulus in a given test. Expanding that to a two test paradigm (e.g. Kime et al. 1998), we pre dicted that 25% of
42 We used a chi square test to determine deviations from these expectations. To further assess female responses, we calculated the difference in time spent within 30 cm of the speaker broadcasting each stimulus, averaged across the two trials and within each stimulus type, and then performed a paired t test to ask whether females spent more time with Fast or Slow stimuli. We used the same analysis to examine the latency to approach each speaker. Positive values from paired t tests indicated that females showed a prefe rence for Fast stimuli, negative values indicated that females showed a preference for Slow stimuli, and 0 indicated no preference for either stimulus (either moving back and forth to each stimulus or exhibiting place preference after the first trial). In a similar manner, we tested for order effects by assessing differences in approach latencies and time spent near each speaker within and among trials. Lastly, we examined the relationship between female responses to the Fast stimulus and its performance sc ore. We within each stimulus type and regressed this average against the performance score for the corresponding stimulus. Similarly, we regressed the average time spent ne ar the All statistical analyses were performed in SAS Version 9.2 and JMP, Version 7 (SAS Institute Inc., Cary, NC, 1989 2008). We report means standard errors (SE) thro ughout the text unless otherwise noted.
43 Results Population Performance L imit We found a negative relationship between trill rate and frequency bandwidth from Costa Rican and Panamanian field populations (y = 466.8x + 30080, R 2 = 0.10, F 1,101 = 11.45, P = 0.001 ; Figure 3 2 A ). Regression of maximum bandwidth against trill rate revealed a negative relationship, operationally defined as the performance limit (y = 1191.4x + 42692, R 2 = 0.72, F 1,8 = 20.12, P = 0.002). We found greater among vs. within individ ual variation in vocal performance ( F 101,280 = 4.80, P < 0.0001), which showed a moderate level of repeatability (0.52). Body condition showed a significant but weak positive association with performance scores of field captured animals ( R 2 = 0.05, F 1,96 = 5.01, P = 0.027). Androgenic Effects on Performance S cores We found no significant group differences in pre surgery performance scores as expected given random assignment of treatments (ANOVA: F 2,24 = 1.71, P = 0.20). However, performance scores of emp ty implanted males fell significantly away from the population performance limit following castration (Paired t test: t 8 = 3.01, P = 0.01; Figure 3 2 B ). Conversely, performance scores of DHT and T treated animals remained at pre castration levels (DHT: t 8 = 0.49, P = 0.63; T: t 8 = 0.47, P = 0.64). The influence of androgens was reflected in a significant treatment effect on post castration performance (ANOVA: F 2,24 = 3.89, P = 0.03). Both T and DHT treated animals differed marginally from empty implant ed controls (T vs. Empty, t 17 = 3.27, P = 0.09; DHT vs. Empty, t 17 = 6.72, P = 0.02), but did not differ from one another ( t 17 =1.05, P = 0.34).
44 Female P honotaxis Three females did not approach either stimulus during the trials. The remaining 15 females responded in both trials. For the 15 responsive females, we found FF:FS:SS = 8:4:3. This differed significantly from the null expectation of no preference ( 2 2 = 6.6, P = 0.037). In general, females were twice as likely to approach the Fast stimulus as the Slow stimulus (F:S=20:10). F emales also had shorter latencies to approach Fast speakers and spent more time near Fast stimuli than Slow stimuli (Latency Fa st : 135.8 20.48 s and Latency Slow : 189.4 20.17 s, Wilcoxon signed rank test : t 5 = 10.5, P = 0.03; Time Fast : 105.9 18.09 s and Time Slow : 29.9 6.75 s; Paired t test: t 5 = 6.71, P = 0.001; Figure 3 3 B ). Order of stimulus presentation did not influ ence approach latency or time spent near speakers (Wilcoxon signed rank tests : t 5 =0.86, P = 0.43 and t 5 = 1.4, P = 0.22. respectively ). Time spent near Fast stimuli was positively correlated with variation in performance scores among trials ( R 2 = 0.69, F 1,4 = 9.12, P = 0.04; Figure 3 3 C ), whereas latency to approach Fast speakers showed a negative association ( R 2 = 0.57, F 1,4 = 5.24, P = 0.08). Discussion We documented a vocal performance trade off in the trills of male Neotropical singing mice and found tha t androgens play an important role in modulating performance scores. In turn, females showed shorter latencies and spent more time near speakers broadcasting high performance trills that resembled vocalizations of androgen treated males. These data demonst rate that androgens influence both the structure and efficacy of mouse vocalizations.
45 Singing mice exhibited a negative relationship between trill rate and frequency bandwidth, suggesting a trade off similar to that documented in a variety of birds (Podos 1997; Janicke et al. 2008; Podos et al. 2009). Frequency bandwidths are much higher in mice (17 30 kHz) than in birds (1 8 kHz; Podos 1997; Ballentine et al. 2004; Illes et al. 2006; Janicke et al. 2008), reflecting divergent mechanisms of vocal productio n (Fitch and Hauser 2003). Notably, high frequency sound production did not seem to correspond to large gape widths ( Figure 3 2 A ), suggesting gape width functions differently across taxa. Indeed, vocal performance trade offs may not arise from gape width per se, but from additional motor constraints governing sound producing organs (i.e. syrinx and larynx) and surrounding musculature (Suthers and Goller 1997; Nowicki et al. 1992; Podos and Nowicki 2004; Nelson et al. 2005). While the origins of the trade off in mice remains to be elucidated, such convergence across taxa suggests fundamental constraints on mechanisms of trill production. Androgens played a strong role in modulating vocal performance in singing mice and appear to be necessary for the prope r performance of trills. Both trill rate and frequency bandwidth decreased following castration and administration of empty implants, whereas T and DHT implants maintained both measures. The ability of DHT, a non aromatizeable androgen, to maintain vocaliz ations suggests that androgens act directly on androgen receptors rather than being aromatized to estrogens. However, administration of aromatase inhibitors would help to clarify the possible contributions of extra gonadal (e.g. adrenal) sources of aromati zeable androgens. To our knowledge, this is the first study to experimentally manipulate androgens levels to assess their impact on vocal performance. Circulating levels of testosterone are correlated with the
46 type similar in structure to trills and used in aggressive contexts (Galeotti et al. 1997). In singing mice, androgens also modulate aggressive behaviour, spontaneous song rate, and dominant frequency of trills (Pasch et al. 2011). These extensive effects on a suite of behavioural phenotypes suggest that androgens act on motivational centers in the brain as well as the larynx and its surrounding musculature. In anurans, androgens alter the expression of laryngeal myosin heavy chain isoforms that influence the velocity and force of muscular contractions (Fischer et al. 1993). In addition, musculature of both the diaphragm and jaw are known to be androgen responsive in mammals (Lyons et al. 1986; Prezant et al. 1997; Eason et al. 2000), providing additional possible routes for androgen influences over vocal performance. Future studies that disentangle central and peripheral targets of androgens will be valuable in understanding their dynamic effects on vocal signals that reflect performance. While androgen s had a strong influence on vocal performance in experimental animals in the laboratory, field captured males showed a significant but weak relationship between body condition and performance score. Such weak correlations between condition and performance are often found in birds (Ballentine et al. 2004; Beebee 2004; Janicke et al. 2008). We speculate that this weak correlation emerges because males modulate vocal effort based on their standing relative to competitors, rather than based on their absolute bo dy condition. In contrast to residual mass, androgen levels have a strong influence over vocal performance because they more reasoning draws heavily on the well established abil ity of sexual and aggressive
47 success to elicit testosterone surges in many taxa (Wingfield 1987; Wingfield et al. 1990; Oliveira 2004). Such successes provide a male with direct measures of his ability to monopolize access to resources like mates, space, o r food. Androgen responses to success may then act on androgen responsive tissues in the brain and body to translate individual differences in RHP into an appropriate level of reproductive investment. To assess whether androgens are indeed signaling indivi dual differences in RHP to the performance measures in the field. In practice, however, the episodic release of androgens may make it difficult to detect relationship s between plasma titers and behavior, and measures such as fecal metabolites may prove more stable predictors. It will also be necessary to manipulate both social experience and testosterone surges (e.g. Oyegbile and Marler 2005) to observe the impact each has over vocal performance. The fact that females showed preference for trills that had higher performance scores provides preliminary evidence that vocal performance may be used in mate choice, though we cannot distinguish among preferences for duration, trill rate, or vocal performance per se. We speculate that trill duration may not be as informative as other measures because female birds and frogs tend to prefer longer vocalizations (not shorter ones, as found herein; e.g. Clayton and Prove 1989; Ryan and Keddy Hector 1992; Neubauer 1999; Gerhardt et al. 2000). Decoupling the importance of vocal performance from trill rate will require additional playback experiments that manipulate frequency bandwidth while holding trill rate constant. Nevertheless, ou r experiment mirrors avian studies that demonstrate female preference for high performance songs
48 ( Draganoiu et al 2002; Ballentine et al. 2009; Caro et al. 2010) and recent work in lab mice ( Mus ) showing female approach to male vocalizations used in sexua l contexts (Hammerschmidt et al. 2009). More generally, our data add to a growing body of literature suggesting that female mate choice is based upon assessment of male motor performance (reviewed in Byers et al. 2010). The correlation between female respo nses and performance scores among trials suggests that females not only attend to variation between two given trills, but increase their response to high performance trills independent of the stimulus against which it is compared (Vallet et al. 1998; Balle ntine et al. 2004). Understanding how such preferences contribute to reproductive success under more natural conditions will provide important insight into the evolution of male trills. Our study highlights links between hormones coordinating reproductive behaviours and trade offs inherent to the production of physically challenging displays. Collectively, our findings support the classification of vocal performance as an index signal, with performance scores putatively reflecting the combination of social and physiological factors that contribute to androgen release (Vehrencamp 2000; DuBois et al. 2008). In turn, females attend to variation in male vocal performance and use it to guide their behaviours. These findings elucidate how male reproductive status can be translated into biologically meaningful variation in signal form.
49 Figure 3 1. Design of female preference experiment. A) Experimental arena and B) spectrograms of Slow and Fast stimuli used in the female preference study. Mouse was redrawn from Reid (1998).
50 Figure 3 2 Vocal perfo rmance in male singing mice. A ) Average trill rate (notes/s) vs. frequency bandwidth (kHz) of male S. teguina from Costa Rican and Panamanian field populations (filled circles; N = 102). The upper bound regressi on on trills with the largest frequency bandwidth per 1 note/s bin (open circles) represents a performan ce limit on trill production. B ) Trill rate (notes/s) vs. frequency bandwidth (kHz) of male S. teguina pre and post treatment ( N = 9 per group) plotte d against performance limit of males from Costa Rican and Panamanian field populations from (a). Error bars are 1 SD. P < 0.05 for deviation of performance score (orthogonal distance from upper bound regression).
51 Figure 3 3 Female response to exp erimentally manipulated male trills. A ) Trill rate (notes/s) vs. frequency bandwidth (kHz) of male S. teguina from C erros La Carpintera males ( N = 6 ). Arrows illustrate the experimental manipulation of vocal performance from Slow (open circles) to Fast (fi lled circles) stimuli. B) Time spent near Slow and Fast stimuli 1 SE. P < 0.05. C) Amount of time females spent near Fast stimuli against male performance score among trials.
52 Object 3 ong. Filmed at 100 frames per second and played back at 30 frames per second. (.avi file 11.7 MB)
53 CHAPTER 4 INTERSPECIFIC DOMINA NCE VIA VOCAL INTERA CTIONS MEDIATES ALTITUDINAL ZONATION IN NEOTROPICAL SINGI NG MICE Background Competition for limited resourc es often results in the evolution of behavioral displays that minimize direct physical combat. In many species, advertisement signals are used to communicate sender identity to a network of potential competitors over long distances (Bradbury and Vehrencamp 1998 ). By encoding information about resource holding potential (Parker 1974), signals reduce escalation of aggressive contests and help shape the spatial distribution of animals across landscapes. Thus, ecological spacing patterns are intrinsically linke d to communication systems in a variety of taxa. S ignals used to settle aggressive conflicts within species may be an important component of aggression between species (Bradbury and Vehrencamp 1998; Ord and Stamps 2009; Peiman and Robinson 2010). In partic ular, advertisement vocalizations that mediate competitive interactions among conspecific males can also contribute to interspecific dominance and territoriality w hen ecologically similar species come into contact ( Orians and Willson 1964; Murray 1971, 198 1; Rice 1978; Catchpole and Leisler 1986; Prescott 1987; Gerhardt and Huber 2002 ). In birds, response to heterospecific signals reflects underlying dominance interactions; subordinate species are silenced and/or repelled, whereas dominant species respond a ggressively (Robinson and Terborgh 1995; Martin et al. 1996). At small spatial scales, this type of asymmetric vocal dominance can result in local habitat segregation (Martin and Martin 2001a, b). At broader scales, interspecific communication may facilita te competitive exclusion of subordinate taxa across environmental gradients ( Robinson and Terborgh 1995; Jankowski et al 2010). Acoustic signaling thus represents an important yet understudied
54 mechanism mediating interspecific aggression that may contribut e to spatial segregation at species limits. The abrupt replacement of species along altitudinal gradients has long fascinated biologists (Humboldt 1807; Merriam 1895; Shreve 1915, 1922) and provided an opportunity to study how biotic factors limit geogra phic distributions (Brown 1971; Heller 1971; Heller and Gates 1971; Chappell 1978; Hairston 1980a, b, 1981; Repasky and Schluter 1994; Carothers et al. 2001; Twomey et al. 2008 ). Interspecific competition is posited to be a major factor delimiting the ran ge edge of closely related parapatric species (Bull 1991), though most evidence is based on expansion of altitudinal ranges in the absence of putative competitors ( Lack and Southern 1949; Diamond 1970, 1973; Terborgh 1971; Terborgh and Weske 1975; Mayr and Diamond 1976; Remsen and Graves 1995; but see Cadena 2007 and Cadena and Loiselle 2007 ). However, recent experimental work in Neotropical birds suggests that interspecific aggression maintains altitudinal range margins (Jankowski et al. 2010), h ighlightin g the importance of behavioral interactions in limiting species distributions. Neotropical singing mice ( Scotinomys ) are diurnal, insectivorous rodents distributed throughout the mountains of Central America. S. teguina ; 9 12 g ) is a widespread species inhabiting mid to high altitude cloud forest, forest edge, and abandoned pastures from southeastern Mxico to western Panam (1000 2900 m), whereas the Chiriqu singing mouse ( S. xerampelinus ; 13 15 g ) is restricted to the highest for ested summits and pramo of Costa Rica and Panam (2200 3300m; Wilson and Reeder 2005; Hooper and Carleton 1976; Figure 4 1 ). The two ecologically similar congeners occupy parapatric distributions in the Cordillera de
55 Talamanca and Cordillera Central, wher e the upper distribution of S. teguina contacts the lower distribution of S. xerampelinus between 2200 2900 m (Enders and Pearson 1939; Hooper 1972; Hooper and Carleton 1976 ). Interspecific competition and species differences in thermal tolerance are hypot hesized to maintain this narrow contact zone (Hill and Hooper 1971; Hooper and Carleton 1976). Males of both species commonly sing species specific trills (Miller and Engstrom 2007; Campbell et al. 2010) that are modulated by androgens and used in mate att raction and male male aggressio n (Pasch et al. 2011a, b ; Object 4 1 ). Vocalizations of S. teguina are long (4 7 s), decelerating trills of moderate power (54 dB SPL at 1 m) and high dominant frequencies (22 26 kHz), whereas S. xerampelinus trills are short er (1 2 s), louder (63 dB SPL at 1 m), and lower in frequency (16 20 kHz: Campbell et al. 2010; Object 2 1 and Object 4 2 ). The first goal of this study was to identify the altitude of species replacement along three slopes in the Cordillera de Talamanca and Cordillera Central We placed data loggers at replacement zones to provide insight into the role of temperature in limiting species distributions. Similar temperature regimes at replacement zones on different mountains would suggest that abiotic factor s are important in limiting one or both species. Our second goal was to evaluate the importance of species interactions in mediating altitudinal range limits. If interspecific interactions mediate species replacement, we predicted that a) species would exh ibit interspecific aggression in staged behavioral trials and b) removal of one species would result in competitive release of the opposite species (Connell 1983) Finally, if interspecific vocal communication contributes to species replacements, we predic ted that response to playback of heterospecific song would reflect underlying dominance interactions.
56 Materials and Methods Altitudinal Replacement Zones and Temperature R egimes In June August 2007, we established trapping transects to demarcate zones of a ltitudinal replacement on two mountain slopes in Costa Rica (Cerro de la Muerte CM; Volcn Iraz VI) and one slope in Panam (Volcn Bar [VB]; Figure 4 1 ). Hooper and selection of study s ites. A 200 m x 30 m grid of 44 traps (spaced 20 m perpendicular and 10 m parallel to the cline, respectively) was placed in the understory for 20 days at each site. If S. teguina was captured, the entire grid was moved 50 m upslope the following day. In addition, sporadic trap lines were placed at altitudes 300 m from the trapping grid to validate species distributions. Sherman live traps (Model SFA) were baited with peanut butter and oats and checked daily. We defined the altitude of replacement as the trapping grid where S. xerampelinus was captured and S. teguina was not captured over the course of 5 trap days. To provide a general estimate of ambient temperatures experienced by each species, temperature loggers (Hobo Pendant, Model UA 001 64, Onset Computer) were placed in the understory at replacement zones to collect ambient temperature every hour for 1 to 3 years. At one site (VI), we placed a temperature logger at the top of the mountain to characterize temperature regimes where S. xerampelinus o ccurs in allopatry. Along the same slope, we estimated the lowest altitude of S. teguina occurrence by broadcasting conspecific song at 100 m altitudinal intervals. Response to song indicated species presence (Hooper and Carleton 1976). Broadcast songs we re randomly selected from laboratory recordings of S. teguina captured higher on the same mountain (Campbell et al. 2010). We played stimuli at 54 dB SPL at 1 m from a Pioneer
57 TS 250 speaker (flat frequency response 6 40 kHz) via an external soundcard (Edi rol FA 66) connected to a laptop running Adobe Audition. Songs were broadcast at 3 min intervals followed by 2 min of silence for 20 min between 0700 1000. A temperature logger was placed at the lowest site where the vocal response of S. teguina was record ed. Interspe cific Behavioral T rials To examine whether singing mice exhibited interspecific aggression, we staged interspecific interactions between adult males of S. teguina and S. xerampelinus captured at the contact zone at CM and VI in June August 2 007 ( n = 9 trials/site). One individual of each species was placed on opposite sides of an 80 x 50 x 40 cm arena separated by an opaque partition for 48 hrs after capture. The arena was isolated in an open air laboratory on site and animals were fed ad lib itum with cat chow and water. The partition was removed between 0700 0900 to allow animals to interact for 10 min. We videotaped all trials and scored the number of aggressive attacks and latency to attack for each species. Reciprocal Removal E xperiment If interspecific interactions mediate species replacement, we predicted that removal of one species would result in competitive release of the opposite species in the form of immigration. In May August 2008, we performed a reciprocal removal experiment at the zone of replacement on VI, which coincided with a transition from secondary forest scrub (solely occupied by S. teguina ) to primary forest (solely occupied by S. xerampelinus ; Appendix A). We established twelve 32 m x 24 m plots of 20 traps spaced 8 m apart in a rectangular grid. We paired control and removal plots for a given resident species ( S. teguina or S. xerampelinus ) with 3 replicates per species. Pairs of
58 control removal plots were spaced 100 m apart and were separated from all other paired plo ts by > 500 m. The vertical distance between S. teguina and S. xerampelinus plots was ca. 30 m and spanned 15 m in elevation Trapping was conducted for 3 consecutive days and every 5 days thereafter over 9 weeks. On control plots, resident animals were c aptured, marked for unique identification, and released on site. On removal plots, individuals of the resident species were captured, uniquely marked, and either transplanted 2 km or used in laboratory playback experiments (below); individuals of the colon izing immigrant species were uniquely marked and released on site. Temperature loggers were placed in the center of all plots to record temperature every hour for 14 days. Temperatures were averaged across control and removal plots for all replicates. Fie l d Playback E xperiments If vocal communication contributes to species replacement, we predicted that response to playback of heterospecific song would reflect underlying dominance interactions. During the removal experiment, we identified traps on control plots where resident males were captured most often. At the end of the experiment, we visited those trap sites ( n = 8 per species) between 0600 1000 and waited 10 min before initiating playbacks. Three types of stimuli were played during each trial: Contr ol stimulus, heterospecific song, and conspecific song. Songs used in playback experiments were recorded from adult males of the same population during the previous year (Campbell et al 2010). We selected at random 2 songs from each of 8 males per species and used songs of a different male for each trial. Control stimuli were generated in Adobe Audition using band pass filtered white noise that matched the spectral envelopes, duration, and root mean square (RMS) amplitude of corresponding conspecific songs. A Bruel and
59 Kjr sound level meter (Type 2219) and calibrator (Type 4230) were used to calibrate the amplitude of stimuli to match average songs of S. teguina (54 dB at 1 m) or S. xerampelinus (63 dB at 1 m). Each playback trial lasted 26 min and was rec orded with a Sony PCM 50 digital recorder We began with a 5 min pre stimulus period to record background levels of singing. At 5 and 6 min we broadcast each of two heterospecific songs, and recorded the number of songs evoked over a five minute interval ( 7 12 min); at 12 and 13 min we broadcast each of two control stimuli and recorded responses for five minutes (14 19 min); and at 19 and 20 min we broadcast each of two conspecific songs and recorded responses from 21 26 min. In half of the trials, control stimuli were presented at 5 and 6 min, and heterospecific stimuli at 12 and 13 min. Conspecific songs were always played last to avoid carryover effects that were apparent in preliminary studies (also see Martin and Martin 2001b). Species identity of respo nding individuals was verified from analysis of digital recordings, with S. teguina songs being consistently longer than S. xerampelinus songs at VI (5.37 0.95 s vs. 1.77 0.41 s, respectively; Campbell et al. 2010). Laboratory Playback E xperiments To validate playback experiments in the field and quantify animal movement associated with vocal behavior, we used adult males captured on removal plots to perform playbacks in a controlled laboratory setting. Animals were housed singly in 28 x 28 x 28 cm PV C coated wiremesh cages and isolated in anechoic chambers for 24 hrs prior to initiating trials between 0600 1000. Individuals of each species were assigned randomly to playback of 1 of 3 stimuli: Control stimuli, heterospecific song, or conspecific song ( S. teguina : n = 8 per treatment; S. xerampelinus ; n = 5 per treatment)
60 We selected 3 songs from each of 4 individuals of each species from laboratory recordings of the same population (Campbell et al. 2010). Control stimuli were synthesized by generating white noise bounded within species specific frequency envelopes ( S. teguina 10 42 kHz, S. xerampelinus 10 38 kHz) We played stimuli at 50 dB SPL at 1 m from one of four randomly selected Pioneer TS 250 speakers placed on each side of the mesh cage. Mice were recorded for 20 min to obtain pre stimulus song rate. At 20, 21, and 22 min, we broadcast each of 1 song followed by 2 minutes of silence. The sequence was repeated three times with songs from different males. For control treatments, the single contro l stimulus was broadcast at the same rate as conspecific and heterospecific song treatments. We recorded the total number of songs emitted by focal animals in the 20 min period following the initial stimulus. We also calculated the amount of time animals s pent on the half of the cage that held the speaker emitting the stimulus. Statistical A nalyses We analyzed response to the removal experiment using a generalized linear model (GLM) with a Poisson distribution (log link) performed in SAS/STAT v 9.2 (SAS In stitute 2008). Explanatory variables included species identity of resident animals ( S. teguina or S. xerampelinus ), species identity of colonizing animals ( S. teguina or S. xerampelinus ), and an interaction term. Plot replicate was initially incorporated a s a random effect but was removed because the model did not converge, indicating little variance among plots (Bolker et al. 2008; SAS Institute 2008). For each species, the response variable was calculated by subtracting the number of colonizers on contro l plots from their paired removal plots following the initial 3 day trapping period.
61 To analyze response to reciprocal playback experiments in the field, we used a Poisson distributed GLM. Individual was initially coded as a random effect but was removed because the estimated variance was zero (Bolker et al. 2008; SAS Institute 2008). The number of songs did not differ between prestimulus and control periods for either species (Kruskal 2 = 0.53 and 0.27, df = 1, P = 0.46 and 0.60 for S. teguina and S. xerampelinus respectiv ely). Therefore, we removed the prestimulus period from our model. Order effects were not significant ( P = 0.49) and were excluded from the model, resulting in stimulus type (control, conspecific, or heterospecific), species identity ( S. teguina or S. xera mpelinus ), and an interaction term as explanatory variables Contrast statements were used to test for significant differences among groups after adjusting for overdispersion using the Pearson 2 statistic (Littell et al. 2002) To analyze responses to playbacks in the laboratory, a multi factor ANOVA was applied to ranked data using the Scheirer R ay Hare technique (Sokal and Ro h lf 1995, Dytham 1999). The number of songs evoked post stimulus wa s subtracted from pre stimulus levels to obtain the response variable; species identity, stimulus type, and an interaction term were included as explanatory variables. Kruskal Wallis tests were then used to test for significant differences among treatment groups. To test differences among treatment groups in time spent near speakers, we used ANOVA followed by corrected post hoc comparisons. Results Altitudinal Replacement Z ones The altitude of replacement differed among mountains (VI: 2920 m; C M : 2610 m; VB: 2150 m; Figure 4 2 A ) yet temperature regimes that characterized replacement
62 zones were qualitatively similar (Figure 4 2 B ). Daily temperature profiles showed little variation and decreased with increasing altitude. Across their altitudinal distributions, S. xerampelinus experiences a narrower range and cooler temperatures than does S. teguina ( S. xerampelinus : 7.3 9.9 C; S. teguina : 9.9 18.7 C; Figure 4 2 B ). On the mountain with the highest altitude of replacement (VI), the replacemen t zone was concomitant with a change in vegetation from early succession scrubland at lower elevations to matu re forest at higher elevations Such vegetation transitions were not readily apparent at other localities. Interspecific Behavioral T rials In 17 of 18 trials, S. xerampelinus initiated more attacks than S. teguina (4.8 0.9 and 0.67 0.3 attacks; t = 4.91, df = 17, P < 0.0001 ; Figure 4 3 A ). In one trial, neither S. teguina nor S xerampelinus exhibited aggression. Similarly, S. xerampelinus show ed shorter latencies to attack than did S. teguina (156.9 41.6 and 442.4 55.6 s, respectively; Wilcoxon signed rank = 59, P = 0.001; Figure 4 3 B ). Reciprocal Removal E xperiment We trapped for 26,880 trap hours over 9 weeks and captured 121 S. teguin a (removed 65) and 71 S. xerampelinus (removed 34) S. teguina plots were warmer and showed more diel variation than S. xerampelinus plots ( S. teguina : 11.5 3.65 C; S. xerampelinus : 9.4 0.78 C; Wilcoxon signed rank test, 14771, P < 0.0001 ; Figure 4 4 A ). We found an asymmetrical response to removals, indicated by a significant 2 = 3.83, df = 1, P = 0.05; Figure 4 4 B ). S. teguina colonized S. xerampelinus removal plots within 3
63 weeks and remained there for the duration of the study ( n = 19), whereas no S. xerampelinus colonized S. teguina removal plots (Figure 4 4 C, D ). Field Playback E xperiments 2 = 32.53, df = 2, P < 0.001). Compared to control stimuli, S. teguina were more likely to sing in response to conspecifics and less likely to respond to heterospecifics 2 = 6.28 and 25.65 respectively, df = 1, P < 0.01; Figure 4 5 A ). Conversely, S. xera mpelinus were more likely to sing in response to both conspecifics and heterospecifics compared 2 = 5.07 and 3.14, P = 0.02 and 0.07 respectively; Figure 4 5 B ). Laboratory Playback E xperiments Similar to field playback experiments, we found a significant interaction between stimulus type and species identity ( 2 = 9.05, df = 2, P < 0.02 ). S. teguina were more likely to sing in response to conspecifics and less likely to respond to heterospecifics when compared to control stimuli ( H = 1 0.52 and 8.68 respectively, df = 1, P < 0.003; Figure 4 6A). Accordingly S. teguina spent more time near the speaker broadcasting conspecific stimuli (73 6%) and less time near heterospecific stimuli (12 6%) compared to controls (44 11%, F 2,21 = 14. 45, P > 0.001; Figure 4 6 B ). S. xerampelinus showed a tendency to respond to both conspecifics and heterospecifics, but our sample size was insufficient to detect differences (both H = 2.22, df =1, P = 0.13 ; Figure 4 6 A ). However, S. xerampelinus spent mor e time near speakers broadcasting conspecific and heterospecific stimuli (70 9% and 73 6%, respectively) compared to controls (32 4 %, F 2,12 = 11.67, P = 0.001; Figure 4 6 B ).
64 Discussion We investigated patterns of altitudinal zonation in Neotropi cal singing mice and found similar temperature regimes across different replacement sites, indicating that abiotic factors may be important in limiting one or both species. Subsequent studies revealed that the larger, higher altitude species, S. xerampelin us was behaviorally dominant to the smaller, lower altitude S. teguina Removal of S. xerampelinus resulted in colonization of S. teguina into cooler microclimates; conversely, S. xerampelinus did not colonize S. teguina removal plots characterized by wa rmer, more variable temperatures. Similarly, asymmetrical response to playback experiments in the field and laboratory indicated that acoustic communication reflects underlying dominance interactions; S. xerampelinus responded to heterospecific song, where as S. teguina was silenced by song of dominant heterospecifics. Our findings are consistent with aggressive exclusion of S. teguina from cooler environments at higher elevations. The study highlights how behavioral interactions mediated by interspecific co mmunication can influence geographic distributions of ecologically similar congeners. The altitude at which S. xerampelinus replaces S. teguina differed greatly across study sites. However, similar temperature regimes at each site suggested an important role for abiotic factors in contributing to species range limits. Indeed, the strong climatic stratification along tropical mountain slopes is thought to limit species distributions via selection on physiological tolerance to temperature (Janzen 1967; Ghal ambor et al. 2006; Tewksbury et al. 2008; McCain 2009 ). Hooper and Carleton (1976) found that the contact zone between S. xerampelinus and S. teguina tracked abiotic conditions as determined by the interaction of topography (e.g. slope, aspect) and climati c patterns.
65 For example, the replacement zone decreases from 2930 m to 2350 m on the eastern slopes of VI, where moisture laden wind and cloud cover provide conditions for epiphyte rich vegetation that confer cool microclimates suitable to S. xerampelinus (Hooper and Carleton 1976). We confirmed this pattern quantitatively, and proceeded to test whether biotic interactions contribute to the abrupt replacement zones. At the highest zone of contact (VI), the replacement of S. teguina by S. xerampelinus coinc ided with an abrupt transition between two vegetation types. Here, temperature regimes differed substantially acros s a small spatial scale (Figure 4 4 ). This pattern suggests that species differences in distribution may be driven by habitat rather than alt itude per se However, such vegetation transitions occur commonly at lower elevations where S. teguina inhabits a wide variety of vegetation types, including primary and secondary forest, and abandoned pastures. Furthermore, S. xerampelinus is not restrict ed to primary forest but occurs in pramo and forest edge on the highest peaks of the Cordillera de Talamanca (Hooper 1972; Hooper and Carleton 1976). In our study, S. teguina rapidly colonized cooler S. xerampelinus plots, whereas S. xerampelinus showed n o colonization into warmer S. teguina plots. These data provide evidence that S. xerampelinus responds to abiotic conditions and/or other factors that correlate strongly with temperature regimes. Our findings mirror results of a laboratory study investigat ing species differences in temperature tolerance (Hill and Hooper 1971). While both species showed thermal tolerances consistent with temperature regimes commonly experienced at native altitudes, experimentally elevating ambient temperatures to 35 C resul ted in death of 66% (2 of 3) of S. xerampelinus vs. 0% (0 of 3) of S. teguina In contrast, only 12.5% (1 of 8) of S. teguina and 0% (0 of 7) of S.
66 xerampelinus died when ambient temperatures were lowered to 0 5 C (Hill and Hooper 1971). Despite small sam ple sizes, the data indicates that S. teguina is a more able to acclimate to a wider range of ambient temperatures. In concert with our study, we infer that S. xerampelinus is relegated to cold summits, whereas S. teguina is less sensitive to prevailing ab iotic conditions and can inhabit cooler environments in the absence of dominant competitors. Interference competition is a predominant mechanism mediating interspecific interactions among small mammals, especially insectivores ( Grant 1972; Dickman 1991; R ychlik and Zwolak 2006). In many communities, ecologically similar species maximize foraging efficiency by exploiting similar prey items and dominance status is dictated by body size (Morse 1974; Dickman 1991; Martin and Martin 2001b). Smaller subordinate species tolerate interspecific interference because the benefits of gaining temporary access to resource rich microenvironments outweigh the costs of vigilance and evasion. In contrast, the cost of interference to larger dominant species is low compared to the benefits of exclusive access to prey rich microenvironments (Chappell 1978; Dickman 1991). Consequently, smaller bodied species are aggressively repelled by larger bodied congeners and show rapid competitive release upon their removal (Dickman 1986a, b, 1988). In our system, the larger S. xerampelinus feeds primarily on nutrient rich larval insects, whereas S. teguina includes a greater proportion of seed, berries, and fibrous plant materials (Hooper and Carleton 1976). S. xerampelinus is also behavior ally dominant, as demonstrated by more aggression and shorter latencies to attack in interspecific trials. Moreover, subordinate S. teguina exhibited rapid colonization into S. xerampelinus removal plots, whereas S. xerampelinus showed no
67 response to S. te guina removal. Such asymmetric interactions suggest that S. teguina is from S. xerampelinus presumably for access to resource rich microenvironments. Size dependent interspecific dominance i s prevalent in other taxa and may be a common biotic mechanism mediating geog raphic range limits (Morse 1974; Persson 1985; Glazier and Eckert 2002). Our findings mirror a widespread phenomenon in biology reflecting asymmetric patterns of interspecific dom inance relative to physiological tolerance, especially among closely related species that segregate along environmental gradients (e.g. barnacles: Connell 1961; birds: Martin and Martin 2001a; bees: Bowers 1985; chipmunks: Heller 1971, Chappell 1978; crayf ish: Bovbjerg 1970; fish: Hixon 1980; salamanders: Jaeger 1971 a, b: Griffis and Jaeger 1998 ). Dominant species are often unable to live in areas occupied by subordinate species, whereas subordinate species can thrive in a variety of habitats but are exclud ed from areas occupied by dominant species (Morse 1974). In small mammals, this pattern is associated with species differences in geographic range size and degree of specialization ( Glazier and Eckert 2002). Dominant species tend to have restricted geograp hic ranges and evolve in stable environments that favor ecological specialization, which in turn may select for higher aggression to defend narrow niches from invasion by generalists (Miller 1967; Morse 1974; Colwell and Fuentes 1975; Glazier 1980 ). Unders tanding the ubiquity of this pattern will be important in the attempt to model species responses to atmospheric warming. For example, rising temperatures are predicted to force montane species upward to eventual extinction, placing tropical mountaintop org anisms at high risk (Colwell et al. 2008; Lenoir et al.
68 2008). and account for variation in species responses to climate change (Davis et al. 1998; Marris 2007; Moritz et al. 2008; Jankowski et al. 2010). Role of A co ustic Communication in Interspecific I nteractions Acoustic communication is a common mechanism mediating intrasexual competition in numerous taxa. In insects and anurans, advertisement signals contribute to inter male spacing as males aggressively compete for females ( Gerhardt and Huber 2002 ). In birds and fish, vocalizations help deter intruder entry into territories containing limited resources in addition to mates (Collins 2004; Bass and Ladich 2008). When ecolo gically similar species come into contact, heterospecific aggression can result in interspecific territoriality, and response to reciprocal playbacks reflects underlying competitive interactions ( Robinson and Terborgh 1995; Martin and Martin 2001b). In thi s study, we found that S. xerampelinus responded to both conspecifics and heterospecifics. Conversely, S. teguina responded to conspecifics but were silenced by heterospecific song. This pattern matches asymmetric dominance interactions observed in the lab oratory and competitive release by S. teguina in response to S. xerampelinus removal in the field. Insectivores (e.g. shrews and moles) and other insectivorous small mammals aggressively defend territories containing sparse and slowly renewing food supplie s (Hawes 1977; Dickman 1986a, b ). In S. xerampelinus females and some males exhibit exclusive use of core areas consistent with territoriality (Blondel et al. 2009). Future work will need to characterize the mating systems of both species to fully underst and how social organization contributes to vocal interactions at the range edge. Likewise, replacement of speakers in areas where S. xerampelinus is removed will elucidate whether songs enforce species limits by deterring S. teguina entry.
69 Heterospecific p layback experiments to S. teguina in allopatry and manipulation of dominance status in the laboratory will also provide insight into whether aggressive experience is a necessary component of vocal discrimination, as appears to be the case in birds (Jankow ski et al. 2010). Conclusion Our study highlights the interaction of physiological and behavioral factors in mediating species range limits. Asymmetric patterns of interspecific dominance relative to physiological tolerance appear to be a common phenomenon when ecologically similar species segregate along environmental gradients Moreover, our findings suggest that asymmetries in interspecific communication reflect underlying dominance interactions and contribute to spatial segregation at the range edge. In terspecific communication may be an important yet overlooked mechanism structuring animal communities.
70 Fi gure 4 1. Latitudinal and altitudinal distributions of Neotropical singing mice ( Scotinomys ). S. xerampelinus (b lue) replaces S. teguina (red) on the highest mountains of Costa Rica and Panam. VI Volcn Iraz, CM Cerro de la Muerte, VB Volcn Bar.
71 Figure 4 2. Altitudinal distributions and temperature regimes of Neotropical singing mice A) Altitude of r eplacement (black line) on two mountains in Costa Rica (Volcn Iraz and Cerro de la Muerte) and Panam (Volcn Bar). B) Average temperature regimes at the site of replacement (black). At Volcn Iraz, average temperature regimes of S. xerampelinus (blue) and S. teguina (red) in allopatry.
72 Figure 4 3. Interspecific aggression in Neotropical singing mice. A) The number of attacks initiated by both species during staged trials. B) The latency to attack by both species during staged trials. Error bars are 1 s.e. P < 0.00 1
73 Figure 4 4. Reciprocal removal experiment on Volcn Iraz, Costa Rica. A) Average temperature regimes on S. xerampelinus (blue) and S. teguina (red) plots. Shading represents 1 s.e. B) Average number of colonizers on control and removal plots for S. teguina and S. xerampelinus C) Accumulation plot depicting the number of S. xerampelinus (blue) colonizing S. teguina removal plots as S. teguina (red) is removed over time (days). D) Accumulation plot depicting the number of S. teguina (red) colonizing S. xerampelinus removal plots as S. xerampelinus (blue) is removed over time (days). Error bars are 1 s.e.
74 Figure 4 5. Response to playback trials in the field at Volcn Iraz, Costa Rica. A) Number of songs produced by S. teguina (red) to playback of control, conspecific, and heterospecific (blue) stimuli. B) Number of songs produced by S. xerampelinus (blue) to playback of control, conspecific, and heterospecific (red) stimuli. Dashed line and shading represents the nu mber of songs produced by focal males prior to playback of stimuli. Error bars are 1 s.e. P < 0.05
75 Figure 4 6. Response to playback trials in the laboratory at Volcn Iraz, Costa Rica. A) Number of songs produced by S. teguina and S. xerampelin us pre (shaded bars) and post (open bars) playback of control, conspecific, and heterospecific stimuli. B) Percent time spent near speakers broadcasting sound stimuli. Error bars are 1 s.e. P < 0.05. Letters not sharing the same letter are significant ly different.
76 Object 4 1. Vocal response of a male S. teguina ) to playback of conspecific song. (.avi file 20 MB) Object 4 2. Song of a male Chiriqu singing mouse ( S xerampelinus ) recorded at 96 kHz. (.wav file 0.49 MB)
77 CHAPTER 5 AGONISTIC CHARACTER DISPLACEMENT IN SOCI AL COGNITION IN A TR OPICAL MOUSE Background Interspecific aggression is a prevalent but poorly studied phenomenon that may shape niche use in a variety of taxa (Ord and Stamps 2009; Peiman and Robinson 2010 ). When ecologically similar species come into secondary contact, agonistic interactions emerging from interference competition can cause selection on traits that affect fighting ability, encounter rates, or competitor recognition (Grether et al. 2009). Su ch agonistic character displacement (ACD ; Grether et al. 2009 ) effectively reduces interspecific interference competition and may facilitate competitiv e exclusion of subordinate taxa, thereby influencing community structure (Miller 1964; Bovbjerg 1970; Gil l 1974 ; Murray 1981; Peiman and Robinson 2010). Agonistic character displacement often involves divergence of signals used in competitor recognition ( Grether et al. 2009 ). For example, Hetaerina damselfl y wing coloration spots involved in territorial aggre ssion exhibit shifts in sympatry to minimize interspecific fighting (Anderson and Grether 2010). Alternatively, ACD may involve shifts in receiver responses to agonistic signals. M ale poison dart frogs ( Allobates ) in sympatry with heterospecifics that emit aggressive calls at similar frequencies exhibit increased discriminative ability compared to allopatric populations, reflecting release from competitive masking interference (Amzquita et al. 2006). However, little is known about the proximate mechanisms mediating competitor recognition systems (Grether 2011). Elucidating the neural mechanisms underlying ACD can provide important insight into the genetic basis and/ or phenotypic plasticity involved in interspecific recognition (ten Cate and Rowe 2007; Greth er et al. 2009).
78 Among animals that use acoustic signals to repel rivals and attract mates the peripheral auditory system is often tuned to match the frequencies contained in conspecific vocalizations (Capran ica 1965; Capranica and Moffat 1983 ; Ryan and K eddy Hector 1992; Riabinina et al. 2011 ) Efficient encoding of conspecific signals is also facilitated by tonotopic neurons in the auditory cortex, inferior colliculus, and medial geniculate (Ehret and Giessler 2005; Malmierca et al. 2008 ). S ympatric spec ies shift the tuning of their auditory systems to reduce responses to heterospecific signals, a phenomenon known as a sensory filter (Capranica and Moffat 1983 ) Such filtering occurs in intersexual contexts to reduce the costs of interspecific mating. I n principle, however, animals may preserve the function of peripheral sensory systems but assign species specific stimuli different motivational values, a phenomenon we refer to as a The latter scenario seems more plausible when hete rospecifics are ecologically salient and represent a potential threat. Neotropical singing mice are diurnal insectivorous rodents that inhabit montane cloud forests throughout Central America. S. teguina ; 9 12 g ) is a widespread sp ecies inhabiting mid to high altitude cloud forest, forest edge, and abandoned pastures from southeastern Mxico to western Panam (1000 2900 m), whereas the Chiriqu singing mouse ( S. xerampelinus ; 13 15 g ) is restricted to the highest forested summits a nd pramo of Costa Rica and Panam (2200 3300m; Wilson and Reeder 2005; Hooper and Carleton 1976 ). Males of both species commonly emit rapid frequency modulated trills that function in male male aggression and female mate choice. Vocalizations of S. teguin a are long (4 7 s), decelerating trills of moderate power (54 dB SPL at 1 m) and high dominant frequencies (22 26 kHz), whereas S.
79 xerampelinus trills are shorter (1 2 s), louder (63 dB SPL at 1 m), and lower in frequency (16 20 kHz: Campbell et al. 2010). On the highest peaks of Costa Rica and Panam S. teguina is abruptly replaced and aggressively excluded from higher elevations by the larger behaviorally dominant S. xerampelinus (Chapter 4). In a large scale analysis of geographic variation in songs, w e found no character displacement in the structure of trills between sympatric and allopatric populations of S. teguina (Campbell et al. 2010). However, the behaviorally subordinate S. teguina reduces song rate when broadcast trills of S. xerampelinus sug gesting an ability to distinguish between conspecifics and heterospecifics in sympatry (Chapter 4). In this study, we tested for ACD in song recognition by investigating auditory perception of S. teguina. We performed playback studies in the lab using mal e S. teguina from a population that co occurred with S. xerampelinus (sympatric) and those that did not (allopatric). Induction of the immediate early gene egr 1 was quantified in the auditory cortex and medial geniculate to identify mechanisms of vocal r ecognition. We predicted that the auditory system would increase its sensitivity to heterospecifics in sympatry to accommodate the ecological salience of the stimulus. Materials and Methods Reciprocal Playback E xperiments We live captured animals in Sh erman traps baited with peanut butter and oats on Volcn Iraz (VI sympatry) and Cerros La Carpintera (LC allopatry), Carta go Province, Costa Rica (Figure 5 1) Adult males were housed singly in 28 x 28 x 28 cm PVC coated wiremesh cages and isolated in an echoic chambers for 24 hrs prior to initiating laboratory playback trials between 0600 1000. Individuals were assigned randomly to playback of 1 of 3 stimuli: Control stimuli, heterospecific song, or conspecific song ( n = 8
80 per treatment at each site ) We selected 3 songs from each of 4 individuals of each species from laboratory recordings of the respective focal population (Campbell et al. 2010). Control stimuli were synthesized by generating white noise bounded within species specific frequency envelopes (10 42 kHz ). At LC, heterospecific stimuli were selected from S. xerampelinus recorded at VI. We played stimuli at 50 dB SPL at 1 m from one of four randomly selected Pioneer TS 250 speakers placed on each side of the mesh cage. Mice were recorded for 20 min to obtain pre stimulus song rate. At 20, 21, and 22 min, we broadcast each of 1 song followed by 2 minutes of silence. The sequence was repeated three times with songs from different males. For control treatments, the single control stimulus was broadc ast at the same rate as conspecific and heterospecific song treatments. We recorded the total number of songs emitted by focal animals in the 20 min period following the initial stimulus. We also calculated the amount of time animals spent on the half side of the cage that held the speaker emitting the stimulus. Tissue P reparation From a subset of individuals (n = 4 per treatment per population), we extracted b rains from euthanized animals 60 min after broadcast of the final stimulus. Brains were drop fixe d in 4% paraformaldehyde in phosphate buffered soluti on (PBS) and stored at 4 C for 14 d at each field site. Brains were then cryoprotected for 24 48 hrs in 30% sucrose made in 4% paraformaldehyde PBS and stored at 80 C until sectioning. Coronal sections from the olfactory bulbs through the cerebellum were cut in 40 m sections on a cryostat (Leica 1900, Leica, Mic rosystems, Heidelberg, Germany). Sections were collected into 24 well plate s containing a cryoprotectant solution (PBS containing 30% sucrose, 3 0% ethylene glycol, and 10% polyvinylpryrrolidone) and
81 stored at 20C until processing for egr 1 immunoreactivity. Egr 1 Immunocytochemistry Floating sections were washed in PBS, incubated for 15 minutes in 3% H 2 O 2 in 0.01 M PBS (pH 7.2 7.4) to quench en dogenous peroxidase activity, washed in PBS, blocked for 1 hour in 10% normal goat serum (NGS) with 3% Triton X in PBS (PBSTX), and washed in PBS. Section s were then incubated overnight at 4C with rabbit polyclonal antibody directed against egr 1 (Santa C ruz Biotechnologies sc 189 ), diluted 1:250 in 5% NGSPBSTX. Control sections were incubated without the primary antibody The next day sections were washed in PBS, incubated for 1 hour with biotinylated goat antirabbit lgG, rinsed in PBS and incubated for 1 hour in avidin biotin horseradish peroxidase solution (Leinco, St. Louis, MO; R106 and A106, both diluted 1:200 in 5% NGS PBSTX). We used the peroxidase/diaminobenzidine (DAB) method with nickel intensification to visualize e gr 1 p rotein Sections were immersed in 0.07% DAB (Sigma, St. Louis, MO) in PBS with 0.1% nickel chloride and 0.03% H 2 O 2 and developed for 2 minutes until antibody antigen binding sites were stained black. The reaction was stopped in PBS. Rinsed sections were mounted onto slides, ai r dried overnight, dehydrated in serial alcohol dilutions, and coverslipped from xylene with Eukitt (Calibrated Instruments, Hawthorne, NY). A nalysis of Egr 1 Immunoreactivity The auditory cortex and medial geniculate were identified using landmarks of Sc otinomys (Campbell et al. 2009), supported by the mouse and rat atlas es (Paxinos and Watson, 1998 ; Paxinos and Franklin, 2001; approx. Bregma 3.16 mm) We selected at random 2 3 alternate sections per region. Projected images of s ections were visualized at 20X using a n Axiocam Zeiss light microscope mounted to a CCD
82 camera ( Leica DFC 310FX ) and interfaced with a computer. Automatic adjustment of light intensity normalized background levels. A 350 x 350 m box was placed on digitized sections for each individual and images were constructed in Adobe Illustrator CS3 (San Jose, CA, v. 13.0.0). A researcher blind to experimental conditions quantified the number of egr 1 stained nuclei. Statistical Analyses W e used the Scheirer R ay Hare technique (Sokal and Ro h lf 1995) to test for significance of playback treatment on response (post pre stimulus) between sympatric and allopatric populations of S. teguina Analysis of variance was used to assess differences i n time spent near speakers broadcasting the stimuli and to test for differences among treatment groups in egr 1 induction. We conducted post hoc Averages are reported SE. Results Response of S. teguina to heterospecific st imuli depended upon whether they were sympatric or allopatric with S. xerampelinus (T reatment x population interaction, H = 12.36, P < 0.002). S. teguina decreased singing after heterospecific playback in sympatry (Pre: 2.13 0 .35; Post: 1.0 0.4 songs ) but incre ased singing in allopatry (Pre: 0.38 0 .18 songs; Post 1.38 0.26 songs; P < 0.05; Figure 5 2). Similarly, animals spent less time near heterospecific stimuli in sympatry (Control: 43.8 11.2%; Conspecific: 73.3 5.8%; Heterospecific: 11.8 6 .0%) but spent similar amounts of time near conspecific and heterospecific stimuli in allopatry compared to controls (Control: 33.1 6.9%; Conspecific: 64.7 9.2%; Heterospecific: 56.2 6.6%; P < 0.05; Figure 5 2).
83 Both populations exhibited significan t induction of egr 1 in response to conspecific songs when compared to control stimuli Surprisingly, heterospecific songs elicited stronger responses in sympatry than in allopatry despite the fact that these songs elicit les s singing behavior in sympatry (Figure 5 3). Discussion We found that allopatric S. teguina responded to heterospecifics, whereas sympatric animals decreased singing. Behavioral responses were reflected in neural auditory tuning. Such agonistic character displacement in social cogniti on reflects emancipation from interspecific aggression in sympatry. Our data suggest that the auditory system can increase its response to heterospecific vocalizations in ecologically relevant contexts Recent work demonstrates that both reward and punish ment can increase mammalian auditory responses to sounds associated with either positive or negative outcomes (e.g. Resnik et al. 2011) We suggest that differences in aggressive experience with heterospecifics can modify auditory tuning to match ecologica lly important stimuli. Results from reciprocal transplant experiments, where allopatric S. teguina experience heterospecific aggression paired to S. xerampelinus song and sympatric S. teguina are removed from interspecific aggression, will help elucidate t he role of learning in shaping brain and behavioral responses.
84 Figure 5 1. Altitudinal distributions of Neotropical singing mice ( Scotinomys ) at our study site in Costa Rica. S. xerampelinus (blue) replaces S. teguina (red) on Volcn Ir az (VI). S. teguina is allopatric on Cerros La Carpintera (LC).
85 Figure 5 2. Response to playback trials in the laboratory at Volcn Iraz, Costa Rica (sympatry) and Cerros La Carpintera (allopatry). A) Number of songs produced by S. teguina pre (sha ded bars) and post (open bars) playback of control, conspecific, and heterospecific stimuli. B) Percent time spent near speakers broadcasting sound stimuli. Error bars are 1 s.e. P < 0.05. Letters not sharing the same letter are significantly different P < 0.05.
86 Figure 5 3. Neuronal immediate early gene expression in response to playbacks in the laboratory at Volcn Iraz (sympatry) and Cerros La Carpintera (allopatry), Costa Rica. A) Brain regions and representative Egr 1 expression in response to playback of stimuli. B) Egr 1 expression in the medial geniculate. C) Egr 1 expression in the auditory cortex. Error bars are 1 s.e. Letters not sharing the same letter are significantly different.
87 CHAPTER 6 CONCLUSIONS My dis sertation addressed proxim ate and ultimate factors mediating acoustic communication in Neotropical singing mice ( Scotinomys ) with the goal of understanding how biotic and abiotic factors limit their altitudinal distributions. The work incorporated biogeographical surveys, long term environmental monitoring, reciprocal removal experiments, behavioral trials, pla yback experiments in the field and laboratory, hormonal manipulation, and analysis of neuronal immediate early gene expression. The integrative nature of the study enabled me to link changes in perception at the neural level to changes in distribution at the geographic level. I highlight my findings below. Androgens are an important class of steroid hormones involved in modulating the expression and evolution of male secondary sex characters. Vocalizations used in the context of aggression and mate attraction are among the most elaborate and diverse androgen dependent animal displays as reflected in a rich tradition of studies on bird song and anuran calls. Male g mice ( S. teguina ) commonly emit trilled songs that appear to function in male male aggression. In this study, we experimentally manipulated androgens in singing mice to assess their role in modulating aggression and song effort. Testosterone and DHT tre ated animals retained aggressive and song attributes similarly. However, castrated mice administered empty implants showed more subordinate behavior and sang fewer songs that were shorter, lower in power, higher in frequency, and less stereotyped. The ext ensive effects of androgens on a suite of phenotypes highlight their role in linking gonadal status with decisions about investment in reproductive behaviors.
88 Vocalizations used in aggressive and mating contexts often convey reliable information about sign aler condition when physical or physiological limitations constrain signal expression. In vertebrates, androgens modulate the expression of vocal signals and provide a proximate link between male condition and signal form. In many songbirds, assessment of males is based on production of trills that are constrained by a performance trade off between how fast notes are repeated and the frequency bandwidth of each note. In this study, we recorded trills of male Neotropical singing mice to ask if they exhibit ed a similar performance trade off, and then manipulated androgen levels to assess their role in modulating vocal performance. Lastly, we broadcast experimentally manipulated trills to females to determine if they preferred versions resembling those of and rogen treated males. Singing mice exhibit a vocal performance trade off similar to birds. Males treated with androgens maintained vocal performance, but castrated mice administered empty implants produced trills with lower performance. Females approached h igh performance trills more rapidly and spent more time near corresponding speakers. Together, our results demonstrate that androgens modulate the production of physically challenging vocalizations, and the resulting signal variation influences female rece iver response. Interspecific aggression between ecologically similar species may influence geographic limits by mediating competitive exclusion at the range edge. Within species, male advertisement vocalizations mediate competitive and dominance interactio ns. Between species, vocalizations may likewise contribute to dominance interactions and influence species distributions. We studied the mechanisms underlying altitudinal range limits in Neo tropical singing mice a genus of muroid rodent in which males voc alize to
89 repel rivals and attract mates We first delineated replacement zones and described temperature regimes on three mountains in Costa Rica and Panam where Chiriqu singing mice ( S. xerampelinus S. teguina ). Next, we conducted interspecific behavioral trials and reciprocal removal experiments to test whether interspecific aggression mediated species replacement. Finally, we performed reciprocal playback experiments to investigate whether response to song match ed competitive interactions. Biogeographical surveys suggest that S. xerampelinus may be limited by temperature tolerance. Behavioral trials and removal experiments indicate that S. xerampelinus is behaviorally dominant and excludes S. teguina from higher altitudes. P layback experiments demonstrated that S. teguina is silenced and repelled by heterospecific song, whereas S. xerampelinus responded to heterospecifics. Thus, response to song reflects underlying dominance and suggests that interspecific communi cation mediates altitudinal zonation of ecologically similar congeners. Species recognition touches on deep themes in diverse domains, ranging from the origin of species to the nature of social cognition. While evolutionary and ecological aspects of specie s recognition have received much attention, the neural substrates underlying receiver decisions are far less explored. When related species occur in sympatry receiver responses to heterospecific signals are often altered i n comparison to allopatric popula tions Such character displacement in communication presents an opportunity to explore ecological variation in the mechanisms of social cognition. Among animals that use acoustic signals to identify conspecifics, such as insects and anurans, sympatric spec ies often shift the tuning of their auditory systems to reduce responses to
90 heterospecific signals, a phenomenon known as a sensory filter In principle, it is possible that animals could preserve the function of peripheral sensory systems, but assign sp ecies specific stimuli different motivational values, a phenomenon we refer to s singing mouse ( Scotinomys teguina ) Neotropical singing mice are diurnal insectivorous ro dents that inhabit montane cloud forests throughout Central America. Males commonly emit rapid frequency modulated trills that function in male male aggression and female mate choice. At some sites, S. teguina co occurs with the larger behaviorally domina nt Chiriqui singing mouse ( S. xerampelinus ). We performed playback studies in the lab and field using male S. teguina from populations that co occurred with S. xerampelinus (sympatric) and those that did not (allopatric). We found that allopatric male S. teguina sang back in response to both S. teguina and S. xerampelinus songs. In contrast, male S. teguina from populations sympatric with dominant S. xerampelinus seem to suppress song in response to heterospecific playback To examine whether species speci fic responses were modulated by the tuning of auditory system, we examined song evoked induction of egr 1 an immediate early gene often used in the study of bird song, in the medial geniculate and auditory cortex of S. teguina males. Both populations exhi bited significant induction of egr 1 in response to conspecific songs when compared to control stimuli. Surprisingly, we found that heterospecific songs elicited stronger responses in sympatry than in allopatry, despite the fact that these songs elicit les s singing behavior in sympatry. These data refute the sensory filter hypothesis for this species, and suggest that the auditory system increases its response to heterospecific vocalizations. Recent
91 work demonstrates that both reward and punishment can incr ease mammalian auditory responses to sounds associated with either positive or negative outcomes. We suggest that differences in aggressive experience with heterospecifics can modify auditory tuning to match ecologically important stimuli.
92 LIST OF REFERENCES Adkins Regan, E. 2005 Hormones and animal social behavior. Princeton University Press, Princeton, NJ. Akcam, E., T. Bolu, A. Merati C. Durmus M. Gerek and Y. Ozkaptan. 2004. Voice changes after androgen thera py for hypogonadotrophic hypogonadism. Laryngoscope 114: 1587 1591. Amzquita, A., W. Hdl L. Castellano s, A. Lima, L. Erdtmann, and M. C. Ara jo 2006. Masking interference and the evolution of the acoustic communication system of the Amazonian frog Ep ipedobates femoralis Evolution 60: 1874 1887. Anderson C N ., and G. F. Grether. 2010 Character displacement in the fighting colours of Hetaerina damselflies. Proc eedings of the Royal Society B: Biological Sciences 277: 3669 3675. Andersson, M. and L. W. Simmons. 2006. Sexual selection and mate choice. Trends in Ecology and Evolution 21: 296 302. Apicella, C. L., D. R. Feinberg, and F. W. Marlowe. 2007. Voice pitch predicts reproductive success in male hunter gatherers. Biology Letters 3: 682 684. Ball, G. F., and J. Balthazart. 2008. Individual variation and the endocrine regulation of behaviour and physiology in birds: a cellular/molecular perspective. Philosophical Tran sactions of the Royal Society B: Biological Sciences 363: 1699 1710. B all, G. F., C. B. Castelino D. L. Maney D. Appeltants, and J. Balthazart. 2003. The activation of birdsong by testosterone: Multiple sites of action and role of ascending catecholamine projections. Pages 211 231 in G. C. Panzica, and R. Melcangi eds. Steroids and the Nervous System. Annals of the New York Academy of Sciences. Ballentine, B., J. Hyman, and S. Nowicki 2004. Vocal performance influences female response to male bird song: an experimental test. Behavioral Ecology 15:1 63 168. Ballentine B. 2009. The ability to perform physically challenging songs predicts age and size in male swamps sparrows ( Melospiza georgiana ). Animal Behaviour 77: 973 978. Bass, A. H., and F. Ladich. 2008. Vocal acoustic communication: From behavior to neurons. P ages 253 278 in A. Popper, R. Fay, and J. Webb, eds. Fish Bioacoustics. Springer, New York.
93 Bass, A. H., Remage Healey, L., 2008. Central pattern generators for social vocalization: androgen dependent neur ophysiological mechanisms. Hormones and Behavio r 53: 659 672. Beach, F.A., 1948. Hormones and Behavior. Paul B. Hoeber, New York. Beani, L., G. Panzica F. Briganti, P. Persichella, and F. Dess Fulgheri 1995. Testosterone induced changes of call structure midbrain and syrinx anatomy in partridges Physiology and Behavior 58:1 149 1157. Beckford, N. S., D. Schain S. R. Roor, and B. Schanbacher 1985 Androgen stimulation and laryngeal development. Annals of Otology, Rhinology, and Laryngology 94: 634 640. Beebee, M.D. 2004. Variation in vocal pe rformance in the songs of a wood warbler: evidence for the function of distinct singing modes Ethology 110: 531 542. Benjamini Y., and Y. Hochberg. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Roya l Statistical Society Series B 57:289 300. Bensch, S., and D. Hasselquist. 1991. Evidence for active female choice in a polygynous warbler. Animal Behavior 44: 301 311. Blackburn, T. M., J. H. Lawton, and J. N. Perry. 1992. A method of estim ating the slope of upper bounds of plots of body size and abundance in natural animal assemblages. Oikos 65: 107 112. Blondel, D. V. 2004. Social organization of a species of singing mouse, Scotinomys xerampelinus M.S. thesis, University of Florida. Bl ondel, D. V., J. Pino, and S. M. Phelps. 2009. Space use and social structure of long tailed singing mice ( Scotinomys xerampelinus ). Journal of Mammalogy 90:715 723. Botero, C. A., R. J. Rossman L. M. Caro L. M. Stenzler I. J. Lovette S. R. de Kort, and S. L. Vehrencamp. 2009. Syllable type consistency is related to age, social status and reproductive success in the tropical mockingbird. Animal Behaviour 77: 701 706. Bovbjerg, H. 1970. Ecological isolation and competitive exclusion in two crayfish ( Orconectes virillis and Orconectes immunis ). Ecology 51 :225 236. Bowers, M. A. 1985. Experimental analysis of competition between two species of bumble bees (Hymenoptera: Apidae). Oecologia 67 :224 230. Bradbury, J.W., and S. L. Vehrencamp. 1998. Princ iples of animal communication. Sinauer Associates, Inc., Sunderland, MA.
94 Brown, J. H. 1971. Mechanisms of competitive exclusion between two species of chipmunks. Ecology 52:305 311. Brudzynski, S. M. 2010. Handbook of Mammalian Vocalization: An Integra tive Neuroscience Approach. Academic Press, London. Brumm, H., and D. Todt. 2004. Male male vocal interactions and the adjustment of song amplitude in a territorial bird. Animal Behaviour 67: 281 286. Bull, C. M. 1991. Ecology of parapatric distribution s. Annual Review of Ecology and Systematics 22:19 36. Byers, J., E. Hebets, and J. Podos. 2010. Female mate choice based upon male motor performance Animal Behaviour 79: 771 778. Calder, W. A. 1970. Respiration during song in the canary ( Serinus canaria ) Comparative Biochemistry and Physiology 32: 251 258. Cadena, C. D. 2007. Testing the role of interspecific competition in the evolutionary origin of elevational zonation : An example with Buarremon brush finches (Aves, Emberizidae) in the neotropical mountains. Evolution 61:1120 1136. Cadena, C. D., and B. Loiselle. 2007. Limits to elevational distributions in two species of emberizine finches: disentangling the role of interspecific competition, autoecology, and geographic variation in the environme nt. Ecography 3 0:491 504. Campbell, P., B. Pasch, J. L. Pino, O. L. Crino, M. Phillips, and S. M. Phelps. 2010. Geographic variation in the songs of Neotropical singing mice: testing the relative importance of drift and local adaptation. Evol ution 64:19 55 1972. Campbell, P. R. L. Reep, M. L. Stoll, A. G. Ophir, and S. M. Phelps. 2009. Conservation and diversity of Foxp2 expression in muroid rodents: Functional I mplications. Journal of Comparative Neurology 512:84 100. Capranica, R. R. 1965. The evoked vocal response of the bullfrog: A s tudy of communication by s ound MIT Press, Cambridge, MA. Capranica, R. R. and Moffat, A. J. M. 1983. Neurobehavioral correlates of s ound communication in anurans. Pages 701 730 in J. P. Ewert, R R. Capranica, and D. J. Ingle, eds. Advances in Vertebrate Neuroethology Plenum Press New York. Carothers, J. H., F. M. Jaksic, and P. A. Marquet. 2001. Altitudinal zonation among lizards of the genus Liolaemus : questions answered and unanswered questions. Revista Chilena de Historia Natural 74:313 316.
95 Catchpole, C. K., and B. Leisler. 1986. Interspe cific territorialism in Reed Warblers: A local effect revealed by playback experiments. Animal Behaviour 34:299 300. C atchpole, C. K., B. Leisler and J. Dittami. 1986. Sex ual differences in the response of captive great reed warblers ( Acrocephalus arundinaceus ) to variation in song structure and repertoire size. Ethol ogy 73: 69 77. Caro, S. P., K. B. Sewall K. G. Salvante, and K. W. Sockman. sparr ows modulate their behavior in response to variation in male song quality. Behavioral Ecology 21: 562 569. Charlton, B. D., J. L. Keating, D. Kersey, L. Rengui, Y. Huang, and R. R. Swaisgood. 2011 Vocal cues to male androgen levels in giant pandas. Biol ogy Letters 7:71 74. Chappell, M. A. 1978. Behavioral factors in the altitudinal zonation of chipmunks ( Eutamias ). Ecology 59 :565 579. Clayton, N. and E. Prove 1989. Song discrimination in female zebra finches and Bengalese finches. Animal Behaviour 38: 352 354. Collins, S. A., 2004. Vocal fighting and flirting: the functions of birdsong. Pages 39 79 in P. Marler, and H. Slabbekoorn, eds. Elsevier Academic Press, San Diego Colwell, R. K., and E. R. Fuente s. 1975. Experimental studies of the niche. Annual Review of Ecology and Systematics 9:281 310. Colwell, R. K., G. Brehm, C. L. Cardels, A. C. Gilman, and J. T. Longino. 2008. Global warming, elevational range shifts, and lowland biotic attrition in th e wet tropics. Science 322:258 261. Connaughton, M. A., M. L. Fine and M. H. Taylor. 1997. The effects of seasonal hypertrophy and atrophy on fiber morphology, metabolic substrate concentration and sound characteristics o f the weakfish sonic muscle. J ournal of Experimental Biology 200: 2449 2457. Connell, J. H. 1961.The influence of interspecific competition and other factors on the distribution on the barnacle Chthamalus stellatus. Ecology 42:710 723. Connell, J. H. 1983. On the prevalence and relat ive importance of interspecific competition: evidence from field experiments. American Naturalist 122:661 696. Crews, D., V. Traina, F. T. Wetzel and C. Muller. 1978. Hormonal control of male reproductive behavior in the lizard, Anolis carolinensis : Ro le of testosterone, dihydrotestos terone, and estradiol. Endocrinology 103: 1814 1821.
96 Cynx, J., N. J. Bean, and I. Rossmann. 2005. Testosterone implants alter the frequency range of zebra finch songs Hormones and Behavior 47: 446 451. Dabbs, J. M., Jr., and A. Mallinger, 1 999. High testosterone levels predict low voice pitch among men. Personality and Individual Differences 27: 801 804. Davis, A. J., J. H. Lawton, B. Shorrocks, and L. S. Jenkinson. 1998. Individualistic species responses invalidate simp le physiological models of community dynamics under global environmental change. Journal of Animal Ecology 67 :600 612. de Kort, S. R., E. R. B Eldermire, S. Valderrama C. A. Botero, and S. L. Vehrencamp, 2009. Trill consistency is an age related asses sment signal in banded wrens. Proceedings of the Royal Society B: Biological Sciences 276: 2315 2321. Diamond, J. 1970. Ecological consequences of island colonization by Southwest Pacific birds. I. Types of niche shifts. Proceeding s of the National Acade my of Sciences of the USA 67:529 536. Diamond, J. M. 1973. Distributional ecology of New Guinea birds. Science 179:759 769. Dickman, C. R. 1986a. An experimental study of competition between two species of dasyurid marsupials. Ecological Monographs 56: 221 241. Dickman, C. R. 1986b. An experimental manipulation of the intensity of interspecific competition: effects on a small marsupial. Oecologia 70:536 543. Dickman, C. R. 1988. Body size, prey size, and community structure in insectivorous mammal s. Ecology 69:569 580. Dickman, C. R. 1991. Mechanisms of competition among insectivorous mammals. Oecologia, 85:464 471. Draganoiu, T. I., L. Nagle, and M. Kreutzer. 2002. Directional female preference for an exaggerated male trait in ca nary ( Serinus canaria ) song. Proceedings of the Royal Society B: Biological Sciences 269: 2525 2531. DuBois, A. L., S. Nowicki, and W. A. Searcy. 2008. Swamp sparrows modulate vocal performance in an aggressive context. Biology Letters 5 : 163 165. Dytham, C. 1999. Blackwell Science, Oxford.
97 Eason, J. M., G. Schwartz K. A. Shirley, and A. W. English. 2000. Investigation of sexual dimorphism in the rabbit masseter muscle showing different effects o f androgen deprivation in adult and young adult animals. Archives of Oral Biology 45: 683 690. Ehret, G., and D. B. Geissler. 2005 Communication call representation in the m ouse auditory cortex: Perception vs. r ecognition Pages 85 96 in J. Syka and M M. Merzenich, eds. Plasticity and Signal Representation in the A udi tory S ystem Springer, New York. Emerson, S. B., 2001. Male advertisement calls: Behavioral variation and physiological processes. Pages 36 44 in M. J. Ryan, ed. Anuran communication S mithsonian Institution Press, Washington, DC. Enders, R. K., and O. P. Pearson. 1939. Three new rodent subspecies of Scotinomys from Chiriqu, Panama. Notulae Naturae of the Academy of Natural Sciences of Philadelphia 34:1 4. Evans, S., N. Neave D. Wakelin and C. Hamilton. 2008. The relationship between testosterone and vocal frequencies in human males. Physiology and Behavior 93: 783 788. Fernandez Vargas, M. 2006. Olfactory responses of neotropical singing mice ( Scotinomys teguina ) to the odors of the mid ventral sebaceous gland M.S. thesis, University of Missouri, St. Louis. Fernandez Vargas, M., Tang Martinez, Z., Phelps, S.M., 2008. Factors influencing singing behavior in the male and female Neotropical short tailed singing mouse ( Scoti nomys teguina ). 12th International Behavioral Ecology Congress, Ithaca, NY. Fischer, L., D. Catz, and D. Kelley 1993. An androgen receptor mRNA isoform associated with hormone induced cell proliferation. Proceedings of the National Academy of Sciences of the USA 90: 8254 8258. Fitch, W. T. and M. D. Hauser. 2003. and the evolution of acoustic signals. Pages 65 137 in A. Simmons, A. Popper and R. Fay eds. Acoustic Communication Springer, New Y ork. Floody, O. R. 1981. The hormonal control of ultrasonic communication in rodents. Integrative and Comparative Biology 21: 129 142. Forstmeier, W., B. Kempenaers A. Meyer and B. Leisler 2002. A novel song parameter correlates with extra pair pater nity and Proceedings of the Royal Society B: Biological Sciences 269: 1479 1485.
98 Fusani, L., L. Beani, and F. Dessi Fulgheri. 1994. Testosterone affects the acoustic structure of the male call in the grey partridge ( Perdix perdix ). Behavio ur 128: 301 310. Galeotti, P., N. Saino R. Sacchi, and A. P. Mller 1997. Song correlates with social context, testosterone and body condition in male barn swallows. Animal Behaviour 53: 687 700. Gerhardt, H. C., and F. Huber. 2002. Acoustic communic ation in insects and anurans. University of Chicago Press, Chicago. Gerhardt, H. C., S. D. Tanner C. M. Corrigan, and H. C. Walton 2000. Female preference functions based on call duration in the gray tree frog ( Hyla versicolor ). Behavioral Ecology 1 1: 663 669. Ghalambor, C. K., R. B. Huey, P. R. Martin, J. J. Tewksbury, and G. Wang. 2006. Are and Comparative Biology 46:5 17. Gill, D. E. 1974. Intrinsic rate of incre ase, saturation density, and competitive ability. II. The evolution of competitive ability. American Naturalist 108:103 116. Glazier, D. S. 1980. Ecological shifts and the evolution of geographically restricted species of North American Peromyscus (mice ). Journal of Biogeography 7 :63 83. Glazier, D. S., and S. E. Eckert. 2002. Competitive ability, body size and geographical range size in small mammals. Journal of Biogeography 29:81 92. Goller, F. and R. A. Suthers. 1996. Role of syringeal muscles in gating air fl ow and sound production in singing brown thrashers. Journal of Neurophysiology 75: 867 876. Goodale, E., G. Beauchamp, R. D. Magrath, J.C. Nieh, and G. D. Ruxton. 2010. Interspecific information transfer influences animal community structu re. Trends in Ecology and Evolution 25:354 361. Grant, P.R. 1972. Interspecific competition among rodents. Annual Reviews of Ecology and Systematics 3 :79 106. Grether, G. F. 2011. The neuroecology of competitor recognition. Integrative and Comparativ e Biology 51:807 818. Grether G. F., N. Losin C. N. Anderson and K. Okamoto. 2009. The role of interspecific interference competition in character displacement and the evolution of competitor recognition. Biological Reviews 84:617 635.
99 Griffis, M. R ., and R. G. Jaeger. 1998. Competition leads to an extinction prone species of salamander: interspecific territoriality in a metapopulation. Ecology 79 :2494 2502. Hairston, N. G. 1980 a The experimental test of an analysis of field distributions: compe tition in terrestrial salamanders. Ecology 61:817 826. Hairston, N. G. 1980 b Evolution under interspecific competition: field experiments on terrestrial salamanders. Evolution 34:409 420. Hairston, N. G. 1981. An experimental test of a guild: salamand er competition. Ecology 61:65 72. Hamlin, H J., M. Milnes, C. M. Beaulaton, L. Albergott i and L. J. Guillette, Jr. 2011 Gonadal stage and sex steroid correlations in Siberian sturgeon, Acipenser baeri habituated to a semi tropical environment. Journal of the World Aquaculture Society 42: 313 320. Hammerschmidt, K., K. Radyushkin H. Ehrenreich, and J. Fischer 2009. Female Biology Letters 5: 589 592. Hartgens, F., and H. Kuip ers. 2004. Effects of androgenic anabolic steroids in athletes. Sports Medicine 34: 5213 554. Hartley, R. S. and R. A. Suthers 1989. Airflow and pressure during canary song: evidence for mini breaths. Journal of Comparative Physiology A 165: 15 26. Hawe s, M. L. 1977. Home range, territoriality, and ecological separation in sympatric shrews, Sorex vagrans and Sorex obscurus Journal of Mammalogy 58:354 367. Heller, H. G. 1971. Altitudinal zonation of chipmunks ( Eutamias ): interspecific aggression. Eco logy 52:312 319. Heller, H. C., and D. M. Gates. 1971. Altitudinal zonation of chipmunks ( Eutamias ): energy budgets. Ecology 5 2 :424 433. Hill, R. W., and E. T. Hooper. 1971. Temperature regulation in mice of the genus Scotinomys Journal of Mammalogy 5 2: 806 816. Hixon, M. A. 1980. Competitive interactions between California reef fishes of the genus Embiotica. Ecology 61 :918 931. Hoese, W. J., J. Podos N. C. Boetticher, and S. Nowicki 2000 Vocal tract function in birdsong production: experimental manipulation of beak movements. Journal of Experimental Biology 203: 1845 1855.
100 Hooper, E. T. 1972. A synopsis of the rodent genus Scotinomys Occasional Papers of the Museum of Zoology, University of Michigan 665:1 32. Hooper, E. T., and M. D. Carlet on. 1976. Reproduction, growth and development in two contiguously allopatric rodent species, genus Scotinomys Miscellaneous Publications of the Museum Zoology, University of Michigan 151:1 52. Humboldt, A. v. 1807. Ideen zu einer Geographie der Pflan zen nebst einem Naturgemlde der Tropenlnder. Bey F. G. Cotta, Tbingen, Germany. Hurd, P. L. and M. Enquist. 2005 A strategic taxonomy of biological communication. Animal Behaviour 70: 1155 1170. Illes, A. E., M. L. Hall, and S. L. Vehrencamp 2006. Vocal performance influences male receiver response in the banded wren. Proceedings of the Royal Society of London B: Biological Sciences 27 3: 907 1912. James, P. J., J. G. Nyby, and G. A. Saviolakis. 2006. Sexually stimulated testosterone releas e in male mice ( Mus musculus ): Roles of ge notype and sexual arousal. Hormones and Behavior 50: 424 431. Jaeger, R. G. 1971 a Competitive exclusion as a factor influencing the distributions of two species of terrestrial salamanders. Ecology 52 :535 546. J aeger, R. G. 1971 b Moisture as a factor influencing the distributions of two species of terrestrial salamanders. Oecologia 6 :191 207. Janicke, T., S. Hahn, M. S. Ritz, and H. Peter 2008. Vocal performance reflects individual quality in a nonpasserine. Animal Behaviour 75: 91 98. Jankowski, J. E. S. K. Robinson, and D. J. Levey. 2010. Squeezed at the top: Interspecific aggression constrains elevational ranges in tropical birds. Ecology 91:1877 1884. Janzen, D. H. 1967. Why mountain passes are higher in the tropics. American Naturalist 101:233 249. Jenkins, J. S., 1998. The voi ce of the castrato. Lancet 351: 1877 1880. JMP, Version 7. SAS Institute Inc., Cary, NC, 1989 2007. Kao, M. H., and M. S. Brainard. 2006. Lesions of an avian basal ga nglia circuit prevent context dependent changes to song variability. J ournal of Neurophysiology 96: 1441 1455.
101 Ketterson, E. D., V. Nolan, Jr., M. J. Cawthorn P. G. Parker and C. Ziegenfus. 1996. Phenotypic engineering: using hormones to explore the mechanistic and functional bases of phenotypic variation in nature. Ibis 138: 1 17. King, A. M., J. Ashby and C. Nelson. 2001. Effects of testosterone replacement on a male professional singer. Journal of Voice 15: 553 557. Kime, N. M., A. S. Rand M. Ka pfer, and M. J. Ryan. 1998. Consistency of female choice in the tngara frog: a permissive preference for complex characters Animal Behaviour 55: 641 649. Klump, G. M., and H. C. Gerhardt. 1987. Use of non arbitrary acoustic criteria in mate choice by femal e gray tree frogs. Nature 326: 286 288. Kroodsma, D. E., and E. H. Miller. 1996. Ecology and Evolution of Acoustic Communication in Birds. Cornell University Press, Ithaca. Lack, D., and H. N. Southern. 1949. Birds on Tenerife. Ibis 91:607 626. Lam brechts M. M 199 6. Organization of birdsong and constraints on performance. Pages 305 320 in D. E Kroodsma, and E. H. Miller, eds. Ecology and Evolution of Acoustic Communication in Birds. Cornell University Press, New York. Leary, C. J. T. S. Jesso p, A. M. Garcia, and R. Knapp. 2004. Steroid hormone profiles and relative body condition of calling and satellite toads: implications for proximate regulation of behavior in anurans. Behavioral Ecology 15: 313 320. Lenoir, J., J. C. Ggout, P. A. Marquet P. de Ruffray, and H. Brisse. 2008. A significant upward shift in plant species optimum elevation during the 20th century. Science 320:1768 1771. Lessells, C. M., and P. T. Boag. 1987. Unrepeatable repeatabilities a common mistake. Auk 104: 116 121. Littell, R. C., W. W. Stroup, and R. J. Freund. 2002. SAS for linear models, Fourth edition SAS Institute, Inc. Cary, North Carolina, USA. Lyons, G. E., A. M. Kelly, and N. A. Rubenstein 1986. Testosterone induced changes i n contractile protein isofo rms in the sexually dimorphic temporalis muscle of the guinea pig. Journal of Biological Chemistry 261: 13278 13284. Mager, J. N., C. Walcott and W. H. Piper. 2007. Male common loons, Gavia immer communicate body mass and condition through dominant freq uencies of territorial yodels. Animal Behavior 73: 683 690.
102 Malmierca M S M. A. Izquierdo S. Cristaudo O. Herna ndez, D. Pe rez Gonzalez E. Covey, and D. L. Oliver. 2008. A discontinuous tonotopic organization in the in ferior colliculus of the rat Journal of Neuroscience 28:4767 4776. Marler, C., and M. J. Ryan. 1996 Energetic constraints and steroid hormone correlates of male calling behaviour in the tungara frog Journal of Zoology (London) 240: 397 409. Marris, E. 2007. The escalator effect. Nature Reports Climate Change doi:10.1038/climate.2007.70 Martin, P. R., J. R. Fotheringham, L. Ratcliffe, and R. J. Robertson. 1996. Response of American redstarts (suborder Passeri) and least flycatchers (suborder Tyranni) to heterospecific playbac k: the role of song in aggressive interactions and interference competition. Behavioral Ecology and Sociobiology 39:227 235. Martin, P. R., and T. E. Martin 2001a. Ecological and fitness consequences of coexistence in two congeneric wood warbler s (Parulidae: Vermivora): a removal experiment. Ecology 82:189 206. Martin, P. R., and T. E. Martin. 2001b. Behavioral interactions between coexisting species: song playback experiments with wood warblers. Ecology 82:207 218. Maynard Smith, J. and D. G C. Harper 1995. Animal signals: Models and terminology. Journal of Theoretical Biology 177: 305 311. Mayr, E., and J. M. Diamond. 1976. Birds on islands in the sky: origin of the montane avifauna of Northern Melanesia. Procee ding of the National Academ y of Sciences of the USA 73:1765 1769. the tropics. Ecology Letters 12:550 560. Merriam, C. H. 1895. The geographic distribution of animals and plants in North Amer ica. Yearbook of the United States Department of Agriculture 1894: 203 214. Miller, J. R. and M. D. Engstrom. 2007. Vocal stereotypy and singing behavior in baiomyine mice. Journal of Mammalogy 88: 1447 1465. Miller R. S. 1964. Ecology and distributio n of pocket gophers (Geomyidae) in Colorado. Ecology 45 :256 272. Miller, R. S. 1967. Pattern and processes in competition. Advances in Ecological Research 4:1 74
103 Moore, F. L., S. K. Boyd, and D. B. Kelley 2005. Historical perspective: hormonal regu lation of behaviors in amphibians. Hormones and Behavior 48: 373 383. Moritz, C., J. L. Patton, C. J. Conroy, J. L. Parra, G. C. White, and S. R. Beissinger. 2008. Impact of a century of climate change on small mammal communities in Yosemite National Park USA. Science 322:261 264. Morse, D. H. 1974. Niche breadth as a function of social dominance. American Naturalist 108:818 830. Morton, M. L., L. E. Peterson D. M. Burns, and N. Allan. 1990 Seasonal and age related changes in plasma t estosterone lev els in mountain w hite crowned sparrows. Condor 92 : 166 173 Murray, B. G. 1971. The ecological consequences of interspecific territorial behavior in birds. Ecology 52:414 423. Murray, B. G. 1981. The origins of adaptive interspecific territorialism. Biol ogical Reviews 56:1 22. Nelson, B. S., G. J. L. Beckers, and R. A. Suthers 2005. Vocal tract radiation in a songbird. Journal of Experimental Biology 208: 297 308. Nelson, R.J., 2005. An Introduction to Behavioral Endocrinology. 3rd edition Sinauer Associates, Inc. Sunderland, MA Neubauer, R. L. 1999. Super normal length so ng preferences of female zebra finches ( Taeniopygia guttata ) and a theory of the evolution of bird song. Evolutionary Ecology 13: 365 380. Nowicki, S., M. W. Westneat, and W. J. Hoese 1992. Birdsong: motor function and the evolution of communi cation. Seminars in Neuroscience 4: 385 390. Nunez, A. A., J. Nyby, and G. Whitney. 1978. The effects of testosterone, estradiol, and dihydrotestosterone on male mouse ( Mus musculus ) ultrasonic vocalizations. Horm ones and Behavior 11: 264 272. N yby, J., J. A. Matochik, and 1992. Intracranial androgenic and estrogenic stimulation of male typical behaviors in house mice ( Mus domesticus ). Horm ones and Behavior 26: 24 45. Oliveira, R. F. 2004. Social modulation of androgens in vertebrates: mec hanisms and function. Advances in the Study of Behavior 3 4: 165 219. Ord, T. J., and J. A. Stamps. 2009. Species identity cues in animal communication. American Naturalist 174:585 593.
104 Orians, G. H., and M. F. Willson. 1964. Interspe cific territories of birds. Ecology 45: 736 745. Oyegbile, T. O., and C. A. Marler 2005. Winning fights elevates testosterone levels in California mice and enhances future ability to win fights. Hormones and Behavior 48: 259 267. Parker, G. A. 1974. Assessment strategy an d the evolution of fighting behaviour. Journal of Theoretical Biology 47: 223 243. Pasch, B., A. S. George, P. Campbell, and S.M. Phelps 2011a. Androgen dependent male vocal performance influences female preference in Neotropical singing mice. Animal Behaviour 82:177 183. Pasch, B., A. S. George, H. J. Hamlin, L. J. Guillette, Jr., and S.M. Phelps. 2011b. Androgens modulate song effort and aggression in Neotropical singing mice. Hormones and Behavior 59: 90 97. Paxinos, G and K. B. J. Franklin. 2001. The mouse brain in stereotaxic coordinates, 2 nd ed ition Academic Press New York Paxinos G and C. Watson 1998. The rat brain in stereotaxic coordinates, 4th edition Academic Press, New York. Peiman, K. S., and B. W. Robinson. 2010. Ecology a nd evolution of resource related heterospecific aggression. Quarterly Review of Biology 85:133 158. Persson, L. 1985. Asymmetrical competition: are larger animals competitively superior? American Naturalist 126:261 266. Podos, J. 1997. A performance constraint on the evolution of trilled vocalizations in a songbird family (Passeriformes: Emberizidae). Evolution 51: 537 551. Podos, J., D. C. Lahti, and D. L. M oseley 2009. Vocal performance and sensorimotor learning in songbirds. Advances in the Stud y of Behavior 40: 159 195. Podos, J. and S. Nowicki. 2004. Performance limits on birdsong. Pages 318 342 in P. Marler, H. Slabbekoorn, eds. The Science of Bird Song Elsevier Academic Press California Prescott, D. R. C. 1987. Territoria l responses to song playback in allopatric and sympatric populations of Alder ( Empidonax alnorum ) and Willow ( E. traillii ) flycatchers. Wilson Bulletin 99:611 619. Prezant, D. J., M. L. Karwa H. H. Kim D. Maggiore, V. Chung, and D. E. Valentine 1997. Short and long term effects of testosterone on diaphragm in castrated and normal male rats Journal of Applied Physiology 82: 134 143.
105 Reby, D., K. McComb B. Cargnelutti C. Darwin W. T. Fitch, and T. Clutton Brock. 2005. Red deer stags use formants a s assessment cues during intrasexual agonistic interactions. Proceedings of the Royal Society of London, Series B 272: 941 947. Reed, W. L, M. E. Clark P. G. Parker, S. A. Raouf N. Arguedas D. S. Monk E. Snajdr, V. Nolan, Jr. and E. D. Ketterson. 200 6. Physiological effects on demography: A long term experimental study of testosterone's effects on fitness. American Naturalist 167: 667 683. Reid, F. A. 1998. A Field Guide to the Mammals of Central America and Southeast Mexico. Oxford University Press UK Remage Healey, L. and A. H. Bass 2007. Plasticity in brain sexuality is revealed by the rapid actions of steroid hormones Journal of Neuroscience 27: 1114 1122. Remsen, J. V., Jr., and W. S. Graves. 1995. Distribution patterns and zoogeography o f Atlapetes Brush Finches (Emberizinae) of the Andes. Auk 112:225 236. Repasky, R. R., and D. Schluter. 1994. Habitat distributions of wintering sparrows along an elevational gradient: tests of the food, predation and microhabitat structure hypotheses. Journal of Animal Ecology 63:569 582. Resnik, J., N. Sobel and R. Paz. 2011 Auditory aversive learning increases discrimination thresholds. Nature Neuroscience 14: 791 797. Riabinina, O., M. Dai, T. Duke, and J. T. Albert. 2011. Active process mediate s species specific tuning of Drosophila ears. Current Biology 21:658 664. Rice, J. 1978. Ecological relationships of two interspecifically territorial vireos. Ecology 59:526 538. Riede, T., R. A. Suthers N. H. Fletcher and W. E. Blevins. 2006. Songbi rds tune their vocal tract to the fundamental frequency of their song. Proceedings of the National Academy of Sciences of the USA 103: 5543 5548. Robinson, S. K., and J. Terborgh. 1995. Interspecific aggression and habitat selection by Amazonian birds Journal of Animal Ecology 64 :1 11. Ryan, M. J., and A. Keddy Hector 1992. Directional patterns of female mate choice and the role of sensory biases. American Naturalist 139: S4 S35. Rychlik, L., and R. Zwolak. 2006 Interspecific aggression and behaviour al dominance among four sympatric species of shrews. Canadian Journal of Zoology 84:434 448.
106 Sales, G. D. 1972. Ultrasound and mating behavior in rodents with some observations on other behavioural situations. Journal of Zoology (London) 168: 149 164. SAS Institute. 2008. SAS/STAT users guide, version 9.2. SAS Institute, Cary, North Carolina, USA. Schoener, T. W. 1983. Field experiments on interspecific competition. American Naturalist 122:240 285. Shreve, F.1915. The vegetation of a desert mountain range as conditioned by climatic factors. Carnegie Institute of Washington, Publication No. 217. Shreve, F. 1922. Conditions indirectly affecting vertical distribution on desert mountains. Ecology 3:269 274. Sokal, R. R., and F. J. Rohlf. 1995. Biome try. 3rd ed ition. W. H. Freeman, San Francisco. Suthers, R. A., and F. Goller. 1997. Motor correlates of vocal diversity in songbirds. Pages 235 288 in J. Nolan, ed. Current Ornithology Plenum Press, New York. Suthers, R. A., F. Goller, and C. Pytte. 1999. The neuromuscular control of birdsong. Philosophical Transactions of the Royal Society B : Biological Sciences 354: 927 939. Taylor, A. M., and D. Reby 2010. Contribution of the source filter theory to the study of mammal vocal communication. Jour nal of Zoology (London) 280: 221 236 ten Cate, C., and C. Rowe. 2007. Biase s in signal evolution: learning makes a difference Trends in Ecology and Evolution. 22:380 387. Terborgh, J. 1971. Distributions on environmental gradients: theory and a prelimi nary interpretation of distributional patterns in the avifauna of the Cordillera Vilcabamba, Peru.Ecology 52:23 40. Terborgh, J., and J. S. Weske. 1975. Role of competition in the distribution of Andean birds. Ecology 56:562 576. Tewksbury, J. J., B. H. Raymond, and C. A. Deutsch. 2008. Putting the heat on tropical animals. Science 320:1296 1297. Twomey, E., V. Morales, and K. Summers. 2008. Evaluating condition specific and asymmetric competition in a species distribution context. Oikos 117:1175 11 84. Vallet, E. M., M. L. Kreutzer I. Beme, and L. Kiosseva 1997. Sexy syllables in male canary songs: honest signals of motor constraints on male vocal production? Advances in Ethology 32: 132.
107 Vehrencamp, S. L. 2000. Handicap, index, and conven tional signal elements of bird song. Pages 277 300 in Y. Espmark, T. Amundsen and G. Rosenqvis t, eds. Animal Signals: Signalling and Signal Design in Animal Communication Tapir Academic Press, Trondheim Vleck, C. M. and J. L. Brown. 1999. Testosterone and s ocial and reproductive behaviour in Aphelocoma jays Animal Behaviour 58: 943 951. Westneat, M. W., J. H. Long W. Hoese, and S. Nowicki 1993. Kinematics of birdsong functional correlation of cranial movements and acoustic features in sparrows. Journal of Experimental Biology 182: 147 171. Welch, A. M., R. D. Semlitsch and H. C. Gerhardt 1998. Call duration as an indicator of genetic quality in male gray tree frogs. Science 20: 1928 1929. Wilczynski, W., K. S. Lynch 2005. Current research in amphibians: studies integrating endocrinology, behavior and neurobiology. Hormones and Behavior 48: 440 450. Wilson, D., and D. M. Reeder. 2005. Mammal species of the world. A taxonomic and geographic reference, 3rd ed itio n. Johns Hopkins University Press, Baltimore. Wingfield, J. C. 1985. Short term changes in plasma levels of hormones during establishment and defense of a breeding territory in male song sparrows, Melospiza melodia Hormones and Behavior 19: 174 187. Wingfield, J. C. 19 87. Changes in reproductive function in free living birds in response to environmental perturbations. Pages 121 148 in M. Stetson, ed. Processing of environmental information in vertebrates Springer, New York. Wingfield, J. C., and M. C. Moore 1987. Hor monal, social, and environmental factors in the reproductive biology of free living male birds. Pages 149 175 in D. Crews, ed. Psychobiology of Reproductive Behavior: An Evolutionary Perspective Prentice Hal l, New Jersey. Wingfield, J. C., R. E. Hegner A. M. Dufty Jr., and G. F. Ball systems and breeding strategies. American Naturalist 136 : 829 846. Woolley, S. C., and A. J. Doupe. 2008. Social c ontext induced song variation affects female behavior and gene expression. PLoS Biology 6: 525 537. Wyman, M.T., M. S. Mooring B. McCowan M. C. T. Penedos and L. A. Hart. 2008. Amplitude of bison bellows reflects male quality, physical condition and motivation. Anim al Behaviour 76: 1625 1639.
108 Yamaguchi, A. and D. B. Kelley 2002. Hormonal mechanisms of acoustic communication. Pages 275 323 in A. Simmons, A. Popper eds. Acoustic Communication Springer, New York
109 BIOGRAPHICAL SKETCH Bret w as born in Point Pleasant, New Jersey and grew up in nearby Whiting on the northern terminus of the Pine Barrens. His exploration of the fauna inhabiting the forest, streams, and bogs served as initial inspiration for his work in biology. He attended Manch ester Township High School before enrolling as a biology major at Ursinus College, Collegeville, PA. Two courses made a significant impact on his development: Biology of tropical rainforests in Costa Rica brought him to the tropics for the first time, and Mammalogy served to formalize his interest in the study of organismal biology. Under the mentorship of Dr. Ellen Dawley, Bret conducted independent research on salamanders and tree squirrels Dr. Joh n Koprowski at the University of Arizona in Tucson. Here he spent many days exploring the sky islands of the desert southwest, with most time in pursuit of Apache fox squirrels in the Chiricahua Mountains. His time in Arizona was capped off by enrollment i n a flagship course in tropical biology through the Organization for Tropical Studies, where he first captured and recorded Neotropical singing mice in the Talamancas of Costa Rica. He went on to enroll in the Department of Biology at the University of Flo rida and spent most of his time in the mountains of Panam and Costa Rica studying the biology of Neotropical singing mice. Bret currently resides in Gainesville, FL with his wife Kate, his daughter Mary, and their two blue dogs, Arlo and Badger.