The acoustic structure of chimpanzee pant-hooting facilitates chorusing

Duetting or chorusing behaviour occurs in a wide variety of animals and is posited to fulfil various important functions including territory defence and social bonding. The structure of calls produced in choruses might be shaped in a way that facilitates such joint vocal displays. In this study, we test the hypothesis that flexibility to modify the temporal structure of chimpanzee pant-hoots, vocalisations often given jointly with other individuals, facilitates chorusing. The results of this study, which was conducted on two communities of wild chimpanzees in Uganda, support this hypothesis. First, the duration of the build-up phase of the pant-hoot correlated with the latency with which the partner joined in the call, suggesting that males prolong the duration of the build-up to allow others to join in the call and to increase the likelihood of a chorus occurring. Second, the loud climax phases were significantly longer when produced in choruses than alone, which suggests that males prolong this part of the call when calling in choruses. Within chorus pant-hoots, there was a positive relationship between the number of climax elements given by two calling partners, suggesting that males adjust the temporal structure of their call to mirror their partner's call. We conclude that the basic acoustic structure of chimpanzee pant-hoots and the flexibility with which males adjust the duration of the constituent phases promote chorusing, and that the temporal structure of this rather stereotyped vocalisation is sensitive to fine details of the vocal responses of the audience.     

Female mate choice in the gray treefrog (Hyla versicolor) in three experimental environments

We studied female mate choice by Hyla versicolor in three venues to examine how acoustic and spatial complexity, background noise, and the calling behavior of males might influence preferences manifest in previous laboratory two-stimulus choice tests. Our laboratory-based two-stimulus choice tests with and without broadcasts of chorus noise demonstrated that females prefer long calls relative to short calls when calling efforts of alternatives are equivalent. Background noise impaired the ability of females to discriminate in favor of longer over shorter calls, but the magnitude of the effect was small. Captures of females at eight speakers broadcasting 6- to 27-pulse calls at the edge of a pond revealed strong discrimination against only the shortest call variant. In natural choruses, females may only rarely encounter males using such unattractive vocalizations. Female phonotaxis at an artificial pond with caged and electronically monitored calling males also indicated that consequences of female preferences for temporal aspects of calling observed in two-stimulus choice tests are much attenuated in choruses and explain only small portions (<10%) of the variation in male mating success. Nevertheless, relatively high call duration and calling effort increased male attractiveness. Acoustic interference emerged as another significant factor influencing male mating success and possibly the differences in female choice observed in laboratory and chorus settings. We suggest that the bias of females against both overlapped and very short calls may help explain why males lengthen their calls but lower their rate of delivery in response to increases in chorus size.     

Determinants of chorus tenure in barking treefrogs (Hyla gratiosa)

In all species of anuran amphibians studied to date, male mating success is positively correlated with male chorus tenure (the number of nights that a male is present in breeding aggregations), yet males spend only a small portion of the total breeding season in choruses. Three hypotheses might explain abbreviated chorus tenure in anuran amphibians: (1) mortality, (2) energy limitations, and (3) movement among choruses. I tested these hypotheses in a study of the barking treefrog (Hyla gratiosa), a species with very abbreviated chorus tenure. Mortality accounted for the short tenures of some males: an estimated 20% of males died while calling in the chorus. Energy limitations were an important determinant of chorus tenure. Males lost weight over successive nights in the chorus, and males with longer chorus tenures were in better condition on their first night in the chorus, lost condition more slowly, and were in poorer condition on their last night in the chorus than were males with shorter chorus tenures. Males that I fed crickets as they left the chorus returned sooner and for more nights than did control (unfed) males. Movement between ponds did not explain abbreviated chorus tenure: less than 16% of the males called at more than one pond, and these males increased their chorus tenures by a median of only 2.5 nights by calling in another chorus. Thus, abbreviated chorus tenure in H. gratiosa appears to be primarily the result of energy limitations, with mortality playing a secondary role.     

Male calling behavior,female discrimination and acoustic interference in the Neotropical treefrog Hyla microcephala under realistic acoustic conditions

Summary Anuran choruses are acoustically complex assemblages of calling males. Little is known about the behavior of males or females in such natural sound environments. I studied calling behavior of males of Hyla microcephala in nature by using an interactive computer-based system that allowed me to simulate call interruptions by a number of males. I also monitored the calling behavior of groups of four to six males. When a male is interrupted by the call of another frog, he increases the spacing between the notes of his call. Responses of this kind are strongest to the loudest neighbor, and some males may ignore interruptions by all but a single close male. Interruptions using synthetic calls with silent gaps indicated that males respond vocally to reductions in sound intensity as brief as 20 ms. This ability helps to explain how males can rapidly alternate notes during pairwise interactions. Amounts of acoustic overlap between pairs of males in the choruses were usually below 10% of an individual's total calling time during bouts. The time a male spent calling that was free of acoustic interference by any other male ranged from 34–92% of his total calling time. When group size was decreased, this unobstructed calling time increased. Previous research showed that females of H. microcephala discriminate against calls that overlap so that the call pulse-train structure is degraded. Here I show that a 6 dB difference in intensity between the overlapped calls is sufficient to reduce the degradative effect of call interference. Females were also given a choice between interfering calls broadcast from two adjacent and two widely separated speakers. An angular separation between speakers of 120° was insufficient to elicit a preference for the separated sources. Together, data on behavior of males and females indicated that males actively reduce acoustic interference with those loud individuals most likely to degrade seriously the temporal structure of their calls.     

The sex in short supply for matings varies over small spatial scales in a katydid (Kawanaphilanartee, Orthoptera: Tettigoniidae)

In katydids such as Kawanaphilanartee, a female bias in the operational sex ratio (OSR) results in female competition for mates and male choice of mates. Previous work showed that the excess of sexually active females occurs when food availability is low, in part because less food increases the propensity of females to mate as they forage for the large edible spermatophores produced by males. In this study with K.nartee, a pollen-feeding species, we estimate natural variation in numbers of sexually active males and females by assessing male calling activity and the propensity of females to respond to experimental calling males. We found an excess of sexually active males at a site with many flowers and an excess of sexually active females at a site with few flowers about 900 m away. Between-site differences in gut masses of calling males were consistent with the hypothesis that pollen availability controls OSR. Finally, at a third site where flowers were at first scarce, we found that the initial excess in sexually active females changed to an excess of sexually active males after a clump of grass-trees flowered. The mean gut mass of all sampled males from this site increased after flowering. The large variation in OSR that we document for K. nartee highlights the importance of identifying the appropriate spatial and temporal scales over which OSRs are measured in studies of factors controlling sexual selection. Received: 13 May 1997 / Accepted after revision: 27 October 1997     

Growth rate and the cost of calling activity in male carpenter frogs,Rana virgatipes

Summary Energy stress during the breeding season and relationships between calling activity and growth were investigated in male carpenter frogs, Rana virgatipes. This species has a prolonged breeding season of up to three months in Southern New Jersey. Monthly collections made in 1985 revealed that both dry mass and percent body lipid decreased throughout the breeding season but sharply increased at the end of the breeding season. Observations of free-living males showed that small males were more likely to gain mass than large males during the breeding season. All males gained mass at higher rates after the breeding season. A simultaneous record of calling activity and mass change was obtained for 42 males. Males called on 95% of nights, indicating that they rarely ceased their reproductive activities. Small males tended to have low calling efforts and high growth rates. When the effect of initial mass was removed, growth rate was negatively correlated with calling effort among small males. This is the first demonstration of a direct tradeoff between a reproductive activity and growth in an anuran.     

Spatio-temporal vigilance architecture of an Australian flying-fox colony

The social structure of animal aggregations may vary considerably in both space and time, yet little is known about how this affects vigilance. Here, we investigate the vigilance architecture of a colony of wild-living grey-headed flying-foxes (Pteropus poliocephalus) in Australia and examine how spatial as well as temporal variation in social organization influences social and environmental vigilance. We sampled color-marked individuals at different stages of the reproductive cycle and the year and at different locations in the colony to examine the effects of temporal and spatial factors on social and environmental vigilance. We found that vigilance architecture reflected the social structure of the colony, with the highest environmental vigilance being displayed by bats at the periphery of the colony, and the highest social vigilance by bats that roosted at intermediate distances from the colony’s edge. Furthermore, we found that vigilance levels reflected changes in reproductive state, with social vigilance increasing toward the mating season, particularly in males. Our findings show that spatial and temporal variation in social structure can have differential effects on social and environmental vigilance. This highlights the necessity to differentiate between functions of vigilance to understand fully vigilance architecture in aggregations of social animals.     

The effect of social interactions on calling energetics in the gray treefrog (Hyla versicolor)

Summary The vocal behavior of Hyla versicolor was studied in the field by means of behavioral observations and playback experiments, and these data were coupled with measurements of oxygen consumption in calling frogs to estimate the effect of social interactions on calling energetics. Male gray treefrogs have intense calls (median peak SPL=109 dB, fast RMS SPL=100 dB at 50 cm). At an air temperature of 23° C, males produced an average of 1,200–1,300 calls/h for 2–4 h per night. Calling rates and call durations differed among individuals, but were relatively constant for each male during periods of sustained calling. Males in dense choruses gave calls about twice as long as isolated males, but produced calls at about half the rate. Consequently, total calling effort and estimated aerobic costs were largely independent of chorus density. Playbacks of recorded calls to males in the field elicited increases in call duration and decreases in calling rate, regardless of the rate or duration of the stimulus. Males gave longer calls in response to long calls or to stimuli presented at high rates, but they did not precisely match either stimulus rate or duration. Calling effort and estimated oxygen consumption changed only slightly during stimulus playbacks. These results indicate that male-male competition elicits pro-found changes in the vocal behavior of calling males, but these changes have little effect on energy expenditure. We estimated that most calling males had metabolic rates of about 1.7–1.8 ml O₂/(g\h), or about 280 J/h for an average size (8.6 g) male at 20° C. Although changes in call duration and calling rate did not affect aerobic costs of calling, males producing long calls at slow rates called for fewer hours per night than males producing shorter calls at higher rates. This suggests that calling time may be limited by the rate at which muscle glycogen reserves are depleted.     

Display rate by male fairy-wrens (Malurus cyaneus) during the fertile period of females has little influence on extra-pair mate choice

Empirical and theoretical studies have only recently begun to examine how females use complex multi-component displays when selecting mates. Superb fairy-wrens are well suited to the study of female choice because females have control over extra-group paternity and cuckold their mates at high rates, while males possess a variety of sexually selected traits. Available evidence suggests that females base their extra-group mate choice on the timing of male moult into breeding plumage or the onset of display. However, males continue to perform elaborate displays throughout the season, and direct most displays to females during their fertile period. We therefore conducted focal observations on fertile females to quantify the frequency of male display and used microsatellite genotyping to compare the role of display rate during the breeding season and the timing of male moult on female mate choice. We show that the addition of data on male display rate does not improve our ability to predict which males obtain extra-group paternity. The timing of male moult into breeding plumage remains the only predictor of male extra-group reproductive success. Nevertheless, we found that males displayed more to females that were unable to select extra-group mates on the basis of the timing of moult or the onset of display. This raises the possibility that there are circumstances when females use display rate to discriminate between potential extra-group sires. Overall this study supports the theoretical prediction that females are more likely to base their mate choice on reliable indicators of male quality such as fixed morphological traits and displays of endurance, in this case an early moult into breeding plumage and the performance of an elaborate display during the winter, than a flexible behavioural trait such as display rate during the breeding season. Received: 26 January 2000 / Revised: 1 August 2000 / Accepted: 26 August 2000     

Alternative mating tactics in the Italian treefrog, <Emphasis Type="Italic">Hyla intermedia</Emphasis>

As in many lekking anurans, Italian treefrog males use two mating tactics: they can attract females by calling vigorously or be satellites, that is, they can remain silent in proximity of a calling male and try to intercept females attracted by their neighbour. We investigated the factors that affected the expression of this mating tactic. Consistent with the conditional mating tactic hypothesis, satellites were smaller than average and smaller than their parasitised calling males. They spent a larger-than-average number of nights at the breeding site, where most of them were also observed calling. Moreover, satellites showed lower call rates and lower mating success than those of males they parasitise but not lower than those of males they did not parasitise. Overall, these results, together with those derived from the analyses of the seasonal and spatial distribution of males, provide evidence for a non-random association between satellites and calling males and are consistent with the hypothesis that satellites have spectral and temporal acoustic preferences that parallel those of females. By adopting the less-successful satellite mating tactic, competitively inferior males can nevertheless maximise their potential reproductive fitness by sexually parasitising the most attractive chorusing males.     

Environmental and social determinants of anuran lekking behavior: intraspecific variation in populations at thermal extremes

Environmental and social factors are critical to determine the timing and duration of lekking behavior since they provide species with signs to maximize benefits over costs in sexual displays. However, these factors have rarely been studied under different environmental conditions, and thus, it remains unclear whether exogenous factors affecting group displays show a general species-specific pattern or whether they are population-specific. Using audio-trapping techniques, we compared factors influencing the daily occurrence and duration of lekking behavior in two populations of Hyla molleri and two populations of Hyla meridionalis located at the thermal extremes (coldest vs. hottest) of their Iberian distribution range. From 12,240 hourly recordings over one season, multimodel inference revealed that the major determinants of chorus occurrence were similar between populations and species (i.e., chorus size the previous day, daytime air temperature, relative humidity, and barometric pressure), and accounted for 51–79 % of its deviance. In contrast, the major determinants of chorus duration differed between populations and species (i.e., chorus size, number of day, and air temperature and relative humidity at the onset of the chorus), and accounted for 38–69 % of its variance. Our findings suggest that the decision making related to lek attendance is environment-dependent, takes place at time between lekking events, and is associated with exogenous factors that may be both stable across species ranges and population-specific when populations are under different climatic conditions. This intraspecific variation might be underlain by plasticity mechanisms providing tree frogs with means to cope with changing environments. Moreover, social facilitation related to male-male acoustic competition seems to play a relevant role on the daily time invested by males in lek attendance.     

Here comes the sun: multimodal displays are associated with sunlight incidence

Conspicuousness of courtship signals in animals likely evolved to maximize mate attraction while minimizing predator detection. It is assumed, though largely unknown, that environmental and luminosity conditions affect the detectability of ornaments and motor displays and could thus strongly influence behavior. We combined visual models and behavioral observations to test whether the multimodal display of the blue-black grassquit (Volatinia jacarina) is influenced by environmental conditions, predicting that males should display more often in moments with high sunlight incidence upon their bodies. By displaying in such a context-dependent fashion, males would be able to maximize conspicuousness of their iridescent blue-black plumage and avoid displaying continuously, since the latter would involve higher energetic investment. We recorded the rates of both complete displays (leap with singing) as well as incomplete ones (singing while perched) for males during repeated 30-min focal observations in varying environmental situations in the field. We found that when bathed directly in sunlight, males increased their rates of displays, tending to exhibit more complete rather than incomplete displays in this condition and suggesting a potential trade-off between display types. Our results suggest that animals may adjust the timing and precise location of signal activity to improve the efficiency or likelihood of detection of some signal elements.     

Carbohydrate and calling: Depletion of muscle glycogen and the chorusing dynamics of the neotropical treefrog Hyla microcephala

Chorusing males of the neotropical treefrog Hyla microcephala call in distinct bouts punctuated by periods of silence, a pattern known as unison bout singing. Schwartz (1991) previously tested and refuted the hypotheses that males periodically stop calling either because of a female preference for males that call cyclically, or because high ambient noise levels inhibit vocal activity. Males of H. microcephala are vocally responsive to the calls of other males, and during calling bouts their rate of note production can exceed 10,000 per hour. In natural choruses females preferentially pair with males that call at the higher rates. Because females can pair with males over many hours, males may stop calling periodically to save energy so they can continue to call for the entire period that females are available. We directly tested this energy conservation hypothesis by collecting samples of males early in the evening just after chorusing commenced and later when chorusing had ended for the night. Trunk muscles (internal and external oblique), which are responsible for the airflow associated with note production, were dissected, frozen, and their glycogen content measured. Data on calling behavior were obtained for late-evening samples. Individual calling behavior was not correlated with a males final glycogen level. In addition, many males ended their calling before glycogen reserves were exhausted, indicating that factors other than energy can determine when males finally stop chorusing, However, the biochemical assays supported the energy conservation hypothesis. Unless chorusing was punctuated by pauses, most males would have been unable to sustain high rates of calling for an entire evening without exhausting glycogen reserves in their trunk muscles. Because the time females pair with males is probably unpredictable to males, the ability to call for long periods may improve a males chances of mating.     

Loud calls as indicators of dominance in male baboons (Papio cynocephalus ursinus)

Adult male baboons (Papio cynocephalus) give loud two-syllable 'wahoo' calls during dawn choruses, interactions between groups, when chasing females, and in aggressive interactions with other males. These 'contest' wahoos are acoustically different from 'alarm' wahoos given to predators. In a study of free-ranging baboons in the Okavango Delta, Botswana, we found no significant correlations between the acoustic features of wahoos and adult male size; however, acoustic features were correlated with male dominance rank, age, and calling bout length. Here we show that other measures of calling behavior also appear to function as honest indicators of stamina and competitive ability. High-ranking males were more likely than middle- or low-ranking males to participate in wahoo bouts. They called at significantly higher rates, and their bouts were longer and contained more calls. All males were significantly more likely to participate in wahoo bouts with another male if their opponent's rank was similar to their own. Bouts involving males of similar ranks were longer, contained more wahoos, and involved calling at higher rates, than other bouts. In contests between males of similar ranks, the subordinate and dominant were equally likely to end the bout, whereas in contests involving males of disparate ranks, subordinates were significantly more likely to end the bout. Bouts involving males of similar rank were significantly more likely than others to escalate and result in physical fighting.     

Spacing and distribution of bowers in Macgregor's Bowerbird (Amblyornis macgregoriae)

Summary Bowerbirds (Ptilonorhynchidae) have been described as exhibiting exploded lek mating patterns. Quantitative documentation and analysis of spatial dispersion, however, is lacking for most species despite its importance to a fundamental understanding of social organization. We studied Macgregor's Bowerbird (Amblyornis macgregoriae) in eastern New Guinea in 1980 and 1981 to quantify spatial distribution and selection of bower sites in relation to ecological and social factors. Forty-two bower sites were located in approximately 750 ha of mid-montane forest. They were linearly and regularly spaced along ridges with a mean inter-bower distance of 169±64 m SD (range 75–348 m). Statistical analysis showed males to be regularly spaced within available habitat. Discriminant analysis of seven habitat variables measured along utilized ridge lines showed that bower sites differed significantly from potentially available sites with respect to slope and width of ridge, closure of canopy, and density of saplings. Placement of bowers with respect to these factors was not correlated with inter-bower distance. Most males maintained only one bower; four (10%) maintained two bowers simultaneously. Males spent an average of 54% of daylight hours within an activity center of 15–20 m radius surrounding the bower and aggressively defended this area against conspecific males. Marauding pressure on bowers was high, and 39% of all observed mating attempts (n=18) were disrupted by neighboring males. Macgregor's Bowerbird exhibits a social system that appears to be intermediate between lek behavior and territoriality, combining the social dynamics and opportunism of the former and the uniform field of display sites and nests of the latter. We hypothesize that dispersion in this species is partly due to males buffering their display space against intruder pressure. We suggest that strict adherence to the dispersion criterion in classifying promiscuous mating systems may be inappropriate.     

Multiple sexual advertisements honestly reflect health status in peacocks (Pavo cristatus)

The evolution of multiple sexual traits remains controversial and poorly understood in evolutionary biology. In many bird species, males exhibit complex courtships involving feather ornaments and behavioral display. Multiple traits may convey information on the genetic and phenotypic quality of males. In particular, fixed characters, such as feather ornaments (produced once a year during the annual molt in many bird species) might convey information about past male condition (at the time of trait development); whereas flexible traits such as behavioral displays should be sensitive to present condition. Here we show that both flexible behavioral displays and fixed feather ornaments of peacocks, used by females to choose a mate, honestly reflect health status. Correlative data showed that peacocks with high display rate (the number of behavioral displays per hour) and a large number of tail eyespots had low levels of circulating heterophils, suggesting better health status. Experimental activation of the immune system, through LPS injection, significantly reduced display rate compared to a control group. However, the sensitivity of a male display rate to the immune challenge was dependent on the number of tail eyespots: males with higher number of tail eyespots were better able to cope with the experimental immune challenge, and maintained their display rate at levels closer to the levels performed before the immune activation. These results are consistent with the idea that multiple signaling might enhance information reliability.     

Variation in signal timing behavior: implications for male attractiveness and sexual selection

Mate choice often takes place in group settings, such as leks or choruses, in which numerous individuals display and compete for mates simultaneously. In addition to well-known preferences for male traits like size and signaling rate, females of group-displaying species often show preferences that are based on the relative timing of male signals, generally preferring the leading signal. Variation in male signal timing behavior may therefore affect male attractiveness and, ultimately, reproductive success. I used acoustic communication in green treefrogs (Hyla cinerea) to assess the amount of signal timing variation found in natural choruses and to test hypotheses about the sources of variation in signal timing behavior. I recorded dyads of vocally interacting males in the field to describe patterns of variation in signal timing behavior. Incorporating information about female signal timing preferences, I also assessed the amount of variation in signal timing that is likely to be selectively neutral, as well as the amount that is likely under selection by female choice because it involves attractive or unattractive signal placement. I show that there is considerable variation in signal timing behavior, particularly involving neutral signal timing placement. I also show that between-male variation in other traits (size, signal period) is partly linked to variation in signal timing behavior and discuss these findings in terms of male attractiveness and sexual selection in group-displaying organisms.     

Calling energetics of a neotropical treefrog,Hyla microcephala

Summary We investigated the calling energetics of Hyla microcephala, a neotropical treefrog with an unusually complex vocal repertoire. Males respond to the calls of other individuals by adding secondary click notes to their calls, thereby increasing the total number of notes given per minute. Rates of oxygen consumption of males calling in metabolic chambers were 0.41–2.80 ml O₂/(g·h), corresponding to calling rates of 205–6330 notes/h. Note rate explained 95% of the variance in meta-bolic rate; the effect of variation in body size and temperature was small. Data from playback experiments with males in the field showed that note rate increased as as the number of notes in a stimulus call increased, and this resulted in substantial increases in the cost of calling. Average metabolic rates for males in the field were about 1.7 ml O₂/(g·h), for a net cost of calling of about 20 J/h for an average-size male. However, estimated metabolic rates varied by more than 300% and were strongly influenced by the proximity and calling activity of other males in the chorus. Male H. microcephala appear to conserve energy by reducing calling rates when only a few males are active and increasing calling efforts only when vocal competition among males is intense.     

Territory establishment in lekking marine iguanas, Amblyrhynchus cristatus: support for the hotshot mechanism

The territory establishment of male marine iguanas and their subsequent mating success were analysed to identify spatial spillover (hotshot) and temporal spillover effects on lek formation. Males started to establish small display territories 2 months ahead of the mating season. Males did not establish territories in temporal synchrony and did not settle at sites where the probability of encountering females was highest. However, males arriving later preferentially established their territories in the neighbourhood of already established territories independently of the density of female-sized iguanas in these territories. Although settling in close proximity, there were no fights between those males. The number of fights between territorial males increased towards, and peaked during, the mating season. Fights did not result in the transfer of space, indicating that space per se was no resource. Instead, fights were directed towards central (hotshot) males. These central males had higher mating success than marginal males. Female density during the time of territory establishment did not predict the mating success of males, because females changed their spatial preferences between early establishment and mating periods. Similarly, the areas where males achieved the highest numbers of copulations changed during 4 years of our study. Thus, there was no evidence for temporal spillover between subsequent seasons. However, most male-male interactions served to distract successful males and may lead to spatial spillover of females into territories of unsuccessful males. In marine iguanas, territorial establishment appears largely governed by hotshot processes.     

Parental investment, potential reproductive rates, and mating system in the strawberry dart-poison frog, Dendrobates pumilio

We studied the effect of relative parental investment on potential reproductive rates (PRRs) to explain sex differences in selectivity and competition in the dart-poison frog Dendrobates pumilio. We recorded the reproductive behavior of this species in a Costa Rican lowland rainforest for almost 6 months. Females spent more time on parental care than males, and `time out' estimates suggest that PRRs of males are much higher than than those of females, rendering females the limiting sex in the mating process. Males defended territories that provide suitable calling sites, space for courtship and oviposition, and prevent interference by competitors. Male mating success was highly variable, from 0 to 12 matings, and was significantly correlated with calling activity and average perch height, but was independent of body size and weight. Estimates of opportunity for sexual selection and variation in male mating success are given. The mating system is polygamous: males and females mated several times with different mates. Females were more selective than males and may sample males between matings. The discrepancy in PRRs between the sexes due to differences in parental investment and the prolonged breeding season is sufficient to explain the observed mating pattern i.e., selective females, high variance in male mating success, and the considerable opportunity for sexual selection. Received: 9 June 1998 / Received in revised form: 27 March 1999 / Accepted: 3 April 1999