CHARACTERIZATION OF FEMALE PREFERENCE FUNCTIONS FOR DROSOPHILA MONTANA COURTSHIP SONG AND A TEST OF THE TEMPERATURE COUPLING HYPOTHESIS

Transcription

1 Evolution, 55(4), 2001, pp CHARACTERIZATION OF FEMALE PREFERENCE FUNCTIONS FOR DROSOPHILA MONTANA COURTSHIP SONG AND A TEST OF THE TEMPERATURE COUPLING HYPOTHESIS MICHAEL G. RITCHIE, 1,2 MARI SAARIKETTU, 3 SUSAN LIVINGSTONE, 1 AND ANNELI HOIKKALA 2 1 Environmental and Evolutionary Biology, Bute Medical Building, University of St. Andrews, St. Andrews, Fife KY16 9TS, Scotland 2 mgr@st-and.ac.uk 3 Department of Biology, University of Oulu, PO Box 3000, Oulu, Finland Abstract. Female mate preferences are a major cause of diversity and elaboration in male sexual traits. Here we characterize the shape of female preference functions for pulse length and carrier frequency of the courtship song of Drosophila montana by fitting both parametric and nonparametric functions to the incidence of female receptive gestures to synthetic song. Preference functions for both traits are strongly directional. That for pulse length is linear and favors short pulses, whereas that for carrier frequency is stabilizing in shape, but would exert directional preferences favoring males with high carrier frequency. The preference for carrier frequency has probably evolved under sexual selection, but reasons for the preference for short pulses are less apparent. We also examine the effect of ambient temperature on the carrier frequency of male song and on the preference function for carrier frequency. For many similar acoustic communication systems, temperature coupling, a compensatory effect of temperature on preference functions, is thought to maintain coordination between preferences and signals. However, although the carrier frequency of D. montana song is highly dependent on environmental temperature, there is no temperature coupling of the female preference function. We suggest that temperature coupling may often arise due to a common effect of temperature on song and preference, rather than be an advantageous characteristic whose function is to maintain coordination in temperature-affected communication systems. Key words. Courtship songs, Drosophila montana, female preferences, mate recognition systems, temperature. Female mate preferences are a major cause of diversity and elaboration in male sexual traits (Bakker and Pomiankowski 1995). To understand how sexual selection by female choice has evolved, it is important to determine not only the variance, repeatability, and inheritance of female preferences, but also female sampling strategies and the manner in which environmental conditions affect the ability of females to discriminate between male traits (Jennions and Petrie 1997; Wagner 1998). Female preference functions describe how the strength of a female s mating response to a male trait varies with the degree of expression of the trait. The amount of overlap between the male trait and female preference distributions will determine the extent to which sexual selection will arise in a communication system (Lande 1981; Ritchie 1996). Preference functions are often described as directional or stabilizing, and preferences favoring the mean or extremes of traits can influence the variability (Gerhardt 1991) and possibly the heritability (Pomiankowski and Møller 1995) of the male trait. Progress is being made with the characterisation of preference functions in a few model species (e.g., Basolo 1999; Shaw and Herlihy 2000) and their inheritance (Ritchie 2000). Few studies have been made of the consistency of preference functions under a range of environmental conditions (their reaction norms). This is important, because some preference strategies should vary with environmental condition and their genetic reference templates can be developmentally flexible (Widemo and Sæther 1999). Three strategies of mate choice have been modeled (Lande 1981; Maynard Smith 1987), although these are not exclusive. Threshold and absolute choice are based on an internal (genetic) standard of the female and suggest that females should mate with males exceeding a threshold. Threshold 2001 The Society for the Study of Evolution. All rights reserved. Received November 1, Accepted November 27, choice occurs when any male who exceeds a threshold value is acceptable, whereas absolute choice occurs when only males who match a particular, narrow range of values are acceptable. Relative choice, on the other hand, relies on comparisons so the choice made might vary depending on the number and range of available males (e.g., when following best-of-n rules; Jennions and Petrie 1997). Relative mate choice strategy is thus the most flexible method of assessing male signals if the signals vary for environmental reasons. Most models of sexual selection emphasise the honesty or reliability of male mating signals as indicators of quality (Andersson 1994). A major class of signal used in sexual communication of poikilotherms, acoustic signals, are profoundly influenced by environmental temperature because they rely on the rate of muscle contraction (Shorey 1962; Bailey 1991). If the signals of individual males vary due to environmental conditions, their role in mate choice may be questionable. Even insects that show a degree of thermoregulation (e.g., the tettigoniid cricket Noeconocephalus robustus, as described in Heinrich 1993) show temperature dependence of song traits. The problem of matching internal standards to such labile signals may be overcome if temperature has similar effects on the male signal and the female preference function (Walker 1957). Gerhardt (1978) termed such common effects temperature coupling. Numerous examples of this have been described, most often in insects and frogs (Walker 1957; Gerhardt 1978; Doherty 1985; Pires and Hoy 1992a). Drosophila montana (a member of the D. virilis group) is especially well suited for studies of female preferences for male song. In this species courtship song, produced by wing vibration, is an obligatory component of courtship without which females do not mate (Liimatainen et al. 1992). Female

2 722 MICHAEL G. RITCHIE ET AL. preferences can be characterized by monitoring female receptive gestures (wing spreading) to synthetic courtship songs (Ritchie et al. 1998). Drosophila montana has a circumarctic distribution, and courtship and mating occur during a short period in spring when the environmental temperature is highly variable (Aspi et al. 1993). Preference of D. montana females for high-frequency song has been detected both in the wild (Aspi and Hoikkala 1995) and the laboratory (Ritchie et al. 1998; Hoikkala and Suvanto 1999), and females have also been shown to gain indirect benefits from mate choice based on song variation (Hoikkala et al. 1998). A previous study with artificial song, which classified pulse length and carrier frequency into three categories (high, medium, and low), found that the most favored song depended on an interaction between these traits. The most stimulating song had short pulses combined with high carrier frequency (Ritchie et al. 1998). In the present study, our aim is to characterize the female preference functions for these two male song traits when they are varied continuously and independently. Because songs were presented singly, this experiment cannot detect relative choice criteria. To determine if temperature coupling occurs in this species, we also examine the influence of environmental temperature on the carrier frequency of male song and the female preference function for this trait. EXPERIMENT 1: PREFERENCE FUNCTIONS FOR SONG PULSE LENGTH AND CARRIER FREQUENCY Methods The objective of experiment 1 was to characterize the preference functions for pulse length and carrier frequency separately. We synthesized songs that systematically varied either parameter using the Signal system (Engineering Design, Belmont, MA) with an A/D rate of 4 khz. Pulse envelopes took 3 msec each to reach and decline from maximum amplitude. The interpulse interval was 35 msec and song was patterned into bursts of 1-sec duration followed by 1 sec of silence. This represents a prolonged bout of courtship, during which female wing spreading and mounting might occur. To examine the effects of pulse length, 11 different songs were synthesized with pulse lengths from 6 to 26 msec (in 2-msec increments) and a constant carrier frequency of 300 Hz. To examine the effects of carrier frequency, an additional 11 songs were synthesized that varied carrier frequency between 200 to 500 Hz (in 30-Hz increments) and a constant pulse length of 8 msec. Songs were replayed directly from a PC (sampling rate 4 khz) via a FERN EF5-20, filter (East Kilbride, Scotland) bandpassed between 100 and 1500 Hz and a Sony Corporation (Staines, Middlesex, U.K.) SRS-38 loudspeaker. The sound pressure level (SPL) was set to 80 db at the chamber and measured using a Realistic (Barrie, Ontario) sound level meter (calibrated with a Brüel and Kjær [Nærum, Denmark] 2231M sound level meter and 4155 microphone). Flies originated from Oulanka in Finland (Oulu code 1251/ 20) and were reared at room temperature under constant light (normal for the Finnish summer). Playback experiments took place in a sound attenuated chamber at 22 1 C. Virgins were collected and used once only at an age of 3 5 weeks, when the flies are sexually mature. Females were held in a chamber measuring cm with a nylon gauze floor and roof, suspended over the loudspeaker. The chambers were held in a drilled plastic frame and were equidistant from the speaker. There was no detectable variation in SPL according to chamber position. A wingless virgin male was included in all trials, because this increases the overall level of responsiveness but not the pattern of preference, presumably by providing important nonacoustic cues (Ritchie et al. 1998). Depending on availability, the number of pairs used in an individual trial varied, to a maximum of six pairs. Playbacks were continued for 10 min, and females scored as responding if they raised their wings for approximately 2 sec (Vuoristo et al. 1996; Ritchie et al. 1998) or as not responding. A total of 12 pairs of females were presented with each of the 11 synthetic songs (presented in a random sequence) giving a total of 132 females for each set of songs. Preference functions are analyzed both by model fitting and nonparametrically (analogous to recent studies of selection surfaces, Brodie et al. 1995). The model-fitting procedure involved fitting a model of the form: preference trait (trait) e, (1) where 1 estimates the linear component of the preference function, 2 the quadratic component, 0 is a constant, and e the error term. 1 will predominate linear preference functions, whereas 2 will predominate stabilizing (or disruptive) functions. Because preference is measured as a binomial response (wing-lifting vs. not wing-lifting) against a continuous trait, model fitting was carried out using generalized linear modelling with a logit link function with Genstat5 (ver. 4.1; Genstat 5 Committee 1993). This provides analyses of the contribution of each term to the total deviance of the model, which can be tested for significance as deviance is distributed as 2. Preference functions are also estimated by fitting cubic splines to the responses. This analysis is particularly appropriate because it is nonparametric, involving no a priori assumptions concerning the shape of the preference function. Potentially any shape up to straight lines between consecutive points can be fitted (Schluter 1988; Ritchie 1996). Before fitting splines a random error of 2.5% was introduced to the song value to avoid ties, and the best smoothing parameter found by iteration. Bootstrapping (1000 replicates) was used to find mean splines and standard errors (Schluter 1988). Results Table 1 shows the results of the model-fitting process for the female responses versus song varying in pulse length or carrier frequency. Both models are significant, but only the linear regression coefficient is significant for the pulse length preference function, whereas only the quadratic regression coefficient is significant for the preference function for carrier frequency. The coefficients indicate that females prefer short pulse lengths and intermediate carrier frequency. This is confirmed by the nonparametric estimation of preference functions. Figure 1a shows the shape of the spline fitted to the female responses to song models varying in pulse length, which shows a highly directional preference function for shorter pulse lengths. Males from this population have a mean

3 FEMALE PREFERENCE FUNCTIONS 723 TABLE 1. Model fitted to the female responses to songs varying in either pulse length or carrier frequency, with associated analysis of deviance. Model Pulse length: Total Coefficient Deviance (df) (1) 0.01 (1) (2) Carrier frequency: Total (1) 8.42 (1) 8.43 (2) P ns ns pulse length of 21.3 msec at this temperature (Suvanto et al. 2000), which is in the least preferred range of pulse length tested here. Figure 1b shows the spline fitted to the female responses to song models varying in carrier frequency. This implies a stabilizing preference function, peaking around 350 Hz. This is toward the upper limit of song frequencies produced by males at this temperature, and so would in practice also produce directional preferences within the population. In combination, the results of this experiment demonstrate clearly how these two preference functions underlie the previous finding of an interaction effect between pulse length and carrier frequency in determining responses to songs that varied in both characteristics the best song has a short pulse length and relatively high carrier frequency. EXPERIMENT 2: THE EFFECT OF AMBIENT TEMPERATURE ON SONG AND PREFERENCE Methods Given the strongly directional preferences for pulse length found in experiment 1, there is no clear prediction of how temperature might influence female preference. Therefore, we concentrated here on the effects of temperature on carrier frequency and preference. The same strain and rearing procedures were used as in experiment 1, except that fly rearing and playback experiments were carried out in a constanttemperature room that maintained temperature to 0.5 C. General fly rearing was carried out at 20 C, although song recording and playback were carried out at a range of temperatures. During these experiments the rather noisy temperature control equipment was switched off. The room held temperature for sufficiently long to complete each experiment. Experiments were discontinued if the temperature rose by more than 2 C. To record song, virgin males were placed with a female in a small chamber within a customized Insectavox microphone (Gorczyca and Hall 1987) and a Marantz (Eindhoven, The Netherlands) CP430 cassette recorder was used to record a few bursts of courtship song. Temperature inside the Insectavox was noted ( 0.1 C, measured using a Portec [Wrestlingworth, Bedfordshire, U.K.] P9005 digital thermometer) while the male was singing. Males were between 22 and 26 days old. Nineteen males were used, and each was recorded FIG. 1. Mean cubic splines fitted to female responses when presented with song varying in (a) pulse length (smoothing parameter 16.69); and (b) carrier frequency (smoothing parameter 10.53). Dotted curves are 1 standard error and the shaded blocks on the abscissae indicate approximate 95% confidence intervals of the trait in males. on two or three occasions on different days and at different temperatures (mean number of recordings per male 2.8). Songs were digitized at 4 khz using the Signal program, and the peak of a maximum of 10 separate FFT transforms (mean 4.13) taken across different areas of song from each recording. The mean of these peaks was used as a single measure of song carrier frequency for each recording. To examine the effect of ambient temperature on song preference, we synthesized song that varied in carrier frequency between 170 and 500 Hz in 30-Hz increments (at a sample rate of 4 khz) with a constant pulse length of 13 msec. These songs were recorded onto reel-to-reel tape using a TASCAM 22-2 (Tokyo) and replayed using a subwoofer of a Boston Acoustics, Inc. (Peabody, MA) MicroMedia PC sound system, which reproduces the low-frequency pulses typical of fly song particularly well. The aim was to test 12 females in groups of three or four with each song. One randomly selected combination of song and temperature was carried out each day. Due to mortality, the mean sample size for each song was (range 9 12). Different females were used for each song, but females were not unique across temperatures. Other procedures were as before, except that flies were left overnight to equilibrate to temperature before playback.

4 724 MICHAEL G. RITCHIE ET AL. FIG. 2. Mean carrier frequency of pulse song versus environmental temperature. Points are derived from the analysis of 19 males, and the fitted line is a simple linear regression (frequency 16.01[temperature] 7.85; r ; see text for full analysis). Results Figure 2 shows mean song carrier frequency per recording plotted against ambient temperature. Carrier frequency strongly covaries with environmental temperature (F 1, , P 0.001, from a generalized linear model with the terms male, temperature, age, and male temperature interaction, all other terms in the model were not significant). Figure 3 shows the nonparametric curves fitted to the female responses at 15, 20, and 25 C. These are broadly similar. The curve from females at 20 C is less strongly peaked than the rest, but this is unlikely to be a genuine biological difference (experiment 1 was carried out at around this temperature). Quadratic regression models were fitted to the data as before to provide a quantitative test of the temperature coupling hypothesis. Differentiation of the fitted curves allowed the peak of the preference functions to be calculated (Table 2). The predicted value of males at the corresponding temper- FIG. 3. Cubic splines fitted to female responses at each temperature (open circles, 15 C; gray circles, 20 C; filled circles, 25 C). The smoothing parameter (6.2) was chosen from a single analysis across all temperatures. Horizontal lines indicate the mean carrier frequency of males at the same temperatures. Bootstrapped error limits for the curves (not shown on the figure for clarity) were all overlapping. See also Table 2. TABLE 2. Analysis of the degree of correspondence between male carrier frequency (CF) and the peak of female preferences at three temperatures. Comparison Peak preference (Hz) Fitted female preference function Temp. ( C) Z-score P Predicted CF (Hz) (95% CI) Temperature ( C) ns ( ) ( ) ( ) 11.35

5 FEMALE PREFERENCE FUNCTIONS 725 ature, together with their predicted distributions, was calculated from simple regression analysis of the data shown in Figure 2. Z-scores allow the probability of the observed peak of female preference to be calculated if the temperature coupling hypothesis applied (i.e., the fitted peak preference should have been from the distribution of carrier frequencies of males at the same temperature). Table 2 summarizes the results. Female peak preferences only poorly match that predicted from the song data each peak being unlikely to correspond to male song at the same temperature (indeed, females at 15 and 25 C would be most likely to respond to males at 20 C, whereas females at 20 C are most likely to respond to males at 15 C). There is no support for the temperature coupling hypothesis. DISCUSSION The wing-spreading receptive gesture of D. montana females allows accurate characterization of their preferences for male song. The response reliably predicts actual mating behavior because analysis of mating pairs in the field implicate the same song traits in mating success as evoke responses in the laboratory (Aspi and Hoikkala 1995). A previous playback experiment implied that the attractiveness of song was due to an interaction between carrier frequency and pulse length, with short high-frequency song being favored (Ritchie et al. 1998). The independent preference functions described here reveal why, showing that preferences are strongly directional for short pulse length and high carrier frequency. The preference function for carrier frequency is not open-ended, but tails off at the upper limits of the distribution of the male trait. However, the selection pressure this would exert on contemporary males is effectively directional (Fig. 1a). Preference favoring high-frequency song has probably evolved under sexual selection. Offspring of males with highfrequency song have increased larval survival (Hoikkala et al. 1998). Although the biomechanics of song production in Drosophila is poorly understood (Bennet-Clark and Ewing 1968; Ewing 1989), it seems likely that high-frequency song is more costly to produce and may therefore be a reliable indicator trait (Andersson 1994). Repeatability analysis indicates that the trait is highly repeatable only after males have overwintered. Winter dormancy is probably a stressful event, affecting male condition (Hoikkala and Isoherranen 1997). The directional function favoring short pulses over the range analyzed here is more surprising because pulse length does not correlate with offspring quality (Hoikkala et al. 1998). Because interpulse interval is held constant in our synthetic songs, this function also means that females are favoring song stimuli with the lowest duty cycle, which is rather unusual for acoustic insects (Hedrick 1986; Yang and Greenfield 1996; Shaw and Herlihy 2000) and animal communication in general (Bradbury and Vehrencamp 1998). One possibility is that short pulses are more stimulatory for neuromechanical reasons (flies detect pulses with antennal aristae; Ewing 1978), but the attack (build-up) of the sound pulse did not vary between our synthetic songs, and shorter pulses are not more attractive in D. melanogaster (Bennet-Clark and Ewing 1969). Males from this strain of D. montana are clearly suboptimal in terms of their attractiveness (Fig. 1a). Sound pulses of wild-caught males vary between 19 to 21 msec (Suvanto et al. 1999). Whether D. montana males might be constrained to have longer pulses is not known; males of other Drosophila species have shorter pulses. Nine species of the virilis group have shorter pulses than D. montana, measuring from 14 to 18 msec (Hoikkala et al. 1982). Temporal components of acoustic signals of poikilotherms are usually temperature dependent (Bailey 1991; Pires and Hoy 1992a), and this is clearly the case for the carrier frequency of D. montana song, which increases by about 15 Hz per degree centigrade. Similar results have been found for other virilis-group species (Hoikkala 1985). The carrier frequency of fly song results from minor wing vibrations within each major wing beat (Bennet-Clark and Ewing 1968) and is therefore muscular rather than resonant in origin, so the effect of temperature is not surprising (in contrast, the carrier frequency of many acoustic insects such as Orthoptera is resonant in origin and therefore less likely to vary with temperature). Temperature coupling between preferences and signals is usually regarded as a means of maintaining coordination between male and female components of such communication systems. It has been almost ubiquitously demonstrated in a range of organisms including grasshoppers, crickets, frogs, vibrationally signaling spiders, and visually signaling fireflies (Walker 1957; Carlson et al. 1976; Gerhardt 1978; Doherty 1985; Pires and Hoy 1992a; Shimizu and Barth 1996). The grasshopper Chorthippus biguttulus, in which females recognize the song of males at any temperature, is one other exception (von Helversen and von Helversen 1981, 1983). Would the lack of temperature coupling demonstrated here present any difficulties for acoustic communication in D. montana? Temperature-induced song variation may be largely irrelevant to mate choice if animals only court within a narrow temperature range under field conditions. Many flies have a strong diurnal peak of mating activity, which may reduce the range of relevant temperatures. However, Liimatainen and Hoikkala (1998, unpubl. obs.) observed D. montana courtship in the field and found that copulations occurred over the temperature range C (with the majority between 15 C and 20 C), mating success being independent of environmental temperature. The preference functions found here will exert directional selection effectively over this range of temperatures and could contribute to the lack of successful courtship outwith this range. M. Noor (pers. comm.) found that D. pseudoobscura mate in the field within a similar temperature range. Other field studies of Drosophila courtship that we are aware of have not mentioned environmental temperatures (Partridge et al. 1987; Gromko and Markow 1993). General considerations lead us to question if temperature coupling is an effective means of adjusting female preferences. The preference function of D. montana for song carrier frequency has presumably evolved under selection to maximize the likelihood of mating with males with higher frequencies. Temperature coupling might therefore be beneficial if it would lessen the probability of females mating with lowquality males who are warmer than others. However, females

6 726 MICHAEL G. RITCHIE ET AL. are unlikely to be able to assess the temperature of singing males except through self-reference. Recent studies of insect temperature regulation imply that it is very unlikely that courting individuals will be at precisely the same temperature, especially because singing will generate an endogenous temperature increase (Heinrich 1993). Singing cicadas commonly have body temperatures as much as 10 C higher than inactive cicadas (Sanborn et al. 1995). Although this might be less severe with the large surface area:volume ratio of Drosophila, temperature increases are still likely to occur. Females of phonotaxing species are almost certain to be in a different microclimate from calling males (e.g., male frogs who call from within ponds to attract females across a terrestrial environment or male tettigoniid crickets calling from atop bushes and receiving direct insolation, attracting females who remain under cover of vegetation). The problem of females not knowing male temperature accurately is clearly seriously compounded if the choice tactic is relative and comparisons are sequential rather than simultaneous. Females of D. montana are known to use relative criteria when assessing mates, if the signals of courting males are within an acceptable range (Hoikkala and Aspi 1993). Perhaps the only effective sexually selected choice tactic for temperature-dependent traits is to have extremely directional preferences, such as we have found for D. montana. A requirement for accurate temperature coupling will be greatest when the function evolves under stabilizing selection rather than directional selection, for example, when there is little or no sexual conflict over the information content (Maynard Smith 1991). This may be the case for species-recognition signals, such as those that have evolved by character displacement to prevent hybridization. Although these are uncommon, convincing examples have recently been described (Marshall and Cooley 2000). If this is true, the near ubiquity of temperature coupling in a wide range of signaling species might often result from common physiological processes underlying song and preference generation, rather than represent an attribute evolved to maintain signaling efficiency (see also Pires and Hoy 1992a). Both song and preferences are neuronal in origin, although separate neurons are involved (Bauer and von Helversen 1987; Pires and Hoy 1992b) and musculature is likely to play a greater role in song. Although song and preference have often been shown to vary with temperature, in some species the covariance is low, sometimes so much so that they are mismatched over a wide range of temperatures (e.g., Skovmand and Pederson 1983; Shimizu and Barth 1996). In such cases, it has been argued that temperature-independent ratios between song traits may play a role in song recognition (although evidence for this is also ambiguous). Testing whether temperature coupling is selected for or an arbitrary by-product of signal and preference physiology should be carried out. One possibility would be to test if coupling is more effective where character displacement produces greater selection on the sexual communication system, along with analysis of the temperature covariation between courting individuals (which might be rather low). To our knowledge, no such tests have been carried out, although Walker s original article (Walker 1957) emphasized the importance of temperature coupling for species recognition in sympatric species of Oecanthus. In conclusion, we have shown that female preference functions of D. montana favor males with short song pulse lengths and a high intrapulse carrier frequency. Carrier frequency is strongly influenced by environmental temperature, but female preferences probably exert directional selection effectively over the relevant range of temperatures. Consideration of this suggests that the concept of temperature coupling of preferences and song as a naturally selected and advantageous means of maintaining coordination in acoustic communication systems needs reevaluation and more rigorous testing. Accurate characterization of female preferences will help resolve many such issues in sexual selection and communication (Bakker and Pomiankowski 1995; Jennions and Petrie 1997). ACKNOWLEDGMENTS This work was partially supported by a research grant (project 36166) from the Academy of Finland to AH. We are grateful to C. Gerhardt, J. Ghazoul, A. Snedden, and two reviewers for discussion and comments on the manuscript. LITERATURE CITED Andersson, M Sexual selection. Princeton Univ. Press, Princeton, NJ. Aspi, J., and A. Hoikkala Laboratory and natural heritabilities of male courtship song characters in Drosophila montana and D. littoralis. 70: Male mating success and survival in the field with respect to size and courtship song characters in Drosophila litoralis and D. montana (Diptera: Drosophilidae). J. Insect Behav. 8: Aspi, J., J. Lumme, A. Hoikkala, and E. Heikkinen Reproductive ecology of the boreal riparian guild of Drosophila. Ecography 16: Bakker, T. C. M., and A. Pomiankowski The genetic basis of female mate preferences. J. Evol. Biol. 8: Bauer, M., and O. von Helversen Separate localisation of sound recognizing and sound producing neural mechanisms in a grasshopper. J. Comp. Physiol. A. 161: Bailey, W. J Acoustic behaviour of insects: an evolutionary perspective. Chapman and Hall, London. Basolo, A. L Evolutionary change in a receiver bias: a comparison of female preference functions. Proc. R. Soc. Lond. B. 265: Bennet-Clark, H. C., and A. W. Ewing The wing mechanism involved in the courtship of Drosophila. J. Exp. Biol. 49: Pulse interval as a critical parameter in the courtship song of Drosophila melanogaster. Anim. Behav. 17: Bradbury, J. W., and S. L. Vehrencamp Principles of animal communication. Sinauer, Sunderland, MA. Brodie, E. D., A. J. Moore, and F. J. Janzen Visualising and quantifying natural selection. Trends Ecol. Evol. 10: Carlson, A. D., J. Copeland, R. Raderman, and A. G. M. Bulloch Role of interflash intervals in a firefly courtship (Photinus macdermotti). Anim. Behav. 24: Doherty, J. A Trade-off phenonema in calling song recognition and phonotaxis in the cricket Gryllus bimaculatus (Orthoptera, Gryllidae). J. Comp. Physiol. 156: Ewing, A. W The antenna of Drosophila as a love song receptor. Physiol. Entomol. 3: Arthropod bioacoustics: neurobiology and behaviour. Edinburgh Univ. Press, Edinburgh, Scotland. Genstat Genstat 5 release 3 reference manual. Oxford University Press. Oxford, U.K.

7 FEMALE PREFERENCE FUNCTIONS 727 Gerhardt, H. C Temperature coupling in the vocal communication system of the gray tree frog, Hyla versicolor. Science 199: Female mate choice in treefrogs: static and dynamic acoustic criteria. Anim. Behav. 42: Gorczyca, M., and J. C. Hall The Insectavox, an integrated device for recording and amplifying courtship songs of Drosophila. Dros. Inf. Serv. 66: Gromko, M. H., and T. A. Markow Courtship and remating in field populations of Drosophila. Anim. Behav. 45: Hedrick, A. V Female preference for male calling bout duration in a field cricket. Behav. Ecol. Sociobiol. 19: Heinrich, B The hot-blooded insects. Springer-Verlag, Berlin. Hoikkala, A Genetic variation in the male courtship sound of D. littoralis. Behav. Genet. 15: Hoikkala, A., and J. Aspi Criteria of female mate choice in Drosophila littoralis, D. montana and D. ezoana. Evolution 47: Hoikkala, A., and E. Isoherranen Variation and repeatability of courtship song characters among wild-caught and laboratoryreared Drosophila montana and D. littoralis males (Diptera: Drosophilidae). J. Insect Behav. 10: Hoikkala, A., and L. Suvanto Male courtship song frequency as an indicator of male mating success in Drosophila montana. J. Insect Behav. 12: Hoikkala, A., S. Lakovaara, and E. Romppainen Mating behavior and male courtship sounds in the Drosophila virilis group. Pp in S. Lakovaara, ed. Advances in genetics, development and evolution of Drosophila. Plenum, New York. Hoikkala, A., J. Aspi, and L. Suvanto Male courtship song frequency as an indicator of male genetic quality in an insect species, Drosophila montana. Proc. R. Soc. Lond. B 265: Jennions, M. D., and M. Petrie Variation in mate choice and mating preferences: a review of causes and consequences. Biol. Rev. 72: Lande, R Models of speciation by sexual selection on polygenic traits. Proc. Natl. Acad. Sci. USA 78: Liimatainen, J., and A. Hoikkala Interactions of the males and females of three sympatric Drosophila virilis group species, D. montana, D. littoralis and D. lummei, in intra- and interspecific courtships in the wild and in the laboratory. J. Insect Behav. 11: Liimatainen, J., A. Hoikkala, J. Aspi, and P. Welbergen Courtship in Drosophila montana: the effects of male auditory signals on the behaviour of flies. Anim. Behav. 43: Marshall, D. C., and J. R. Cooley Reproductive character displacement and speciation in periodical cicadas, with description of a new species, 13-year Magicicada neotredecim. Evolution 54: Maynard Smith, J Sexual selection a classification of models. Pp in J. W. Bradbury and M. B. Andersson eds., Sexual selection: testing the alternatives. Wiley, Chichester, U.K Theories of sexual selection. Trends Ecol. Evol. 6: Partridge, L., A. Hoffmann, and J. S. Jones Male size and mating success in Drosophila melanogaster and D. pseudoobscura under field conditions. Anim. Behav. 35: Pires, A., and R. R. Hoy. 1992a. Temperature coupling in cricket acoustic communication. I. Field and laboratory studies of temperature effects on calling song production and recognition in Gryllus firmus. J. Comp. Physiol. A 171: b. Temperature coupling in cricket acoustic communication. II. Localization of temperature effects on song production and recognition networks in Gryllus firmus. J. Comp. Physiol. A 171: Pomiankowski, A., and A. P. Møller A resolution of the lek paradox. Proc. R. Soc. Lond. B 260: Ritchie, M. G The shape of female mating preferences. Proc. Natl. Acad. Sci. U.S.A. 93: The inheritance of female preference functions in a mate recognition system. Proc. R. Soc. Lond. B 267: Ritchie, M. G., R. M. Townhill, and A. Hoikkala Female preference for fly song: playbacks confirm correlational evidence of the targets of sexual selection. Anim. Behav. 56: Sanborn, A. F., M. S. Heath, J. E. Heatch, and F. G. Noriega Dirunal activity, temperature responses and endothermy in three south american cicadas (Homoptera: Cicadidae: Dorsiana bonaerensis, Quesada gigas and Fidicina mannifera). J. Therm. Biol. 20: Schluter, D Estimating the form of natural selection on a quantitative trait. Evolution 42: Shaw, K. L., and D. P. Herlihy Acoustic preference functions and song variability in the Hawaiian cricket Laupala cerasina. Proc. R. Soc. Lond. B 267: Shimizu, I., and F. G. Barth The effect of temperature on the temporal structure of the vibratory courtship signals of a spider (Cupiennius salei Keys.). J. Comp. Physiol. A 179: Shorey, H. H The nature of the sound produced by Drosophila melanogaster during courtship. Science 137:677. Skovmand, O., and S. B. Pedersen Song recognition and song pattern in a shorthorned grasshopper. J. Comp. Physiol. A 153: Suvanto, L., J. Liimatainen, and A. Hoikkala Variability and evolvability of male song characters in Drosophila montana populations. Hereditas 130: Suvanto, L., J. O. Liimatainen, T. Tregenza, and A. Hoikkala Courtship signals and mate choice of the flies of inbred Drosophila montana strains. J. Evol. Biol. 13: von Helversen, D., and O. von Helversen Korrespondenz zwischen gesand und auslösendem schema bei feldheuschrecken. Nova acta Leopold. 54: Species recognition and acoustic localisation in acridid grasshoppers: a behavioral approach. Pp in F. Huber and H. Markl, eds. Neuroethology and behavioral physiology. Springer-Verlag, Berlin. Vuoristo, M., E. Isoherranen, and A. Hoikkala Female wing spreading as acceptance signal in the Drosophila virilis group of species. J. Insect Behav. 9: Wagner, W. E. J Measuring female mating preferences. Anim. Behav. 55: Walker, T. J Specificity in the response of female tree crickets (Orthoptera, Gryllidae, Oecanthidae) to calling songs of the males. Ann. Entomol. Soc. Am. 50: Widemo, F., and S. A. Sæther Beauty is in the eye of the beholder: causes and consequences of variation in mating preferences. Trends Ecol. Evol. 14: Yang, Y., and M. D. Greenfield Ultrasonic communication and sexual selection in waxmoths: female choice based on energy and asynchrony of male signals. Anim. Behav. 51: Corresponding Editor: H. L. Gibbs