ANIMAL BEHAVIOUR, 1998, 56, 713–717 Article No. ar980799 Female preference for fly song: playback experiments confirm the targets of sexual selection MICHAEL G. RITCHIE*, RICHARD M. TOWNHILL* & ANNELI HOIKKALA² *Environmental & Evolutionary Biology, University of St Andrews ²Department of Genetics, University of Oulu (Received 19 September 1997; initial acceptance 1 December 1997; final acceptance 14 February 1998; MS. number: 5651) ABSTRACT The courtship song of Drosophila is thought to be involved in sexual selection and species recognition. Because of the mating system of flies, however, directly demonstrating that song influences female preference is difficult. The majority of previous studies have used an experimental design that potentially confounds male and female reactions to song. In D. montana, correlational evidence has suggested that males that produce short sound pulses consisting of a high number of sound cycles (i.e. a high carrier frequency) have a higher mating success than other males. In this study, we played synthetic song that varied in pulse length and carrier frequency to individual females in the laboratory, both alone and in the presence of mute males. We scored female preference via an acceptance posture, ‘wing spreading’, which the females of this species usually display prior to mounting by a male. Females responded to synthetic song in the absence of males. Th e presence of mute males significantly increased their overall responsiveness, but the relative effectiveness of the songs did not change, eliminating male reaction to song as a possible confounding factor in the results. The interaction between pulse length and carrier frequency determined the discrimination between song types, with females responding most readily to song consisting of short pulses with a high carrier frequency. Thus, direct examination of female preferences supports the previous studies of male mating success, and confirms female song preference as a likely determinant of male mating success.  1998 The Association for the Study of Animal Behaviour Studies of the courtship songs of fruit flies have most often concentrated on their potential role in speciation (e.g. Bennet-Clark & Ewing 1969 ; Kyriacou & Hall 1982 ; Ewing & Miyan 1986 ; Tomaru & Oguma 1994a ; Tomaru et al. 1995 ) or on their genetic control (Cowling 1980 ; Kyriacou & Hall 1986 ; Hoikkala & Lumme 1987 ; Tomaru & Oguma 1994b ; Pugh & Ritchie 1996 ; Ritchie & Kyriacou 1996 ). These songs, however, will also play an important role in sexual selection within the species if the female preference discriminates between male songs found within a population. Analysing female preferences for songs is notoriously difficult in Drosophila, as the mating system of fruit flies does not easily lend itself to acoustic playback experiments (Crossley & Bennet-Clark 1993 ). Some experiments have suggested that differences in song of the magnitude seen between species influence female mating speed (Bennet-Clark & Ewing 1969 ; Kyriacou & Hall 1982 ; van den Berg 1988 ; Greenacre et al. 1993 ; Tomaru et al. 1995 ). These experiments have typi- cally been carried out by confining a group of females with mute males and recording the mating speed of the flies while playing different songs through a loudspeaker. One serious drawback in mass mating trials is that one can measure only average group mating speeds. These averages are confounded by marked day-to-day variation in mating speed, requiring large numbers of replicates. Th ese experiments have usually implied that female preferences are broadly rather than narrowly tuned (e.g. van den Berg 1988 ). An interesting exception is Tomaru et al.’s (1995) demonstration that heterospecific song inhibits mating in D. biauraria. It is, however, difficult to distinguish whether the female preferences are really broad, or whether the preferences cannot be precisely determined with the method used. An additional problem is that most of these experiments also fail to distinguish between the effect of song on females and Correspondence: M. Ritchie, Environmental & Evolutionary Biology, Bute Medical Building, University of St Andrews, St Andrews, Fife KY16 9TS, U.K. (email: mgr@st-andrews.ac.uk). A. Hoikkala is at the Department of Genetics, University of Oulu, Linnanmaa, SF-90571 Oulu, Finland. 0003–3472/98/090713+05 $30.00/0  1998 The Association for the Study of Animal Behaviour 713