on dy, S

R st he tia 20 he s v pa es. pec te le.spent more time near this stimulus. In contrast, females of the gynogenetic poecilid Poecilia formosa showed an equal preference for a live and video image of a P. mexicana male, suggesting a response to live animals as strong as to video images. We discuss differences between the species that may explain their opposite reaction to video images. 2000 The Association for the Study of Animal Behaviour Playbacks have been widely used in the study of animal communication. They are potentially the most powerful experimental tool available to investigate the significance of animal signals (McGregor et al. 1991). Audio playbacks have been successfully applied to the study of acoustic communication for more than 50 years (Falls 1991) and issues such as female mate choice and species recognition (e.g. Myrberg et al. 1986) or individual recognition (e.g. Myrberg & Riggio 1985) have been addressed with this technique. Video playback methods have recently become popular because of developments in video and computer tech-nology. Experimental set-ups in which the video play-back technique is used instead of a live stimulus have many advantages: it allows the manipulation of a single variable in the stimulus (e.g. size, behaviour) while con-trolling for all others; it eliminates variation in motiv-ation, physical appearance or effects of experience of the stimulus individuals; from an ethical standpoint, fewer manipulated animals can be used or predators can be simulated; and chimera stimuli can be created (e.g. a female with male-like behaviour; Kodric-Brown & Nicoletto 1997; D’Eath 1998; Körner et al. 1999; Landmann et al. 1999). This technique is controversial, however (D’Eath 1998; Fleishman et al. 1998) and has produced conflicting results. Successful experiments have been conducted on several groups of animals including spiders (e.g. Clark & Uetz 1990), birds (e.g. Evans & Marler 1992), reptiles (e.g. Macedonia & Stamps 1994) and fish (e.g. Rosenthal et al