A computer program was used to generate "populations" of dimorphic "prey" on the screen of a colour monitor. Different subjects were presented with the prey at seven different frequencies and were asked to use a light pen to remove each prey they detected. They all received the same two types of prey but 70 had them presented against a matching background and 49 had them against a background that made them conspicuous. The results showed that apostatic selection occurred when the prey were inconspicuous but not when they were conspicuous. There is evidence that the apostatic selection was caused by some effect of the difficulty in detecting the prey when they were cryptic.
INTRODUCTIONA preference for common prey-"apostatic selection" (Clarke, 1962) or "switching" (Murdoch, 1969)-has been demonstrated in several studies using artificial prey (e.g., Allen and Clarke, 1968;Allen, 1972Allen, , 1974Allen, , 1976Manly, Miller and Cook, 1972;Fullick and Greenwood, 1979;Raymond, 1984) and natural prey (Den Boer, 1971;Maskell et a!., 1977;Cornell and Pimentel, 1978;Akre and Johnson 1979;Bergelson, 1985). The magnitude of the response has been shown to be affected by the prey's density (Cook and Miller, 1977), palatability (Greenwood, Wood and Batchelor, 1981) and degree of crypsis (Cooper, 1984;Bond, 1983).Theoretical work has demonstrated that apostatic selection has the potential to maintain colour polymorphisms in prey species (Haldane and Jayakar, 1963;Clarke and O'Donald, 1964;Cook, 1965), although its efficacy in nature has yet to be proven (Cain, 1983). Very little is understood of its behavioural basis. Clarke (1962) suggested that the development of specific search images (Tinbergen, 1960) is the primary cause but it is now clear that a wide range of behaviours can in theory give rise to frequency-dependent selection by predators (Murdoch, Avery and Smythe, 1975;Greenwood, 1984 The experiments described in this paper were designed to test whether humans select in an apostatic manner when "killing" computer-generated dimorphic "prey" on the screen of a colour monitor. Despite the obvious pitfalls, the use of humans has given valuable insights into the development of predator-prey models (Holling, 1959;Gendron and Staddon, 1984). The computercontrolled technique has the advantage of giving the experimenter considerable control over important variables, as well as providing a means for the automatic and accurate recording of data. In the experiments the subjects simulated predation by using a light pen to remove prey from the screen until the trial was stopped automatically at the point when half the prey had been eliminated.There is evidence from wild birds (Cooper, 1984) and captive pigeons (Bond, 1983) that a matching background increases the magnitude of the strength of apostatic selection. The behavioural basis of this effect is in doubt, but it could result from a greater ability of the predator to learn to distinguish prey from the background when they are common ratherthan rare (Bond, 1983;Cooper, 1984). T...