Bird Flocking Revisited: The Case with Polymorphic Prey

Abstract

In recent years it has been argued that predators may select for visual polymorphism in their prey species. This form of selection, known as apostatic selection (Clarke 1962), occurs when the predator takes common morphs in greater proportion than their occurrence in the population and rare morphs in lower frequency than they occur in the population. This phenomenon has been invoked, for example, to explain the maintenance of polymorphisms in several species of snails (Cain & Sheppard 1954; Clarke 1960, 1969; Owen 1965). Apostatic selection on artificial baits has been demonstrated for various bird species (Allen & Clarke 1968; Allen 1972; Manly, Miller & Cook 1972). Confounding the situation are the difficulties of accounting for density-independent preferences of the predator (Manly et al. (1972) present a method of handling this difficulty) and the apparent disadvantage of polymorphism to the prey species at high densities (Allen 1972; Greenwood 1969; Holling 1965) and possibly at very low densities as well (Greenwood 1969). Finally, we note that in Croze's (1970) study the combined attack rate on a trimorphic population was smaller than on any of the three monomorphic populations (overall density being kept constant). Accepting that polymorphism may be an effective strategy against a bird predator, one may ask what counterstrategies are available to the birds. In this study we investigate one such possibility, examining the value of flocking as a means to increase the birds' feeding success on a polymorphic prey population.* The study was carried out using a simulation model developed to investigate the survival value of bird flocking in relation to prey capture (Thompson, Vertinsky & Krebs 1974). In the previous study using monomorphic prey populations at low densities, it was found that flocking neither increased nor decreased mean prey capture rate over a wide range of prey distributions, but that flocking did serve to reduce the risk to a bird of failing to feed at a level sufficient for winter maintenance. As the model is described in detail in the cited work, only a brief summary of key features will be given here. The simulated birds are visually discriminating predators which form short-term prey preferences, corresponding to a combination of search images (e.g. Tinbergen 1960; Dawkins 1971a, b) and of 'prey niches' (e.g. Royama 1970). A modelled bird making two successive captures of the same prey type forms a prey preference for that type. Hence, we can expect that they will be less effective predators on a polymorphic prey population than on a monomorphic one provided that the prey encounter rate exceeds the reinforcement rate required to maintain the prey preference. When the encounter rate is too low, prey preferences are rarely * Since the distinction in the model is between prey 'types', these types may represent different similar prey species rather than morphs of a single species. Hence, this study also pertains to the coexistence of separate prey species as a result of density dependent-predation (e.g. Murdoch 1969). Throughout the paper we will use the term polymorphism to refer to either of these cases.