On the spurious occurrence of Tit for Tat in pairs of predator-approaching fish

An experimental analysis of the movements of predator-approaching fish is presented. The experiments evaluated two competing hypotheses. (1) Predator-approaching fish play the gametheoretical strategy Tit for Tat. Alternatively, (2) the movements of predator-approaching fish superficially resemble Tit for Tat, because fish independently orient to a predator and simultaneously attempt to stay close together. Experimental subjects were mosquito fish, Gambusia affinis, approaching a green sunfish, Lepomis cyanellus. Two experiments were performed. Experiment 1 replicated results of Milinski (1987) and Dugatkin (1991), showing that Gambusia come closer to a visible predator when a mirror is oriented parallel to their direction of travel. Experiment 2 attempted to separate the effects of common orientation and social cohesion in accounting for the frequency of Tit-for-Tat-like motions in pairs of predator-approaching Gambusia. Results of experiment 2 suggest that a simple additive combination of the effects of (1) social cohesion in the absence of a visible predator and (2) orientation to a visible predator in the absence of a visible companion can account for the observed frequency of Tit-for-Tat-like motions for pairs of predator-approaching Gambusia. It is concluded that predator approach in shoaling fishes is probably a simple by-product mutualism, rather than cooperation maintained by reciprocity in a Prisoner’s Dilemma. ? 1997 The Association for the Study of Animal Behaviour In 1981, Axelrod & Hamilton published an influential paper that offered a seemingly general framework for understanding the evolution of non-kin cooperation. Axelrod and Hamilton’s framework has two components: a game and a strategy. First, they argued that Flood and Drescher’s ‘Prisoner’s Dilemma’ game typified the problems of achieving cooperation in a selfish world (see Mesterton-Gibbons 1992 for a precise definition of the Prisoner’s Dilemma and a review of technical issues). Second, they argued that stable mutual cooperation could be achieved in this paradigmatic game if (1) the game was played repeatedly, and (2) the players adopted a strategy they called ‘Tit for Tat’. A player adopting the Tit-for-Tat strategy cooperates on the first ‘play’ of a sequence and copies its opponent’s behaviour on all subsequent plays; for example, a Tit-forCorrespondence: D. W. Stephens, Nebraska Behavioral Biology Group, School of Biological Sciences, University of Nebraska, Lincoln, NE 68588-0118, U.S.A. (email: [email protected]). 0003–3472/97/010113+19 $25.00/0/ar960283 ? 1 11 Tat player cooperates on the tenth play, if its opponent cooperated on the ninth, and so on. Predator Approach as Tit for Tat Milinksi (1987) offered what seemed to be the first plausible experimental system for the analysis of Axelrod and Hamilton’s framework. Sticklebacks, and many other small fishes, will often approach larger potentially predacious fish rather than fleeing, as the naive observer might expect. Milinski studied the phenomenon of predator approach in the stickleback Gasterosteus aculeatus. Milinski, following Pitcher et al. (1986), favoured an ‘inspection’ interpretation of this approach behaviour, in which approaching sticklebacks are thought to obtain information about the state of the predator (e.g. whether it is hungry or aggressive). With this interpretation in mind, he argued that the economics of pair-wise inspection represented a Prisoner’s Dilemma, and he therefore predicted that sticklebacks approaching a predator would play Tit for Tat. Milinski 997 The Association for the Study of Animal Behaviour