Who Dares, Wins: the Function and Evolution of Predator Inspection Behaviour in Shoaling Fish

Abstract

This paper examines the experimental evidence which underpins our current understanding of inspection behaviour in fish shoals, and reviews ideas about how predator inspection behaviour could have evolved. For fish which are vulnerable to attack in the presence of a predator, inspection is critical in the recognition of danger and in the precise assessment of risk on a second-to-second time scale. Up to a threshold, inspection rates increase with danger, partly because new fish in the shoal begin to inspect. Inspection appears to have a genetic basis, although experience with predators during development can fine-tune it. Inspection is evidently a risky business, as shown indirectly by attack-cone avoidance and directly by predator strikes in experiments. An appraisal of the current state of a predator is transferred from inspectors to other individuals in the group: we do not know whether such information transfer is active or passive, or whether it involves an element of manipulation. In multispecies groups, some species seem to benefit from the inspection of others. There are large individual differences in inspection, so groups probably comprise two types of individuals with differing trade-offs between risk and information. A key question is how such apparently altruistic behaviour can have evolved. Among putative functions which have been examined, inspection seems to facilitate anticipation of attack, but hypotheses of attack inhibition, attack invitation, and approach conditional on danger have not been supported unequivocally by experimental results. The tit-for-tat strategy of alternating cooperation between pairs of fish is a strong candidate explanation for inspection, but frequent observations of singletons and large groups inspecting are not easily compatible with tit-for-tat. The reason for the anomaly may be that inspection is opportunistic: depending on the circumstances, inspecting individuals may choose from a range of mechanisms that mitigate risk of attack. Risk dilution (safety in numbers) operates to form large inspecting groups when perceived risk is high, and cannot be entirely ruled out as an explanation for behaviour in other sizes of shoals. Singletons may bear the costs of inspection alone in order to manipulate the rest of the group to their advantage. Through rapid serial repetition of a standard act at a rate proportional to perceived danger, inspection may itself be an easily detected, discriminated and remembered signal. Its unhurried form may reflect the evolution of deception and self-deception in shoaling fish.