Pairs of cleaner fish prolong interaction duration with client reef fish by increasing service quality

Few biological examples of cooperation seem to precisely fit the assumptions of an iterated prisoner’s dilemma. In an attempt to increase biological validity, one model altered the assumption that cooperating is an all-or-nothing decision to a situation where benefits are a function of interaction duration, which in turn is a function of the levels of cooperation. A potential application involves pairs of cleaner fish coinspecting a client fish. In this mutualism, clients visit cleaners to have ectoparasites removed but a conflict of interest exists, as cleaners prefer to eat client mucus, which constitutes cheating. As large clients often flee in response to a cleaner cheating, pair inspections lead to a dilemma: the cheater obtains the benefit while both cleaners share the cost of the client leaving. The model predicts that pairs of cleaners behave more cooperatively toward reef fish clients than when inspecting alone, to entice clients to profit from the increased parasite removal rate and keep interaction duration almost constant. Here, we present field experiments that first replicate results that pairs behave indeed more cooperatively than when inspecting alone and second show that levels of cooperation quantitatively predict the duration of cleaning interactions. We also found that several additional variables may affect the duration of cleaning interactions, such as a client’s willingness to interact with a cleaner, identity of interaction terminator, and the presence of bystanders. In conclusion, introducing benefits as a function of interaction duration into the prisoner’s dilemma framework provides a biologically relevant framework to study cooperation.