Client fish traits underlying variation in service quality in a marine cleaning mutualism

Abstract

Game-theoretical models help us understand how and when cooperation can evolve and persist. However, current models fall short of explaining the striking amount of variation in cooperation levels that we observe in nature, even within a system. For example, an animal's ability to choose partners with which to interact can explain the maintenance of cooperative interactions, but not variation in cooperation levels among individuals with similar partner choice options. Here, we explored how partner choice options and other key partner traits predict the quality of the service provided by cleaner fish, Labroides dimidiatus (the ‘chosen’ partner) to their client reef fishes (the ‘choosy’ partner). In this marine cleaning mutualism, cleaner fish sometimes cheat and eat mucus from their clients rather than cooperate by removing ectoparasites. We examined 13 client fish species and assessed the influence of nine variables on cooperation levels by cleaner fish (i.e. the frequency of cheating events) using information theory. Variables characterizing clients and the interactions themselves were used as proxies for a client's nutritional value, partner choice options, ability to punish a cheating partner and a cleaner's temptation to cheat its client (some of these traits were characterized by more than one variable). We found that six of the nine variables were equally important in explaining variation in client jolts: the duration of an interaction, and traits of client species including their body size (a proxy for the quality of a food patch), gnathiid ectoparasite load (nutritional value), partner choice options, agility (ability to control an interaction via avoidance or termination) and mucus characteristics (a measure of the cleaners' temptation to cheat). These findings suggest that numerous factors influence variation in cooperation levels, showing a parallel with sexual selection theory, where female mate choice does not depend on a single factor such as the operational sex ratio, but also on other key variables including a male's quality relative to other males, and a female's ability to avoid male coercion.