An experimental test of whether cheating is context dependent.

Microbial cells rely on cooperative behaviours that can breakdown as a result of exploitation by cheats. Recent work on cheating in microbes, however, has produced examples of populations benefiting from the presence of cheats and/or cooperative behaviours being maintained despite the presence of cheats. These observations have been presented as evidence for selection favouring cheating at the population level. This apparent contradiction arises when cheating is defined simply by the reduced expression of a cooperative trait and not in terms of the social costs and benefits of the trait under investigation. Here, we use two social traits, quorum sensing and iron-scavenging siderophore production in Pseudomonas aeruginosa, to illustrate the importance of defining cheating by the social costs and benefits. We show that whether a strain is a cheat depends on the costs and benefits associated with the social and abiotic environment and not the absolute expression of a cooperative trait.