Resolving the paradox of local warning signal diversity: A framework for hypothesis testing

Abstract Defended prey often evolve conspicuous colorations that advertise their unprofitability to predators. These visual advertisements act as honest warning signals that predators can use to avoid these prey through associative learning. Müller's pioneering idea aptly explains the often‐observed signal convergence among unprofitable prey (i.e., Müllerian mimicry) as the result of positive frequency‐dependent selection. We therefore expect aposematic prey, at least those targeted by the same predators, to be homogeneous in signal appearance at local scales. Yet, it is common to observe diverse warning signals coexisting within communities. The simultaneous occurrence of warning signal convergence and diversity at local scales represents a stark contrast with theoretical expectations and has long stood as a paradox. Despite much interest in this phenomenon, there has not been an article that summarizes possible resolutions of the paradox and, more importantly, offers an empirical framework for hypothesis testing. We aim to fill these knowledge gaps by summarizing possible resolutions to this paradox, including (1) relaxed predation, (2) microhabitat segregation, (3) source–sink dynamics, (4) retentive predators, (5) immigration boost, (6) immigration rescue and (7) imminent signal extinction. We also propose a hierarchical hypothesis‐testing framework leveraging data from population surveys, genetic analyses and field learning experiments to elucidate the plausibility and contribution of these mechanisms. Our proposed framework would help resolve this long‐standing paradox in natural communities. By applying this framework to communities harbouring diverse warning signals and differing in their ecological environments, we can test hypotheses regarding how local warning signal diversity is being maintained at larger scales, and how ecology might have shaped the nature of resolutions. Read the free Plain Language Summary for this article on the Journal blog.