Predator learning can resolve the paradox of local warning signal diversity

Abstract

Coexistence of distinct warning signals at local scales has long stood as a paradox, as selection is expected to preserve only the most common signal. So far, there has not been an explanation that is both broadly applicable and testable. This study presents a novel and generalizable resolution to this paradox (the Unforgetful Predator Hypothesis) by showing that prey displaying a rarer warning signal can persist if predators have low enough forgetting rates relative to prey generation time. In addition, inducing a high level of predator avoidance facilitates warning signal diversity when prey do not compete. In the presence of prey competition, however, warning signal diversity is more likely to occur if prey elicit intermediate levels of avoidance, such that the competitive disadvantage for individuals displaying the rare signal can be offset by predation. This hypothesis can be tested by quantifying predator avoidance and forgetting rate in laboratory and field experiments. As the level of predator avoidance is also crucial in determining the fate of rare signal in communities, I performed a meta-analysis to examine the determinants of its variation and found that higher predator avoidance can be observed in the following situations - when prey unprofitability is due to toxicity rather than mere unpalatability, when predators search widely for prey and/or prey aggregate, and when predators could differentiate between unprofitable and profitable prey using only color or pattern. In addition to resolving the paradox, these findings help inform the types of communities in which distinct warning signals can stably coexist. SignificanceCoexistence of distinct warning signals within a community represents an evolutionary paradox that still awaits a resolution that is both general and testable. I used ecologically realistic simulations to show that warning signal diversity can occur if predators have long enough memory and if prey elicit either low or moderate levels of avoidance, depending on whether they compete. A meta-analysis further shows that higher levels of predator avoidance tend to occur when unprofitable prey are highly unprofitable, when predators search widely for prey, when prey aggregate, and when prey profitability could be discerned by only color or pattern. These findings offer a testable resolution to the paradox and inform the types of communities where warning signal diversity may occur.