Color change and color-dependent behavior in response to predation risk in the salamander sister species Ambystoma barbouri and Ambystoma texanum.

Although many organisms show multiple types of trait responses to predation risk (e.g., shifts in behavior, morphology, color, chemistry or life history), relatively few studies have examined how prey integrate these multiple responses. We studied the joint expression of color and behavioral responses to predation risk in two sister species of salamander larvae that live in habitats with different selection pressures. We examined responses to predation risk in three situations that differed in availability of refuge and substrate color heterogeneity, and thus availability of behavioral options for reducing risk. Relative to Ambystoma texanum, A. barbouri larvae were darker in color and showed a greater range of color change. With no variation in background color or refuge available, both species exhibited color change to better match the available background. The degree of color change showed by both species, however, did not depend on predation risk. Given the option to choose between light and dark substrates, A. texanum exhibited behavioral background matching (i.e., they preferred substrates that matched their own body color), while A. barbouri's substrate preferences did not depend on their initial body color. Instead, A. barbouri responded to risk by showing a strong preference for dark substrates, followed by a change to a darker body color. With refuge available, A. texanum's refuge use was color-dependent; larvae that were well camouflaged spent less time in refuge. In contrast, A. barbouri showed strong refuge use in response to risk, regardless of their body color. Overall, these results reflect how conflicting selection pressures (predation risk, habitat ephemerality, risk of UV damage) and species differences in mean color and ability to change color can govern the interplay of complementary and compensatory behavioral and color responses to predation risk.