COS 104-3
The struggle for safety: Effectiveness of caterpillar defenses against bird predation

Thursday, August 14, 2014: 8:40 AM
315, Sacramento Convention Center
Isaac H. Lichter Marck, Biology Department, Wesleyan University, Middletown, CT
Micah Wylde, Department of Mathematics and Computer Science, Wesleyan University
Eric Aaron, Department of Computer Science, Vassar College
Jeffrey C. Oliver, Department of Zoology, Oregon State University
Michael S. Singer, Biology Department, Wesleyan University, Middletown, CT
Background/Question/Methods

The effectiveness of anti-predator traits has rarely been quantified in a complex ecological community. Here we quantify the effect of several putative anti-predator traits of caterpillars on bird predation risk. We synthesize eight years of field and laboratory study of a temperate forest community, including a four-year bird exclusion experiment that provided comparative measures of bird predation risk for thirty-eight caterpillar species. We then conducted a comparative analysis of warning signals (aposematism or mimicry), camouflage (crypsis or masquerade), and behavioral responses to physical attack as determinants of bird predation, while also testing for effects of sample size, maximum population density, mean bodysize, phenology, and phylogenetic relationships of caterpillar species.

Results/Conclusions

Consistent with aposematism theory, warning signals had the largest defensive effect, especially for the most behaviorally specialized caterpillar species. Superior camouflage was also associated with reduced bird predation in some analyses. Caterpillar behavioral responses to physical attack, however, did not predict bird predation risk. Therefore, anti-predator traits involving visual signaling were the most important defenses against bird predation. These findings show the functional significance of anti-predator traits of externally feeding caterpillars in the context of a complex community and provide a framework for future study of focal taxa or particular traits.