Meeting Abstract
59.2 Thursday, Jan. 6 A test of a model of visual signal efficacy under natural conditions using Anolis GUNDERSON, Alex R.*; LEAL, Manuel; Duke University; Duke University arg12@duke.edu
Two variables that can impact the probability that a signal is perceived by an intended receiver are the physical properties of the environment and the sensory physiology of the receiver. Over the last decade the development of context-dependent signal detection models, particularly with respect to visual displays, have facilitated important advances in our understanding of signal evolution. Surprisingly, however, the predictions of these models are rarely corroborated under natural conditions. Such validation is critical because of the complex and dynamic visual mosaic (i.e., noise) that pervades most natural habitats. In this study, we experimentally tested the predictions of a signal detectability model in the field under natural habitat light conditions using the lizard Anolis cristatellus as a model. Anoles posses a colorful throat fan called a dewlap that is displayed during social interactions. Anolis cristatellus exhibits geographic variation in dewlap coloration that correlates with habitat type. Populations in dark mesic forests have highly reflective and transmissive dewlaps, while populations in bright xeric forests have darker, more opaque dewlaps. A model of A. cristatellus visual perception predicts that mesic and xeric dewlaps are most detectable in their native habitat at the expense of detectability in the other habitat. We created fake dewlaps that mimic the spectral properties of real mesic and xeric dewlaps, and presented them to lizards in each habitat using a remote controlled “dewlap display apparatus”. Consistent with model predictions, the native fake dewlap had a higher detection probability than the non-native fake dewlap in each habitat. However, this pattern was statistically significant only in the mesic habitat. We discuss the potential implications of these findings for the sensory drive model of signal evolution.
