Meeting Abstract
Understanding the relationship between form and function is a major goal in evolutionary biology. In complex mechanisms, function is the product of several phenotypic components, and evolution of the functional system is affected by the interactions between those components. Therefore, the key to understanding the evolution of complex functional systems lies in the ability to predict how the multiple phenotypic characters interact in determining performance. Performance landscapes enable the mapping of several phenotypic characters to their performance value. In this work, we constructed a performance landscape to describe the relationship between phenotypic characters and performance in the suction feeding mechanism of fish. Morphology and kinematics were collected in the field for 5 populations of Chromis viridis feeding in their natural coral reef habitat, using high-speed three-dimensional video footage. Using random phenotypic combinations, sampled as permutations from the observed phenotypic distribution, we constructed a continuous, multi-dimensional performance landscape, statistically modeling the relationship between multiple feeding-related phenotypic traits and feeding performance. We then used randomization tests to determine the effect of natural selection, portrayed by the performance landscape, on the intra-specific phenotypic distribution of this functional system. Our results indicate that the suction-feeding mechanism evolves under a complex selective regime, driven by the adaptive value of specific phenotypic combinations, and constrained by several correlations between phenotypic characters. Our study demonstrates that functional systems can evolves on a performance landscape, and it is now possible to apply our framework for other systems.
