Meeting Abstract
Performances surfaces are representations of the relationship between phenotype and performance in a particular function. They are often employed to explore patterns of phenotypic diversification. However, many structures perform multiple functions and are associated with multiple performance surfaces. There is no consensus regarding methods for combining multiple performance surfaces in diversification analysis, partially due to the difficulty of determining how to weight performance surfaces. I present three methods for combining performance surfaces, none of which require known estimates of the relative importance of functions. Performance surfaces can be used to identify regions of phenotypic space in which simultaneous increases in performance for all functions are possible, and hence where species are not expected to occur. They can be used to identify functional “ridges” along which organisms are expected to occur, for all possible relative weights of different functions. Finally, they can be used to estimate the relative importance of various functions for individual species, as well as the amount of variation which cannot be explained by performance for any of the functions under consideration. These techniques are illustrated on shell shape data for 261 species of turtle and performance surfaces for shell strength and hydrodynamics. The surfaces predict the distribution of turtles fairly well, but only approximately 30% of shell shape variation is explained by performance on these two functions. The combination of performance surfaces has enormous potential, not only for determining whether certain functions have influenced patterns of phenotypic diversification, but also for illustrating which aspects of variation remain unexplained by functional performance.
