Cautionary notes on the descriptive analysis of performance curves in reptiles

Organismal performance curves are important functions for the study of reptilian ecology and evolution, but their interpretation can be affected strongly by the choice of analytical approach. We first use an example from the literature to demonstrate that excluding biologically meaningful data alters the description of performance and leads to non-sensical inferences. We then use fictional data to show that linear models (ANOVA) commonly used in the descriptive analysis of performance curves can also be biologically misleading or can lack biological relevance. Out-examples demonstrate that fitting non-linear curves to performance data is more meaningful and avoids erroneous representation and interpretation of these important biological functions. (c) 2006 Elsevier Ltd. All rights reserved.