Comparing colors using visual models

An outstanding challenge for the study of color traits is how best to use “colour spaces” to represent their visual perception, particularly when asking questions of color difference (e.g. the (dis)similarity of males and females, mimics and models, or sister species, to a given viewer). We use simulations to show that existing methods fail to statistically and biologically estimate the separation of groups in color space, and we suggest a flexible, robust, alternative that avoids those pitfalls.