Use of spectroscopy for assessment of color discrimination in animal vision

Abstract

Animals use color vision for a number of tasks including food localization, object recognition, communication, and mate selection. For these and other specific behaviors involving the use of color cues, models that quantify color discriminability have been developed. These models take as input the photoreceptor spectral sensitivities of the animal and radiance spectra of the surfaces of interest. These spectra are usually acquired using spectroscopic instruments that collect point-by-point data and can easily yield signals contaminated with neighboring colors if not operated carefully. In this paper, I present an equation that relates the optical fiber diameter and numerical aperture to the measurement angle and distance needed to record uncontaminated spectra. I demonstrate its utility by testing the discriminability of two solid colors for the visual systems of a dichromatic ferret and a trichromatic frog in (1) a conspicuous scenario where the colors have little spectral overlap and (2) a perfect camouflage scenario where the spectra are identical. This equation is derived from geometrical optics and is applicable to spectroscopic measurements in all fields.