“Using human vision to detect variation in avian coloration: How bad is it?”
Zachary T. Bergeron and Rebecca C. Fuller (Feb 2018)
Human vision detects most of the variation in animal color in the visible range
Biologists have long known that animals differ in their visual systems. Species often differ in the number and types of cone cells in their eyes. These differences in visual systems have led to the conclusion that human vision can never be used to assess variation in animal coloration. This idea stands in contrast with the fact that many productive research programs on color pattern evolution have relied on human vision. Are humans missing the majority of color variation in nature and, instead, only investigating a narrow subset of variation? Admittedly, humans cannot detect ultraviolet, far-red or polarized signals. However, the question remains as to whether humans are failing to detect the bulk of color variation in the visible range.
Bergeron and Fuller addressed this question by using a methodology that should be flawed by human subjectivity and comparing it to a methodology that is free of human perceptual biases. They compared the coloration of bird specimens from a museum with the coloration of bird images from a field guide (see image at right). They found that field guide images detected the vast majority of the variation in coloration that was present in the museum specimens. This means that human vision detects the major patterns in coloration in the visible range. Human vision cannot be used to say which color patterns are more or less conspicuous to another species, but it can be used to detect major patterns in animal coloration in the visible spectrum in nature.
Assessing variation in animal coloration is difficult as animals differ in their visual system properties. This has led some to propose that human vision can never be used to evaluate coloration, yet many studies have a long history of relying on human vision. To reconcile these views, we compared the reflectance spectra of preserved avian plumage elements with two measures that are humans biased: RGB values from digital photographs and the corresponding reflectance spectra from a field guide. We measured 73 plumage elements across 14 bird species. The field guide reflectance spectra were drastically different from that of the actual birds, particularly for blue elements. However, principal components analyses on all three data sets indicated remarkably similar data structure. We conclude that human vision can detect much of the variation in coloration in the visible range, providing fodder for subsequent studies in ecology, evolution, behavior, and visual ecology.