The unsuitability of html-based colour charts for estimating animal colours – a comment on Berggren and Merilä (2004)

BACKGROUND: A variety of techniques are used to study the colours of animal signals, including the use of visual matching to colour charts. This paper aims to highlight why they are generally an unsatisfactory tool for the measurement and classification of animal colours and why colour codes based on HTML (really RGB) standards, as advocated in a recent paper, are particularly inappropriate. There are many theoretical arguments against the use of colour charts, not least that human colour vision differs markedly from that of most other animals. However, the focus of this paper is the concern that, even when applied to humans, there is no simple 1:1 mapping from an RGB colour space to the perceived colours in a chart (the results are both printer- and illumination-dependent). We support our criticisms with data from colour matching experiments with humans, involving self-made, printed colour charts. RESULTS: Colour matching experiments with printed charts involving 11 subjects showed that the choices made by individuals were significantly different between charts that had exactly the same RGB values, but were produced from different printers. Furthermore, individual matches tended to vary under different lighting conditions. Spectrophotometry of the colour charts showed that the reflectance spectra of the charts varied greatly between printers and that equal steps in RGB space were often far from equal in terms of reflectance on the printed charts. CONCLUSION: In addition to outlining theoretical criticisms of the use of colour charts, our empirical results show that: individuals vary in their perception of colours, that different printers produce strikingly different results when reproducing what should be the same chart, and that the characteristics of the light irradiating the surface do affect colour perception. Therefore, we urge great caution in the use of colour charts to study animal colour signals. They should be used only as a last resort and in full knowledge of their limitations, with specially produced charts made to high industry standards.