The evolution of visual pigment spectral sensitivities is probably influenced by the reflectance spectra of surfaces in the animal's environment. These reflectances, we conjecture, fall into three main classes:
i. Most inorganic and many organic surfaces, including tree bark, dead leaves and animal melanin pigmentation, whose reflectance increases gradually as a function of wavelength,
ii. Living leaves, which contain chlorophyll, have a sharp reflectance peak at about 555 nm.
iii. Flowers, fruit and other signaling colours that have co-evolved with animal vision typically do not reflect strongly at the same wavelength as leaves, and present a colour contrast against a leafy background. These three spectral functions we call

grey-red

,

leaf-green

and

leaf-contrast

respectively. This simple categorisation allows us to interpret the spectral tuning of human cone pigments in a way that might not seem possible given the wide variety of colours present in nature. In particular
L-(red) cones will capture the highest possible proportion of photons reflected by leaves, and
M-(green) cones will capture about 10% fewer photons both from leaves and from

grey-red

surfaces. These observations have some clear implications for our understanding of the evolution of trichomacy and the trade-off between chromatic and luminance vision in Old-World Primates.