The role of detectability in the evolution of avian-dispersed fruit color.

ABSTRACT

If the primary function of avian-dispersed fruit coloration were the maximization of detectability, then the commonest avian-dispersed fruit colors should be the ones most detectable to birds. We tested this prediction by photographing 63 fruit species primarily dispersed by birds, in situ in Sweden and Australia, with a multispectral camera closely mimicking the predominant spectral sensitivities of birds, including both UVS and VS (peak ultraviolet sensitivity ∼370 and 409 nm respectively) visual systems. Fruits were classified into nine distinct color categories based on different patterns of cone excitations, and were named by combining human color names with fruits' UV reflective properties. For example, a bluish-UV fruit would be a fruit that excited the avian UV cone the most, but that also strongly excited the blue cone. Color and achromatic contrasts were calculated between each fruit color and common background objects, and compared to the relative abundance of the different fruit colors. Although red was highly detectable and the commonest color, the second and third commonest colors, purplish-UV and bluish-UV (often termed "black" by humans), were the least detectable. Although these latter two colors were more detectable to UVS than to VS birds, they were the least detectable to both visual systems. Rare fruit colors, such as UVish-purple, pink, and orange, were highly detectable to both visual systems. The lack of correlation between fruit color abundance and detectability suggests that the maximization of detectability has not been the primary driving force behind the evolution of fruit color.