Body size, not habitat or sex, best explains the extent of ultraviolet fluorescence in African dwarf chameleons (Bradypodion)

Abstract

It has been hypothesized that biofluorescence is a trait linked to intraspecific signaling in many taxa, especially those with enhanced modes of conspecific signaling in complex habitats. Chameleons possess bone-based fluorescent tubercles (FTs) on their head ornaments that purportedly facilitate intraspecific signaling. We investigated the hypothesis that dwarf chameleons (Bradypodion) use biofluorescence for signaling by testing if the number of FTs associated with their ornaments can be explained by sexual dimorphism or ecological variation in five species from various habitats (i.e. fynbos, Afrotemperate forest, and shrublands). If the trait is used for signaling, we would expect males to have more FTs than females due to sexual selection, and/or forest species/populations to have more FTs than open-habitat species/populations via natural selection because forests are expected to be the most conducive terrestrial environment for fluorescent signals. Our results revealed that the number of FTs was greater for the larger sex (regardless of the direction of size dimorphism) but was not significantly different between sexes when adjusted for body size or head area. Forest species had more FTs than smaller-bodied fynbos species but fewer than the large-bodied shrublands species in absolute number, but there were no differences in FTs across species from different habitats when corrected for size and phylogeny. Moreover, there were no differences in FTs between natural and urban populations when correcting for body or head size. These findings suggest that larger-bodied species have more FTs than smaller-bodied species regardless of the conduciveness of their habitats toward facilitating biofluorescence. Therefore, FT trait magnitude is likely explained best by chameleon size rather than natural or sexual selection for increased signaling capability between sexes, species, or populations. We interpret these findings to suggest that it is unlikely that Bradypodion use biofluorescence as a signaling mechanism.