Optimization, constraint, and history in the evolution of eyes.

Several features of the evolution of eyes and photoreceptors are examined in an effort to explore the relative roles of adaptation and historical and developmental constraints. Optical design shows clear evidence of adaptation, which in some respects approaches optima predictable from physics. The primate fovea, on the other hand, illustrates how adaptation can be channeled by developmental heritage. The primary structures of opsins reveal multiple evolutionary lineages within both Drosophila and humans. The pigments of vertebrae rods comprise a subset of opsins whose evolutionary relationships map onto the phylogeny of the parent species. The evolutionary reasons for why most rod pigments absorb maximally at 500 +/- 10 nm are obscure, as there is no convincing explanation based on adaptation alone. Rods are appropriately distinguished from cones on the basis of which opsin gene is expressed. This criterion is likely to be in conflict with other definitions in phyletic lines (e.g., geckos, snakes) that have long diurnal or nocturnal histories accompanied by loss of one or more opsin genes, followed by a secondary adaptation to life in a different photic environment. Color vision--a generalizable perception associated with the spectral composition of light--is usefully distinguished from wavelength-specific behaviors. The latter are also based on multiple visual pigments and more than one spectral class of receptors but cannot be altered by learning. The distinction is particularly forceful in bees, which exhibit both kinds of behavior. The evolution of primate color vision has been shaped by historical factors involving an extensive period of early mammalian nocturnality. Birds, by contrast, have more elaborate cones and a richer set of visual pigments. Avian color space can be represented in a tetrahedron.