This study addresses the correlation of retinal topography with factors such as the visual environment, life style, and behavior for a major mammalian group, the artiodactyls. To provide a broader basis for semiquantitative comparison, short-wavelength-sensitive (S)- and middle-to-long-wavelength-sensitive (M)-opsin cone receptor populations from 25 species from five artiodactyl families and of the African elephant were labeled and sampled. The resulting topographic maps were analyzed with respect to the position and extension of high-density regions. For better parameter differentiation, systematic relationships were statistically normalized. In all species examined, two classes of cones have been detected. In most species, the S-cone maxima were located in the temporodorsal retina, but there are exceptions such as the roe deer with accumulation in the ventral retina. For M-cones, as a consequence of their role in terrain/food assessment and predator detection, the standard topography is L-shaped: a horizontal visual streak including a temporal area centralis is extended by a temporal rim. Its extension is correlated with the animal's body height (P = 0.0017): small species (pudu, mouse deer) tend to have a visual streak only, whereas the giraffe shows a complete dorsal arch of elevated densities. Furthermore, a size-independent habitat correlation was revealed for a similar M-cone pattern (P < 0.0001): mountainous species show a striking specialization around the dorsal retina, pointing to the importance of the inferior visual field in precipitous terrain.

In mammals, cone photoreceptor subtypes are thought to establish topographies that reflect the species-relevant properties of the visual environment. Middle- to long-wavelength-sensitive (M) cones are the dominant population and in most species they form an area centralis at the visual axis. Short-wavelength-sensitive (S) cone topographies do not always match this pattern. We here correlate the interrelationship of S and M cone topographies in representatives of several mammalian orders with different visual ecology, including man, cheetah, cat, Eurasian lynx, African lion, wild hog, roe deer, and red deer. Retinas were labeled with opsin antisera and S and M cone distributions as well as S/M cone ratios were mapped. We find that species inhabiting open environments show M cone horizontal streaks (cheetah, pig, deer). Species living in structured habitats (tiger, lynx, red deer) have increased S cone densities along the retinal margin. In species with active vision (cheetah, bear, tiger, man), S cone distributions are more likely to follow the centripetal M cone gradients. Small species show a ventral bias of peak S cone density which either matches the peak of M cone density in a temporal area centralis (diurnal sciurid rodents, tree shrews) or not (cat, manul, roe deer). Thus, in addition to habitat structure, physical size and specific lifestyle patterns (e.g. food acquisition) appear to underlie the independent variations of M and S cone topographies.