The effects of aging on human visual function are poorly understood; some have been attributed to neural changes in the retina such as neuronal losses. The potential losses may be detected in measurements of the angular density of cones and ganglion cells. These measurements and their comparison are hindered by differences in retinal location and areal magnification, which are due to a radial displacement between cones and postreceptoral cells reaching a peak in the fovea, and which affect postreceptoral density measurements in terms of visual angle (cells/degree square). Previously collected retinas from four human donors with an age range of 75-85 years, who lacked history of eye disease, were histologically processed in a study of cone and ganglion-cell densities. Initial results show a mean peak density of 16,000 cones/degree square. Comparisons with the results of others show that the data are consistent with those from younger human donors of recent studies and reveal an error in the central data of a well-known, older study by Osterberg, which is still occasionally used as a reference. In contrast, the ganglion-cell results suggest a decline in extrafoveal cell density with age; comparisons with older studies are hindered by the absence of proper magnification and histological shrinkage corrections in such studies and, in one case, by a computational error. Measurements of the displacement between cone pedicles and inner segments along the part of the retina of donor eye age 85 years are similar to those obtained in more complete measurements of macaque retina. To compare angular densities, such displacements are used to compensate for differences in retinal location between cones and ganglion cells, and their associated difference in areal coverage can then be compensated by a factor representing the square of the ratio of the separation between two points at the cone level to that between two corresponding points at the ganglion-cell level. At the foveal level, the ratio of cone to (corrected) ganglion-cell density was 0.83; given the data dispersion of 8 percent to 16 percent of the mean and the estimation by others that 10 percent of the density in the ganglion-cell layer is due to displaced amacrine cells, the results cannot exclude a unity density ratio in the fovea. The ratio is larger than unity at higher retinal eccentricity, however, because cone density declines less rapidly than that of ganglion-cells, and this effect appears to be exaggerated by peripheral ganglion-cell losses in older donors.
