We strictly limited small injection of horseradish peroxidase (HRP) to lamina A of the lateral geniculate nucleus of cats. This was done to label retrogradely only the alpha (Y) and beta (X) classes of retinal ganglion cell. Eighty-six such injections at a range of matched eccentricities were made bilaterally in 9 normal adult cats, 7 cats reared from birth to adulthood with monocular lid suture, and 9 normal kittens at 4 weeks of age; 5348 alpha and beta cells were retrogradely labeled from these injections. Quantitative measurements were made from these labeled cells and compared among 4 experimental conditions, these being normal adult retinas, the nondeprived and deprived retinas of lid sutured cats, and the retinas of kittens. Each injection led to a similar relative ratio of labeled alpha and beta cells (typically 5-15% alpha cells) that did not differ significantly among the experimental conditions, but further analysis suggested a slight diminution of labeled alpha cells in deprived retinas. Because the larger arbors of retinogeniculate Y axons are more likely to penetrate small geniculate HRP injection sites from eccentric locations than would be the case for the more restricted arbors of X axons, a normal tendency resulted for the peripheral halo of zones of retrograde labeling to be dominated by alpha cells. Thus a more accurate reflection of the relative numbers of labeled alpha and beta cells would result from considering only the core of zones of retrograde labeling. When this is done, deprived retinas exhibited relatively fewer labeled alpha cells than did normal, nondeprived, or kitten retinas. This may relate to prior observations (Sur et al. 1982) that abnormally few Y axons from the deprived retina innervate lamina A. No statistically significant differences in alpha or beta cell size were seen among normal, nondeprived, and deprived retinas, although both of these cell types in the kittens were equally smaller than their normal adult counterparts. This is particularly interesting in view of the postnatal growth of retinogeniculate axon arbors (Sur et al. 1984). The results are not surprising for alpha cells, since retinogeniculate Y axon arbors grow considerably after 4 weeks of age, but they are surprising for beta cells, since retinogeniculate arbors of X axons decrease after 4 weeks of age. This suggests no clear, general relationship between soma size and the extent of a cell's axonal arbor.(ABSTRACT TRUNCATED AT 400 WORDS)