We studied the effects of early unilateral blur on the anatomical organization of the visual pathways in 8 macaque monkeys. Blur was induced in one eye, beginning 2-14 d after birth, by 0.5% atropine twice a day. Atropinization was stopped at 6-8 months of age, and the animals were studied for anatomy 3-24 months later. The retina and all other eye tissues showed normal histology. In the dorsal lateral geniculate nucleus (LGN), cells in parvocellular layers receiving input from the atropine-treated eye were 9-32% smaller and were more lightly stained than those in layers innervated by the untreated eye. These changes were generally larger in the LGN ipsilateral to the treated eye. LGN cell size changes were absent or much smaller in the magnocellular layers. In the striate cortex, the distribution of the oxidative enzyme cytochrome oxidase (CO) was markedly altered in layer 4C beta. Layer 4C beta is uniformly stained in normal animals, but showed a distinct pattern of alternating high and low CO bands in the atropine-treated animals; the bands of higher CO activity were narrower than the bands of lower activity and had a 857-1050 micron repeat. Fainter banding was seen in layers 4A, 4C alpha, and 6, but the density of the rows of dark CO-stained dots in layer 3 was unaffected. Double-labeling revealed that the narrow dark CO bands were associated with the centers of the ocular dominance columns devoted to the atropine-treated eye. The distribution of 14C-2-deoxyglucose uptake in visual cortex produced by 4.5-9 c/deg spatial frequency stimulation was strongly biased toward the untreated eye. The treated eye could, however, elicit reasonably strong uptake when stimulated with patterns containing lower spatial frequencies. These results suggest that unilateral neonatal blur preferentially affects the parvocellular layers of the LGN and layer 4C beta of striate cortex, which are the portions of the central visual system associated with the processing of information concerning fine spatial detail. These anatomical changes are consistent with the high spatial frequency loss of vision demonstrated behaviorally and electrophysiologically in the atropine eye-driven visual system of these same animals.