In the adult cat injections of retrograde fluorescent tracers near the border between areas 17 and 18 and extending to the underlying white matter label neurons in restricted parts of nine other ipsilateral visual areas. A very similar, restricted distribution of retrograde labeling is found in newborn kittens when injections near the 17/18 border are confined to the cortical gray matter. When, however, the neonatal 17/18 border injection reaches the underlying white matter, more visual areas and numerous nonvisual areas become labeled, each of them over nearly its whole tangential extent. Labeled nonvisual areas include the primary and secondary auditory areas, the auditory areas of the posterior ectosylvian gyrus, areas 7 and 5, the cingulate gyrus, and the primary and secondary somatosensory areas. The widespread labeling in kittens was not due to larger or differently placed injections, since the distribution and extent of retrograde labeling in the ipsilateral lateral geniculate nucleus were similar at all ages. The transitory projections from the auditory and somatosensory areas are not reciprocated by a projection from areas 17 or 18. In kittens injected around the end of the first postnatal month the distribution of labeled association neurons is similar to that found in the adult; i.e., many of the juvenile projections have been eliminated. Only a few of the transitory axons to areas 17 and 18 enter the gray matter; the others remain confined to the white matter. Some of these axons were anterogradely labeled with rhodamine-B-iso-thiocyanate from the auditory cortex; they show bulbous endings, some of which are probably growth cones. Retrograde double-labeling experiments showed that, in the newborn, some neurons on the lateral sulcus have at least two long collaterals, one running rostrally, the other caudally; such branching is not observed in adults. In conclusion, areas 17/18 receive at birth from a large, continuous territory including areas, or parts of areas, which will later eliminate these projections. Most of the transitory projections do not appear to enter the cortex to any great extent. The major reshaping of association projections occurs before the end of the first postnatal month. The development of association projections resembles that of callosal projections.