Cell suspensions of embryonic occipital cortex were transplanted into newborn rats with large unilateral visual cortex lesions. When the animals were adults, they were tested on a difficult visual discrimination, and subsequently their brains were analyzed for possible neurotrophic effects of the transplants on nonvisual cortical areas which normally form connections with the occipital cortex. Behaviorally, animals with lesions and transplants learn to discriminate between columns and rows of squares at a rate which is identical to normal rats while animals with lesions and no transplants are impaired. Volume and cell-density measures show that the transplants also rescue neurons in cortical area 8 that would normally degenerate following the cortical lesion. No such neurotrophic effect of the transplants is found in cortical area 24 or area 17 contralateral to the lesion. In rats with lesions and no transplants, there is a significant correlation between the amount of area 8 remaining after the lesion and trials to criterion on the columns-rows discrimination, a relationship that does not exist in transplant animals because of their normal learning curve and the consistent sparing of area 8. Injections of HRP into the visual cortex contralateral to the lesion result in variable numbers of labeled cells within the transplant. However, there is no consistent relationship between the number of transplant cells which project to the opposite hemisphere and learning of the discrimination. It is suggested that the learning deficit following the lesion is largely attentional and that the sparing of cortical area 8 (which in rats may include the analog of the frontal eye fields present in the primate cortex) contributes to the sparing of function.