1. Two independent but neighbouring visual areas, V3 and V3A, sharing a common cytoarchitectural plan, but in each one of which the visual fields are separately represented, have been studied anatomically, functionally, and in combined anatomico-physiological experiments. 2. The properties of single cells in the two areas are so similar, judged by the techniques used in this study, that it is often impossible to tell whether any one penetration was sampling from cells in V3 or V3A. This is especially so if the cells have receptive fields in the lower hemi-quadrants, since the vertical meridian of the lower visual fields is represented along the V3-V3A boundary and since a transition from V3 to V3A along this border is not accompanied by a shift in receptive field positions of cells. 3. Since the visual fields, including the vertical meridian, are separately represented in these two areas, and since regions of vertical meridian representation are callosally connected, a simple and certain method of specifying the boundary between V3 and V3A is to examine the degeneration following section of the callosal splenium. A heavy patch of degeneration then marks the V3-V3A boundary. Within this patch, however, is a sub-patch containing fewer callosal fibres, or none at all. The boundary between V3 and V3A was taken to be at this subpatch. 4. Since the horizontal meridian is represented at the V2-V3 boundary, and since V1 projects to both these areas, sending coarse fibres to V3 and fine fibres to V2, it was found that the boundary between V2 and V3 could be precisely drawn by making a lesion in the horizontal meridian representation in V1 and noting where, in the prestriate cortex, fine fibres give way to coarse ones, without an intervening gap. 5. Double tracer anatomical experiments, in which tritiated proline was injected into V1 of animals whose callosal splenium had been sectioned, showed that whereas V3 receives a direct input from V1, V3A does not. V3A, instead, was found to receive an input from V3. Double tracer anatomical experiments were undertaken to study a possible input from V2 to V3A. Although such experiments did not reveal a direct input from V2 to V3A, they were not entirely conclusive. 6. The vast majority of cells in V3 and V3A were binocularly driven, without obvious monocular preferences. Some cells, however, though responding to stimulation of the individual eyes, summated their responses to binocular stimulation. Others responded only when both eyes were simulataneously stimulated. In any oblique penetration, cells preferring binocular stimulation only occurred either singly or in groups. 7. In an oblique penetration, the shift from a cell responding to binocular stimulation only to one responding equally well to stimulation of either eye was not necessarily accompanied by a shift in orientational preferences, shifts in the former...