Laminar distributions in cat striate cortex were prepared for neurons classed according to their functional type, ordinal position, or the type of afferent input they receive from the thalamus. A frequency histogram plotted as a function of depth in the cortex for the total population of 231 neurons showed that cells of lamina 5 and at the border of laminae 3 and 4 are sampled more often by the recording electrodes, possibly because they are large pyramidal neurons. The laminar distribution of different cell types across cortex was prepared for 218 neurons classified as S, S(h), B, C, or cells with nonoriented or concentric receptive field (N-O/conc), following the terminology of Henry. S-cells were found to congregate in laminae 4 and 6, C-cells in laminae 5 and upper 6, and B-cells in laminae 3 and 5. Cells with nonoriented or concentric receptive fields were most numerous in the lower part of lamina 4 and S(h)-cells, in the lower part of lamina 6. Cells that could not be driven by visual stimulation were concentrated in laminae 2 and 3. The laminar distribution of neurons of different ordinal position was examined for a sample of 223 units. Cells receiving a monosynaptic input from the thalamus were found in laminae 4, 5, and 6 but were absent from laminae 2 and 3. Cells receiving an indirect input from the thalamus were concentrated in laminae 2 and 3 and 5 and 6 and were rare in lamina 4. Cells that were not activated by electrical stimulation of the optic radiations were most numerous in laminae 2 and 3. It is concluded that lamina 4 acts as a primary receiving layer, with possibly all the cells having a direct input from the dLGN. There are indications that some serial processing takes place in laminae 5 and 6, whereas laminae 2 and 3 appear to be the main locations for higher order neural processing. Cells receiving a monsynaptic input from the fast and slow streams were found to congregate at different levels in cortex. Neurons receiving a monosynaptic input from the fast stream (group I cells) were concentrated in the upper half of lamina 4 and in laminae 5 and upper 6. In contrast, neurons receiving a monosynaptic input from the slow stream (group II cells) were most numerous in the lower three-fourths of limina 4 and in the upper part of lamina 6. The fast stream appears to synapse on S-cells in lamina 4 and on C-cells in laminae 5 and 6. The slow stream provides the input to S, S(h), and N-O/conc cells in lamina 4 and to S-cells in lamina 6. The terminal arborizations of axons infiltrated with horseradish peroxidase (HRP) after injection into the optic radiations were traced for 25 axons in lamina 4. Large axons (diameter > 1.7 μm) were found to branch in the upper half of lamina 4, whereas medium-size axons (0.9 < diameter < 1.4 μm) arborized in the lower three-fourths of this lamina. The cortical distribution of the terminal arborization of medium-size axons fits well with that of group II cells (receiving from the slow stream). In contrast, terminal branching of larger axons extended deeper in lamina 4 than expected from the distribution of group I cells (receiving from the fast stream). The significance of these results and those of the two preceding papers are discussed in relation to the processing of information within cat striate cortex. It is suggested that the two afferent streams (fast and slow) remain separate within the cortex. These streams undergo further branching to form cortical tributaries and consideration is given to the possible course taken by these tributaries through the cortex.
