The size and response magnitude of receptive fields were evaluated for cells in the rat cortical barrel-field by using standard vibrissal deflections of 1.14 degrees. Such stimuli fell within the plateau region of stimulus-response curves. The response of all neurones to all vibrissae within and surrounding centre-receptive fields were analysed for probability and latency of response. It was found that cells in supragranular layers had small centre-receptive fields (average 1.6 vibrissae) with small excitatory surrounds (1.5 vibrissae) while cells in the granular layers had small, powerful centre-receptive fields (1.4) with moderately large excitatory surrounds (2.6). Neurones in infragranular layers possessed large but weak centre-receptive fields (2.6) with large surrounds (3.5). Sixty-four neurones in layer IV were studied, the precise locations of which were identified by using dye lesioning and cytochrome oxidase staining. There were no differences in receptive field size for cells within septa and barrel hollows, but the latter were twice as likely to produce two or more spikes per stimulus from the principal vibrissa (65% against 33%). Histological analysis showed that the principal vibrissa was synonymous with the appropriate vibrissa for the barrel on 86% (55 of 64) of occasions. A quantitative analysis of convergent input to three neighbouring barrels (E1, E2, and D1) showed considerable graded overlap of receptive field surrounds, although facial hair adjacent to the mystacial pad only influenced cells on the edge of the barrel-field. Individual vibrissae exhibited significant divergent input to adjacent inappropriate barrels, being preferentially directed to distant septal rather than barrel hollow cells. An analysis of latencies showed that 40% of barrel hollow cells and 48% of barrel septal cells responded at short (less than 10 ms) latencies to their appropriate vibrissa. In contrast, responses to inappropriate vibrissae were overwhelmingly of long latency (10-greater than 30 ms), only 2% of inappropriate responses from barrel hollow cells and 13% from septal cells being of short latency. These results suggest that direct inputs largely project to appropriate barrels. The possibility that divergent inputs are generated by intracortical mechanisms is discussed.