1. gamma-Aminobutyric acid-A (GABAA) receptor-mediated synaptic currents evoked by intracortical stimulation in rat somatosensory cortical slices maintained in vitro were studied using the whole cell patch-clamp technique. All anatomically identified pyramidal neurons of layer II-III (SG neurons), layer IV (IV neurons), and layer V (IG neurons) generated evoked inhibitory postsynaptic currents (eIPSCs) that were blocked by bicuculline. At threshold, eIPSCs had kinetic properties (rise time of 0.9 ms and decay time constant of 9 ms) similar to those of spontaneous IPSCs generated in the same cells. 2. The strength of inhibition was quantified by determining the stimulus threshold for evoking responses and the relationship between stimulus strength and eIPSC peak amplitudes (input/output curve). For eIPSCs recorded in control solution, the input/output curve was about four times steeper than for eIPSCs recorded in the presence of the ionotropic glutamate receptor antagonists 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and D-2-amino-5-phosphonovalerate (D-AP5), suggesting the dependence of GABAA inhibition on synpatic excitation of interneurons. 3. In the presence of CNQX and D-AP5, monosynaptic IPSCs, evoked by stimulation close to the recording patch pipette, had similar input/output curves in SG and IG neurons. This suggests that the level of monosynaptic inhibition generated in these two populations of cells is similar. 4. When the stimulus was moved to a distant site > 350 microns from the recorded neuron, either in vertical or in horizontal direction, the stimulus intensity required for evoking IPSCs was higher, and the input/output curve was less steep. This suggests that the density of GABAergic somata and axons projecting to the recorded neuron is lower at these distances than at more proximal sites. 5. The maximum horizontal distance over which IPSCs could be evoked ("horizontal field") was larger in layer V than in other layers. The horizontal field (distance between stimulating and recording pipettes) was 600 microns in layer II-III, 580 microns in layer IV, and 720 microns in layer V. Anatomic identification of the somatosensory cortical barrels indicated that the extent of GABAergic projections was larger than the barrel hollow and might thus form a substrate for interbarrel inhibition in layer IV during cross-wisker stimulation. 6. The maximum vertical inhibitory field was larger than the maximum horizontal field. IPSCs could be evoked in layer V neurons by layer I stimuli, showing that a powerful interlaminar inhibition is present that may play a role in synchronizing the activity of neurons in a column. IPSCs evoked by layer I stimulation frequently had slower kinetics than those elicited by stimulation at sites close to the soma. 7. These findings suggest that functional GABAergic projections are characterized by a large degree of convergence. Quantification of GABAA-mediated IPSCs indicates that this zone of inhibitory synaptic convergence onto a given pyramidal neuron is subdivided into a powerful local inhibitory zone and a surrounding area of long-range, less effective, inhibitory projections. Potential roles for these concentric inhibitory areas in cortical processing of sensory information are discussed.