1. The effect of auditory stimulation with click trains, noise bursts, amplitude-modulated noise bursts, and amplitude-modulated tone bursts on the correlation of firing of 1,290 neuron pairs recorded on one or two electrodes in primary auditory cortex of the cat was investigated. A distinction was made between neural synchrony (the correlation under stimulus conditions) and neural correlation (the correlation under spontaneous or under stimulus conditions after correction for stimulus-related correlations). For neural correlation 63% of the single-electrode pairs showed a unilateral excitation component, often combined with a common-input peak, and only 11% of the dual electrode pairs showed this unilateral excitation. 2. Under poststimulus conditions the incidence of correlograms with clear peaks was high for single-electrode pairs (80-90% range) and somewhat lower for dual-electrode pairs (50-60% range). The strength of the neural correlation for poststimulus conditions, from 0.5 to 2 s after a 1-s stimulus, was comparable with that obtained for 15-min continuous silence, suggesting that aftereffects of stimulation had largely disappeared after 0.5 s. A stationary analysis of the correlation coefficient corroborated this. 3. Two stimulus-correction procedures, one based on the shift predictor and the other based on the joint peristimulus-time histogram (JPSTH) were compared. The mean value of the neural correlation under stimulus conditions obtained after applying the poststimulus time (PST) predictor was on average 20% larger than the mean value obtained after application of the shift predictor; however, this was not significantly different at the 0.05 level. There were no differences in the shape of the correlograms. This suggests that the less time-consuming shift predictor-based stimulus-correction procedure can be used for cortical neurons. 4. Under stimulus conditions neural correlation coefficients could be < or = 50% smaller than for spontaneous conditions. The strength of the stimulus-corrected neural correlation was inversely related to the relative size of the stimulus predictor (compared with the neural synchrony) and thus to the effectiveness of stimulation. This suggests that the assumption of additivity of stimulus and connectivity effects on neural synchrony is generally violated both for shift predictor and PST predictor procedures. 5. The neural correlogram peaks were narrower for single-electrode pairs than for dual-electrode pairs both under stimulus and spontaneous conditions. Under stimulus conditions the peaks were generally narrower than under spontaneous firing conditions.(ABSTRACT TRUNCATED AT 400 WORDS)