The precision of spike response timing of 94 primary auditory cortex neurons was studied using conventional extracellular recording techniques in barbiturate-anesthetized cats to which tone- and/or noise-burst stimuli were presented using sealed sound delivery systems. Precision of spike timing was indexed using the standard deviation of the first-spike latent period in responses evoked by repeated presentation of tonal stimuli systematically varied in frequency, amplitude, and/or repetition rate. Within a neuron, variability of first-spike timing was usually proportional to the mean first-spike latency, in agreement with previous reports. In cases where there was a systematic relation between the precision of response timing and the mean latency, a linear correlation accounted for up to 90% of the data variance. Across the 94 neurons, standard deviations seen in responses of minimum latency were related to minimal mean latencies, and were typically in the range from 0.15-1.5 ms. The data suggest that responses to transients in the cortex show a precision of spike timing which is only slightly worse than that seen in cochlear-nerve fibers. This, however, is in dramatic contrast to previous evidence on the steady-state temporal response of cortical cells, which is at least an order of magnitude poorer than that seen in auditory-nerve fibers and many cochlear nucleus cells. These observations may be directly relevant to the known consequences of auditory cortex pathology in man.