1. Neural activity was recorded extracellularly with two independent microelectrodes aligned in parallel and advanced perpendicular to isofrequency sheets in cat primary auditory cortex. Multiunit activity was separated into single-unit spike trains using a maximum variance spike sorting algorithm. Only units that demonstrated a high quality of sorting and a minimum spontaneous firing rate of 0.2 spikes/s were considered for analysis. The primary aim of this study was to describe the effect of periodic click train and broadband noise stimulation on short-time-scale (< or = 50 ms) bursts in the spike trains of single auditory cortical units and to determine whether stimulation influenced the occurrence, spike count, and/or temporal structure of burst firing relative to a spontaneous baseline. 2. Extracellular recordings were made in 20 juvenile and adult cats from 69 single auditory cortical units during click train stimulation and silence, and from 30 single units during noise stimulation and in silence. In an additional 15 single units the effect of both click train and noise stimulation was investigated. The incidence, spike count, and temporal structure of short-time-scale burst firing in the first 100 ms following stimulus presentation was compared with burst firing in the period starting 500 ms after stimulus presentation and with spontaneous burst firing. In addition, the serial dependence of interspike intervals within a burst was tested during periods of stimulation. 3. Burst firing was present in the stimulation, poststimulation, and spontaneous conditions. Longer bursts (consisting of > or = 3 spikes) were more commonly observed in the poststimulation and spontaneous conditions than in the stimulation condition. This effect was most pronounced during click stimulation. A period of elevated firing activity was present in a subset of units 0.5-1.5 s after stimulus presentation, indicating prolonged effects of stimulation on single-unit firing behavior. 4. For both stimuli, the proportion of single-unit responses composed of bursts was significantly greater in poststimulation and spontaneous periods than during stimulation. Burst rate was higher in post-click-train stimulation and spontaneous periods than during periods of click stimulation. The isolated spike rate was significantly higher during periods of noise and click stimulation than in the poststimulation and spontaneous periods. 5. An examination of the autocorrelograms and higher-order interspike interval histograms of single-unit responses during click train stimulation indicated that 25% of single-unit spike trains contained an excess of brief first-order intervals and 14% of spike trains contained a shortage of long higher-order interspike intervals relative to a spontaneous baseline. During noise stimulation, 10% of single-unit responses contained an excess of short intervals relative to baseline. Interspike intervals of short-duration bursts were not serially dependent during periods of stimulation. 6. A comparison of the autocorrelograms and higher-order interval histograms of single-unit responses in the poststimulation and spontaneous conditions indicated that 20% of single-unit spike trains contained an excess of short first-, second-, and third-order intervals following stimulation. This subgroups of single units could not be distinguished on the basis of the age of the animal or the depth at which the recording was made. 7. The low incidence of burst firing during stimulation opposes the view that bursts serve as a mechanism to emphasize or amplify particular stimulus-related responses in the presence of ongoing spontaneous activity in the primary auditory cortex. Moreover, there is little evidence to support the notion that brief bursts represent neural codes, because intraburst intervals are not serially dependent. It is suggested that pyramidal burst firing may be an effective way to evoke postsynaptic firing in inhibitory interneurons and subsequ