A quantitative ethological analysis of rodent aggression was performed in order to characterize the aggression-heightening effects of alcohol in certain individuals. In dyadic confrontations, a resident rat pursues, threatens and attacks an intruder, who reacts with defensive, flight and submissive behaviors. The behavioral data from five series of experiments conducted from 1984 through 1989 were subjected to a lag sequential analysis that identified highly predictable sequences of aggressive behavior, and to interval analysis that delineated a burst pattern of aggressive behavior. These analyses revealed a distinct behavioral sequence of pursuit----sideways threat----attack bite----aggressive posture that occurs in bursts with an inter-event interval of less than 6.6 s. In the total population, alcohol heightened attack behavior at low acute doses (0.1, 0.3, 1.0 g/kg) in 47% of the animals (n = 44), suppressed reliably attack behavior in another 25% (0.1-3.0 g/kg; n = 23) and had unreliable effects in the remaining 28% (n = 24). The peak enhancement of aggressive behavior was seen over more than a log cycle of alcohol doses (0.1, 0.3 or 1.0 g/kg) in different individuals. In an additional group of rats (n = 20), individuals were identified according to whether or not acute low alcohol doses enhanced or suppressed the frequency of attack bites. In the subgroup of five rats who doubled their attack frequency upon acute alcohol challenge, this aggression-heightening effect was confirmed on repeated occasions. The aggression-heightening effects of alcohol were seen during the high-rate interactions in the initial phase of the confrontation and particularly during the lower level of fighting later on. Regardless of alcohol dose and subgroup, the highly predictable sequence of pursuit----sideways threat----attack bite----aggressive posture remained intact as long as the individual was able to fight. The present analysis identifies those individuals in whom low alcohol doses increase the frequency of attack behavior, the number of aggressive elements in bursts and particularly the "time in burst". Alcohol produces these changes without altering the latency to initiate aggressive behavior, the rate of aggressive behavior within a burst or the number of bursts in an encounter. Alcohol may lengthen aggressive bursts by preventing termination of longer aggressive sequences rather than by altering the initiation of this behavior.