Placebo and triazolam (0.125, 0.25 and 0.5 mg/70 kg of body weight) were administered to male subjects under double-blind conditions prior to experimental sessions which provided three operant response options. These options were: 1) responding maintained by the presentation of points exchangeable for money, 2) responding which ostensibly resulted in the subtraction of points from a fictitious person was termed aggressive since this responding resulted in the delivery of an aversive stimulus to another person, and 3) responding which ostensibly protected the subject's point counter from subtractions initiated by the other person and was termed escape. Aggressive and escape responding were initiated by subtracting points from the subject. Point subtractions were attributed to the other person. Aggressive and escape responding were maintained by initiation of provocation-free intervals (PFI), during which no further point subtractions were presented. Triazolam produced dose-dependent decreases in point-maintained and escape responding. The effects of triazolam on aggressive responding varied across subjects.