Sixty-four rats were trained to jump onto a safety ledge to avoid a -5 mA shock, to a criterion of 10 successive avoidance responses, and with a 30-sec inter-trial interval. They were then immediately assigned (randomly) to one of four treatments during which shocks never occurred: massed-trials extinction, in which the inter-trial interval was reduced to 3 sec; response-prevention one, in which the safety ledge was withdrawn; and response prevention two, in which the safety ledge was made inaccessible by a transparent slide. The fourth group served as a control for confinement in the responseprevention groups. The duration of the treatment was 9 minutes for all groups. Central to this study was that the duration of exposure to the fear conditioned stimulus (CS) was held constant across groups. Following treatment, all subjects were immediately retrained to the active avoidance criterion and then given normal avoidance extinction. Multivariate analysis showed that massed-trials extinction produced a greater attenuation of avoidance responding than did response-prevention, and that the responseprevention groups did not differ from each other. This confirmed the finding of Baum & Oler (1968) that massed-trials extinction is superior to response-prevention even when CS exposure duration was held constant. It is congruent with a modified two-process theory of avoidance extinction, in that massed-trials extinction provided more opportunity for the discriminative function of the CS to be modified.