Single-subject methods were used to study response speed in three younger men (M = 22 years) and three older men (M = 71 years). Monetary reinforcement depended on predetermined sequences of 10 responses. Training sessions consisted of alternating components; in one the sequence changed each session, but in the other the sequence never changed. During self-paced sessions stable patterns of acquisition (increasing rates of sequence completion, decreasing error rates) and of performance (uniform rates) emerged in the two components. During paced sessions responses had to occur within an interval ranging from .5 to 3 sec across conditions. Pacing had little effect on the younger men but disrupted the behavior of the older men, with increases in both omission and commission errors. Acquisition of new sequences was disrupted far more than performance of the established sequence. In both cases, however, the older men adjusted to pacing over the course of most sessions.