Previous studies have shown that a variety of local anesthetics including procaine are self-administered at high rates by rhesus monkeys. In the present study two rhesus monkeys were given a mutually exclusive choice between various doses of intravenous cocaine and procaine. In almost all comparisons cocaine was preferred even when the procaine dose was 16 times that of cocaine. Other measures of performance such as rate of responding did not vary systematically with preference. These data provide further support for the idea that rate of responding under simple schedules of drug delivery is an unreliable measure of relative reinforcing efficacy. In addition, the consistent preference for cocaine over procaine in monkeys suggests that the infrequent abuse of procaine by humans may be related to its low reinforcing efficacy relative to drugs such as cocaine.