The role of corticosterone in cocaine reinforcement was investigated in rats exposed to either response-contingent electric footshock, noncontingent shock or no shock prior to the initiation of testing for intravenous cocaine self-administration. Although rats from the two shock groups were consistently more sensitive to cocaine, plasma corticosterone was always significantly higher in all rats that subsequently self-administered low doses of cocaine compared to the rats that did not, regardless of the treatment condition. In fact, plasma corticosterone was always greater than approximately 150 ng/ml for rats that self-administered low doses of cocaine, suggesting that this stress hormone must be increased above a critical threshold for stable low-dose cocaine self-administration to subsequently occur. Plasma corticosterone was also measured following exposure to cocaine and was significantly elevated in rats from all three treatment groups during cocaine self-administration, provided that doses of cocaine that would maintain responding were tested. When the dose would not maintain self-administration, plasma corticosterone was markedly lower, suggesting that at higher concentrations, the cocaine injections alone were sufficient to increase plasma corticosterone above a critical threshold, even for rats which had low precocaine levels of the hormone. These data suggest a significant role for corticosterone in both the acquisition and maintenance of cocaine reinforcement in rats.