We previously reported that adrenalectomized rats given constant corticosterone via a sc pellet (B-PELLET) hypersecrete ACTH in response to stress. Although lacking a feedback signal, B-PELLET rats do not secrete ACTH indefinitely after stress; plasma ACTH levels in these animals returned to those in sham-operated (SHAM) rats within 1-4 h after 2-min restraint. To distinguish between the requirement for circadian or stress-induced increases in corticosterone, we compared changes in ACTH and corticosterone levels after stress in SHAM and B-PELLET rats with those in cyanoketone-treated rats (CK) and adrenalectomized rats given corticosterone in their drinking fluid (B-WATER). B-WATER rats exhibited sustained increases in plasma corticosterone after lights-off, correlating with the nocturnal feeding period. Morning plasma corticosterone levels in B-WATER rats were constant and even lower than those in B-PELLET rats; however, B-WATER rats did not differ from SHAM rats in their ACTH response to ip injection. CK rats, which have an approximately normal circadian corticosterone rhythm but do not have significant corticosterone responses to acute stimuli, also exhibited plasma ACTH levels similar to those of SHAM rats at all times after 5-min restraint. Compared with SHAM and B-WATER rats in the same experiment, B-PELLET rats tended to hypersecrete ACTH 60 min after 5 min of restraint, but only had significantly elevated plasma ACTH relative to both groups 45 min after 10 min of restraint. We conclude that circadian, rather than stress-induced, increases in corticosterone may be sufficient for normal termination of ACTH responses to stress.