Intracerebroventricular (i.c.v.) administration of corticotropin-releasing factor (CRF) biphasically affects performance in tests of learning and memory. In the present study, we used CRF, urocortin (Ucn), a recently cloned CRF homologue, and CRF receptor antagonists, to determine which CRF receptor subtype(s) mediate the memory modulating effects of CRF receptor agonists in male Wistar rats. Under difficult learning conditions (massed trials), i.c.v. pretreatment with CRF or Ucn facilitated the acquisition of spatial navigation in the Morris water maze in a non-dose-dependent fashion (optimal doses of 0.1 and 0.03 microg, respectively). Under less difficult learning conditions (spaced trials), both peptides impaired water maze performance. In addition, with i.c.v. posttraining treatment, the peptides were equipotent (1.0 microg) in facilitating the consolidation of passive avoidance learning. The performance-enhancing effects of Ucn in both water maze and passive avoidance paradigms were reversed by i.c.v. pretreatment with D-Phe CRF(12-41) (2.5, 5 microg), a broad CRF(1)/CRF(2) receptor antagonist, or antalarmin (10 microg), a potent, nonpeptide, CRF(1) selective receptor antagonist. Thus, Ucn shares CRF's memory-modulating effects, and these effects appear to be mediated via the CRF(1) receptor. These findings are consistent with the hypothesis that CRF receptor agonists affect performance in tests of learning and memory by increasing arousal.