This study examined systemic testosterone concentrations in rams that were classified according to their sexual behavior and partner preference as either female-oriented (FOR), male-oriented (MOR), or asexual (NOR). For this purpose, we measured testosterone concentrations under three separate conditions: in conscious rams during the nonbreeding season (June) and breeding season (November), and in anesthetized rams during the breeding season. Basal testosterone concentrations in conscious rams were not different among the three groups (P > 0.05) in either season. However, when rams were anesthetized, mean systemic concentrations of testosterone in FORs (mean +/- SEM, 13.9 +/- 7.4 ng/ml serum) were greater (P < 0.05) than in NORs (0.9 +/- 0.1 ng/ml), but not in MORs (2.2 +/- 6.2 ng/ml), whereas testosterone concentrations were not different between MORs and NORs (P > 0.05). Concentrations of testosterone in the spermatic vein of FORs (127 +/- 66 ng/ml) were greater (P < 0.05) than in MORs (41 +/- 10 ng/ml) and NORs (19 +/- 7 ng/ml). Serum LH concentrations were not different. Cortisol was higher (P < 0.05) in anesthetized MORs (25.1 +/- 4.2 ng/ml) and NORs (27.2 +/- 4.4 ng/ml) than in FORs (10.9 +/- 1.8 ng/ml). These results demonstrate that circulating testosterone concentrations are related to sexual behavior only when rams are bled under anesthesia. Thus, differences in basal androgen concentrations in adulthood cannot be responsible for expression of male-oriented preferences or low libido in sheep. Instead, functional differences must exist between the brains of rams that differ in sexual preference expression.