To examine the role of vasopressin V(1) and V(2) receptors, nitric oxide and prostanoids in the cerebrovascular effects of arginine vasopressin, cerebral blood flow was electromagnetically measured in awake goats. In 16 animals, vasopressin (0.03 - 1 microg), injected into the cerebral circulation, caused increments of resting cerebrovascular resistance which ranged from 18% (0.03 microg, P<0.01) to 79% (1 microg, P<0.01). Desmopressin (0.03 - 1 microg, four goats) did not affect significantly cerebrovascular resistance. The cerebrovascular resistance increases by vasopressin were reduced significantly by the antagonist for vasopressin V(1) receptors d(CH(2))(5)Tyr(Me)-AVP in a rate depending way (five (six goats) and 15 (four goats) microg min(-1)), and by the mixed antagonist for vasopressin V(1) and V(2) receptors desGly-d(CH(2))(5)-D-Tyr(Et)Val-AVP (5 microg min(-1), four goats), and they were not significantly affected by the antagonist for vasopressin V(2) receptors d(CH(2))(5), D-Ile(2), Ile(4)-AVP (5 microg min(-1), four goats). The inhibitor of nitric oxide synthesis N(w)-nitro-L-arginine methyl ester (L-NAME, 47 mg kg(-1) i.v., five goats) augmented cerebrovascular resistance by 130% (P<0.01), and for 24 h after this treatment the cerebrovascular effects of vasopressin were potentiated. The inhibitor of cyclo-oxygenase meclofenamate (6 mg kg(-1) i.v., five goats) did not modify significantly resting haemodynamic variables measured or the cerebrovascular effects of vasopressin. Therefore, the vasopressin-induced cerebral vasoconstriction may be mediated by vasopressin V(1) receptors, without involvement of vasopressin V(2) receptors, and may be modulated by nitric oxide but not by prostanoids.