Rats were tested two or three days after bilateral abdominal vagotomy or a laparotomy control procedure for their drinking responses to subcutaneous (1 mg-kg-1) or intracerebroventricular (100 ng) injections of angiotensin II. Vagotomy delayed the initiation of drinking and decreased 60-min water intake after subcutaneous, but not after intracerebroventricular, angiotensin II. This is the shortest postoperative interval in which the decrease in drinking after systemic injection of angiotensin II by abdominal vagotomy has been observed. The failure of vagotomy to decrease the response to intracerebroventricular angiotensin II demonstrates that the deficit after subcutaneous injection was not a nonspecific effect of recent vagotomy. These results, therefore, suggest that the abdominal vagus is necessary for normal drinking in response to circulating angiotensin II. Furthermore, the selective and acute onset of the deficit is consistent with the loss of a specific, rather than tonic facilitatory, vagal mechanism for drinking after elevation of circulating angiotensin II levels. Finally, the results imply that the physiological mechanisms which mediate the drinking responses to central and peripheral angiotensin are not identical.