In an attempt to define more precisely the various mechanisms involved in antidiuretic hormone (ADH) release during positive end-expiratory pressure ventilation (PEEP), experiments were performed on seven groups of dogs. PEEP-10 and PEEP-15 cmH2O caused significant elevations of plasma ADH from basal values of 24.9 +/- 5.2 pg/ml (mean +/- SE) to 64.6 +/- 14.2 and 106.0 +/- 20.6, respectively (P < 0.02, P < 0.005). The ADH levels returned to basal values after cessation of PEEP. This rise in ADH levels was prevented by an infusion of dextran prior to PEEP. The fall in blood pressure and cardiac output that occurred during PEEP was also prevented by the dextran infusion. Changes in ADH levels were unrelated to lung volume, left transmural pressure, and serum osmolality. Bilateral vagotomy and carotid sinus denervation was followed by an attenuated rise in ADH levels in terms of the percent rise above base line, but it did not significantly alter the absolute rise in ADH during PEEP. ADH levels were, however, reduced significantly by decreasing intracranial pressure by the removal of cerebrospinal fluid during PEEP. Propranolol administration prior to PEEP completely blocked plasma renin activity. Although the peak ADH levels were unaffected by propranolol, the rise was delayed. The results obtained indicate that a number of physiological factors may affect plasma ADH levels during PEEP. These include the carotid body and aortic arch baroreceptors as wells as sensors of intracranial pressure.