The confluent sinus pressure was measured in eight mongrel dogs in the head-up position to compare the effectiveness of positive end-expiratory pressure (PEEP) and jugular venous compression in increasing cerebral venous pressure. When the head was elevated 30 cm above the heart, confluent sinus pressure decreased from 9.6 +/- 1.8 (mean +/- SEM) to -5.3 +/- 0.5 mmHg. At constant arterial carbon dioxide tension (PaCO2 = 28 +/- 2 mmHg), PEEP (20 cmH2O) did not increase cerebral venous pressure. However, when the jugular veins were compressed with a neck tourniquet with pressures of 20-140 mmHg, cerebral venous pressure increased rapidly. When neck tourniquet pressure was maintained at 40 mmHg, confluent sinus pressure in all dogs was increased and sustained at 2.4 +/- 0.8 mmHg. Carotid artery pressure measured distal to the tourniquet was not altered. The efficacy of extrathoracic venous pressure elevation (neck tourniquet) is greater than intrathoracic (PEEP), and this may relate to the Starling resistor effects of neck veins and the presence of jugular venous valves. We conclude that prophylactic use of PEEP in the prevention of air embolism during the sitting position may not be as effective as jugular venous compression.