The Valsalva maneuver (VM) is frequently used to test autonomic function. However, the VM is also affected by changes in blood volume and by blood volume redistribution. We hypothesized that even a standardized VM may produce a wide range of thoracic blood volume shifts. Larger blood volume shifts in some normovolemic individuals may be sufficient to induce decreases in blood pressure which preclude autonomic restoration of BP in phase II of VM. To test this hypothesis we studied 17 healthy volunteers aged 15-22 years. All had similar supine and upright vasoconstrictor responses and normal blood volume. We assessed changes in thoracic blood volume by impedance plethysmography before and during the VM performed supine. In some subjects, large decreases in BP were produced by thoracic hypovolemia. The maximum fractional decrease in BP correlated well (r2=0.64, p<.001) with thoracic hypovolemia and with systolic blood pressure at the end of phase II of the Valsalva maneuver (r2=0.67, p<.001). The blood pressure overshoot in phase IV of the maneuver was uncorrelated to phase II changes suggesting intact autonomic vasoconstriction. We conclude that the blood pressure decrease during the Valsalva maneuver is related to a variable decrease in thoracic blood volume which may be sufficient to preclude pressure recovery during phase II even with normal resting peripheral vasoconstriction. The Valsalva maneuver depends on vascular as well as autonomic activation, which broadens its utility but complicates its analysis.