Eight men were studied to determine the effect of cycling exercise on plasma volume (PV) during water immersion to the xiphoid process (WIX). In all protocols the subjects were seated upright. After 30 min of rest, subjects were immersed in 34.5 degrees C water and seated on a cycling ergometer. During three 1 h WIX protocols, subjects either remained at rest (No Ex) or pedaled from minutes 20 to 30 at 38% (Ex1) or 62% (Ex2) of peak oxygen consumption (VO2peak). Hematocrit (Hct) and hemoglobin concentration [( Hb]) from venous blood samples were compared pre-WIX and at minutes 20, 30, 40, and 60. Percent change in PV (delta PV) was calculated from pre-WIX Hct and [Hb] within each protocol. Hct and [Hb] decreased after 20 min of resting WIX (P less than 0.017). In the No Ex protocol, there were no further significant changes in these variables, with delta PV values of +10.4% at minute 20 and at a peak of +13.5% at minute 40. In Ex1 and Ex2, cycling increased Hct and [Hb] (P less than 0.01, minute 30 vs No Ex), with delta PV values at minute 30 of +3.7% and -0.9%, respectively, vs +12.8% in No Ex. Minute 60 values between protocols were not significantly different (mean delta PV of +10.8 +/- 0.6% SD). The hemodilution associated with WIX was either partially or completely attenuated by cycling exercise; the degree of hemoconcentration was related to exercise intensity. The exercise-induced hemoconcentration was reversed by 30 min of resting WIX. Exercise during WIX appears to cause similar decreases in PV, as does exercise in air provided that postural hemoconcentration prior to exercise is not already maximal.