We found a difference between the venous hematocrits of immersed and nonimmersed arms during immersion of the lower body in cold water but not during a comparable exposure to warm water. Fourteen healthy men were exposed to three different experimental conditions: arm immersion, body immersion, and control. The men always sat upright while both upper extremities hung vertically at their sides. During arm immersion, one forearm was completely immersed for 30 min in either cold water (28 degrees C, n = 7) or warm water (38 degrees C, n = 7). This cold-warm water protocol was repeated on separate days for exposure to the remaining conditions of body immersion (immersion of 1 forearm and all tissues below the xiphoid process) and control (no immersion). Blood samples were simultaneously drawn from cannulated veins in both antecubital fossae. Hematocrit difference (Hct diff) was measured by subtracting the nonimmersed forearm's hematocrit (Hct dry) from the immersed forearm's hematocrit (Hct wet). Hct diff was approximately zero when the men were exposed to the control condition and body immersion in warm water. In the remaining conditions, Hct wet dropped below Hct dry (P less than 0.01, 3-way analysis of variance). The decrements of Hct diff showed there were differences between venous hematocrits in immersed and nonimmersed regions of the body, indicating that changes of the whole-body hematocrit cannot be calculated from a large-vessel hematocrit soon after immersing the lower body in cold water.