This study assessed the need to replace sodium in endurance exercise less than or equal to 6 h in duration by comparing responses to fluid replacement with water, saline (25 mmol.l-1), or no fluid. Eight subjects (five male, three female) participated in three 6-h exercise trials on an electrically braked cycle ergometer at 55% VO2max, at 30 degrees C and 50% r.h. In the water (W) and saline (S) trials, sufficient fluid was ingested to balance sweat and urinary fluid losses, while in the third trial, no fluid (NF) was ingested. Plasma sodium less than or equal to 130 mmol.l-1 was a criterion for trial termination. In the NF trial, heart rate, rectal temperature, plasma sodium, plasma aldosterone, and rating of perceived exertion were all significantly higher (P less than 0.001) than during W and S, whereas plasma volume was lower (P less than 0.001). On average, subjects terminated this trial 1.5 h prior to its scheduled completion, having lost 6.4% body weight. In contrast, no significant differences between fluid replacement with W or S were detected, although the effect of time on all aforementioned variables was highly significant (P less than 0.001). Saline intake was not associated with significantly higher plasma sodium during exercise than was water intake: plasma sodium decreased significantly during both W (to 135.5 +/- 0.5 mmol.l-1) and S (to 137.3 +/- 0.7 mmol.l-1). No subject had to terminate a trial based on plasma sodium less than or equal to 130 mmol.l-1.(ABSTRACT TRUNCATED AT 250 WORDS)