Ten experienced male marathon runners ran 20 miles (32.18 km) on an outdoor course in a warm climate to measure responses in selected physiological variables as a result of drinking water, an electrolyte-glucose solution (ERG), or a caffeine solution (5 mg X kg-1 body weight) before and during the run. The caffeine solution and water were colored and flavored to resemble the electrolyte-glucose solution so that a double-blind condition could be maintained. Subjects ingested a different fluid in each of the three trials in a counterbalanced design. Parameters studied were: heart rate; rectal temperature; body weight; hemoglobin and hematocrit; serum glucose, sodium, potassium, chloride, and free fatty acids; perceived exertion; respiratory exchange ratio (R); and fractional utilization of VO2max (percentage of VO2max). Both R and percentage of VO2max were higher in subjects who drank the caffeine solution compared to those who drank water. Although post-run free fatty acid mean values were significantly higher than pre-run levels, there were no statistically significant differences among the fluid treatments. Since no other differences were observed, we concluded that, under the outdoor road-running conditions encountered here, these fluid replacement treatments did not differ in their effects on the parameters studied.