We hypothesized that central fatigue may have a role in limiting the endurance capacity of horses. Therefore, we tested the effect of infusing tryptophan and/or glucose on endurance time and plasma concentrations of free tryptophan and other substrates thought to affect tryptophan uptake into the brain of seven mares (3-4 yr of age, 353-435 kg) that ran on a treadmill at 50% of maximal O2 consumption to fatigue. With use of a counterbalanced crossover design, the horses were infused with tryptophan (100 mg/kg in saline solution) or a similar volume of saline solution (placebo) before exercise. During exercise, horses received infusions of glucose (2 g/min, 50% wt/vol) or a similar volume of saline. Thus the treatments were 1) tryptophan and glucose (T & G), 2) tryptophan and placebo (T & P), 3) placebo and glucose (P & G), and 4) placebo and placebo (P & P). Mean heart rate, hematocrit, and concentration of plasma total solids before and during exercise were similar for all trials. Mean time to exhaustion was reduced (P < 0.05) for T & P and T & G compared with P & P [86.1 +/- 6.9 and 87.1 +/- 6.8 vs. 102.3 +/- 10.3 (SE) min], whereas endurance for P & G (122.4 +/- 11.9 min) was greater than for all other trials (P < 0.05). Compared with nontryptophan trials, during the tryptophan trials plasma prolactin increased (P < 0.05) nearly threefold before exercise and almost twofold early in exercise. Muscle glycogen concentrations were reduced (P < 0.05) below preexercise values in the P & G and P & P trials only. However, glucose infusions (P & G) did not affect (P > 0.05) concentrations of plasma free fatty acids or ratios of branched-chain amino acids to free tryptophan. In conclusion, tryptophan infusion reduced endurance time, which was consistent with the central fatigue hypothesis. The failure of glucose infusion to alleviate the effects of tryptophan and the absence of significant muscle glycogen reduction in the tryptophan trials suggest that the early onset of fatigue in the tryptophan trials is not due to a lack of readily available substrate.