Body mass (BM), water intake (WI), and plasma osmolality (P(osm)) and electrolyte concentrations were measured in six 2-year-old Arabian horses provided either 4 l, 8 l, or an unlimited amount of water (UW) for drinking during the initial 5 min of recovery from 45-km of treadmill exercise. After weighing, horses were placed in a stall and further WI between 20 and 60 min of recovery was measured. During exercise, horses lost 3.3+/-0.3%, 3.2+/-0.1%, and 3.3+/-0.2% (P>.05) of BM and P(osm) increased by 7.2+/-0.5, 7.9+/-0.8, and 7.7+/-0.5 mOsm/kg (P>.05) for 4 l, 8 l, and UW, respectively. WI during the first 5 min of recovery was 4.0+/-0.0, 8.0+/-0.0, and 9.0+/-1.3 l and was accompanied by 2.4+/-0.4, 5.8+/-0.9, and 6.1+/-0.7 mOsm/kg decreases (P<.05) in P(osm) for 4 l, 8 l, and UW, respectively. Between 20 and 60 min of recovery, WI was 6.2+/-1.5, 1.2+/-0.6, and 1.0+/-0.7 l (P<.05) for 4 l, 8 l, and UW, respectively. Thus, total WI was 10.2+/-1.5, 9.2+/-0.6, and 10.0+/-1.1 l (P>.05) for 4 l, 8 l, and UW, respectively. After 60 min of recovery, persisting BM loss was 1.3+/-0.5%, 1.1+/-0.2%, and 1.0+/-0.2% (P>.05) for 4 l, 8 l, and UW, respectively and P(osm) had returned to pre-exercise values for all treatments. In conclusion, limiting the volume of water initially provided to horses dehydrated by endurance exercise had no significant effect on total WI during the initial 60 min of recovery; however, persisting BM loss was observed with all treatments. Further, following exercise-induced dehydration, the primary stimulus of thirst was an increase in plasma tonicity rather than hypovolemia.