This study provides baseline information on the potential use of propofol as a general anesthetic for horses. Using a Latin square design, propofol (2, 4, and 8 mg/kg) was administered intravenously on three separate occasions to six mature horses. Information about anesthetic induction, duration, and recovery was recorded along with results of rectal temperature, heart rate, respiratory rate, pHa, PaCO2 and PaO2. Statistical analysis included a mixed model analysis of variance, a general linear model analysis and least square means test for post hoc comparisons. A P < .05 was considered significant. The quality of induction of anesthesia varied from poor to good. Two horses were not recumbent following the lowest dose of propofol. Brief paddling limb movements occurred occasionally and unpredictably after recumbency induced by all three doses. During recovery, horses were uniformly calm and coordinated in their moves to stand. Duration of recumbency (minutes) was dose related; 15.05 +/- 1.58 (mean +/- SD) following 2 mg/kg, 31.06 +/- 5.56 following 4 mg/kg, and 47.85 +/- 13.63 following 8 mg/kg. During recumbency at all doses, heart rate significantly increased from a predrug value of 40 +/- 6 beats per minute. Substantial respiratory depression, characterized by a significant decrease in respiratory rate (from 11.7 +/- 2.9 to 3.7 +/- 1.6 breaths per minute) and increased PaCO2 (from 44.5 +/- 2.5 to 52.7 +/- 8.0 mm Hg) was seen only after 8 mg/kg. A significant decrease in PaO2 was observed throughout the recumbency induced by 8 mg/kg, and also at 3 and 5 minutes following induction of anesthesia with 4 mg/kg propofol.(ABSTRACT TRUNCATED AT 250 WORDS)