The authors used the absence of vocalization and movement during application of a hemostat clamped to the first ratchet at the base of the tail as an indication of anesthesia for evaluation of the anesthetic properties of fentanyl in the dog. Eighty-six unpremedicated, unrestrained, untrained mongrel dogs were given one of eight doses of fentanyl citrate (125, 250, 500, 750, 1000, 1500, 2000, and 3000 micrograms/kg) as a single intravenous bolus injection. Dogs breathed spontaneously without oxygen supplementation. Anesthesia was assessed every 5 min until absence of anesthesia was recorded for two consecutive evaluations. Venous plasma samples were obtained in two or three dogs receiving each of the doses of fentanyl 5 min after fentanyl injection and again when application of the tail clamp elicited either vocalization or movement (positive response). Fentanyl resulted in recumbency in all animals except two receiving 125 micrograms/kg. Although all doses of fentanyl produced anesthesia in at least one animal 5 min after injection, the duration of anesthesia was short, responses unpredictable, and anesthesia achieved in all animals only with a dose of 3000 micrograms/kg. Increasing doses of fentanyl resulted in higher plasma fentanyl concentrations 5 min after injection and at the time of the first positive response to tail clamp but there was great variability. All doses of fentanyl caused statistically significant decreases in heart and respiratory rates but none produced apnea or a PaCO2 higher than 67 mm Hg.(ABSTRACT TRUNCATED AT 250 WORDS)