Five horses with Australian stringhalt were treated with 15 mg/kg phenytoin orally for 2 weeks. During the second week of the trial, 3 of the horses were given an additional dose of 10 mg/kg phenytoin. The response to treatment was clinically assessed by grading the severity of the gait abnormality at the walk, trot, turning and backing twice daily. There was a significant (P less than 0.05) improvement in the gait abnormality when pre-treatment values were compared with the mean of the last 3 assessments before treatment stopped. When reassessed 2 weeks after treatment ceased, there remained a significant (P less than 0.05) improvement compared with pre-treatment values at the trot and on backing, but not at the walk or turning. Surface electromyographic recordings were made weekly from the long digital extensor muscle, and there was a change to a near normal recording by the end of treatment. Plasma phenytoin concentrations were monitored during the trial, and the dose rates used achieved a steady state with a mean plasma level of 37 +/- 7 mumol/l. There was wide variability between plasma concentrations in different horses, although there was no difference in absorption between administration of the phenytoin as a paste, or when it was mixed in the feed. Although mild tranquilization was seen after treatment, there were no clinical, haematological or biochemical signs of toxicity from the phenytoin therapy.