Using three horses we imvestigated responding on several concurrent variable-interval schedules. Each horse was first trained, using the method of successive approximations, to press a response lever with its upper lip. Following successful acquisition of the lip-press response, horses spent several days on pretraining schedules. These included a continuous reinforcement schedule and three variable-interval concurrent schedules: VI 15-sec VI 15-sec, VI 30-sec VI 30-sec, and VI 45-sec VI 45-sec. Horses were then exposed to the experimental sessions; sessions were conducted daily in the horse's home stall and lasted until 60 grain reinforcers had been received. All horses were exposed to five variable-interval schedules: VI 30-sec VI 90-sec, VI 45-sec VI 75-sec, VI 60-sec VI 60-sec, VI 75-sec VI 45-sec, and VI 90-sec VI 30-sec. Like previous experimenters, we found a close matching relationship between the horse's relative rates of responding and the relative rates of reinforcement. When each horse's data was fitted with a regression line, using the least squares method, more than 97% of the variance was accounted for. These results expand the generality of the matching law to include horses as well as providing some other information about the response characteristics of horses.