One Gotland pony, one New Forest pony, and one warm-blooded horse exercised one after the other on a 6450 m long track. The horses walked, trotted and cantered according to a predetermined scheme, which took about 50 minutes including three stops. The scheme was repeated on five consecutive days when horses were: 1) unclipped 2) unclipped？+？blanket during recovery, 3) left or right side clipped, 4) clipped, and 5) clipped？+？riding blanket？+？blanket during recovery. Heart rate (HR) was measured with telemetry, respiratory rate (RR) by counting flank contractions, skin temperatures by thermistor probes, and rectal temperature with a digital thermometer. Skin wetness (SW) was estimated by ocular inspection (dripping？=？5, dry？=？0).Mean outdoor temperature varied from -1.1 to - 8.7°C. HR increased progressively during exercise with no difference between treatments. Maximum RR was 77 ± 30 breaths/min (unclipped) and 49 ± 27 breaths/min (clipped). The lowest skin temperature was 17.5 ± 2.7°C in a hind leg during exercise, which increased to 34.5 ± 0.1°C during recovery. Rectal temperature was elevated during recovery in unclipped, but not in clipped horses and skin temperature at base of tail was elevated during recovery except in unclipped horses without blanket. Moisture after exercise scored 3.2 ± 0.8 in unclipped and zero in clipped horses.Leg skin temperature initially dropped at onset of exercise in clipped horses, and then increased after about 30 minutes due to internal heat from the working muscles. These changes were not significant when clipped horses had riding blankets, whereas unclipped horses became overheated as judged from respiratory rate and elevated rectal temperature. Providing clipped horses with blankets dampened the changes in leg skin temperature during exercise.