Abstract: Nine yearling horses, stabled in individual stalls, were used in a trial to determine the diurnal pattern of wood chewing and the effects of exercise on this behavior. The trial was a Latin square design conducted over three 2-wk periods during which each horse was exposed to each of the three following treatments: 1) no exercise (NE), 2) exercise after the morning feeding (AM), and 3) exercise in the afternoon (PM). Horses were fed a complete pelleted feed in the morning and both pelleted feed and long-stemmed hay in the afternoon. Exercise consisted of 45 min on a mechanical walker followed by 45 min in a paddock with bare soil. Each stall was equipped with two untreated spruce boards during each period for wood chewing. Wood chewing was evaluated by videotaping each horse for 22 h during each period, determining the weight and volume of the boards before and after each period, and by visual appraisal of the boards. Intake of trace mineralized salt was also measured. Wood chewing occurred primarily between 2200 and 1200. All measures of wood chewing were correlated when totals for the entire 6 wk were analyzed. When analysis was performed on 2-wk values, videotape results were not correlated with volume or weight loss of boards. Horses chewed more when on the NE treatment (511 s/d) than when on AM or PM (57 and 136 s/d, respectively; P less than .05). Salt intake tended to be greater for NE than for the other treatments (P less than .10).(ABSTRACT TRUNCATED AT 250 WORDS)