In this study we tested the hypothesis that exercise induces an adaptive response in the developing skeleton which may be monitored in vivo by measuring biochemical markers of bone metabolism. The effects of exercise on two biochemical markers of bone formation were determined; the carboxy-terminal propeptide of type I procollagen (PICP), and the bone-specific isoenzyme of alkaline phosphatase (BAP), and one putative marker of resorption, the pyridinoline crosslinked telopeptide domain of type I collagen (ICTP). All three markers were measured for a year in 2-year-old thoroughbred horses exercised three times a week on a treadmill, and values compared to a control group of age-matched animals. Levels of all three markers fell in both exercised and control groups over the 12-month period reflecting normal age changes. However, there were differences between groups in the pattern of this decrease. When expressed as a percentage of baseline values, BAP was higher (p < 0.05) at 2 months and both BAP and the PICP were higher at 4 months (p < 0.01 and p < 0.05, respectively) in the exercised group, reflecting an increase in bone turnover in this group in the early stages of training. PICP levels were also elevated in the exercised group at 10 months and this result indicates an increase in bone turnover at this time. The changes in ICTP were different; at 2 months, levels were higher in exercised animals than in controls, but there was no significant difference between the two groups at 4 and 6 months. After 8 months, ICTP levels in the exercised group increased returning to near baseline values at 10 months.(ABSTRACT TRUNCATED AT 250 WORDS)