We tested the hypothesis that training would result in more rapid increase in V̇O2 and V̇CO2 following the onset of submaximal exercise. Six Standardbred horses were used and trained at submaximal intensities for 5 weeks. Horses were trained 6 days per week on an inclined treadmill (10% slope) at intensities between 40 and 70% V̇CO2max. At the onset of training horses ran 2600 m/day which progressively increased to 4500 m/day by week 5. Testing was carried out before training and after weeks 3 and 5 of training. The test involved exercise at 2 speeds, 4 m/s and 8 m/s, both for 3 min with one h recovery between the 2 speeds. All horses performed at both speeds but the initial speed was randomly allocated. An open flow gas collection system was used to measure V̇O2 and V̇CO2 every 15 s for the first min of exercise and every 30 s thereafter. After 5 weeks training, mean body weight and V̇O2max (mean ± s.e.) of the group had altered from 433 ± 5 kg to 430 ± 8 kg (P > 0.05) and from 104 ± 5 ml/kg/bwt/min to 116 ± 6 ml/kg/bwt/min (P < 0.05), respectively. There were no significant training effects on the rate of change of V̇O2 or V̇CO2 at either of the test speeds. By 45 s, the V̇O2 and V̇CO2 values had reached 85 to 95% of mean steady state values at both intensities of exercise. In general, steady state V̇O2 and V̇CO2 values were evident by 60 s of exercise. We conclude that horses have rapid increases in gas exchange at the start of exercise and these are not influenced by low intensity training.