Twenty-three Thoroughbred racehorses that were presented for a standard incremental exercise test also performed a test at an intensity equivalent to 115% V̇O2max in an effort to evaluate their anaerobic capacity by measuring the maximal accumulated oxygen deficit (MAOD). Submaximal V̇O2 values for speeds 3 to 10 m/s and V̇O2max were determined for horses exercising on a treadmill inclined at a 10% slope. An individual regression equation of speed and V̇O2 was used to calculate the speed for each horse to exercise at 115% V̇O2max and the energy demand for exercise at this intensity. The horses underwent a warm-up period consisting of 5 min at 50% V̇O2max followed by walking for 5 min at 1.5 m/s. The treadmill was then acclerated as rapidly as possible until a speed of 115% V̇O2max was attained. During the test, expired gas samples were collected at 0, 15, 30, 45, 60 s and every 30 s until fatigue. MAOD was calculated by subtracting the measured oxygen uptake from the calculated oxygen demand. A 5 min post exercise blood sample was collected for measurement of plasma [lactate]. The MAOD values ranged from 36 to 94 mlO2 equivalents/kg and the V̇O2max values ranged from 128 to 170 ml/kg/min. The mean ± s.e. MAOD values were 59 ± 3 mlO2 equivalents/kg and mean V̇O2max was 146 ± 2 ml/kg/min. Mean plasma [lactate] was 23.0 ± 1.2 mmol/l. MAOD was weakly correlated with V̇O2max (r = 0.514, P < 0.05), but was not correlated with run time, plasma lactate or treadmill work. The test of MAOD was suitable for use during clinical exercise testing.