Abstract: REASONS FOR PERFORMING STUDY: Running ability of Thoroughbred racehorses is correlated with maximal oxygen uptake, and the velocity at maximal oxygen uptake is highly correlated with the velocity at maximal heart rate (VHRmax). OBJECTIVE: To investigate the relationship between VHRmax and racing performance, expressed as 'peak dollars earned per race start'. METHODS: Heart rate (HR) and velocity were recorded in 25 Thoroughbred racehorses during trotting and subsequent fast gallops in the field at velocities of 15-16 m/sec. Velocity was recorded by a global positioning system (GPS). Maximal HR (HRmax) and maximal velocity (Vmax) were identified, and a linear regression of HR on velocity for trotting and galloping data was constructed to derive VHRmax. Horses followed the training programme designed by one trainer, had at least 6 race starts and were clinically sound at the time of testing. Race earnings were expressed as the peak dollars per start in the horse's race career. Data were normalised using the results for the square root of 'peak dollars earned per race start' and the significance of associations between variables was determined by correlation coefficient and least square analyses. RESULTS: Horses with higher VHRmax earned significantly more dollars per race start (r = 0.41, P<0.05), and horses with VHRmax less than 14.5 m/sec had mean earnings of less than A$2500 per race. There were no correlations between race earnings and either HRmax or Vmax. CONCLUSION: Field studies of the relationship between HR and velocity with a GPS enable identification of horses with limited earnings. POTENTIAL RELEVANCE: This study demonstrates that a field test of fitness of Thoroughbred racehorses that correlates with retrospective racing ability is feasible. The technique has potential application in commercial training environments assisting with decisions concerning racing careers of individual racehorses.