Abstract: REASONS FOR PERFORMING STUDY: Field based studies of changes in fitness are difficult to conduct in galloping Thoroughbreds. Measurements of heart rate (HR) and real time velocity with a global positioning system (GPS) could provide a method for routine field studies of fitness. OBJECTIVE: To investigate measurements of fitness in the field without using multiple, standardised steps of increasing velocity in the exercise test. METHODS: Twelve healthy 3- to 5-year-old Thoroughbreds were used in the study of reliability, and twelve 2-year-old Thoroughbreds in the investigation of the effect of training. Five second averages of velocity and HR were recorded during a typical fast exercise training session, using a GPS and Polar HR monitor. A standardised exercise test protocol was not used. Regression analyses using trot and gallop data were used to calculate velocities at HR of 200 beats/min (V200), and at maximal HR (VHRmax). Data were collected on consecutive 'fast' day training sessions to assess the reliability of measurements. The effect of training was investigated with fitness tests in weeks 2 and 6 of 'fast' training. Absolute and relative differences were calculated to evaluate reliability, and paired t tests were used to detect an effect of training. RESULTS: VHRmax, V200 and HRmax were reliable measurements of fitness, with mean differences of 2% or less. Reliability of VHRmax was not dependent on VHRmax. VHRmax and V200 both increased significantly with training (P<0.01), but there was no effect of training on HRmax. There were no significant changes in the slope or intercept of the regression equations after training. CONCLUSIONS: Velocity and HR measurements during field gallop exercise provided reliable measures of fitness which enabled a measurement of the response to training. POTENTIAL RELEVANCE: This approach offers a simple, noninvasive method for monitoring adaptations to training in the field.