It has long been known that body mass and, more specifically, lean body mass are strongly correlated with maximal oxygen uptake (V̇O2max) in man and animals. However, there are no data to date describing this phenomenon in the horse. The purpose of this paper is to examine the relationship between body composition and V̇O2max in the horse. Twenty-three healthy and unfit Standardbred mares performed an incremental exercise test (GXT) to measure V̇O2max. Rump fat thickness (RTH), a measure of fat covering, was measured using B-mode ultrasound. Plasma volume, total blood volume and red cell volume were determined, using the Evan's Blue dye dilution technique and packed cell volume. V̇O2max was correlated with body mass (r = 0.541; P<0.01) and exercise haematocrit (exHCT; r = 0.407; P<0.05) but not RTH or the other haematological variables. To eliminate the influence of body mass on the individual variables, a regression analysis was performed on the mass-residuals of V̇O2max, RTH, plasma volume and exHCT. The residuals of V̇O2max were correlated negatively with the residuals of RTH (r = −0.687; P = 0.0003) and positively with the residuals of exHCT (r = 0.422; P = 0.045) but not plasma volume. V̇O2max could be predicted from a linear combination of the residuals of RTH and exHCT (r = 0.767; P<0.0001). These data indicate that V̇O2max in the horse is significantly related to fat-free mass (FFM), independent of body mass. Red blood cells from the splenic reserve constitute an important factor in the horse's ability to achieve a high V̇O2max. Therefore, lean body mass may be a more appropriate basis for assessing metabolic function in the athletic horse.