In human and veterinary medicine, arterial and venous haemoglobin oxygen saturations are often used to estimate the severity of a disease and to guide therapeutic decisions. In veterinary medicine, haemoglobin oxygen saturation (SO(2)) is usually calculated using a blood gas analyser and algorithms developed for humans. It is possible, therefore, that the values obtained in animals may be distorted, particularly in animals with a high haemoglobin oxygen affinity, like young calves. In order to verify this hypothesis, we compared the arterial (SaO(2)) and venous (SvO(2)) haemoglobin oxygen saturations calculated using three different algorithms, and the oxygen exchange fraction (OEF) at the tissue level, which is the degree of haemoglobin desaturation between arterial and venous blood (SaO(2)-SvO(2)), with the values obtained from the whole bovine oxygen equilibrium curve (OEC) determined by a reference method. The blood gas analysers underestimated SvO(2) values; consequently, the OEF was overestimated (by about 10%). Two methods of reducing these errors were assessed. As the haemoglobin oxygen affinity decreases during the first month of life in calves a relationship between PO(2) at 50% haemoglobin saturation (P50) and age was established in order to correct the calculated values of venous and arterial SO(2), taking into account the estimated position of the OEC. This method markedly reduced the error for SvO(2) and OEF. Secondly, the SO(2) was calculated using a mathematical model taking into account the age of the animal and the specific effects of pH, PCO(2), and temperature on the bovine OEC. Using this method, the mean difference between the OEF values calculated using the mathematical model and those calculated by the reference method was close to zero. The errors produced by blood gas analysers can thus be minimised in two ways: firstly, by simply introducing a P50 estimated from the age of the calf into the analyser before the measurement; and secondly, by calculating the SO(2) using a mathematical model applied to the bovine OEC.