The aim ofthis study was to evaluate the practicality and accuracy of different attachment sites for the optodes of a pulse oximeter (measuring arterial haemoglobin oxygen saturation) in healthy cattle, and to assess the accuracy of pulse oximetry in diseased cattle with low haemoglobin oxygen saturation values caused by respiratory disease. The tail, the nasal septum and the genital mucosa offemales provided a continuous, stable and intense signal. The smallest bias, and no significant difference between measurements of arterial haemoglobin (mSpO2) with the pulse oximeter and measurements of arterial haemoglobin oxygen saturation (SaO2) with a blood gas analyserwas obtained when the probe was attached to the tail. This site was used to evaluate the accuracy of pulse oximetry in animals with respiratory disease. There was a small bias between the measurements of SaO2 and mSpO2, with a tendency for pulse oximetry to underestimate higher values and to overestimate lower values. The precision of pulse oximetry decreased substantially with the values for SaO2 >80 per cent, which lies outside the clinically relevant range.