The aim of this study was to compare axillary and tympanic membrane (TM) temperature measurements to rectal temperature in a large group of clinical canine patients. We also sought to ascertain whether certain factors affected the differences between the measurements and to compare the ease of measurement. Axillary temperatures were easy to obtain but tended to be lower than rectal readings (median difference 0.6°C). In 54.7 per cent of dogs there was a difference of >0.5°C between the two readings. Weight, coat length, body condition score and breed size were significantly associated with the difference between the rectal and axillary temperature. TM temperatures were more similar to rectal temperatures (median difference 0°C) but in 25 per cent of dogs, there was a difference of >0.5°C between rectal and TM readings. TM measurements were less well tolerated than axillary measurements. None of the factors assessed were associated with the difference between the rectal and TM temperature. As a difference of >0.5°C has previously been described as unacceptable for different methods of temperature measurement, neither axillary nor TM temperatures are interchangeable with rectal temperatures for the measurement of body temperature.