Several studies suggest that the temperature of the inner canthus of the eye (T ca ), determined with infrared thermal imaging, is an appropriate method for core temperature estimation in mass screening of fever. However, these studies used the error prone tympanic temperature as a reference. Therefore, we compared T ca to oesophageal temperature (T es ) as gold standard in 10 subjects during four conditions: rest, exercise, recovery and passive heating. T ca and T es differed significantly during all conditions (mean DT es 7 T ca 1.80 + 0.898C) and their relationship was inconsistent between conditions. Also within the rest condition alone, intersubject variability was too large for a reliable estimation of core temperature. This poses doubts on the use of T ca as a technique for core temperature estimation, although generalization of these results to fever detection should be verified experimentally using febrile patients.