Many birds and mammals show substantial circadian variation in body temperature, which has been attributed to fluctuations in ambient temperature and energy reserves. However, to fully understand the variation in body temperature over the course of the day, we also need to consider effects of variation in work rate. We made use of a dataset on body temperature during the resting and active periods in female marsh tits (Poecile palustris) that bred in a temperate area and were subjected to experimental changes in reproductive investment through brood size manipulations. Furthermore, the amplitude increased with daytime, but were unaffected by nighttime, ambient temperature. Amplitudes in females with manipulated broods were 44% above predictions based on inter-specific allometric relationships. In extreme cases, amplitudes were > 100% above predicted values. However, no individual female realised the maximum potential amplitude (8.5 °C, i.e. the difference between the highest and lowest body temperature within the population) but seemed to prioritise either a reduction in body temperature at night or an increase in body temperature in the day. This suggests that body temperature amplitude might be constrained by costs that preclude extensive use of both low nighttime and high daytime body temperatures within the same individual. Amplitudes in the range found here (0.5–6.7 °C) have previously mostly been reported from sub-tropical and/or arid habitats. We show that comparable values can also be found amongst birds in relatively cool, temperate regions, partly due to a pronounced increase in body temperature during periods with high work rate.