The circadian rhythm of core temperature depends upon several interacting rhythms, of both endogenous and exogenous origin, but an understanding of the process requires these two components to be separated. Constant routines remove the exogenous (masking) component at source, but they are severely limited in their application. By contrast, several purification methods have successfully reduced the masking component of overt circadian rhythms measured in field circumstances. One important, but incidental, outcome from these methods is that they enable a quantitative estimate of masking effects to be obtained. It has been shown that these effects of activity upon the temperature rhythm show circadian rhythmicity, and more detailed investigations of this have aided our understanding of thermoregulation and the genesis of the circadian rhythm of core temperature itself. The observed circadian rhythm of body temperature varies with age; in comparison with adults, it is poorly developed in the neonate and deteriorates in the aged subject. Comparing masked and purified data enables the reasons for these differences--whether due to the body clock, the effector pathways or organs, or irregularities due to the individual's lifestyle--to begin to be understood. Such investigations stress the immaturity of the circadian rhythm in the human neonate and its deterioration in elderly compared with younger subjects, but they also indicate the robustness of the body clock itself into advanced age, at least in mice.