It is widely believed that body fatness (and hence total body mass) is regulated by a lipostatic feedback system. This system is suggested to involve at least one peripheral signalling compound, which signals to the brain the current size of body fat stores. In the brain the level of the signal is compared with a desirable target level, and food intake and energy expenditure are then regulated to effect changes in the size of body fat stores. There is considerable support for this theory at several different levels of investigation. Patterns of body-mass change in subjects forced into energy imbalance seem to demonstrate homeostasis, and long-term changes in body mass are minor compared with the potential changes that might result from energy imbalance. Molecular studies of signalling compounds have suggested a putative lipostatic signal (leptin) and a complex network of downstream processing events in the brain, polymorphisms of which lead to disruption of body-mass regulation. This network of neuropeptides provides a rich seam of potential pharmaceutical targets for the control of obesity. Despite this consistent explanation for the observed phenomena at several different levels of enquiry, there are alternative explanations. In the present paper we explore the possibility that the existence of lipostatic regulation of body fatness is an illusion generated by the links between body mass and energy expenditure and responses to energy imbalance that are independent of body mass. Using computer-based models of temporal patterns in energy balance we show that common patterns of change in body mass following perturbation can be adequately explained by this 'non-lipostatic' model. This model has some important implications for the interpretations that we place on the molecular events in the brain, and ultimately in the search for pharmaceutical agents for alleviation of obesity.