Body composition of animals and man is generally assumed to be a regulated phenomenon. Understanding the factors that are involved in such regulation is important in at least two different contexts. The obesity epidemic sweeping across developed nations has been described by the WHO as the greatest health threat facing Western societies. Understanding the underlying physiological factors that lead sectors of the population to fail in their attempts to regulate body mass and composition are of key importance in the drive to develop pharmaceutical remedies for this serious condition. A knock on effect for the agricultural sector however is the consumer demand for leaner animal products. This places a premium on understanding how body composition is regulated in livestock. Moreover the financial rewards for improving the energetic efficiency of production provide an additional incentive for understanding the details of energy regulation that underlie control of body composition. While direct genetic and physiological studies of livestock are feasible the mouse provides a convenient model animal that can inform our understanding of the conserved physiological mechanisms in both man and other animals. The key advantages of using the mouse in this context are that its small size and short breeding cycle enable experiments to be performed rapidly. Measuring physiological components of energy balance in mice is easily performed. In addition the mouse genome has been sequenced and the tools for performing large scale gene expression studies are already commercially available, opening up the capacity to perform integrative physiological studies form the level of the genome to the whole animal. Several genetic models of mice are known which have monogenic forms of extreme obesity. The best known of these is the ob/ob mouse, which is deficient in production of the adipokine leptin. However, many other single gene mutants are known which generate similar effects. Dissecting the loci of such effects has enabled us to construct a working model of how adiposity is signalled in the brain via the melanocortin system and how feedback loops including this system may contribute to regulation of energy balance. While some individuals have been identified that have similar genetic disruptions these are very rare, and it is clear that variation in human and livestock body composition generally reflects polygenic effects. More useful models may therefore be mice that have been selected for many generations for traits that impact on their body composition. The short generation time of the mouse has been extremely useful in the generation of such lines. My own group has been collaborating with the University of Edinburgh where such long-term selection experiment was initiated in the 1970s. We have in particular been quantifying the energy balance of mice long term selected for fat and lean body composition. We first showed that these polygenic effects do not seem to include polymorphisms within the leptin signalling system. By comparing total energy intake, with measures of resting energy expenditure the contribution of different energy compartments to energy balance can be assessed. These studies indicated that the major difference in energy regulation between the lines is that the lean line diverts substantially more energy into activity than the fat line. Direct measures of activity have been made to confirm this effect with mixed results. Using passive infrared detection devices the converse result is found – that the lean mice are less active, but using running wheels the logged activity matches the measured energy balance. These data indicate that the factors influencing energy balance reflect complex polygenic effects that reach far outside the leptin signalling system. The mouse is an ideal tool to explore these effects with implications for both the livestock industry and the obesity epidemic.