A computer model of body mass and composition in relation to gross energy balance is constructed. The model is built using conventional empirical physiological formulae rather than statistical or analytical mathematical techniques. The model is applied to the Minnesota and other experiments and produces as good or better simulations of observed values of changes in body weight than reported for other formulae or models. Alternative physiological mechanisms concerning metabolic adaptions to starvation, changes in time activity budgets and the energy equivalents of weight loss offer equally good simulations of experimental results. The present analysis highlights the survival value of a basal metabolic depression during starvation and indicates an optimal body composition of 10% mobilizable fat for starvation survival for a 70 kg man. Proper quantification of the effects of the physiological mechanisms involved depends on new experimental data, however. Long term continuous monitoring of time activity budgets are a necessary part of such experiments.