The development and occurrence of anorexia, the voluntary reduction in food intake during parasitic infections in animals, is somewhat paradoxical and contrary to conventional wisdom and expectation. We take the view that its occurrence is an evolved, costly behavioural adaptation which serves a function. Five such functional and general hypotheses to account for it are developed: (1) anorexia is induced by the parasite for its own benefit; (2) food intake decreases to starve parasites; (3) the negative effect on the host's energetic efficiency during parasitic diseases has a direct effect on food consumption; (4) food intake decreases for the purpose of promoting an effective immune response in the host; and (5) anorexia allows the host to become more selective in its diet, and thus select foods that either minimize the risk of infection or are high in antiparasitic compounds. Only hypotheses (4) and (5) survive the comparison for consistency with the physiological, metabolic and behavioural alterations that occur during the development of parasitic infections, and with the rule of generality (i.e. account for its occurrence in both protozoan and helminth infections). Both surviving hypotheses will need further experimental testing for their support or rejection, and such experiments are proposed. Also, the advantages and consequences of viewing anorexia during parasitic infections within a functional framework are discussed. These arise from the recognition that anorexia is a disease-coping strategy, part of the mechanism of recognition of parasite invasion by the immune system, which leads to a modification of the host's feeding behaviour. Copyright 1998 The Association for the Study of Animal Behaviour