When an animal has to meet increased demands on its working capacity, for example, for thermoregulation or parental care, two strategies are available. The animal can reallocate energy from costly maintenance processes – such as immunological defence or DNA repair systems (compensation hypothesis) – or it may try to increase the rate of energy intake or efficiency of digestion by increasing the size of the alimentary tract (increased–intake hypothesis). By manipulating brood size, I affected parental effort among marsh tits (Parus palustris) as demonstrated by a significant increase in parental feeding rate with experimental brood size. Basal metabolic rate (BMR) increased both with manipulated brood size and individual feeding rate, supporting the predictions from the increased–intake hypothesis. Furthermore, I found a direct positive relation between BMR and energy expenditure, measured with the help of the doubly labelled water technique. The cost of achieving a higher working capacity is substantial since BMR increases more quickly than the surplus energy available for work. Since the cost of a high sustained workload was not primarily dependent on a reallocation of energy away from maintenance, such a cost should be searched for among the detrimental effects of a high metabolic rate per se, for example, an increased oxidative damage to DNA, proteins and lipids.