The commercial restricted feeding programme of broiler breeders has a major negative effect on welfare, as the birds are continuously hungry. Objective parameters of hunger are needed to evaluate new management or feeding systems that may alleviate hunger and thus improve broiler breeder welfare. The aim of this experiment was to develop such parameters. Female broiler breeders (Hybro G), n = 10 per treatment, were subjected to different levels of feed restriction and we assumed that we thus induced different levels of hunger. Treatment groups were fed ad libitum and at 90%, 70%, 50%, 35% and 25% of ad libitum. In addition, female broiler breeders that were less intensively selected on growth and food conversion (JA57, n = 10 per treatment) were fed ad libitum and at 70% of ad libitum. At 6-7 weeks of age, home pen behaviour, behaviour in the open field and baseline plasma corticosterone, glucose and nonesterified fatty acid (NEFA) concentrations were determined. Thereafter, birds were subjected to the feed intake motivation (FIM) test that measures compensatory feed intake. Linear relationships between the level of restriction and the compensatory feed intake on the first days of the FIM test and the glucose/NEFA ratio were found which indicate that these parameters can be used to quantify the level of hunger. Curvilinear relationships between sitting and standing in the home pen, walking in the open field and level of restriction were found. A third-grade hyperbolic relationship between the level of restriction and the baseline plasma corticosterone concentration was found. Thus, these parameters do not seem to represent a straightforward relation with the level of hunger. Intensive selection on growth and food conversion seems to have affected the behaviour but not the physiological responses to feed restriction, because the JA57 birds responded behaviourally, but not physiologically, different to feed restriction as compared with the Hybro birds. (C) 2003 Elsevier Science Inc. All rights reserved.