1. Animals frequently exhibit consistent among-individual differences in behavioural and physiological traits that are inherently flexible. Why should individuals differ consistently in their expression of labile traits? Recently, positive feedbacks between state and behaviour have been proposed as a parsimonious explanation for the maintenance of consistent among-individual differences in both state and behaviour. If state affects behaviour, and behaviour reciprocally affects state, then even chance differences in either state or behaviour that arise among-individuals could be maintained over extended periods of time. 2. We tested for positive feedbacks experimentally using wild-caught red knots (Calidris canutus islandica). In the wild, knots exhibit consistent among-individual differences in digestive physiology (gizzard mass) and foraging behaviour (diet), two inherently labile traits. 3. Experimentally manipulated diet quality had a large effect on gizzard mass. Experimentally manipulated gizzard mass reciprocally influenced total food eaten during ad libitum trials. 4. The effect of gizzard mass on diet choice, though in the predicted direction, was not statistically significant. Individuals exhibited consistent differences in foraging behaviour (of unknown origin) independent of current gizzard mass, as well as large residual (unexplained) variance in foraging behaviour. These sources of variation in foraging behaviour overruled the gizzard mass-dependent foraging behaviour and hence eroded the treatment-related differences in gizzard mass. 5. We conclude that positive feedbacks between diet choice and gizzard mass play at best a limited role in maintaining among-individual variation in gizzard mass in knots. Furthermore, we suggest that many models of state-behaviour feedbacks likely overestimate their potential importance in maintaining long-term among-individual variation in labile traits because they fail to account for the effects of additional factors that may act to disrupt the feedback loops. 6. The among-individual differences in diet choice observed during solitary foraging trials eroded the consistent among-individual differences in gizzard mass observed following periods of staple diet treatments in which knots foraged in social groups. As such, we suggest that social foraging interactions may play an important role determining the expression of foraging behaviours (e.g. intake rate) that themselves influence gizzard mass. Further studies are needed to experimentally test this proposed mechanism.