We used ultrasonic imaging to monitor short-term changes in the pectoral muscle size of captive red knots Calidris canutus. Pectoral muscle thickness changed rapidly and consistently in parallel with body mass changes caused by flight, fasting and fuelling. Four knots flew repeatedly for 10 h periods in a wind tunnel. Over this period, pectoral muscle thickness decreased in parallel with the decrease in body mass. The change in pectoral muscle thickness during flight was indistinguishable from that during periods of natural and experimental fasting and fuelling. The body-mass-related variation in pectoral muscle thickness between and within individuals was not related to the amount of flight, indicating that changes in avian muscle do not require power-training as in mammals. Our study suggests that it is possible for birds to consume and replace their flight muscles on a time scale short enough to allow these muscles to be used as part of the energy supply for migratory flight. The adaptive significance of the changes in pectoral muscle mass cannot be explained by reproductive needs since our knots were in the early winter phase of their annual cycle. Instead, pectoral muscle mass changes may reflect (i) the breakdown of protein during heavy exercise and its subsequent restoration, (ii) the regulation of flight capacity to maintain optimal flight performance when body mass varies, or (iii) the need for a particular protein:fat ratio in winter survival stores.