There have been few comparisons between the relationship between the mechanical power requirements of flight and flight speed obtained using different approaches. It is unclear whether differences in the power–speed relationships reported in the literature are due to the use of different techniques for determining flight power or due to inter-specific differences. Here we compare the power–speed relationships in cockatiels ( Nymphicus hollandicus ) determined using both an aerodynamic model and measurements of in vitro performance of bundles of pectoralis muscle fibres under simulated in vivo strain and activity patterns. Aerodynamic power was calculated using different ranges of values for the coefficients in the equations: induced power factor ( k 1.0–1.4), the profile ( C D,pro 0.01–0.03) and parasite drag ( C D,par 0.05–0.195) coefficients. We found that the aerodynamic power-speed relationship was highly sensitive to the values assumed for these coefficients and best fit the power calculated from in vitro muscle performance when k =1.2, C D,pro =0.02 and C D,par =0.13.