Extended Data Fig. 8: Additional data on the architecture of the pectoralis muscle in the red-tailed hawk and in non-soaring birds.

a, Three specimens of soaring birds (B. jamaicensis) do not differ significantly from each other in their mean pectoralis fascicle lengths (n = 80 fascicles per muscle), but each have significantly shorter fascicle lengths than all four non-soaring species measured here. b, Functional morphospace (normalized mean fascicle length versus normalized PCSA) plots of the pectoralis muscle in soaring and non-soaring birds illustrating their similar PCSAs, but shorter fascicle lengths in soaring birds. This suggests a more force-based architectural specialization in soaring muscles versus a more power/displacement specialization in non-soaring species. c,d, The strong cranial-to-caudal variation in fascicle length noted at the whole-muscle level (see Fig. 3c in main text), with significant increases in fascicle length in caudal regions, is present in both the superficial (c) and the deep (d) (whiskers denote min and maximum data range, n = 40 per bird, n = 10 per region) layers of the pectoralis (although note that the overall fascicle lengths are consistently longer in superficial than in deep parts of the muscle). Open circles indicate non-soaring birds, whereas filled circles indicate the red-tailed hawks in c,d. Two-way factorial ANOVA test statistics are presented in c,d, with statistical significance described as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.