Fig. 6. Acquiring and suppressing merging patterns for higher running efficiency.
a Running efficiency of different subsets of adults already possessing the efficiency-enhancing S0-7 + 11 (E1), 5 + 6 + 12 or 5 + 6 + 8 inclusive (E2), E1 or E2 inclusive, or E1 and E2, but with or without the efficiency-reducing S0-3 + 12 (R1) or 4 + 5 + 7 (R2). In each of the eight pairs of bars (mean ± SE), the left denotes values in subjects with (Wi) the indicated R combination, and the right, subjects without (Wo) the combination. P values for the four sets with number of Wi subjects ≥3 were (left to right), p = 0.040, 0.17, 0.0071, 0.0015 (2-tailed Mann–Whiney; **p < 0.01; *p < 0.05; NS, p > 0.05). When E was present, the presence of R decreased the running efficiency. b Running efficiency of different subsets of adults already not possessing E1, E2, E1 or E2 inclusive, or E1 and E2, with or without R1 or R2. P values for the eight sets were (left to right) p = 0.12, 0.17, 0.85, 0.97, 0.27, 0.38, 0.86, 0.048 (2-tailed Mann–Whitney; *p < 0.05; NS, p > 0.05). When E was absent, the absence of R only increased efficiency modestly without statistical significance. c Running efficiency of adults without the R’s, but with either E1 only (blue), E2 only (orange), or E1 and E2 (green). P values were (left to right) p = 0.21, 0.31, 0.38 (Kruskal–Wallis; NS, p > 0.05). The presence of any E combinations increased efficiency about equally when R was absent. d–f Prevalence (% subjects in the group with combination), across subject groups, of combinations E1 only (blue), E2 only (green), E1 and E2 (magenta), and E1 or E2 (red). Percentages were calculated without excluding subjects with R (d), after excluding those with R1 (e) or those with R1 or R2 inclusive (f) (** in d, r across adults = 0.96, p = 4.5 × 10−4, 2-tailed t-test; ** in e, r = 0.91, p = 0.0039; *r = 0.86, p = 0.013). Acquisition of E1 or E2 together with suppression of R1 correlated best with running experience (**). In f, the decrease of the E1-or-E2 frequency in Exp reflects the prominence of R2 – and hence the lower efficiency – in the group (Supplementary Fig. 3B). Source data for all panels are available as a Source Data file.