Table 1.

Results from tests of H2 and H3, from a linear mixed effects model analysis (see Methods), focusing on four kinematic parameters: log velocity (u), log Froude^0.5 ($\hat{u}$), duty factor (DF) and relative stride frequency (RSF).

Test + data	log(u)	log($\stackrel{\mathit{ˆ}}{\mathit{u}}$)	DF	RSF
H2: dataset 3	−0.06 ± 0.03	−0.19 ± 0.03	0.07 ± 0.01	−0.01 ± 0.02
maximal u	(p = 0.180)	(p = 0.026)	(p = 0.013)	(p = 0.417)
H2: dataset 2	0.01 ± 0.05	−0.12 ± 0.04	0.02 ± 0.01	−0.02 ± 0.01
running (DF < 0.50)	(p = 0.806)	(p = 0.044)	(p = 0.029)	(p = 0.014)
H3: Alli-symm	0.05 ± 0.21ab	−0.41 ± 0.22ab	0.57 ± 0.03a	0.30 ± 0.03a
H3: Crocs-symm	−0.05 ± 0.11a	−0.48 ± 0.12a	0.56 ± 0.02a	0.30 ± 0.02a
H3: Crocs-asymm	0.46 ± 0.14b	0.04 ± 0.14b	0.46 ± 0.02b	0.41 ± 0.02b
H3: Overall p-value	<0.0001	<0.0001	<0.0001	<0.0001

Residual variances were allowed to vary depending on the recording frequency (Hz). Bold font emphasizes where p < 0.05. For H2, relationships with log(body mass) are based only on Crocodyloidea and asymmetrical gait data. Adjusted regression coefficients (i.e. slope of the regression line) ± standard errors are shown. Results from dataset 3 are compared with those from dataset 2, in which the model accounted for repeated measures from the same subjects. For H3, we present a comparison of three groups (“alli-symm” = Alligatoroidea symmetrical gaits; “crocs-symm” = Crocodyloidea symmetrical gaits; “crocs-asymm” = Crocodyloidea asymmetrical gaits), using dataset 1 (all valid strides), focusing on adjusted means ± standard errors. Individual number was used as a random effect in the analysis. There was no statistical difference between groups sharing the same letters (superscript a or b).