Table 2. PGLS models with manipulation complexity as response variable and brain size as explanatory variables, terrestriality and diet quality as covariates singly and as combined models (n = 34, Homo sapiens excluded).

Model	P-value model	λ	adj. r2	AIC	∆AIC	Predictor variables	Estimate	Std. error	P-value
model 1	<0.001	0	0.728	60.773	–	log brain	1.286	0.387	0.002
						log body	−0.551	0.312	0.087
						terrestriality	0.948	0.328	0.007
model 2	<0.001	0	0.721	62.463	1.690	log brain	1.376	0.429	0.003
						log body	−0.618	0.346	0.081
						terrestriality	0.927	0.335	0.010
						log diet quality	−0.313	0.607	0.610
model 3	<0.001	0	0.663	67.105	6.332	log brain	1.274	0.430	0.006
						log body	−0.416	0.343	0.234
model 4	<0.001	0	0.659	68.438	7.665	log brain	1.422	0.474	0.005
						log body	−0.531	0.376	0.165
						log diet quality	−0.513	0.666	0.447
model 5	<0.001	0.147	0.567	68.498	7.725	terrestriality	0.863	0.374	0.028
						log body	0.439	0.098	<0.001
model 6	<0.001	0.211	0.461	73.783	13.010	log diet quality	−0.069	0.659	0.918
						log body	0.531	0.101	<0.001

Body mass is always included as covariate.

Significant effects and best-fitting models are highlighted in bold face.