Table 3.

Statistical model for altitude-related variation in vein trait as a function of aridity index (AI), growth form (GF), and their interaction.

Fixed Effect	df	VD				VT				VV
		Estimate	Std.Error	P	SS%	Estimate	Std.Error	P	SS%	Estimate	Std.Error	P	SS%
Intercept	118	0.63	0.14	0.036		1.43	0.08	< 0.001		0.39	0.15	0.009
GF	118	0.28	0.09	0.003	0.47	−0.31	0.13	0.018	4.00	-0.38	0.23	0.105	6.39
AI	118	0.16	0.14	0.339	1.12	−0.05	0.09	0.571	0.53	0.05	0.16	0.764	1.22
GF×AI	118	−0.35	0.12	0.005	1.11	0.57	0.16	0.001	9.95	0.86	0.29	0.004	12.47
R2 m		0.06				0.15				0.15
R2 c		0.21				0.15				0.15

df, degree of freedom; SS%, percentage of sum of squares explained. All the trait data are log₁₀-transformed prior to analysis and their abbreviations are in Table 1 .

Linear mixed-effects model was fit by restricted maximum likelihood. Random effects in model were ‘site’. Marginal R² (R² _m) is concerned with variance explained by fixed factors, and conditional R² (R² _c) is concerned with variance explained by both fixed and random factors.