Table 3.

Results of the standardised major axis regression models, testing whether relationships between dipterocarp hydraulic traits are modulated by forest type (A = alluvial, S = sandstone and K = kerangas).

Hydraulic traits		Coordination‐P			Intercept			Slope			Slope‐P	Intercept–P	Pairwise‐P			r 2
		A	S	K	A	S	K	A	S	K			A–S	A–K	S–K	A	S	K
P50	P88	< 0.001	< 0.001	< 0.001	−0.23	0.09	0.07	0.69	0.69	0.69	0.2	0.021	0.02	0.06	0.76	0.84	0.55	0.70
P50	K s	0.02	0.36	0.067	−3.32		−4.54	0.04		0.06	0.004		0.001	0.36	0.04	0.11		0.11
P50	K sleaf	0.002	0.64	0.55	−3.04			178.2			< 0.001		< 0.001	0.70	< 0.001	0.20
P50	LS	0.003	0.024	0.19	−1.31	−1.48		−0.00014	−0.00014		0.33	< 0.001	0.15	< 0.001	0.001	0.19	0.15
P50	WD	0.86	< 0.001	0.025		0.33	0.77		−5.83	−5.83	0.30	0.01	0.1	0.09	0.03		0.19	0.15
P88	K s	0.02	0.41	0.56	−4.47			0.064			0.014		0.011	0.97	0.018	0.10
P88	K sleaf	0.01	0.27	0.95	−3.01			2.32			0.33	< 0.001	0.003	< 0.001	0.003	0.13
P88	LS	0.015	0.99	0.15	−1.45			−0.0002			0.034		0.25	0.08	0.01	0.13
P88	WD	0.42	0.012	0.017		−0.21	−4.6		−6.93	−13.3	0.005		0.46	0.013	0.002		0.11	0.15
K s	K sleaf	< 0.001	< 0.001	< 0.001	96.5	95.1	91.1	12.3	12.3	12.3	0.08	0.002	0.47	< 0.001	0.04	0.59	0.41	0.43
K s	LS	0.56	0.54	0.016			39.0			0.003	0.008		0.002	0.08	0.15			0.13
K s	WD	0.12	0.63	0.58
K sleaf (×103)	LS	< 0.001	< 0.001	< 0.001	9.4	5.3	10.5	0.007	0.003	0.009	< 0.001		< 0.001	0.29	< 0.001	0.37	0.33	0.59
K sleaf (×103)	WD	0.33	0.97	0.49
LS	WD	0.67	0.09	0.11

Models were fitted with forest type affecting both the intercept and slope of the trait–trait relationships. When trait–trait relationships were significant in at least one forest type (coordination‐P < 0.05), we tested for the effect of forest type on the slope of trait coordination. Slope‐P is the probability of the trait–trait relationship of the different forest types having the same slope. When forest type did not have a significant effect on the slope of the trait coordination, we refitted the models for the effect of forest type on the intercept of trait relationship. Intercept‐P is the probability of the relationship between traits in different forests having the same intercept. Pair‐wise‐P is the probability of two different forest types having similar slopes (if slope differences are significant) or intercepts (if intercept differences are significant). r ² is the coefficient of determination of the different models. Bold values highlight significant forest‐level interactions, and italics indicate marginally significant interactions (P < 0.07). For nonsignificant relationships, entries are left blank. Note that the left hydraulic traits column was fitted as the y‐axis in the model.

A, alluvial forest; K, kerangas forest; K _s, maximum hydraulic specific conductivity (kg m m⁻² s⁻¹ MPa⁻¹); K _sleaf, maximum hydraulic leaf‐specific conductivity (kg m m⁻² s⁻¹ MPa⁻¹); LS, leaf : sapwood area ratio (m² m⁻²); P50, water potential causing a 50% reduction in hydraulic conductivity (MPa); P88, water potential causing an 88% reduction in hydraulic conductivity (MPa); S, sandstone forest; WD, wood density (g cm⁻³).