Data analysis
Anatomic traits and their contrast between earlywood and latewood were checked for normality using the ‘Sharpiro.test’ function in R (R Core Team 2018). Phylogenetic independent contrast (Pics) of tracheid characters and maximum plant height were calculated using the R package ‘ape’ (Paradis and Schliep 2018), and their correlations were tested using the R package ‘psych’ (Revelle 2018). Phylogenetically paired T-test was performed to test for a significant difference between corresponding traits in earlywood and latewood using the R package ‘phytools’ (Revell 2012), which is similar to paired t-test but take phylogeny into account.
As all of the five tracheid traits were positively correlated (Appendix 3), we applied a Phylogenetic principal component analysis (pPCA) to aggregate traits into PCs and to detect non-independent values of variables (traits or PCs) with the phylogenetic relationship (phylogenetic autocorrelation) between species. Phylogenetic principal components were analyzed using the R package ‘adephylo’ (Jombart, Balloux, and Dray 2010), in which phylogenetic proximities were calculated using Abouheif’s proximity, and the resulting matrix of phylogenetic proximities was used to calculate phylogenetic autocorrelation, i.e., Moran’s value (Jombart et al. 2010, Zheng and Martínez-Cabrera 2013). Phylogenetic generalized least square regression (PGLS) was used to build bi-variate and multi-variate models in a phylogenetic context using the R package ‘nlme’ (Pinheiro et al. 2015), in which cell traits were expressed as a function of environmental variables, i.e., mid-point latitude, MAT, MAP, and interactions between climatic variables. R2 for phylogeny and environmental variables in each model were calculated using the R package ‘rr2’ (Ives 2019).