Statistical analysis
Pagel’s lambda (λ) is a robust estimate of the strength of phylogenetic signal in a continuous trait (Münkemüller et al., 2012; Pagel, 1999; Molina-Venegas & Rodríguez, 2017). In our study, Pagel’s λ can range from 0 to 1, a λ of 0 indicates that there is no phylogenetic signal in the focal traits, whereas a λ of 1 indicates high phylogenetic signal in which the focal trait evolved according to Brownian motion. We calculated Pagel’s to quantify the strength of the phylogenetic signal in plant traits (plant height, leaf size, seed mass, fruit width, blade length and stem height) of palms, other monocots, dicots and gymnosperms (Cadotte et al., 2013). We evaluated the importance of through randomized tests implemented in the function phylosig of the R package ‘phytools’ (Revell, 2011).
We used a Gaussian distribution with phylogenetic trees, implemented in the R packages ‘phyr’ and ‘ape’. The multivariate phylogenetic generalized linear mixed models (PGLMM) were used to test the effects of leaf size and plant height on seed mass while controlling for phylogeny.
The location of palms, other monocots, dicots, and gymnosperms in a multivariate trait space illustrated by the first two axes of the PCA based on traits of seed mass (fruit width), leaf size (blade length), and plant height (stem height) (Duras, 2020).
The partial R2 for the logistic regression model (Ives & Helmus, 2011) implemented by the R package “rr2” was used to tease apart the relative contributions of leaf size, plant height and phylogeny to the variation in seed mass of palms, other monocots, dicots, and gymnosperms. The partial R2lik for each factor was calculated by comparing the full model with reduced models in which a given factor was removed, and measuring the consequent reduction in the likelihood.