Skip to main content
. 2020 Apr 28;20:185. doi: 10.1186/s12870-020-02390-8

Table 2.

Variable loading scores of 16 selected global traits (12 root-related and four shoot traits) and the proportion of variation of each principal component

Trait PC1 PC2 PC3 PC4 PC5
SRLZ1 0.60 0.67 0.03 −0.21 0.21
SRLZ2 0.52 0.66 0.06 −0.07 0.06
MRD 0.62 0.72 0.04 −0.18 0.18
SRN 0.26 −0.36 0.42 −0.25 0.09
RL 0.89 0.02 0.03 0.37 0.20
RD −0.34 −0.53 −0.09 −0.02 0.36
SRL −0.34 0.50 0.38 0.65 −0.14
RLI 0.59 −0.50 0.01 0.55 0.10
RTD 0.46 −0.22 −0.25 −0.69 −0.19
RM 0.93 − 0.18 − 0.15 0.05 0.25
SM 0.91 −0.24 0.20 − 0.02 − 0.08
TDM 0.95 −0.22 0.06 0.01 0.05
RMR −0.19 0.07 −0.70 0.14 0.54
SH 0.34 −0.22 0.73 −0.09 0.13
LN 0.59 −0.01 −0.35 0.14 −0.55
TN 0.71 −0.05 −0.39 0.23 −0.28
Variation proportion
Eigenvalue 6.28 2.57 1.75 1.57 1.08
Variance (%) 39.3 16.2 11.0 9.8 6.7
Cumulative variability (%) 39.3 55.5 66.5 76.3 83.0

Rotation converged in 25 iterations using Varimax with Kaiser normalization. For each trait, the large variable loading score crossing the six components appears in bold. Three principal components with eigenvalues > 1 were extracted and considered significant (Tabachnik and Fidell, 1996)