Fig. 3

Biplot of nine treatments and all studied traits. Principal Component Analysis (PCA) was then applied to the standardized dataset using the variance–covariance matrix as the basis for eigen-decomposition. PCA biplots, simultaneously displayed the positions of sample points (scores) and variable vectors (loadings) in the reduced-dimensional space defined by the first two principal components, thereby facilitating the identification of traits that contribute most to nine treatments differentiation. The environments E1, E2, E3, E4 are defined in Table 3, Grain yield: GRY, Leaf proline content: PRN, Soluble leaf sugars: CBD, Anthocyanin: ANT, Phenolic compound content: PNL, Root colonization percentage: CPR, Polyphenol oxidase: POX, Superoxide dismutase: SOD, Catalase: CAT, Ascorbate peroxidase: APX, Guaiacol peroxidase: GPX, OM: magnesium oxide treatments, MI: Mycorrhizal fungi treatments.