Top left, Bayesian principal component analysis (bPCA) with missing value estimation on samples with less than 50% of missing values, resulting to a dataset with less than 25% of NA (n = 149). Two principal components were imposed, with a R2 (the importance of component) of 0.337 for the first (horizontal) axis and 0.135 for the second (vertical) axis. Top right, discriminant analysis (DA) on samples without NA values (n = 30), with two of the three axes obtained representing inertia of 0.988 and 0.764 (and the third, not represented here, 0.473); Bottom, boxplots for morphometrical values per taxa. LmsL, length of median stem leaf (cm); Lls, length of leaf sheath (cm); L1bls, length of first basal leaf segment (cm); tLblb, total length of basal leaf blade (cm); cL, carpophore length (cm); cD, length of carpophore division (cm); Ns, number of segments within the last-order division on the basal leaf; P, number of subterminal umbels; Nr, number of umbel rays; Nb, number of involucellar bracteoles. Dark green dots represent samples of C. hirsutum var. calabricum considered as C. hirsutum var. hirsutum samples in the present study.

 
 
  Part of: Reinhart T, Guillon L, Begoc T, Chapotin P, Reduron J-P, Espinosa Prieto A, Hardion L (2024) An integrative taxonomic revision of the Chaerophyllum hirsutum complex (Apiaceae) using morphological and molecular markers. Plant Ecology and Evolution 157(3): 399-406. https://doi.org/10.5091/plecevo.124907