Table 1.
Characteristics of individual datasets: alignment length, the number and percentage of constant, variable and potentially parsimony-informative (Pi) sites, the number of coded indel(s), and the best-fit nucleotide substitution model determined by BIC.
Table 2.
P values of pairwise ILD tests for the nuclear ETS, ITS, PGDH, TRPT and SQD1 datasets.
Fig 1.
Bayesian tree based on the concatenated nuclear ITS, ETS, PGDH, TRPT and SQD1 sequences.
The Bayesian posterior probabilities are above the branches, and the maximum parsimony (left) and maximum likelihood (right) bootstrap support values are below the branches. Dashes indicate branches that are not found in the maximum parsimony tree or the maximum likelihood tree.
Table 3.
P values of pairwise ILD tests for the plastid psbA-trnH, trnC-petN, trnL-trnF, trnS-trnG and petN-psbM datasets.
Fig 2.
Bayesian tree based on the concatenated plastid psbA-trnH, trnC-petN, trnL-trnF, trnS-trnG, and petN-psbM sequences.
The Bayesian posterior probabilities are above the branches, and the maximum parsimony (left) and maximum likelihood (right) bootstrap support values are below the branches. A dash indicates a branch that is not found in the maximum parsimony tree.
Fig 3.
Summary of the incongruent position of the Hedysarum s.s. clade in the nuclear (left) and the plastid (right) trees, with selected legume species from each clade (middle).
The legumes: a, Eversmannia subspinosa B.Fedtsch.; b, Onobrychis atropatana Boiss.; c, O. viciifolia Scop.; d, Corethrodendron multijugum (Maxim.) B.H.Choi & H.Ohashi; e, Greuteria membranacea (Coss. & Balansa) Amirahm. & Kaz.Osaloo; f, Hedysarum semenowii Regel & Herder; g, H. americanum (Michx.) Britton; h, H. kumaonense Benth. ex Baker; i, H. tibeticum (Benth.) B.H.Choi & H.Ohashi; j, H. gremiale Rollins; k, H. petrovii Yakovlev; l, H. boveanum Bunge ex Basiner; m, H. callithrix Bunge ex Boiss.; n, Taverniera diffusa (Cambess.) Thulin; o, T. glauca Edgew.; p, T. lappacea DC.; q, T. longisetosa Thulin; photoed by PLL and LD.