A phylogenetic analysis of Bromus (Poaceae: Pooideae: Bromeae) based on nuclear ribosomal and plastid data, with a focus on Bromus sect. Bromus.

To investigate phylogenetic relationships among and within major lineages of Bromus, with focus on Bromus sect. Bromus, we analyzed DNA sequences from two nuclear ribosomal (ITS, ETS) and two plastid (rpl32-trnLUAG , matK) regions. We sampled 103 ingroup accessions representing 26 taxa of B. section Bromus and 15 species of other Bromus sections. Our analyses confirm the monophyly of Bromus s.l. and identify incongruence between nuclear ribosomal and plastid data partitions for relationships within and among major Bromus lineages. Results support classification of B. pumilio and B. gracillimus within B. sect. Boissiera and B. sect. Nevskiella, respectively. These species are sister groups and are closely related to B. densus (B. sect. Mexibromus) in nrDNA trees and Bromus sect. Ceratochloa in plastid trees. Bromus sect. Bromopsis is paraphyletic. In nrDNA trees, species of Bromus sects. Bromopsis, Ceratochloa, Neobromus, and Genea plus B. rechingeri of B. sect. Bromus form a clade, in which B. tomentellus is sister to a B. sect. Genea-B. rechingeri clade. In the plastid trees, by contrast, B. sect. Bromopsis species except B. tomentosus form a clade, and B. tomentosus is sister to a clade comprising B. sect. Bromus and B. sect. Genea species. Affinities of B. gedrosianus, B. pulchellus, and B. rechingeri (members of the B. pectinatus complex), as well as B. oxyodon and B. sewerzowii, are discordant between nrDNA and plastid trees. We infer these species may have obtained their plastomes via chloroplast capture from species of B. sect. Bromus and B. sect. Genea. Within B. sect. Bromus, B. alopecuros subsp. caroli-henrici, a clade comprising B. hordeaceus and B. interruptus, and B. scoparius are successive sister groups to the rest of the section in the nrDNA phylogeny. Most relationships among the remaining species of B. sect. Bromus are unresolved in the nrDNA and plastid trees. Given these results, we infer that most B. sect. Bromus species likely diversified relatively recently. None of the subdivisional taxa proposed for Bromus sect. Bromus over the last century correspond to natural groups identified in our phylogenetic analyses except for a group including B. hordeaceus and B. interruptus.