Briza and other Briza-like genera, and Parafestuca {=Koeleria) and Torreyochloa have Aveneae-type plastids (Soreng and Davis 2000; Soreng et al. 2007) and are placed in other subtribes. Plastid DNA evidence from the single species of the Australian genus Dryopoa Vickery (Vickery 1963) places it close to Festuca (Gillespie and Soreng 2005); and Soreng et al. 2007) suggested that it does not belong close to Poinae but to Loliinae (as postulated by Clayton and Renvoize in testing the monophyly of Poa found support for the subtribe Puccinelliinae as the sister group of a clade Poinae, Alopecurinae and Miliinae (PAM).Plastid data supported the monophyly of Poa (Gillespie and Soreng 2005; Gillespie et al. 2007), with some modifications.
Plastid data confirmed that it belongs in Poa, and it is now treated as part of Poa (Gillespie and Soreng 2005; Gillespie et al. 2007; Soreng and Gillespie 2007). Representatives of 21 genera accepted or generally separated from Poa and 67 species of Poinae (sensu Soreng et al. 2007) were sampled (Tables 1 and 4); this includes all currently (and many previously) recognized Poinae genera with the exception of four rarely collected monotypic genera from the eastern Mediterranean region for which no material was available {Lindbergella, Libyella, Nephelochloa and Oreopoa). Sampling in the largest genus Poa (40 species sampled) included species belonging to all major sections and previously identified plastid DNA clades (Table 4) (Gillespie et al. 2007), focusing on broad phylogenetic relationships.
Species and assemblages with length variation in a mononucleotide repeat include Poa alpina L. {Gillespie et al. 6749. The tribe Poeae has undergone numerous changes in circumscription and intra-tribal classification in recent years (Soreng et al. 2007). Using nuclear ribosomal ITS sequence data (here, and Quintanar et al. 2007), nuclear subtribes of Aveneae are nested between elements of Poeae's.
These results were consistent with those of Quintanar et al. 2007) in their study focusing on elements previously separated in the tribe Aveneae. Based on our new ITS and TLF data. and Quintanar et al. 2007) Cinna clearly belongs to elements of Poinae (see subsection on subtribe Cinninae below). In contrast, in this analysis of plastid DNA and in the analyzes of Sorengo et al. 2007), the Sesleriinae are separated in the Poeae-type plastid lineage on a branch with the subtribe Loliinae.
Our ITS result supports the refinement of the original concept of Puccinelliinae (Soreng et al. 2003úí) to include Hyalopoa (probably placed in Poinae by Soreng et al. 2007) and exclude Colpodium s. Zingeria was resolved in a well-supported polytomy with the genera Puccinelliinae in the plastid analysis of Döring et al. 2007).
However, further research is needed because Döring et al. 2007) resolved a second species of Hyalopoa with Poa pratensis L. In the most recent definition, Miliinae included only Milium, Colpodium and Zingeria (Soreng et al. 2007). At this point it remains plausible that the subgroup Miliinae sensu Soreng et al. 2007) is monophyletic, but far from certain.
Die plastiedanalise van Döring et al. 2007) supports Zingeria in a polytomy with Puccinelliinae genera, en Milium onder Poinae en Alopecurinae element.
ÏT'
It is difficult enough to imagine how such a complex series of character state transitions evolved from states generally considered plisiomorphic in PAM (Soreng et al. 2007) or PPAM, or Poeae s. In the plastid analysis of Soreng et al. 2007), there was strong support for the separation of Phleum, Alopecurus and Beckmannia among elements of Poinae. A Bayesian analysis of ITS data (Quintanar et al. 2007) also rejected the association of Phleum and Alopecurus.
In the meantime, we propose to continue to retain Alopecurinae in the current sentence (Soreng et al. 2007), pending additional sampling and stronger data that contradict this morphologically well-defined group. In Flora Europaea (Tutin et al. 1980) Arctagrostis and Cinna are placed under genera of Poeae s. Cinna, Cyathopus and Limnodea have spikelets that disarticulate below the glume, an apomorphic condition in PPAM (Soreng et al. 2007).
Aniselytron was included in Calamagrostis within Aveneae by Clayton and Renvoize (1986), but they noted that it was aberrant {Calamagrostis belongs to the Aveneae core clade in TLF and ITS analyses; Quintanar et al. 2007). In Rodionov et al. 2005) ITS analysis revealed a cluster of Alopecurus, Arctagrostis, Arctophila, Arctopoa (as Poa schischkinii), Dupontia and Phleum, but Quintanar et al. 2007) analysis included too few strictly Poinae elements to assess groupings. The systematics and origin of the putative hybrid genus Dupontia was studied in depth by Byrsting et al.
At the same time as Soreng et al. 2007) presented new plastid data for Gaudiniopsis, treating it as sister to Apera and Bellardiochloa, with strong support for bootstrapping. Recent evidence from ITS and trnL-trnF suggests that they were misplaced in Austrofestuca and should be recognized as separate genera (Hunter et al. 2004). The ITS and TLF data presented here and the plastid study by Sorengo et al. 2007) strongly support their assumption that the first two genera are closely related to each other and distantly related to Austrofestuca.
Hookerochloa and Festucella appear to be closely related to Nicoraepoa and Arctagrostis in our analyzes and those of Sorenga et al. Taxonomic aspects and a detailed discussion of morphology are the focus of a separate paper (Jacobs et al. 2008), where Festucella is subsumed into Hookerochloa. Within Poa, we provide new or additional data to include more genera; see Table 3 for subgeneric classification of Poa and clade acronyms published by Gillespie et al.
Pseudopoa) and Parodiochloa in Poa, based on plastid DNA data and their recent recognition as sections within Poa (Gillespie and Soreng 2005; Gillespie et al. 2007), is now confirmed by ITS data. Most elements of the revised subgeneric classification system, based on morphology and plastid DNA data, presented in Gillespie et al.
