A REVISED INFRAGENERIC CLASSIFICATION OF DIMERIA R . BR . ( POACEAE : ANDROPOGONEAE )

The four sections of the little known genus Dimeria R. Br. of the rather anomalous paleotropical subtribe Dimeriinae Hack. (Poaceae–Panicoideae– Andropogoneae) are revised. A key is provided. Three Peninsular Indian species, viz. Dimeria sivarajanii, D. idukkiensis and D. borii are treated here as subspecies of D. bialata, D. kurumthotticalana and D. mooneyi respectively; and five, viz. D. chelariensis, D. copei, D. eradii, D. jayachandranii and D. kollimalayana are reduced to synonymy.

Hackel (1889) had 12 species, 2 subspecies, and 10 varieties and regarded it as close to the Sacchareae Dumort., and very much to the Tristegineae Nees.For a discussion on the latter tribe see Veldkamp (2015).48 KIRAN RAJ et al. Roberty (1960) found the genus Dimeria so homogeneous, that in his "cohors" Dimeriastreae he accepted only a single species, D. avenacea (Retz.)C.E.C. Fisch., with not less than 22 subvarieties all with invalid names, because he did not follow the Linnean infraspecific classification.Clayton and Renvoize (1986) regarded it as derived from the Ischaeminae J. Presl by suppression of the sessile spikelet.This might as well be loss of the pedicelled ones, as was observed by Miquel (1851: "rudimentary pedicels"), but by no one else.Kellogg and Watson (1993) in their phylogenetic analysis based on morphological data treated Dimeria as a sister group of Cleistachne Benth. of the subtribe Sorghinae Bluff et al. with both genera nesting in a clade.Estep et al. (2014) in a nuclear molecular study found Dimeria nested in a clade within Ischaemum, but with little basal support, so a reduction of Dimeria to Ischaemum seems premature.

Sectional classification of Dimeria by Bor (1953)
The first infrageneric classification of the genus was by Bor (1953), who treated the species for India, Sri Lanka (Ceylon), and Myanmar (Burma) as belonging to three sections viz.Dimeria sect.Annulares Bor, sect.Capillares Bor, and sect.Loriformes Bor and the sections were recognized based on rachis and pedicel characters.As the type species, D. acinaciformis R. Br., is from Australia, he did not mention a section Dimeria in his treatise.Bor (1953) pointed out the necessity of a further detailed study of more specimens of all species for a better understanding of diversity and extent of intraspecific variation.After 1953, a fairly large number of new species have been described, and it was found after field work and morphological examinations in the herbarium that some could not be properly assigned to a section.Also, the Capillares and Loriformes contained species with strictly triquetrous racemerachises, overlapping spikelets, and pedicels closely appressed to the rachis.Dimeria acinaciformis is characterized by the presence of triquetrous raceme-rachises and compressed pedicels.
Considering all the above aspects, a revised infrageneric classification of Dimeria is proposed here.
The raceme structure of the representative taxa of the sections is illustrated in Fig. 2 as an aid for easy understanding of the diagnostic characters.I. Dimeria R. Br. sect.Dimeria Type: Dimeria acinaciformis R. Br.Annuals or perennials.Racemes 2 or 3, rarely 1; rachis of raceme triquetrous, trigonous in cross section, occasionally zigzag, usually wingless, if winged, only at the internodes; spikelets closely packed on the rachis, usually overlapping; raceme internodes c. 0.5 mm long; glumes slightly diverging at anthesis; pedicels 0.3−0.5 mm long, trigonous to flat, closely appressed to the rachis.Bor (1953)  Type: Dimeria woodrowii Stapf Annuals.Racemes 2 or 3, peduncle bent downwards or erect at maturity; rachis of raceme compressed, trigonous on one side and convex on the other, straight when young and curved at maturity to form a 'globule', or a single or double 'ringlet' carrying the spikelets along the inner side; raceme internodes up to 1 mm long; spikelets distantly arranged along the rachis; upper glume distinctly winged, or minutely winged, or wingless; pedicels terete, not compressed.

Distribution
Distribution: Two species in Peninsular India (Table 1), and so far known only from lateritic plains of the Northern Western Ghats.
Distribution: Restricted to Indian subcontinent (Western Ghats region of Peninsular India, Sri Lanka and Myanmar); three species in Peninsular India (Table 1).
Notes: Bor (1953) included seven species in Dimeria.sect.Capillares, and the type was not designated and hence the present lectotypification.In the present classification, three species are transferred to sect.Dimeria (Table 1).
Distribution: Peninsular India, Myanmar and Sri Lanka; mostly occurring in Peninsular India with 17 species.