A revision of Dimeria ( Gramineae - Dimeriinae ) in malesia with a note on Cymbachne

In Malesia there are 4 species of Dimeria ( Gramineae ). A brief history of the genus with a key, nomenclature, descriptions, and notes is provided. Dimeria ciliata, D. dipteros, and D. fuscescens are reduced to D. chloridi­ formis , D. leptorhachis and its var. velutina to D. gracilis, D. monostachya to D. kurzii. Dimeria ornithopoda var. gra­ cillima is not recognised. Cymbachne is doubtfully referred to Ischaemum .

Although rather anomalous, it clearly belongs to the Andropo goneae Dumort.It is immediately distinct by the tenacious, espatheate, solitary or digitate racemes with solitary hermaphrodite spikelets lateral to the rachis, shortly pedicelled, laterally compressed, and florets with 2 anthers (there are some reports of 3, but this has not been seen in this study).The authorship of the subtribe Dimeriinae is sometimes given as C.E. Hubbard (1934), but he explicitly cited Hackel (1889).
It was first described from Australia with as the only species D. acinaciformis R.Br. which Brown placed between Imperata Cirillo and Ischaemum L. He also mentioned an unnamed species from India.The name alludes to the paired racemes (dimerous: δις μερις, 'two parts': 'spica duplex').Bor (1953: 554) has suggested that Brown had inspected an abnormal specimen with 3 anthers.
The genus was revised for India by Bor (1953) with 25 species and more recently Kiran Raj (2008) treated the subtribe with 2 genera and 43 taxa in an unpublished thesis.
For the Malesian taxa the diagnostic characters are but few; hence several are here reduced.
For the record it may be noted that Roberty (1960)

TAxonomIC poSITIon
The Dimeriinae generally have been regarded as a member of the Andropogoneae Dumort.with Dimeria as the only genus, until Nanooravia Kiran Raj & Sivad. (in Kiran Raj et al. 2013a, b) with one species from India was described.Affinities have been pointed out with e.g.Ischaemum L., Saccharum L., and Sorghum Moench.Clayton & Renvoize (1986) regarded it as derived from the Ischaeminae J.Presl by suppression of the sessile spikelet, as the remaining spikelets are shortly pedicelled.This might as well be a loss of the pedicelled ones, as was observed by Miquel (1851;'rudimentary pedicels'), but by no one else.In the Andropogoneae reduction of the pedicelled spikelet is the usual trend, while the 'sessile' spikelets sometimes may be shortly pedicelled.The reduction of the pedicelled spikelets usually is from male, neuter, to absent.Kellogg & Watson (1993: 296) in a 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 Saccharinae clade.Estep et al. (2014) in a nuclear molecular study, however, found Dimeria nested in a clade within Ischaemum, but with little basal support, so a reduction of Dimeria to Ischaemum seems premature.Morphologically the two are very dissimilar.Teerawatananon et al. (2011aTeerawatananon et al. ( , 2014) ) regarded the Dimeriinae as monophyletic and most related to Eulalia Kunth s.l. and Ischaemum.Kellogg (2015) actually reduced the Dimeriinae to Ischaeminae, but remarked "The subtribe lacks obvious morphological synapomorphies, though the lower glume often has a winged keel".The latter feature is highly variable in both genera.However, she also pointed out the presence of "leaf epidermis with intercostal cells with several papillae per cell".Whether this is indeed a synapomorphy remains to be seen.Bor (1953)  Distribution -Sri Lanka, Nepal, India (Andhra Pradesh, Assam, Bengal, Kerala, Manipur, Meghalaya, Nagaland, Orissa, Tamil Nadu) to Thailand (NE: Loei; E: Chayaphum; Pen.: Satun; therefore may be expected in N Malay Pen.), N Vietnam, Micronesia (Guam, Yap), Australia (N Territory, Queensland); Malesia: Philippines (Palawan), New Guinea: Irian Jaya (Baliem River); Papua New Guinea (Central, W-, S Highlands, Western Prov.).
Notes -The var.heteromorpha is merely a rather glabrous form with an extended wing on the upper glume.
An isotype in A of Dimeria dipteros is a small specimen, but with cataphylls and extra-vaginal branching, hence not an annual.There were hardly discernible wings on the glumes.Hence the species is reduced here.
The Continental Asian representatives are known as D. fusce scens Trin.( 1832) but I have failed to find any convincing differences.Roberty (1960: 402) cited his subvar.ciliata for Deccan (India) and the Philippines.The first surely is a misidentification.Reeder (1948: 325) and Jansen (1953: 267) noted that the peduncle would be "short pubescent below the inflorescence".I have not seen this.
The development of a wing on the keel of the upper glume is apparently not such a reliable character as some keys suggest.Merrill (1914) did not note the upper glume to be winged, Reeder said prominently in upper part.Actually it may be absent to broad, while the plants in other respects do not seem to differ.Although said to be common in e.g.New Guinea (Henty 1969: 80), only two specimens were received from K. Possible specimens in L could not be found.Distribution -Sri Lanka, W India (Goa, Karnataka, Maharasthra), Burma (Tenasserim), Thailand (Trat), Cambodia (Kampong Speu), Vietnam (Dalat), Malesia: Malay Pen.(Kedah).
Notes -The differences between D. leptorhachis and its var.velutina is only in the pubescence of the sheaths and leaves, which is insufficient to recognise these forms at any level.The differences with D. gracilis were equally unimportant.
Notes -Reeder compared his D. monostachya to D. sinen sis Rendle, a synonym of D. kurzii.The only difference outside the very disjunct distribution is that the awns are shorter than in D. kurzii.The isotype specimen in A does not have broadly winged upper glumes, as described.I have seen specimens with such short awns also from Vietnam (e.g.Schmid s.n. 3 Feb. 1954 Awn present or absent, (0-)4-6.5(-10)mm long, incl.1-3.5 mm long column.Anthers 0.25-0.8mm long.2n = 14, 32.Distribution -India to Japan, N Australia, Malesia: throughout.
Habitat -Sunny, infertile soil, grass fields, road sides, river banks, abandoned rice fields in the dry season, resistant to mowing and grazing, locally abundant, 0 -1800 m altitude.
Notes -This species is rather variable and Hackel (1889) optimistically and based on only a few specimens distinguished 6 varieties and 4 subvarieties, warning that intermediary forms occur.Bor (1953Bor ( , 1960) ) for India distinguished four.Jansen (1953) for Malesia provisionally distinguished three.Duistermaat (2005) regarded them as extremes of a range and I agree.
Specimens with spikelets 1-1.5 mm long and anthers up to 0.25 mm have been distinguished as var.gracillima Bor (1953: 576), and e.g.Teerawatananon et al. (2014: 143, lectotype!), but otherwise they cannot be distinguished.Some collections appear to be just dwarfed forms.
Forms without awns or mucronate ones (var.glabra) occur within the same population (see also Ohwi 1965, Duistermaat 2005), but I have not seen mixed collections.BS 336719 (Ra mos & Edaño) (SING) from Luzon had spikelets without awns or with short, simple, filiform ones in the same racemes.
Note - Roberty (1960: 398) recorded it as present in Malesia; the actual distribution is S India (Kerala, Tamil Nadu).

Cymbachne
Cymbachne Retz.( 1791) was described with as the single species C. ciliata represented by a single (!) specimen collected by König in Bengal.Retzius material may be expected to be in Lund (LD), but it has been long lost.It was not found by Fischer (1932), but as noted by Bor (1960: 139) a specimen is present in the Zetterstedt herbarium (LD 1228279).It was annotated not by Retzius, but by Zetterstedt (Messrs.P. Frodén and P. Lassen, LD, in litt.).Bor identified it as D. alata Hook.f., a Sri Lanka endemic (Fig. 1, 2).
-C10016737 with a label by Hackel 'Andropogon cymbachne Hack./ Cymbachne ciliaris Retz.' and a reference to Hofman Dory (??) (K neg. 19384).The 'Rottboellia' in pencil is to be neglected.A note by Clayton states "does not match description" (Fig. 4).These specimens belong to Andropogon canaliculatus Schumach.This is an African species ranging from Mali to Tanzania and Zimbabwe and never could have been collected by König.Obviously, they were mislabelled and very well could be isotypes of Schumacher's species described from Ghana, the type of which has otherwise not been found as yet.However, the type of A. eucnemis Trin.( 1832) may very well be an isotype of this, and is here designated as the neotype.
Note that later typifications do not make a name superfluous (ICN Art.52, Note 2).
Hackel for some reason with a query included Arthrostachys Desv.(1831) with A. gracilis.This was described without prove-However, this cannot be original material.From the reference to Sprengel (1824) on the label and the absence of any mention of König it would seem that after 1824 Zetterstedt tried to identify a collection, and besides the incongruence with the original description, this cannot be the type of C. ciliata.That it came from Bengal, as is noted on the label, might well have been a deduction from the provenance cited by Retzius and subsequent authors.
Bor wrote "the description might easily apply to a species of Dimeria … no need for a change of name".However, the description mentions the presence of paired spikelets, whereby it can never refer to Dimeria R.Br. ( 1810), which is one of the few andropogonoid genera characterised by solitary ones.Dimeria spikelets have only 2 stamens, 3 are described here, and depicted by Presl (1830; but described as two!).This is most fortunate, as Cymbachne has priority over Dimeria, and there is now no need for a proposal for conservation to prevent about 70 new combinations.
Probably following Bor, Clayton & Renvoize (1986: 376) have suggested that it was based on a damaged specimen of Dimeria R.Br. and thus it is perpetuated in the literature.
Willdenow (1797) unseen and without argumentation transferred it to Rottboellia L.f. as R. cymbachne, a superfluous name, as at the time there was no R. ciliata.There is no duplicate in the Willdenow herbarium (B).This combination, without any further reference, was used by Steudel (1854: 362).
Hackel (1889: 450) stated to have seen a specimen collected by König in Copenhagen (C), labelled as Cymbachne ciliata.He observed (my translation) that "the description is quite obscure and by no means without errors, for what he calls the female flowers seem to be nothing else than thick pedicels, remaining after de male spikelets have fallen off.Besides what else Retzius adds, agree well with the Koenig specimen and I do not doubt that I have described the same as Retzius' plant.A specimen in the herbarium Retzius, now in Lund, is to be wished for".
In C there are two (!) sheets, which makes them suspect, because a single sheet was expected in Lund: nance or collector and is immediately distinct by the articulate inflorescence axes.It was not mentioned by Steudel (1854).Clayton & Renvoize (1986: 349) and Soreng & Pennington (2003) regarded it as a synonym of Andropogon with which I can agree.Arthrostachya Link (1827) is not an earlier homonym (Art.53.3.Ex.12).