A revision of Mnesithea ( Gramineae – Rottboelliinae ) in Malesia and Thailand

You are free to share to copy, distribute and transmit the work, under the following conditions: Attribution: You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work). Non-commercial: You may not use this work for commercial purposes. No derivative works: You may not alter, transform, or build upon this work. For any reuse or distribution, you must make clear to others the license terms of this work, which can be found at http://creativecommons.org/licenses/by-nc-nd/3.0/legalcode. Any of the above conditions can be waived if you get permission from the copyright holder. Nothing in this license impairs or restricts the author’s moral rights. Blumea 58, 2013: 277–292 www.ingentaconnect.com/content/nhn/blumea http://dx.doi.org/10.3767/000651913X678257 RESEARCH ARTICLE


INTRODUCTION
Back in the 1980s, initial studies in the Rottboelliinae J.Presl (Gramineae) were started during courses in Plant Taxonomy at the Rijksherbarium, now Naturalis Biodiversity Center, section Botany.These led to papers on the generic delimitations in the subtribe (Veldkamp et al. 1986) and revisions of Thaumastochloa C.E.Hubb.and Heteropholis C.E.Hubb.(De Koning et al. 1983).The various analyses indicated that the generic differences between Coelorachis Brongn., Hackelochloa Kuntze, Heteropholis, Mnesithea Kunth, Ratzeburgia Kunth and Rottboellia formosa R.Br. were artificial and a much expanded Mnesithea, the oldest generic name, was proposed (Veldkamp et al. 1986).This then now contains about 33 species worldwide, of which 11 occur in Malesia and Thailand.Veldkamp et al. (1986) provided an extensive discussion on the morphology of the spikelets.In Mnesithea the spikelets are in pairs or triads: a sessile and a pedicelled spikelet or two sessile and one pedicelled ones.The latter condition seems exceptional in the Andropogoneae (e.g. also in Polytrias Hack.), but as inflorescences in some species are mixed in various degrees it is not an apomorphic generic character for this alliance.
The sexuality of the lower floret of the sessile spikelet turned out to be of importance in the generic delimitation which was a rather surprising observation.
In the Panicoideae -Andropogoneae to which Mnesithea belongs the spikelets are 2-flowered, with a lower and an upper floret.The lower one shows a range in reduction running from bisexual (the presumed plesiomorphic state, but now rare) to male, to sterile, to loss of the lower palea and exceptionally to the total absence of a lower floret (e.g. in Microstegium Nees).In general, the various stages of reductions can be observed within a genus, within a single species, or sometimes even in spikelets in the same inflorescence; each depending on the taxonomic circumstances.As with all evaluations of taxonomic characters this must be viewed at an ad hoc basis and that differences in reduction might be of generic significance in some place may not a priori be ruled out.
And so it was concluded that the only remaining difference between a much expanded Mnesithea and reduced Rottboellia L.f. was "the presence of a sterile lower floret in the sessile spikelet of the first genus against a male floret in the second one" (Veldkamp et. al. 1986: 289, 297, 298, (key 4b), 302, 305).
In view of the cline in reduction elsewhere, this conclusion appeared too good to be true, but during the present study when more material could be studied in depth it was observed again that it is constant and the only character that seems to separate Mnesithea from Rottboellia.
Many of these are drought indicators.As research in Malesian grasses continues more examples are expected, especially among the Andropogoneae, which are prominent in dry areas.
These examples hint at the presence of an ancient drought refuge in the Banjarmasin area as was also suggested by Van Steenis (1979).No doubt there will be more examples in other families.
Many species of the Rottboelliineae and the apparently unrelated Chionachninae Clayton have a remnant of the vascular bundle at the base of each joint.Beumée (1927) has observed that ants readily carried away the joints of M. mollicoma ('M.pubescens') and Polytoca macrophylla Benth.after they had inspected the knobs, thus assisting in the dispersal of the diaspores.Apparently they were attracted by the presence of cells that contain large drops of oil.He therefore regarded this knob as an elaiosome, a term we have adopted here for this structure.A considerable amount of oil was also present in the elaiosomes of Sclerachne punctata R.Br., a variable amount possibly due to the state of maturity in those of Polytoca bracteata R.Br., and only some oil in those of M. glandulosa, M. laevis, Ophiuros exaltatus (L.) Kuntze, Rottboellia cochinchinensis (Lour.)Clayton (R. exaltata L.f.).This may be present in other species as well and would be an interesting subject for study.Collect both the ant and the grass!
Distribution -Pantropical with about 32 species, of which 10 species and 1 variety in Malesia and Thailand.

KEY TO THE TAXA IN MALESIA AND THAILAND
NB: Unless otherwise stated, spikelet parts refer to those of the sessile spikelet.Habitat -Open grassy pine, dry dipterocarp, Tectona forests, on sand, roadsides, 120 -1400 m altitude.
Rottboellia cancellata and derivatives were based on the collection in SING, while R. foveolata etc. was based on the isotype in K and thus the names are not strictly homotypic, but Art. 9 Note 2 makes them superfluous.
Most similar is M. glandulosa:  Notes -For Queensland Domin (1915) distinguished between a var.typica with a forma subglabra and a var.pilosissima based on the density of the pubescence of the plant.This turned out to be a very variable feature and these taxa cannot be maintained.

Mnesithea geminata
Notes -The above account is based on four collections only, all in K: Ridley 11,     Banda, Ceram, Halmaheira, Nusa Laut), New Guinea: Aru, Waigeo.The record for Australia (e.g.Merrill 1921) probably is based on a Von Mueller collection cited as R. muricata by Bentham (1878), type of R. ophiuroides Benth.var.commutata Hack., a synonym of R. rottboellioides, q.v.
Uses -Yields a considerable amount of leaf, but the old ones and the culms are very hard.According to analyses the nutritional value is also low.
Notes -Rottboellia mutica was mentioned in a list of names with various literature references, e.g. to Sprengel's Systema vegetabilium, but this one without any, suggesting that it was a new species.It is stated to grow in sugarcane fields.Fernandez-Villar (1882: 314) 'corrected' it to Rottboellia muricata, a name often misapplied to Mnesithea glandulosa.The 'true' R. muricata (Eremochloa muricata (Retz.)Hack.) does not occur in the Philippines, while M. glandulosa has not been reported for sugarcane fields (f.Backer 1928).The identity of R. mutica can therefore not be established.Buse (1854) distinguished a form from Java with the pedicelled spikelets reduced to two glumes which was formally named var.javanica Buse by Miquel (1857), who cited besides the Junghuhn specimen referred to by Buse also Zollinger ( 352), the type.
The var. bandanensis would be distinct by the presence of a stronger indument on the lower glume of the sessile spikelets.Similar hairy glumes were observed in Jacobs 5645 (Brunei) (but also some nearly glabrous), Alston 13411 (W Kalimantan), SAN 33502 (Meijer) (Sabah), Buwalda 6032 (Ceram) and Van Royen 5558 (New Guinea, Vogelkop).In view of the fact that within the material there seems to be a more or less continuous cline from subglabrous to pubescent glumes, being sometimes variable within a single inflorescence (see Jacobs 5645 mentioned above) and that such specimens occur within at least a considerable part of the range, it must be concluded that this variety should not be maintained.
In the Philippines the species is restricted to the southern and central Philippines and has not been found in the north-east, except for a single collection from Benguet, Luzon.
The species is easily recognized by the small appendages along the edges of the lower glume of the sessile spikelet.These are also sometimes found in M. helferi, which otherwise differs only slightly.See the key.
Habitat -In open place at mountain ridge, open and fireinduced grassland, weed on sandy soil, near the road side, partly shaded areas among bamboo, somewhat disturbed place in the deciduous dipterocarp oak forest, common in evergreen forest, swampy ground, granite bedrock, 0 -1450 m altitude.

. . H. porifera
However, when viewed over the whole range these characters appear to break down (Veldkamp et al. 1986: 294).
Differs from all congeners by the suborbicular, cancellate lower glume of the 1.5 -2.6 mm long sessile spikelet and the 2.2-4 mm long pedicelled spikelet.
Distribution -Pakistan to Sri Lanka, Burma, Thailand (N: Chiang Mai, Sukhothai, Nakhon Sawan, Kamphaeng Phet; E: Chaiyaphum, Nakhon Ratchasima, Buri Ram; SE: Chon Buri, Kanchanaburi; Peninsular: Songkhla), Laos (Champasak, Saravan), Malesia: Java (Ceribon, Surabaya, Besuki), Madura, Celebes (Rapang), Bali.The variety has been recorded for Laos, Cambodia and S Vietnam, but material from these countries as far as we have seen it all belonged to var.cochinchinensis.The genus was not recorded for Afghanistan by Bor (1970) although it is a country cited for this species and where it possibly may occur.
Habitat -On rather open grassy slope, dry savannah forest, sunny places, in a dipterocarp-savannah area, along the road, stony soil, up to 1500 m altitude.
Uses -Readily eaten by cattle, at least when young.Little valued by the Javanese because of its relative little yield of rather tough leaves mixed with tough culms.Analyses showed a very little nutritional value with an excess of fibre (Backer 1950: 177, sub M. laevis).
Notes -The species is distinct by the smooth, unappendaged lower glume of the sessile spikelets, the adnate pedicel, and the much reduced, small pedicelled spikelets.Jansen (1953) described a var.hirta on a Celebes collection which notably differed from the generally glabrous Java material by its rather dense bulbous-based bristly indument of the blades and peduncles.Specimens variable in this feature have been seen from Sri Lanka (Clayton 5845, Davidse 7598, Lazarides 7262), while others from Thailand (Van Beusekom & Phengkhlai 1192, Maxwell 75-433) were intermediary.It therefore seems to us that a taxonomic distinction for such hairy specimens cannot be maintained.
According to the general literature M. laevis would have spikelets in triads.In fact this is not always so.Within the same spike both triads and paired spikelets occur, with a general tendency for the triads to be present at its base and the pairs more upwards.Sometimes the most basal spikelets are paired, followed by triads, and distally again paired.This basal grouplet usually persists after the spike has broken up, and, when consisting of paired spikelets, may cause confusion in identification.In poorly developed specimens the triads may be very few to even totally absent (cf.Kievits 1543, see the discussion by Veldkamp et al. (1986) 284).The distinction against the var.cochinchinensis then becomes very difficult.The var. laevis seems to be distinct by being more robust (60-170 cm tall against up to 50 cm in var.cochinchinensis), while it grows in less disturbed places.

mm b a
The general distribution is also different (see map in Veldkamp et al. 1986: 285).
In Van Steenis 6648 and 7532 from Ceribon the terminal joints have one sessile and two pedicelled spikelets, the more 'normal' type of triad often found in the Andropogoneae.The first collection, moreover, has doubly pedicelled spikelets, i.e. there is an additional internode between the pedicel and the spikelet, which then reaches to 0.5 -0.67th of the next higher joint fitting into a shallow excavation there.
Notes -The records for India and Java cited and depicted by De Koning et al. (1983) refer to var.laevis (Veldkamp et al. 1986).
Differs from the Malesian / Thai congeners in: Nodes glabrous.Margins of the blades at base pectinate.Lower glume of sessile spikelet 7-nerved.Pedicelled spikelets reduced to 1 small scale.
Uses - Backer (1950: 176, sub M. pubescens) noted that a large amount of soft leaf was produced which was well-liked by cattle.Analyses showed a very satisfactory nutrition value.The yield was much lower than that of Panicum maximum Jacq., however, and did not increase after manuring.
Notes -The type of Rottboellia mollicoma is represented in K by two sheets.One has no original label, but is annotated '1862' and bears a note by C.B. Clarke "This is the type".Hance said he had collected the species first in October 1861.The other sheet has an original label which bears the date August 1869.Roberty with an exclamation mark following Clarke's selection designated the G duplicate as the lectotype ('type').However, if a specimen with the date October 1861 does exist that should take precedence as the lectotype.
Mnesithea mollicoma differs mainly from M. cancellata by the pubescence of the spikelet and the shorter pedicelled spikelets.
Msesithea mollicoma is usually densely hairy, but some exceptions were observed: Backer s.n.(1918) from Java has the lower glumes of the sessile spikelets almost glabrous to nearly so with hardly excurrent nerves.This is the only specimen seen from Java.Beumée (1927) and Backer (1950) have suggested that it was imported, e.g. with rubber seeds from Malaya; in 1918 it had locally become quite common in Imperata-fields.More recent collections are lacking.
Ledesma s.n., the type of Rottboellia triflora, has nearly glabrous culms and blades, but the nodes are hairy.The upper glume is flatter than usually seen, slightly keeled with ciliolate margins.The dimensions of the width of the blade, the length of the glumes, lemmas, paleas are all rather small.This form seems to be linked to the more 'normal' representatives of M. mollicoma by Smitinand 1827 from Thailand which has nearly glabrous culms, nodes and blades.Moreover, the leaves of this collection are exceptionally narrow (4 -7 mm) and the plant is only 40 cm tall.Because of this transitional form, albeit with quite a different provenance, Rottboellia triflora has been included in this species.More material from Palawan is required to see how variable that population is and whether or not it should have a separate status after all.
Notes -The species is easily recognized by the smooth lower glume of the sessile spikelets apically with distinct auricles.
The var. commutata would be distinct by bulbous-based bristles on the lower glume of the sessile spikelets.This feature is variable within a single spike and therefore there seems to be no other reason to maintain this taxon.
A form described as var.vestita by Domin from Australia and as var.hirsuta by Jansen from New Guinea would be distinct by the presence of densely hairy sheaths (especially the upper ones) and blades (e.g.Carr 11134, lectotype of the second), but Hoogland 4690 and NGF 22056 (Gillison) have glabrous sheaths and hairy blades.Many specimens have blades that are pubescent on the upper surface only, and several have completely glabrous blades.It seems therefore impossible to maintain this variety.
In the Philippine var.intermedia the lower glume of the sessile spikelet is only shortly winged above the middle.The expression of this feature, however, is quite variable, and insufficient to base a taxon on.
Most similar is M. glandulosa:     Distribution -Burma (Mergui) to S Vietnam and S China (S Yunnan); Thailand (N: Chiang Mai, Chiang Rai, Lamphun; NE: Khonkaen; SW: Kanchanaburi; Peninsular: Phangnga).Mentioned for Singapore by Wallich and so cited by others, but Wallich 8876 is M. mollicoma.It is erroneously cited as the type of M. striata by Roberty (1960).Noltie (2000: 829, t. 62k) reported the species for 'Terai', an area along the S border of Bhutan with Assam.
Habitat -Sunny places, old clearings, in the area between the deciduous dipterocarp oak-forest and a bamboo hardwood zone, moist savannah on clay, dry evergreen or mixed (Quercus) forest, along the road, granite bedrock, 25-1325 m altitude, locally common.
Notes -The species is best recognized by the glabrous lower glume of the sessile spikelets without appendages and  e with lines of small slits (which, however, are nearly absent in var.laevis Bor).For differences with M. helferi, see there.
The reasons for the lectotypification of M. striata and the reduction of the varieties has been discussed by Veldkamp et al. (1986: 292).
Part of Helfer 457 (syntype of M. merguensis) has pedicelled spikelets on two-jointed pedicels.One joint had four sessile and two pedicelled spikelets.Some other joints had two sessile and one pedicelled spikelets.The lower glumes have some pusticules with hairs.
Two collections from Thailand (Van Beusekom et al. 3597 (L,P) and Iwatsuki et al.T-10929 (L) have the pedicel fused to the joint in the lower part to completely so, while in the other specimens seen they were usually completely free.
Geesink & Santisuk 5004 (L, P) has lower glumes with pusticules on the keels.Tran Van Hing 196 has some joints with three sessile and two pedicelled spikelets.This specimen has hairy culms, sheaths and blades.It is the only collection seen from Vietnam, all other references seem to refer to M. cancellata (q.v.).
Habitat -Common in rice fields, c. 100 m altitude.
Notes -Only known from the type collection.A request to BK (24 Jan. 2012) for additional data or a loan remained unanswered.
The description above had to be summarised from the publication and is not congruent with the other ones given here.A later visit by the authors to the area showed that it had been converted to industrial development.
Differs from the Malesian / Thai congeners in: Ligule 0.5 mm long.Joints 2-2.5 mm long, hairy in the lower half.Upper glume of sessile spikelet hairy along the keel.