Murdannia aff. triquetra (Wall. ex C.B.Clarke) G. Brueckn ., Nat. Pfl.-Syst. (ed. 2) 15a: 173. 1930.

Pellegrini, Marco Octavio de Oliveira, Faden, Robert B. & Almeida, Rafael Felipe de, 2016, Taxonomic revision of Neotropical Murdannia Royle (Commelinaceae), PhytoKeys 74, pp. 35-78 : 61-65

8. Murdannia aff. triquetra (Wall. ex C.B.Clarke) G. Brueckn., Nat. Pfl.-Syst. (ed. 2) 15a: 173. 1930. Fig. 9

Phaeneilema triquetrum (Wall. ex C.B.Clarke) G. Brückn., Notizbl. Bot. Gart. Berlin-Dahlem 10: 56. 1927.

Aneilema triquetra Wall. ex C.B.Clarke, Monogr. Phan. 3: 208. 1881. Lectotype (designated by Ancy et al. 2015): BANGLADESH. India Orientalis, in Prov. Sylhet, fl., fr., s.dat., N. Wallich 5220 (B barcode B100367814!: isolectotypes: E barcode E00393352!, GDC barcode GDC00489348!; K n.v.).

Diagnosis.

Herbs ca. 10.0-20.0 cm tall, annual, without a definite base, rooted emergent in flooded fields. Roots thin, fibrous, medium to dark brown, densely to sparsely pilose with medium to dark brown hairs, emerging from the basalmost nodes. Rhizomes absent. Stems trailing, floating on water with ascending apex, succulent, densely branched at the base, glabrous or with minute eglandular hairs. Leaves spirally-alternate, evenly distributed along the stems; sheaths 0.7-1.0 cm long, glabrous; lamina 2.0-4.5 × 0.6 cm, narrowly lanceolate to lanceolate-oblong, glabrous membranous, slightly canaliculate, green on both sides, base rounded to amplexicaul, margins glabrous, sometimes undulate, apex acute to acuminate. Inflorescences 1-3, terminal or axillary in the distalmost (up to 4) nodes, fascicle-like, sessile, enclosed by the leaf-sheaths, composed of 1-2-(3) verticillate cincinni; peduncle absent; basal bract inconspicuous; cincinni bracts absent; cincinni 1-flowered, erect, straight, peduncle ca. 3.0 mm long, glabrous, internodes inconspicuous; bracteoles absent. Flowers male or bisexual, actinomorphic, barely exserted from the sheath; floral buds ellipsoid, light green; pedicels ca. 3 mm long, erect and elongate in fruit; sepals 4.0-5.5 mm long, linear-elliptic, cucullate, light green to pale pink, glabrous; petals equal, elliptic, slightly cucullate, white to pale lilac or pale pink, glabrous; androecium not determinable; ovary ellipsoid, tapering into the style, 3-locular, light green, smooth, glabrous, style straight, 1.7 mm long, glabrous, stigma capitate. Capsules 4.5-5.5 × 2.0-2.5 mm, ob longoid to ellipsoid, 3-locular, 3-valved, apiculate due to persistent style, light brown when mature, smooth, glabrous, locules 3-seeded (only 1 counted). Seeds (only 1 mature seed seen) transversely ellipsoid, ca. 1.5 × 0.9 mm, testa brown, with deep dorsal pits and longitudinal furrows, farinose only around the embryotega, appendage absent; embryotega lateral, inconspicuous, without a prominent apicule; hilum linear, less than ½ the length of the seed, borne on a ridge.

Specimen seen.

VENEZUELA. Tachira. Distr. Liberatador: 10 km S of El Piñal, 71°55'W, 7°27'N, alt. 250 m, 7 Nov. 1982, G. Davidse & A. C. González 21663 (US).

Distribution and habitat.

Known for certain only from this collection. The general habitat was recorded as "partially inundated forest remnant with slow stream and pools of standing water" and for this collection as "stems floating in pool of creek." A photograph of a plant from Colombia, which may or may not be the same species, was sent to the first author, but without a corroborating specimen, so it has not been considered for this description. However, we have illustrated it in Fig. 9 to encourage collectors to look for it.

The Murdannia keisak complex is widespread in Asia, ranging from India to China and Japan, growing in flooded grasslands and disturbed areas. In South America, it is known from only two collections, one from Venezuela and one from Colombia. Unfortunately, it seems that the specimen from Colombia went astray during shipping, since it was never received by the first author.

Phenology.

It was found in bloom and fruit in November.

Conservation status.

Following the IUCN recommendations ( IUCN 2001), this species should be considered Data Deficient. Correspondence by the second author with the collector Gerrit Davidse, indicated that this was not a disturbed habitat in which one would expect to find introduced weeds. However, the habitat was under great pressure and possibly no longer exists.

Nomenclatural notes.

Nandikar and Gurav (2015) designated the specimen at CAL (CAL0000025807) as the lectotype for Aneilema triquetrum . Nevertheless, after analyzing the specimen, comparing it to the protologue and to the remaining specimens, it became clear that the specimen at CAL is not conspecific to the specimens at B, E and GDC. Ancy et al. (2015), unaware of the article published just few months earlier by Nandikar and Gurav (2015), designate the specimen at B (B100367814) as the lectotype for Aneilema triquetrum . Their choice matches perfectly the protologue, and thus should be followed instead of the lectotypification made by Nandikar and Gurav (2015). Nonetheless, if ever found, the specimen at K would make a much better choice of a lectotype. At the time of the description of Aneilema triquetrum and the completion of his monograph (i.e. 1881), Clarke was working at K, and would had access to a possible specimen in the Wallich Herbarium, housed at Kew.

Discussion.

This is a widely distributed species complex, being very common and well collected in Asia. Nevertheless, the morphologic limits between Murdannia keisak and Murdannia triquetra , as well as the application of these names, varies greatly according to each author. In Flora of China ( Hong and DeFilipps 2000), both species are accepted, although somewhat tentatively, and are separated by the length of the sepals, shape and size of the capsule, and number and shape of the seeds. The authors also state that the morphologic differences seem to be associated with the geographic distribution of the taxa. Nevertheless, both descriptions overlap with the description presented by Faden (2000) for Murdannia keisak , in North America. Ancy (2014), in her unpublished Ph.D. thesis, presents a thorough taxonomic account on Murdannia from India. Her description of Murdannia triquetra matches very closely the two specimens known for South America, in sepal, petal and fruit morphology. Nonetheless, Ancy (2014) describes the filaments as being glabrous, contrary to the bearded filaments known for the South American specimens. The author also omits the description of the antherodes, which in the South American specimens are yellow and cordate. Nevertheless, some young flower buds dissected by the second author lacked hairs on the filaments of the stamens and completely lacked staminodes, but that might have been a developmental stage and thus may not be a discrepancy. This could be related to the extremely immature state of the dissected buds, and could explain the discrepancy of our description and the description presented by Ancy (2014). Nandikar and Gurav (2015) published a second account on the Indian species of Murdannia . In their treatment, Murdannia triquetra differs greatly from the South American specimens. However, it matches very closely the description presented by Hong and DeFilipps (2000), Faden (2000) and Chowdhury et al. (2015) for Murdannia keisak . In these descriptions, the antherodes are described as sagittate and ranging from lilac to purple, and clearly do not match the South American specimens.

It is the authors opinion that a study focusing on the specific boundaries between these taxa is necessary. Nevertheless, since this species complex is only invasive in the New World, we also believe that the required investigation should be carried out in the plants native range. It is also possible that these Neotropical collections represent a distinct taxon, not closely related to the other native South American species. But a much better South American sampling for comparison and a much more detailed would be required. Field work, better sampling of herbaria specimens, detailed study of reproductive morphology, analysis of the protologues, and population studies might shed a light on the issue.

Figures

Figure 9. Murdannia aff. triquetrum (Wall. ex C. B. Clarke) G. Brueckn., from Esteros de Arauca, Colombia. A Detail of a stem, showing an apical fruit B detail of an internode, showing a side view of a male flower. Photographs by M. Fernandez.