Resurrection of the genus Haplanthus ( Acanthaceae : Andrographinae )

A systematic morphological study of Andrographis (Acanthaceae: Andrographinae) in India has revealed that the genus Haplanthus is distinct from Andrographis. We resurrect the genus Haplanthus here with four species, one of which contains three varieties. Five new combinations are proposed: H. laxiflorus, H. laxiflorus var. parishii, H. laxiflorus var. recedens, H. ovatus, and H. rosulatus. In addition, the following four names are lectotypified here: Gymnostachyum andrographioides, G. ovatum, G. parishii, and Haplanthus tener var. elongatus.


INTRODUCTION
The genus Haplanthus Nees was established with a single species, H. tener Nees and distinguished from Andrographis Wall.ex Nees by having the corolla subactinomorphous, five-lobed with a curved tube, and monothecous anthers that are hairy throughout the connectives (Nees 1832).Subsequently, Nees (1847) transferred two more species of Justicia to Haplanthus: H. tentaculatus (L.) Nees and H. verticillaris (Roxb.)Nees.Anderson (1867) transferred H. tener, the type of Haplanthus, to Andrographis but renamed it as A. tenuiflora T.Anderson, based on a nomen nudum, Justicia tenuiflora Wall., which has never been validly published, but with reference to a description of H. tener.According to Art. 55 of ICN (McNeill et al. 2012), the name A. tenuiflora is an illegitimate superfluous name because the epithet 'tener' ought to have been adopted.Accordingly, Kuntze (1891) proposed a new combination A. tenera (Nees) Kuntze. However, Bremekamp (1948) pointed out that this plant was described as early as 1826 by Blume under the name of Justicia laxiflora Blume. Later, Lindau (1895) transferred it to Andrographis as A. laxiflora (Blume) Lindau, which is presently treated as the correct name for this species (Sreemadhavan 1969, Karthikeyan et al. 2009, Hu et al. 2011).Anderson (1867) Anderson (1867) non Nees (1832) was thus published by him inadvertently.In consequence of Andersons error, Kuntze (1903) proposed a replacement name Haplanthodes Kuntze for this later homonym but failed to propose new combinations for the species concerned.Later, Sreemadhavan (1964) proposed another replacement name, Bremekampia Sreem.for Haplanthus T.Anderson, probably unaware of the earlier substitute name Haplanthodes.However, according to Art. 52.2 of ICN (McNeill et al. 2011), the name Bremekampia is not a superfluous illegitimate name because it does not include all the original elements of Haplanthodes Kuntze.The effort of Santapau (1967) to conserve the name Haplanthus Nees ex T.Anderson against Haplanthus Nees for nomenclatural stability was rejected by the committee for spermatophyta who instead recommended to accept Haplanthodes Kuntze as the correct name (McVaugh 1968).Subsequently Majumdar (1971) and Panigrahi & Das (1981) made the necessary combinations for all four recognised species under Haplanthodes.Li (1983) described a new genus Haplanthoides H.W.Li with the sole species H. yunnanensis H.W.Li from Yunnan, China.The name Haplanthoides differs from Kuntzes name Haplanthodes only in the presence of one character 'i'.Therefore these two names are sufficiently alike to be confused and might be treated as homonyms.However, Haplanthoides was later treated as a synonym of Andrographis, with H. yunnanensis considered to be a synonym of A. laxiflora (Blume) Lindau (Chu 1991, Hu 2002, Hu & Cui 2006, Hu et al. 2011).
More recently McDade et al. (2008) included five species from the subtribe Andrographinae in a molecular phylogenetic study on the family Acanthaceae and confirmed its monophyletic nature.However, they have highlighted the need for extra denser sampling from Andrographinae and a critical assessment of morphological characters that may delineate different genera of this group.

MATERIALS AND METHODS
The present systematic study on Andrographis (Acanthaceae: Andro graphinae) in India (Gnanasekaran 2015) is primarily based on the critical examination of fresh specimens collected from different states of India, deposited at MH and herbarium specimens housed at B*, BM, BSI, C*, CAL, CALI, E*, FRC, FRLH, G*, K*, KUN, L*, MH, P*, RHT, S*, SKU, and TBGT using the optical microscope (Nikon SMZ1500) coupled with digital DS-Fi1 camera.In addition, micro-morphological characters of pollen grains and seeds were examined using the Scanning Electron Microscope (Evo M18, Carl Zeiss).

RESULTS
In the present study, the three allied genera, namely Haplanthus, Haplanthodes, and Andrographis, were compared using macro-and micro-morphological characters and the distinguishing characters are summarised in Table 1.The genus Haplanthus can be distinguished from Andrographis by having the following characters (Fig. 1): i. corolla subactinomorphic vs zygomorphic; ii.corolla tube curved vs straight; iii.stamens included vs exserted; iv.filaments pouched at apex vs not pouched; v. anther connectives hairy throughout dorsally vs hairy only at the base or glabrous (A.lawsonii); vi.pollen grains oblate vs prolate or subprolate vii.seeds compressed and not distinctly grooved vs not to hardly compressed with a distinct groove; and viii.seeds with an oblong vs almost circular outline in cross section.
From Haplanthodes, differs in the following characters: i. cladodes (reduced abortive branchlets) in inflorescence absent vs present; ii.ovary with 6 -8 vs 3 -4 ovules per locule, corresponding to 8-16 vs 6 -8 seeds per capsule; iii.seeds compressed and not distinctly grooved vs hardly compressed with two distinct grooves; and iv.seeds without hygroscopic hairs vs with hygroscopic hairs.

TAXONOMIC TREATMENT Haplanthus
Distribution -Bangladesh, Bhutan, Cambodia, China, India, Malesia, Myanmar, Thailand, Vietnam.Note -Scrutiny of literature and examination of specimens at various herbaria revealed that the species of this genus exhibit high levels of morphological variation.Therefore, it is prerequisite to incorporate molecular datasets along with these morphological characters to have better understanding and delimitation of species in this group.

Distribution -Myanmar.
Note -This species was originally placed in Haplanthus T.Anderson when described.Sreemadhavan (1964) did not transfer this species to Bremekampia but later it was treated under Andrographis by Kress & DeFilipps (2003).However, examination of type specimen revealed that this species should be treated under Haplanthus Nees.Under Art.55. 1 of ICN (McNeill et al. 2012), the name H. hygrophiloides T. Anderson (1867) is legitimate even though it was published under the illegitimate superfluous generic name Haplanthus T.Anderson (1867).(Blume)  Distribution -Bangladesh, Bhutan, Cambodia, China, India, Malesia, Myanmar, Thailand, Vietnam.

Haplanthus laxiflorus
Notes -The present study corroborates the views of Hansen (1985) and Hu et al. (2011) that this species is highly variable in habit, leaf shape and size, and structure of inflorescence.The indumentum pattern on the leaves, inflorescence rachises, pedicels, bracts, bracteoles, calyces and capsule also varies considerably.Clarke (1884) recognised two varieties in this species, var.tenuiflora and var.recedens C.B.Clarke.He further distinguished var.tenuiflora into three distinct variations: tenuiflora, parishii and andrographioides.Here, we recognise var.parishii and var.recedens as varieties distinct from var. laxiflorus.
The name Haplanthus tener var.elongatus is lectotypified here.Nees (1832) described this taxon based on the collections of Wallich Numer.List No. 7185 b & c.An examination of these specimens shows that the specimen '7185b' has two gatherings collected from Prome marked as '7185b 1' (K000014474) and Taong Dong marked as '7185b 2' (K000014473) and specimen '7185c' collected from Tavoy (K000014472).Of these, the specimen '7185b 2' is selected here as the lectotype for this name since it is complete with flowers and also matches with the description provided in the protologue.
Note -Anderson (1867) treated H. tener var.elongatus as a synonym of G. parishii whereas the type specimen of the former name matches well with var.laxiflorus.Therefore H. tener var.elongatus is here treated as synonym of Haplanthus laxiflorus var.laxiflorus.Examination of specimens cited by Anderson (1867) under G. parishii reveals that they are a mixture of two distinct taxa.Helfer s.n. is chosen here as the lectotype because all the other syntypes are identical with var.laxiflorus.The variety parishii can be distinguished from var. laxiflorus by being glabrous throughout, with very loose, filiform racemose inflorescences with a solitary flower in each node of the rachis.Note -This variety has not appeared in any of the later works after it was originally described by Clarke (1884) but it is recognised here as a distinct variety under H. laxiflorus.It can be distinguished from the typical variety by having glabrous filiform habit with a very lax compound panicle with clusters of flowers in each node of the rachis.
cited 'Hab.Assam, Masters!; Burmah, Griffith!' in the protologue without stating any other details such as field numbers and place of herbarium.A thorough search of these specimens at different herbaria resulted in locating only the Griffith s.n.collected from Burma at CAL (Acc.No.: 341233) with the name of this species annotated by the original author.Therefore, this specimen is chosen here as the lectotype of this name.a. var.parishii (T.Anderson) Gnanasek., G.V.S.Murthy & Y.F.Deng, comb.nov.