The red algae occur widely in marine habitats and less commonly in terrestrial and freshwater habitats as unicellular, colonial, simple filamentous, flat-bladed and more complex pseudo-parenchymatous forms with quite varied reproductive patterns. The order Batrachospermales is an exclusively freshwater group comprised of three distinct evolutionary lineages, all with uniaxial shoots, pseudoparenchymatous morphologies. The molecular phylogeny and classification are somewhat controversial. The order includes two families, Batrachospermaceae with 6 genera (Batrachospermum, Nothocladus, Petrohua, Psilosiphon, Sirodotia and Tuomeya) and Lemaneaceae with two genera [Lemanea Bory (1808) and Paralemanea (Silva) Vis and Sheath (Vis and Sheath 1992)]. Lemanea and Paralemanea occur in boreal to warm temperate regions of the northern hemisphere (Vis and Sheath 1992, Eloranta and Kwandrans 2007) with a single record of Paralemanea in the southern hemisphere from Brazil (Necchi and Zucchi 1995). Petrohua Saunders (in Vis et al. 2007), a Lemanea-like monotypic genus from Chile, has structural and reproductive features similar to Lemaneaceae but differs molecularly. Psilosiphon, a monotypic genus from Australia (Entwisle 1989), is related t othese other genera and was placed previously in its own family, Psilosiphonaceae (Sheath et al. 1996, Vis et al. 1998), but currently is placed in the family Batrachospermaceae (Entwisle et al. 2009).

Khan (1973) was the first to report a member of the Lemaneaceae (Lemanea mamillosa Kützing) in the Indian subcontinent. Among the approximately 70 taxonomic entities distributed in 19-20 genera of non-marine red algae recorded in India, Lemanea has been reported for more than 4 decades (1973-2015) under 5 specific epithets (Ganesan and West 2013a, 2013b). Judging from the descriptions and figures for Lemanea plants in India and comparing them with current literature, the generic disposition and specific identities of these Indian records are problematic.

The precise taxonomic identity of the Indian Lemanea specimens is relevant in view of their various uses: 1) as an edible alga (called Nungsham-“hair of stones”) (Fig. 1A) collected in rivers of Manipur and sold fresh in local markets (Fig. 1B) and in dried packets (Fig. 1C) also sold on the internet, 2) medical uses (Bhosale et al. 2012) by numerous tribal populations of the remote northeast states of India, and 3) potential for biofuels (Rout et al. 2011). Based on freshly collected material (Fig. 1A) from Manipur state, we presented preliminary morphological observations on this entity (Ganesan and West 2013b). Manipur (Fig. 1D) is in the Indo-Burma Biodiversity Hotspot region and requires more extensive sampling. The previous morphological observations are supported here with molecular rbcL data to show that these Manipur specimens belong to the genus Lemanea and represent a new species.

Lemanea specimens were collected by J. Rout and M. Thajamandhi in the Chakpi River, Thoubal district, Imphal, Manipur, India on December 23, 2013. For morphological studies, the materials were preserved in 4% formalin, voucher specimens were mounted on herbarium paper and specimens for molecular analyses were blotted dry with tissue and placed in small snap lock plastic bags containing silica gel beads.

For microscopic observations a mounting medium including stain was prepared as 0.02% aniline blue WS in 50% corn syrup and 0.5% phenol (to prevent contamination by bacteria and fungi). Surface views and cross sections were prepared on a microscope slide, removing water carefully with blotting paper, adding a drop of mounting medium and arranging the specimens with a fine needle and forcep tip. The medium was stirred gently to allow good penetration of the stain without disrupting the sections. The material was allowed to absorb the stain for about thirty minutes. A cover slip was then gently placed on the slide. After 1-2 days each slide was sealed with a clear fingernail polish to prevent the medium from drying and receding under the cover slip margins. Photographs were taken with a Zeiss binocular research compound microscope (Carl Zeiss, Jena, Germany) and Canon G3 camera (Canon, Tokyo, Japan). The holotype photo (Fig. 2A) was taken with a Canon G10 camera. Photographic plates were made using Adobe Photoshop CS4.

Total DNA was isolated from silica gel-dried material using a modified CTAB procedure (Zuccarello and Lokhorst 2005). Amplification and sequencing of the plastid-encoded large subunit of the ribulose bisphosphate carboxylase / oxygenase gene (rbcL) used amplification primers presented in Freshwater and Rueness (1994) (F8-R753, F577-rbcL-st). Amplified products were checked for correct length (approximate 1,000 bp length), purity and yield on 1% agarose gels. Polymerase chain reaction products were cleaned using ExoSAP-IT (USB, Cleveland, OH, USA) and commercially sequenced (Macrogen Inc., Daejeon, Korea). The outgroups used were Thorea and Nemalionopsis of the sister order Thoreales (Müller et al. 2002, Johnston et al. 2014).

Sequences were edited, assembled and aligned using the Geneious software package (Biomatters, available from http://www.geneious.com/). Alignment was straight forward as no gaps were found in the data set. The program jModeltest version 2 (Darriba et al. 2012) was used to find the model of sequence evolution. Maximum likelihood (ML) was performed with RAxML 7.2.8 (Stamatakis 2006). RAxML was performed with all threes codons partitioned and the GTR + gamma model and 500 non-parametric bootstrap replicates.

Bayesian inference was performed with MrBayes v3.1.2 (Ronquist and Huelsenbeck 2003). Analyses consisted of two independent simultaneous runs of one cold and three incrementally heated chains, and 3 × 10⁶ generations with sampling every 1,000 generations. Codons were partitioned. The log files of the runs were checked with Tracer v1.5 (Rambaut and Drummond 2007) and a burn-in sample of 100 trees was removed from each run before calculating the majority rule consensus tree.

Our results indicate that the species collected in the Chakpi River is a new species of Lemanea that we describe here.

Diagnosis. Plants to 10 cm long, growing in clusters on stones, green to dark purple, with or without a basal constriction to form a stalk, shoots to 650 µm diam., with sparse irregular branching in the lowermost region of the erect shoots (Figs 1A & 2A), ultimate branches with attenuated apices (Fig. 2B); basal Chantransia phase sometimes evident; hair cells absent, four ray cells of a single type formed at right angles on each cell of the single central axial filament (Fig. 2C & D), internal corticating filaments lacking (Fig. 2C); outer cortex cells cylindrical, compact (7-9 µm diam. × 11-12 µm long) in side view, irregular to quadrate shape in surface view (Fig. 1B & C); monoecious, spermatangia (3-5 µm diam. × 10-12 µm long) in irregular sub-prominent patches or incomplete irregularly shaped rings (to 130 µm wide) around the axial nodes (Fig. 3A-C); carposporophytes projecting into the thallus cavity, gonimoblast filaments short and simple to repeatedly branched (Fig. 3D-G), each terminal cell forming a single, large, elongate cylindrical carpospore with round ends, 38-54 µm diam. and 85-115 µm long, filled with abundant, colourless spheres (3-5 µm diam.) and a prominent central nucleus (Fig. 3F). There is no ostiole opening (Fig. 3D) for carpospore release so presumably release occurs when the gametophyte disintegrates.

The holotype specimen is deposited in the Royal Botanic Gardens Melbourne, National Herbarium of Victoria, Australia (MEL 2381649). Isotype specimens are also deposited in the University of Michigan Herbarium (MICH 1210723) and University / Jepson Herbaria, University of California, Berkeley (UC / JEPS). Specimens (Figs 1A & 2A) were collected December 23, 2013, near Chakpi village on the Chakpi River (24°16′ N, 93°52′ E) growing profusely on stones at 755 m above sea level (water temperature was 14℃ at the time of collection). The river flows through the villages of Serou and Sugnu in the Thoubal district, 75 km from the capital city Imphal of Manipur state. L. manipurensis is probably more widespread in Manipur river systems, but more sampling is needed to confirm this. The species is named for the state of Manipur, the only location in India where Lemanea is currently known.

Phylogenetic analysis. The phylogenetic tree (Fig. 4) of the Batrachospermales and Thoreales using rbcL gene sequences clearly shows a monophyletic Lemaneaceae, including both Lemanea and Paralemanea. The phylogeny reveals that L. manipurensis is molecularly distinct and in a sister lineage to other Lemanea samples accessed from Genbank, with weak support (ML BP 68). The rbcL phylogeny shows great genetic variation in the genus Lemanea and the most commonly sequenced species (L. fluviatilis) is not monophyletic based on that species name assigned to nucleotides in Genbank.

GenBank accession number. KP407869.

Morphologically the specimens we examined clearly belong to the genus Lemanea (Table 1).

Among the approximately 19-20 non-marine red algal genera known hitherto from the Indian sub-continent (Ganesan and West 2013a, 2013b), Lemanea is reported only from Manipur state, and occurs mostly during the colder months (December to February) when harvesting, drying and packaging are done for sale in local markets and online stores (Fig. 1A-C) as food and for ethnobotanical / medical uses (Bhosale et al. 2012). Manipur state is located at elevations of 50-3,300 m above sea level) and has various rivers, of which the Chakpi, Imphal, Iril, Manipur and Thoubal are well known (24°16′ N, 93°52′ E) (Singh and Gupta 2014, Fig. 1). Lemanea specimens have been collected and reported from all five rivers (Singh and Gupta 2010a, 2010b, 2011, Gupta and Singh 2012). Critical observations on the anatomical and reproductive details, supplemented by molecular sequences data are lacking for specimens collected here. Five different specific names, i.e., L. australis Atkinson [now Paralemanea grandis (Wolle) Kumano], L. catenata Kützing [now Paralemanea catenata (Kützing) M. Vis et R. Sheath], L. fluviatilis (Linneaus) C. Agardh, L. mamillosa Kützing, and L. torulosa Sirodot have been assigned during more than four decades (1973-2015) for the Indian Lemanea / Paralemanea populations (Table 2).

We believe the binomial L. mamillosa was misapplied by Khan (1973). He said “from the outer ends of the T shaped cells, ascending and descending lateral filaments were formed.” More importantly, Khan (1973, p. 173, Fig. B) misjudged large carpospores with numerous, small, colourless spheres and a large central nucleus as spermatangia.

Deb et al. (1986) questioned Khan’s identification of L. mamillosa and referred their specimens to L. australis. A careful analysis of the description and figures by Deb et al. (1986) also raised serious doubts about the taxonomic identity of specimens they examined. The central axial filament was covered by slender descending corticating filaments, two types of ray cells were seen and the outermost cortex with spermatangial nodal rings that are features of Paralemanea (Deb et al. 1986, Pl. II, Fig. 2). They pointed out that the carpospores of the Indian material were comparatively larger (45-63 × 72-126 μm) than those of L. australis given by Atkinson (1890) for the North American specimens. Voucher specimens of Deb and others were stated to have been deposited at the Botanical Survey of India: “Sugnu, Feb 1971 & Dec 1973, Sh. Singh. S. n. (BSIS).” However, these were not traceable during our (EKG and JW) visit to the BSIS Herbarium, Kolkata, W. Bengal in November 2013. Sugnu is near where our specimens for the present study were collected.

Desikachary et al. (1990) concluded, based on observations of Deb et al. (1986), that they were L. mamillosa. The issue was further clouded by Desikachary et al. (1990, Fig. 20A-P) stating that several figures of North American L. mamillosa populations (Mullahy 1952, Figs 17-23 & 25-33) were similar to the Indian populations. Both sets of figures clearly show that axial filaments with many down-growing cortical filaments characteristic of Paralemanea.

Bhosale et al. (2012) studied the morphology of abundant fresh material from the same general area as our present collection site (Chakpi River, Thoubal District) and concluded that the Indian specimens were misidentified earlier and that they should be referred to the widely distributed Lemanea fluviatilis. However, these authors failed to observe some critical anatomical and reproductive details, e.g., there is no mention or illustration to show if the axial filament is naked or surrounded by cortical filaments. More importantly, the nature of the mature carpospores in their studies (Bhosale et al. 2012, Fig. 2G) clearly showed a single, large and terminal carpospore on each gonimoblast branch which is in agreement with our specimens. In other words, the nature, shape and size of carpospores shown by Bhosale et al. (2012) are in sharp contrast to L. fluviatilis, where carpospores are generally spherical and in simple or branched chains (see Sheath and Sherwood 2011, Pl. 42E & F, Guiry and Guiry 2015). Identification as L. fluviatilis by Bhosale et al. (2012) was followed subsequently in other studies on biofuel production (Rout et al. 2011), nutrient content (Singh and Gupta 2010a, 2010b), treatment of diabetes (Devi 2011), trace elements (Bino-Devi et al. 2011), biosynthesis of gold nanoparticles (Sharma et al. 2014) and anti-urolithiatic agent (Mikawlrawng et al. 2014).

Singh and Gupta (2010a, 2010b, 2011, 2014) and Gupta and Singh (2012) extensively studied the ecology of Lemanea populations and identified five different species. These are, in addition to L. fluviatilis from the Thoubal River, (1) L. australis, (2) L. torulosa from Manipur and Imphal Rivers, (3) L. mamillosa from Iril River, and (4) L. catenata from Iril River in the Manipur River systems. However, critical details on thallus anatomy and reproductive features were not provided. Gupta and Singh (2012, p. 23, Table 2) and Singh and Gupta (2010a, p. 52, Table 2) further confused this issue showing in a comparative table that some populations had axial cortical filaments (L. mamillosa and L. catenata), while in others (L. australis, L. torulosa, and L. fluviatilis) these filaments were not shown. Also, excepting L. fluviatilis, all the other species had spermatangia in rings. The bionomial L. catenata was used throughout their studies but the present binomial is Paralemanea catenata (Kützing) Vis et Sheath. Most of their illustrations were habit views without anatomical details in illustrations i.e., with or without axial filaments in the 5 species they identified. Singh and Gupta (2010a, p. 54, Pl. 5b) illustrated “spermatangial sori in rings between the internodes” in L. catenata. These appear to be the large carpospores characteristic of L. manipurensis. Finally, Deb et al. (1986), Bhosale et al. (2012), Gupta and Singh (2012), and Singh and Gupta (2010a) did not provide appropriate magnification scales in their photographic and line-drawing illustrations. Many of the above identifications referred to species unknown in India and lacked photographic evidence of critical details.

Among the records of the Lemanea / Paralemanea species complex in countries neighbouring India, Faridi (1971) mentioned Lemanea and four other freshwater red algae, Chroodactylon (as Asterocytis), Audouinella, Batrachospermum, and Compsopogon, in a diagnostic key to the freshwater red algae of Pakistan. Later (Faridi 1975) mentioned L. mamillosa from Northwest frontier (Parachinar), but without description or illustration of the specimens he examined. Nizamuddin (1988) remarked that “Faridi’s collections are not available to make any comments on their occurrence in Pakistan.”

Xie et al. (2004) studied 4 species of Lemanea and 2 species of Paralemanea from China. Lemanea simplex C. -C. Jao and L. sinica C. -C. Jao were described by Jao (1941) and L. crassa S. L. Xie et Z. X. Shi and L. ramosa S. L. Xie et Z. X. Shi were described by Xie et al. (2004). Since there are no other reports of these species they may be endemic to different regions of China. The carposporophyte structure and carpospore shape and size of L. simplex and L. sinica are similar to those of L. manipurensis specimens. The specific differences between L. manipurensis and L. sinica are shown in Table 3. Except for thallus height and branching mode, L. manipurensis is remarkably similar to L. sinica although carpospore size is slightly larger. No other reports of L. sinica are known in China or elsewhere. Future molecular studies on L. sinica would undoubtedly reveal whether or not it differs from L. manipurensis. The type locality of L. sinica is Chun-tien, Yunnan (32°03′01″ N, 110°36′50″ E), while that of the Indian species is Chakpi, Thoubal district, Manipur (24°16′ N, 93°52′ E), perhaps not a very significant geographical distance but the many diverse river systems and mountain ranges between these localities may be formidable barriers for dispersal.

In the five Manipur river systems, Lemanea populations may be abundantly distributed in the different river segments. Hence the possibility of cryptic speciation among the different populations cannot be ruled out. Only critical molecular studies would clarify this point. Lemanea and Paralemanea are known to grow together in the same ecological habitat and there are instances where the same herbarium packet contained more than one species (Palmer 1941). Harvested material of Lemanea from different places and the processed dried packets sold in the local markets may be helpful on morphological and molecular bases in resolving whether Paralemanea species also occur in India.

Lemanea, Paralemanea, Petrohua, and Psilosiphon species offer few reliable morphological characteristics to separate them at the generic level. As stated by Sheath (2003, pp. 208, 216 & 217) the presence or absence of axial filament cortication and the configuration of spermatangial rings or patches should be the most useful features in defining Lemanea and Paralemanea, but these 2 generic attributes apparently are not consistent, thus raising doubts about the generic status of Paralemanea as different from Lemanea. In any case molecular analyses in this and other published work suggest Paralemanea is paraphyletic leaving the issue unresolved.

The landmark publication by Jao (1941) on L. sinica is overlooked in many later publications on the Lemaneaceae. Jao (1941, p. 271, Pl. VII, Figs 50-52) described L. sinica specimens with “the antheridial bands are mostly in complete rings around the antheridial zones. In older portions of the thallus they are occasionally slightly interrupted but also kept a band-shape. Probably the interrupting is due to the hypertrophy of the tissue beneath the antheridial bands. In addition to the ordinary antheridial bands, small antheridial sori sometimes appear on the old carpogonial zones.” Xie et al. (2004) said in their Lemanea species key that L. sinica has “Spermatangia in perfect rings” but they did not obtain any new L. sinica specimens. Hu and Wei (2006) stated “the antheridia are always in well-defined bands” without further evidence. Jao (1941) clearly mentioned the “naked central axis” lacks cortical filaments and commented that it is a “peculiar combination of the characteristics of two genera” in L. sinica. Kučera and Marvan (2004), in a taxonomic study of the Lemaneaceae in the Czech Republic, characterized Lemanea with a central axis with or without cortical filaments. According to them, “the regular formation of spermatangia in rings varies with age and the magnification used.” These variable features in spermatangial formation in their material of L. fluviatilis were also shown in illustrations (Kučera and Marvan 2004, p. 168, Fig. 3A-D, Simić and Djordjević 2011, p. 515). Xie et al. (2004) in their study of Chinese Lemaneaceae, mentioned that in Paralemanea parvula (Sirodot) S. L. Xie & Z. X. Shi, spermatangia are found in patches (Xie et al. 2004, p. 886, Fig. 13). Eloranta et al. (2011, p. 96) preferred to keep P. parvula in Lemanea as an uncertain species, but notably remarked that “spermatangial papillae form separate or sometimes coalescent tubercles.” Anatomical details, i.e., whether cortical filaments are present or not in P. parvula is not known in Chinese and European specimens. Eloranta et al. (2011) did not include the reference of Xie et al. (2004). The above mentioned instances clearly show the ambiguity of using some anatomical and reproductive features to differentiate Lemanea and Paralemanea.

It is significant that the five species of Lemanea (L. crassa, L. manipurensis, L. ramosa, L. simplex, and L. sinica) from the Southeast Asian continent are remarkably similar, producing just a single and elongate terminal carpospore on each short simple to branched carposporophyte filament with narrow cylindrical cells. Most other species of Lemanea and Paralemanea produce spherical carpospores serially in simple or branched chains (catenate). Detailed and comparative morphological and molecular data would clarify whether this group of Lemanea species warrants segregation into a new genus, different from closely related genera like Lemanea, Paralemanea, and Petrohua. A case in point is that extensive molecular and morphological data on Batrachospermum species sensu lato resulted in the segregation of Kumanoa from Batrachospermum with spirally twisted carpogonial branches and axial carposporophytes (Entwisle et al. 2009) and Sheathia with hetero-corticated thalli, carpogonia on undifferentiated cells and exserted carposporophytes (Salomaki et al. 2014).

It is noteworthy that Necchi et al. (2013) found a very low genetic diversity between species of the two genera Composopogon and Compsopogonopsis (Compsopogonales, Rhodophyta) based on global sampling and molecular data. Species of these 2 genera were distinguished hitherto by caeruleus morphology (without cortication) for Compsopogon species and leptoclados morphology (with cortication) for Compsopogonopsis species. Necchi et al. (2013) concluded that only one genus, Compsopogon, can be recognized.

It is still too early to provide a complete molecular and morphological story concerning generic and species delimitations for Lemanea and Paralemanea.