The freshwater red algae (Batrachospermales, Rhodophyta) of Africa and Madagascar I. New species of Kumanoa, Sirodotia and the new genus Ahidranoa (Batrachospermaceae)

Our knowledge of the diversity of African freshwater red algae is rather limited. Only a few reports exist. During our field work in the last five years we frequently encountered freshwater red algae in streams in Rwanda and Madagascar. Here we describe four new species and one new genus of freshwater red algae from the Batrachospermales, based on morphological and molecular evidence: Kumanoa comperei from the Democratic Republic of the Congo and Rwanda is related to K. montagnei and K. nodiflora; Kumanoa rwandensis from Rwanda is related to K. ambigua and K. gudjewga; Sirodotia masoalen­ sis is related to S. huillensis and S. delicatula; and the new genus and species Ahidranoa madagascariensis from Madagascar is sister to Sirodotia, Lemanea, Batrachospermum s.str. and Tuomeya. There is also evidence for the presence of Sheathia, which was recorded as yet-unidentifiable Chantransia stages. These are among the first new descriptions since 1899 from the African continent and since 1964 from Madagascar. A short history of the exploration of freshwater red algae from Africa and Madagascar is provided. All new taxa are accompanied by illustrations and observations on their ecology.


Introduction
The history of the exploration of freshwater red algae in Africa and Madagascar 'Very little is known concerning the freshwater algae from Africa' (West & West 1897). This is still true more than 120 years after that statement. While interest in the long-neglected group of freshwater red algae has increased in the Northern Hemisphere (e.g., Entwisle et al. 2009;Eloranta et al. 2011;Vis et al. 2012;Knappe & Huth 2014;Salomaki et al. 2014), South America (e.g., Necchi 1990;Necchi et al. 2010Necchi et al. , 2019 and Australia (e.g., Entwisle & Foard 2007;Entwisle et al. 2016), there are only limited data available for Africa.
The first collection of a freshwater red alga from Africa was made by W. G. Schimper, who collected a Lemanea in 1840 in the Simien Mountains of Ethiopia ('in latere medio boreali montis Silke'). This specimen was subsequently described by Kützing (1849) as Lemanea abyssinica. The first species of Batrachospermum was published as B. patens (Suhr 1840) from Kroemsrivier (today Kromrivier) in the Cape region of South Africa, collected by Drège (without date). Szinte et al. (2020) place Batrachospermum patens into synonymy with Tor ularia atra (see below).
The most extensive collections of freshwater red algae in Africa were made by Friedrich Welwitsch between 1855 and 1860 in Angola. The results were published by West & West (1897) almost 40 years later. The recorded species are Batrachospermum angolense [= Sirodotia angolensis according to Guiry (2019) but considered a heterotypic synonym of Torularia atra (= Atrophycus ater) in the previous literature, e.g., Entwisle 1992], Batrachosper mum nigrescens (Sirodotia nigrescens = Torularia puig gariana), Batrachospermum gracillimum (= Kumanoa gracillima), Batrachospermum huillense (= Sirodotia huil lensis), Hildenbrandia rivularis and H. angolensis. West & West (1897) remark that it is 'interesting to note that the earliest collection of algae made in Africa has been found to be more extensive and representative than any hitherto described' (West & West 1897: 1-2). None of these species has been re-collected from Africa but they were later recorded from South America or Europe. One of the last new descriptions from the African continent is Batra chospermum bohneri (Schmidle 1899a) from Cameroon, which is assigned by its author to section Helminthosum and is said to have a large stalked trichogyne. Unfortunately it is not mentioned in the subsequent literature and is also omitted from AlgaeBase (Guiry 2019). As the type probably has been destroyed and the protologue is not accompanied by any illustration, it is considered a doubtful name here. Separately published illustrations from this material (Schmidle 1899b) led Szinte et al. (2020) to identify it as Montagnia macrospora. A collection from Rwanda by Johannes Mildbraed in 1907 was identified as Batrachospermum bohneri (Lemmermann 1914: Rugege Wald, kalter Quellbach des Rukarara) but the specimen was destroyed in Berlin and the record is considered doubtful. Several other red algae were collected at this locality during our research but no specimen matched the description of Batrachospermum bohneri.
In his review of the subaerial and freshwater algal flora of the Tropics, Fritsch (1907) listed for Africa the five species described by West and West (1897) and Schmidle (1899a) but ignored the records of Lemanea. Thus, nine species of freshwater red algae were known from tropical Africa in 1907, and for 112 years not much progress was made. Borge (1928) listed two species of Batra chospermum, B. dillenii (= Torularia atra) and B. vagum (= B. turfosum) from Tanzania (eastern Usambara, Kilimandscharo) but the identifications are considered doubtful, as no specimens could be retrieved, so they serve only as an indication that there are freshwater red algae present in the mountains of eastern Tanzania. The same is true for the record of Tuomeya fluviatilis (= T. americana) from eastern Usambara (Borge 1928). In some papers on the diversity of freshwater algae, Compère (1975) recorded Audouinella hermannii from Chad, and Da et al. (1999) and Alika & Akoma (2012) listed Batrachospermum tur fosum (as B. vagum) for the Ivory Coast and Nigeria, these latter records probably erroneous. By coincidence, another paper has just appeared independently (Szinte et al. 2020) in which the authors describe the new species Kumanoa bouwmanii, Sheathia murpheyi and Sirodotia kennedyi, all from Zambia, and they name a Chantransia stage as C. azurea from South Africa. They also give an overview of available herbarium specimens, mainly from BR and PC (abbreviations after Thiers 2019), the majority of which are not identifiable.
For South Africa, the taxon described by Suhr (1840) and the two species described by Rabenhorst (1855) are known, with the recent additions of Sirodotia suecica (Lam et al. 2012) and Kumanoa iriomotensis, collected in Western Cape Province in 2006 . The only records of freshwater red algae from Madagascar are Batrachospermum gelatinosum (Fritsch 1914, record probably erroneous) and Nothocladus afroaustralis (Skuja 1964), known only from the Fort Dauphin region in southeastern Madagascar.

Phylogenetic relationships of Batrachospermum and related genera
With about 112 species recognized, the genus Batrachos permum with the type species B. gelatinosum was considered the most species-rich genus of freshwater red algae, and was divided into two subgenera (Batrachospermum, Acarposporophytum), the former containing eight sections (Kumano 2002). However, based on molecular data from phylogenetic studies (Vis et al. 1998;Vis & Entwisle 2000;Vis et al. 2005;Entwisle et al. 2009), this genus has been shown to be paraphyletic. Early accepted segregates are Sirodotia and Nothocladus. Entwisle et al. (2009) proposed acceptance of different sections for the time being: Acarposporophytum, Aristata, Batrachospermum, Helminthoidea, Macrospora, the informal 'Australasica group', Setacea, and Turfosa and Virescentia. The sections Contorta and Hybrida were placed in the new genus Kumanoa (Entwisle et al. 2009). Subsequently, Salomaki et al. (2014) placed members of section Helminthoidea in the new genus Sheathia. More recently the new genus Nocturama ) was erected, now comprising two species from Australia and South America ). In the first paper the genus Nothocla dus, formerly containing three species from Madagascar and Australia (which now constitute section Nothocla dus) was enlarged to comprise most of the old informal 'Australasica-group' with the new sections Australasicus, Theaquus and Kraftii, and also section Setaceus. Contrary to the arguments of Entwisle et al. (2016), Rossignolo and Necchi (2016) subsequently raised section Setacea to generic level and included three species. For nomenclatural reasons the generic name Setaceus had to be replaced by Atrophycus (Rossignolo et al. 2017). Wynne (2019) showed that the name Torularia (Bonnemaison 1828) has priority over Atrophycus, and thus the three recognized members of the former section Setacea were placed in the genus Torularia. Section Virescentia was raised to genus level under the name Virescentia (Necchi et al. 2018). Sections Acarposporophytum and Aristata were raised to generic status under the names Acarposporophycus and Visia (Necchi et al. 2019b), and section Macrospora to genus Montagnia (Necchi et al. 2019a). Two new genera were recently added: Volatus ) with three species from North America and Europe, and the monotypic Lympha with Lympha mucosa (Evans et al. 2017). Thus the genus Batrachospermum s.str. comprises only Batrachospermum gelatinosum with a few related species, while section Turfosa is still not assigned to a definite genus and is thus kept as Batrachospermum s.l. For a summary of the taxonomic changes and the currently accepted taxonomy we refer to Table 1. Since 2011, several freshwater red algae have been collected from tropical Africa (Gabon, Rwanda, Burundi) and Madagascar. Most of them could not be identified with the available literature. The aim of this study is to reconstruct the phylogenetic relationships between these taxa based on molecular and morphological evidence, and to characterize the taxa collected in tropical Africa and Madagascar. In this first paper we deal with collections of the genera Ahidranoa, Kumanoa and Sirodotia from Rwanda and Madagascar, and we describe one genus and four species new to science. We also provide evidence for the occurrence of the genus Sheathia, albeit only collected as Chantransia stages.

Sampling and investigations
Freshwater red algae were sampled between 2014 and 2018 from four locations in Rwanda and two in Madagascar. Herbarium vouchers with specimens preserved in 70% alcohol were deposited in BR, TAN and KOBL (abbreviations after Thiers 2019, Table 2). Their morphology was examined and photographed with a KEYENCE VHX-S15 digital microscope.

DNA isolation
Silica-dried plant material was homogenized in a 2 ml Eppendorf cap (round bottom) with two glass beads (ø 5 mm) and a small amount (tip of spatula) of autoclaved sand at 30 hz for 2 min. using a TissueLyser II (QIA-GEN, Venlo, Netherlands). From the obtained powder, genomic DNA was extracted using a NucleoSpin Plant II Kit (Marchery-Nagel, Düren, Germany), following the customized protocol of the supplier.

Amplification and sequencing
Amplification of COI followed Saunders (2005) using the primers GazF1 and GazR1, while new primers were designed for amplification of rbcL (rbcL-redF: TGCYAAAATGGGWTAYTGG; rbcS-redR: For amplification of COI the following PCR profile was used: 94°C 3 min. followed by 34 cycles (94°C 60 s, 45°C 60 s, 72°C 60 s) and final extension of 10 min. at 72°C. Newly generated sequences were deposited in GenBank (Table 2).

Contig assembly, alignment and phylogenetic analyses
Quality control of the pherograms, contig assembly and alignment was done in PhyDE1 (available at www.phyde. de). In order to place the African taxa in phylogenetic context, a representative set of freshwater red algae rbcL and COI sequences were downloaded from GenBank (Table 3) and aligned with the newly generated sequences, using PhyDE1. Sampling was guided by Entwisle et al. (2009). Maximum likelihood (ML) analyses were performed using RAxML-NG (Kozlov et al. 2019) via the RAxML BlackBox (raxml-ng.vital-it.ch), applying the GTR + Γ + I model. Bootstrap analysis was performed with the automatic bootstrapping option in effect at a cutoff of 0.3. Bayesian analyses were performed with MrBayes v.3.2.5 (Ronquist et al., 2012), applying the GTR + Γ + I model. Four runs with four chains (10 7 generations each) were run simultaneously, with chain sampling every 1000th generation. Tracer v.1.7.1 (Rambaut & al., 2018) was used to examine log likelihoods to determine the effective sampling size and stationarity of the MCMC search. Calculations of the consensus tree, including clade posterior probability (PP), were performed using the relburnin function in MrBayes; that is, the first 25% of the trees were discarded. Consensus topologies and support values were compiled and drawn using TreeGraph v.2 (Stöver & Müller, 2010). Bootstrap support (BS > 50) is depicted above and posterior probability (PP > 0.8) below the branches of the maximum likelihood tree. Support values in the text are given as BS/PP.

Phylogenetic discussion
The well-resolved and highly supported phylogeny of the Batrachospermales indicates that the current concept of the generic boundaries seems to be settled. All genera in which more than one taxon was included, such as Kumanoa, Batrachospermum, Lemanea, Siro dotia, Sheathia, Nocturama, Torularia and Nothocladus, are monophyletic and receive significant if not maximal support. Specifically, this new analysis supports recognition of the genus Torularia, as proposed by Rossignolo & Necchi (2016), which based on an earlier tree  would have resulted in a paraphyletic Notho cladus. However, our tree has limited taxa sequenced within the Nothocladus clade, and recognition of Toru laria still runs counter to the concern raised in Entwisle et al. (2016) around the proliferation of less informative small genera. With respect to Batrachospermum more analyses need to be done, as indicated by Entwisle et al. (2009). The resolution of both Madagascan specimens within the 'BatrachospermumLemaneaSirodotia' clade points to an interesting aspect related to the speciation and biogeography of red algae on the island, requiring further investigations.

Taxonomy
The genus Kumanoa The genus Kumanoa (Entwisle et al. 2009) was proposed to accommodate the members of Batrachospermum sections Contorta and Hybrida. One of the main distinguishing characters is the twisted or curved carpogonial branch. An expanded molecular phylogeny of Kumanoa was provided by Vis et al. (2012), with additions by Necchi et al. Description. Plants moderately mucilaginous, thalli 3-14 cm high, branching irregular and abundant. Whorls reduced, composed of primary fascicles, 157-389 µm in diameter, barrel-shaped or obconic, slightly distant to confluent with age. Internode 120-190 µm long. Primary fascicles straight, with 7-8 cell-storeys. Secondary fascicles abundant, covering the entire internode, usually shorter than the primary fascicles, 2-3(-8) cell-storeys.
Ecology and distribution. Kumanoa comperei occurs in small acidic streams in montane forest between 1100 and 2100 m. The streams are slow-flowing and half-shaded, with usually no aquatic vegetation except an unidentified Cladophora species at very low abundance. No aquatic bryophytes have been observed. So far the new species is known only from two localities in Nyungwe National Park in Rwanda and one locality in Kahuzi-Biéga-National Park in the Democratic Republic of the Congo.

The new genus Ahidranoa
While comparing two samples of a freshwater red alga, they seemed at first similar to the Southern Hemispherean genus Nocturama that was established to accommodate a single species from Australia and New Zealand, Batra chospermum antipodites (Entwisle 1995). This species was formerly included in a cluster of Australian and New Zealand species that was distinguished by carpogonia subtended by a relatively short filament of modified cells (Entwisle & Foard 1997, 2007. The other taxa included in this cluster are now placed in the expanded genus Nothocladus . Necchi et al. (2016) described a second species of Nocturama, N. novamun densis  from Brazil, Rio Grande do Sul. Molecular evidence, however, supports the recognition of a new genus. This taxon is well supported as sister to a clade with Sirodotia, Lemanea, Batrachospermum s.str. and Tuomeya (Fig. 1) Description. Thalli brownish to reddish-brownish, moderately mucilagineous, abundantly and irregularly branched, up 5-8 cm long. Rhizoidal cells not inflated, cylindrical, 120-130 µm in diameter. Whorls spherical or barrel-shaped, 150-500 µm in diameter. Internodes 160-280 µm long. Primary fascicles curved, composed of 10-13 cell-storeys, distal cells ellipsoid or obovoid, 380-410 × 77-80 µm. Secondary fascicles present, below the whorls of primary fascicles, composed of 2-8 cell-storeys, not covering the whole internode.
Ecology and distribution. Ahidranoa madagascariensis occurs in the Marojejy Massif in a small shaded stream just below Cascade d´Humbert. The second locality is the River Ambanizana on Masoala presque-île, which is quite sun-exposed. Ahidranoa madagascariensis, however, occurs between the leaves of a member of Podostemaceae and thus is also growing in shade. At the same locality, the new species Sirodotia masoalensis (see below) is growing in full sun. Ahidranoa madagascariensis is so far restricted to northeastern Madagascar.

Evidence for the occurrence of the genus Sheathia in Central Africa
Two specimens of a bluish grey acrochaetioid alga were collected and tentatively identified as Audouinella. However, molecular data suggest that these are Chantransia stages of the genus Sheathia, and the two accessions accordingly cluster with other Sheathia species, but no mature gametophyte could be detected.  Skuja (1934) assumed that only the red forms should be included in the genus Audouinella, and that the blue forms are young Chantransia stages of the genus Batrachosper mum. Necchi & Zucchi (1997) support this suggestion. Also, Chen et al. (2014)

The genus Sirodotia
Sirodotia was segregated from Batrachospermum, mainly due to the asymmetrical base of the carpogonium and the indeterminate gonimoblast filaments in the carposporophyte (Lam et al. 2012). From the eight taxa accepted by Kumano (2002), Sirodotia goebelii (Entwisle & Foard 1999) was placed in synonymy with S. suecica, and S. tenuissima was confirmed as a synonym of S. suecica (Lam et al. 2012). In Africa, Sirodotia huillensis was described from Angola (see above) but has never been re-collected near the type locality. Material with DNA data has so far only been studied from the United States and Mexico (see Lam et al. 2012). A specimen identified as Sirodotia aff. huillensis from South Africa differed from the samples from Texas, Arizona and Mexico, and could well represent the typical Sirodotia huillensis. Sirodotia suecia is widespread and recorded from the United States, Europe, Australia and New Zealand, and one sample from South Africa. A new species, Sirodotia kennedyi (Szinte et al. 2020), has been described from northern Zambia and is sister to our sample but morphologically differs distinctly, and Sirodotia masoalensis is described here as a new species. Description. Thalli robust, brownish to reddish-brownish, moderately mucilagineous, abundantly and irregularly branched, up 7-14 cm long. Whorls reduced, obconical to barrel-shaped, 115 µm in diameter. Internodes 250-380 µm. Primary fascicles straight, composed of 5-6 cell-storeys, branching 3-4 times, distal cells ellipsoid or obovoid, terminal hairs lacking. Secondary fascicles numerous, below the whorls of primary fascicles, composed of 2-3 cell-storeys, straight, up to 20-22 µm long, not covering the whole internode.

Ecology and distribution.
Known only from small tributaries of the River Ambanizana on the Masoala Peninsula in Madagascar, where the species is associated with various green algae including Chara sp.
Notes. Sirodotia masoalensis is part of the Sirodotia clade ( Fig. 1) and sister to S. kennedyi. Both cluster with S. huil lensis and S. delicatula. Several characters of Sirodotia masoalensis, such as the origin of gonimoblast filaments, are not observed, but the new species differs from all congeners in the reduced primary fascicle whorls. ologique de Tsimbazaza (PBZT, Antananarivo) for their constant support, Claudia Schütte for lab assistance and for cheering up Johanna when she defended her BSc thesis on this project, and Timothy Entwisle and an anonymous reviewer for valuable comments on the manuscript. We thank the Madagascar National Parks (MNP) at Antananarivo and Maroantsetra for research permits, the Direction des Eaux et Forêts (Antananarivo) for export permits, the Rwanda Development Board, Tourism and Conservation (RDB) for good cooperation and for research and export permits, and BMUB (Project 16_III_083_RWA_A_Cyamudongo Regenwald) for financial support of field trips to Rwanda. The first author thanks the Akademie der Wissenschaften und Literatur Mainz for funding support of field trips to Madagascar and Rwanda.