The genus Lithophyllum in the north-western Indian Ocean , with description of L . yemenense sp . nov . , L . socotraense sp . nov . , L . subplicatum comb . et stat . nov . , and the resumed L . affine , L . kaiseri , and L . subreduncum ( Rhodophyta , Corallinales )

Based on literature, the genus Lithophyllum was represented in the whole Indian Ocean by 14 taxa, mostly in need of revision in a modern context. Molecular analyses integrated with morpho-anatomical comparisons between the recently revised type material of Lithophyllum kotschyanum Unger, and the related infraspecific taxa, lead to a resumption of L. affine, L. kaiseri, and L. subreduncum, and the description of L. socotraense sp. nov., L. yemenense sp. nov., and L. subplicatum comb. et stat. nov. Detailed accounts are provided for each species, including keys, along with information on synonymy, examined collections, distribution, habitat as well as sequence data of the 5’ end of the LSU from the type specimens. The anatomical features of the tetrasporangial conceptacle that were considered collectively diagnostic for species identification are: the mean diameter of the tetrasporangial conceptacles, the length of the pore-canal in the tetrasporangial conceptacles (with the number of cells in roof filaments), the occurrence of a depression at the top of the conceptacle roof in relation to the pore opening, and the number of cells from the floor of the tetrasporangial conceptacle chamber to the thallus surface. This study provides evidence of a previously unsuspected diversity within the Red Sea and NW Indian Ocean species of Lithophyllum.


Introduction
The coralline flora of the Indian Ocean is insufficiently known. A survey of the literature available for the area revealed that the subfamily Lithophylloideae was represented by 19 species and infraspecific taxa, and that 8 of them were in need of a revision in a modern taxonomic context. In particular, only Lithophyllum kotschyanum Unger, and its infraspecific taxa L. kotschyanum forma affine and L. kotschyanum forma subplicatum, together with L. orbiculatum and L. okamurae, were listed from the NW Indian Ocean (including Red Sea, Gulf of Aden, Arabian Sea, and Persian Gulf, in the boreal emisphere, and toward the east to include the western Indian coasts; Table 1).
In the framework of a large-scale investigation on Red Sea and Indian Ocean corallines, we collected some common Lithophyllum plants, bearing abundant trichocytes (Basso et al. 2014). In order to correctly identify the species, we provided morphological and molecular data (the 5'end of the nuclear LSU) to compare the recently revised type material of L. kotschyanum Unger, and the related infraspecific taxa, namely L. kotschyanum forma affine, L. kotschyanum forma madagascarense, L. kotschyanum forma subplicatum and L. kotschyanum forma subreduncum, housed at TRH (abbreviations following Holmgren continuously updated). The objective of this study is to use both morphological and molecular data to assess species delineation within this taxon. This study contributes to the discussion on coralline biodiversity in the north-western Indian Ocean and Red Sea, focusing on the genus Lithophyllum.  Silva et al. 1996;Womersley 1996;Maneveldt et al. 2008;Basso et al. 2014. An asterisk before the binomial indicates those species that should belong to the genus Titanoderma, based on Bailey (1999), and Schneider & Wynne (2007). In the column of the Indian Ocean distribution, the type locality is italicized.
Unchecked reports of the Lithophyllum species discussed here are in square brackets. Note that only L. kotschyanum, its infraspecific taxa L. kotschyanum f. affine and L. kotschyanum f. subplicatum, L. orbiculatum and L. okamurae were reported from the NW Indian Ocean before this study. Indian Ocean Lithophyllum species Indian Ocean distribution modern account

Material and Methods
Plants were collected at several localities along the Yemeni coast of the Southern Red Sea, Gulf of Aden, Arabian Sea and Persian Gulf. Corallines were air dried prior transport to the Milan facilities and the selected type material was deposited in TRH.
Air-dried specimens for scanning electron microscopy (SEM) were fractured and mounted on aluminium stubs with silver glue. The samples were gold coated and examined in a SEM Vega Tescan TS5136XM at 20 kV at the Milano-Bicocca facilities.
For light microscopy (LM), pieces of thallus were decalcified with Tellyesniczkyi's solution (Bressan 1974) for 12-48 h, washed in distilled water, dehydrated through a graded ethanol series and embedded in methacrylate resin (Technovit 7100, Heraeus Kulzer, Wehrheim, Germany). Preparation of histological slides of serial sections follows Basso & Rodondi (2006). Permanent slides for LM were examined and photographed with a Leica DMRB photomicroscope.
The type material of L. madagascarense is a very small fragment that would have been completely destroyed for any analysis. Therefore no investigation was performed on L. madagascarense, following the recommendations of the TRH curator. Anatomical terminology follows Adey & Adey (1973) and Woelkerling (1988), and that for growth forms follows Woelkerling et al. (1993). Conceptacle measurements follow the system of Adey & Adey (1973) and were made directly from the SEM or under LM using a calibrated eyepiece micrometer. Cell measurements follow Basso et al. (2004).
For molecular analyses, specimens were sampled under the dissecting scope using a driller. Total DNA was extracted, using a DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany), according to the manufacturer's instructions and negative controls with no tissue were included. Two reverse primers were designed to amplify with the forward primer T01N (Harper & Saunders 2001), small fragments of ca. 200 and 300bp of the 5' end of the nuclear LSU (28S): TR191 (5'CCACGGTACRCAWTTCCATG3') and TR273 (5'TCTCAAGCTACCCGACTC3') respectively. Purification and sequencing reactions were performed by Macrogen. Forward and reverse electropherograms were edited and assembled with the software Codoncode (Dedham, MA). Distance analysis was conducted on the LSU sequences from the type specimens to assess their divergence and Maximum likelihood analysis was conducted on a selection of taxa from GenBank and on the two recently collected type specimens to assess their phylogenetic relationship.
HABITAT AND PHENOLOGY: plants attached to corals, in shallow water (depth of collection unknown). GEOGRAPHIC DISTRIBUTION: Lithophyllum affine occurs in the southern Red Sea (Massawa, Eritrea). The occurrence of L. affine outside this area is unknown.
MATERIAL EXAMINED: Pacific Ocean, Samoa Islands, the lectotype TRH A20-1286. HABITAT AND PHENOLOGY: unknown. GEOGRAPHIC DISTRIBUTION: L. subplicatum is known only from the type locality, Samoa Islands. Other entries need verification (Table 1).
HABIT AND VEGETATIVE STRUCTURE: Plants non-endophytic, with fruticose growth-form. The long protuberances are cylindrical and dichotomously branched, diverging, about 3-4 mm in diameter, flattened and anastomosing at their tips. The plant is 7 cm high and 11 cm in diameter (Fig. 6C).
REPRODUCTION: Uniporate tetrasporangial conceptacle chambers protruding over the surrounding thallus surface, becoming buried in the thallus, 180-250 μm in diameter and 60-90 μm high, with pore-canal 45-60 μm long (  HABIT AND VEGETATIVE STRUCTURE: The coralline may be attached on corals or other biogenic substrate or form unattached nodules (rhodoliths). Thallus non-endophytic, encrusting, lumpy to fruticose. The protuberances, up to about 10 mm long and about 2-5 mm wide, have a smooth surface, are cylindrical or compressed, branched, sometimes fused and apically enlarged and flattened (Fig. 11A).
A ML phylogenetic analysis inferred from psbA (Fig. 13) revealed that the two last species (L. socotraense and L. yemenense) were only distantly related to the specimens so far sequenced for this gene and available in GenBank. Sample DB569, slide DB569-10y5.

Discussion
Trichocytes occur in all the Lithophyllum species dealt with in this paper (Basso et al. 2014). After some nomenclatural changes (details in Woelkerling et al. 2005), Lithophyllum kaiseri Heydrich (Heydrich) was treated by Foslie (1909) as heterotypic synonym of L. kotschyanum (Woelkerling et al. 2005). The syntype of Lithothamnion kaiseri Heydrich (TRH, A20-1264) includes two specimens and two slides. Foslie annotated that the largest specimen was not confidently identified, probably exchanged with another box (Woelkerling et al. 2005). For this reason we selected the smallest fragment as the lectotype of Lithophyllum kaiseri. Lemoine (1965) already questioned the validity of the synonymy of L. kaiseri with L. kotschyanum, and the results of our morphological analyses support the separation of the two species. In particular, the morphological characters that can be comparatively used to identify L. kaiseri are: the smaller conceptacles almost flush with the surrounding thallus surface, the conceptacle floor placed 11 cells from thallus surface (16-17 in L. kotschyanum), and only 4-5 cells in roof filaments (7 in L. kotschyanum, Tab. 2).
In his protologue of Lithothamnion affine, Foslie (1897) described the new species with details concerning its habit: plant attached and growing on corals, surrounding the coral branches, or rhodoliths "fastened to smaller stones …forms roundish or somewhat lobed balls on the bottom, 3-5 cm in diameter…". Later (1898) Foslie transferred Lithothamnion affine into Lithophyllum, without change in rank, and finally (Foslie 1909;Prinz 1929) reduced it to the rank of form as Lithophyllum kotschyanum f. affine (Woelkerling 1993;Woelkerling et al. 2005). He also described the two forms complanata and tuberosa on the basis of their growth form and shape of protuberances. The form complanata corresponds to the lectotype (= Lithophyllum affine f. affine, Woelkerling et al. 2005).
Our results, however, point to a separation of the two species, since L. affine Foslie (Foslie) is morphologically separated from the holotype of L. kotschyanum on the basis of its smaller tetrasporangial conceptacles with a pore opening on an apical depression of the conceptacle roof, the conceptacle floor placed 11 cells from thallus surface (16-17 in L. kotschyanum), and only 5-6 cells in roof filaments (7 in L. kotschyanum, Tab. 2; Fig. 1).
L. kotschyanum Unger, L. kaiseri (Heydrich) Heydrich and L. affine (Foslie) Foslie have similar and variable growth-form, although the integration of morphological and molecular results support their separation ( Fig. 1;  Tab. 2). Morphologically, the mean diameter of the tetrasporangial conceptacles, the length of the pore-canal in the tetrasporangial conceptacles (with the number of cells in roof filaments), the occurrence of a depression at the top of the conceptacle roof and the number of cells from the floor of the tetrasporangial conceptacle chamber to the thallus surface are the characters that appear collectively diagnostic for their segregation (Tab. 2 and dichotomous key). In this respect, it is worth mentioning that counting the cells at the periphery of the tetrasporangial conceptacle chamber is a very delicate operation, whose precision is easily compromised by the imperfect orientation of the thallus. Therefore we recommend to use this character with caution and, when possible, in combination with other diagnostic characters.
The holotype of Lithophyllum subreduncum Foslie 1901b (TRH, A20-1291) has growth form and microscopical anatomy very similar to that of the holotype of L. kotschyanum Unger, including the trichocyte occurrence and morphology (Fig. 9, Tab. 2). This observation explains Foslie's conclusions, who reduced it to the rank of form after eight years from the description of Lithophyllum subreduncum as a new species, under the combination Lithophyllum kotschyanum Unger forma subreduncum Foslie (Foslie 1909;Woelkerling et al. 2005). Our analyses demonstrate that the two species differ from L. kotschyanum Unger on the basis of LSU divergence and microscopic anatomy ( Fig. 1; Tabs 1-2). The comparison of the infraspecific variability of fresh collections from the type localities is a task beyond the aim of the present study. Pending further studies, we decide to follow a conservative approach by keeping the original combination for L. subreduncum Foslie 1901b, and considering L. kotschyanum f. subplicatum as a separate entity, under the new status and combination L. subplicatum (Foslie).
The reverse primers designed in this study successfully amplified the 5'end of the LSU in herbarium specimens older than one century. The strategy to amplify short fragments to overcome the issue of degraded DNA in old specimens was appropriate and opens new perspectives to rationalize taxonomic studies of coralline algae. Amplified fragments displayed divergence up to 5.1% between L. affine and L. kaiseri, L. kotschyanum, L. subreduncum and L. socotraense. It is noteworthy to stress that L. subreduncum and L. socotraense have identical sequences, nonetheless, the nuclear LSU has been shown to sometimes lack of variation among closely related species (e.g. Le Gall & Saunders 2010). Therefore, different nuclear LSU strongly supports species separation, and on the contrary, the apparent conspecificity based on this poorly variable LSU fragment appears questionable. In this case, morphology supports species separation: L. subreduncum differs from L. socotraense in having much larger and higher tetra-bisporangial conceptacle chamber, a longer pore canal and 15-16 cells from conceptacle floor instead of 9-11 (Tab. 2), and, moreover, the two specimens have been collected in different geographic regions. We have no data about the natural variability of L. subreduncum at its type locality (Hawaii), and the state of conservation of the type material does not allow to achieve further molecular results, nevertheless the decision to consider L. subreduncum and L. socotraense as separate species is presently the most conservative. The anatomical features of the tetrasporangial conceptacle that are here considered collectively diagnostic for species identification in Lithophyllum are: the mean diameter of the tetrasporangial conceptacles, the length of the porecanal in the tetrasporangial conceptacles (with the number of cells in roof filaments), the occurrence of a depression at the top of the conceptacle roof and the number of cells from the floor of the tetrasporangial conceptacle chamber to the thallus surface (dichotomous key).
We demonstrate here a previously unsuspected diversity within the Red Sea and NW Indian Ocean species of Lithophyllum, that increase in number from 3 (L. kotschyanum, L. orbiculatum and L. okamurae) to 7 (with the addition of L. affine, L. kaiseri, L. socotraense and L. yemenense). The occurrence of L. subplicatum and L. subreduncum outside their type localities in the Pacific Ocean requires further confirmation. On the basis of literature data (Silva et al. 1996;Guiry & Guiry 2015), more than one hundred species of Lithophyllum are presently considered taxonomically accepted, although many of them, including L. orbiculatum and L. okamurae, lack a modern account or need confirmation of their geographic distribution. The complete revision of this material is beyond the aim of this work, and constitutes a challenge for future studies.