DNA barcoding of the German green supralittoral zone indicates the distribution and phenotypic plasticity of Blidingia species and reveals Blidingia cornuta sp. nov.

In temperate and subarctic regions of the Northern Hemisphere, green algae of the genus Blidingia are a substantial and environment-shaping component of the upper and mid-supralittoral zones. However, taxonomic knowledge on these important green algae is still sparse. In the present study, the molecular diversity and distribution of Blidingia species in the German State of Schleswig-Holstein was examined for the first time, including Baltic Sea and Wadden Sea coasts and the off-shore island of Helgoland (Heligoland). In total, three entities were delimited by DNA barcoding, and their respective distributions were verified (in decreasing order of abundance: Blidingia marginata , Blidingia cornuta sp. nov. and Blidingia minima ). Our molecular data revealed strong taxonomic discrepancies with historical species concepts, which were mainly based on morphological and ontogenetic characters. Using a combination of molecular, morphological and ontogenetic approaches, we were able to disentangle previous misidentifications of B . minima and demonstrate that the distribution of B . minima is more restricted than expected within the examined area. Blidingia minima , the type of the genus name Blidingia , is epitypified within this study by material collected at the type locality Helgoland. In contrast with B . minima , B . marginata shows a higher phenotypic plasticity and is more widely distributed in the study area than previously assumed. The third entity, Blidingia cornuta sp. nov., is clearly delimited from other described Blidingia species, due to unique characters in its ontogenetic development and morphology as well as by its tufA and rbcL sequences. Supporting found in the at


■ INTRODUCTION
Green algae of the genus Blidingia Kylin are a substantial component of the upper and mid-supralittoral zones of the Northern Hemisphere, where they can be found as dense mats on various natural and artificial surfaces and additionally as epiphytes on other macrophytobenthic organisms. Blidingia spp. can withstand extreme conditions such as periodic and sustained desiccation, heat waves or frost periods with snow cover, and marine flooding as well as extended freshwater immersion. Thus, representatives of the genus Blidingia often indicate the transition between the marine or estuarine zone and the terrestrial zone. Blidingia species shape these particular environments and give them a unique and important ecological character by providing habitats for small invertebrates. However, in-depth understanding of the genetic species diversity within the genus and hence about geographic species distribution is still sparse. The recognition of Blidingia spp. is largely based on morphological characters of mature thalli and ontogenetic stages, while molecular knowledge is limited. Kylin (1949) proposed the new genus Blidingia for Enteromorpha minima Nägeli ex Kütz., based on observations made by Bliding (1938). One of the main characteristics that distinguishes Blidingia from the closely related and morphologically similar genus Enteromorpha Link (nowadays Ulva L.) is its small cells, with a diameter less than 10 μm (Kylin, 1949). Both genera include species with monostromatic, tubular thalli; however, their ontogenetic development was also found to differ significantly (Kylin, 1949). The motile swarmers of Blidingia have no eyespot, in contrast to the motile swarmers of Ulva. The settled Blidingia swarmer grows into an elongated tube that incorporates the intracellular spore contents, and the empty spore sleeve is separated by a transverse membrane. A prostrate disc develops, and from the expanded centre of this partly bi-layered disc, a monostromatic tube begins to emerge (Bliding, 1938;Kylin, 1949). While several studies supported these ontogenetic findings (Dangeard, 1961;Gayral, 1967), others emphasized variations in the ontogenetic development between different Blidingia species (Kornmann & Sahling, 1978;Tatewaki & Iima, 1984;Iima, 1989), particularly during spore sleeve formation (Bliding, 1963;Kornmann & Sahling, 1978). Bliding (1963) stated, "the germinating tube of the swarmer is mostly divided in a basal empty-cell and an upper cell containing all the cytoplasm", suggesting that differences in spore sleeve formation have been observed. Detailed observations of early ontogenetic developmental patterns on spore sleeve formation were made by Kornmann & Sahling (1978), who focused on the different Blidingia species from Helgoland (Heligoland). Whereas half of the observed entities (B. minima (Nägeli ex Kütz.) Kylin, B. chadefaudii (Feldmann) Bliding) cut off an empty spore sleeve (called embryospore), this pattern was not observed for B. subsalsa (Kjellm.) Kornmann & Sahling ex Scagel & al. and B. marginata (J.Agardh) P.J.L.Dang. ex Bliding. Thus, it was suggested that early ontogenetic stages were suitable criteria for species delimitation (Kornmann & Sahling, 1978).
Three different ontogenetic developmental patterns of prostrate discs were observed in studies focused on the sexual reproduction and early development of Blidinga minima from Japan (Tatewaki & Iima, 1984;Iima, 1989). The authors distinguished compact discs with short cells (D-type), open discs with longer cells (F-type) and an intermediate disc type (Mtype) (Iima, 1989). Notably, European specimens that exhibited thallus discs similar to the F-type were assigned to a newly differentiated species, namely B. chadefaudii (Bliding, 1938;Chadefaud, 1957;Kornmann & Sahling, 1978). However, due to the interfertility of specimens forming the different disc-types identified by Iima (1989), it was suggested that B. chadefaudii should be considered as a variant of B. minima (Woolcott & al., 2000). Those authors undertook a molecular reassessment of the taxonomic affiliation of individuals exhibiting different disc-morphotypes based on nuclear rDNA ITS sequences. By showing that the allegedly ontogenetic criteria that were used for species delimitation encompass an integrated continuum of variation, together with the fact that molecular analysis and interbreeding-experiments did not demonstrate a species boundary between Japanese B. minima and B. chadefaudii, the authors concluded that the two taxa are conspecific. Thus, the observations of Woolcott & al. (2000) blur the boundary between B. minima and B. chadefaudii and underline that morphological characters within most Ulvales are highly variable and that molecular sequencing is needed to delimit species boundaries.
Currently  (Guiry & Guiry, 2020). However, molecular knowledge of these species is sparse and most have primarily been distinguished by morphological and ontogenetic traits.
The present study aimed to examine the diversity of Blidingia species in the northern German State of Schleswig-Holstein.
Schleswig-Holstein's coastline is of limited length, but is structurally very diverse, including a south-east section of the fully marine North Sea and a south-west section of the brackish Baltic Sea, as well as various estuaries and lagoons. The area also includes the North Sea island of Helgoland, the focus of phycological research in the mid-19th century (Reinke, 1889) and among the best-studied seaweed habitats in Europe (Bartsch & Kuhlenkamp, 2000). Long-term observations of the benthic flora of Helgoland have shown that the establishment of artificial substrata resulted in an increased abundance of Blidingia spp. (Bartsch & Kuhlenkamp, 2000). Subsequent to C.W. Nägeli collecting the holotype of B. minima on Helgoland around the mid-19th century (historically: Enteromorpha minima, suppl. Fig. S1), the first recordings of other Blidingia species for the island were made by Wollny (1881). More recently, Kornmann & Sahling (1978) documented the Blidingia species of Helgoland in a detailed synopsis, focusing on their developmental differences. The study emphasized the presence of four Blidingia species (B. chadefaudii, B. marginata, B. minima, B. subsalsa). The authors observed severaland often vaguemicro-morphological differences between the four species, but also one commonly shared macro-morphological character: All species were characterized as unbranched tubes (Kornmann & Sahling, 1978).
According to species inventories, only B. minima and B. marginata have been recorded from Schleswig-Holstein's mainland coasts (Schories & al., 2009). However, due to the aforementioned taxonomic difficulties, historical records of Blidingia species require confirmation, and the first molecular assessment of the distribution of Ulvales and Ulotrichales in the region (Steinhagen & al., 2019a,b) has indicated that the present distribution of Blidingia spp. may not accord with recent inventories. However, there has not been a molecular characterization of the northern European Blidingia species or a review of their allegedly significant identification criteria undertaken until now.
A crucial point for molecular studies using DNA barcoding is always the choice of suitable marker genes that are on the one hand side variable enough to delimit species but also stable within the respective species group. A marker gene that has been widely used within molecular studies of green algae is the coding gene for the ribulose-bisphosphate carboxylase large subunit rbcL (Saunders & Kucera, 2010). Later, the plastid encoded tufA gene was promoted as the most suitable marker for delimitation of green algae, providing the highest amplification success and the largest barcode gaps for green macroalgae (Saunders & Kucera, 2010). Genetic databases such as GenBank include numerous sequences for both markers. Reference sequences, however, need to be selected with caution, as type specimens of Ulvales have rarely been successfully sequenced.
By combining tufA and rbcL sequencing with morphological and ontogenetic observations, we have assessed the diversity of Blidingia species within Schleswig-Holstein and have recognized an undescribed new species that is relatively abundant in the area. In addition, our study highlights past difficulties in distinguishing Blidingia species.

■ MATERIALS AND METHODS
Field collection and sample preparation. -Sites along the North Sea and Baltic Sea coasts of Schleswig-Holsteinincluding the island of Helgolandwere repeatedly visited in the years 2014-2017 (see also Steinhagen & al., 2019a). In 2016, sampling also covered the heavily trafficked Kiel Canal, which connects both sea areas (Fig. 1, sites 14-16, see also Steinhagen & al., 2019b). Upper littoral and supralittoral zones were checked for macroalgal growth with a focus on freshwater inflows (e.g., drainages, river inflows, beach showers). Several sites were re-visited in the years 2018 and 2019, to verify the presence of populations and obtain material for cultivation (Table 1). Altogether, Blidingia spp. were observed at 19 of the visited sites ( Fig. 1; Table 1).
Specimens were collected, placed into sealed plastic bags, and stored on ice until further processing in the lab. Most samples were preserved as herbarium vouchers (see Table 1) and lodged in the Natural History Museum Denmark, Copenhagen (C). A subsample was divided, with part of it stored at 4 C or −20 C for subsequent morphological observation and part of it stored in a microreaction tube at −80 C for genomic DNA extraction. Light microscopy observations were undertaken, and microscopic documentation was carried out on the remaining part, using a digital camera (Nikon DS-Vil) attached to a light microscope (Nikon Eclipse TS 100). Salinity was measured using a WTW portable conductivity meter (Xylem Analytics, Weilheim, Germany).
In addition to the field collected samples, herbarium specimens in the Herbarium of the Helgoland Biological Station of the Alfred Wegener Institute (BRM) were included in our analyses (barcode numbers: BRM007967 and BRM008079; see also Table 1). Several additional herbarium specimens were investigated, including the type specimen of B. minima (Naturalis Biodiversity Center, Leiden, Netherlands [L], barcode L 0054691; suppl. Fig. S1). However, they yielded insufficient amounts of DNA due to the vouchers being pre-treated with fixation solutions, or the amplification of marker genes was impossible due to impurities of the herbarium vouchers (diatoms, multi-algal vouchers, etc.). Thus, these herbarium vouchers were excluded from the molecular analyses of this study.
DNA extraction, amplification and sequencing. -Genomic DNA was isolated from the lyophilized algal tissue with an Invisorb Spin Plant Mini Kit (Stratec, Birkenfeld, Germany) following the manufacturer's protocol. Extracted DNA was stored at −80 C and used for amplification of the rbcL and tufA genes. PCR amplifications of the rbcL gene used the primer pairs rbcLstart and R750, as well as F650 and rbcLend (Shimada & al., 2003). The PCR reactions were performed as follows: 94 C for 1 min; 35 cycles at 94 C for 30 s, at 56.3 C for 30 s, and at 72 C for 1 min; and a final extension step at 72 C for 7 min. PCR amplification of the tufA gene followed the detailed description of Steinhagen & al. (2019a). PCR products were purified using the QUIAquick PCR Purification Kit (Quiagen, Hilden, Germany). Subsequent sequencing of the purified amplicons was provided by GATC Biotech (Konstanz, Germany). Forward and reverse sequence reads were assembled to produce contigs in Sequencher (v.4.1.4, GeneCodes, Ann Arbor, Michigan, U.S.A.), and a multiple sequence alignment was constructed for each gene region using MAFFT v.7.402 (Katoh & al., 2002) (for Alignments, see supplementary Appendices S1 and S2). Sequences obtained in this study are publicly available in GenBank (for accession numbers, see Table 1).
Phylogenetic analyses. -RbcL and tufA sequences were analysed in separate datasets. Newly generated sequences Fig. 1. Sites of the Blidingia samples in northern Germany processed in this study. Overview map about northern Germany with numbered sampling sites at the Wadden Sea (no. 1-10), on Helgoland (no. 11-13), within the Kiel Canal (no. 14-16) and in the Baltic Sea (no. 17-19).  Version of Record were aligned with reference sequences downloaded from GenBank and used for further phylogenetic analysis. Particularly, sequences of specimens of the genera Ulva L., Kornmannia (Kjellm.) Bliding and Monostroma Thuret were included to assess relationships with closely related taxa, whereas sequences of Protomonostroma undulatum (Wittr.) K.L.Vinogr. (rbcL: HQ603387; tufA: HQ619275, MH475501) were chosen as outgroups. Preference for reference sequence selection was given to peer-reviewed sequences. The models that best fit our data were found under the Akaike information criterion by employing MrModeltest v.2.2. (Nylander, 2004). For both datasets, the optimal substitution model was determined and found to be GTR+Γ+I. Maximum likelihood (ML) analyses were then carried out using RAxML v.8 (Stamatakis, 2014), employing the chosen substitution model with 1000 bootstrap replicates for each alignment.

Version of Record
Cultivation. -Complete thalli of selected mature Blidingia specimens were washed thoroughly and repeatedly with sterile seawater (salinity of the respective collection site) to remove dirt and adhering impurities and were isolated into cultures. Clean thalli were transferred into polystyrene 24well plates and were incubated in sterile artificial seawater adjusted to the salinity of the respective sites at 15 C under a photon flux density of 40-70 μmol m −2 s −1 and a 17 : 7 h light : dark photo regime. To prevent the growth of diatoms, 1 mg l −1 GeO 2 was added. The thalli were examined daily for sporulation events. After sporulation had taken place, adult thalli were removed from the wells, and spore development was observed with an inverted microscope (Nikon Eclipse TS 100) and photographed (Nikon DS-Vil).

■ RESULTS
Phylogeny. -The phylogenetic analyses performed on datasets of the rbcL and tufA markers resulted in comparable and nearly identical results, and almost equivalent evolutionary relationships of both investigated marker genes were discovered (Fig. 2). The rbcL alignment consisted of a total of 695 positions (suppl. Appendix S1), whereas the tufA gene dataset was 771 basepairs long (suppl. Appendix S2).
The node separating the genus Blidingia from closely related and outgroup taxa received full bootstrap support for both marker genes and unequivocably confirmed the taxonomic position of Blidingia as a separate genus. Whereas the tufA dataset reveals Ulva as sister clade to Blidingia (bootstrap support: 99), the rbcL phylogram displays a reference sequence of Kornmannia leptoderma (AF499677) as the closest relative of Blidingia before Ulva. Both analyses resolve the German Blidingia samples from the examined area in three clades (Blidingia marginata, Blidingia sp. 1, Blidingia sp. 2) with high (>85) to full bootstrap support (Fig. 2). The clades delimiting B. marginata showed low intraspecific genetic variability (tufA: 0%-0.3%; rbcL: 0%-0.2%) (see also Table 2) and could be resolved with reference sequences (rbcL: HQ603379, Canada; tufA: HQ610237, Canada). A reference sequence that seemed to be incorrectly assigned to B. minima (MG721599) clustered within the clade representing B. marginata.
The phylograms of both the tufA (KT290281, HQ610239) and rbcL genes (AF499676, MF90430, MF90429) include clades of downloaded reference sequences from GenBank that were supposed to represent Blidingia minima (Fig. 2). However, the topology of these clades and thus their placement within the trees is not congruent. One of the specimens assigned to the clade representing B. minima within the tufA phylogram originates from a site 30 km to the east in the neighbouring German State of Mecklenburg-Vorpommern, Wohlenberg, which was observed by us in a previous study (Steinhagen & al., 2019a).
Notably, GenBank entries of the rbcL and tufA genes of Blidingia minima showed significant differences in their nucleotide reads (indicating different species are combined in GenBank under this species name). Since most of the uploaded sequences contain several ambiguous bases and are rather short (250-500 bp), we decided to exclude them from the displayed results. However, within this study, we were not able to validate a genotype representing any of the genotypes associated with B. minima in GenBank, not even at the type locality of B. minima on Helgoland. Intra-and interspecific divergence values of the here investigated Blidingia entities are summarized and listed in Table 2.
Both herbarium specimens from Helgoland (Table 1) showed no agreement between their previous identification based on morphological traits and their molecular identification by DNA barcoding. Within our phylogenetic analysis of the tufA gene, voucher BRM008079 (morphologically identified as Blidingia minima, GenBank accession no.: MT076212) clustered within the clade representing Blidingia sp. 2, and voucher BRM007967 (morphological identity B. marginata, GenBank accession no.: MT076211) was assigned to the clade of Ulva intestinalis (Fig. 2).
Morphological characterization, habitat and distribution. -In the following section the Blidingia species present in the study area of northern Germany are described in more detail. Macro-and micromorphological observations including respective ontogenetic features, as well as ecological and distribution information, are presented for each species: Blidingia marginata (J.Agardh) P.J.L.Dang. ex  Habitat, seasonality and distribution. -Blidingia marginata is the most common Blidingia species within the study area and has the widest distribution. It is abundant on Baltic Sea and Wadden Sea coasts and at Helgoland (Table 1). This species can be found in remote as well as anthropogenically strongly impacted habitats (see also Steinhagen & al., 2019a,b) and inhabits fully marine and brackish water ecosystems. However, B. marginata was seldomly observed in water bodies Fig. 2. Comparative maximum likelihood phylograms of tufA and rbcL sequences from taxa of Blidingia from northern Germany. The grey-shaded boxes indicate clades that were present in the study area and connects the respective clades of species in both phylogenetic analyses for direct comparison. Dashed boxes indicate the ambiguous use of the historic construct of "Blidingia minima" and highlight reference sequences identified as such. Numbers at nodes indicate bootstrap values. Poorly supported nodes (<70% bootstrap support) are not labelled. Branch lengths are proportional to sequence divergence. # symbol marks herbarium samples (see also Table 1).
Version of Record below 5 PSU, and it was not detected in the Kiel Canal. As is typical for the genus, B. marginata was observed growing as dense turf mats in the upper and middle intertidal zone ( Fig. 3A-F). Adults of B. marginata can be found throughout the year, but their peak abundance was observed during early and mid-summer (May-July). In late summer and fall (August-October), dense stands began to bleach and populations shrank in size, so that in winter and spring only a few smaller individuals were encountered. Individuals were always found attached to the substratum and only observed detached after extreme weather events. Especially in the North Sea areas, B. marginata was the most abundant alga in the intertidal zone. It occupied a variety of natural and artificial hard substrata (stones, cobble, breakwaters, wooden piles, etc.) and was often found growing epiphytically on other macrophytes (e.g., Fucus spp.) and higher plants (Phragmites sp.) in the intertidal (Fig. 3B,C). The same distribution patterns were encountered in the Baltic Sea. Here, however, B. marginata was found in mixed stands with Ulva intestinalis, and in some cases, young individuals of U. intestinalis were difficult to distinguish from B. marginata by morphological characters alone. Individuals of B. marginata could resist strong UV-radiation and desiccation in summer as well as snow cover and long frost periods in winter (Fig. 3F).
Morphology. -Individuals of Blidingia marginata exhibited all the morphological features described for this species (Bliding, 1963), but also some differences. The long (1-12 cm) and narrow thallus was tubular and usually formed dense mats ( Fig. 3D-F). Broader thalli were often wrinkled and twisted and resembled Ulva intestinalis (Fig. 3G-I). Contrary to the observations of Bliding (1963), who stated that specimens of B. marginata exhibited rare small proliferations, 80% of the individuals of the investigated material had microscopically visible branches or branchlet-like appendages (Fig. 3H). These branchlets were often uni-or biseriate with a single apical cell. Macroscopic branching in the middle or apical thallus parts was rarely observed, whereas some specimens had macroscopic branches in the rhizoidal zone. In young, narrow thalli, the cells were arranged in distinct rows (Fig. 3J,K); however, mature or broader thalli exhibited only short cell rows or no cell organization. Cells were quadratic to rectangular, sometimes of amorphous shape, 4-10 μm long and 2-9 μm wide in surface view. The chloroplast filled the whole cell with one central pyrenoid (Fig. 3J,K).
Ontogeny. -As also described by Kornmann & Sahling (1978), the quadriflagellate spore settled on the substratum and immediately developed a germination tube without cutting off an empty cell (Fig. 4A,B). The first mitotic cell divisions resulted in the formation of a prostrate disc (Fig. 4C,D), which then in most of the cases became 2-layered (Fig. 4E). The tubus typically started to develop from lateral initial cells and was rarely observed to develop form the disc's centre.
Blidingia sp. 1 Habitat and distribution. -Specimens of this species were also found on Baltic Sea and Wadden Sea coasts and on Helgoland (Table 1). However, dense, turf-like populations of Blidingia sp. 1 were not as frequent and abundant as those of B. marginata, and they were more clearly restricted to the upper supralittoral zone. Notably, most of the observed populations of Blidingia sp. 1 grew in the direct vicinity of freshwater inflows (drain pipes, beach showers, stream run-off, etc.; Fig. 5A). As a consequenceand in contrast with B. marginata and Blidingia sp. 2specimens of Blidingia sp. 1 were rarely found desiccated during the summer months, despite their location in the upper supralittoral. When freshwater inflows were more located towards the medio-or infralittoral, Ulva intestinalis was the prevailing species, and Blidingia sp. 1 was absent. The species was also absent from more inland freshwater inflows that had no direct connection to the sea. Blidingia sp. 1 was present throughout all seasons; however, it was more common during July to August. Values for interspecific comparison of "Blidingia minima" were calculated on the respective historically mis-identified clades represented in Fig. 2 (dashed box tufA; rbcL lower dashed box). Therefore, they are framed by double quotation marks within this table. Morphology. -The morphology of Blidingia sp. 1 shows similarities with the description of Bliding's (1963) "Blidingia minima var. ramifera" (not validly published, type not indicated), which was later "raised to species rank" by Garbary & Barkhouse (1987; as "Blidingia ramifera"). "Blidingia ramifera" is currently regarded as a synonym of B. marginata (Guiry & Guiry, 2019). However, there are distinct differences: The small, tubular and branched thalli were compact and reached 1-10 mm in length (rarely taller) (Fig. 5B,E) whereas thalli of "Blidingia ramifera" were distinctly larger (Bliding, 1963;Garbary & Barkhouse, 1987). The width of the thallus increased as it proceeded from the rhizoidal zone to the tip (0.08-0.8 mm wide). Branches were antler shaped, uni-to multiseriate and present across the whole thallus (Fig. 5B,C). Unbranched individuals were rarely encountered. In addition to the clearly formed branches, spine-like microscopic appendages (10-60 μm) were frequently found across the whole thallus (Fig. 5E,G). Branches and microscopic appendices were blunt-ended. Thalli grew in tufts that formed dense stands, and several individuals were connected by their rhizoidal zones. Cells formed clear, longitudinal rows ( Fig. 5F) that sometimes blurred in broader thallus areas of the apical region (Fig. 5G), whereas no distinct cell arrangement was observed in mature individuals of "B. ramifera" (Bliding, 1963;Garbary & Barkhouse, 1987). Individuals with unordered cell arrangements were observed infrequently. The cells were quadratic, rectangular, often polygonal with blunt to rounded corners and 3-8 μm long and 2-8 μm broad. The chloroplast filled the cell or was rarely parietal, with 1 (rarely 2) central pyrenoid(s) (Fig. 5F,H). Ontogeny.
-After the quadriflagellate spore had settled, a germination-tube began to form (Fig. 4F). The cell content migrated through the germination-tube and formed a germinating cell, detaching the initial spore sleeve and in most cases also parts of the germination tube (Fig. 4G,H). By mitotic cell divisions, a monostromatic, relatively open disc developed, and its cells were not as dense as in other entities (Fig. 4I,J). After becoming distromatic, the disc bulged out, and an erect tube formed.
Molecular analysis.
-The sequence divergence of the clade representing Blidingia sp. 1 from other species within the genus, in combination with morphological differences, indicates that Blidingia sp. 1 is genetically and Fig. 4. Ontogenetic development of German Blidingia spp. After the spore of B. marginata had settled, a germination-tube began to form (A). No cutting-off of an empty cell was observed (B). A prostrate disc began to form (C-E) with densly packed cells in its centre (E). The early development of Blidingia sp. 1 started with the formation of a germination-tube after the spore had settled (F). The spores cell content migrated through the germination tube and formed a germinating cell while detaching the empty initial spore sleeve (G & H). Open discs were formed (I & J). The settled spore of Blidingia sp. 2 formed a small germination tube (K) that then divided into an empty initial cell and a cell containing the cell content (L). A monostromatic disc with dense cell arrangement formed (M) that in its centre became distromatic (N & O)  Version of Record intact branch apices, attached by rhizoids to substratum, 1-10 mm (mean ± 6 mm; rarely >1 cm) long, 0.08-0.8 mm (mean ± 0.3 mm; rarely >1 mm) broad. Main axis increasing in breadth from 10-35 μm (rhizoidal zone) to 0.3 mm (middle and apical thallus). Middle thallus often with spine-like appendices (20-35 μm in length), these appendices shorter than real branches. Cells in surface view in clear longitudinal rows in basal and middle thallus parts, short longitudinal rows in apical regions, square, rectangular or polygonal with blunt to round corners, most commonly 3-8 μm long and 2-8 μm broad. The chloroplast is cell-filling (rarely parietal covering most of the cell wall) and cells containing 1 (rarely 2) central pyrenoid(s). Reproduction by quadriflagellate spores. From the spore a germination-tube arises, and the cell contents migrate to a germinating cell, detaching the initial spore sleeve and germination-tube. By mitotic cell divisions a monostromatic disc develops, and after becoming distromatic, the disc bulges out and an erect tube is formed.
Etymology. -The species epithet cornuta ("horned" in Latin) refers to the antler-like morphology of branches.
GenBank accessions. -MH475459 represents the sequence of the tufA marker gene, and MN258049 is the respective rbcL sequence.
Type locality. -Brunsbüttel harbour, Schleswig-Holstein, Germany (N 53.889 E 9.101133 ). Thalli were growing as dense turf in the high intertidal zone directly under a drainage with constant freshwater seepage from land on a bulkhead (Fig. 5A). The site is part of the Elbe estuary, close to the outlet of the Kiel Canal.

Blidingia sp. 2
Habitat and distribution. -This entity was only observed on Helgoland and in the northeastern Wadden Sea (Table 1) and was not present in the Baltic Sea. It inhabited the upper supralittoral zone and was found growing as turfs, but more often it was observed as small patches on stones, concrete, wooden piles or other hard substrates (Fig. 6A,B). When growing epiphytic on macrophytobenthic species (e.g., Fucus spp.), Blidingia sp. 2 did not cover the host like B. marginata. Instead, single individuals were found to be scattered across the host plants.
The thalli of Blidingia sp. 2 were mostly only few millimetres long (rarely taller than 1 cm) and 50-300 μm wide (single individuals had broader thalli up to 700 μm) (Fig. 6C-E). No branches in the middle or apical thallus parts were observed, however the base sometimes exhibited branches (Fig. 6D,E). Thalli were most often compressed, but inflated individuals were also present. Whereas cells form clear and distinct longitudinal rows in the basal thallus parts (Fig. 6F), the arrangement of cells is less organised in the middle and apical thallus parts (Fig. 6G). Cells were of various shapes, quadratic to polygonal with rounded corners, 4-8 μm long and 4-6 μm wide in surface view. No thickened cell walls, nor any lamellar internal structures were observed. The chloroplast was parietal or filled the cell, with one central pyrenoid (Fig. 6H).
Ontogeny. -The main ontogenetic patterns discovered in this study agree with the developmental descriptions of Blidingia chadefaudii (rather than B. minima) made by Kornmann & Sahling (1978). From the attached quadriflagellate spore, a small germination tube was formed (Fig. 4K). The tube divided into an empty initial cell and an upper cell containing the cell contents (Fig. 4L). A monostromatic disc with a dense cell arrangement was formed (Fig. 4M,N), which became distromatic in its centre (Fig. 4N,O) and gave rise to an erect tube.

■ DISCUSSION
We here provide a strongly revised picture of the diversity, distribution and morphological variability of Blidingia spp. in Schleswig-Holstein, Germany, a study area that includes coastal sections of two important ecosystems in northern Europe, the North and Baltic Seas (Fig. 1). In total, 30 populations of Blidingia spp. at 19 sites were analysed to cover the full diversity in the study area. However, whereas older studies and recent species inventories mention four Blidingia species as abundantly present along the coastlines of Schleswig- Holstein (Kornmann & Sahling, 1978;Schories & al., 2009), we only discovered three. Our findings suggest that this discrepancy results from the exclusive use of morphological traits and ontogenetic developmental characters as identification criteria in the past. These criteria were probably invalid in part, due to morphological variability within species. That morphological identification criteria for Blidingia species are not in concomitance with more clear-cut molecular identification criteria should be discussed in detail.
Of the four Blidingia species listed for Schleswig-Holstein, only B. marginata was frequently observed. Blidingia marginata was found to be the dominant species within the intertidal zone, and it was frequently observed at Baltic Sea and Wadden Sea coasts and on Helgoland. The species is represented by a well-delimited clade with low intraspecific variation (0%-0.3% tufA and 0%-0.2% rbcL) within both provided phylogenetic trees (Fig. 2) and was well resolved to the species level by including several peer-reviewed reference sequences (GenBank accessions HQ610237 tufA and HQ603379 rbcL). Its closest relative was Blidingia sp. 2, and both entities had a sequence dissimilarity of 8.6%-9.4% within the tufA gene and 3.2%-3.7% in their rbcL sequences (Table 2). However, the combination of morphological and molecular techniques revealed that B. marginata exhibited a larger morphological variability than expected (Fig. 3).
Individuals of Blidingia marginata exhibited all the morphological features described for this species (Bliding, 1963), and our ontogenetic observations (Fig. 4) are in accordance with previous findings (Bliding, 1963;Kornmann & Sahling, 1978). The tubular thalli where either thin and elongated or had a corkscrew-like morphology, and their cells were arranged in clear longitudinal rows. It should be noted that mature thalli are morphologically similar to small Ulva intestinalis and could easily be confused with that species. Correspondingly, a herbarium sample from Helgoland that was in concordance with the morphological characters described for B. marginata (Biological Station Helgoland Herbarium BRM007967) was identified as U. intestinalis by DNA barcoding (Fig. 2).
Contrary to the observations of Bliding (1963), who stated that specimens exhibited rare small proliferations, more than 80% of the individuals of the examined material had microscopic branches or branchlet-like appendages (Fig. 3H). Bliding (1963) assigned specimens with branchlet-like structures to Blidingia marginata subsp. subsalsa (Kjellm.) Bliding. Later, Scagel & al. (1989) raised B. marginata subsp. subsalsa to species level, as already suggested by Kornmann & Sahling (1978). Our observations indicate that B. marginata can exhibit various morphologies, including those assigned to B. subsalsa; however, such morphological differences were not reflected by molecular differences of the marker genes tufA or rbcL, and no delimitations of the different morphotypes within our phylogenetic analyses were observed (Fig. 2). Kornmann & Sahling (1978) observed that cultivated swarmers, released by material identified as Blidingia subsalsa, did not grow into the "naturally looking wildtype forms" of B. subsalsa and that their first-generation offspring were macro-morphologically rather identical with individuals of B. marginata. Based on our results we conclude that the morphological spectrum of B. marginata is broader than previously expected, as it includes the morphologies assigned to B. subsalsa.
Based on literature (Kornmann & Sahling, 1978;Pankow, 1990;Schories & al., 2009), Blidingia minima was expected to have the widest distribution in northern Germany. However, our study revealed that this species is restricted to Helgoland and some other North Sea locations. Past records, in particular those from Baltic Sea locations, may represent misidentifications due to flawed historic species concepts.
This view is strongly supported by our analysis of available barcoding sequences. Screening of databases such as GenBank for sequences of Blidingia minima provides several hits for tufA and rbcL sequences of various length. However, within our phylogenetic analyses (Fig. 2), different sequences from around the globe and identified as B. minima fell in several well delimited clusters (Fig. 2, dashed boxes) that are not necessarily closely related or even cluster with B. marginata. This unequivocally confirms that the recent species concept of B. minima is ambiguous, combines genetically distinct entities, and needs clarification.
Even though sampling sites were chosen in a way that distances between the sites did not exceed 25 km (see also Steinhagen & al., 2019a,b), no genotypes in accordance with any GenBank entries for Blidingia minima were observed in the investigated area. One of the reference sequences allegedly representing B. minima (KT290281) originates from Wohlenberg, 30 km to the east in the neighbouring German State of Mecklenburg-Vorpommern and was observed in a previous study. However, sequences from the type locality of B. minima, Helgolandseveral hundred kilometres away from Wohlenberg -, were so far missing, which gains importance when we consider that this island was extensively studied within our survey (Fig. 1). A frequently found entity on Helgoland that could not be resolved to species level due to the absence of any similar GenBank entries was Blidingia sp. 2 (Fig. 2). Blidingia sp. 2 delimits in a unique clade and is the next closest relative to B. marginata (Fig. 2, Table 2).
Based on the results obtained within our study we conclude: (1) The type locality of Blidingia minima (as (Kylin, 1949;Bliding, 1963;Kornmann & Sahling, 1978). As another unique trait, thalli exhibited a specific antler-like branching pattern, together with a cell arrangement in longitudinal rows that has not previously been observed in any other Blidingia species. It should be mentioned, however, that "Blidingia minima var. ramifera" (Bliding, 1963) exhibits some similaritiesbut also significant differenceswith the newly described B. cornuta. The thalli described by Bliding (1963) were distinctly taller in size and reached a length of up to 50 cm, whereas thalli of B. cornuta are less than 1 cm in length (rarely taller). The branches of "B. minima var. ramifera" were described as elongate, while the branching pattern in B. cornuta is antler-like, and the respective branches are compact rather than elongate. Bliding (1963) reported that cells of "B. minima var. ramifera" were predominantly unarranged or only arranged in short rows, whereas most specimens of B. cornuta have clearly visible longitudinal cell rows that are evident throughout the thallus. Garbary & Barkhouse (1987) treated "B. minima subsp. ramifera" at species level. Their description of "B. ramifera" encompassed the same striking differences from B. cornuta as previously described for "B. minima subsp. ramifera". Hence, branched specimens of "B. ramifera" were regarded as morphotypes of B. marginata (Burrows, 1991;Guiry & Guiry, 2019). However, the genetic distinction of B. cornuta seems indeed reflected in a unique branching pattern. In this light, the status of "B. ramifera" should be reevaluated based upon support with genetic markers. Several unidentified Blidingia sequences are available via GenBank that hint a hidden diversity that is still to be discovered. Our study indicates once again that it can be rewarding to reassess the exact taxonomic relationships of allegedly well-known species groups within the Ulvophyceae based on molecular markers and subsequent phylogenetic techniques to reveal exact species relationships, potential morphological plasticity, and species-specific ecological traits.
We can conclude that the historical species concepts for several Blidingia spp. are flawed and problematic for species occurring in northern Germany. Misinterpretation of phenotypic plasticity in mature thalli, and to some degree also in ontogenetic developmental stages, has led to misidentifications in the past, and species delimitation based on morphological traits is often impossible. Thus, our findings support the use of molecular methods for correct and clear species identification and devalue the use of morphological characters alone.