Taxon‐rich phylogeny and taxonomy of the genus Phacus (Euglenida) based on morphological and molecular data

Morphological and molecular features were analyzed for a species of Phacus to better understand the phylogenetic relationships among them and establish the taxonomy. Phylogenetic analyses were based on nSSU rDNA and the research resulted in 55 new sequences. The study included species available in algal collections and those isolated directly from the environment in Poland and the Czech Republic. As a result, the obtained phylogenetic tree of Phacus includes 50 species, out of which 7 are represented on a tree for the first time (Phacus anacoelus, P. anomalus, P. curvicauda, P. elegans, P. lismorensis, P. minutus and P. stokesii) and many have been taxonomically verified. For all verified species, diagnostic descriptions were amended, the naming was reordered and epitypes were designated.

Phacus was described in the 19 th century (Dujardin 1841) and currently includes approximately 300 species names, of which 174 are taxonomically accepted (http://www.algaebase.org; Guiry and Guiry 2020). The number of species changes as taxonomic verifications (based on morphological and DNA sequence data) are ongoing (Kosmala et al. 2007, Karnkowska-Ishikawa et al. 2010, Linton et al. 2010, Bennett and Triemer 2012, Kim and Shin 2014, Łukomska-Kowalczyk et al. 2015. As a result, our understanding of the phylogenetic relationships among them is also constantly changing. The research cited above did not validate Phacus' separation into two subgenera, Chlorophacus (green forms) and Hyalophacus (colorless) as suggested by Pochmann (1942), or into four sections for the green forms (Proterophacuscells flattened and leaf-shaped; Pleuraspisspherical or flat with ribbing instead of striae; Acanthochlorisslightly flattened with papillae on the surface; Kampylopterround, triangular in crosssection). First, it was shown that species from the Pleuraspis section are in fact representatives of Monomorphina (Marin et al. 2003). Later, when both Lepocinclis salina and Euglena limnophila, neither of which have flattened cells, were included in Phacus, the description of the genus underwent changes (Linton et al. 2010). Recently, Phacus horridus (=P. hispidulus from the Acanthochloris section) was reclassified as a member of Lepocinclis Triemer 2012, Łukomska-Kowalczyk et al. 2020). As a result of all previous morphological-molecular studies, descriptions were emended, epitypes were designated, many species were renamed, and nine new taxa were described , Łukomska-Kowalczyk et al. 2015. Currently, Phacus is characterized by semi-rigid to rigid cells, usually laterally compressed (with the exception of P. limnophilus and P. salinus), in most cases leaf-shaped, sometimes twisted. They possess numerous chloroplasts of uniform shape (numerous, small, disc-shaped, parietal, without pyrenoids). All species (with the exception of P. salinus) have dimorphic in size paramylon grainsin most cases the large grains are plate-or ring-shaped. Colourless forms are known (e.g., Phacus ocellatus; Marin et al. 2003). Phacus is currently classified in the Phacaceae family, together with the representatives of the Lepocinclis and Discoplastis genera .
The most recently published phylogenetic trees of Phacus include 43 species , Łukomska-Kowalczyk et al. 2015, some of which have been misidentified. We conducted this study to increase the number of species and strains representing Phacus in the phylogenetic trees and to perform comparative morphological and DNA sequence studies on new strains (=isolates). This will allow an estimation of morphological and genetic diversity and verification of morphological diagnostic features for particular taxa (well-established clades) on a molecular phylogenetic tree; (b) the reconstruction of phylogenetic relationships and (c) taxonomic verification, amending diagnoses and designating epitypes for well-distinguished taxa.

MATERIAL AND METHODS
Sampling and morphological study. During nine seasons (2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019), plankton samples were collected from 37 eutrophic water bodies located in Poland and one located in the Czech Republic using a plankton net with a mesh size of 10 µm (Fig. 1). Samples were screened in terms of their species diversity, and then a number of cells (7-300; see Table S1) exhibiting the same morphology were isolated from the sample using a micromanipulator (MM-89 Narishige) with a micropipette installed on a Nikon Ni-U microscope (Nikon, Tokyo, Japan). Morphological studies (descriptions and measurements) and documentation (photographs and video clips) of the isolated cells (isolates) of the Phacus morphotypes and those from six strains from algae collections (ACOI 1753, ACOI 1754, ACOI 1755, ACOI 1757, ACOI 3237, AICB 324) were executed with a NIKON Eclipse E-600 microscope with differential interference contrast, equipped with the NIS Elements Br 3.1 software (Nikon) for image processing and recording. Photographs (and video clips) were taken using a NIKON DX-1200 digital camera connected to the microscope. The NIS Elements Br measurement program was used for morphometric studies; three parameters were measured for the cells of each isolate (strain)cell length, cell width, and tail length (which was defined as the hyaline projection); measurements were conducted from photographs of the isolates (strains). The data was analysed using the R v.3.2.0 software (R Development Core Team, 2008); means and standard deviations are given in Table 1. The isolates (i.e., morphologically identical cells) were transferred through multiple drops of sterile media (to clean the sample), and kept frozen at À80°C until DNA extraction.
DNA isolation, amplification, and sequencing. Isolation of total DNA from samples (environmental samples and laboratory cultures of strains) was performed as described previously (Zakry s et al. 2002). An additional step of whole genome amplification was carried out in the case of environmental samples (according to Bennett and Triemer 2012). PCR amplification of the nSSU coding region, purification, and sequencing of the PCR products were performed by standard methods as previously described (Zakry s et al. 2002). To obtain some sequences, the nested PCR method was used as described previously (Łukomska-Kowalczyk et al. 2020).
Sequence accession numbers, alignment, and sequence analyses. Fifty-five new sequences (6 from strains and 49 from environmental isolates) used in the present study were submitted to GenBank with the following accession numbers: MN149548 -MN149602 and MN326770. Information about the accession numbers for the nSSU rDNA sequences reported here and those used for phylogenetic analyses (136 in total, of which 129 for Phacus) are shown in Table S1. Identical sequences from different strains/isolates were represented as single terminals in the phylogenetic analyses. The nSSU rDNA sequences were aligned using the MAFFT v.7.2 software (Katoh and Standley 2013) with E-INS-I strategy. Alignment was inspected by eye and corrected if necessary. Regions of doubtful homology between sites were removed from the alignments using TrimAl 1.2 with the option "automat-ed1" (Capella-Gutierrez et al. 2009). After trimming, 2,157 out of 4663 positions were left. Sequence diversity of nSSU rDNA was calculated using the Mega X software (Kumar et al. 2018) as pairwise distance based only on unambiguously aligned positions. The alignment and corresponding phylogenetic tree have been submitted to TreeBase (Study Accession URL: http://purl.org/ phylo/treebase/phylows/study/TB2:S25743).
Phylogenetic analyses: The model of sequence evolution was selected using jModeltest 2.1.7 (Darriba et al. 2012) and the GTR+I+G model was chosen based on AIC, BIC or DT criteria (with -lnL = 48697.4; Lanave et al. 1984, Tavare 1986, Rodriguez et al. 1990 and was used to calculate maximumlikelihood (ML) and Bayesian trees. The ML tree was inferred using RAxML 8.2.11 (Stamatakis 2014) using 1,000 rapid bootstrap inferences.
Bayesian inference (BI) with default priors was performed in MrBayes 3.2.6 software (Ronquist and Huelsenbeck 2003). A gamma correction with eight categories and proportion of invariable sites were used. Two independent runs with four Markov chains were performed. In each run, the chains lasted for 10,000,000 generations and trees were sampled every 1,000 generations. The first 25% of trees were discarded as burn-in. Convergence among runs was assumed as the average standard deviation of split frequencies was below 0.01. The trees were visualized using FigTree v.1.4.2 (available at http://tree.bio.ed.ac.uk/software).

RESULTS
Phylogenetic analyses and morphological characteristics. Phylogenetic trees obtained by Bayesian (not shown herein) and maximum likelihood methods have a very similar topology, with 11 main clades and two branches with single sequences (Fig. 2, Fig. S1  , sister to all other Phacus sequences, is not strongly supported, but consists of two maximally supported subclades. It includes sequences of P. arnoldii and P. limnophilus originated from both laboratory cultures and environmental isolates. The genetic variability of nSSU rDNA does not exceed 4% for either species (Table S2). Representatives of both species morphologically resemble members of genus Lepocinclisthe three-ridged cell shape of P. arnoldii makes it similar to L. tripteris, while the unflattened cells of P. limnophilus are typical for the majority of the representatives of the Lepocinclis genus (for details see Discussion). The clade corresponds to clade G (Kim and Shin 2014) and I6 .
The next clade (B), complementing clade F in Kim and Shin (2014) and clade I5 in Kim et al. (2015), includes three strongly supported subclades corresponding to the species: Phacus pleuronectes, P. minutus, and P. acuminatus. The cells of all three species are leaf-like (flat), widely oval with a short, pointy tail. Sequences of P. pleuronectes are located in two groups: one with sequences of only environmental isolates and the second with all six strain sequences (out of which three are identical) and only one environmental sequence, sister to them. All isolates are characterized by a trapezoidal shape of cells in comparison to the ovoid cells of the cultivated strains. The genetic variability in the species does not exceed 3% and is lower than 0.5% among the strain sequences (Table S2). Phacus acuminatus and P. minutus are morphologically almost undistinguishable (see Taxonomic revisions), but they do not form a common clade and the genetic variability between them varies between 8.1% and 9.6% and is lower than 3% and 0.1% within the species respectively (Table S2).
Strongly supported clade C (1/100), that corresponds to clade D1 in Kim and Shin (2014) and to one subclade of the I3 clade in Kim et al. (2015), consists of sequences of ten species characterized by large cells (>75 µm) with long tails. In the clade there are three groups of sequences: (1) representatives of Phacus ranula, P. convexus, P. tortus, P. longicauda, P. cordatus, P. circumflexus with one sequence of the strain of unknown morphology (Phacus sp. Burni081809), (2) representatives of P. cristatus, P. crassus, P. helikoides, and (3) as sister clade to all other sequences, the subclade including two sequences of Polish environmental isolates of P. elegansthe species that is represented on the phylogenetic tree for the first time.
Clade E (1/100) includes four maximally supported subclades that correspond to four species: Phacus raciborskii and P. mariae branching together with one sequence of a strain of unknown morphology (Phacus sp. ASW 08004) and a pair of sister species: P. caudatus and P. manginii. Species located in this clade are easily distinguishable on the basis of morphology and their common diagnostic feature is flat cells with a conspicuous dorsal crest, which elongates into a sharp hyaline tail. The P. manginii subclade consists exclusively of sequences originated from environmental isolates (from Poland and South Korea), and the other subclades include sequences of both strains and isolates. Genetic diversity in P. caudatus and P. manginii does not exceed 1.8% and in the P. raciborskii clade there are five sequences with their genetic diversity below 0.3% and one (ACOI 1758) that is 4.5-5.0% divergent from all other (Table S2). The E clade corresponds to clade E in Kim and Shin 2014 and to clade I4 in Kim et al. 2015. Strongly supported clades F, G, and H form a common clade in the maximum likelihood analyses only and in the bayesian analyses, relationships between them are not defined.
Clade F, absent in Kim and Shin (2014) and corresponding to a fragment of the I3 clade in Kim et al. (2015), consists of three identical sequences of Phacus gigas (one of the strain MSU and two Polish isolates), one sequence of a strain of unknown morphology (Phacus sp. ACOI 1138) and a subclade of three P. hamatus sequences. The P. gigas and P. hamatus individuals are flat, wide-oval in general overview, and terminate with a sharp hyaline tail.
Clade G, not present in either Kim and Shin (2014) or in Kim et al. (2015), groups sequences of Phacus salinus: five from environmental isolates and one from a cultivated strain. Despite all of them being morphologically indistinguishable (spherical, egg-like cells with both ends widely rounded), their genetic variability is very diverse (up to 11.3%) and the sequences vary considerably in length. The longest fragment of 18S rDNA belongs to the isolate UW1948Wos (3560 bp), and the sequences of other P. salinus isolates and strains are also fairly long: 2,124 bp for strain SAG 1244-3, 2,151 bp for UW1970Chod, 2,778 bp for UW2392IK2. Usually the length of this fragment varies between 1,900 and 2,050 bp in most other Phacus species, and exceeds 2,100 bp only for P. helikoides, P. crassus, P. cristatus, and P. elegans.
Clade H, corresponding to the B clade in Kim and Shin (2014) and to clade I2 in Kim et al. (2015), consists of sequences of five species: Phacus ankylonoton, P. tenuis, P. applanatus, P. hamelii, and P. segretii, which can be easily distinguished based on morphology and molecular differences (see Figs. 3, q and r, 4f, and also fig. 1, a and b in Kosmala et al. 2007).
Two sequences of environmental isolates of Phacus lismorensis form a strongly supported clade I, however, that does not define their position. The species has a very characteristic morphology (cells bent like a bow; Fig. 4n) and its sequences were not hitherto present on any phylogenetic tree.
Clade J consists of two subclades corresponding to two sister species: Phacus orbicularis and P. paraorbicularis, which are characterized by wide, ovoid cells ending with a sharp curved tail; both had been previously taxonomically verified (Kosmala et al. 2007, Kim andShin 2014). Two sequences of Polish isolates of P. paraorbicularis and one of P. orbicularis are included in the subclades. Variability in the whole clade does not exceed 1.5% (Table S2). The clade corresponds to clade C in Kim and Shin (2014) and to one subclade of the I3 clade in Kim et al. (2015).
Maximally supported clade K, corresponding to clade D2 in Kim and Shin (2014) and to one subclade of the I3 clade in Kim et al. (2015), includes sequences of four species. The genetic diversity in that clade is lower than 5% and in most cases (in spite of the variability of the sequence of Phacus curvicauda UW2163Wos) does not exceed 0.9% (Table S2). All representatives of the clade are characterized by thick cells with well-developed ventral and dorsal sides, ending with a short tail curved toward the dorsal side. However, the four species are easily recognizable based on such features as the shape of the large paramylon grains and their location in the cell. Sister clades with the sequences of P. anacoelus and P. anomalus are maximally supported and the other two subclades have lower support -P. curvicauda: 1/96 and P. alatus: 0.82/96. All sequences in clade K (aside from the three of P. alatus) stem from environmental isolates from Poland.
The single sequence of Phacus stokesii has a sister position to clade K, although it is not strongly supported (0.98/73). It represents one out of two (along with P. segretii) tailless species (Fig. 4d).
Taxonomic revisions. Phylogenetic and morphological analyses and a review of the literature enabled the taxonomic identification of almost all clades present on the phylogenetic tree. For new or misidentified clades, as well as those represented by single sequences a taxonomic revision had been conducted. Its basis was to define a species as a group of singular morphotypes, which create a wellsupported clade on the phylogenetic tree. For such clades (morphologically well-distinguished taxa) diagnostic descriptions were emended, epitypes designated and the nomenclature has been reordered.
Due to the lack of morphological data, the taxonomic affiliation of 5 strains (Jigok090112 as Phacus ranula on the phylogeny tree; Leynes012810F as P. mariae; Burni081809, ASW 08004 and ACOI1138 as Phacus sp.) was not verified.
Given that the isolated cells were destroyed for DNA extraction, the photographs are designated as epitypes (see International Code of Nomenclature for algae, fungi, and plants [Shenzhen Code]; chapter II, section 2, article 9.9; Turland et al. 2018).

1142
Comments: Following the nomenclatural priority rule, this morphological form (cells small, almost spherical, flat, and terminate with a short, sharp, straight or bent sideways, and wedge-shaped tail) has been assigned the name Phacus acuminatus. The individual seen in Stokes' drawing (1885a, fig. 1) corresponds with the aforementioned characteristic, similarly to the cells from the UTEX 1317 strain (chosen as the representative strain), from which the epitype originates. However, due to the high morphological similarity to P. minutus, only molecular identification allows the differentiation of the two species. In this situation, designation of epitypes for both species seems justified (see Comments for P. minutus). The taxa that constitute P. acuminatus synonyms are those whose cells have a morphology similar to P. acuminatus and without proper diagnostic traits in their original descriptions.
Basionym  56)a thick cell (only slightly flattened, not leaf-like flat) with a deep furrow and parietal paramylon grainsresembles the Polish representatives of P. alatus. Nonetheless, both this cell and others shown in the remaining drawings display a high similarity to P. curvicauda and P. anomalus due to the presence of a wide furrow dividing the cells, which are slightly twisted longitudinally, in halves. The morphological research presented herein demonstrates that the large paramylon grain location and shape are good diagnostic features for discriminating those three species (for more details see Discussion). The epitype designation for all three species will allow their proper identification. The taxa that constitute synonyms of P. alatus are those whose slightly flattened cells posses a furrow and two large, ring-or shieldlike paramylon grains located parietally.
Phacus anacoelus A.Stokes 1885b: 19, fig. 2  (Fig. 3l-o) Emended diagnosis: Cells almost spherical (42-57 9 25-40 µm), slightly flattened, with the ventral and dorsal sides clearly visible; at the back cells terminate with a sharp, hyaline tail (10-12 µm long) bent toward the dorsal side; two high crests run along each side (four crests in sum), and as the cells are spirally twisted, when mobile it seems that the crests are evenly distributed; two large, spherical paramylon grains are located in the center of the cell.
Holotype  fig. 2) also highly resembles the species described later (P. asymmetricus, P. quinquemarginatus) (for more details see Discussion). Due to the aforementioned, emending the diagnostic description and designating an epitype will allow its proper identification.
Lectotype: designated herein, Pochmann 1942, fig. 37d Epitype: Figure 3q designated herein that supports the lectotype (Pochmann 1942, fig. 37d) Representative DNA sequence: GenBank KF744064 Representative strain: CCAC 0043 PHYLOGENY AND TAXONOMY OF PHACUS Type locality: pond in small town Wolice, Eastern Poland (now Ukraine) Comments: Due to its size and the general cell shape, P. ankylonoton is similar to P. caudatus, the only difference being the width of the oval cell, which in cross-section is triangular. This particular trait is represented only in two of Pochmann's drawings ( fig. 37d and e), which are in fact an original drawing by Dre_ zepolski (Dre_ zepolski 1925, fig. 112 two cells). It is Dre_ zepolski's drawing that Pochmann refers to when describing the new taxonone of them has been designated herein as the lectotype ( fig. 37d). The remaining individuals in Pochmann's drawings (37a and b) resemble more P. caudatus due to the elongated oval-like cell shape. The designation of an epitype will allow the proper identification of P. ankylonoton.
Phacus anomalus Fritsch & Rich 1929: 73, figs. 24h-n ( Fig. 3g and h) Emended diagnosis: Cells wide-egg or trapezoidshaped (30-50 9 20-35 µm) in general overview, thick, with a furrow running down the entire length of the cell and dividing it into halves, which are slightly twisted longitudinally; two large, spherical paramylon grains are located in the widest and thickest bottom part of the cell, which elongates into a sharp tail (on average 6-8 µm long) bent toward the dorsal side.
Lectotype: designated herein, Fritsch and Rich 1929, fig. 24h Epitype: Figure 3g designated herein that supports the lectotype (Fritsch and Rich 1929, fig. 24h) Representative DNA sequence: GenBank MN149560 Type locality: Griqualand Westthe interior of South Africa Representative locality: Freshwater, Ciela z dz village (51°42 0 50.9" N, 20°20 0 44.7" E) Heterotypic  Popova and Safonova 1976: 58, pl. 12, figs. 1-22. Comments: In most of the drawings by Fritsch and Rich (1929) the large, spherical paramylon grains are positioned centrally in the cell, which causes P. anomalus to be practically indistinguishable from P. curvicauda (both have thick cells with a furrow and the tail bent toward the dorsal side). Meanwhile, the study presented herein proved that in P. anomalus the large paramylon grains are located in the widest and thickest, bottom part of the cell. The lectotype designated hereinthe cell visible in the drawing 24 h (Fritsch and Rich 1929) has the most antapical paramylon grains (located the closest to the posterior of the cell). Taxa that constitute P. anomalus synonyms are those whose cell morphology corresponds to the emended diagnostic description (for more details see Discussion). Pochmann 1942: 152, fig. 42 (Fig. 3r) Emended diagnosis: Cells elongated oval-like (40 9 20 µm on average), flat, terminate with a rather long, straight tail (6-7 µm on average).

Phacus applanatus
Holotype : Pochmann 1942, fig. 42 Epitype: Figure 3r designated herein that supports the holotype (Pochmann 1942, fig. 42) Representative DNA sequence: GenBank EU624031 Representative strain: CCAC 2604 B Type locality: Sandberg pond near Teplitz-Sch€ onau, Germany Comments: The morphological study of the CCAC 2604B strain (isolated in the Netherlands, channel in Leiden) shows that its representatives have cells terminated with a tail that is twice as long (6-7 µm) in comparison with the individual in the drawing by Pochmann (1942, fig. 42).
Comments: In the original description, "round, flat cells with a high, S-shaped crest" and "wide, sparsely arranged periplast strips" are mentionedthanks to these features the Polish strains could be identified. As our research has shown, cells appear triangular with concave sides in cross section that gives the impression of three wings (ridges) and is a diagnostic trait of P. arnoldii. Meantime, in Svirenko's drawing (1915a, pl. 3, fig. 1) only a high crest is visible that does not looks like one of the three equally sized wings. Furthermore, we have shown that a more or less round shape of the cells has no diagnostic meaning, which is why P. arnoldii var. ovatus has been included as a synonym of P. arnoldii.
Holotype: H€ ubner 1886, fig. 5 Epitype: Figure 3t designated herein that supports the holotype (H€ ubner 1886, fig. 5 fig. 105. Comments: Phacus caudatus is very similar to P. manginii and P. ankylonoton due to the presence of a crest that elongates into the hyaline tail. Meanwhile, both the diagnostic description as well as H€ ubner's drawing of P. caudatus (H€ ubner 1886, fig. 5) does not contain any diagnostic features that would allow the distinguishment of P. caudatus from the two species described below (for more details see Discussion). Due to the aforementioned, the designation of an epitype seems justified. The taxa that constitute P. caudatus synonyms are those whose cell morphology corresponds to the emended diagnosis.
Phacus  (1942) is invalid, as a previously described taxon of the same name (though of a different morphology from P. longicauda subs. rotundus) exists (P. rotundus Brabez 1941). We propose P. convexus nomen novum for the morphological form known previously as P. longicauda subs. rotundus. The latin name convexus refers to the spoon-shaped (convex) cells.
The proposed lectotype, Stein's drawing (1878 (pl. 20, fig. 2), was chosen not only because it is one of the few Pochmann (1942, p. 201) referred to when describing the subspecies rotundus, but also mainly because Stein (1878)  Emended diagnosis: Cells wide-oval or wide eggshaped (22-37 9 16-27 µm), thick, divided into two longitudinally twisted parts by a furrow; terminate with a sharp, short hyaline tail bent toward the dorsal side (on average 2.5-5 µm); two large, spherical paramylon grains positioned centrally in the cell.
Epitype: Figure 3e designated herein that supports the lectotype (Svirenko 1915a, fig. 13 Comments: In Svirenko's drawings, the large paramylon grains are located centrally in the cells, which allowed us to identify the Polish strains of Phacus curvicauda. However, in cell shape, P. curvicauda is very similar to P. anomalus and P. alatus PHYLOGENY AND TAXONOMY OF PHACUS (they all have thick cells with a long furrow). The best diagnostic trait for differentiating the three species is cell shape, as well as the location and shape of the large paramylon grains (for more details see Discussion and Comments for P. anomalus and P. alatus). The drawing indicated as the lectotype (Svirenko 1915a, fig. 13) shows a long furrow and round paramylon grainsin the remaining images (figs. 14, 15, 16), the individuals either have short furrows, or ring-like paramylon grains. The designation of an epitype will allow the proper identification of P. curvicauda.
Holotype: Pochmann 1942, fig. 107. Epitype: Figure 4a designated herein that supports the holotype (Pochmann 1942, fig. 107) Representative DNA sequence: GenBank MN149571 Type locality: Southern Germany, peatbog Rauhen Wiese near B€ ohmenkirch (Schwarzwald) Representative locality: Freshwater, Urwitałt, pond 19 (53°50 0 38.0" N, 21°38 0 06.5" E) Comments: The Polish populations of this species have been identified based on the shape of the cells (elongated oval-like with a visibly asymmetrical anterior end, ending with a long hyaline tail). According to Pochmann, Phacus elegans can be distinguished from P. lismorensis based on "a more elongated cell form and a shorter tail." Furthermore, we have established that the presence of a crest that elongates into a long, straight hyaline tail is a diagnostic morphological feature; in P. lismorensis representatives the tail is ventrally bent (for more details see Discussion).
Holotype: Cunha 1913, pl.10, fig. 3 Epitype: Figure 4j designated herein that supports the holotype (Cunha 1913, pl.10, fig. 3) Representative DNA sequence: GenBank MN149572 Type locality: Manguinhos, Rio de Janeiro Representative locality: Freshwater, Oracze, farm pond (53°52 0 36.6" N, 22°20 0 41.2" E) Comments: Phacus gigas has large, flat, almost round cells, and the Polish populations have been identified based on these traits. The study presented herein shows that the diagnostic feature allowing the proper identification of P. gigas (and therefore making the differentiation between other similar species, e.g., P. hamatus, P. paraorbicularis, or P. orbicularis possible) is the posterior end asymmetry as well as the tail bent sideways. As the cell asymmetry is not mentioned either in the diagnostic description or in Cunha's drawing (1913, pl.10, fig. 3), we believe that the designation of an epitype is justified.
Epitype: Figure 4k designated herein that supports the lectotype (Pochmann 1942, fig. 86b) Representative DNA sequence: GenBank AJ532473 Representative strain: CCAC 2605B (=ASW 08032) Type locality: pond in small town Dobrostany, Eastern Poland (now Ukraine) Comments: The characteristic shape of the Phacus hamatus cells (spoon-shaped [convex]) visible in all of Pochmann's drawings (1942, fig. 86, a-f) allowed the identification of the species. On the other hand, the cell asymmetry visible in all drawings and stressed in the original diagnosis, is confusing, as our research has shown that it concerns only P. gigas. Phacus hamatus has symmetrical cells, but both species have so far been difficult to distinguish and often are confused (for more details see Discussion  fig. 1a.
Representative DNA sequence: GenBank DQ397673 Comment: As the epitypification statement published in Kosmala et al. 2007 (p. 1078)) does not include the phase "designated here" (or on equivalent), thus the nomenclatural act has not been effected in accordance with ICN Art. 7.11 (Turland et al. 2018 fig. 1b. Representative DNA sequence: GenBank FJ590503 Comment: As the epitypification statement published in Karnkowska-Ishikawa et al. 2010 (p. 178) does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act has not been effected in accordance with ICN Art. 7.11 (Turland et al. 2018) and is conducted herein.
Basionym  fig. 205) are confusing in terms of the number and location of the large paramylon grainsthe description mentions one grain positioned at the back of the cell (behind the nucleus) or two grains placed at the nucleus's level on both sides (the left and the right). The second option is depicted in Lemmermann's drawing (1913, fig. 205). Meanwhile the rod-like, large paramylon grains always exist as a pairone in front of the nucleus, and the other behind; there is no room in the narrow, cylindrical cells of Phacus limnophilus for the paramylon grains to be located on nucleus level. Many euglenid researchers, including ourselves, have noted this (Pringsheim 1956, Popova and Safonova 1976, Tell and Conforti 1986. Phacus lismorensis Playfair 1921: 125, pl. 5 fig. 14  (Fig. 4n) Emended diagnosis: Cells long ovate (70-134 9 24-40 µm), flat (no crest), bent like a bow; anterior end wide and rounded; posterior end gradually narrowed and terminated with a long, sharp, pointy tail (on average 40-50 µm long); the tail is set almost at a right angle and does not straighten even when the cells are immobile (not swimming).
Holotype: Playfair 1921, pl. 5 fig. 14. Epitype: Figure 4n designated herein that supports the holotype (Playfair 1921, pl. 5 fig. 14 fig. 30. Comments: Polish reresentatives of Phacus lismorensis have a long tail heavily bent toward the ventral side. In Playfair's drawing, it is only slightly bent which is why P. lismorensis might be mistaken with P. elegans, to which it is very similar both in cell shape and size (for more details see Discussion and Comments for P. elegans). The taxa that constitute P. lismorensis synonyms are those whose cells are long ovate, flat and terminate with a long tail heavily bent ventrally.
Phacus longicauda ( Type: (see Kim and Shin 2014), permanently preserved material from the strain Buan092609I deposited at the Chungnam National University, as number CNU 025428. figure 2D in Kim and Shin 2014 shows the type.
Holotype: Pochmann 1942, fig. 85. Epitype: Figure 3d designated herein that supports the holotype (Pochmann 1942, fig. 85) Representative DNA sequence: GenBank MN149551 Representative strain: ACOI 1755 Type locality: Australia, Sydney, botanical garden Comments: Strain ACOI 1755 has been identified as Phacus minutus due to its shape and cell size. Meanwhile, the morphological study presented has shown, that P. minutus has a tail bent toward the dorsal side that makes it different from P. acuminatus (bent sideways). Both species are very similar in size and shape, which is why the designation of their epitypes seems justified (for more details see Discussion).
Phacus orbicularis H€ ubner 1886: 5, fig. 1 (Fig. 3k Comments: Phacus heimii has been included as a synonym of P. orbicularis, as the sequence of the Polish isolate (UW2362Choc), the cells of which matched the morphological form described as P. heimii, was located in the orbicularis clade (for more details see Discussion). As the epitypification statement published in Kosmala et al. 2007Kosmala et al. (p. 1077 does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act has not been effected in accordance with ICN Art. 7.11 (Turland et al. 2018) and is conducted herein.
Representative DNA sequence: GenBank FJ590499 Comment: Because the epitypification statement published in Karnkowska-Ishikawa et al. 2010 (p. 177) does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act had been not effected in accordance with ICN Art. 7.11 (Turland et al. 2018) and that's why it's done now.
Representative DNA sequence: GenBank AJ532472 Comment: Because the epitypification statement published in Karnkowska-Ishikawa et al. 2010 (p. 177) does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act had been not effected in accordance with ICN Art. 7.11 (Turland et al. 2018) and that's why it's done now.
Representative DNA sequence: GenBank FJ590498 Comments: As the epitypification statement published in Karnkowska-Ishikawa et al. 2010 (p. 178) does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act has not been effected in accordance with ICN Art. 7.11 (Turland et al. 2018) and is conducted herein.
Representative DNA sequence: GenBank AF190815 Comment: As the epitypification statement published in Karnkowska-Ishikawa et al. 2010 (p. 178) does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act has not been effected in accordance with ICN Art. 7.11 (Turland et al. 2018) and is conducted herein.
Phacus raciborskii Drezepolski 1925: 234 and 266, fig. 113 (Fig. 4l and m) Emended diagnosis: Cells flat, cylindrical (35-60 9 8-15 µm), slightly U-bent and spirally twisted; anterior rounded, narrowing toward the end in a wedge-like manner and terminated with a sharp hyaline tail (on average 9-13 µm long); low crest running along the entire cell length; periplast longitudinally striated. Two large, ring-like paramylon grains with one placed in front of the nucleus, and the other behind it.
Representative DNA sequence: GenBank EU624028 Representative strain: SAG 1244-3 Comments: A new diagnostic traitthe lateral reservoir openinghas been included in the description. Moreover the measurements of the cells of the Polish populations have been taken into account, together with literature data. This seems justified, particularly that no such emendment had been done by Linton et al. (2010) when moving Phacus salinus from Lepocinclis to Phacus.
Representative DNA sequence:FJ597146 Comment: As the epitypification statement published in Karnkowska-Ishikawa et al. 2010 (p. 178) does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act has not been effected in accordance with ICN Art. 7.11 (Turland et al. 2018) Karnkowska-Ishikawa et al. 2010 (p. 179) does not include the phrase "designated here" (or an equivalent), thus the nomenclatural act has not been effected in accordance with ICN Art. 7.11 (Turland et al. 2018) and is conducted herein.
Holotype: Lemmermann 1910, fig. 9 Epitype: Figure 4d designated herein that supports the lectotype (Lemmermann 1910, fig. 9) Representative DNA sequence: GenBank MN149602 Type locality: North America, Teichen, ponds Comments: The morphology of the cells of Polish populations of Phacus stokesii (thick, widely rounded at both ends, without a tail) corresponds to the original diagnosis. The drawing by Lemmermann (1913, fig. 231), on the other hand, is confusing as it shows an almost round form, while in reality the shape of the cells of this species is wide-oval. The taxa that constitute P. stokesii synonyms have wide-oval cells (40-46 9 30-35 µm), thick and rounded at both ends (for more details see Discussion).
Lectotype: designated herein, Svirenko 1915b, fig. 17. Epitype: Figure 4f designated herein that supports the lectotype (Svirenko 1915b, fig. 17 fig. 107. Comments: According to Svirenko (1915b) P. tenuis differs from P. caudatus by having more oval cells and a lower crest (the latin name "tenuis" means flat), and this has been confirmed by our research. Meanwhile the individuals in Svirenko's (1915b) drawings: P. caudatus ( fig. 8) and P. tenuis (figs. 17, 18) are so similar, that the differentiation of the two species is PHYLOGENY AND TAXONOMY OF PHACUS almost impossible. figure 17 has been indicated as the lectotype, as it is in this drawing that the cell has the most oval shape and has the lowest crest.

DISCUSSION
On the oldest molecular phylogenetic trees, Phacus was recognized as a monophyletic member of the family Phacaceae ), but further investigation suggested that the genus was paraphyletic Shin 2014, Karnkowska et al. 2015). In those two studies Phacus limnophilus and Phacus arnoldii (=P. warszewiczii; Karnkowska et al. 2015) or only P. limnophilus (Kim and Shin 2014) sequences formed a sister clade to the clade grouping other sequences of Phacus and Lepocinclis. Later, the increased number of species and utilization of multi-marker analyses confirmed the monophyly of Phacus . Our research, based on nSSU rDNA only, but including more sequences and Phacus taxa (136 sequences, 50 species), yielded a phylogenetic tree with a topology similar to those published previously. Most clades present on our tree consisted of sequences that were also together on previous trees. This includes our clade D (P. oscillans group), that underwent taxonomic revisions earlier (Karnkowska-Ishikawa et al. 2010, Kim andShin 2014) and is always strongly supported , clade J with sequences of P. orbicularis and P. paraorbicularis , Kim and Shin 2014 and clade C (P. longicauda and P. helikoides group; Kim and Shin 2014, Łukomska-Kowalczyk et al. 2015 that was extended by adding sequences of P. elegans. Clades F, B, E, and H were also present on previous trees Shin 2014, Kim et al. 2015), but herein they were verified and new sequences were added.
Clade I (Phacus lismorensis) is represented for the first time, clade G (P. salinus) was previously represented only by one sequence, and clade K was represented at most by three sequences of only one species (Łukomska-Kowalczyk et al. 2015), now includes sequences of four species.
In the following part, the details concerning nomenclature and history of particular species and groups of species are discussed. Phacus acuminatus. Over 30 intraspecific taxa (subspecies, forms, and varieties) have been described for Phacus acuminatus (see: http://www.algaebase.org), mainly based on cell shape and size, tail length and shape as well as the number and morphology of the large paramylon grains. Many authors view such differences as intraspecific variability arising from either ontogenesis or environmental conditions (Huber-Pestalozzi 1955, Popova andSafonova 1976 among others). This is further confirmed by our research. In regard to the shape and size of the cells, the obtained results (on average 30 9 20 µm) were compatible with the rich literature (e.g., Popova and Safonova 1976: 25-33 9 19-22 µm; short tail, no fold; Table 1) and the observed differences in each strain were the result of ontogenesis the individuals tended to enlarge (particularly their width) right before cell division. Some dissimilarities were also found between strains, but they had no DNA sequence supportall strains occurred in a subclade of clade B (Fig. 2) with nSSU rDNA diversity below 3%. The UTEX 1317 strain (in the collection as P. brachykentron; Fig. 3a) was located among them. This morphological form, described by Pochmann (1942) as P. brachykentron, had been distinguished due to its slight cell asymmetry. For the aforementioned reasons, many taxa described in the literature as having a morphology similar to P. acuminatus and without proper diagnostic traits in their original descriptions, have been placed in synonymy.
Phacus alatus, P. anomalus, and P. curvicauda. These three species are closely related (clade K in Fig. 2) and morphologically similar. They all possess thick cells (not flat) divided into two parts by a furrowthe halves are also twisted in the longitudinal axis. Phacus alatus was the first to be described (Klebs 1883), when Klebs came to the conclusion that the morphological form identified by Stein (1878) as P. triqueter (pl. 19, figs. 55-57) represented a species new to science. In the original diagnosis, Klebs drew attention to the atypical, characteristic shape ("widened, wing-like sides of the cell which overlap; one from the ventral side, the other from the dorsal") and the lateral placement ("in both wings") of two large, discoidal paramylon grains. Since then, many taxa had been described at various taxonomical ranks (species, varieties and forms) of a similar morphology, the proper identification of which is practically impossible. It even came to the point, where Pochmann (1942) described P. platyaulax for the same exact morphological form previously described by Klebs as P. alatus, i.e., P. triqueter from Stein's original work (1878;see p. 165 in Pochmann 1942). As a result, based on the aforementioned reasons and the research presented herein, many of those forms have been placed in synonymy under P. alatus.
Two more species were described in the 20 th century (Phacus curvicauda and P. anomalus) that differ from P. alatus by having spherical large paramylon grains positioned in the center of the cell (not parietal). As the criteria for their distinguishment were ambiguous, researchers tended to interpret their taxonomic rank in various ways (e.g., see Pochmann 1942;Safonova 1976, Starmach 1983). The species can be easily morphologically identified (Fig. 3, e-j); on the phylogenetic tree there is a clade with sequences of P. anacoelus sister to P. anomalus. Thus, distinguishing three separate taxa at the rank of species is justified. The shape and placement of the large paramylon grains are decent diagnostic traits that allow correct identification.
All three species are considered common and cosmopolitan (e.g., Safonova 1976, Starmach 1983). In Poland, Phacus curvicauda and P. anomalus are commonplace and are often found in large densities. Phacus alatus, on the other hand, is found rarely and only in low densities (B. Zakry s, pers. obs.). One other species -P. drezepolski (Stawi nski 1969) had been described from Poland based on a single cellhowever, as it is not different from P. anomalus, it has been listed as one of its synonyms.
Phacus anacoelus. This euglenid of a particularly characteristic shape (almost spherical, slightly flattened, with 4 spiral ribs; Fig. 3, l-o) was described from the USA at the end of the 19 th century by Stokes (1885b). Forty years later, Sokoloff (1933) described a similar species from ponds in Mexico, but due to the asymmetry of the cell the species was named P. asymmetricus. Jahn and Shawhan (1942) described P. quinquemarginatus with 5 ribs from canals and ponds in Iowa (USA). All of the aforementioned species are of a similar size (on average around 40 9 35 µm; Table 1). Some authors of critical monographs recognized only P. anacoelus, deeming it a very rare species (Pochmann 1942;Popova and Safonova 1976). Phacus quinquemarginatus and P. asymmetricus had been reported from Argentina (Tell and Conforti 1986), while P. anacoelus had been recorded in Russia (Popova and Safonova 1976), Slovakia (Wołowski and Hind ak 2005) and Poland (Dre_ zepolski 1925). We found two populations in Mazovia (Regn ow and Izdebno Ko scielne 1; see Fig. 1 and Table S1), one of which, in a small field pond in Izdebno Ko scielne village, has been persisting for the past several years and appears regularly during summertime in high densities. The Polish populations have cells with bilateral symmetry visible in a cross section, but less so when observing mobile cells (Fig. 3, l-o). Thus, the term "asymmetrical" might be interpreted in various ways, and there is no basis to distinguish these species based on their imprecise, original diagnoses. However, we continue to recognize these three taxa pending the study of additional material. The sequences of P. anacoelus are located on the phylogenetic trees in a subclade within clade K (Fig. 2) along with P. curvicauda, P. anomalus and P. alatus, despite that "at first glance" these species appear morphologically very dissimilar (Fig. 3, e-j, l-o). However, they have a similar cell shape: slightly flattened, with well-developed ventral and dorsal sides, terminating with a sharp, hyaline tail bent toward the dorsal side.
Phacus ankylonoton. This morphological form was described from Poland by Dre_ zepolski (1925) as Phacus caudata var. polonica, and later was recognized as separate species (P. ankylonoton) by Pochmann (1942). It differs from P. caudatus in the width of the cell, which in cross-section is triangular (Fig. 3q, Table 1). The validity of distinguishing this species is further confirmed by two DNA sequences (strain CCAC 0043 from Germany and Mokpo033107E from South Korea), creating a common clade (H) on the phylogenetic tree at a distant position from the sequences of P. caudatus (clade E in Fig. 2). The species is rareit has been noted only a few times in Europe, Asia and South America (Shi et al. 1999, Popova and Safonova 1976, Starmach 1983, Tell and Conforti 1986. In Poland it was previously reported (Stawi nski 1969, Wołowski 1998), but we never found it ourselves.
Phacus applanatus. Two strains, CCAC 2604B (=ASW 08023) isolated in the Netherlands (channel in Leiden) and Beopsu030709E isolated in South Korea, occurred in clade H ( Fig. 2 and Table S1). The morphological study of the CCAC 2604B strain demonstrates that it is an elongated-oval, flat cell that terminates with a straight tail (Fig. 3r). These are the only reports of this species across the world.
It is yet to be found in Poland.
Phacus arnoldii and P. limnophilus. These two taxa represent the basal branch of the Phacus phylogenetic tree (clade A in Fig. 2), despite being morphologically "more similar" to Lepocinclis. The greatest similarity is the three-ridged cell shape of P. arnoldii, alike L. tripteris, while the spindle-shaped (unflattened) cells of P. limnophilus are typical of the majority of Lepocinclis species (Fig. 4, b, c and g-i). Increasing the number of DNA sequences only further supported this positioning. Phacus arnoldii had been described from Ukraine in 1915, but its presence is often reported in the literature as P. warszewiczii Drezepolski; named after Poland's capital city, Warsaw ("Warszawa" in Polish). In Poland P. arnoldii occurs rarely and always in small densities, in contrast to P. limnophilus, which is much more common and may create dense populations (B. Zakry s, pers. obs.). Both species are considered cosmopolitan (e.g., Popova and Safonova 1976, Starmach 1983, Tell and Conforti 1986, Shi et al. 1999. The variety described from Russia (Phacus arnoldii var. ovatus; Popova 1947) has been included as a synonym of P. arnoldii, as minute differences in cell shape are not diagnosticthe sequence of the Polish strain (one with more rounded cells) is sister to a strain isolated in Austria (CCAC 2432E, =ASW 08064), the shape of which (less round, more elongated) seems to be a result of culturing conditions.
Phacus caudatus and P. manginii. These two species are closely related and morphologically similar due to their well-developed (tall) dorsal fold that elongates into a sharp, hyaline, slightly ventrally bent tail, but they do differ in cell shape and size (subclades in clade E in Fig. 2; Fig. 3, s-u, Table 1). Phacus caudatus was described at the end of 19 th century (H€ ubner 1886). Later, several varieties were distinguished (see http://www.algaebase.org) based on minute cell shape and size differences. We did not confirm the validity of several of these forms, which is why they are being synonymized under P. caudatus. The presence of Austrian, Korean, Romanian and Polish strains in the P. caudatus clade affirms the cosmopolitan nature of this species. Daepyeong101908G, as well as ASW 08020 (=CCAC 2415 B) strain, identified as P. carinatus in previous works, and likewise Suwol060709C strain, identified as P. swirenkoi Shin 2014, Kim et al. 2015), have been placed in synonymy under P. caudatus based on their placement in the P. caudatus clade as well as their morphology (Table 1, Fig. 3, s and t). Moreover P. carinatus is a very dubious species. This morphological form was first identified in Australia by Playfair (1921, as P. triqueter) and later considered a new species (P. carinatus) by Pochmann (1942). It has not been found since. Moreover it is not present in any critical monograph on euglenids (Popova and Safonova 1976, Tell and Conforti 1986, Shi et al. 1999. In Poland, Phacus caudatus is common and occasionally can be found in dense populations, similarly to P. manginii described from Indochina (L ef evre 1931). The latter has rarely been mentioned in the literature (from China, Shi et al. 1999; and Argentina, Tell and Conforti 1986 among others), most likely due to being mistaken for P. caudatus. However, its presence in Europe does affirm its cosmopolitan status. The Gungnamji052507C strain (identified as P. triqueter) is renamed as P. manginii due to the position of its sequence in the P. manginii clade.
The variety described by Dre_ zepolski from Poland as Phacus caudatus var. polonica was recognized by Pochmann (1942) as separate species (P. ankylonoton), which has been supported by this researchthe sequence of CCAC 0043 strain (mistakenly identified in the collection as P. ranula) occurred in a clade with the Korean strain located far from the P. caudatus clade.
Phacus elegans and P. lismorensis. According to Pochmann these two very characteristic, yet still very similar morphological forms represent two species -Phacus elegans can be distinguished from P. lismorensis based on "a more elongated cell form and a shorter tail" (Pochmann 1942: 199). The Polish strains representing both morphological forms occur in different clades, Phacus elegans close to species with long tails (P. longicauda, P. helikoides among others), whereas P. lismorensis occurs in clade I in Figures 2 and S1. Furthermore, we have established that cell shape is a diagnostic morphological feature with a bow-like arch and a visibly ventrally bent tail in P. lismorensis and straight (elongated) in P. elegans (Fig. 4, a and n). Additionally, P. elegans has a low crest that causes it to be slightly thicker than P. lismorensis. In Poland both species occur rarely and in minute densities. In the literature P. lismorensis is mentioned often and is considered to be cosmopolitan (e.g., Popova and Safonova 1976, Starmach 1983, Tell and Conforti 1986, Shi et al. 1999). On the other hand, P. elegans is rarely noted, which might be the result of the doubt over the validity of distinguishing the two species Safonova 1976, Tell andConforti 1986).
Phacus gigas and P. hamatus. These two species occur in the same clade (F in Fig. 2), despite being morphologically different both in cell shape and size -Phacus gigas is flat, slightly asymmetrical and large (the largest of all species in Phacuson average 100 9 75 µm), whereas P. hamatus has symmetrical, much smaller (on average 50 9 30 µm) and spoonshaped (convex) cells (Table 1, Fig. 4, j and k). Despite such characteristic features, until recently some researchers questioned the validity of distinguishing P. gigas (e.g., Popova 1947, Popova andSafonova 1976). However both Bennett and Triemer (2012) and herein, in which the number of sequences for both species has been increased, unambiguously support two distinct species with interspecific nSSU rDNA variability above 7%. Phacus hamatus (as P. pleuronectes var. citriformis) has been described from Poland (Dre_ zepolski 1922), while P. gigas from Brazil (Cuncha 1913). Both species are cosmopolitan (Starmach 1983, Tell and Conforti 1986, Shi et al. 1999. In Poland P. hamatus is common, whereas P. gigas is far more rare, however both may occur in large densities (B. Zakry s, pers. obs.).
Phacus granum. This species had been described from Poland (Dre_ zepolski 1925) based on the presence of large paramylon grains in the cell that in the light of current knowledge is not a diagnostic trait, as it depends on the physiological state of the organism. Sadly, due to the small cell size and its morphological similarity to other representatives of the "small Phacus" group (P. brevisulcus, P. claviformis, P. hordeiformis, P. longisulcus, P. minimus, P. oscillans, P. parvulus, P. polytrophis, P. pusillus, and P. skujaeclade D in both Figs. 2; S1), we have failed to isolate material from environmental samples. Most likely, based on the same reasons, it is often omitted in the literature, despite being reported as cosmopolitan by some authors (Popova and Safonova 1976, Starmach 1983, Tell and Conforti 1986. Phacus minutus. This morphological form has been described from Australia (a botanical garden in Sydney) as a variety of Phacus pleuronectes (P. pleuronectes var. minuta; Playfair 1921) due to their similar, ovoid cell shape, but of a smaller size (P. pleuronectes: 39-55 9 22-33 µm; P. minutus: 20-30 9 11-25 µm). Later, it was raised to the rank of species by Pochmann (1942). In his diagnosis, Pochmann stressed additional differences from 1152 P. pleuronectes, such as a more flattened cell shape. The validity of distinguishing this taxon is confirmed by our research. Although two sequences (Korean and Portuguese) form a sister clade with sequences of P. pleuronectes (Fig. 2, clade B subclades), the sequence divergence (2.4-4%) and morphological features (cell size and shape) allowed us to decide that P. pleuronectes and P. minutus should be accepted as distinct species (Fig. 3, c, d and p), at least pending the inclusion of new strains morphologically similar to P. minutus in the analyses. Moreover cells of the Portuguese strain (ACOI 1755) correspond morphologically to the original description by Pochmann with the exception of the tail, which is bent toward the dorsal side rather than sideways.
Phacus orbicularis and P. paraorbicularis. The two common, cosmopolitan and morphologically similar species are sister taxa (sequence divergence 0.7-1.4 %) and have already been verified taxonomically (Kosmala et al. 2007, Kim andShin 2014). Phacus orbicularis was described in the 19th century (H€ ubner 1886), whereas P. paraorbicularis was described recently based on morphological and DNA sequence data (Kim and Shin 2014). In Poland, P. paraorbicularis occurs commonly and often in high densities. Phacus orbicularis is far rarer, and its populations are much more varied in cell shape and size (Kosmala et al. 2007). The form we collected in Poland is often known in the literature as P. heimii (symmetrical cells [31 9 23 µm] with a short tail [4 µm]; Fig. 3k). Its SSU sequence (MN149583 of the isolate UW2362Choc) places it in the P. orbicularis clade, confirming that P. heimii is a synonym of P. orbicularis (subclade in clade J in Figs. 2, S1).
Phacus pleuronectes. This species was taxonomically verified (Kosmala et al. 2007). After the addition of new sequences, it remains in a wellestablished clade, in which two groups can be distinguishedone consists of the sequences from culture strains and one from Poland, while the other groups Polish strains only (Fig. S1). The cells from the Polish populations have a slightly more trapezoid-like shape and are slightly smaller (30-37 9 20-25 µm; Fig. 3p, Table 1) in comparison with cultured cells (38-55 9 26-28 µmsee table 2 in Kosmala et al. 2007). As genetic diversity in the clade is low, those morphological differences might be a result of different living conditions. The trapezoid-like form had been described from Poland under the name of P. trapezoides (Stawi nski 1969) and from China as P. trapezialis (Shi 1987). Both names are now included in the list of synonyms for P. pleuronectes. It is cosmopolitan and a common species in Poland.
Phacus raciborskii. The species of Phacus with flat and permanently spirally twisted cells are located in different clades (see P. inflexus, P. smulkowskianus, P. tortus or P. helikoides -Figs. 2, S1). Another taxon of such morphology is P. raciborskii, described from Poland (Dre_ zepolski 1925). Four sequences of Polish strains have joined the Portuguese (strain ACOI 1758), two Korean (Gungnamji052507B as P. trimarginatus and Leynes012810F as P. mariae in Shin 2014, Kim et al. 2015) and the ASW 08004 (as Phacus sp.) strain sequences, creating a well-supported clade (Fig. 2, clade E subclade, Fig. S1). The Portuguese sequence had been verified based on images posted on the website of the ACOI culture collection (http://acoi.ci.uc.pt/spec_detail.php?cult_ id=1442). In the literature, we have found one more species (P. trimarginatus), described from ponds in Iowa (USA), that is morphologically indistinguishable from P. raciborskii and therefore has been included as a synonym. Phacus raciborskii is widely accepted as cosmopolitan and common (Pochmann 1942, Popova and Safonova 1976, Starmach 1983. It had been noted in Poland for many years (Dre_ zepolski 1925, Czosnowski 1948, Stawi nski 1969 and is often found in dense populations (B. Zakry s, pers. obs.).
Phacus salinus. This very characteristic morphological form had been appearing in the literature since the 19 th centuryfirst under the name Crumenula texta (Dujardin 1836: 204, pl. 9, fig. M), later as Euglena texta (H€ ubner 1886) or Lepocinclis texta (Lemmermann 1901). None of the aforementioned descriptions refer to the striation direction. Only Fritsch (1914) drew attention to this feature and described a species new to science (L. salina) due to its periplast striation (striae running from right to left). Popova (1955) however considered it to be a variant of L. texta (L. texta var. salina). Unexpectedly, based on DNA sequence data, L. salina is transferred to Phacus (P. salinus), in spite of being morphologically (oblong, rigid cells ; Fig. 4e) more similar to Lepocinclis (Linton et al. 2010). Until now only one sequence had been available (EU624028 of the strain of SAG 1244-3), and its position was not well supported (Linton et al. 2010). Five additional sequences form a common, strongly supported clade with the previous sequence, but the position remains uncertain (clade G in Fig. 2,  Fig. S1). Phacus salinus occurs commonly in Poland, often in large densities. Both forms (texta and salina) are universally accepted as cosmopolitan and common (e.g., Popova and Safonova 1976, Starmach 1983, Tell and Conforti 1986, Shi et al. 1999. In spite of that, we have never found texta (with periplast striation running from left to right), despite having inspected over 100 populations.
Variability of nSSU rDNA fragments (lengths and sequences) of Phacus salinus is much higher than the observed morphological diversity. There are also no morphological features that could be used to distinguish the species. This occurs in several common euglenid species with very high intraspecific genetic diversity, e.g., Lepocinclis fusiformis up to 7.8%, L. hispidulus up to 5.1% (Łukomska-Kowalczyk et al. 2020), and P. circumflexus up to 4.9% (Łukomska-Kowalczyk et al. 2015).

PHYLOGENY AND TAXONOMY OF PHACUS
The sequence MN149599 of Phacus salinus UW1948Wos (3,560 bp) is almost the longest known nSSU rDNA in autotrophic euglenids, only that of Euglena sanguinea is longer-the fragment exceeds 6,000 bp (Karnkowska-Ishikawa et al. 2013).
Phacus stokesii and P. segretii. Despite the two species being very similar morpologically (wide-oval cells without a tail and with a long apical furrow (Fig. 4d and the photo of Phacus segretii strain ACOI 1337 on the collection's website: http://acoi.ci.uc.pt/spec_ detail.php?cult_id=815), their sequences are found in separate clades (in clade H and as a sister to clade K in Fig. 2, Fig. S1). The most significant difference is cell size -P. stokesii is almost twice as big (40-46 9 30-35 µm) as P. segretii (22-28 9 20-22 µm; Table 1). In the literature, there are several other species of a similar morphology described, out of which five (P. fominii, P. aspidion, P. balatonicus, P. starmachii, and P. betkowski) are herein designated as synonyms of P. stokesii; one other (P. stokesii f. minor) has been synonymized under P. segretii due to a similar size and a lack of any other well-defining feature. Popova and Safonova (1976: 62-63) listed the same synonyms, except for P. fominii. Both species (P. stokesii and P. segretii) are considered cosmopolitan, but P. stokesii is far more common (Popova and Safonova 1976, Starmach 1983, Tell and Conforti 1986, Shi et al. 1999. Historically, only P. stokesii has been reported in Poland (Dre_ zepolski 1925, Stawi nski 1969, Wołowski 1998, which currently is very uncommon and never found in large densities (B. Zakry s, personal observation) Phacus tenuis. The Portuguese strain ACOI 1757 (as Phacus caudatus in ACOI) occurs outside of the "caudatus" clade (clade H in Fig. 2), despite being morphologically very similar to P. caudatus (cells flat, elongated-oval in shape; Figs. 3, s and t, 4f). It does however differ in terms of having a lower crest and a straight tail, which in turn makes it similar to P. applanatus (Fig. 3r). A detailed literature study shows, that such a morphological form was described from Ukraine by Svirenko (1915a), first as P. caudatus var. tenuis, and in the same year was raised to the rank of species (P. tenuis) by Svirenko (1915b). The research presented herein supports such a decision. Phacus caudatus var. tenuis is reported from various countries: Russia (Popova and Safonova 1976), the Czech Republic (Wołowski 1992), Slovakia (Wołowski and Hind ak 1996) and Poland (Stawi nski 1969).
Phacus triqueter. The diagnostic trait for this species is its high crest that causes the cell's triangular appearance when cross-sectioned. This three-ridged cell shape of Phacus triqueter makes it similar to Lepocinclis tripteris or P. arnoldii. The very characteristic morphological form was described by Ehrenberg (1838) as Euglena triquetra and later moved to Phacus as P. triqueter (Dujardin 1841). Many authors considered it to be cosmopolitan (e.g., Pochmann 1942, Popova and Safonova 1976, Starmach 1983, Tell and Conforti 1986, Shi et al. 1999. Meanwhile, it has yet to be found in Poland by us, despite previous reports of its presence (Czosnowski 1948, Stawi nski 1969, Wołowski 1998. Only one sequence of P. triqueter (from strain SAG 1261-8) was included in the phylogenetic analysis (Fig. 2). Based on the image posted on the Culture Collection of Algae and Protozoa (CCAP) website, it appears that the strain had been verified correctly (https://www.ccap.ac.uk/ results2014.php?mode=attr&Environment=All&Coun try=All&Pathogen=All&Type_Culture=All&Genus_ Name=Phacus&Strain_Name=Phacus+triqueter&Stra in_No=1261%2F8). CONCLUSION As a result of the presented studies the phylogenetic trees of Phacus now includes 50 species represented by 129 sequences of SSU rDNA, of which seven species (Phacus anacoelus, P. anomalus, P. curvicauda, P. elegans, P. lismorensis, P. minutus, and P. stokesii) and 55 sequences are new for the tree. The new sequences were obtained from laboratory strains (6) and those isolated directly from the environment (49). The environmental isolates came from 37 eutrophic water bodies located in Poland and one located in the Czech Republic. Comparative morphological and molecular studies on new laboratory strains and environmental isolates as well as research in the literature, have allowed (a) an estimation of morphological and genetic diversity and verification of morphological diagnostic features for particular taxa (well-established clades on a molecular phylogenetic tree) and as a result their proper identification; due to the lack of morphological data, the taxonomic affiliation of five strains (Jigok090112 as P. ranula on the phylogeny tree; Leynes012810F as P. mariae; Burni081809, ASW 08004 and ACOI1138 as Phacus sp.) was not verified; (b) the reconstruction of phylogenetic relationships among 50 species of Phacus (c) taxonomic verification, emending diagnoses, and designating epitypes for well-distinguished taxa (19 species).

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web site: Figure S1. The Maximum Likelihood phylogenetic tree based on 136 nSSU rDNA sequences (of which 129 represent Phacus). The Bayesian posterior probability (pp) and the bootstrap (bs) values obtained by maximum likelihood analysis are marked at the nodes. The pp <0.75, bs values <50 and clades not present in the particular analysis are marked with a hyphen (-). The sequences obtained in this study are indicated in bold type. Scale bar represents number of substitutions per site. Table S1. List of species and sampling data of isolates/strains used in this study. GenBank accession numbers and their nuclear SSU rDNA gene sequenced are given, with new sequences indicated in bold type. Table S2. The nuclear SSU rDNA pair-wise sequence distances (%) among strains and isolates.