American Journal of Botany 95(9): 1079–1095. 2008.

I N SEARCH OF MONOPHYLETIC TAXA IN THE FAMILY DESMIDIACEAE (ZYGNEMATOPHYCEAE, VIRIDIPLANTAE): THE GENUS COSMARIUM 1

Andrey A. Gontcharov2 and Michael Melkonian

Botanisches Institut, Lehrstuhl I, Universit ä t zu K ö ln, Gyrhofstr. 15, D-50931 K ö ln, Germany

Nuclear-encoded small subunit (SSU) rDNA, 1506 group I introns, and chloroplast rbcL genes were sequenced from 97 strains representing the largest desmid genus Cosmarium (45 spp.), its putative relatives Actinotaenium (5 spp.), Xanthidium (4 spp.), Euastrum (9 spp.), Staurodesmus (13 spp.), and other Desmidiaceae (Zygnematophyceae, Streptophyta) and used to assess phylo- genetic relationships in the family. Analyses of single genes and of a concatenated data set (3260 nt) established 10 well-supported clades in the family with Cosmarium species distributed in six clades and one nonsupported assemblage. Most of the clades con- tained representatives of at least two genera highlighting the polyphyletic nature of the genera Cosmarium , Euastrum , Staurodes- mus, and Actinotaenium. To enhance resolution between clades, we extended the data set by sequencing the slowly evolving chloroplast-encoded large subunit (LSU) rRNA gene from 40 taxa. Phylogenetic analyses of a concatenated data set (5509 nt) suggested a sister relationship between two clades that consisted mainly of Cosmarium species and included C. undulatum , the type species of the genus. We describe molecular signatures in the SSU rRNA for two clades and conclude that more studies in- volving new isolates, additional molecular markers, and reanalyses of morphological traits are necessary before the taxonomic revision of the genus Cosmarium can be attempted.

Key words: Actinotaenium; clades; Cosmarium ; Desmidiaceae; Euastrum ; molecular phylogeny; molecular signatures; poly- phyly; taxonomy.

The conjugating green algae (Zygnematophyceae, Viri- known in microalgal systematics in which descriptions of gen- diplantae) represent the most species-rich lineage in the Strepto- era have often been based on single or very few morphological phyta except for the embryophytic land plants ( Gerrath, 1993 ). characters visible in the light microscope without a careful in- They share with most embryophytes not only a common ances- vestigation of phylogenetic signifi cance (one related example is try but also the absence of fl agellate reproductive stages. A pe- that of the coccoid green algal genus Chlorella and its relatives; culiar mode of sexual reproduction (i.e., conjugation) sets the Krienitz et al., 2004; Luo et al., 2006). We have therefore started class apart from other streptophytes and may have contributed to evaluate the genus concept in the most species-rich family to their successful diversifi cation (Brook, 1981). On the basis of of the Zygnematophyceae, the Desmidiaceae (desmids), using ultrastructural analyses of mitosis and cytokinesis, zygnemato- taxon-rich sampling and multigene phylogenetic analyses. Al- phycean algae have been recognized as members of the strepto- though the traditional genus Staurastrum Meyen ex Ralfs was phyte algae known also as Charophyceae sensu Stewart and shown to be polyphyletic, a monophyletic core of the genus Mattox (Pickett-Heaps, 1975). Molecular phylogenetic analy- could be identifi ed using such an approach ( Gontcharov and ses of the Zygnematophyceae corroborated these results Melkonian, 2005). It is anticipated that once a phylogenetic ( Chapman and Buchheim, 1992; Surek et al., 1994; McCourt framework of a desmid genus has been established, a reinvesti- et al., 2000; Gontcharov et al., 2003 ), although the phylogenetic gation of its morphological traits should lead to the recognition position of the class among the streptophyte algae still remains of hitherto overlooked, but more sound and reliable generic unresolved (Chapman et al., 1998; Karol et al., 2001; Lewis and morphological characters. McCourt, 2004 ; Turmel et al., 2007 ). Here, we address the phylogenetic status of the genus Cos- Phylogenetic analyses of the Zygnematophyceae using a marium Corda ex Ralfs (Desmidiaceae, Zygnematophyceae), broad taxon sampling and multigene data sets have more re- the most species-rich desmid genus with more than 1000 spe- cently led to the conclusion that many traditional genera of the cies described (Gerrath, 1993). Together with Staurastrum class are polyphyletic, suggesting that the characters used to (~700 spp.) it constitutes about half of the total number of spe- delineate these taxa are either plesiomorphic, homoplasious or cies in the otherwise species-poor streptophyte green algae. It unreliable ( Gontcharov et al., 2003 , 2004 ; Gontcharov and should be mentioned that Cosmarium has always been regarded Melkonian, 2005 ; Hall et al., 2008 ). This situation is well as an artifi cial genus and thus taxonomically problematic ( West and West, 1905 , 1908 ; Fritsch, 1953 , Hirano, 1959a ; Krieger and Gerloff, 1962, 1965, 1969; Prescott et al., 1981, Croasdale and Flint, 1988; Brook and Johnson, 2002; Gerrath, 2003). It 1 Manuscript received 7 February 2008; revision accepted 11 June 2008. was poorly circumscribed by a vague diagnosis (Ralfs, 1848) The authors thank P. C. Silva and F. A. C. Kouwets for discussion on the and linked morphologically to other genera such as Euastrum type species of Cosmarium. This work was supported by DFG grant Ehr. ex Ralfs and Xanthidium Ehr. ex Ralfs. Thus, although ME-658/26-1. 2 Author for correspondence (e-mail: [email protected]); over the 160 years since its description, numerous taxa have permanent address: Institute of Biology and Soil Science, 690022, been added to Cosmarium , the defi nition of the genus (“ the Vladivostok-22, Russia. fronds are minute, simple, constricted in the middle; the seg- ments are generally broader than long and infl ato-compressed, doi:10.3732/ajb.0800046 but in some species orbicular or cylindrical; they are neither 1079 1080 American Journal of Botany [Vol. 95 emarginate at the end nor lobed at the sides, and have no spines criminate these genera on the phylogenetic tree revealed exten- or processes” ; Ralfs, 1848, p. 91) has not changed and its dis- sive homoplasy, calling into question the current genus concept criminatory power diminished. Although widely acknowledged in the Desmidiaceae. to be polyphyletic, the genus is still adopted to date in its origi- nal sense ( Lenzenweger, 1999 ; Brook and Johnson, 2002 ; Gerrath, 2003; Coesel and Meesters, 2007). Unfortunately, attempts MATERIALS AND METHODS during the 19th century to resolve the taxonomic problems in Cosmarium and establish more “ natural ” (morphologically uni- Cultures— One hundred twenty-seven strains of Desmidiaceae and Pe- form) taxonomic entities were unsuccessful (e.g., N ä geli, 1849 ; de niaceae used for this study were obtained from different sources (Appendix 1) Bary, 1858; Lundell, 1871; Kirchner, 1878; Gay, 1884; Hansgirg, and grown in modifi ed WARIS-H culture medium (McFadden and Melkonian, − 2 − 1 1888; de Toni, 1889; Raciborski, 1889; Turner, 1892). The 1986) at 15 °C with a photon fl uence rate of 40 µ mol ⋅ m ⋅ s in a 14/10 h light/ dark cycle. The taxonomic designation of all strains was verifi ed by light mi- novel taxa were based on simple morphological features such croscopy prior to DNA extraction ( Krieger and Gerloff, 1962 , 1965 , 1969 ; as cell and semicell shape, ornamentation of the cell surface, Prescott et al., 1981 ; Croasdale and Flint, 1988 ; Brook and Johnson, 2002 ; degree of cell constriction, and chloroplast shape that occur in Coesel and Meesters, 2007 ). any combination in the genus. In 1954, Teiling established a new genus, Actinotaenium Teil ., for taxa with smooth-walled, DNA extraction, amplifi cation, and sequencing— After mild ultrasonica- elongated cells, circular in apical view, and displaying a shal- tion to remove mucilage, total genomic DNA was extracted using the Qiagen low sinus. Although Actinotaenium is often regarded as a “ nat- (Hilden, Germany) DNeasy Plant Mini Kit. Nuclear-encoded (nu) SSU rDNA ural group” (Prescott et al., 1981, p. 1), its members are distinct (including the 1506 group I intron) and chloroplast-encoded (cp) rbcL and LSU only in the combination of characters that individually occur in rDNA were amplifi ed by polymerase chain reactions (PCR) using published protocols and 5′ -biotinylated PCR primers (Marin et al., 1998, 2005; many Cosmarium taxa. Another consequence of the unsatisfac- Gontcharov, et al., 2004). PCR products were purifi ed with the Dynabeads tory taxonomic status of Cosmarium is the fact that only one M-280 system (Dynal Biotech, Oslo, Norway) and used for bidirectional se- unfi nished attempt of a monography of the genus dealing with quencing reactions (for protocols, see Hoef-Emden et al., 2002). Gels were run fewer than half of the described species exists ( Krieger and on a Li-Cor IR2 DNA sequencer (Li-Cor, Lincoln, Nebraska, USA). Gerloff, 1962, 1965, 1969). The lack of type material and the inaccessibility or vagueness of many original descriptions further Sequence alignments and tree reconstructions— Sequences were manually complicates a critical assessment of species that are distin- aligned using the SeaView program (Galtier et al., 1996). For coding regions of guished largely on the basis of the shapes of cells, semicells, the nu SSU rDNA, cp LSU rDNA, and noncoding 1506 group I introns, the and chloroplasts; features of cell wall ornamentation; and zy- alignment was guided by primary and secondary structure conservation ( Bhattacharya et al., 1994, 1996 ; Wuyts et al., 2000, 2001 ; Gillespie et al., gospore shape. The extent of the variability of these characters 2006 ). All three codon positions of the rbcL gene were used for analyses. The is poorly known, and their taxonomic signifi cance has never alignments are available at http://srs.ebi.ac.uk/, accessions ALIGN_001252, been assessed within a cladistic framework. ALIGN_001253 and ALIGN_001254. The fi rst tests of the genus concept of Cosmarium with mo- The amount of phylogenetic signal vs. noise in our nu SSU rDNA, rbcL , and lecular tools confi rmed the expected polyphyly of the genus, cp LSU rDNA data were assessed by plotting the uncorrected against corrected but this result was based on a very limited taxon sampling, and distances determined with the respective model of sequence evolution esti- mated by the program MODELTEST version 3.06 (Posada and Crandall, 1998). the phylogeny was affected by long-branched taxa and the lim- The selected models and model parameters are summarized in Table 1. Also, ited phylogenetic resolution of the marker used (nuclear-en- the measure of skewness (g1-value calculated for 10c000 randomly selected coded small subunit [SSU] rDNA). Six Cosmarium sequences trees in the program PAUP* version 4.0b10; Swofford, 2002 ) was compared representing distinct morphotypes within the genus had no rela- with the empirical threshold values ( Hillis and Huelsenbeck, 1992 ) to verify the tionship to each other; instead, some species formed clades with nonrandom structuring of the data. To quantify the extent of substitution satura- members of other genera ( Gontcharov et al., 2003 ). Conversely, tion in data sets, we calculated the I ss statistic for the individual and combined data sets with the program DAMBE ( Xia and Xie, 2001 ). phylogenetic analyses of two taxa of Cosmarium with the same Phylogenetic trees were inferred with maximum likelihood (ML), neighbor- morphotype, revealed only a distant relationship between them, joining (NJ) distance, and maxiumum parsimony (MP) optimality criteria using questioning the validity of morphological characters tradition- PAUP* 4.0b10 and Bayesian inference (BI) using the program MrBayes ver- ally used in the taxonomy of the genus but confi rming a previ- sion 3.1.2 ( Huelsenbeck and Ronquist, 2001 ). Evolutionary models (for ML ously reported relationship between some smooth-celled species and NJ analyses) were selected by the Akaike information criterion in of Cosmarium and spine-bearing Staurodesmus Teil. taxa MODELTEST. ML and MP analyses used heuristic searches with a branch- swapping algorithm (tree bisection-reconnection); distances for NJ analyses ( Gontcharov et al., 2003 ; Gontcharov and Melkonian, 2005 ). were calculated by ML. In BI, two parallel MCMC runs were carried out for The primary goal of this study was to defi ne major mono- two million generations sampling every 100 generations for a total of 20 000 phyletic lineages within the traditional genus Cosmarium , re- samples. The fi rst 500 – 1500 samples were discarded as burn-in, and the re- lated genera, and the family Desmidiaceae and generate a maining samples were analyzed using the sumt command in MrBayes. The ro- hypothesis of their phylogenetic relationship. To this end, we bustness of the trees was estimated by bootstrap percentages (BP; Felsenstein, sampled more than 120 species, representing major morpho- 1985 ) using 1000 (NJ and MP) or 100 (ML) replications and by posterior prob- abilities (PP) for BI. BP < 50% and PP < 0.95 were not taken into account. types of Cosmarium , and its putative relatives, Staurodesmus , In MP, the stepwise addition option (10 heuristic searches with random taxon Euastrum , Xanthidium, and Actinotaenium . Nuclear-encoded input order) was used for each bootstrap replicate. The ML bootstrap used a SSU rDNA, noncoding 1506 group I introns, chloroplast large single heuristic search (starting tree via stepwise addition) per replicate. subunit (LSU) rDNA and protein-coding rbcL gene sequences Previous molecular phylogenetic studies resolved the families Peniaceae were obtained from these taxa and analyzed individually and in (the genus Penium ) and Desmidiaceae as sisters and have shown that they concatenation using different phylogenetic methods. We identi- evolved with comparable evolutionary rates unlike the more distantly related, fast-evolving desmid families Gonatozygaceae and Closteriaceae ( Besendahl fi ed 10 clades in the Desmidiaceae, with the genus Cosmarium and Bhattacharya, 1999 ; McCourt et al., 2000; Denboh et al., 2001 ; Gontcharov distributed over six of these. Furthermore, Cosmarium , Euas- et al., 2003 ). Later Penium was found to not be monophyletic, and only one trum , Staurodesmus, and Actinotaenium were shown to be sublineage was sister to the Desmidiaceae ( Gontcharov et al, 2004 ; Hall et al., polyphyletic. Mapping traditional morphological traits that dis- 2008 ). Therefore, only P. margaritaceum and P. spirostriolatum , comprising September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1081 0001 ( Fig. 4 ) rbcL cp LSU rDNA + nu SSU rDNA (with intron) + + nu SSU rDNA cp LSU rDNA 40 taxa cp LSU rDNA rbcL + nu SSU rDNA (with intron) nu SSU rDNA ( Fig. 3 ) rbcL + nu SSU rDNA (with intron) nu SSU rDNA ed by the program MODELTEST for different data sets used for Figs. 2 – 4 and for additional analyses. 4 – Figs. 2 data sets used for for different ed by the program MODELTEST

97 taxa rbcL group I intron) nu SSU rDNA (with 1506 nu SSU rDNA ( Fig. 2 ) 127 taxa rbcL 0204/0766, < 0001< 0204/0766, 0001 0001< < 0185/0788, 0,21/0766, 0001< 0195/0806, 0001< 0001 0222/0806, < 0231/0793, < 0163/0812, 86 119 90 209 140 231 371 a T = 1.00]) 416 281 402 683 155 514 669 a ↔ -value of 32 taxon -value P + c, C] G] T] 0.9147 3. 6638 1.2907 1.5573 3.7675 2.1920 0.9356 3.9135 1.4815 1.2953 4.0542 1.9766 0.5478 1.1043 1.6874 1.2349 3.6121 2.0898 0.9605 2.5445 2.0619 G] T] 1.0244 7.1143 0.5683 10.2938 0.9365 7.5798 0.7169 8.8722 0.1431 2.9835 0.6480 8.3522 0.4238 6.8163 ss 1. lnL) and model parameters identifi – ( values models, log likelihood Evolutionary I ↔ ↔ ↔ / ↔ ↔ ss I Informative and uninformative (but not invariable) characters in MP not invariable) (but and uninformative Informative a data subsets) ( ss

[A [A [C Aligned nt [C Constant nt MP-informative 1339 836 1921 1524 1339 846 3260 2370 2255 1960 3254 2509 5509 4469 [A Table Model – lnLIGBase frequencies A C G T GCRate matrix ([G GTR+I+G 15157.271 GTR+I+G 0.5644 0.9369 12888.799 0.3002 0.1632 GTR+I+G 0.1857 13512.904 0.7145 0.3509 0.3489 0.4982 0.2485 GTR+I+G 0.2108 27175.160 0.2646 0.5701 0.2512 1.0241 0.4754 GTR+I+G 0.2975 0.1708 7373.987 0.1884 0.3433 0.6583 0.3592 GTR+I+G 0.6482 17786.400 0.2652 0.1937 0.2359 0.3052 0.7864 0.4296 GTR+I+G 0.5697 25536.258 0.2783 0.2301 0.3100 0.6779 0.1816 0.6572 0.5401 0.2628 0.1981 0.2442 0.2949 0.4423 0.7197 0.6007 0.2694 0.2128 0.2710 0.2467 0.4838 MP-uninformative MP-uninformative Measure of skewness (g1) I − 0324 − 0396 − 0451 − 0406 − 0564 − 0526 − 0655 Model and model parameter 1082 American Journal of Botany [Vol. 95 this distinct clade, were used as an outgroup for the Desmidiaceae in our analy- The majority of Cosmarium species were distributed among ses, and three more species of Penium were regarded as ingroup taxa (see seven clades (ARTHR, CO1, CO2, CO3, CO4, STD2, and om- Results). niradiate) and one assemblage (multicellular 2). Most of these clades (except for CO2 and CO 3) also included species of other Combined analyses— For concatenated analyses, partitions were fused and genera, namely Actinotaenium , Euastrum , Spondylosium , and analyzed using a single “ concatenated model” with averaged parameters. Be- fore that, models for individual partitions ( Table 1 ), ML topologies, and ML/ Staurodesmus ( Fig. 2 ). Two Cosmarium taxa, C. decedens and NJ(ML)/MP bootstrap support (Table 2) were obtained and compared to reveal C. ralfsii, were placed in the Euastrum and Micrasterias clades, signifi cant discrepancies. We also assessed incongruence between the data sets respectively. In both cases, affi liation of the Cosmarium species by the incongruence length difference (ILD) test ( Farris et al., 1994 ) in PAUP* to these lineages, however, was only weakly (60/55% BP) or (partition homogeneity test with 1000 replicates). moderately (86/78% BP) supported. In addition, the long- The concatenated data set of nu SSU rDNA (including intron), rbcL , and cp branched C. depressum diverged basally in the family together LSU rDNA was analyzed by BI using specifi c model parameters for each partition. with Actinotaenium cruciferum and three Penium species (CAP clade; Fig. 2 ). Cosmarium was most prominent (26 of 68 species analyzed) RESULTS in clade CO2. Strain SVCK482, identifi ed as C. undulatum, the type species of the genus ( Silva, 1952 ; Gerrath, 1993 ), was also Quality of the molecular data— Apart from the differences a member of this clade. Beside Cosmarium , this species-rich in the alignment length, base and substitution frequencies, num- clade accommodated three Actinotaenium and four Euastrum ber of parsimony-informative positions, pattern of substitution species (Fig. 2). CO2 was further divided into several subclades distribution (G parameter), and proportion of invariable sites and some single branches, but their relationships were largely between our data sets, the most complex GTR+I+G model of unresolved likely due to low sequence divergence in the clade. sequence evolution was identifi ed by the MODELTEST as the Cosmarium species also formed the bulk of the taxa in clade best model fi tting the data ( Table 1 ). A test of the data sets for omniradiate, that also included Spondylosium panduriforme substitution saturation (distribution of the uncorrected vs. cor- and two additional Actinotaenium species ( Fig. 2 ). In the omni- rected distances; Fig. 1) revealed a nearly linear correlation in radiate clade, the Cosmarium species formed four well-sup- the SSU rDNA data indicating low saturation. The saturation ported subclades whose relationships, however, were unresolved plot of rbcL was somewhat leveled off, suggesting the presence in the rbcL phylogeny ( Fig. 2 ). of some saturation that would be expected from the third codon Our analyses also revealed well-supported relationships be- position of this protein-coding gene ( Gontcharov et al., 2004 ). tween some Cosmarium and Staurodesmus species. Members However, according to the I ss statistics, neither of the data sets of Staurodesmus were recovered in three distinct clades (STD1, was saturated (P < 0,001; Table 2 ). STD2, and ARTHR), and two of these clades also contained Comparison of the skewness of the tree length distribution species of Cosmarium . In both STD2 and ARTHR, Cosmarium (g1 value) of random trees of all data sets with the empirical species diverged in paraphyletic succession before the Stau- threshold values ( Hillis and Huelsenbeck, 1992 ) showed that rodesmus taxa. Only in STD2 clade the derived position of the the length distributions were considerably left-skewed, indicat- Staurodesmus species was supported by high bootstrap values ing that the alignments were signifi cantly more structured than ( Fig. 2 ). random data and likely contained a strong phylogenetic signal Two Cosmarium strains, M 2717 and SVCK 570, both iden- ( Table 1 ). tifi ed as C. punctulatum , were found in two distantly related Noise assessment in the data sets without the outgroup (two clades, CO2 and CO3, respectively. Light microscopic exami- Penium species) yielded results identical to those obtained with nation revealed minor differences in cell dimensions and pat- the complete data sets, suggesting that the outgroup did not in- terns of the cell wall ornamentation between the strains, which, terfere with the phylogenetic signal (not shown). nevertheless, fi tted the rather broad species diagnosis (Prescott et al., 1982). A similar discrepancy between taxonomic desig- rbcL data set (127 taxa)— ML analyses of 127 rbcL se- nation and phylogenetic position in the tree was observed for quences (1339 nt, GTR+I+G model; Table 1 ) yielded the phy- Staurodesmus extensus and S. extensus var. joshuae that showed logenetic tree in Fig. 2 . Although resolution of internal branches little affi nity to each other in clade STD1 ( Fig. 2 ). It is likely was mostly low, 11 terminal clades that included most of the that the small difference in cell morphology between these con- taxa studied were resolved and designated STD1, STD2, CO1, specifi c strains masked the conspicuous genetic distances be- CO2, CO3, CO4, ARTHR, Euastrum , Micrasterias, omniradi- tween the two taxa that should warrant the status of separate ate, and CAP ( Fig. 2 ). Most clades were well supported; only species STD2, Euastrum , omniradiate, and CAP attained weak to mod- erate support (Fig. 2; Table 2). In addition, two lineages con- Concatenated data set (SSU rDNA + rbcL)— To increase taining most of the multicellular (fi lamentous or colonial) taxa phylogenetic resolution and probe the clades established in the analyzed and some Cosmarium species were recovered, the rbcL analyses with another marker, we combined SSU rDNA larger multicellular 2 assemblage, and a long-branched Spon- (including 1506 group I introns) and rbcL sequences obtained dylosium planum/S. secedens pair, multicellular 1. Species of from the same strain. The taxon sampling was reduced to 97 the genus Xanthidium (including Staurastrum tumidum ; species in this data set because in some taxa the SSU rRNA Gontcharov et al., 2003; Gontcharov and Melkonian, 2005) gene could not be amplifi ed using various primer combinations were also split between several individual branches and one (e.g., most members of the CO3 and CO4 clades) or their se- clade. Finally, Haplotaenium minutum formed an unresolved quences formed extremely long branches in the SSU rDNA assemblage with two Staurastrum species; however, bootstrap phylogeny (e.g., Cosmarium ovale and Euastrum moebii ). Con- analyses favored monophyly of Staurastrum with weak support catenation of SSU rDNA and rbc L sequences resulted in a data ( Fig. 2 ; Table 2 ). set consisting of 3260 nt ( Table 1 ). September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1083

Table 2. Signifi cances (ML/NJ(ML)/MP/BI) for the clades and branches (encircled numbers in Figs. 2 – 4 ) in different analyses. If there is no number before the slash, no ML bootstrapping was done for the data set.

127 taxa 97 taxa 40 taxa nu SSU rDNA nu SSU rDNA nu SSU rDNA cp LSU rDNA + nu SSU rDNA Clade/branchrbcL ( Fig. 2 ) (with 1506 group I intron)rbcL (with intron) + rbcL ( Fig. 3 ) cp LSU rDNA (with intron) + rbcL (with intron) + rbcL ( Fig. 4 )

ARTHR / 100/100/1.00 99/98/97/1.00100 100 – /89/100/ – 100 100 STD1 /100/100/1.00 100 100 100 96/100/100/1.00 100/96/100/1.00 98/100/100/1.00 CO1 /100/100/1.00 93/93/91/1.00100 100 N/A N/A N/A CO2 /96/78/0.99 86/79/73/1.00 83/99/80/1.00100 97/99/100/1.00 99/100/99/1.00100 CO3 /100/100/1.00 59/64/ − / − 100 100 N/A N/A N/A CO4 /92/85/0.99 N/A N/A N/A N/A N/A N/A Xanthidium / − / − / − -/60/-/1.00 56/63/68/- 78/96/85/1.00 -/58/100/- 95/100/100/1.00 98/100/100/1.00 STD2 /77/84/ − − / − / − / − 75/79/83/0.99 92/96/94/1.00 71/83/90/0.99 84/96/78/1.00 98/100/90/1.00 Multicellular 1+2 / − / − / − − / − / − / − − / − / − / − − / − / − / − − / − /67/- − / − /57/ − − / − /67 – Euastrum /60/55/0.99 − / − / − / − − / − / − / − − / − / − / − − / − / − / − 84/93/82/0.95 60/ − /51/0.99 Staurastrum /63/53/- 61/ − / − /- 57/60/ − / − 73/73/64/1.00 N/A N/A N/A Micrasterias /86/78/1.00 80/93/69/1.00 93/91/93/1.00 99/100/93/1.00 − / − /97/0.99 99/100/100/1.00 91/98/97/1.00 Omniradiate /54/ − /0.99 68/71/ − /1.00 − / − / − /0.98 91/82/85/1.00 96/73/99/1.00 99/92/90/1.00 100/100/99/1.00 CO2 + CO3 / − / − / − − / − / − /1.00 − / − / − / − − / − / − / − − /62/67/ − − / − / − / − − /78/67/ − CO2 + CO3 + / − / − / − − / − / − / − − / − / − / − − /63/ − / − − / − /68/ − − /70/67/ − 71/68/79/1.00 Xanthidium 1 /51/ − /0.95 − / − / − / − − / − / − 0.99 − / − / − / − − / − /52/ − 62/ − /56/0.97 78/ − /52/1.00 2 /56/56/0.97 − / − / − / − 63/53/55/0.99 56/51/ − / − 92/81/95/ − 87/81/88/1.00 98/97/95/1.00 Notes: N/A, not accessed; 100 =100/100/100/1.00

Comparison of the SSU rDNA- and rbcL -based topologies combined data set to pass the ILD test (P = 0.003). Only when for the 97 taxa data set and their bootstrap supports (Table 2) long branches such as C. depressum , Actinotaenium cruciferum , showed general agreement between the markers (results not and members of the Euastrum and multicellular assemblages shown). Only two clades attained somewhat weaker or no sup- were removed from the data set did it pass the test (P > 0,005). port in the SSU rDNA phylogeny compared to the rbcL data set This result suggests that the nuclear and chloroplast data sets (CO3 and STD2, respectively). Both partitions recovered two are congruent for the data set as a whole. Problems with the ILD nonsupported assemblages, one containing eight of nine Euas- test are well known (e.g., Dolphin et al., 2000 ; Yoder et al., trum species, the other the multicellular species and associated 2001 ; Darlu and Lecointre, 2002 ; Dowton and Austin, 2002 ; Cosmarium taxa ( Table 2 ). Most members of these assemblages Quicke et al., 2007 ), so failures due to localized cases of incon- had accelerated evolutionary rates in the SSU rRNA gene or in gruence are not surprising (Thornton and DeSalle, 2000). Phy- both genes that likely affected bootstrap support for their group- logenetic analyses of the data set without the long-branched ing. Different placement of these unresolved taxa/branches in taxa mentioned revealed that they have no effect on the support the SSU rDNA and rbcL trees was responsible for failure of the levels of other terminal clades and the general tree topology.

Fig. 1. Analyses of saturation in the nu SSU rDNA, rbcL , and cp LSU rDNA data (uncorrected vs. corrected distances). Corrected distances were calculated with the GTR+I+G model estimated by MODELTEST for each partition ( Table 1 ). (A) 97-taxa alignment ( Fig. 3 ), (B) 40-taxa alignment ( Fig. 4 ). 1084 American Journal of Botany [Vol. 95 September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1085

In the combined analyses the clades ARTHR, STD1, CO1, the cell and chloroplast of each species, on the rbcL topology CO2, and CO3 attained 100%BP/1.00 PP in all analyses ( Fig. 3 ; (NJ[ML] bootstrap consensus; Fig. 2 ) in all clades/assemblages Table 2 ). The support increased also from weak or moderate to that contain Cosmarium species ( Fig. 5 ). The following mor- high ( > 90% BP) for the omniradiate and STD2 clades. The ear- phological characters were computed and mapped: (1) apical lier unresolved Xanthidium lineage attained moderate bootstrap view of the cell (circular [omniradiate], elliptical [biradiate] or values, whereas the two multicellular assemblages remained triangular), (2) ornamentation of the cell wall (smooth-walled, without support. The more divergent SSU rDNA sequences ornamented [granular/spiny] or with [a few] stout spines), (3) contributed also to the resolution of the terminal branches width of the isthmus (narrow [ < 1/2 of the cell width] or broad within some clades, in particular CO2 ( Fig. 3 ). [> 1/2 of the cell width]), and (4) type of the chloroplast (axial or parietal). In addition, distinct features of specifi c genera, e.g., Concatenated data set (nu SSU rDNA+rbcL+cp LSU a lobed cell outline (L), incision of the apical lobe (I), and pres- rDNA)— Although phylogenetic analyses using SSU rDNA ence of fi laments (F) or colonies (C) were also recorded ( Fig. 5 ). and rbcL led to the recognition of 10 well-resolved clades in the Most members of clades that consist predominantly of Cos- Desmidiaceae, resolution of internal branches was low; thus re- marium species (CO2, and omniradiate) display some sort of lationships among clades remained unclear ( Figs. 1, 2 ; Table cell wall ornamentation, i.e., granules or warts, arranged in 2 ). We therefore increased the data set by adding a more slowly taxon-specifi c patterns. However, smooth-walled species (e.g., evolving marker, the chloroplast-encoded LSU rRNA gene C. hammerii , C. cucumis , Actinotaenium spp.) were also mem- ( Table 1 ). For these analyses, taxon sampling within the well- bers of the same clades. None of the prominent Cosmarium lin- supported clades was reduced to 1– 2 representatives per clade eages was distinct in the degree of radiation. Biradiate cells are (a total of 40 taxa). generally more common in the genus and in our clades, but om- The PCR products obtained by amplifi cation of cp LSU niradiate species also occur and often constitute distinct sub- rDNA with primers that bound in the B19 and G20 domains of clades (e.g., Actinotaenium spp.; Fig. 5). Taxa with the same the gene (nomenclature after De Rijk et al., 2000), varied sig- chloroplast morphology, width of isthmus (which may infl u- nifi cantly in length between species. This inconsistency was ence the pattern of cytokinesis; see Meindl (1986) , H ö ftberger due to the presence of a single intron in 16 taxa that occurred at and Meindl (1993) ) or cell/semicell shape are also not confi ned four different insertion sites ( Fig. 4 ). BLAST searches indicated to single clades (Fig. 5). Characters are mostly nonlinked and that the introns in the cp LSU rDNA of desmids are similar to occur together in various combinations. However, the 12 taxa group IA1, IA3, and IB4 introns from other green algal chloro- characterized by circular in vertical view cells (and distributed plasts and likely contain group I homing endonucleases over fi ve clades/assemblages and fi ve genera: Cosmarium , Ac- ( Turmel et al., 1993 , 1995 ). tinotaenium , Bambusina , Groenbladia , and Spondylosium ) all As expected, the phylogeny recovered with the concatenated had cells with a wide isthmus, whereas the opposite does not data set including the cp LSU rDNA (a total of 5509 nt) was apply ( Fig. 5 ). generally congruent with the phylogenies obtained in the previ- ous analyses ( Fig. 4 ). Compared to a concatenated data set of Synapomorphies in the SSU rRNA— Mapping morphologi- the nu SSU rDNA and rbcL with a congruent taxon sampling cal characters traditionally used to distinguish desmid genera (40 taxa: Table 2), the cp LSU rDNA data set added phyloge- on the phylogenetic tree clearly demonstrated that these charac- netic signal (and thus enhanced bootstrap values) to some inter- ters cannot be used to circumscribe the clades identifi ed in this nal branches, in particular the Desmidiaceae (exclusive of the study (Fig. 5). To initiate a molecular circumscription of the Penium clades and Actinotaenium cruciferum ; 98% BP in ML; clades, we searched for the presence of molecular signatures, Fig. 4), the omniradiate clade (100% BP in ML) and its basal i.e., nonhomoplasious synapomorphies (NHS; according to position in the Desmidiaceae (78% BP in ML for the Desmidi- Marin et al., 2003) in the SSU rRNA. We restricted the analysis aceae excluding omniradiate and the Penium clades + A. cruci- to clades that contained a large number of Cosmarium species. ferum), and a putative novel clade consisting of CO2, CO3, and Screening of the secondary structure-based alignment revealed Xanthidium (71% BP in ML). These results suggest that the cp several substitutions in the SSU rRNA that characterize the om- LSU rRNA gene could be a useful molecular marker for future niradiate and CO2 clades ( Fig. 6 ). phylogenetic analyses in the Desmidiaceae, especially when the All species comprising the omniradiate clade were distinct in focus is on resolving the deeper divergences in the family. two transversions in an internal loop of Helix 25 (H673 after Gillespie et al. (2006) ; nucleotides 30 (G→ A) and 36 (A→ U); Mapping morphological characters that defi ne genera on Fig. 5A ), whereas all other members of the Desmidiaceae (and the phylogenetic tree— Most of the 10 clades identifi ed in the also the Desmidiales) reveal the plesiomorphic state (G, A) at Desmidiaceae by our molecular phylogenetic analyses, com- these positions. The two substitutions in the clade omniradiate prised representatives of several desmid genera demonstrating thus represent nonhomoplasious synapomorphies of this clade. the artifi cial nature of these taxa and the inadequacy of the mor- A compensatory base change (A-U→ G-C) in base pair 38 of phological features used to defi ne them. As a fi rst step to char- Helix 49 (bp 47 of H1399; according to Gillespie et al. (2006) ) acterize the new clades, we mapped morphological features of differentiated all members of clade CO2 from the rest of the

← Fig. 2. The rbcL phylogeny of desmids (Desmidiaceae, Zygnematophyceae) based on 127 sequences (1339 nt, maximum likelihood [ML] topology, for model and model parameters see Table 1 ; two Penium spp. as an outgroup). Nodes are characterized by bootstrap percentages (BP) (≥ 50%) and posterior probabilities (PP) ( ≥0.95): neighbor joining (NJ[ML])/MP/Bayesian inference (BI). Branches with 100% BP in all methods and 1.00 PP are shown bold- face. For clade names see Results. Clades containing Cosmarium species are underlined. For species represented by more than one accession, strain data are given. Cosmarium undulatum , the type species of the genus, is in bold. 1086 American Journal of Botany [Vol. 95 September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1087 family ( Fig. 6B ). The A-U pair is conserved across the family associated with other clades (Euastrum , Micrasterias) or formed and the order Desmidiales making the CBC in clade CO2 an a novel clade (CO4). NHS for this clade. One reason for the rampant polyphyly of the genus Cosmar- ium lies in its vague diagnosis that lacks any distinct synapo- morphic or even typological characters. When describing DISCUSSION Cosmarium, Ralfs (1848) stressed the lack of some characters typical for other genera (e.g., incision at the apex, lateral lobes, In this study, two ribosomal genes (nu SSU rDNA and cp spines or processes), but he did not present a single character or LSU rDNA), the protein-coding rbcL , and the noncoding 1506 a combination of features that is unique to the genus. We have group I intron of the SSU rDNA were used to assess phyloge- shown that the morphology of the genus Cosmarium as defi ned netic relationships among species of the Desmidiaceae (Zygne- by Ralfs is one of the most common in the family and has likely matophyceae, Streptophyta). The phylogenetic structure of the arisen independently in several lineages ( Fig. 5 ). Its evolution- most species-rich genus in the family, Cosmarium , and its puta- ary status (ancestral vs. derived) could be also different. Note- tive relatives, Actinotaenium , Euastrum , Staurodesmus , and worthy in this respect is the close relationship between Xanthidium was the major focus of our analyses. A large data smooth-walled Cosmarium and spine-bearing Staurodesmus set and extensive taxon sampling revealed 10 distinct clades in taxa in their common clades. It is very likely that in the STD2 the Desmidiaceae. clade, a triradiate cell with a stout spine at each angle, typical Most of the clades established here were well supported for the genus Staurodesmus, was derived from a smooth-walled (Figs. 2, 3). Comparison of individual topologies and their biradiate “Cosmarium ” taxon ( Figs. 2 – 4 ), whereas in the AR- support values showed that the different markers used THR clade spines may have been lost in some members, in contained a similar phylogenetic signal and yielded largely which case the “ Cosmarium ” -type morphology may be the de- congruent phylogenetic relationships between desmid taxa rived state ( Fig. 2 ). ( Table 2 ). In general, the rbcL gene provided better support The presence of several morphotypes in most Cosmarium - for most of the clades but was less informative in resolving containing clades demonstrates homoplasy of morphological relationships among species within clades (e.g., CO2), features thought to be important for desmid taxonomy. Tradi- whereas divergence of the nu SSU rDNA and the 1506 group tionally, semicell shape, degree of radiation (biradiate vs. omni- I intron was sometimes too high to recover a clade (e.g., CO3, radiate), cell wall features (smooth vs. ornamented), and STD2; Table 2). Differences in evolutionary rates between chloroplast morphology have been used to distinguish intrage- markers were particularly pronounced in clade CO2, in which neric taxa in Cosmarium ( de Bary, 1858 ; Lundell, 1871; Hans- divergence of rbcL at the species level was particularly low girg, 1888 ; De Toni, 1889 ; Raciborski, 1889 ; Turner, 1892 ; compared to SSU rDNA (compare Figs. 2 and 3). The concat- West and West, 1905 ; Hirano, 1959a , b ; Bourrelly, 1966 ). enated analyses of SSU rDNA and rbcL resulted in increased However, our study revealed that each of these morphological support for clades compared to phylogenetic analyses of indi- features has a mosaic distribution in the tree and does not ex- vidual genes ( Table 2 ). Relationships among clades, however, plicitly characterize a specifi c clade ( Fig. 5 ). Our phylogenies remained largely unresolved even when fast and slow evolv- revealed a patchy distribution of the omniradiate morphotype ing sequences were combined (Fig. 4). In the latter case, over our trees in several unrelated clades (Figs. 1– 4 ). More- though, taxon sampling was smaller, and phylogenetic reso- over, we found that some species that differ in their degree of lution may be much better, when taxon sampling is increased radiation have nearly identical sequences (e.g., Cosmarium to the same level as in the concatenated SSU rDNA and rbcL portianum [biradiate]/C. bisphaericum [omniradiate] in the analyses. omniradiate clade; C. lundelli and C. pachydermum [biradiate]/Actinotaenium turgidum [omniradiate] in CO2 (Fig. Phylogeny of Cosmarium— Not surprisingly, the results 2 ), and C. connatum [biradiate]/C. pseudoconnatum [pseudo- presented here confi rmed the long-anticipated artifi cial nature omniradiate] in CO3, Fig. 2 ). The latter case is a good example of the genus (e.g., West and West, 1905 , 1908 ; Fritsch, 1953 , for the likely derived nature of the omniradiate morphotype at Krieger and Gerloff, 1962 ; Prescott et al., 1981 ). Early molecu- least in some Cosmarium strains. In conclusion, the degree of lar phylogenetic studies were indicative in this respect, but the radiation is a homoplasious feature in the family Desmidiaceae limited taxon sampling, presence of long-branched taxa, and and its taxonomic signifi cance has been largely overestimated. the relatively low phylogenetic signal of the markers made the Morphological heterogeneity of all clades containing Cos- conclusions preliminary ( Lee, 2001 ; Nam and Lee, 2001 ; marium taxa (Fig. 5) and lack of information on other pheno- Gontcharov et al., 2003; Moon and Lee, 2003; Gontcharov and typical features currently does not allow conclusions to be Melkonian, 2005; Hall et al., 2008). In the current study, 45 (68 drawn on the morphological characters that may unite their in the rbcL analyses) species/strains of Cosmarium were dis- members or, more importantly, that differentiate the clades. tributed over six clades, one nonsupported assemblage (multi- The characters traditionally used to defi ne taxa above the spe- cellular), and a long-branched basal lineage ( Fig. 3 ). Four of the cies level are not suitable for this purpose, in particular features six clades also included species of other desmid genera. In the of cell wall ornamentation and chloroplast morphology (Fig. 5). rbcL analyses with 127 taxa, additional species of Cosmarium Many Cosmarium species have distinct patterns of granules,

← Fig. 3. Phylogeny of Desmidiaceae (Zygnematophyceae) based on combined analyses of nu SSU rDNA, 1506 group I intron and rbcL sequences (97 taxa, 3260 nt, maximum likelihood [ML] topology, for model and model parameters see Table 1 ). The tree was rooted with two Penium spp. Nodes are characterized by bootstrap percentages (BP ≥ 50%) and Bayesian posterior probabilities (PP ≥ 0.95): neighbor joining (NJ[ML])/MP/Bayesian inference (BI). Branches with 100% BP in all methods and 1.00 PP are boldfaced. See Fig. 2 for further details. 1088 American Journal of Botany [Vol. 95 September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1089 spines, or warts on the cell surface, and it is not clear whether Cosmarium species from other clades that have no affi nity to these types and patterns are homologous and what features CO2/CO3 should in the future either be classifi ed together with would represent character states here. Also very little is known representatives of the genera in their clades or recognized as about such a distinct characteristic of the desmid cell wall as the new genera. pores, their types, functions, signifi cance of different distribu- The omniradiate clade is phylogenetically most distant from tion patterns. Preliminary studies already revealed quite a di- the rest of the genus Cosmarium and is likely one of the basal versity of pore patterns (Cout é and Tell, 1981; Neuhaus and branches in the Desmidiaceae ( Gontcharov et al., 2003 , 2004 ; Kiermayer, 1981; Coesel, 1984; Gontcharov et al., 2002), how- this study). Like other clades containing Cosmarium spp., this ever, their possible suitability for desmid taxonomy above the clade includes a number of morphotypes distinct in cell/semi- species level is virtually unknown. cell shape, chloroplast morphology, and cell wall ornamenta- Our knowledge about the diversity of chloroplast structures tion ( Fig. 5 ). Membership of two omniradiate Actinotaenium in desmids is also very limited and has not been extended sig- (formerly Penium) species with elongated, weakly constricted nifi cantly beyond the careful early studies of L ü tkem ü ller cells and Spondylosium (formerly Cosmarium ) panduriforme (1893 , 1895 ) and Carter (1919a , b , 1920a , b ). Teiling (1952) forming short fi laments further complicates the circumscription typifi ed chloroplast shapes of desmids and presented a scenario of the clade. However, we discovered several nonhomoplasious of their hypothetical evolution correlating it with cell morphol- molecular synapomorphies (NHS; for defi nition, see Marin et ogy, mostly radiation, but chloroplast morphology is still poorly al., 2003) in the SSU rDNA that discriminate members of the known in many Cosmarium species. omniradiate and CO2 clades from all other taxa in the family Desmidiaceae and that might be useful in future taxonomic re- Type species of Cosmarium— Affi liation of a type species is visions ( Fig. 6 ). crucial for the identity of a genus. In Cosmarium this issue is The taxonomic affi nity of Cosmarium species belonging to complicated by the uncertainty with the designation of the type the clades ARTHR and STD2, and the nonsupported assem- species. The Index Nominum Genericorum (ING; http://botany. blage multicellular, is not yet clear. The distinctness of the si.edu/ing/) recognizes C. margaritiferum as the type although ARTHR and STD2 clades from STD1 that contains the type in the earlier version of ING C. undulatum had been suggested species of the genus Staurodesmus, Std. triangularis ( Gontcharov (Silva, 1952). The designation of C. margaritiferum as the type and Melkonian, 2005 ), received further support with the ex- is credited to N ä geli (1849, p. 114). However, Nä geli consid- tended taxon sampling and the larger data sets in this study. The ered Cosmarium as the subgenus of Euastrum and referred to E. current lack of synapomorphic phenotypic characters for these margaritiferum Ehr. Ehrenberg (1835) regarded his alga identi- clades calls for further phenotypic study before taxonomic con- cal with Ursinella margaritifera Turpin (1820) , so the correct clusions should be made. citation should be Euastrum margaritiferum (Turpin) Ehr. Be- Within the assemblage of multicellular desmids, the Cos- cause the publications by Turpin and Ehrenberg were published marium species either showed affi nity to the colonial Heiman- before the starting point of desmid taxonomy ( Ralfs, 1848 ), sia ( C. sinostegos ), the fi lamentous Spondylosium pulchellum N ä geli ’ s designation of the type species is invalid (ICBN, Art. ( C. regnellii) or formed an independent lineage (C. diffi cile , C. 7.7). Obviously, Nä geli was not familiar with Ralfs’ s publica- dilatatum; Figs. 2, 3). Most taxa comprising this cluster are tion at that time, and his typifi cation had no relation to the ge- characterized by fast evolutionary rates in all three genes and nus Cosmarium Corda ex Ralfs. were long-branched in the individual as well as the concate- Moreover, the alga described by Turpin is not identifi able nated analyses ( Figs. 2 – 4 ). One may hypothesize that the mor- and obviously not identical to two or likely three species illus- phological diversifi cation of Desmidiaceae from a unicellular trated by Ralfs under the name C. margaritiferum (1848, tables to a multicellular life habit, a consequence of a peculiar cell XVI, XXXIII; fi gs. 2a – d , 3 a, b). In contrast, the choice of C. division mode ( Gerrath, 1970 , 1973 ; Krupp and Lang, 1985a , undulatum as the type of Cosmarium was prompted by the fact b ), may have been accompanied by accelerated rates of evolu- that it is the most clearly known of the species included in the tion in several genes. These long branches likely affected to- genus by Corda ( Silva, 1952 ). We agree with Silva that N ä geli ’ s pologies and may have been responsible for lack of support for typifi cation should be rejected because it is based on an inval- this assemblage in almost all data sets ( Table 2 ). idly published name, and following Silva (1952), we regard C. Identifi cation of monophyletic entities consisting of species undulatum as the type species of the genus Cosmarium . of the traditional genus Cosmarium should stimulate the search Our analyses placed C. undulatum in the CO2 clade ( Figs. for synapomorphic phenotypic characters distinguishing these 2– 4 ) linking the genus name to this clade. The CO2 clade is a lineages. member of a large, weakly supported assemblage that unites two more clades, CO3 and the genus Xanthidium ( Fig. 4 ). The Euastrum— The current study substantiated the polyphyletic species richness of the CO2 and CO3 clades and the diversity nature of the traditional genus Euastrum ( Hall et al., 2008 ) as- of morphotypes in the clades suggest that these lineages signing its species to several clades (see Results). Most Euas- will accommodate the majority of the existing Cosmarium trum taxa studied formed an assemblage that received low or no species. support with the different data sets ( Figs. 2 – 4 ; Table 2 ). The

← Fig. 4. Phylogenetic tree of 40 species representing major lineages (see Results) of desmids based on comparisons of concatenated nu SSU rDNA (including 1506 group I intron), cp rbcL and LSU rDNA sequences (5509 aligned nt). The tree was constructed with maximum likelihood (ML) (GTR+I+G, for model parameters, see Table 1 ). BP ≥ 50% for ML/neighbor joining (NJ)(GTR+I+G)/MP and PP ≥ 0.95 (Bayesian inference) values are given for the nodes. Branches with bootstrap percentages (BP) 100% in all methods and 1.00 posterior probabilities (PP) are boldfaced. Presence of group I introns in the cp LSU rDNA is indicated by asterisk and their positions relative to Escherichia coli 23S rDNA are given. 1090 American Journal of Botany [Vol. 95 September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1091

Fig. 6. Nonhomoplasious synapomorphies in the SSU rRNA molecule characterizing the “ Cosmarium ” clades (A) omniradiate and (B) CO2 of the Desmidiaceae (see Results). Alignments contained only representative taxa for the clades. Taxa and nucleotides characterized by the synapomorphy are boldfaced. SSU rRNA secondary structure after Wuyts et al. (2000) ; diagrams are based upon the last taxon in the alignment. The nomenclature of nucle- otides (nt, single stranded spacers and loops) and base pairs (bp, stem regions of Helices) depends on the polarity of the RNA: increasing numbers indicate the 5 ′ → 3 ′ direction. relatively high divergence of Euastrum sequences may have CO2 and CO4. In most cases, their phylogenetic position could been responsible for the uncertain status of this assemblage, so be accessed with rbcL only; therefore, we consider their posi- that it requires further study with a more comprehensive taxon tion as tentative. However, high support values for the clades sampling. It appears that the Euastrum assemblage is split into that include Euastrum substellatum , E. verrucosum , E. ger- two morphologically distinct lineages. One lineage comprises manicum , E. spinulosum , E. moebii, and E. prowsei ( Fig. 2 ) large-celled taxa ( > 50 – 60 µ m long) with smooth cell walls and suggest that these species are distinct from other Euastrum a more or less regular porous cell surface often provided with taxa. several large facial protrusions and excavations, e.g., E. ob- longum and E. affi ne. In these species, the polar lobe has a nar- Actinotaenium— This study confi rmed the polyphyletic sta- row vertical incision with parallel margins. A similar tus of the genus Actinotaenium as suggested previously morphology is typical for a number of other Euastrum taxa, and (Gontcharov et al., 2003). In our analyses, six Actinotaenium their affi liation with this lineage is anticipated (Gontcharov species were distributed among three only distantly related lin- et al., 2003 ). eages including the genus Penium (e.g., A. cruciferum ; Figs. The second lineage contains relatively small-celled (typi- 2– 4 ). Their position in the tree highlights the homoplasious na- cally < 50 µ m long) and variously ornamented taxa. Their apical ture of the characters used to defi ne the genus (elongated, om- incision is less pronounced and often appears as a V-shaped niradiate cells with two stellate [ A. cucurbita , A. cf. wollei ] or invagination. numerous parietal [A. turgidum ] chloroplasts and smooth cell Euastrum taxa placed outside the Euastrum assemblage are walls ( Teiling, 1954 ). large-celled, with granules or spines arranged in specifi c pat- The status of phenotypic characters in A. phymatosporum terns. Their semicells are divided into one or two basal lobes and A. silvae-nigrae , members of the omniradiate clade, and and a polar lobe, as is typical for the genus, but the polar lobe is particularly A. cruciferum, having a weak affi nity to one of the concave to nearly straight and lacks an apical incision. Species Penium clades (Figs. 2– 4 ), is unclear. The morphology of the with these characteristics were distributed between two puta- former species (elongate-elliptical, omniradiate cells with a tively related clades consisting mostly of Cosmarium taxa, i.e., very weak median constriction and axial chloroplasts) is similar

← Fig. 5. Bootstrap consensus topology of 10 major clades of desmids (Desmidiaceae) based on neighbor joining (NJ[ML]) analyses of rbcL sequence data ( Fig. 2 ) with mapped features of cell and chloroplast morphology traditionally used to defi ne genera. 1092 American Journal of Botany [Vol. 95 to that of the genus Penium in which they were placed until Chapman , R. L. , M. A. Buchheim , C. F. Delwiche, T. Friedl, V. A. R. relatively recently ( Kouwets and Coesel, 1984 ), and these char- Huss, K. G. Karol, L. A. Lewis, et al . 1998 . Molecular systematics acteristics may be pesiomorphic. of the green algae. I n D. E. Soltis and J. J. Doyle [ eds .], Molecular sys- The affi liation of the type species of the genus, A. curtum , tematics of plants , 508 – 540. Kluwer , Boston , Massachusetts, USA. remains currently unknown (Gontcharov et al., 2003), and taxo- Coesel , P. F. M. 1984 . Taxonomic implications of SEM revealed cell wall sculpturing in some small-sized desmid species (Chlorophyta, nomic revision of Actinotaenium must await clarifi cation of its Conjugatophyceae). Acta Botanica Neerlandica 33 : 385 – 398 . phylogenetic position. Coesel , P. F. M. , and K. J. Meesters . 2007 . Desmids of the lowlands. Mesotaeniaceae and Desmidiaceae of the European lowlands. KNNV Conclusions— Phylogenetic analyses based on several genes Publishing, Zeist, Netherlands. from two genomes and extensive taxon sampling confi rmed the Cout é , A. , and G. Tell . 1981 . Ultrastructure de la paroi cellulaire des anticipated artifi cial nature of the desmid genera Cosmarium, Desmidiacees au microscope electronique a balayage. Nova Hedwigia Euastrum, Staurodesmus, and Actinotaenium. Our results high- 68 ( Beiheft ): 1 – 228 . Croasdale , H. T. , and E. A. Flint . 1988 . Flora of New Zealand. light the inadequacy of morphological features such as cell and Freshwater algae, Chlorophyta, Desmids II. Caxton Press, semicell shape, degree of radiation, chloroplast morphology, Christchurch, New Zealand. and cell wall ornamentation, traditionally used to discriminate Darlu , P. , and G. Lecointre . 2002 . When does the incongruence length these genera and revealed numerous cases of homoplasy in dis- difference test fail? Molecular Biology and Evolution 19 : 432 – 437 . tantly related lineages. Many clades established during this De Bary , A. 1858 . Untersuchungen ü ber die Familie der Conjugaten study deserve to be recognized as new genera but cannot be (Zygnemeen und Desmidieen), F ö rstner, Leipzig, Germany. formally described in the framework of the current morpholo- Denboh , T. , D. Hendrayanti , and T. Ichimura . 2001 . Monophyly gy-based taxonomic concept of desmids because either their of the genus Closterium and the order Desmidiales (Charophyceae, phenotypical features are insuffi ciently known or they do not Chlorophyta) inferred from nuclear small subunit rDNA data. Journal differ from those of other clades and genera. Molecular signa- of Phycology 37 : 1063 – 1072 . tures may be used to circumscribe the new taxa in the future, De Rijk , P. , J. Wuyts , Y. Van De Peer , T. Winkelmans , and R. De Wachter . 2000 . The European large subunit ribosomal RNA data- but their utility is not straightforward in a group that accounts base. Nucleic Acids Research 28 : 177 – 178 . for more than 2000 species. Although many desmid strains are De Toni , G. B. 1889 . Sylloge algarum omnium hucusque cognitarum, vol. 1. available in culture collections, the strains still represent only a Chlorophyceae. Privately published, Padova, Italy. minute fraction of the taxa described, the unknown diversity of Dolphin , K. , R. Belshaw , C. D. L. Orme , and D. L. J. Quicke . 2000 . Noise desmids that may exist in the environment notwithstanding. and incongruence: Interpreting results of the incongruence length dif- Many more isolates of desmids are needed as well as refi ned ference test. Molecular Phylogenetics and Evolution 17 : 401 – 406 . analyses of morphological traits and additional molecular mark- Dowton , M. , and A. D. Austin . 2002 . Increased congruence does not ers. At the current state of knowledge, desmid systematics has necessarily indicate increased phylogenetic accuracy — The behavior arrived at the beginning. of the incongruence length difference test in mixed-model analyses. Systematic Biology 51 : 19 – 31 . Ehrenberg , G. C. 1835 . Dritter Beitrag zur Erkenntniss grosser LITERATURE CITED Organisation in der Richtung des kleinster Raumes. Abhandlungen Kö niglich Preuß ische Akademie der Wissenschaften zu Berlin 1833 : Besendahl , A. , and D. Bhattacharya . 1999 . Evolutionary analyses 149 – 336 . of small-subunit rDNA coding regions and the 1506 group I introns Farris , J. S. , M. K ä llersj ö , A. G. Kluge , and C. Bult . 1994 . Testing of Zygnematales (Charophyceae, Streptophyta). Journal of Phycology signifi cance of incongruence. Cladistics 10 : 315 – 319 . 35 : 560 – 569 . Felsenstein , J. 1985 . Confi dence limits on phylogenies: An approach us- Bhattacharya , D. , S. Damberger , B. Surek , and M. Melkonian . ing the bootstrap. Evolution 39 : 783 – 791 . 1996 . Primary and secondary structure analysis of the rDNA group I Fritsch , F. E. 1953 . Comparative studies in a polyphyletic group: The introns of the Zygnematales. Current Genetics 29 : 282 – 286 . Desmidiaceae. Proceedings of the Linnean Society of London 164 : Bhattacharya , D. , B. Surek , M. Rü sing , S. Damberger , and M. 258 – 280 . Melkonian. 1994 . Group I introns are inherited through common an- Galtier , N. , M. Gouy , and C. Gautier . 1996 . SeaView and Phylowin, cestry in the nuclear-encoded rRNA of Zygnematales (Chlorophyta). two graphic tools for sequence alignment and molecular phylogeny. Proceedings of the National Academy of Sciences, USA 91 : 9916 – 9920 . Computer Applications in the Biosciences 12 : 543 – 548 . Bourrelly , P. 1966 . Les algues d’ eau douce. Tome I, Les algues vertes. Gay , F. 1884 . Essai d ’ une Monographie locale des Conjug é es. Montpellier N. Boubee and Cie, Paris, France. Revue des Sciences Naturelle s, III 3: 187 – 228; 285 – 335. Brook , A. J. 1981 . The biology of desmids. Botanical Monographs 16. Gerrath , J. F. 1970 . Ultrastucture of the connecting strands in Blackwell, Oxford, UK. Cosmocladium saxonicum de Bary (Desmidiaceae) and a discussion Brook , A. J. , and L. R. Johnson . 2002 . Order Zygnemales. In D. M. of the taxonomy of the genus. Phycologia 9 : 209 – 215 . John, B. A. Whitton, and A. J. Brook [eds.], The freshwater algal lora Gerrath , J. F. 1973 . Notes on desmid ultrastructure. Nova Hedwigia 42 of the British Isles. An identifi cation guide to freshwater and terres- ( Beiheft ): 103 – 113 . trial algae, 479 – 593. Cambridge University Press, Cambridge, UK. Gerrath , J. F. 1993 . The biology of desmids: A decade of progress. In Carter , N. 1919a . Studies on the chloroplasts of desmids. I. Annals of F. E. Round and D. J. Chapman [eds.], Progress in phycological re- Botany 33 : 215 – 254 . search, vol. 9, 79 – 192. Biopress, Bristol, UK. Carter , N. 1919b . Studies on the chloroplasts of desmids. II. Annals of Gerrath , J. F. 2003 . Conjugating green algae and desmids. In J. D. Wehr Botany 33 : 295 – 304 . and R. G. Sheath [eds.], Freshwater algae of North America, ecology Carter , N. 1920a . Studies on the chloroplasts of desmids. III. The chlo- and classifi cation, 353 – 381. Academic Press, San Diego, California, roplasts of Cosmarium. Annals of Botany 34 : 265 – 285 . USA. Carter , N. 1920b . Studies on the chloroplasts of desmids. IV. Annals of Gillespie , J. J. , J. S. Johnston , J. J. Cannone , and R. R. Gutell . Botany 34 : 303 – 319 . 2006 . Characteristics of the nuclear (18S, 5.8S, 28S and 5S) and Chapman , R. L. , and M. A. Buchheim . 1992 . Green algae and the evo- mitochondrial (12S and 16S) rRNA genes of Apis mellifera (Insecta: lution of land plants: Inferences from nuclear-encoded rRNA gene Hymenoptera): Structure, organization, and retrotransposable ele- sequences. Bio Systems 28 : 127 – 137 . ments. Insect Molecular Biology 15 : 657 – 686 . September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1093

Gontcharov , A. A. , D. L. Findlay , H. J. Kling , and M. M. Watanabe . Lenzenweger , R. 1999 . Desmidiaceenfl ora von Ö sterreich, Teil 3. J. 2002 . Desmids (Desmidiales, Streptophyta) from the Experimental Cramer, Berlin, Stuttgart, Germany. Lakes Area, Ontario, Canada. The genera Actinotaenium and Lewis , L. A. , and R. M. McCourt . 2004 . Green algae and the origin of Cosmarium. Algological Studies 105 : 51 – 78 . land plants. American Journal of Botany 91 : 1535 – 1556 . Gontcharov , A. A. , B. Marin , and M. Melkonian . 2003 . Molecular Lundell , P. M. 1871 . De Desmidiaceis quae in Suecia inventae sunt, phylogeny of conjugating green algae (Zygnemophyceae, observationes criticae. Nova Acta Regiae Societatis Scientiarum Streptophyta) inferred from SSU rDNA sequence comparisons. Upsaliensis . Ser. 3 : 1 – 100 . Journal of Molecular Evolution 56 : 89 – 104 . Luo , W. , S. Pflugmacher , T. Pr ö schold , N. Walz , and L. Krienitz . Gontcharov , A. A. , B. Marin , and M. Melkonian . 2004 . Are com- 2006 . Genotype versus phenotype variability in Chlorella and bined analyses better than single gene phylogenies? A case study using Micractinium (Chlorophyta, Trebouxiophyceae). Protist 157 : 315 – 333 . SSU rDNA and rbcL sequence comparisons in the Zygnematophyceae Lü tkem ü ller , J. 1893 . Beobachtungen ü ber die Chlorophyllk ö rper einiger (Streptophyta). Molecular Biology and Evolution 21 : 612 – 624 . Desmidiaceen. Ö sterreichische botanische Zeitschrift 43: 5– 11, 41– 44. Gontcharov , A. A. , and M. Melkonian . 2005 . Molecular phylogeny of Lü tkem ü ller , J. 1895 . Ü ber die Gattung Spirotaenia Breb. Staurastrum Meyen ex Ralfs and related genera (Zygnematophyceae, Ö sterreichische botanische Zeitschrift 45: 1 – 6, 51 – 57, 88 – 94. Streptophyta) based on coding and noncoding rDNA sequence com- Marin , B. , M. Klingberg , and M. Melkonian . 1998 . Phylogenetic re- parisons. Journal of Phycology 41 : 887 – 899 . lationships among the Cryptophyta: Analysis of nuclear-encoded SSU Hall , J. D. , K. G. Karol , R. M. McCourt , and C. F. Delwiche . rRNA sequences support the monophyly of extant plastid-containing 2008 . Phylogeny of the conjugating green algae based on chloroplast lineages. Protist 149 : 265 – 276 . and mitochondrial nucleotide sequence data. Journal of Phycology Marin , B. , E. C. M. Nowack , and M. Melkonian . 2005 . A plastid in the 44 : 467 – 477 . making: Evidence for a second primary endosymbiosis. Protist 156 : Hansgirg , A. 1888 . Prodromus der Algenfl ora von B ö hmen. Archiv fü r 425 – 432 . die naturwissenschaftliche Landesdurchforschung von Bö hmen 6: Marin , B. , A. Palm , M. Klingberg , and M. Melkonian . 3 – 9, 97-288. 2003 . Phylogeny and taxonomic revision of plastid-containing eugle- Hillis , D. M. , and J. P. Huelsenbeck . 1992 . Signal, noise and reliability in nophytes based on SSU rDNA sequence comparisons and synapomor- molecular phylogenetic analyses. Journal of Heredity 83 : 189 – 195 . phic signatures in the SSU rRNA secondary structure. Protist 154 : Hirano , M. 1959a . Flora desmidiarum japonicarum, no. 5. Contributions 99 – 145 . from the Biological Laboratory , Kyoto University 7 : 226 – 301 . McCourt , R. M. , K. G. Karol ., J. Bell, K. M. Helm-Bychowski, Hirano , M. 1959b . Flora desmidiarum japonicarum, no. 6. Contributions A. Grajewska, M. F. Wojciechowski, and R. W. Hoshaw . from the Biological Laboratory , Kyoto University 9 : 302 – 386 . 2000 . Phylogeny of the conjugating green algae (Zygnemophyceae) Hoef-Emden , K. , B. Marin , and M. Melkonian . 2002 . Nuclear and nu- based on rbc L sequences. Journal of Phycology 36 : 747 – 758 . cleomorph SSU rDNA phylogeny in the Cryptophyta and the evo- McFadden , G. I. , and M. Melkonian . 1986 . Use of HEPES buffer lution of cryptophyte diversity. Journal of Molecular Evolution 55 : for microalgal culture media and fi xation for electron microscopy. 161 – 179 . Phycologia 25 : 551 – 557 . H ö ftberger , M. , and U. Meindl . 1993 . Cell differentiation, ultrastruc- Meindl , U. 1986 . Autonomous circular and radial motions of the nucleus ture and nuclear migration in the desmid Xanthidium armatum. Nova in Pleurenterium tumidum and their relation to cytoskeletal elements Hedwigia 56 : 75 – 88 . and the plasma membrane. Protoplasma 135 : 50 – 66 . Huelsenbeck , J. P. , and F. Ronquist . 2001 . MrBayes: Bayesian in- Moon , B. , and O.-M. Lee . 2003 . A phylogenetic signifi cance of several ference of phylogenetic trees. Bioinformatics (Oxford, England) 17 : species from genus Cosmarium (Chlorophyta) of Korea based on 754 – 755 . mitochondrial cox III gene sequences. Algae (Korean Phycological Karol , K. G. , R. M. McCourt , M. T. Cimino , and C. F. Delwiche . 2001 . The Society) 18 : 199 – 206 . closest living relatives of land plants. Science 294 : 2351 – 2353 . N ä geli , C. 1849 . Gattungen einzelliger Algen, physiologisch und sys- Kirchner , O. 1878 . Kryptogamenfl ora von Schlesien, II. Band, I. Heft. tematisch bearbeitet. Neue Denkschriften der Allgemeinen schweiz- Algen. Breslau, Germany. erischen Gesellschaft fü r die gesamten . Die Naturwissenschaften 8 : Kouwets , F. A. C. , and P. F. M. Coesel . 1984 . Taxonomic revision of 1 – 139 . the conjugatophycean family Peniaceae on the basis of cell wall ultra- Nam , M. , and O.-M. Lee. 2001 . A comparative study of the morphologi- structure. Journal of Phycology 20 : 555 – 562 . cal characters and sequence data of rbc L gene in Cosmarium species. Krieger , W. , and J. Gerloff . 1962 . Die Gattung Cosmarium. Liefer 1. J. Algae (Korean Phycological Society) 16 : 349 – 361 . Cramer, Weinheim, Germany. Neuhaus , G. , and O. Kiermayer . 1981 . Formation and distribution of Krieger , W. , and J. Gerloff . 1965 . Die Gattung Cosmarium . Liefer 2. cell wall pores in desmids. In O. Kiermayer [ed.], Cytomorphogenesis J. Cramer, Weinnheim, Germany. in plants, 216 – 228. Wien, New York, Springer, Germany. Krieger , W. , and J. Gerloff . 1969 . Die Gattung Cosmarium . Liefer Pickett-Heaps , J. D. 1975 . Green algae. Sinauer, Sunderland, 3+4. J. Cramer, Weinnheim, Germany. Massachusetts, USA. Krienitz , L. , E. H. Hegewald , D. Hepperle , V. A. R. Huss , T. Rohrs , Posada , D. , and K. A. Crandall . 1998 . MODELTEST: Testing the and M. Wolf . 2004 . Phylogenetic relationship of Chlorella model of DNA substitution. Bioinformatics (Oxford, England) 14 : and Parachlorella gen. nov. (Chlorophyta, Trebouxiophyceae). 817 – 818 . Phycologia 43 : 529 – 542 . Prescott , G. W. , C. E. M. Bicudo , and W. C. Vinyard . 1982 . A syn- Krupp , J. M. , and N. J. Lang . 1985a . Cell division and fi lament forma- opsis of North American desmids. Part II. Desmidiaceae. Section 4. tion in the desmid Bambusina brebissonii (Chlorophyta). Journal of University of Nebraska Press, Lincoln, USA. Phycology 21 : 16 – 25 . Prescott , G. W. , H. T. Croasdale , W. C. Vinyard , and C. E. M. Bicudo . Krupp , J. M. , and N. J. Lang . 1985b . Cell division and the role of 1981 . A synopsis of North American desmids, part II. Desmidiaceae, the primary wall in the fi lamentous desmid Onychonema laeve section 3. University of Nebraska Press, Lincoln, Nebraska, USA. (Chlorophyta). Journal of Phycology 21 : 316 – 322 . Quicke , D. L. J. , O. R. Jones , and D. R. Epstein. 2007 . Correcting Kusel-Fetzmann , E. , and M. Schagerl . 1992 . Verzeichnis der the problem of false incongruence due to noise imbalance in the Sammlung von Algen-Kulturen an der Abteilung f ü r Hydrobotanik incongruence length difference (ILD) test. Systematic Biology 5 6 : am Institut f ü r Pfl anzenphysiologie der Universit ä t Wien. Phyton 32 : 496 – 503 . 209 – 234 . Raciborski , M. 1889 . Desmidyje nowe. Pamietnik Akademii Umiejtnosci Lee , O.-M. 2001 . The nucleotide sequences variability in ITS and 5.8S re- w Krakowie 17: 73 – 113. gions of the nuclear rDNA among Cosmarium species. Algae (Korean Ralfs , J. 1848 . The British Desmidieae. Reeve, Benham and Reeve, Phycological Society) 16 : 129 – 136 . London, UK. 1094 American Journal of Botany [Vol. 95

Silva , P. C. 1952 . A review of nomenclatural conservation in the algae Turmel , M. , J. F. Pombert , P. Charlebois , C. Otis , and C. Lemieux . from the point of view of the type method. University of California. 2007 . The green algal ancestry of land plants as revealed by the Publications in Botany 25 : 381 – 395 . chloroplast genome. International Journal of Plant Sciences 168 : Surek , B. , U. Beemelmanns , M. Melkonian , and D. Bhattacharya . 679 – 689 . 1994 . Ribosomal RNA sequence comparisons demonstrate an evo- Turner , W. B. 1892 . Algae aquae dulcis Indiae orientalis. The freshwa- lutionary relationship between Zygnematales and charophytes. Plant ter algae (principally Desmidiaceae) of East India. Kongliga Svenska Systematics and Evolution 191 : 171 – 181 . Vetenskapsakademiens Handlingar 25 : 1 – 187 . Swofford , D. L. 2002 . PAUP*: Phylogenetic analysis using parsimony Turpin, P. J. F. 1820 . Dictionnaire des Sciences Naturelles. Planches, part (*and other methods), version 4b10. Sinauer, Sunderland, MA, USA. 2. R è gne organis é . V é g é taux acotyl é dons. Levrault, Paris, France. Teiling , E. 1952 . Evolutionary studies on the shape of the cell and of the West , W. , and G. S. West . 1905 . A monograph of the British Desmidiaceae, chloroplast in desmids. Botaniska Notiser 105 : 264 – 306 . vol. II. The Ray Society, London, UK. Teiling , E. 1954 . Actinotaenium , genus Desmidiacearum resuscitatum. West , W. , and G. S. West. 1908 . A monograph of the British Botaniska Notiser 1954 : 376 – 426 . Desmidiaceae, vol. III. The Ray Society, London, UK. Thornton , J. W. , and R. Desalle. 2000 . A new method to localize Wuyts , J. , P. De Rijk , Y. Van de Peer , G. Pison , P. Rousseeuw , and R. and test the signifi cance of incongruence: detecting domain shuf- De Wachter . 2000 . Comparative analysis of more than 3000 se- fl ing in the nuclear receptor superfamily. Systematic Biology 4 9 : quences reveals the existence of two pseudoknots in area V4 of eu- 183 – 201 . karyotic small subunit ribosomal RNA. Nucleic Acids Research 28 : Turmel , M. , V. Cote , C. Otis , J.-P. Mercier , M. W. Gray , K. M. 4698 – 4708 . Lonergan , and C. Lemieux . 1995 . Evolutionary transfer of ORF- Wuyts , J. , Y. Van de Peer , and R. De Wachter . 2001 . Distribution of containing group I introns between different subcellular compartments substitution rates and location of insertion sites in the tertiary structure (chloroplast and mitochondrion). Molecular Biology and Evolution of ribosomal RNA. Nucleic Acids Research 29 : 5017 – 5028 . 12 : 533 – 545 . Xia , X. , and Z. Xie . 2001 . DAMBE: Data analysis in molecular biology Turmel , M. , R. R. Gutell , J.-P. Mercier , C. Otis , and C. Lemieux . and evolution. Journal of Heredity 92 : 371 – 373 . 1993 . Analysis of the chloroplast large subunit ribosomal RNA gene Yoder , A. D. , J. A. Irwin , and B. A. Payseur . 2001 . Failure of the ILD from 17 Chlamydomonas taxa. Journal of Molecular Biology 232 : to determine data combinability for slow loris phylogeny. Systematic 446 – 467 . Biology 50 : 408 – 424 .

A ppendix 1. Strain information and EMBL/GenBank accession numbers for taxa used in this study. A dash ( — ) indicates the region was not sampled. New sequences are boldfaced. ACOI = Coimbra Collection of Algae, University of Coimbra, Portugal (http://www1.ci.uc.pt/botanica/ACOI.htm); ASW = Sammlung von Algen-Kulturen, University of Vienna, Austria (Kusel-Fetzmann and Schagerl, 1992); CCAC = Culture Collection of Algae at the University of Cologne, Germany (http://www.ccac.uni-koeln.de); M = Culture Collection Melkonian, Botanical Institute, University of Cologne, Germany (strains available upon request); NIES = Microbial Culture Collection at National Institute for Environmental Studies, Tsukuba, Japan (http://www.nies.go.jp/biology/mcc/home.htm); SAG = Sammlung von Algenkulturen, University of Gö ttingen, Germany (http://www.epsag.uni-goettingen.de/html/sag.html); SVCK = Sammlung von Conjugaten-Kulturen, University of Hamburg, Germany (http://www.biologie.uni-hamburg.de/b-online/d44_1/44_1.htm). Taxon names in parentheses correspond to those used in the culture collection catalogue.

Taxon ; strain; nu SSU rDNA+1506 group I intron (if a separate entry); rbcL ; cp LSU rDNA.

Actinotaenium cruciferum (de Bary) Teil.; M2025; AM920332 ; AM911235 ; AM919465 . C. granatum Br é b. in Ralfs; M2127; AM920364; AM911282; AM919466 . A. cucurbita (Ralfs) Teil.; M1199; AJ428099+AM910431 ; AM919445 . C. hammeri Reinsch; ACOI349; AM920386; AM911302 ; AM911236 ; AM919439 . A. phymatosporum (Nordst) Coes. et Kouwets; — . C. holmii Wille; M1211; AM920387; AM911303; AM919461 . C. M1368; AJ428088+AM910432; AM911233; AM919434 . A . ( Penium ) impressulum Elfv.; SVCK58; AM920362; AM911279 ; — . C. laeve silvae-nigrae (Raban.) Kouwets. et Coes. var. parallelum . (Krieg.) Rabenh.; SVCK35; AM920363; AM911280 ; — . C. lundellii Delp.; Kouwets. et Coes.; SVCK295; AM920333; AM911234 ; — . A. turgidum SVCK357; AJ428113+AM910446; AM911310 ; — . C. maculatum (Ralfs) Teil. ex Rouzicka et Pouzar; M1192; — ; AM911238 ; — . A. cf. Turn. ; SVCK422; — ; AM911315 ; — . C. margaritiferum Menegh. ex wollei (W. et G. S. West) Teil. ex Rouzicka et Pouzar; M2945; AM920384; Ralfs ; SVCK88; — ; AM911311 ; — . C. meneghinii Br é b. ex Ralfs; AM911237 ; — . Bambusina borreri (brebissonii) (Ralfs) Cl.; CCAC SVCK59; AM920366; AM911284; AM919458 . C. notabile de Bary var. 0045; AJ428118+AM910433 ; AJ553935; AM919437 . Cosmarium medium (Gutw.) Krieg. et Gerloff; ACOI936; AM920369; AM911287 ; amoenum Br é b. in Ralfs; M1172; — ; AM911322 ; — . C. angulosum — . C. obsoletum (Hantz.) Reinsch; M2303; — ; AM911277 ; — . C. Br é b. ; ACOI378; — ; AM911328 ; — . C. binum Nordst.; ACOI896; — ; obtusatum Schidle; M2275; AM920389; AM911305 ; — . C. ochthodes AM911329 ; — . C. bioculatum Br é b. ex Ralfs; CCAP612/17; AM920354; Nordst.; M1205; AM920376; AM911293 ; — . C. ornatum Ralfs; AM911265 ; — . C. biretum Br é b. in Ralfs; M2123; AM920339; SVCK569; AM920372; AM911288 ; — . C. ovale Ralfs; SVCK342; AM911267; AM919440 . C. bisphaericum Printz; SVCK436; AM920338; AJ428114+ AM910447; AM911309 ; — . C. pachydermum Lund.; AM911266 ; — . C. blyttii Wille; CCAP612/19; AM920374; AM911290 ; SVCK24; — ; AM911321 ; — . C. perforatum Lund.; SVCK109; — ; — . C. botrytis Menegh. ex Ralfs; SVCK274; AM920378; AM911295; AM911318 ; — . C. phaseolus Bré b. in Ralfs; M2302; AM920382; AM919462 . C. broomei Thwaites ex Ralfs; M2075; AM920340; AM911299 ; — . C. portianum Archer; M2560; AM920337; AM911273; AM911269 ; — . C. caelatum Ralfs; ACOI826; — ; AM911319 ; — . C. AM919450 . C. protractum (N ä g.) de Bary; SVCK460; AM920381; connatum Br é b. in Ralfs; ACOI1152; — ; AM911317 ; — . C. contractum AM911298 ; — . C. pseudoconnatum Nordst. ; M1272; — ; AM911316 ; Kirchn.; SVCK396; AJ428112+AJ829661; AJ553937; — . C. cf. — . C. pseudonitidum Nordst.; ACOI1160; AM920377; AM911294 ; — . contractum ( hians ) ; NIES452; AM920365; AM911283; AM919459 . C. C. punctulatum Br é b.; M2717; AM920383; AM911300; AM919464 . C. crenatum Ralfs ex Ralfs; M2164; AM920370; AM911268 ; — . C. punctulatum; SVCK570; AM920373; AM911289; AM919451 . C. cucumis Corda ex Ralfs; M2715; AM920334; AM911270; AM919442 . quadratum Ralfs ; SVCK484; — ; AM911323 ; — . C. quadratum; M2946; C. cyclicum Lund.; M1208; AM920388 ; AM911304 ; — . C. decedens — ; AM911313 ; — . C. quadrum Lund. var. sublatum (Nordst.) W. et G. S. (Reinsch) Racib. ; ACOI794; — ; AM911330 ; — . C. depressum (N ä g.) West f. dilatatum Scott et Gr ö nbl.; ACOI368; AM920335; AM911272 ; Lund. ; ACOI1030; — ; AM911325 ; — . C. depressum; ; AM920367; — . C. ralfsii Br é b. in Ralfs ; SVCK300; — ; AM911324 ; — . C. regnellii AM911285 ; — . C. diffi cile L ü tkem.; ACOI403; — ; AM911326 ; — . C. Wille; M2947; AM920350; AM911276 ; — . C. reniforme (Ralfs) Arch.; dilatatum L ü tkem. et Gr ö nbl.; SVCK463; AJ829665; AM911274 ; — . C. SVCK34; AM920336 ; AY964179; AM919447 . C. sinostegos Schaarschm. elegantissimum Lund.; M1887; AJ428115+ AM910434; AM911271; var. obtusius Gutw.; ACOI406; AM920349; AM911275; AM919435 . C. September 2008] Gontcharov and Melkonian — The genus C OSMARIUM Corda ex Ralfs 1095

sportella Br é b.; M2152; AM920390; AM911306 ; — . C. subcrenatum AJ553930+ AM910438 ; AJ553959; AM919467 . P. exiguum W. West; Hantz.; M1200; AM920368; AM911286 ; — . C. subcucumis Schmidle; M2159 ; AJ553929+AM910439 ; AJ553960; — . P. margaritaceum (Ehr.) ACOI103; — ; AM911312 ; — . C. subgranatum (Nordst.) L ü tkem. ; ex Br é b. in Ralfs; SAG 22.82; AF115440; AM911254; AM919469 . P. M2629; — ; AM911281 ; — . C. subochthodes Schmidle; ACOI377; polymorphum (Perty) Perty; M2335; AM920331; AM911255; AM919468 . AM920379; AM911296 ; — . C. subprotumidum Nordst.; SVCK373; P. spirostriolatum Barker.; SVCK189; AJ553928+ AM910440 ; AJ553961; AM920375; AM911292 ; — . C. tesselatum (Delp.) Nordst.; SVCK381; — . Phymatodocis nordstedtiana Wolle; SVCK327; AJ428122+ AM910449 ; — ; AM911320 ; — . C. tetraophthalmum K ü tz. ex Ralfs; SVCK220; — ; AJ553962; — . Spondylosium panduriforme (Heimerl) Teil.; SAG 52.88; AM911314 ; — . C. tinctum Ralfs; M2301; AM920355; AM911278; AJ428124+ AM910441 ; AJ553969; AM919449 . S. planum (Wolle) W. et AM919455 . C. trachypleurum Lund.; ACOI935; AM920391; AM911307 ; G. S. West; SAG 41.81; AJ428123+AM910442; AM911260; AM919441 . — . C. trilobulatum Reinsch; ACOI866; AM920385; AM911301 ; — . C. S. pulchellum Arch.; SVCK365; AJ428130; AM911261 ; — . S. pulchrum undulatum Corda ex Ralfs; SVCK482; AM920380; AM911297; (Bail.) Arch.; SVCK331; AJ428129+AM910443 ; AJ553970.1; AM919463 . C. vexatum West; M2119; AM920392; AM911308 ; — . AM919436 . S. secedens (de Bary) Arch.; SVCK31; AJ428128+AM910445; Cosmarium sp.; M2093; AM920394; FM163363; — . Cosmarium sp.; AM911259 ; — . Staurastrum lunatum Ralfs; SVCK15; M2856; — ; AM911327 ; — . Cosmarium sp.; M2731; AM920395; AJ428106+AJ829640; AJ553971; AM919431 . S. orbiculare Ehr. ex AM911291 ; — . Euastrum affi ne ( humerosum) Ralfs; SVCK185; Ralfs; M2217; AJ829660; AM911331; AM919456 . S. sebaldi Reinsch; AM920342; AM911240; AM919432 . E. bidentatum N ä g.; ACOI282; M1133; AJ829630; AM911332 ; — . S. tumidum Bré b. ex Ralfs; SVCK85; AM920345; AM911243 ; — . E. binale Ralfs. var. gutwinskii (Schm.) AJ428108+AJ829666; AJ553972; AM919452 . Staurodesmus bienianus Homfeld; ACOI488; AM920347; AM911245 ; — . E. biverrucosum (Rabenh.) Florin; M1130; AJ829659; AM911341; AM919457 . S. Gontcharov et Watanabe; SVCK464; AM920346; AM911244 ; — . E. brevispina (Bré b. ex Ralfs) Croasd. var. obversus (West) Croasd.; divaricatum Lund.; SVCK 156; AM920344; AM911242; AM919444 . E. ACOI881; AM920361; AM911340 ; — . S. convergens (Ehr. ex Ralfs) germanicum (Schmidle) W. Krieg.; SVCK461; — ; AM911251 ; — . E. Lillier; M1886; AJ428102+AJ829651; AM911344 ; AM919460 . S. moebii (Borge) Scott et Prescott; SVCK358; AM920393; AM911248 ; — . extensus (Borge) Teil.; ACOI956; AM920358; AM911337 ; — . S. extensus E. oblongum (Grev.) Ralfs; ASW 07018; AJ428095+AM910435; var. joshuae (Gutw.) Teil.; ACOI1000; AM920360; AM911339 ; — . S. AM911239 ; — . E. prowsei Scott et Prescott ; SVCK353; — ; AM911249 ; glaber (Ehr.) Teil.; ACOI954; AM920356; AM911334 ; — . S. — . E. spinulosum Delp. var. henriquesii Sampaio ; ACOI1092; — ; ( Staurastrum ) isthmosus (Heimerl) Croasd.; SVCK466; AM920359; AM911252 ; — . E. subalpinum Messik.; ACOI855; AM920348; AM911338 ; — . S. mucronatus (Ralfs ex Br é b.) Croasd.; M1394; AM911246; AM919432 . E. subhexalobum W. et G. S. West ; ACOI1671; AJ428103+AJ829658; AM911342 ; — . S. omearii (Arch.) Thom.; M0751; — ; AM911253 ; — . E. substellatum Nordst.; SVCK364; AM920371; AJ829655; AM911333; AM919453 . S. spencerianus (Mask.) Teil.; AM911247; AM919446 . E. trigiberrum W. et G. S. West; ACOI1174; ACOI735; AM920357; AM911335 ; — . S. ( Arthrodesmus ) triangularis AM920343; AM911241 ; — . E. verrucosum Lund. ; SVCK798; — ; (Lagerh.) Teil.; SVCK280; AJ829654; AM911336; AM919454 . S. AM911250 ; — . Groenbladia neglecta (Racib.) Teil.; SVCK478; ( Arthrodesmus ) validus (W. et G. S. West) Scott et Gr ö nbl.; SVCK457; AJ428119+ AM910436 ; AJ553943; — . Haplotaenium (Pleurotaenium) AJ829653; AM911343 ; — . Triploceras gracile Bail.; SVCK173; minutum (Ralfs) Bando; SVCK302; AJ428090+ AM910437 ; AJ553947; AJ428089+ AM910444; AM911258; AM919448 . Xanthidium AM919438 . Heimansia (Cosmocladium) pusilla (Hilse) Coes.; SVCK428; antilopaeum var. canadense Josh.; SVCK147; AM920353; AM911264 ; AJ428125+ AM910448 ; AJ553948; — . Micrasterias fi mbriata Ralfs; — . X. armatum (Br é b.) Ralfs; ASW 07059; AJ428094+AJ829667; M1188; AJ428098+AM910450; AM911257 ; — . M. cf. thomasiana Arch. DQ026262; — . X. cristatum Br é b. var. uncinatum Ralfs f. ornatum var. notata (Nordst.) Gronbl.; M2253; AM920341; AM911256; Jackson; SVCK426; AM920352; AM911263 ; — . X. subhastiferum W. AM919430 . Penium cylindrus (Ehr.) Bré b. ex Ralfs; ACOI780; West; CCAP 690/1; AM920351; AM911262; AM919443 .