Sequencing Type Material Resolves the Identity and Distribution of the Generitype Lithophyllum Incrustans, and Related European Species L

SEQUENCING TYPE MATERIAL RESOLVES THE IDENTITY AND DISTRIBUTION OF THE GENERITYPE LITHOPHYLLUM INCRUSTANS, AND RELATED EUROPEAN SPECIES L. HIBERNICUM AND L. BATHYPORUM (CORALLINALES, RHODOPHYTA)1

Jazmin J. Hernandez-Kantun2 Botany Department, National Museum of Natural History, Smithsonian Institution, MRC 166 PO Box 37012, Washington District of Columbia, USA Irish Seaweed Research Group, Ryan Institute, National University of Ireland, University Road, Galway Ireland Fabio Rindi Dipartimento di Scienze della Vita e dell’Ambiente, Universita Politecnica delle Marche, Via Brecce Bianche, Ancona 60131, Italy Walter H. Adey Botany Department, National Museum of Natural History, Smithsonian Institution, MRC 166 PO Box 37012, Washington District of Columbia, USA Svenja Heesch Irish Seaweed Research Group, Ryan Institute, National University of Ireland, University Road, Galway Ireland Viviana Pena~ BIOCOST Research Group, Departamento de Bioloxıa Animal, Bioloxıa Vexetal e Ecoloxıa, Facultade de Ciencias, Universidade da Coruna,~ Campus de A Coruna,~ A Coruna~ 15071, Spain Equipe Exploration, Especes et Evolution, Institut de Systematique, Evolution, Biodiversite, UMR 7205 ISYEB CNRS, MNHN, UPMC, EPHE, Museum National d’Histoire Naturelle (MNHN), Sorbonne Universites, 57 rue Cuvier CP 39, Paris 75005, France Phycology Research Group, Ghent University, Krijgslaan 281, Building S8, Ghent 9000, Belgium Line Le Gall Equipe Exploration, Especes et Evolution, Institut de Systematique, Evolution, Biodiversite, UMR 7205 ISYEB CNRS, MNHN, UPMC, EPHE, Museum National d’Histoire Naturelle (MNHN), Sorbonne Universites, 57 rue Cuvier CP 39, Paris 75005, France and Paul W. Gabrielson Department of Biology and Herbarium, University of North Carolina, Chapel Hill, Coker Hall CB 3280, Chapel Hill, North Carolina 27599-3280, USA

DNA sequences from type material in the belonging in Lithophyllum. As well as occurring as a nongeniculate coralline genus Lithophyllum were subtidal rhodolith, L. hibernicum is a common, used to unambiguously link some European species epilithic and epizoic crust in the intertidal zone names to field-collected specimens, thus providing a from Ireland south to Mediterranean France. A set great advance over morpho-anatomical identifi- of four features distinguished L. incrustans from cation. In particular, sequence comparisons of rbcL, L. hibernicum, including epithallial cell diameter, COI and psbA genes from field-collected specimens pore canal shape of sporangial conceptacles and allowed the following conclusion: the generitype sporangium height and diameter. An rbcL sequence species, L. incrustans, occurs mostly as subtidal of the lectotype of Lithophyllum bathyporum, which rhodoliths and crusts on both Atlantic and was recently proposed to accommodate Atlantic Mediterranean coasts, and not as the common, NE intertidal collections of L. incrustans, corresponded Atlantic, epilithic, intertidal crust reported in the to a distinct taxon hitherto known only from literature. The heterotypic type material of Brittany as the subtidal, bisporangial, lectotype, but L. hibernicum was narrowed to one rhodolith also occurs intertidally in Atlantic Spain. Specimens from Ireland and France morpho-anatomically identified as L. fasciculatum and a specimen from 1 Received 3 November 2014. Accepted 7 May 2015. Cornwall likewise identified as L. duckerae were 2Author for correspondence: e-mails [email protected], [email protected] resolved as L. incrustans and L. hibernicum, Editorial Responsibility: C. Lane (Associate Editor) respectively.

791 792 JAZMIN J. HERNANDEZ-KANTUN ET AL.

Key index words: anatomy; Lithophyllum bathyporum; Mesophyllum lichenoides ( J. Ellis) Me. Lemoine, while Lithophyllum hibernicum; Lithophyllum incrustans; type material of L. decussatum, which consisted of psbA; rbcL; rhodolith; taxonomy; type specimens two small fragments, could not be assigned with confidence to a species or even a genus. Finally, Abbreviations: BI, Bayesian inference; BP, Bootstrap L. lichenoides is currently recognized morpho-ana- value; GTR, general time reversible; ML, maximum tomically as a heterotypic synonym of L. byssoides likelihood; psbA, Photosystem II D1 protein gene; rbcL, (Lamarck) Foslie (Woelkerling and Lamy 1998: 258- ribulose-1,5-bisphosphate carboxylase/oxygenase large 259). subunit gene In Brittany, Crouan and Crouan (1867) described Lithothamnion depressum P.L. Crouan & H.M. Crou- an; this taxon was transferred by Foslie (1898) to As with many other groups of organisms, one of L. incrustans f. depressum (P.L. Crouan & H.M. Crou- the most important and difficult tasks in current an) Foslie, and later attributed to Atlantic crustose taxonomic studies on algae is to link Linnean sys- forms of L. incrustans (Chamberlain and Irvine tem names with molecular phylogenetic clades 1994). Recently, Ballantine et al. (2011) proposed (Verbruggen 2014). Nowhere is this more evident the novel species name Lithophyllum bathyporum than in coralline red algae (subphylum Coral- Athanasiadis & D.L. Ballantine based on the lecto- linophycidae) and for two main reasons: (i) DNA typification of L. incrustans f. depressum (L. depres- sequencing studies have demonstrated a much sum) described in Chamberlain and Irvine (1994). higher diversity of both non-geniculate (Bittner L. incrustans has been recorded not only from the et al. 2011, Kato et al. 2011, 2013, Mateo-Cid et al. Mediterranean Sea (Hamel and Lemoine 1952, 2014, Pardo et al. 2014, Adey et al. 2015) and genic- Bressan and Babbini 2003) and Atlantic coasts of ulate (Hind and Saunders 2013, Hind et al. 2014b, Europe (Adey and Adey 1973, Ford et al. 1983, Pardo et al. 2015) species compared to previous Edyvean and Ford 1986, Chamberlain and Irvine morpho-anatomical studies, and (ii) morpho-ana- 1994), but also from South Africa (Chamberlain tomical characters have proved inadequate to segre- 1996) and the Caribbean Sea (Ballantine et al. gate species and even genera of coralline taxa 2011). It is reported to occur in a wide range of (Gabrielson et al. 2011, Kato et al. 2011, Martone habitats from the mid-intertidal through the subtid- et al. 2012, Hind and Saunders 2013, Hind et al. al and with diverse habits ranging from epizoic or 2014a,b, Pardo et al. 2014, Adey et al. 2015). The epilithic crusts on bedrock in and outside of pools, most unambiguous method to determine the cor- to rhodoliths (Chamberlain and Irvine 1994, Bres- rect application of a name is to amplify a DNA san and Babbini 2003, Ballantine et al. 2011). Keys sequence from a type specimen to compare with provided by Adey and Adey (1973), Chamberlain complementary sequences from field-collected mate- and Irvine (1994) and Bressan and Babbini (2003) rial (Hughey et al. 2001, Gabrielson 2008, Hind suggested that its habitat and morpho-anatomical et al. 2014a). This method has been applied success- features could be used to identify and separate fully for geniculate (Gabrielson et al. 2011, Hind L. incrustans from similar species. However, in a et al. 2014a,b) and non-geniculate corallines (Pena~ recent molecular study on rhodoliths (Hernandez- et al. 2014, Sissini et al. 2014, Adey et al. 2015, Her- Kantun et al. 2015) the morpho-anatomical identifi- nandez-Kantun et al. 2015). Herein, we sequenced cations of L. dentatum (Kutzing)€ Foslie, L. fascicula- type material to begin resolving the genus Lithophyl- tum (Lamarck) Foslie, and L. incrustans did not lum Phillipi, based on its generitype, L. incrustans match monophyletic clades identified by DNA Philippi (type locality: Sicily) and related species in sequencing, indicating the need for further studies Europe. to resolve the identity of these species. Lithophyllum incrustans, one of four species origi- The identity of the generitype L. incrustans based nally described by Philippi (1837), was designated on reliable molecular and morpho-anatomical fea- as the generitype by Foslie (1898). Philippi’s origi- tures is of crucial importance in defining the genus nal material was lost for over 140 years, until redis- Lithophyllum and for the circumscription of the sub- covered by Woelkerling (1983a,b) in the Nationaal family Lithophylloideae and family Corallinaceae. Herbarium Nederland (L). Along with L. incrustans, To date, molecular systematics studies included type material was also found of L. decussatum (Ellis DNA sequences assigned to L. incrustans without the & Solander) Philippi (based on Millepora decussatum corroboration based on type material (Bailey 1999, Ellis & Solander), L. expansum Philippi, and L. li- Harvey et al. 2003, Bailey et al. 2004, Broom et al. chenoides Philippi. After an exhaustive morpho-ana- 2008, Bittner et al. 2011, Kato et al. 2011, Richards tomical analysis of Philippi’s original material, et al. 2014, Hernandez-Kantun et al. 2015). Woelkerling (1983a) concluded that only L. incru- Taxa ascribed to the genus Lithophyllum (the type stans, the generitype, and L. lichenoides agreed with genus of the subfamily Lithophylloideae with 117 most of the features that define the modern conspecific/infraspecific names currently accepted, cept of Lithophyllum. Type material of L. expansum Guiry and Guiry 2014) represent a major ecological agreed morpho-anatomically with type material of group from tropical to temperate coastal areas, TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 793 demonstrating the importance of this genus in Recent collections were amplified, including negative and diversity studies and in understanding its phyloge- positive controls for rbcL, COI and psbA genes at NUIG and for netic relationships. In this study, we obtained diag- rbcL at NCU, while COI and some psbA sequences were ampli- nostic DNA sequences from the type material of fied at MNHN. DNA was extracted using either the NucleoSpinâ 96 Tissue kit (Macherey-Nagel, GmbH and Co. L. incrustans, L. hibernicum and L. bathyporum to KG, D-52313 Duren,€ Germany) following the manufacturer’s compare with comparable sequences from field-col- protocol or the Qiagen DNeasy Blood and Tissue Kitâ (Qiagen, lected specimens morphologically identified as the Crawley, UK) following the modified protocol of Broom et al. same species. Partial sequences of rbcL, COI and (2008), except for the type specimens of L. incrustans and psbA genes from type specimens and field collected L. hibernicum which were extracted using the protocol of Hug- samples were used to establish the identities of hey et al. (2001) as modified by Gabrielson et al. (2011) for coralline algae. The lectotype of L. bathyporum and the histori- molecular clades; longer sequences from rbcL and cal collection of L. depressum were extracted using QIAampâ psbA genes were used to resolve the phylogenetic DNA Micro Kit (Qiagen S.A.S., Les Ulis, France) following the affinities of these northeast Atlantic and western manufacturer’s protocol for tissues. DNA extractions and speci- Mediterranean Lithophyllum species. mens were linked with a code including the letter E and con- secutive numbers (e.g., E158 L. hibernicum BM Box 578) or VPF or LLG for those specimens deposited in SANT and PC, MATERIALS AND METHODS respectively; specimens analyzed at NCU were identified by herbarium accession number (Tables S1 and S2). Specimens. Samples were obtained from museum and field Three regions of the rbcL gene were amplified by PCR collections as detailed in Tables S1 and S2 in the Supporting from field collected material using the following pairs of Information. Collection sites were located in the Republic of primers: the 50 end (1,093 base pairs [bp]) was amplified with Ireland, Northern Ireland, England, France (Atlantic and the combination F57 (forward) - R1150 (reverse) and the 30 Mediterranean), Spain (Atlantic and Mediterranean), Portu- end (~700 bp) with the combination F753 (forward) - RrbcS gal, Italy, and Croatia (Tables S1 and S2). The field sampling start (reverse) (Freshwater and Rueness 1994). A 296 bp frag- strategy was designed to obtain a thorough coverage of geo- ment of rbcL(~20% of the gene in red algae) from each of graphic distributions, habitats (high, mid- and low intertidal the type specimens of L. incrustans, L. hibernicum, L. bathypo- and subtidal), and morphologies (e.g., branched and rum as well as the historical collection of L. depressum was unbranched rhodoliths, epilithic, epiphytic, and epizoic amplified using the primer combination F1150Cor (Sissini crusts) of Lithophyllum species. Samples were collected by et al. 2014) and RrbcS and following the protocol of Hughey hand or using hammer and chisel on rocky shores and on et al. (2001). The primers used obtained longer sequences, sand-covered rocks at low tide, and by dredging and/or but these were truncated at the end of the rbcL gene, result- SCUBA diving in the subtidal (Tables S1 and S2). Samples ing in a 296 bp sequence. The psbA gene was amplified in were air-dried for 1 d or more depending on their volume, one reaction using the primers psbA-F1 and psbA-R2 following then cleaned and stored in labeled bags with silica gel. Vou- the protocols of Yoon et al. (2002) and Bittner (2009). COI cher specimens were deposited in GALW, NCU, PC, SANT gene sequences were obtained using primers pair GazF1 and and US; museum collections included types and other histori- GazR1 (Saunders 2005), the primer pair GWSFn (forward; Le cal specimens preserved in BM, L, PC and TRH (Tables S1 Gall and Saunders 2010) and GWSRx (reverse, Saunders and and S2). Herbarium acronyms follow Thiers (2015) continu- McDevit 2012), or a combination of the forward primer ously updated). GazF1 with the reverse primer GCorR3 (Pena~ et al. 2015), DNA extractions, PCR amplification, and sequencing. Genomic DumR1 (Saunders 2005), or COX1R1 (Saunders 2008). DNA was extracted from specimens with features referable by Three different mixtures for PCR amplification were used morpho-anatomy to current concepts of Lithophyllum byssoides, for field-collected specimens depending on availability in the L. dentatum, L. duckerae, L. fasciculatum, L. hibernicum, L. incru- laboratory and the contribution by coauthors, i.e., the mix- stans, L. orbiculatum (Foslie) Foslie and L. racemus (Lamarck) tures according to Hernandez-Kantun et al. (2015), Pena~ Foslie. DNA extractions were performed using fragments that et al. (2014), or the following mixture, which in a total vol- were clean or were cleared of epiphytes. A total of 168 speci- ume of 25 lL contained 12.5 lL of GoTaqR Green Master mens were successfully sequenced. Mix (Promega Corporation, Madison, WI, USA, including Archival herbarium material was extracted and amplified buffers, taq polymerase, deoxynucleotide triphosphates and TM in different institutions: type material of L. incrustans (L MgCl2), 9 lL of HyPure Cell culture Grade Water (Thermo 943.10.34) and L. hibernicum (TRH A23-1399) at NCU by P. Scientific, Milwaukee, WI, USA), 2.5 lL of DNA template and Gabrielson in 2010, historical collections from BM at the 0.5 lL (10 pM) of each primer (forward and reverse). DNA National University of Ireland, Galway (NUIG) sequenced by templates were usually diluted (1/10), except in the amplifi- J. Hernandez-Kantun in 2010 and lectotype of L. bathyporum cations of the lectotype of L. bathyporum and of the historical (L. depressum – CO02511) plus a historical collection of this collection of L. depressum where undiluted DNA template was species (CO02513) sequenced by L. Le Gall & V. Pena~ at the used. A Techne TC-3000 and Techne TC-3100 thermal cyclers Museum National d’Histoire Naturelle, Paris (MNHN) in (Techne, Stone, UK) and GeneAmpâ PCR System 2700 2014. These extractions and amplifications were performed (Applied Biosystems, Carlsbad, CA, USA) thermal cyclers separate from field-collected specimens (excluding also posi- were employed for all PCR amplifications. tive controls during PCR runs), and they were accompanied Amplicons from fresh and archival material of expected by negative controls in every step. In no case did a negative length, yield and purity from NUIG and NCU were purified control present amplicons, while targeted amplicons showed using the QIAquick PCR Purification Kit (Qiagen). The no double peaks or inconsistencies for forward and reverse NUIG amplicons were sequenced commercially (GATC Bio- reads in the chromatograms. For L. incrustans, type material tech, Konstanz, Germany) using the same primer pairs as for was extracted and amplified at NCU, whereas all field-col- the PCR amplifications, while the NCU amplicons were lected material that proved to be this species was extracted sequenced at the DNA Analysis Core Facility, Center for Mar- and amplified in European labs. ine Sciences, University of North Carolina, Wilmington; PCR 794 JAZMIN J. HERNANDEZ-KANTUN ET AL. products from fresh and type material at MNHN were characters were observed by SEM in a fragment from each sequenced by Genoscope (Bibliotheque du Vivant program, voucher specimen. Each specimen was cut with a razor blade Centre National de Sequenc ßage, France), while those from perpendicular to the thallus surface in areas where concepta- the lectotype of L. bathyporum and historical collection of cles were visible. If the specimen appeared to lack concepta- L. depressum were sequenced by Eurofins (Eurofins Scientific, cles, the sectioned part was selected haphazardly. Fragments Courtaboeuf, France); PCR products from the type specimens were mounted with adhesive film on metal stubs and subse- of L. incrustans and L. hibernicum were sequenced as in Gabri- quently coated with gold. Observations were made on digital elson et al. (2011). Sequences were submitted to GenBank images obtained with a Hitachi S-4700 SEM (Hitachi manu- (accession numbers listed in Tables S1 and S2). facturing, Ibaraki, Japan) at the National Centre for Biomedi- Data sets and molecular analyses. Three data sets were built cal Engineering Science (NCBES), NUIG. Herein, the terms and hereafter called “rbcL-specimens,” “COI-specimens,” and rhodolith and maerl are used synonymously, and include “rbcL-psbA-concatenated.” The “rbcL-specimens” data set used specimens that completely cover a nucleus (e.g., stone, shell, sequences obtained in this study to match type specimens to etc.) and/or a system of branches or a sphere without an field-collected material for subsequent morpho-anatomical apparent nucleus. comparisons. This data set included all specimens with short sequences (296 bp, including type specimens), medium – sequences (700 1,093 bp) and long (1,353 bp) sequences of RESULTS the rbcL gene. After molecular clades were identified unambiguously by comparison with sequences from type speci- Identification of molecular clades using type mate- mens, short sequences that could affect phylogenetic analyses rial. rbcL-296 sequences were obtained from the were removed in later analyses (concatenated data sets). The type specimens of L. incrustans, L. hibernicum, and “COI-specimens” data set for DNA barcoding was built from complementary COI sequences of specimens already L. bathyporum (lectotype of L. depressum) and “L. de- sequenced for rbcL and from additional specimens for which pressum historical collection” CO02513 (alignment only COI sequences could be generated. The “rbcL-psbA-con- in Fig. S1 in the Supporting Information). For all catenated” data set, constructed from complementary species, most of the base pairs changes were sequences of both genes was used for phylogenetic analyses observed in third codon position; only L. incrustans after respective tree reconstructions, were compared for simi- presented two changes in first codon position. lar topologies to prevent bias when concatenating. L. bathyporum could not be included in the rbcL/psbA concatenated Analyses furthermore included specimens identified phylogenetic analysis as no 1,353 rbcL sequence was available. based on morpho-anatomical features as L. byssoides, Sequences of specimens identified by morpho-anatomy as L. orbiculatum and the Mediterranean species L. byssoides and L. orbiculatum were used as outgroup taxa. A L. dentatum and L. cf. racemus for which no type supplementary data set was compiled after obtaining a psbA material was sequenced. The UPGMA reconstruc- sequence from type material of L. bathyporum (see Table S1). tion of this data set, including rbcL-296 from the Data sets were aligned in ClustalW in Mega version 5 (Tam- type specimens of L. bathyporum, L. hibernicum, and ura et al. 2011) using default settings and refined by eye. This L. incrustans, unambiguously joined field-collected software was also employed to analyze the data sets “rbcL-specimens” and “COI-specimens” by UPGMA using uncorrected p- and type specimens (Fig. 1). The sequences of distances with 1,000 re-samplings for statistical support. L. hibernicum and “L. depressum historical collection” Phylogenetic reconstructions for the “rbcL-psbA-concate- were identical, and diverged by 2.8% from the lecto- nated” data set were performed using maximum likelihood type of L. bathyporum (Table 1). (ML) in RAxML 1.3 (Mac version, Silvestro and Michalak Molecular analyses. In all analyses of individual 2012). The data set was partitioned by marker plus codon genes, rbcL UPGMA distance analysis (Fig. 1), COI position (following Verbruggen et al. 2010). The general UPGMA distance analysis (Fig. S2 in the Supporting time reversible (GTR) model was applied with invariant sites and gamma distribution for each of the partitions in the con- Information), and psbA ML phylogenetic analysis catenated data set and statistical support was obtained from (Fig. S3 in the Supporting Information), L. incru- bootstrap (BP) analyses with 1,000 re-samplings. Bayesian stans, L. bathyporum and L. hibernicum are recognized inference (BI) was applied in MrBayes v. 3.2.2 (Huelsenbeck as distinct lineages. The rbcL distance analysis and Ronquist 2001) using model GTR+G+I for each of the revealed two groups among included specimens of partitions. Four Monte Carlo-Markov chains were run for 5 L. incrustans (Fig. 1) that differed by 0.46% million generations, and trees sampled every 1,000 generations. The stationary distribution of the runs was verified with (Table 1), while the same groups differed by 0.49% Tracer v.1.5 (Rambaut and Drummond 2007) before stop- with the psbA gene (Fig. S3). One group comprised ping the program; 1,250 trees were discarded as burn-in, only rhodoliths associated with maerl beds in Ireland using the remaining trees to build the 50% majority-rule con- and France (hereafter called E137 clade); the other, sensus trees. including the holotype specimen, included epizoic Morphological characterization. Morpho-anatomical charac- and epiphytic crustose specimens as well as rhodo- ters traditionally used to segregate species in Lithophyllum (Woelkerling and Campbell 1992, Chamberlain and Irvine liths (Fig. 1; Table S1). Nonetheless, the variation in 1994) were investigated to determine if any were diagnostic rbcL and psbA sequences was not correlated with for the species treated herein (Table S3 in the Supporting either habit or distribution (Fig. 1; Table S1). L. hib- Information). Specimens examined for morpho-anatomical ernicum occurred both as rhodoliths and epilithic features were only those from which rbcL or psbA gene crusts, intertidally and subtidally. L. bathyporum is sequences had been acquired. Because only three specimens currently only known from three specimens includ- of L. bathyporum were obtained from the type material and ing the lectotype, and all are epilithic crusts (Fig. 1; from the field, only a restricted number of anatomical features were obtained (e.g., bisporangia size). Anatomical Tables S1 and S2). In the analyses based on individ- TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 795 ual genes (Fig. 1, Fig. S2 and S3), L. bathyporum set showed high bootstrap support for terminal and formed a weakly supported clade with L. hibernicum. internal nodes (from 89/0.93 to 100/1; Fig. 2), with Likewise, the position of the psbA sequence from the one exception (the node subtending Lithophyllum lectotype of L. bathyporum (Fig. S3) was similar to sp. E149, see below). The L. incrustans clade was that of the rbcL reconstruction shown in Figure 1. well-supported (99/1) and sister to a major clade Congruence of the rbcL and psbA markers, previ- comprising three groups; one lineage contained ous to concatenation was reflected in equivalent Lithophyllum sp. E149 (unsupported), the second topologies and bootstrap supports for both phyloge- group covered specimens identified morpho- nies using RAxML and BI (Fig. S4 in the Support- anatomically as L. dentatum (E152, E219, E218, 93/ ing Information). In general the concatenated data 0.99 support), and the third included L. hibernicum

FIG. 1. Unweighted Pair Group Method with Arithmetic Mean (UPGMA) tree for Lithophyllum species based on rbcL sequences; short sequences of 296 bp (including type specimens), medium of 700– 1,093 bp, and long of 1,353 bp. Bootstrap percent shown at branch nodes; support values below 50% are marked as (-). Type specimens of L. incrustans, L. hibernicum, and L. bathyporum are in bold. Habit of each specimen as sphere = rhodolith and wave = crustose thalli. Non- type specimens identiﬁed by code E# or CO# in Table S1. 796 JAZMIN J. HERNANDEZ-KANTUN ET AL.

TABLE 1. Pair-wise distance of partial rbcL sequences comparing Lithophyllum species (Fig. 1). Distance recovered from sequences near to or 1,353 bp in length plus the rbcL-296 from type sequences.

Uncorrected pair-wise distance Within L. incrustans clade 0–0.46 (E137, E25, E231) Holotype L. incrustans vs E137/E25 0.46 Holotype L. incrustans vs E231 0 Holotype L. incrustans vs type 12.1 L. hibernicum Holotype L. incrustans vs Lithothamnion 12.1 depressum Historical collection Holotype L. incrustans vs Lectotype 10.7 L. bathyporum Within L. hibernicum clade 0–1.4 (E259, E211, E307, E215) Holotype L. hibernicum vs E259 0 Holotype L. hibernicum vs L. depressum 0 Historical collection Holotype L. hibernicum vs Lectotype 2.8 L. bathyporum Lectotype L. bathyporum vs L. depressum 2.8 Historical collection specimens (100/1 support). The remaining Litho- phyllum spp. appeared in fully supported clades morpho-anatomically identiﬁed as L. cf. racemus (E244, E255), L. orbiculatum/L. byssoides (E222+ E150, E72), and only L. byssoides (E150, E72) (Fig. 2).

Morphological features that distinguish L. incrustans FIG. 2. Phylogenetic reconstruction of Lithophyllum spp. based and L. hibernicum. A total of 31 morpho-anatomi- on concatenated rbcL+psbA gene sequences. First number at cal characters and their character states (Table S3) branch nodes is Bootstrap percent (BP) for maximum likelihood were examined to find autapomorphies for L. incru- analysis; second number is Posterior Probability (PP) for Bayesian Inference analysis; (*) represents clades fully supported in both stans and L. hibernicum. Eighteen characters did not reconstructions (100% BP/1.0 PP); support values below 50% are show any character state variation or the states were marked as (-); each specimen is identified by E#. ambiguous for one or both species. Thirteen characters were informative, but only four could be used Crodelia incrustans (Philippi) Heydrich 1911: 16 in combination to distinguish both species. These DNA sequences: The rbcL-296 sequence was diagnos- characters were: epithallial cell diameter, pore canal tic for this species, and see Tables S1 and S2 for shape, and sporangium height and diameter additional sequences of rbcL, psbA and COI. (Table 2). No morpho-anatomical differences were Morphology and Habitat: Non-geniculate coralline observed between the two groups of L. incrustans crust (0.6–1 cm thick) with or without protuber- revealed by the rbcL and psbA analyses. ances/excrescences or globular to sub-globular rho- Taxonomic account: Revised descriptions, habitats doliths (up to 10 cm in diameter), sparsely or and distributions based on sequenced specimens densely branched with compressed and/or cylindri- are provided below for L. incrustans, L. hibernicum cal branches (Fig. 4, A–G); surface smooth when and L. bathyporum. All cell measurements are lumen vegetative but irregular when conceptacle pores dimensions unless otherwise stated. Detailed distri- were present; color varied from almost white to butions of L. incrustans, L. hibernicum and L. bathypo- pink, violet to lilac-gray; rhodoliths occurred in sub- rum are provided in Figure 3. tidal maerl beds (to 27 m depth); crusts were sub- L. incrustans Philippi 1837: 388 (herein Figs. 4, tidal and epilithic, epizoic on dead or live molluscs, A–G and 5, A–I) or epiphytic on Laminaria hyperborea (Gunnerus) Lectotype: L 943.10.34 Thallus completely covering Foslie; rarely in the lower intertidal on the Atlantic a gastropod shell, designated by Woelkerling coast. (1983a; fig. 15). Anatomy: Anatomical features illustrated in Type locality: near Sicily, Italy. Figure 5, A–I. Thallus monomerous; hypothallus un- Homotypic Synonyms: Lithothamnion incrustans (Phi- istratose, attached to substratum by cellular adhe- lippi) Foslie 1895: 122 sion; in transverse view (Fig. 5B), primary pit Hyperantherella incrustans (Philippi) Heydrich connections observed between cells of the same fila- 1901: 190 ment; secondary pit connections observed between TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 797

TABLE 2. Shared and distinct features for Lithophyllum incrustans and L. hibernicum in the Atlantic and Western Mediterra- nean Europe.

Lithophyllum Vegetative L. incrustans L. hibernicum Nature of substratum relationship Rhodolith, less frequent Rhodolith, Epilithic, Epizoic as crust, epilithic, epizoic Habitat Subtidal, rarely intertidal Intertidal, rarely subtidal Thallus thickness 0.6–1 cm 0.10–0.5 cm Epithallial cell length (lm) 3–43–4 Epithallial cell diameter (lm) 6–8 3–4 Cell height in medulla (lm) 6–10 7–20 Cell diameter in medulla (lm) 4–64–7 Reproductive L. incrustans L. hibernicum Number of cells in the pore canal 7–98–23 Rim at the bottom of the pore canal Present Present Rim at the surface of the pore canal Absent Present/Absent Shape of the pore canal Triangular (D) or conical Tube-like (cylindrical) Cells from the floor to 14–22 11–31 surface in conceptacle chamber Columella presence/absence Present Present Roof of conceptacle protruding Absent Absent Conceptacle chamber diameter (lm) 260–352 264–380 Nature of spores in sporangia Usually Bi* Tetra/Bi Tetra/bisporangia height (lm) 100–115 Tetra = 90–150, Bi = 70–87 Tetra/bisporangia diameter (lm) 40–45 Tetra = 50–65, Bi = 30–40 Conceptacles buried within thallus Bi – Present Tetra and Bi – Present *Tetrasporangia only recorded in the type material. Underlined results denote diagnostic features to separate species. Bi, Bisporangia; Tetra, Tetrasporangia. cells of neighboring filaments; in surface view epi- 100–115 lm high and 40–45 lm in diameter thallial cells thick-walled and polygonal (Fig. 5A); in (Fig. 5H); old bisporangial conceptacles deeply bur- transverse view epithallus one to three cell layers ied in thallus (Fig. 5I). Woelkerling (1983a) thick and cells rectangular to flattened, lumen flat, reported tetrasporangial conceptacles in the holo- 3–4 lm high and 6–8 lm diameter, i.e., with thin type specimen. lateral cell walls (Fig. 5B); meristematic cells interca- Distribution: L. incrustans is confirmed from Ire- lary between epithallus and perithallus, ovoid, land, France (Mediterranean and Atlantic), Spain 6–15 lm high and 4–6 lm diameter (Fig. 5B); peri- (Mediterranean and Atlantic) and Italy (Sicily, the thallus formed by cells cut-off proximally from inter- type locality; Fig. 3). Reports of its presence in calary meristematic cells; perithallial cells 5–12 lm South Africa (Chamberlain 1996) and Puerto Rico high and 5–6 lm wide (in middle of rhodolith (Ballantine et al. 2011) based only on morpho-ana- branches 21–38 lm high and 7–14 lm wide); tricho- tomical features were likely incorrect and need to cytes not seen. Gametangial thalli not observed. be reassessed by DNA sequencing. Only uniporate, bisporangial conceptacles seen (no L. hibernicum Foslie 1906; 24 (herein Figs. 6, A–D tetrasporangial conceptacles, Fig. 5, C–I); in surface and 7, A–F) view conceptacle roof not protruding and pore Lectotype: TRH A23-1399, designated herein, and usually slightly lower than thallus surface (Fig. 5, see below under Comment. F–G) with pore up to 45–55 lm diameter and veil Type locality: Fahy Bay, Ballynakill Harbour, Co. covering entrance (Fig. 5E); in transverse view, bisp- Galway, Ireland (Foslie 1906). orangial conceptacle chambers dumbbell-shaped DNA sequences: The rbcL-296 sequence was diagnos- due to dissolution of peripherally of previously tic for this species, and see Tables S1 and S2 for formed perithallial carbonate, 40–180 lm high and additional rbcL, psbA and COI sequences. 260–352 lm wide (Fig. 5C); number of cells from Morphology and Habitat: Non-geniculate coralline chamber floor to thallus surface ranges from 14 to comprising spherical to fruticose rhodoliths (up to 22; calcified columella formed by pre-existing peri- 5 cm diameter) sparsely or densely branched with thallial filaments, without disruption at the concep- cup-shaped apices (Fig. 6A) to encrusting smooth tacle base (Fig. 5C); pore canals 7–9 cells long, 50– thalli (0.1–0.5 cm thick), becoming irregular when 80 lm diameter medially; both openings of pore conceptacle pores were present, or with numerous canal possessed rims (Fig. 5D); pore canal junct- excrescences particularly in confluent areas with ion with conceptacle chamber conical in shape neighboring thalli; specimens pale pink to dark or tapering (Fig. 5D); only bisporangia observed, pink and purple. In the Atlantic, occurred as a 798 JAZMIN J. HERNANDEZ-KANTUN ET AL.

FIG. 3. Sample locations and distributions of Lithophyllum incrustans, L. hibernicum and L. bathyporum based on DNA sequences produced in this study. Ireland: (1) Roundstone Bay, (2) Kingstown Bay (21 km from Ballynakill harbor, type locality of L. hibernicum), (3) Carna, (4) Black Head, (5) Mukinish, (6) Kilkee, (7) Ardgroom, (8) Roaringwater Bay, (9) Lough Hyne, (10) Crosshaven, (11) Dublin; England: (12) Gurnard’s Head, (13) Nanjizal Beach, (14) Hannafore Point; France: (15) Roscoff, (16) Rade de Brest (type locality of L. bathyporum), (17) Douarnenez, (18) Biarritz, (19) Marseille, (20) Port de Sete; Spain: (21) Zumaya, (22) Gaztelugatxe, (23) Zierbena, (24) La arena, (25) Langre, (26) Virgen del Mar/San Roman, (27) Comillas, (28) La Franca, (29) Playa de Aguilar, (30) Foz- Cadavedo, (31) Perbes, (32) Coruna,~ (33) Camelle, (34) Ria de Arousa, (35) Tossa de Mar, (36) El Plomo, Cabo de Gata, (37) La Herradura, (38) Punta Plata; Portugal: (39) Montedor, (40) Playa de Santa Cruz, (41) Almograve, (42) Ingrina and (43) Olhos d’agua. Sicily, Italy not shown but is type locality of L. incrustans. Specimen details in Tables S1 and S2.

common epilithic crust in the mid- to low intertidal observed between cells of neighboring filaments zone on emergent bedrock, in rock pools, and in (Fig. 7A); epithallus two to three cell layers thick tidal canals, as an epizoic crust on mussels and each cell with thick upper calcified walls, some- other bivalves, and also occasionally as an intertidal times broken out flat to triangular, lumen 3–4 lm or subtidal (up to 15 m) rhodolith, if the latter high and 3–4 lm wide (Fig. 7A); meristematic cells sometimes associated with maerl beds (Fig. 6, A–D); intercalary between epithallus and perithallus, rect- in the Mediterranean only found subtidally as both angular to oblong, 12–25 lm long and 6–8 lm rhodoliths, usually in maerl beds, and as crusts wide (Fig. 7A); perithallus formed by ovoid cells, epiphytic on other rhodoliths, for example, on 7–20 lm high and 4–7 lm broad; trichocytes not L. incrustans. seen. Both bisporangial and tetrasporangial unipo- Anatomy: Thallus monomerous with plumose rate conceptacles were observed (Fig. 7, B–F), as multi-layered hypothallus; in transverse view pri- were buried conceptacles; in surface view, concepta- mary pit connections observed between cells of the cle roof flush with thallus surface or/and sunken same filament and secondary pit connections (Fig. 7, B and C), conceptacle pore of 45–55 lm TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 799

FIG. 4. Range of habits of Lithophyllum incrustans rhodoliths. Scales B-I, 1 cm. (A) Lectotype specimen of L. incrustans L 943.10.34. (B) Nucleated rhodolith with smooth surface and some swellings (E276, Ireland). (C) Nucleated rhodolith with smooth surface, swellings and dark dots (bisporangial conceptacles, E309, Ireland). (D) Rhodolith without nucleus and with lamellate branches (E69, Ireland). (E) Nucle- ated rhodolith with smooth surface, swellings and sharp edges (E132, Mediterranean France). (F) Rhodolith without nucleus, expanded branches and with dentate protuberances (E25, Ireland). (G) Flat rhodolith without nucleus, lamellae somewhat dentate (E231, Ireland). with veil covering entrance (Fig. 7D); in transverse of the top right specimen identified as belonging to view, sporangial conceptacle chambers dumbbell- Lithothamnion. shaped due to peripheral dissolution of previously According to Art. 9.14 of the International Code formed perithallial carbonate, 150–190 lm high of Nomenclature (McNeill et al. 2012), “When a and 264–380 lm diameter (Fig. 7B); number of type contains parts belonging to more than one cells from chamber floor to thallus surface ranged taxon, the name must remain attached to the part from 11 to 31 cells; calcified columella formed that corresponds most nearly with the original from pre-existing perithallial filaments (without dis- description or diagnosis.” Foslie’s (1906) description ruption at conceptacle base; Fig. 7, B, C, E); pore could refer to either taxon, as it included informa- canal cylindrical throughout length, 8–23 cells tion only about branch diameter and the size and long, 55–70 lm diameter medially (Fig. 7, B and shape of hypothallial and perithallial cells, none of C); rim present at both openings of pore canal; which was diagnostic to separate the genera Litho- bisporangia 70–87 lm high and 30–40 lm wide phyllum and Lithothamnion in a modern context (Fig. 7E), tetrasporangia 90–150 lm high and 40– (Woelkerling 1988). However, since rbcL-296 65 lm wide (Fig. 7F). Old tetra/bisporangial con- sequences clearly placed one specimen in Lithophyl- ceptacles deeply buried. Gametangial conceptacles lum, and the other in Lithothamnion, we designated not observed. the top left specimen in Figure 6A as the lectotype Distribution: L. hibernicum was confirmed from Ire- of L. hibernicum, thus maintaining the name in the land, England, France (Atlantic and Mediterra- genus to which it is currently assigned. nean), Spain (Atlantic and Mediterranean), and The material referred to herein as “L. depressum Portugal (Fig. 3). historical collection” (CO02513 in CO) was part of Comment: Three specimens illustrated by Printz the original collection of L. depressum, but was (1929) (Pl. LXIII, figs. 11–13) were considered to excluded from the designated lectotype by Cham- constitute the holotype of L. hibernicum (Woelker- berlain and Irvine (1994) who specified bisporangial ling 1993, p. 118, Chamberlain and Irvine 1994, p. material in agreement with the original protologue 74). The top two specimens illustrated in Figure 6A (Crouan and Crouan 1867: 151, Pl. 20, fig. 133). were sequenced. The rbcL 296 sequence from the Chamberlain and Irvine (1994: 79) did not com- top left specimen corresponded to Lithophyllum as ment on this tetrasporangial material that also defined herein by the generitype (L. incrustans), belonged to Lithophyllum, but the rbcL 296 sequence whereas the top right specimen belonged to a clade unambiguously demonstrated that it is L. hibernicum. of NE Atlantic Lithothamnion spp. (data not shown). Both L. hibernicum and L. bathyporum presented sun- The bottom specimen in Figure 6A appeared to be ken conceptacles, and thus this feature was not diag- conspecific with the top left specimen since branch nostic for either of these two species (see more tips are cup-shaped in contrast with the round tips under L. bathyporum description). 800 JAZMIN J. HERNANDEZ-KANTUN ET AL.

FIG. 5. Anatomical features of Lithophyllum incrustans. (A) Surface view of epithallial cells (E308). (B) Transverse view of three layers of epithallial cells with primary pit connections visible between cells in same ﬁlament and subtended by meristematic cell (arrow); secondary pit connections among adjacent ﬁlaments appear as holes (E35). (C) Transverse view of sporangial conceptacle showing pronounced columella (asterisk) (E308). (D) Sporangial pore canal extending over length of seven cells; base wider than apex; note rim-like section at base (arrow, E308). (E) Surface view of sporangial pore canal with veil covering pore (E300). (F) Surface view (different angle) of sporangial pore canal slightly sunken compared to epithallial cells (E308). (G) Surface view of conceptacles (E309). (H) Transverse section of conceptacle with bisporangia; one bisporangium numbered (E309). (I) Deeply buried old conceptacle (E276). TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 801

FIG. 6. Morphological varia- bility in Lithophyllum hibernicum; scale B–D = 1 cm. (A) Type material of L. hibernicum TRH A23-1399, the top left specimen (sequenced) is the lectotype (designated herein); top right specimen (sequenced) an unidentiﬁed Lithothamnion sp. and bottom specimen not sequenced but morphologically in agreement with lectotype specimen. (B) Smooth crust with some protuberances and sharp edges (E211, Ireland). (C) Smooth crust (E322, England). (D) Nucleated rhodolith with smooth surface and some protuberances (E324, Ireland).

L. bathyporum Athanasiadis & D.L. Ballantine DISCUSSION (herein Fig. 8, A–D) Lectotype: CO2511 in PC, L. depressum as lectotyp- Basing coralline genera on morpho-anatomical charac- ified by Chamberlain and Irvine (1994: 79). Fig- ters. There has been much confusion and disagree- ure 8A, top, small specimen. ment in the literature concerning the application of Type locality: Rade de Brest, France. morpho-anatomical characters to define genera of DNA sequences: Partial rbcL and psbA sequences both geniculate (Gabrielson et al. 2011, Hind and from the lectotype specimen were diagnostic for this Saunders 2013, Hind et al. 2014a) and non-genicu- species. See Tables S1 and S2 for additional psbA late corallines (for example, see discussions under and COI sequences. Mesophyllum Me. Lemoine and Synarthrophyton R.A. Morphology and Habitat: Non-geniculate encrusting Townsend in AlgaeBase, accessed 26 March 2015), thallus with smooth surface (Fig. 8A top specimen including genera in the Lithophylloideae (Bailey and label in Fig. 8B). Lectotype specimen presented 1999). DNA-based phylogenies have proved useful bisporangial conceptacles (Fig. 8C) with sunken in beginning to resolve these genera, but many have roof. Only known from three collections, including either not included some generitype species (Bailey the lectotype; it occured as intertidal and subtidal and Chapman 1998, Broom et al. 2008) or have epilithic crusts. assumed that they have sequenced generitype spe- Anatomy: Only bisporangia were observed and cies based on morpho-anatomical similarity (Le Gall measured (97 lm height by 58–63 lm diameter, et al. 2009), even when sequenced specimens were Fig. 8D). collected over 1,000 km from the type locality of Distribution: Known only from Atlantic France and the generitype species and/or from different ocean Spain (Tables S1 and S2) basins (Bailey and Chapman 1998, Le Gall et al. Comment: According to partial rbcL and psbA 2009). Later studies (e.g., Bittner et al. 2011, Kato sequences, the bisporangial lectotype specimen of et al. 2011) have used these earlier-generated Gen- L. depressum selected by Chamberlain and Irvine Bank sequences without verification of the respec- (1994) represented L. bathyporum (See Chamberlain tive taxa, further compounding the problem. and Irvine 1994: 79 and above under L. hibernicum The subfamily Lithophylloideae. By clarifying the comment.) identity of the generitype species of Lithophyllum, 802 JAZMIN J. HERNANDEZ-KANTUN ET AL.

FIG. 7. Anatomical features of Lithophyllum hibernicum. (A) Transverse section with two layers of epithallial cells (numbers); meristematic cell linked by primary pit connection (arrow) to perithallial cell in same filament; holes on lateral walls are secondary pit connections between adjacent filaments (E259). (B) Transverse section of sporangial conceptacle with pronounced columella, pore canal extending for length of about 20 cells (white arrow); proximal portion of canal with homogeneous, cylindrical shape and rim at end (black arrow) (E216). (C) Surface view of sporangial pores with one conceptacle flush with surface (black arrowhead); another pore slightly sunken (white arrowhead; E325). (D) Surface view of veil-covered tetrasporangial pore canal (E259). (E) Conceptacle with bisporangia (spores marked by numbers) (E211). (F) Conceptacle with tetrasporangium (spores marked by numbers) (E259). the type genus of subfamily Lithophylloideae, first which had been designated as the lectotype species by DNA sequences and then with morpho-anatomi- by Foslie (1898) and from which we sequenced cal features, we provide a starting point to resolve about 20% of the rbcL gene. The type specimen of the boundaries and evolutionary relationships of the generitype species of Lithophyllum is thus genera in the subfamily. Our current concept of unequivocally linked to field collected specimens Lithophyllum is based on Woelkerling (1983a), who from which a variety of markers were sequenced. found Philippi’s original collection of L. incrustans, This methodology needs to be applied to all other TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 803

FIG. 8. Lithophyllum bathyporum. (A) The three specimens in type collection; top small specimen is lectotype of Lithothamnion bathyporum (CO2511); two remaining multiporate specimens are Phym- atolithon spp. (left CO02720 and right CO02721). (B) Label of type collection. (C) Surface view of sporangial conceptacles in lectotype. (D) Detail of conceptacle with two bisporangia.

generitype species of genera in Lithophylloideae, The genus Lithophyllum. Recently, Basso et al. whether the genus is currently recognized (e.g., Am- (2014) modiﬁed the diagnosis of Lithophyllum based phiroa J.V.Lamouroux, Lithothrix J.E.Gray) or is con- on the presence of abundant trichocytes in the type sidered a synonym (e.g., Titanoderma N€ageli and material of L. kotschyanum Unger and recently col- Pseudolithophyllum Me.Lemoine), in order to resolve lected samples of Lithophyllum spp. from the NW the phylogenetic relationships among the genera. Indian Ocean. The emendation removed the adjec- An example of the problems involved in cor- tive “rare,” leaving trichocytes as present or absent. rectly identifying specimens of generitype species Based on the specimens examined in this study, nei- and selecting a marker that segregates related spe- ther L. incrustans, the generitype, nor L. hibernicum cies is provided by Bailey (1999). He sequenced a presented trichocytes. More taxa need to be specimen that was collected and identiﬁed by Y. M. sequenced and analysed anatomically to determine Chamberlain as L. incrustans, but this specimen at what rank the presence and absence of tricho- likely was crustose and intertidal, a habit uncom- cytes as well as their features are diagnostic in taxa mon for the species and a habitat in which it does of Lithophylloideae. not occur (unfortunately, a description of the vou- Lithophyllum incrustans, L. hibernicum, and cher was not provided, nor was a voucher cited L. bathyporum. Based on our results and previous that can be reassessed). Nevertheless, the SSU studies for geniculate and non-geniculate species, sequence (GenBank AF093409), which likely was DNA sequencing is unfortunately revealing just how from a specimen of L. hibernicum, a common inter- incorrect our knowledge has been about non-genic- tidal crust, is identical to an SSU sequence from ulate corallines at the species rank, knowledge previ- L. incrustans (data not shown). As Adey et al. ously based on morpho-anatomical characters. (2015) demonstrated for Clathromorphum Foslie, Nowhere is this more apparent than in the present SSU sequences do not allow discrimination of clo- study where DNA sequencing has shown that one sely related species, and the same is true for L. in- name, L. incrustans, has been misapplied for the crustans and L. hibernicum, providing further past 60 years, while another name for a very com- evidence that SSU is not a marker of choice for mon species, L. hibernicum, was lost for nearly separating species within the Lithophylloideae. 90 years, and a recently proposed name, L. bathypo- Instead, based on our results, psbA, rbcL and COI rum, which was based on characters that are not gene sequences provided similar topologies, but diagnostic at the species rank, nevertheless applies enough variation to reliably distinguish Lithophyllum to an uncommon and geographically restricted spe- spp. from Europe. cies. 804 JAZMIN J. HERNANDEZ-KANTUN ET AL.

Our DNA sequencing results clearly demonstrate species were present under the name L. incrustans. that nearly all historical records of L. incrustans They restricted the use of L. incrustans to specimens from around Europe based on morpho-anatomy are from the Mediterranean Sea with elevated concepta- probably incorrect. This species is not the common, cles and a single layered epithallium, and they pro- epilithic, intertidal crust that is described and illus- posed a new name, L. bathyporum, for specimens trated in Chamberlain and Irvine (1994) and in from the British Isles and France with sunken con- Cabioch and Boudouresque (1992), or online in ceptacles and an epithallium of up to four layers. Bressan et al. (2015) (http://www2.units.it/biologia/ L. bathyporum is a substitute name based on Corallinales/home.html, accessed 27 February L. depressum due to the prior existence of Lithophyl- 2015). Instead, it is a mostly subtidal to occasionally lum depressum Villas-Boas, Figueiredo & Riosmena- low intertidal species usually found as a rhodolith, Rodriguez (2009). Based on sequenced material, but may also be epilithic, epizoic or epiphytic neither character that Ballantine et al. (2011) pro- (Table S2). Based on our results, L. incrustans is posed, i.e. conceptacle position (elevated or sun- characterized by the following suite of features: (i) a ken) nor number of layers of epithallial cells, is mostly subtidal habitat, usually occurring as rhodo- diagnostic for European Lithophyllum species, nor is liths but also as epilithic, epiphytic and epizoic L. incrustans restricted to the Mediterranean Sea. crusts, (ii) epithallial cells 6–8 lm in diameter, (iii) Based on DNA sequencing of our numerous and the pore canal of each conceptacle chamber being widespread Northwest Atlantic collections, L. bathyp- triangular/conical or tapering in lateral view and orum is a very uncommon (known only from three (iv) bisporangia 100–115 lm high and 40–45 lm specimens) encrusting, epilithic, bisporangiate spe- broad. Bressan et al. (online) show a L. incrustans cies of the intertidal and subtidal with a distribution conceptacle cross-section with a cylindrical pore restricted to Brittany (France) and Atlantic Spain. canal, a feature characteristic of L. hibernicum, not We also question Ballantine et al.’s (2011) report of L. incrustans, but whether this feature is diagnostic L. incrustans in the Caribbean Sea (Puerto Rico) for L. hibernicum compared to all other European based only on morpho-anatomy; this needs to be Lithophyllum species is unknown. Babbini and Bres- confirmed by DNA sequencing. san (1997) reported records of L. incrustans as far Identification of Atlantic European Lithophyllum rho- east as the western coast of Turkey, but we have not doliths. Three recent papers (Hernandez-Kantun confirmed this species east of Sicily, its type locality. et al. 2014, 2015, Pardo et al. 2014) illustrate the The name L. hibernicum was essentially ignored for impossibility of correctly identifying the vast major- 88 years after its original description until Chamber- ity of rhodolith-forming species using morpho-anat- lain and Irvine (1994) briefly characterized it based omy, regardless of locality. Pardo et al. (2014) on three collections, including the holotype. Herein, sequenced rhodoliths in the Northeast Atlantic from we demonstrated that the holotype comprises two the Svalbard Archipelago (78° N) to the Canary heterotypic elements belonging to different families, Islands (29° S) and recovered 13 species, only six of and we selected one individual as lectotype that which could be assigned names based on morpho- maintains the species in Lithophyllum, the genus to anatomical comparisons with type specimens or spe- which it currently is assigned. This species is one of cies descriptions. After comparing sequences the most common intertidal corallines in temperate reported herein with sequences in Pardo et al. Atlantic Europe from Ireland south to Spain and (2014), we consider specimens named L. fascicula- Portugal. L. hibernicum is characterized by: (i) mid- tum and L. dentatum in their paper to belong to to low intertidal epilithic and epizoic crusts abun- L. incrustans. Hernandez-Kantun et al. (2015) were dant in rock pools, sometimes, as intertidal rhodo- likewise unable to assign names to any of the NE liths, occasionally present as subtidal rhodoliths or Atlantic Lithophyllum rhodoliths included in their epiphytic on L. incrustans rhodoliths, (ii) epithallial DNA analyses, but now those specimens can be cells 3–4 lm in diameter, (iii) pore canal of each identified as follows: Lithophyllum sp. 1 is L. incru- conceptacle chamber with a cylindrical shape and stans and Lithophyllum sp. 3 is L. hibernicum; for (iv) bisporangia 70–87 lm high and 30–40 lm diam- Hernandez-Kantun et al. (2014) as follows: Lithophyl- eter, tetrasporangia 90–150 lm high and 50–65 lm lum sp. 4 is L. incrustans and Lithophyllum sp. 5 is diameter. Only one vegetative feature, the epithallial L. hibernicum. Herein, specimens with the morphol- cell diameter, consistently distinguishes L. incrustans ogy of L. fasciculatum from Ireland and Brittany from L. hibernicum, whereas pore canal shape and were shown by DNA sequences to belong to L. in- the diameter of sporangia also are diagnostic when crustans, thus removing L. fasciculatum from both bi/tetrasporangial conceptacles are present. How- floras. Likewise, the report of L. duckerae (type local- ever, whether these morpho-anatomical features will ity: Sicily) from Cornwall (Chamberlain and Irvine remain diagnostic is unclear, pending the study of 1994: 70-72) is instead confirmed as L. hibernicum, additional European Lithophyllum species. removing L. duckerae from the British Isles flora. Recently, based on morpho-anatomy, but without L. hibernicum was previously known only as a subtidal examining any material, Ballantine et al. (2011) rhodolith (Foslie 1906, Chamberlain and Irvine proposed that in European waters two different 1994), but the present study shows it to be a TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 805 common mid- to low intertidal epilithic and epizoic Barcode et gestion durable des collections,” “Biodiversite actu- crust. The distinctive form of the lectotype species, elle et fossile. Crises, stress, restaurations et panchronisme: le message designated herein, i.e., a branched rhodolith with systematique ” and “Emergences”) and network “Bibliotheque du Vivant” funded by CNRS, Museum National d’Histoire Natu- cup-shaped or flattened branch apices (Fig. 7A), is relle, INRA and CEA (Centre National de Sequenc ßage). PWG so far confirmed only for the type locality, despite thanks Dr. T. Vision (University of North Carolina, Chapel numerous other collections of rhodoliths from the Hill) for use of lab space and equipment and Dr. D.W. Fresh- west coast of Ireland and elsewhere in Atlantic Eur- water (DNA Analysis Core Facility, Center for Marine Sciences, ope that were identified by DNA sequence as University of North Carolina, Wilmington) for DNA sequenc- belonging to L. hibernicum. Based on psbA gene ing. sequences, Hernandez-Kantun et al. (2015) suggested that a common, epilithic, intertidal, encrust- Adey, W. H. & Adey, P. J. 1973. Studies on the biosystematics and ing species also occurred as a much branched, ecology of the epilithic crustose Corallinaceae of the British unattached, rhodolith; the additional markers rbcL Isles. Br. Phycol. J. 8:343–407. Adey, W. H., Hernandez-Kantun, J. J., Johnson, G. & Gabrielson, and COI in this study confirmed this entity to be P. W. 2015. DNA sequencing, anatomy, and calcification pat- L. hibernicum. terns support a monophyletic, subarctic, carbonate reef- These observations have profound implications forming Clathromorphum (Hapalidiaceae, Corallinales, Rhodo- for the identification of corallines worldwide based phyta). J. Phycol. 51:189–203. only on habit (i.e., rhodolith versus attached and Babbini, L. & Bressan, G. 1997. Recensement de Corallinacees de la Mer Mediterran ee et considerations phytogeographiques. encrusting), as well as other, intermediate forms Bibliotheca Phycologica 103:1–421. (e.g., crusts that completely cover pebbles or cob- Bailey, J. C. 1999. Phylogenetic positions of Lithophyllum incrustans ble). Our findings demonstrate that the plasticity in and Titanoderma pustulatum (Coralliaceae, Rhodophta) based gross-morphology and habitats for maerl-forming on 18S rRNA gene sequence analyses, with a revised classifi- cation of the Lithophylloideae. Phycologia 38:208–16. species is varied and complex, even within single Bailey, J. C. & Chapman, R. L. 1998. A phylogenetic study of the species as illustrated by L. incrustans and L. hiberni- Corallinales (Rhodophyta) based on nuclear small-subunit cum. Relying only on morpho-anatomy greatly rRNA gene sequences. J. Phycol. 34:692–705. increases potential misidentifications of species, Bailey, J. C., Gabel, J. E. & Freshwater, D. W. 2004. Nuclear 18S which highlights the necessity of DNA sequences to rRNA gene sequence analyses indicate that the Mastophoroi- deae (Corallinaceae, Rhodophyta) is a polyphyletic taxon. understand non-geniculate coralline biodiversity Phycologia 43:3–12. and biogeography. Ballantine, D. L., Athanasiadis, A. & Ruiz, H. 2011. Notes on the benthic marine algae of Puerto Rico. X. Additions to the We are very grateful to the following curators for the loan of flora. Bot. Mar. 54:293–302. type material essential to this research: J. Wilbraham (BM), Basso, D., Caragnano, A. & Rodondi, G. 2014. Trichocytes in Litho- Dr. E. Smets (L), and Dr. K. Hassel (TRH). Jacques Grall, phyllum kotschyanum and Lithophyllum spp. (Corallinales, Rho- Christine Maggs, Jason Hall-Spencer, Line Le Gall, Conxi dophyta) from the NW Indian Ocean. J. Phycol. 50:711–7. Rodriguez-Prieto, Ignacio Barbara, Pilar Dıaz, Antonio Secilla, Bittner, L. 2009. Phylogenie des Corallinales (Rhodophyta) et analyse de leur diversiteg en etique dans le Pacifique Sud. Meadhbh Moriarty, Ricardo Bermejo Lacida, Alex Wan, Be- Museum National d’Histoire Naturelle, Paris, 318 pp. noit Queguineur, Julia Nunn, and Marie Pazoutova are grate- Bittner, L., Payri, C., Maneveldt, G., Couloux, A., Cruaud, C., de fully acknowledged for the collection of samples and/or Reviers, B. & Le Gall, L. 2011. Evolutionary history of the assistance in the field. We thank P. Lalor for assistance with Corallinales (Corallinophycidae, Rhodophyta) inferred from SEM images, Prof. J. Brodie for revising early versions of the nuclear, plastidial and mitochondrial genomes. Mol. Phyloge- manuscript, and the reviewers for helpful comments and sug- net. Evol. 61:697–713. gestions; JHK was supported by a PhD scholarship from Bressan, G. & Babbini, L. 2003. Biodiversita marina delle coste CONACyT-211950 (National Council for Science and Tech- Italiane: Corallinales del Mar Mediterraneo: guida alla dete- nology, Mexico) and SEP (Secretariat for Public Education, minazione. Biol. Mar. Med. 10:1–237. Mexico). Financial support was received from the Marine Bressan, G., Babbini, L. & Poropat, F. 2015. Available at: http:// Institute (Ireland) as part of the National Marine Biodiscovery www2.units.it/biologia/Corallinales/home.html (accessed 27 Program (Beaufort Award for Marine Biodiscovery to the February, 2015). National University of Ireland, Galway), from the European Broom, J. E. S., Hart, D. R., Farr, T. J., Nelson, W. A., Neill, K. F., Community through the FP6-funded project HIPPOCRATES Harvey, A. S. & Woelkerling, W. J. 2008. Utility of psbA and (NMP3-CT-2003-505758), from the National University of Ire- nSSU for phylogenetic reconstruction in the Corallinales land, Galway through the Thomas Crawford Award, and from based on New Zealand taxa. Mol. Phylogenet. Evol. 46:958–73. the M.I.U.R. (Italian Ministry for Education, Universities and Cabioch, J. & Boudouresque, C.-F. 1992. Guide des algues des Research) through a P.R.I.N. 2011 project (Coastal biocon- mers d’Europe. Lausanne. Chamberlain, Y. M. 1996. Lithophylloid Corallinaceae (Rhodo- structions: structure, function and management). VP acknowl- phyta) of the genera Lithophyllum and Titanoderma from edges support by the postdoctoral program Axudas de apoio a southern Africa. Phycologia 35:2014–21. etapa inicial de formacion posdoutoral do Plan I2C (Xunta de Chamberlain, Y. M. & Irvine, L. M. 1994. Lithophylloideae Setc- Galicia), and projects funded by British Phycological Society hell. In Irvine, L. M. & Chamberlain, Y. M. [Eds.] Seaweeds of (Small Grant Scheme-Project Award 2011-2012) and Spain’s the British Isles. Volume 1 Rhodophyta Part 2B Corallinales, Ministerio de Economıa y Competitividad (CGL2009-09495/ Hildenbrandiales. HMSO, London, pp. 58–112. BOS, partially founded by ERDF); molecular data provided by Crouan, P. L. & Crouan, H. M. 1867. Florule du Finistere . Imp. J. VP were produced at the Service de Systematique Moleculaire B. & Lefournier Aine, Brest, France, 262 pp. of the Museum National d’Histoire Naturelle (CNRS - UMS Edyvean, R. G. J. & Ford, H. 1986. Spore production by Lithophyl- 2700) with funds provided by Action Transversale du Museum lum incrustans (Corallinales, Rhodophyta) in the British Isles. National d’Histoire Naturelle (“Taxonomie moleculaire: DNA Br. Phycol. J. 21:255–61. 806 JAZMIN J. HERNANDEZ-KANTUN ET AL.

Ford, H., Hardy, F. G. & Edyvean, R. G. J. 1983. Population biol- Le Gall, L., Payri, C. E., Bittner, C. E. & Saunders, G. W. 2009. ogy of the crustose red alga Lithophyllum incrustans Phil. Multigene polygenetic analyses support recognition of the Three populations on the east coast of Britain. Biol. J. Lin- Sporolithales, ord. nov. Mol. Phylogenet. Evol. 54:302–5. nean Soc. 19:211–20. Le Gall, L. & Saunders, G. W. 2010. DNA barcoding is a powerful Foslie, M. 1895. The Norwegian forms of Lithothamnion. K. norske tool to uncover algal diversity: a case study of the Phyllophor- Vidensk. Selsk. Skr. 1894:29–208, 23 pls. aceae (Gigartinales, Rhodophyta) in the Canadian flora. Foslie, M. 1898. Systematical survey of the Lithothamnia. K. norske J. Phycol. 46:374–89. Vidensk. Selsk. Skr. 1898:1–7. Martone, P. T., Lindstrom, S. C., Miller, K. A. & Gabrielson, P. Foslie, M. 1906. Algologiske Notiser II. Det Kongelige Norske Vide- W. 2012. Chiharaea and Yamadaia (Corallinales, Rhodophyta) nskabers Selskabs Skrifter 1906:1–28. represent reduced and recently derived articulated coralline Freshwater, D. W. & Rueness, J. 1994. Phylogenetic relationships morphologies. J. Phycol. 48:859–68. of some European Gelidium (Gelidiales, Rhodophyta) species Mateo-Cid, L. E., Mendoza-Gonzalez, A. C. & Gabrielson, P. W. based upon rbcL sequence analysis. Phycologia 33:187–94. 2014. Neogoniolithon (Corallinales, Rhodophyta) on the Atlan- Gabrielson, P. W. 2008. Molecular sequencing of Northeast Paci- tic coast of Mexico, including N. siankanensis sp. nov. Phyto- fic type material reveals two earlier names for Prionitis lyallii, taxa 190:64–93. Prionitis sternbergii and Prionitis jubata (Halymeniaceae, Rho- McNeill, J., Barrie, F. R., Buck, W. R., Demoulin, V., Greuter, W., dophyta). Phycologia 47:89–97. Hawksworth, D. L., Herendeen, P. S. et al. 2012. International Gabrielson, P. W., Miller, K. A. & Martone, P. 2011. Morphomet- Code of Nomenclature for algae, fungi and plants (Melbourne ric and molecular analyses confirm two distinct species of Code) adopted by the Eighteenth International Botanical Calliarthron (Corallinales, Rhodophyta), a genus endemic to Congress Melbourne, Australia, July 2011. Reg. Veg. 154:1–208. the northeast Pacific. Phycologia 50:298–316. Pardo, C., Lopez, L., Pena,~ V., Hernandez-Kantun, J., Le Gall, L., Guiry, M. D. & Guiry, G. M. 2014. AlgaeBase. World-wide elec- Barbara, I. & Barreiro, R. 2014. A multilocus species delimi- tronic publication, National University of Ireland, Galway, tation reveals a striking number of species of coralline algae Ireland. Available at: http://www.algaebase.org (searched on forming maerl in the OSPAR maritime area. PLoS ONE February 2014). 9:3104073. Hamel, G. & Lemoine, P. 1952. Corallinacees de France et d’Afrique Pardo, C., Pena,~ V., Barreiro, R. & Barbara, I. 2015. A molecular du Nord. Archs Mus. nat. Hist. nat. Paris VII Ser. 7 1:15–136. and morphological study of Corallina sensu lato (Corallinales, Harvey, A. S., Broadwater, S. T., Woelkerling, W. J. & Mitrovski, Rhodophyta) in the Atlantic Iberian Peninsula. Cryptogamie P. J. 2003. Choreonema (Corallinales, Rhodophyta): 18S rDNA Algol. 36:31–54. phylogeny and resurrection of the Hapalidiaceae for the sub- Pena,~ V., De Clerck, O., Afonso-Carrillo, J., Ballesteros, E., Bar- families Choreonematoideae, Australithoideae, and Melobe- bara, I., Barreiro, R. & Le Gall, L. 2015. An integrative sys- sioideae. J. Phycol. 39:988–98. tematic approach to species diversity and distribution in the Hernandez-Kantun, J. J., Riosmena-Rodriguez, R., Adey, W. H. & genus Mesophyllum (Corallinales, Rhodophyta) in Atlantic Rindi, F. 2014. Analysis of the cox 2-3 spacer region for pop- and Mediterranean Europe. Eur. J. Phycol. 50:20–36. ulation diversity and taxonomic implications in rhodolith- Pena,~ V., Hernandez-Kantun, J. J., Grall, J., Pardo, C., Lopez, L., forming species (Rhodophyta: Corallinales). Phytotaxa Barbara, I., Le Gall, L. & Barreiro, R. 2014. Detection of 190:331–54. gametophytes in the maerl-forming species Phymatolithon cal- Hernandez-Kantun, J. J., Riosmena-Rodriguez, R., Hall-Spencer, careum (Melobesioideae, Corallinales) assessed by DNA bar- J., Pena,~ V., Maggs, C. & Rindi, F. 2015. Phylogenetic analysis coding. Cryptogamie Algol. 35:15–25. of rhodolith formation in the Corallinales (Rhodophyta). Philippi, R. A. 1837. Beweis, dass die Nulliporen Pflanzen sind. Eur. J. Phycol. 50:46–61. Archiv fur€ Naturgeschichte 3:387–93. Heydrich, F. 1901. Bietet die Foslie’sche Melobesien-Systematik Printz, H. 1929. Foslie: Contributions to a monograph of the Lithotham- eine sichere Begrenzung? Berichte der deutsche botanischen nia. Kong. Nor. Vidensk. Selsk. Museet, Trondheim, 60 pp. Gesellschaft 19:180–94. Rambaut, A. & Drummond, A. J. 2007. Tracer v1.4. Available at: Heydrich, F. 1911. Lithophyllum incrustans Phil. mit einem Nach- http://beast.bio.ed.ac.uk/Tracer. trag uber€ Paraspora fruticulosa (Ktz.) Heydrich. Bibliotheca Richards, J. L., Gabrielson, P. W. & Fredericq, S. 2014. New Botanica 18:1–24. insights into the genus Lithophyllum (Lithophylloideae; Coral- Hind, K. R., Gabrielson, P. W., Lindstrom, S. C. & Martone, P. T. linaceae, Corallinales) from offshore the NW Gulf of Mexico. 2014a. Misleading morphologies and the importance of Phytotaxa 190:162–75. sequencing type specimens for resolving coralline taxonomy Saunders, G. W. 2005. Applying DNA barcoding to red macroal- (Corallinales, Rhodophyta): Pachyarthron cretaceum is Corallina gae: a preliminary appraisal holds promise for future applica- officinalis. J. Phycol. 50:760–4. tions. Phil. Trans. R. Soc. B. 360:1879–88. Hind, K. R., Gabrielson, P. W. & Saunders, G. W. 2014b. Molecu- Saunders, G. W. 2008. A DNA barcode examination of the red lar-assisted alpha taxonomy reveals pseudocryptic diversity algal family Dumontiaceae in Canadian waters reveals sub- among species of Bossiella (Corallinales, Rhodophyta) in the stantial cryptic species diversity. 1. The foliose Dilsea-Neodilsea eastern Pacific Ocean. Phycologia 53:443–56. complex and Weeksia. Botany 86:773–89. Hind, K. R. & Saunders, G. W. 2013. Molecular markers from Saunders, G. W. & McDevit, D. C. 2012. Acquiring DNA sequence three organellar genomes unravel complex taxonomic rela- data from dried archival red algae (Florideophyceae) for the tionships within the coralline algal genus Chiharaea (Coralli- purpose of applying available names to contemporary nales, Rhodophyta). Mol. Phylogenet. Evol. 67:529–40. genetic species: a critical assessment. Botany 90:191–203. Huelsenbeck, J. P. & Ronquist, F. 2001. MRBAYES: Bayesian infer- Silvestro, D. & Michalak, I. 2012. raxmlGUI: a graphical front-end ence of phylogeny. Bioinformatics 17:754–5. for RAxML. Org. Divers. Evol. 12:335–7. doi:10.1007/s13127-011- Hughey, J. R., Silva, P. C. & Hommersand, M. H. 2001. Solving 0056-0. Available at: https://sites.google.com/site/raxmlgui/ taxonomic and nomenclatural problems in Pacific Gigartina- Sissini, M. N., de Oliveira, M. C., Gabrielson, P. W., Robinson, N. ceae (Rhodophytat) using DNA from type material. J. Phycol. M., Okolodkov, Y. B., Riosmena-Rodriguez, R. & Horta, P. A. 37:1091–109. 2014. Mesophyllum erubescens (Corallinales, Rhodophyta)–so Kato, A., Baba, M. & Suda, S. 2011. Revision of the Mastophoroi- many species in one epithet. Phytotaxa 190:299–319. deae (Corallinales, Rhodophyta) and polyphyly in nongeni- Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & culate species widely distributed on Pacific coral reefs. Kumar, S. 2011. MEGA5: molecular evolutionary genetics J. Phycol. 47:662–72. analysis using maximum likelihood, evolutionary distance, Kato, A., Baba, M. & Suda, S. 2013. Taxonomic circumscription and maximum parsimony methods. Mol. Biol. Evol. 28:2731–9. of heterogeneous species Neogoniolithon brassica-florida (Coral- Thiers, B. 2015. (continuously updated) Index Herbariorum: A linales, Rhodophyta) in Japan. Phycol. Res. 61:15–26. global directory of public herbaria and associated staff. New TAXONOMIC IDENTITY OF LITHOPHYLLUM INCRUSTANS 807

York Botanical Garden’s Virtual Herbarium. Available at: http://sweetgum.nybg.org/ih/ L. hibernicum, L. bathyporum, and of “Lithothamnion Verbruggen, H. 2014. Morphological complexity, plasticity, and depressum historical collection” CO02513. Codon species diagnosability in the application of old species names position marked by number above each position in DNA-Based taxonomies. J. Phycol. 50:26–31. Verbruggen, H., Maggs, C., Saunders, G., Le Gall, L., Yoon, H. & (e.g., 1, 2, or 3); asterisks demark differences De Clerck, O. 2010. Data mining approach identifies among sequences. research priorities and data requirements for resolving the red algal tree of life. BMC Evol. Biol. 10:16. Figure S2. UPGMA (Unweighted Pair Group Villas-Boas, A. B., Riosmena-Rodriguez, R., Amado-Filho, G. M., Method with Arithmetic Mean) tree of COI Litho- Maneveldt, G. W. & de O. Figueiredo, M. A. 2009. Rhodo- phyllum gene sequences that have complementary lith-forming species of Lithophyllum (Corallinales; Rhodo- rbcL sequences in Figure 1. phyta) from Espırito Santo State, Brazil, including the description of L. depressum sp. nov. Phycologia 48:237–48. Figure S3. Maximum likelihood (ML) phyloge- Woelkerling, W. J. 1983a. A taxonomic reassessment of Lithophyllum netic reconstruction of Lithophyllum spp. based on Philippi (Corallinaceae, Rhodophyta) based on studies of R. A. Philippi’s original collections. Br. Phycol. J. 18:299–328. psbA gene sequences. Woelkerling, W. J. 1983b. A taxonomic reassessment of Lithotham- Figure S4. Phylogenetic reconstructions of nion Philippi (Corallinaceae, Rhodophyta) based on studies of R. A. Philippi’s original collections. Br. Phycol. J. 18:165–97. Lithophyllum spp. comparing psbA and rbcL genes. Woelkerling, W. J. 1988. The Coralline Red Algae: An Analysis of the Table S1. List of specimens sequenced (except Genera and Subfamilies of Nongeniculate Corallinaceae. British Museum (Natural History) & Oxford University Press, Lon- for those with label CPVP, provided in Pardo don & Oxford, 268 pp. et al. 2014), including assigned name by morpho- Woelkerling, W. J. & Campbell, S. J. 1992. An account of the logical means or DNA from type material, Loca- southern Australian species of Lithophyllum (Corallinaceae, tion & collection data (* indicates rhodolith, if Rhodophyta). Bulletin of the British Museum (Natural History) * Botany 22:1–107. no is crustose), Herbarium accession #, Speci- Woelkerling, W. J. 1993. Type collections of Corallinales (Rhodo- men Code (if given) and GenBank Accession phyta) in the Foslie Herbarium (TRH). Gunneria 67:1–289. number for each gene sequenced. Taxa are listed Woelkerling, W. J. & Lamy, D. 1998. Non-geniculate Coralline Red in alphabetical order. Collection details for speci- Algae and the Paris Museum: Systematics and Scientific History. mens with only COI sequence are available in Publications Scientifiques du Museum/ADAC, Paris, 767 pp. Yoon, H. S., Hackett, J. D. & Bhattacharya, D. 2002. A single ori- Table S2. gin of the peridinin and fucoxanthin containing plastids in Table S2. List of COI sequenced specimens, dinoflagellates through tertiary endosymbiosis. Proc. Natl. Acad. Sci. USA 99:11724–9. their GenBank numbers, herbarium accession numbers and collection information. Table S3. Morpho-anatomical features analysed Supporting Information for Lithophyllum incrustans and L. hibernicum, and their references. Additional Supporting Information may be found in the online version of this article at the publisher’s web site: Figure S1. Alignment of rbcL-296 sequences from type materials of Lithophyllum incrustans,