Molecular Phylogenetics and Evolution 41 (2006) 513–527 www.elsevier.com/locate/ympev A molecular phylogenetic analysis of the Octocorallia (Cnidaria: Anthozoa) based on mitochondrial protein-coding sequences Catherine S. McFadden a,¤, Scott C. France b, Juan A. Sánchez c, Phil Alderslade d a Department of Biology, Harvey Mudd College, 1250 N. Dartmouth Ave., Claremont, CA 91711, USA b Department of Biology, University of Louisiana at Lafayette, Lafayette, LA 70504, USA c Departamento de Ciencias Biológicas, Universidad de los Andes, Bogotá, Colombia d Museum and Art Gallery of the Northern Territory, Darwin, NT 0801, Australia Received 5 January 2006; revised 9 June 2006; accepted 16 June 2006 Available online 18 June 2006 Abstract Despite their abundance and ecological importance in a wide variety of shallow and deep water marine communities, octocorals (soft corals, sea fans, and sea pens) are a group whose taxonomy and phylogenetic relationships remain poorly known and little studied. The group is currently divided into three orders (O: Alcyonacea, Pennatulacea, and Helioporacea); the large O. Alcyonacea (soft corals and sea fans) is further subdivided into six sub-ordinal groups on the basis of skeletal composition and colony growth form. We used 1429 bp of two mitochondrial protein-coding genes, ND2 and msh1, to construct a phylogeny for 103 octocoral genera representing 28 families. In agreement with a previous 18S rDNA phylogeny, our results support a division of Octocorallia into two major clades plus a third, minor clade. We found one large clade (Holaxonia–Alcyoniina) comprising the sea fan sub-order Holaxonia and the majority of soft corals, and a second clade (Calcaxonia–Pennatulacea) comprising sea pens (O. Pennatulacea) and the sea fan sub-order Calcaxonia. Taxa belonging to the sea fan group Scleraxonia and the soft coral family Alcyoniidae were divided among the Holaxonia–Alcyoniina clade and a third, small clade (Anthomastus–Corallium) whose relationship to the two major clades was unresolved. In contrast to the previous studies, we found sea pens to be monophyletic but nested within Calcaxonia; our analyses support the sea fan family Ellisellidae as the sister taxon to the sea pens. We are unable to reject the hypothesis that the calcaxonian and holaxonian skeletal axes each arose once and suggest that the skeletal axis of sea pens is derived from that of Calcaxonia. Topology tests rejected the monophyly of sub-ordinal groups Alcyoniina, Scleraxonia, and Stolonifera, as well as 9 of 14 families for which we sampled multiple genera. The much broader taxon sampling and better phylogenetic resolution aVorded by our study relative to the previous eVorts greatly clarify the relationships among families and subordinal groups within each of the major clades. The failure of these mitochondrial genes as well as previous 18S rDNA studies to resolve many of the deeper nodes within the tree (including its root) suggest that octocorals underwent a rapid radiation and that large amounts of sequence data will be required in order to resolve the basal relationships within the clade.  2006 Elsevier Inc. All rights reserved. Keywords: Alcyonacea; Gorgonian; Molecular systematics; msh1; ND2; Pennatulacea; Sea pen; Soft coral 1. Introduction The anthozoan subclass Octocorallia includes over 3000 described species of soft corals, sea fans, and sea pens (Williams and Cairns, 2005). Octocorals are ecologically diverse and important members of a wide variety of * Corresponding author. Fax: +1 909 607 7172. E-mail address: mcfadden@hmc.edu (C.S. McFadden). 1055-7903/$ - see front matter  2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2006.06.010 marine communities, from shallow tropical coral reefs to the deep sea. For example, soft corals are abundant and ecologically dominant organisms on coral reefs throughout the Indo-West PaciWc, often occupying 50% or more of the available primary substrate (Tursch and Tursch, 1982; Dinesen, 1983; Dai, 1988; Riegl et al., 1995; Fabricius, 1997). Gorgonians (sea fans with a rigid scleroproteinaceous axis) dominate many Caribbean coral reefs (Sánchez et al., 1997, 1998), as well as deep sea communities 514 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 (Heifetz, 2002; Watling and Auster, 2005), where their three-dimensional structure provides critical habitat for associated organisms (e.g., Krieger and Wing, 2002; BuhlMortensen and Mortensen, 2004a,b). Sea pens occupy soft sediment habitats over a wide depth range and are often abundant macrofaunal Wlter-feeders in the deep sea (Tyler, 2003). In contrast to other major groups of cnidarians for which there is a long and rich history of phylogenetic study (e.g., Veron et al., 1996; Collins et al., 2006), octocorals remain a poorly known and little studied group (Bayer, 1981a). Attempts to understand their taxonomy and phylogenetic relationships have been hampered by a paucity of useful morphological characters, widespread homoplasy and intraspeciWc variation in characters such as colony growth form and sclerite morphology, and a poor fossil record (Williams, 1997). To date, cladistic analyses based on morphological characters have been attempted only for the sea pens (Williams, 1994, 1997). Throughout most of the 20th century, the sub-class Octocorallia was subdivided into seven orders (e.g., Hyman, 1940), two of which are clearly distinct morphologically: O. Helioporacea (blue corals), the only group of octocorals to form a massive aragonite skeleton and O. Pennatulacea (sea pens), in which a primary axial polyp (oozooid) diVerentiates into a bulbous peduncle, used to anchor the colony in soft substrate, and a distal rachis from which secondary polyps arise. The remaining Wve orders (Alcyonacea, Gorgonacea, Stolonifera, Telestacea, and Protoalcyonaria) were deWned solely on the basis of diVerences in colony growth morphology. Recognizing that growth morphology represents a continuum and that morphologically intermediate taxa linked each of these groups, Bayer (1981b) combined these Wve orders into a single order, Alcyonacea, a revision that has been widely accepted by modern octocoral taxonomists (Fabricius and Alderslade, 2001). The current classiWcation system therefore divides Octocorallia into orders Alcyonacea (28 families of soft corals and sea fans), Pennatulacea (14 families of sea pens), and Helioporacea (two families of blue corals). The large and morphologically diverse O. Alcyonacea is further subdivided into six sub-ordinal groups that are distinguished on the basis of colony architecture and composition of the skeletal axis, if present (Table 1). The sea fan sub-orders Holaxonia and Calcaxonia (GrasshoV, 1999) represent morphologically discrete entities, but the other four groups (Stolonifera, Alcyoniina, Scleraxonia, and Protoalcyonaria) grade into one another morphologically, and consequently have been classiWed loosely as sub-ordinal “groups” rather than sub-orders (Fabricius and Alderslade, 2001). Several recent molecular phylogenetic studies of class Anthozoa using 18S rDNA or 16S mtDNA sequences have supported the monophyly of Octocorallia and have divided the sub-class into either two or three distinct clades (France et al., 1996; Berntson et al., 1999, 2001; Won et al., 2001; Sánchez et al., 2003a). The most taxonomically comprehensive of these studies found support for three major clades of Octocorallia, with one clade representing most of the sea pens, and each of the other two clades comprising a heterogeneous mix of taxa from most or all of the six sub-ordinal groups of Alcyonacea (Berntson et al., 2001). Although these data suggest a lack of phylogenetic support for the current sub-ordinal taxonomic groupings, relationships among the disparate taxa of Alcyonacea included in each major clade could not be inferred due to insuYcient phylogenetic resolution. Here, we construct a phylogeny for sub-class Octocorallia using partial sequences of two mitochondrial protein-coding genes: ND2 and msh1, a mutS homolog that is found in the mitochondrial genome of all octocorals but no other metazoans (Pont-Kingdon et al., 1995, 1998; Culligan et al., 2000; France and Hoover, 2001). Because rates of octocoral mitochondrial gene evolution are very slow compared to other animals, these genes are informative for genus- and familylevel phylogenetic analyses (France and Hoover, 2001, 2002; McFadden et al., 2004), but lack suYcient resolution to discriminate species within many genera (Sánchez et al., 2003b; Wirshing et al., 2005; Cairns and Bayer, 2005). We compare the resulting mitochondrial gene phylogeny to those obtained Table 1 Current higher taxonomic classiWcation of the anthozoan subclass Octocorallia Taxonomic group N DeWning characteristics O. Pennatulacea [sea pens] O. Helioporacea [blue coral] O. Alcyonacea [soft corals—no skeletal axis] Grp. Protoalcyonaria Grp. Stolonifera Grp. Alcyoniina 14 2 Axial polyp diVerentiated into basal peduncle and distal rachis Massive aragonite skeleton [sea fans—with skeletal axis] Grp. Scleraxonia SO. Holaxonia SO. Calcaxonia 2a 5 5 Solitary polyps Polyps united basally by simple stolons which may fuse to form ribbons Polyps united within Xeshy mass of coenenchyme 7 4 Inner axis (or axial-like layer) consisting predominantly of sclerites Axis of scleroproteinous gorgonin, commonly with small amounts of embedded non-scleritic CaCO3; axis with hollow cross-chambered central core Axis of scleroproteinous gorgonin with large amounts of non-scleritic CaCO3 as internodes or embedded in the gorgonin; axis without hollow cross-chambered central core 5 N, number of described families (from Williams and Cairns, 2005). a One described family may not be valid. C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 previously using 18S rDNA and address the following speciWc questions: (1) What is the phylogenetic relationship between orders Alcyonacea and Pennatulacea? (2) Is there phylogenetic support for the taxonomic division of O. Alcyonacea into 6 sub-ordinal groups? (3) Is there phylogenetic support for the current family-level taxonomy of O. Alcyonacea? Although many aspects of octocoral phylogeny, including the root of the clade, remain uncertain, our study represents the most comprehensive phylogenetic treatment of this group to date, and provides a solid framework for future taxonomic revision and further testing of phylogenetic hypotheses. 2. Materials and methods 2.1. DNA extraction, ampliWcation, and sequencing Specimens used in this study included EtOH-preserved material from museum collections as well as fresh material collected by the authors. We obtained representatives of all three orders of Octocorallia, 5 of 6 sub-ordinal groups, 21 of 28 recognized families of O. Alcyonacea, and 7 of 14 families of O. Pennatulacea. A complete list of specimens including collection information, GenBank, and museum voucher catalog numbers is provided in Appendix A. DNA was extracted from specimens using previously published methods (e.g., Berntson and France, 2001; McFadden et al., 2001; Sánchez et al., 2003b). We used the primers of McFadden et al. (2004) to amplify the 5⬘ end of the NADH-dehydrogenase subunit 2 gene (ND2) (16S647F: 5⬘ACACAGCTCGGTTTCTATCTACCA-3⬘; ND21418R: 5⬘-ACATCGGGAGCCCACATA-3⬘), and primers ND42599F (5⬘-GCCATTATGGTTAACTATTAC-3⬘) (France and Hoover, 2002) or ND42625F (5⬘-TACGTG GYACAATTGCTG-3⬘) (Lepard, 2003) and Mut-3458R (5⬘-TSGAGCAAAAGCCACTCC-3⬘) (Sánchez et al., 2003b) to amplify the 5⬘ end of msh1. PCR protocols followed Sánchez et al. (2003b) and McFadden et al. (2004). For specimens that yielded no visible PCR product, we ran a second PCR reaction using an internal forward primer and 1 L of the original product as template (e.g., Berntson and France, 2001). For msh1, we used ND42625F as the internal primer; for ND2, initial ampliWcation was done using forward primer 16S544F (5⬘-CGACCTCGATGTT GAGTTGCGG-3⬘) and 16S647F was used as the internal primer. Negative (no DNA) controls from the Wrst reaction were re-ampliWed in the second reaction to check for sample contamination. PCR products were puriWed using a PEG-precipitation (Sánchez et al., 2003b) or agarase-digestion protocol (France and Hoover, 2002), cycle-sequenced and run on ABI3100 (PE Applied Biosystems), CEQ8000 (Beckman Coulter) or Global IR2 DNA (Li-Cor) automated sequencers. 2.2. Data analysis LaserGene software was used to translate nucleotide sequences using the cnidarian mitochondrial genetic code 515 (Pont-Kingdon et al., 1994). The amino acid sequences were aligned using ClustalX v. 1.81 (Thompson et al., 1997) and adjusted by eye. Phylogenetic trees were constructed using corresponding nucleotide alignments for each gene separately as well as for the combined dataset. Bayesian phylogenetic analyses were conducted using MrBayes v. 3.04 (Huelsenbeck and Ronquist, 2001) with a GTR + I +  model run for 1.5 £ 106 generations (burnin D 3750 generations) and separate data partitions for ND2 and msh1 in the combined analysis. PAUP* v. 4.0b10 (SwoVord, 2002) was used for both maximum parsimony and maximumlikelihood analyses. For maximum parsimony, we used a heuristic search with TBR branch-swapping; due to computational and time constraints, we ran 100 bootstrap replicates with a maximum of 1000 trees saved per replicate. Additional MP analyses were run with gaps coded as Wfth nucleotides. Maximum-likelihood analyses were run with model parameters (TVM+I+G) chosen using Modeltest v. 3.06 (Posada and Crandall, 1998); both the Akaike Information Criterion (AIC) and hierarchical likelihood ratio test (hLRT) selected this same model. A quartet-puzzling maximum-likelihood approach with the model mtREV24 (Adachi and Hasegawa, 1996) implemented in TREE-PUZZLE 5.2 (Schmidt et al., 2002) was used to analyze amino acid alignments for the purpose of rooting the octocoral tree. We used a weighted SH (WSH) test (Shimodaira and Hasegawa, 1999) and Shimodaira’s (2002) approximately unbiased (AU) test to compare the topology of the best-Wt maximum-likelihood tree to alternative trees in which we constrained speciWed sub-ordinal groups or families to be monophyletic. Both tests were implemented using the program CONSEL run with 10 sets of 100,000 bootstrap replicates (Shimodaira and Hasegawa, 2001). We chose these topology tests over the widely used SH test (Shimodaira and Hasegawa, 1999) because the latter is overly conservative and sensitive to the number of trees being compared (Buckley, 2002; Shimodaira, 2002). 3. Results We obtained both msh1 and ND2 sequences for a total of 103 genera of octocorals, and sequences of one of the two genes for an additional 12 genera (Appendix A). Whenever possible we sequenced at least two representatives of each genus. In most cases, congeners had identical or very similar nucleotide sequences, so we included only a single representative species from each genus in the phylogenetic analysis. Multiple representatives were included, however, for several genera whose relationships with related taxa appeared to be paraphyletic (e.g., Acanthogorgia, Alcyonium, Iciligorgia, and Pseudopterogorgia). Several genera shared identical mtDNA sequences, including Nephthea/ Litophyton, Xenia/Heteroxenia, and Paralemnalia digitiformis/Lemnalia. We included only one representative sequence from each of these pairs in the analysis. The ND2 fragment ranged in length from 148 to 158 amino acids. Relative to the longest sequence (Telestula), 516 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 most species had one or two deletions of 3–6 amino acids each near the 5⬘ end of the gene. The msh1 fragment ranged in length from 235 to 272 amino acids. Much of this length variation was attributable to three genera with large, unique insertions: Lepidisis, with three insertions of 11, 8, and 8 amino acids; Briareum, with insertions of 9, 4, and 4 amino acids; and Telestula, with a single unique 9 amino acid insertion. Despite these and other indels, nucleotide sequences of both genes maintained the correct reading frame, and we have no reason to suspect a loss of coding function for either gene. The Wnal nucleotide alignment of the two genes combined was 1429 bp in length and included 474 bp of ND2 and 955 bp of msh1. Of these 1429 nucleotides, 390 characters were invariant, and 745 of 1039 variable sites were parsimony-informative (including gaps as Wfth characters). 3.1. Rooting Because no other metazoans are known to have the mitochondrial msh1 gene (Culligan et al., 2000), we were unable to root the msh1 tree using a non-octocorallian outgroup. Attempts to use yeast msh1 as an outgroup sequence were unsuccessful due to our inability to align the amino acid sequences with any certainty. Attempts to root the ND2 tree using non-octocoral anthozoan ND2 sequences (Acropora tenuis, Metridium senile) were impaired by length diVerences between octocoral and hexacoral ND2. Because the octocoral ND2 is 70 amino acids longer at its 5⬘ end, we were able to align only 85 amino acids from the 3⬘ end of the ND2 fragment with the Acropora and Metridium sequences. Although maximum parsimony, Bayesian, and neighbor-joining analyses of the corresponding nucleotide alignments all placed Erythropodium at the base of the octocoral tree, analysis of this short fragment provided no support for any internal nodes within the Octocorallia clade. A quartet-puzzling maximum-likelihood analysis of the ND2 amino acid sequence using the mtREV24 model of evolution (Adachi and Hasegawa, 1996) provided somewhat better resolution of internal nodes and placed at the base of the tree an unresolved group of taxa that included both Erythropodium and Briareum. Because the outgrouprooted 18S phylogeny of Berntson et al. (2001) also weakly supported placement of these taxa at the base of the octocoral clade, we have shown our phylogenetic trees rooted with Erythropodium as the sister taxon to the remaining Octocorallia. Mid-point rooting yields a very similar tree topology that does not change our interpretation of clade structure and membership. 3.2. Trees Bayesian, maximum parsimony, and maximum-likelihood analyses all recovered very similar topologies, and a majority of the nodes that was well supported (>90%) by the Bayesian posterior probabilities also had strong support (>70%) from maximum parsimony bootstrap values (Fig. 1). Two distinct clades of octocorals were recovered in all analyses. One very large and well-supported clade included all members of the sea fan sub-order Holaxonia, a majority of the soft corals belonging to sub-ordinal group Alcyoniina, and some taxa of Scleraxonia and Stolonifera. The second clade, very well supported by the Bayesian posterior probabilities but only weakly supported by bootstrap values (Fig. 1), included all sea pens (O. Pennatulacea), the blue coral Heliopora (O. Helioporacea), and all members of the sea fan sub-order Calcaxonia. A third small clade, whose position relative to the two major clades remained unresolved, included several genera belonging to the soft coral family Alcyoniidae as well as the precious coral Corallium. To simplify, herein we refer to these three clades as Holaxonia–Alcyoniina, Calcaxonia–Pennatulacea, and Anthomastus–Corallium. The only discrepancies among phylogenetic methods were in the levels of support for some of the deeper nodes within each of the two major clades. Bayesian analyses strongly supported the separation of a heterogeneous group of taxa at the base of the Holaxonia–Alcyoniina clade, including several stoloniferans (Tubipora, Rhodelinda, and Telesto), scleraxonian sea fans (Iciligorgia, Solenocaulon), nidaliid soft corals (Nidalia, Chironephthya, Siphonogorgia), and the aberrant alcyoniid soft coral Malacacanthus. Although maximum parsimony analyses also positioned a subset of these taxa as a sister clade to the remaining Holaxonia–Alcyoniina, bootstrap values did not support this separation. Several basal nodes within the Calcaxonia–Pennatulacea clade were also more strongly supported by the Bayesian methods than maximum parsimony, in particular, the monophyly of Calcaxonia+Pennatulacea+Helioporacea. Coding gaps as Wfth characters in the maximum parsimony analyses weakened support for these nodes even further, but had little eVect on support for Fig. 1. Phylogenetic relationships among 103 genera in the anthozoan sub-class Octocorallia. Left: Maximum parsimony tree, strict consensus of 38,448 equally parsimonious trees of length 3714; values at nodes are percentages from 100 bootstrap replicates with maxtrees D 1000. Right: Bayesian likelihood tree, 50% majority-rule consensus of 12,251 trees (1.5 £ 106 generations; burnin D 3750); values at nodes are posterior probabilities. Circled numbers indicate the clades discussed in the text: 1, Holaxonia–Alcyoniina; 2, Calcaxonia–Pennatulacea; 3, Anthomastus-Corallium. Clades and taxon names shown in gray indicate regions of disagreement between the two trees. Colored bars indicate higher taxonomic groups: color indicates sub-ordinal group or order, letter abbreviation within bar indicates family. Light blue, Alcyoniina; green, SO. Calcaxonia; red, SO. Holaxonia; sage, Scleraxonia; yellow, Stolonifera; pink, O. Pennatulacea; dark blue, O. Helioporacea. Family abbreviations: Ac, Acanthogorgiidae; Al, Alcyoniidae; An, Anthothelidae; Br, Briareidae; Ce, Coelogorgiidae; Ch, Chrysogorgiidae; Cl, Clavulariidae; Co, Coralliidae; El, Ellisellidae; Go, Gorgoniidae; H, O. Helioporacea; Is, Isididae; Ke, Keroeididae; Ne, Nephtheidae; Ni, Nidaliidae; Pa, Paralcyoniidae; P, O. Pennatulacea; Pl, sF. Plexaurinae (F. Plexauridae); Pr, Primnoidae; St, sF. Stenogorgiinae (F. Plexauridae); Tu, Tubiporidae; Xe, Xeniidae. C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 517 518 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 any of the internal relationships within the Holaxonia– Alcyoniina clade. The position of the stoloniferan genus Clavularia as a sister clade to Holaxonia–Alcyoniina was well supported by all analyses. All phylogenetic methods recovered a monophyletic O. Pennatulacea, but none of the Wve sub-ordinal groups of O. Alcyonacea was monophyletic. The maximum-likelihood tree (¡ln L D 20319.07) (Fig. 2) was signiWcantly better than alternative trees in which the monophyly of Holaxonia, Calcaxonia, Alcyoniina, Scleraxonia, and Stolonifera was sequentially enforced (AU and WSH topology tests, p < 0.01) (Table 2). Representatives of groups Alcyoniina, Scleraxonia, and Stolonifera occurred both in the well-supported Holaxonia–Alcyoniina clade and among the unresolved nodes at the base of the tree. Although there was moderate support for most of the scleraxonians and stoloniferans forming a sister clade to the remaining Holaxonia–Alcyoniina, genera such as Anthothela (Scleraxonia) and Coelogorgia (Stolonifera) were nested well within clades of Alcyoniina, while Erythropodium, Briareum (both Scleraxonia), and Telestula (Stolonifera) fell entirely outside of Holaxonia–Alcyoniina. The sea fan sub-order Calcaxonia was paraphyletic with O. Pennatulacea, with all analyses strongly supporting the family Ellisellidae as the sister taxon to the sea pens. A tree in which the Anthomastus–Corallium clade was constrained to be the sister group to the sea pens had a signiWcantly lower likelihood (Table 2). Maximum-likelihood and Bayesian analyses also placed the blue coral Heliopora (O. Helioporacea) within Calcaxonia–Pennatulacea, but a constrained tree in which Heliopora was excluded from that clade was not signiWcantly less likely than the maximumlikelihood tree (Table 2). Although conWned to the Holaxonia–Alcyoniina clade, members of the sea fan sub-order Holaxonia fell into a number of distinct sub-clades among which basal relationships could not be resolved. When all taxa were included in the analysis, the monophyly of Holaxonia was rejected (Table 2). However, when Thrombophyton, a soft coral genus that nested within a mixed clade of holaxonian sea fans, was removed from the analysis, the maximum-likelihood tree (¡ln L D 20246.31) was only marginally better than a tree with holaxonian monophyly enforced (¡ln L D20262.06) (AU test, p D 0.04; WSH test, p D 0.085). In the absence of Thrombophyton, therefore, we cannot with conWdence reject the monophyly of Holaxonia. In addition to the lack of support for monophyly of most of the sub-ordinal groups of Alcyonacea, most of the families for which we sampled multiple genera were also polyphyletic (Table 2). Genera belonging to the families Alcyoniidae [Alcyoniina], Clavulariidae [Stolonifera], and Anthothelidae [Scleraxonia] were found both within Holaxonia–Alcyoniina and outside of that clade. Topology tests rejected alternative trees that enforced the monophyly of each of these families, along with the holaxonian sea fan families Plexauridae, Gorgoniidae, and Acanthogorgiidae and the soft coral families Nephtheidae, Nidaliidae, and Xeniidae (Table 2). Only the four calcaxonian sea fan families, Ellisellidae, Isididae, Chrysogorgiidae, and Primnoidae, were monophyletic in our analysis. Phylogenetic analyses of each gene separately produced trees that were similar in topology but only weakly supported relative to the combined tree (Fig. 3). Single-gene analyses did, however, allow us to examine the phylogenetic positions of some taxa for which we were able to sequence only one of the two genes. On the basis of analyses of ND2 alone, two additional alcyoniid soft coral genera, Paraminabea and Notodysiferus, belonged to the small Anthomastus–Corallium clade. The xeniid genera Sympodium and EZatounaria were strongly supported as members of the xeniid clade, with EZatounaria positioned basal to a polytomy of seven other xeniid genera, and the sea pen genus Anthoptilum fell within the pennatulacean clade. msh1 sequences placed the scleraxonian genera Mopsella and Asperaxis in a clade with scleraxonians Iciligorgia and Solenocaulon, and the plexaurid Paracis within the Stenogorgiinae–Acanthogorgiidae clade (weakly supported as a sister taxon to Lepidomuricea). The alcyoniid Elbeenus lauramartinae and acanthogorgiid Cyclomuricea both belonged to the Holaxonia–Alcyoniina clade, but neither taxon grouped with other family members. The two genera that comprise the calcaxonian family Ifalukellidae (Ifalukella, Plumigorgia) formed a well-supported clade that grouped with other calcaxonians. 4. Discussion Our mitochondrial gene phylogeny supports the results of previous 18S rDNA and 16S mtDNA studies that divide the subclass Octocorallia into two or three distinct clades (France et al., 1996; Berntson et al., 2001; Won et al., 2001; Sánchez et al., 2003a). One clade includes all members of the sea fan sub-order Holaxonia, a majority of the taxa belonging to the soft coral group Alcyoniina, and most of the taxa in groups Scleraxonia and Stolonifera. This clade corresponds well to clade C of Berntson et al.’s (2001) 18S phylogeny and the “Alcyoniina–Holaxonia” clade of Sánchez et al. (2003a). A second large clade includes all of the sea pens (O. Pennatulacea), blue corals (O. Helioporacea), and the sea fan sub-order Calcaxonia. Combined with the small Anthomastus–Corallium clade, this clade corresponds to the “Calcaxonia” clade of Sánchez et al. (2003a). Berntson et al.’s (2001) 18S phylogeny subdivided the Calcaxonia– Pennatulacea clade into two separate clades, one (B) corresponding to the Pennatulacea+Ellisellidae clade in our tree and the other (A) comprising three calcaxonian families plus the Anthomastus–Corallium clade. Our analyses of mtDNA support a monophyletic Calcaxonia + Pennatulacea that excludes Anthomastus–Corallium, and a tree that enforces the three clades found by Berntson et al. (2001) is signiWcantly less likely than our best ML tree based on the AU test (but not the WSH test; Table 2). One discrepancy between our phylogeny and the less resolved phylogeny of Berntson et al. (2001) is the position C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 519 Fig. 2. Maximum-likelihood phylogram (¡ln L D 20319.07) of relationships among 103 genera in the anthozoan sub-class Octocorallia. Circled numbers indicate the clades discussed in the text: 1 D Holaxonia–Alcyoniina; 2 D Calcaxonia–Pennatulacea; 3 D Anthomastus–Corallium. 520 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 Table 2 Log-likelihood scores of trees with speciWed taxonomic groups of octocorals constrained to be monophyletic Constrained group ¡ln L AU WSH None (best ML tree) Holaxonia Calcaxonia Alcyoniina Stolonifera Scleraxonia (Pennatulacea+Calcaxonia) (Pennatulacea+Anthomastus) Berntson et al. (A+B+C) F. Alcyoniidae F. Nephtheidae F. Nidaliidae F. Xeniidae F. Plexauridae F. Gorgoniidae F. Acanthogorgiidae F. Clavulariidae F. Anthothelidae 20319.07 20402.53 20388.12 20800.90 20559.14 20638.85 20324.35 20401.15 20345.50 20867.73 20498.25 20352.69 20396.59 20693.42 20520.88 20549.58 20521.06 20431.61 ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ns ¤¤¤ ¤¤ ¤¤¤ ¤¤¤ ¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ns ¤¤¤ ns ¤¤¤ ¤¤¤ ns ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ ¤¤¤ Likelihood diVerences between the best ML tree (Fig. 2) and constrained trees were compared using Shimodaira’s (2002) approximately unbiased (AU) test and a weighted Shimodaira–Hasegawa (WSH) test (Shimodaira and Hasegawa, 1999). ¤¤¤p < 0.001; ¤¤p < 0.01; ns: p > 0.05. of Heliopora (O. Helioporacea), which they placed in the Holaxonia–Alcyoniina (clade C) as a sister taxon to Alcyonium gracillimum. They questioned this placement, however, because the Heliopora sequence used in their analyses was very incomplete. Our contrasting placement of Heliopora within the Calcaxonia–Pennatulacea clade was well supported by both ND2 and msh1 sequences. Berntson et al. also failed to recover a monophyletic Pennatulacea due to the inclusion of the sea pen Umbellula in a clade with Anthomastus and Corallium, a result not supported by our mitochondrial data. Finally, their phylogeny included a partial sequence for the calcaxonian genus Plumigorgia (Family Ifalukellidae) within Holaxonia–Alcyoniina. Our analyses of msh1 unite both Plumigorgia and Ifalukella with the other calcaxonian families in Calcaxonia–Pennatulacea (Fig. 3A). All molecular phylogenetic studies conducted to date support the division of the sub-class into at least two major clades, neither of which corresponds to the traditional ordinal divisions within sub-class Octocorallia. Both mitochondrial and nuclear 18S rDNA studies support the separation of sea pens (O. Pennatulacea) and the sea fan sub-order Calcaxonia into one major clade, and the sea fan sub-order Holaxonia plus the majority of soft corals and other groups of Alcyonacea into the other clade. Both types of molecular data also recognize the scleraxonian sea fan Corallium, soft coral Anthomastus, and several other genera as members of a third small clade whose relationship to the two major clades remains uncertain. 4.1. Relationship between Alcyonacea and Pennatulacea On the basis of a cladistic analysis of morphological characters, Williams (1994, 1997) hypothesized that the sea pens, a group with a highly derived colony growth morphology, evolved from a soft coral ancestor similar to the alcyoniid soft coral genus Anthomastus. Although our molecular analyses suggest that Anthomastus may be more closely related to the Calcaxonia–Pennatulacea clade than other soft corals, our data do not support a sister relationship between the Anthomastus–Corallium clade and the sea pens. Instead, our phylogeny strongly supports the calcaxonian sea fan family Ellisellidae as the sister group to O. Pennatulacea, a relationship Bayer (1955) proposed on the basis of observed similarities in the axial structure of the two groups. Observed under SEM, the structure and orientation of calcareous microcrystals in the axis of an ellisellid (Ctenocella sp.) and a pennatulacean (Virgulariidae) appear remarkably similar (J.A. Sánchez, unpubl. data). It is possible, however, that inclusion in our analysis of the genera Veretillum and Echinoptilum, hypothesized by Williams (1994, 1997) to be the most primitive sea pen groups, could change the inferred phylogenetic relationships within the Calcaxonia–Pennatulacea clade. 4.2. Phylogenetic support for sub-ordinal groups of Alcyonacea Our results provide phylogenetic support for only two of the six currently recognized sub-ordinal divisions within O. Alcyonacea (Table 1). These taxonomic divisions are widely acknowledged to be problematic and likely to reXect grades of colony architecture rather than phylogenetic relationships (Fabricius and Alderslade, 2001). Of the six groups, only the sea fan sub-orders Holaxonia and Calcaxonia are morphologically discrete and deWned by synapomorphies. The skeletal axis of all calcaxonians has a core of scleroproteinous gorgonin with large amounts of embedded, nonscleritic calcium carbonate. In contrast, the skeletal axis of holaxonians contains only small amounts of embedded calcium carbonate (mostly non-scleritic, although one group has sclerites in the axis) and has a hollow cross-chambered core. Our phylogeny is consistent with a single-evolutionary origin of the calcaxonian skeletal axis and with the axis of the sea pens having been derived from that of a calcaxonian ancestor. Our phylogeny is also consistent with a singleevolutionary origin of the holaxonian skeletal axis, provided we entertain the possibility that Thrombophyton, a genus of encrusting soft corals nested deep within a clade of holaxonians, may be a holaxonian that has secondarily lost the skeletal axis. Although the basal nodes within our Holaxonia–Alcyoniina clade remain unresolved, we are unable to reject with conWdence the monophyly of Holaxonia when Thrombophyton is removed from the analysis. The molecular phylogeny, therefore, supports the taxonomic division of sea fans into the distinct sub-orders Calcaxonia and Holaxonia. Our molecular data do not, however, support the phylogenetic distinction of the other three sub-ordinal groups of Alcyonacea included in our analysis (Alcyoniina, Scleraxonia, and Stolonifera). Representatives of each of these C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 521 Fig. 3. Phylogenetic relationships among octocoral genera based on single-gene analyses of (A) msh1 and (B) ND2. Maximum-likelihood tree topology is shown (msh1: ¡ln L D 16012.12; ND2: ¡ln L D 5363.58). Asterisks indicate nodes that had maximum parsimony bootstrap values >70% and Bayesian posterior probabilities >90%. The same parameters were used for all analyses as for the combined tree. Boldface indicates species that were not included in the combined analysis. 522 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 groups were distributed throughout the Holaxonia–Alcyoniina clade, the Anthomastus–Corallium clade, and the poorly resolved basal region of the octocoral tree. Our phylogeny therefore supports the assumption, evidenced by the reluctance of octocoral taxonomists to elevate them to the status of sub-orders, that these groups represent grades of morphological construction that have likely evolved repeatedly during the history of Octocorallia. For instance, although the taxa included within the group Scleraxonia share a skeletal axis or axial-like layer containing sclerites, the details of skeletal construction diVer substantially among families (Fabricius and Alderslade, 2001), consistent with multiple independent derivations of this type of axis. Although the majority of the soft coral taxa classiWed as Alcyoniina belonged to the Holaxonia–Alcyoniina clade, a notable subset of genera fell into the small AnthomastusCorallium clade. These genera all belong to the morphologically heterogeneous family Alcyoniidae and include Anthomastus, Paraminabea, Notodysiferus, and an undescribed genus with an encrusting growth form. The scleraxonian genus Corallium was united with these alcyoniids, and the 18S study of Berntson et al. (2001) also placed the scleraxonian Paragorgia in this same clade. Sánchez (2005) noted the similarity and possible homology of the surface sclerites in Corallium and the two paragorgiid genera Paragorgia and Sibogagorgia, sclerite forms that also resemble those found in Anthomastus (Broch and Horridge, 1957). A more obvious morphological trait shared by all of these disparate genera is polyp dimorphism. Our results do not, however, support Berntson et al.’s (2001) suggestion that dimorphism has evolved only once in Octocorallia and is a synapomorphy shared by this clade and the sea pens. We Wnd no support for a sister relationship between the sea pens and the Anthomastus–Corallium clade. Moreover, although Corallium and Paragorgia are the only dimorphic scleraxonians (Bayer, 1964), there are several other dimorphic genera of soft corals (the alcyoniids Sarcophyton and Lobophytum and the xeniid Heteroxenia) that do not belong to the Anthomastus–Corallium clade. Polyp dimorphism, therefore, has clearly evolved independently within multiple lineages of octocorals. 4.3. Family monophyly Any inferences we can make regarding the monophyly of octocoral families are necessarily limited by taxon sampling, as in no case does our phylogeny include all of the genera belonging to a family (with the exception of monotypic families). Despite this limitation, we can nonetheless conclude that most families of octocorals do not represent monophyletic groups (Fig. 1; Table 2). The main exceptions are the calcaxonian sea fan families Ellisellidae, Chrysogorgiidae, Primnoidae, Isididae, and Ifalukellidae, all of which form well-supported clades in our trees. However, we have sampled representatives of only one of four described sub-families of Isididae, which additional molecular data suggest may not constitute a monophyletic group (S.C. France, unpubl. data). The pitfalls of inferring intra-familial relationships when taxon sampling is incomplete are illustrated by our results for the holaxonian sea fan families Plexauridae and Gorgoniidae. Several previous molecular systematic studies have already addressed the relationships among taxa in these two families, concluding that they are paraphyletic and comprise at least three major clades (Gorgoniidae, Plexaurinae, and Stenogorgiinae) plus several minor clades among which the basal relationships are unclear (Sánchez et al., 2003b; Wirshing et al., 2005). Our results support those conclusions. Neither previous study, however, included representatives of the holaxonian family Acanthogorgiidae. Our results indicate that the acanthogorgiid genera Acanthogorgia and Anthogorgia also fall within Stenogorgiinae; moreover, within this clade, Acanthogorgia comprises two distinct lineages that are not closely related. This previously unrecognized paraphyletic relationship between Acanthogorgiidae and Stenogorgiinae further confounds the morphological arguments that have been oVered as support for the resurrection of family Paramuriceidae ( D Stenogorgiinae) (Wirshing et al., 2005). 4.4. Conclusions Our phylogenetic analysis of mitochondrial protein-coding sequences supports the two-clade structure of sub-class Octocorallia seen in previous analyses of 18S rDNA and 16S mtDNA (France et al., 1996; Berntson et al., 2001; Won et al., 2001; Sánchez et al., 2003a). Our study includes the most comprehensive taxon sampling to date and provides strong support for the phylogenetic relationships among genera within many octocoral families. Most of the basal nodes in our tree, however, are still poorly resolved, and we are unable to infer the phylogenetic relationships among most families, especially within the large Holaxonia–Alcyoniina clade. The failure of both 18S rDNA (Berntson et al., 2001) and the more rapidly evolving mitochondrial markers used here to resolve these deeper nodes suggests that the Holaxonia–Alcyoniina clade underwent a rapid radiation (Page and Holmes, 1998). Unfortunately, the lack of a fossil record for the group means we are unable to estimate dates of divergence or corroborate this hypothesis with direct fossil evidence. It is likely that the basal relationships within Holaxonia–Alcyoniina will be resolved only by the eventual accumulation of large amounts of sequence data for a variety of nuclear genes. The root of the Octocorallia phylogeny also remains uncertain, and further molecular studies using genes that are shared and can be aligned with certainty among octocorals and their sister group, Hexacorallia, will be necessary to conWrm the basal topology of the clade. Although our ND2 data and the 18S rDNA data of Berntson et al. (2001) both suggest that the scleraxonian genera Erythropodium and Briareum lie at or near the base of Octocorallia, neither study provides strong support for this root. Our conclusions regarding the division of Octocorallia into two major clades would, however, require re-interpretation only if the 523 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 root was shown to lie within either the Holaxonia–Alcyoniina or Calcaxonia–Pennatulacea clade. Neither study provides strong support for that scenario. Despite the lack of basal resolution in our phylogeny, it is already clear that a number of families in O. Alcyonacea are in need of taxonomic revision. In particular, families Alcyoniidae, Anthothelidae, and Clavulariidae each comprise phylogenetically heterogeneous groups of genera. The holaxonian sea fan families Plexauridae, Gorgoniidae, and Acanthogorgiidae are paraphyletic, and although we recovered several distinct, well-supported holaxonian clades, their membership does not reXect current morphologybased taxonomic divisions. Likewise, the soft coral families Nephtheidae and Paralcyoniidae are paraphyletic, and several genera of Nephtheidae belong to disparate clades. Although the molecular phylogeny does not support the division of O. Alcyonacea into the six sub-ordinal groups currently recognized on the basis of morphology, our results do not suggest any clear alternative classiWcation scheme. The sea pens (O. Pennatulacea) and the sea fan sub-orders Calcaxonia and Holaxonia are well-deWned groups supported by synapomorphies and should be retained as higher taxonomic groups. The morphological characters that loosely deWne the Scleraxonia, Alcyoniina, and Stolonifera (Table 1) apparently have evolved multiple times within Octocorallia, however, and those three subordinal designations are clearly artiWcial. Although a revision of the higher taxonomic levels of Octocorallia will ideally reXect the division of the sub-class into two major clades, the morphological similarities and distinctions among taxa belonging to each of these clades are not yet obvious. With a molecular phylogenetic framework to help guide taxonomic revision, we are hopeful that morphological synapomorphies eventually will be found that will allow us to divide Octocorallia into higher taxonomic units that reXect accurately both the underlying phylogeny and morphological diversity of the group. Acknowledgments We thank Allen Collins for comments on the manuscript and assistance with analyses, Leen van Ofwegen, Peter Wirtz, John Starmer, Ed Seidel, Stephen Cairns, and the Coral Reef Research Foundation (Pat and Lori Colin) for contributing specimens, and Holly Johnsen, Nicholas Johnson, Ian Tullis, M. Breton Hutchinson, Katherine Winner, Jill Sohm, Mercer Brugler, Walter Renne, and Elizabeth Jones for generating some of the sequence data included here. Partial support for this work was provided by a Harvey Mudd College Faculty Research Program award to CSM, a Howard Hughes Medical Institute award to Harvey Mudd College, the Museum and Art Gallery of the Northern Territory, grant no. NA03OAR4600116 from NOAA’s OYce of Ocean Exploration to SCF, a Smithsonian Institution postdoctoral fellowship to JAS, the National Institute of Water and Atmospheric ResearchNIWA New Zealand, and Universidad de los Andes (Facultad de Ciencias). Appendix A Octocoral specimens for which partial msh1 and ND2 sequences were obtained Family [subfamily] and species O. ALCYONACEA [STOLONIFERA] Clavulariidae [Clavulariinae] Clavularia inXata Rhodelinda sp. [Telestinae] Carijoa riisei ?Telesto sp. Telestula sp. Tubiporidae Tubipora sp. Coelogorgiidae Coelogorgia palmosa [ALCYONIINA] Alcyoniidae [Alcyoniinae] Alcyonium digitatum Alcyonium ?aurantiacum Cladiella sp. Dampia pocilloporaeformis Discophyton rudyi Elbeenus lauramartinae Eleutherobia sp. Klyxum sp. Lobophytum ?strictum Collection locality Catalog number Date msh1 ND2 Great Barrier Reef, Qld, AUS Kermadec Islands NTM-C011542 NTM-C010034 1991 1989 DQ302799 DQ302800 DQ302873 DQ302874 Rio de Janeiro, Brazil Gabo I., Victoria, AUS Tasman Sea, AUS CSM-TELCF NTM-C012710 NTM-C014984 2003 2000 2003 DQ302801 DQ302802 DQ302803 DQ302875 DQ302876 DQ302877 Lighthouse Lagoon, Palau UF 1811 2000 DQ302804 DQ302878 Blue Corner East, Palau NTM-C014914 2005 DQ302805 DQ302879 Isle of Man New Zealand Gunn Pt, NT, AUS Rowley Shoals, WA, AUS Tatoosh I., WA, USA Uchelbeluu Reef, Koror, Palau West Channel, Palau Indonesia Agat Bay, Guam SBMNH 360700 NTM-C014988 NTM-C012095 NTM-C005805 CSM-TAN14 NTM-C013108 NTM-C014902 NTM-C012417 NTM-C014937 1992 1999 1993 1987 1991 2001 2005 1996? 1998 AY607777 AF530482 DQ302806 DQ302880 DQ302807 DQ302881 DQ280593 DQ302882 DQ302808 AF530488 DQ536320 N/A DQ302809 DQ302883 DQ302810 DQ302884 DQ280585 AF530495 (continued on next page) 524 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 Appendix A (continued) Family [subfamily] and species Collection locality Catalog number Date msh1 ND2 Malacacanthus capensis Notodysiferus d’hondti Paraminabea aldersladei Rhytisma sp. Sarcophyton glaucum Sinularia maxima Thrombophyton coronatum Alcyoniidae n. gen. [Anthomastinae] Anthomastus ritteri Nephtheidae Capnella imbricata Dendronephthya sp. Gersemia sp. Lemnalia sp. Litophyton sp. Nephthea sp. Neospongodes sp. Paralemnalia digitiformis Scleronephthya sp. Stereonephthya sp. Umbellulifera sp. Nidaliidae [Nidaliinae] Nidalia sp. Pieterfaurea sp. [Siphonogorgiinae] Chironephthya sp. Nephthyigorgia sp. Siphonogorgia sp. Paralcyoniidae Paralcyonium spinulosum Studeriotes sp. Xeniidae Anthelia sp. Asterospicularia randalli Cespitularia sp. EZatounaria sp. Heteroxenia sp. Sansibia sp. Sarcothelia sp. Sympodium sp. Xenia sp. [SCLERAXONIA] Briareidae Briareum polyanthes Anthothelidae [Anthothelinae] Anthothela sp. Erythropodium caribaeorum [Semperininae] Iciligorgia sp. 1 Iciligorgia sp. 2 Solenocaulon sp. Coralliidae Corallium ducale Melithaeidae [Asperaxinae] Asperaxis karenae [Melithaeinae] Mopsella sp. [HOLAXONIA] Acanthogorgiidae Acanthogorgia sp. Algoa Bay, South Africa King George Sound, WA, AUS Chuuk, Micronesia Swain Reefs, Qld, AUS Rowley Shoals, WA, AUS Piti Bay, Guam Santa Catalina I., CA, USA North West Cape, WA, AUS CSM-SAFR155.2 NTM-C014221 NTM-C012053 NTM-C001942 NTM-C010771 UF 3500 SBMNH 145123 WAMZ 13105 1998 1989 1993 1980 1987 1998 1992 2002 DQ302811 N/A N/A DQ302812 DQ280525 DQ302813 DQ302814 DQ302815 DQ302885 DQ302886 DQ302887 DQ302888 DQ302889 DQ302890 DQ302891 DQ302892 Pebble Beach, CA, USA CSM-ANRI 1998 DQ302816 DQ302893 GBR, Qld, AUS Darwin Harbor, NT, AUS Balsfjorden, Norway Ashmore Reef, WA, AUS Cartier I., NT, AUS Andaman & Nicobar I. Darwin Harbor, NT, AUS GBR, Qld, AUS Palau Cartier I., NT, AUS Arafura Sea, NT, AUS NTM-C012235 NTM-C012655 RMNH-Coel.14708 NTM-C011720 NTM-C011318 NTM-C012400 NTM-C013130 NTM-C012309 NTM-C011489 NTM-C011307 NTM-C011063 1992 2000 1978 1986 1992 1995? 2001 1994? 1990 1992 1990 DQ302817 DQ302818 DQ302819 DQ302820 DQ302821 DQ302822 DQ302823 DQ302824 DQ302825 DQ302826 DQ302827 DQ302894 DQ302895 DQ302896 DQ302897 DQ302898 DQ302899 DQ302900 DQ302901 DQ302902 DQ302903 DQ302904 West Channel, Palau Algoa Bay, South Africa NTM-C014876 NTM-C013943 2005 1998 DQ302828 DQ302829 DQ302905 DQ302906 Arafura Sea, NT, AUS Caledon Bay, NT, AUS Rowley Shoals, WA, AUS NTM-C012426 NTM-C011345 NTM-C011159 1995 1991 1987 DQ302830 DQ302831 DQ302832 DQ302907 DQ302908 DQ302909 Sesimbra, Portugal Darwin, NT, AUS CSM-PASP NTM-C011441 2001 1982 DQ302833 DQ302834 DQ302910 DQ302911 Tasman Sea, AUS Double Reef, Guam Semporna I., Malaysia Central GBR, AUS Rowley Shoals, WA, AUS Darwin Harbor, NT, AUS Oahu, Hawaii Central GBR, AUS Ngederak, Palau NTM-C013050 CSM-JS10.12.98.05 NTM-C013542 NTM-C012311 NTM-C010897c NTM-C012955 NTM-C015151 NTM-C012271 CSM-XESP 1997 1998 1999 1994 1987 2001 2000 1994 2000 DQ302835 DQ302836 DQ302837 DQ302838 DQ302839 DQ302840 DQ302841 N/A DQ302842 DQ302912 AF530497 DQ302913 DQ302914 DQ302915 DQ302916 DQ302917 DQ302918 AF530496 AY533653 AY534734 Oceanographer Cyn, NW Atl Bocas del Toro, Panama USNM 1014917 2001 2003 DQ297415 DQ302843 DQ297434 DQ302919 Tasman Sea, AUS NE Kalimantan, Indonesia NE Kalimantan, Indonesia NTM-C014376 RMNH Coel. 33331 RMNH Coel. 33332 2003 2003 2003 DQ302844 DQ302845 DQ302846 DQ302920 DQ302921 DQ302922 Cross Seamount, Hawaii USNM 94456 1993 DQ297416 DQ297435 Port Davey, Tasmania, AUS NTM-C013575 2002 DQ302847 N/A Gulf of Carpentaria, NT, AUS NTM-C014468 2003 DQ302848 N/A Bishop Seamount, Hawaii USNM 94442 1993 AY268461 DQ297436 525 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 Appendix A (continued) Family [subfamily] and species Collection locality Catalog number Date msh1 ND2 Acanthogorgia angustifolia Anthogorgia sp. Calcigorgia sp. Cyclomuricea sp. Gorgoniidae Gorgonia ventalina Guaiagorgia sp. Leptogorgia virgulata PaciWgorgia stenobrochis Phyllogorgia dilatata Pseudopterogorgia acerosa Pseudopterogorgia sp. Pterogorgia citrina Keroeididae Keroeides gracilis Plexauridae [Plexaurinae] Alaskagorgia aleutiana Eunicea knighti Euplexaura sp. Muricea muricata Muriceopsis Xavida Plexaura homomalla Plexaurella nutans Pseudoplexaura crucis [Stenogorgiinae] Plexauridae n. gen. Astrogorgia sp. Lepidomuricea sp. Menella sp. Muriceides sp. Paracis sp. Paramuricea sp. Paraplexaura sp. Swiftia sp. [CALCAXONIA] Chrysogorgiidae Chrysogorgia chryseis Iridogorgia sp. Metallogorgia melanotrichos Radicipes gracilis Ellisellidae Ctenocella barbadensis Nicella sp. Verrucella sp. Ifalukellidae Ifalukella yanii Plumigorgia sp. Isididae [Keratoisidinae] ?Isidella sp. Lepidisis olapa Primnoidae Calyptrophora sp. Narella nuttingi Paracalyptrophora sp. Thouarella sp. Bear Seamount Tasman Sea, AUS Bobrof Is., Alaska, USA Tasman Sea, AUS USNM 100895 NTM-C014983 SCF-BOI09131 NTM-C014571 2000 2003 2003 2003 DQ297418 DQ302849 DQ297419 DQ302850 DQ297437 DQ302923 DQ297438 N/A Cayo Lobo, Puerto Rico Gulf of Carpentaria, NT, AUS North Carolina, USA E PaciWc Rio de Janeiro, Brazil Lee Stocking I., Bahamas Gulf of Carpentaria, NT, AUS Lee Stocking I., Bahamas USNM 1007421 NTM-C014545 USNM 1007414 JAS-27 MNRJ-4336 USNM 1007413 NTM-C014541 USNM 1007406 2000 2003 2000 1996 2000 2000 2003 2000 AY126425 DQ302851 AY126418 AY126420 AY126428 AY126421 DQ302852 AY126402 AY126397 DQ302924 AY126390 AY126392 AY126400 AY126393 DQ302925 AY126374 Tasman Sea, AUS NTM-C014573 2003 DQ302853 DQ302926 Aleutian Islands, AK, USA Cat Island, Bahamas Gulf of Carpentaria, NT, AUS Lee Stocking I., Bahamas San Salvador, Bahamas Florida, USA Lee Stocking I., Bahamas Lee Stocking I., Bahamas USNM 1007125 USNM 1007366 NTM-C014536 USNM 1007340 USNM 1007376 USNM 1007399 USNM 1007378 1994 2001 2003 2000 1999 2001 2000 2000 AY533649 AY126404 DQ302854 AY126408 AY126416 AY126410 AY126415 AY126401 AY534738 AY126376 DQ302927 AY126380 AY126388 AY126382 AY126387 AY126373 Tasman Sea, AUS Tasman Sea, AUS Tasman Sea, AUS Gulf of Carpentaria, NT, AUS Tasman Sea, AUS Tasman Sea, AUS Muir Seamount Gulf of Carpentaria, NT, AUS Tasman Sea, AUS NTM-C014562 NTM-C014408 NTM-C014578 NTM-C014493 NTM-C014445 NTM-C014576 YPM 28867 NTM-C014494 NTM-C014396 2003 2003 2003 2003 2003 2003 2003 2003 2003 DQ302855 DQ302856 DQ302857 DQ302858 DQ302859 DQ302860 DQ297420 DQ302861 DQ302862 DQ302928 DQ302929 DQ302930 DQ302931 DQ302932 N/A DQ297439 DQ302933 DQ302934 Cross Seamount Muir Seamount Manning Seamount Bear Seamount SCF-CR106-2 YPM 28866 SCF-MAN306-1 USNM 100900 1993 2003 2003 2000 DQ297421 DQ297422 DQ297423 DQ297424 DQ297440 DQ297441 DQ297442 DQ297443 Tasman Sea, AUS Tasman Sea, AUS NTM-C014406 NTM-C014982 2003 2003 AY533651 DQ302863 DQ302864 AY534736 DQ302935 DQ302936 Yap, Micronesia Palawan, Philippines 0CDN 3387-Q 0CDN 3101-G 1992 1995 DQ536319 DQ536318 N/A N/A Lanai Is., Hawaii, USA Cross Seamount, Hawaii, USA SCF-LAD25 SCF-CR206-4 1996 1993 DQ297425 DQ297426 DQ297444 DQ297445 Lanai Is., Hawaii, USA Pensacola Seamount, Hawaii Penguin Bank, Hawaii, USA Oceanographer Cyn, NW Atl SCF-LAD36 USNM 94424 SCF-PBS09 USNM 1014915 1996 1993 1996 2001 DQ297427 DQ297428 DQ297429 DQ297430 DQ297446 DQ297447 DQ297448 DQ297449 Tasman Sea, AUS Colombia NTM-C014985 INVEMAR 2003 DQ302865 DQ302937 DQ311678 DQ311679 (continued on next page) O. PENNATULACEA [SESSILIFLORAE] Kophobelemnidae Kophobelemnon macrospinum Sclerobelemnon theseus 526 C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 Appendix A (continued) Family [subfamily] and species Collection locality Catalog number Date msh1 ND2 Tasman Sea, AUS NTM-C014385 2003 N/A DQ302938 Bishop Seamount, Hawaii Tasman Sea, AUS USNM 94465 NTM-C014561 1993 2003 DQ297431 DQ302866 DQ297450 DQ302939 Gulf of Mexico, Florida, USA SCF-FLA 2001 DQ297432 DQ297451 Tasman Sea, AUS NTM-C014384 2003 DQ302867 DQ302940 Tasman Sea, AUS NTM-C014596 2003 DQ302868 DQ302941 Tasman Sea, AUS Tasman Sea, AUS Tasman Sea, AUS NTM-C014392 NTM-C014415 NTM-C014391 2003 2003 2003 DQ302869 DQ302870 DQ302871 DQ302942 DQ302943 DQ302944 Blue Corner East, Palau CRCNI 577 2005 DQ302872 DQ302945 Anthoptilidae Anthoptilum murrayi Protoptilidae Protoptilum sp. Distichoptilum gracile Renillidae Renilla muelleri Umbellulidae Umbellula sp. [SUBSESSILIFLORAE] Halipteridae Halipteris Wnmarchica Pennatulidae Gyrophyllum sp. Pennatula sp. Pteroeides sp. O. HELIOPORACEA Helioporidae Heliopora coerulea Sub-ordinal group is indicated in uppercase boldface and square brackets, family in lowercase boldface. Catalog numbers are indicated for voucher specimens, GenBank Accession Numbers for msh1 and ND2 sequences. CSM, Collection of C.S. McFadden; CRCNI, 0CDN: Coral Reef Research Foundation, Palau; JAS, Collection of J.A. Sánchez; MNRJ, Museu Nacional do Rio de Janeiro, Brazil; NTM, Museum and Art Gallery of the Northern Territory, Australia; RMNH, Nationaal Natuurhistorisch Museum, formerly Rijksmuseum van Natuurlijke Historie, Leiden, The Netherlands; UF, Florida Natural History Museum, FL, USA; USNM, National Museum of Natural History, Smithsonian Institution, USA; SBMNH, Santa Barbara Museum of Natural History, CA, USA; SCF, Collection of S.C. France; WAMZ, Western Australian Museum; YPM, Yale Peabody Museum, New Haven, CT, USA. References Adachi, J., Hasegawa, M., 1996. Model of amino acid substitution in proteins encoded by mitochondrial DNA. J. Mol. Evol. 42, 459–468. Bayer, F.M., 1955. Contributions to the nomenclature, systematics, and morphology of the Octocorallia. Proc. US Natl. Museum 105, 207– 220. , pl. 1–8. Bayer, F.M., 1964. The genus Corallium (Gorgonacea: Scleraxonia) in the western North Atlantic Ocean. Bull. Mar. Sci. Gulf Carib. 14, 465–478. Bayer, F.M., 1981a. Status of knowledge of octocorals of world seas. Seminarios de Biologia Marinha. Academia Brasileira de Ciencias, Rio de Janeiro, pp. 3–11. Bayer, F.M., 1981b. Key to the genera of Octocorallia exclusive of Pennatulacea (Coelenterata: Anthozoa), with diagnosis of new taxa. Proc. Biol. Soc. Wash. 94, 902–947. Berntson, E.A., France, S.C., 2001. Generating DNA sequence information from museum collections of octocoral specimens (Phylum Cnidaria: Class Anthozoa). Bull. Biol. Soc. Wash. 10, 119–129. Berntson, E.A., France, S.C., Mullineaux, L.S., 1999. Phylogenetic relationships within the class Anthozoa (Phylum Cnidaria) based on nuclear 18S rDNA sequences. Mol. Phylogenet. Evol. 13, 417–433. Berntson, E.A., Bayer, F.M., McArthur, A.G., France, S.C., 2001. Phylogenetic relationships within the Octocorallia (Cnidaria: Anthozoa) based on nuclear 18S rRNA sequences. Mar. Biol. 138, 235–246. Broch, H., Horridge, A., 1957. A new species of Solenopodium (Stolonifera: Octocorallia) from the Red Sea. Proc. Zool. Soc. Lond. 128, 149–160. Buckley, T.R., 2002. Model misspeciWcation and probabilistic tests of topology: Evidence from empirical data sets. Syst. Biol. 51, 509–523. Buhl-Mortensen, L., Mortensen, P.B., 2004a. Crustaceans associated with the deep-water gorgonian corals Paragorgia arborea (L., 1758) and Primnoa resedaeformis (Gunn., 1763). J. Nat. Hist. 38, 1233–1247. Buhl-Mortensen, L., Mortensen, P.B., 2004b. Symbiosis in deep-water corals. Symbiosis 37, 33–61. Cairns, S.D., Bayer, F.M., 2005. A review of the genus Primnoa (Octocorallia: Gorgonacea: Primnoidae), with the description of two new species. Bull. Mar. Sci. 77, 225–256. Collins, A.G., Schuchert, P., Marques, A.C., Jankowski, T., Medina, M., Schierwater, B., 2006. Medusozoan phylogeny and character evolution clariWed by new large and small subunit rDNA data and an assessment of the utility of phylogenetic mixture models. Syst. Biol. 55, 97–115. Culligan, K.M., Meyer-Gauen, G., Lyons-Weller, J., Hays, J.B., 2000. Evolutionary origin, diversiWcation and specialization of eukaryotic MutS homolog mismatch repair proteins. Nucleic. Acids Res. 28, 463–471. Dai, C.-F., 1988. Coral communities of southern Taiwan. Proceedings of the 6th International Coral Reef Symposium 2, 647–652. Dinesen, Z.D., 1983. Patterns in the distribution of soft corals across the central Great Barrier Reef. Coral Reefs 1, 229–236. Fabricius, K.E., 1997. Soft coral abundance on the central Great Barrier Reef: eVects of Acanthaster planci, space availability, and aspects of the environment. Coral Reefs 16, 159–167. Fabricius, K., Alderslade, P., 2001. Soft Corals and Sea Fans: A Comprehensive Guide to the Tropical Shallow-Water Genera of the CentralWest PaciWc, the Indian Ocean and the Red Sea. Australian Institute of Marine Science, Townsville. France, S.C., Hoover, L.L., 2001. Analysis of variation in mitochondrial DNA sequences (ND3, ND4L, MSH) among Octocorallia ( D Alcyonaria) (Cnidaria: Anthozoa). Bull. Biol. Soc. Wash. 10, 110– 118. France, S.C., Hoover, L.L., 2002. DNA sequences of the mitochondrial COI gene have low levels of divergence among deep-sea octocorals (Cnidaria: Anthozoa). Hydrobiol. 471, 149–155. France, S.C., Rosel, P.E., Agenbroad, J.E., Mullineaux, L.S., Kocher, T.D., 1996. DNA sequence variation of mitochondrial large-subunit rRNA provides support for a two-subclass organization of the Anthozoa (Cnidaria). Mol. Mar. Biol. Biotech. 5, 15–28. GrasshoV, M., 1999. The shallow water gorgonians of New Caledonia and adjacent islands (Coelenterata: Octocorallia). Senkenberg. Biol. 78, 1– 121. Heifetz, J., 2002. Coral in Alaska: Distribution, abundance, and species associations. Hydrobiol. 471, 19–28. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17, 754–755. C.S. McFadden et al. / Molecular Phylogenetics and Evolution 41 (2006) 513–527 Hyman, L.H., 1940. The Invertebrates, vol. 1. Protozoa through Ctenophora. McGraw-Hill, New York. Krieger, K.J., Wing, B., 2002. Megafauna associations with deepwater corals (Primnoa spp.) in the Gulf of Alaska. Hydrobiol. 471, 83–90. Lepard, A., 2003. Analysis of variation in the mitochondrial encoded msh1 in the genus Leptogorgia (Cnidaria: Octocorallia) and implications for population and systematic studies. M.Sc. University of Charleston, S.C, Charleston, S.C.. McFadden, C.S., Donahue, R., Hadland, B.K., Weston, R., 2001. A molecular phylogenetic analysis of reproductive trait evolution in the soft coral genus Alcyonium. Evolution 55, 54–67. McFadden, C.S., Tullis, I.D., Hutchinson, M.B., Winner, K., Sohm, J.A., 2004. Variation in coding (NADH dehydrogenase subunits 2, 3, and 6) and noncoding intergenic spacer regions of the mitochondrial genome in Octocorallia (Cnidaria: Anthozoa). Mar. Biotech. 6, 516–526. Page, R.D.M., Holmes, E.C., 1998. Molecular Evolution. Blackwell Science Ltd., Malden, MA. Pont-Kingdon, G., Beagley, C.T., Okimoto, R., Wolstenholme, D.R., 1994. Mitochondrial DNA of the sea anemone Metridium senile (Cnidaria): prokaryote-like genes for tRNAf¡met and small-subunit ribosomal RNA, and standard genetic code speciWcities for AGR and ATA codons. J. Mol. Evol. 39, 387–399. Pont-Kingdon, G., Okada, N.A., Macfarlane, J.L., Beagley, C.T., Wolstenholme, D.R., Cavalier-Smith, T., Clark-Walker, G.D., 1995. A coral mitochondrial MutS gene. Nature 375, 109–111. Pont-Kingdon, G., Okada, N.A., Macfarlane, J.L., Beagley, C.T., WatkinsSims, C.D., Cavalier-Smith, T., Clark-Walker, G.D., Wolstenholme, D.R., 1998. Mitochondrial DNA of the coral Sarcophyton glaucum contains a gene for a homologue of bacterial MutS: a possible case of gene transfer from the nucleus to the mitochondrion. J. Mol. Evol. 46, 419–431. Posada, D., Crandall, K.A., 1998. Modeltest: Testing the model of DNA substitution. Bioinformatics 14, 817–818. Riegl, B., Schleyer, M.H., Cook, P.J., Branch, G.M., 1995. Structure of Africa’s southernmost coral communities. Bull. Mar. Sci. 56, 676–691. Sánchez, J.A., 2005. Systematics of the bubblegum corals (Paragorgiidae: Octocorallia: Cnidaria) with description of new species from New Zealand and the Eastern PaciWc. Zootaxa 1014, 1–72. Sánchez, J.A., Zea, S., Diaz, J.M., 1997. Gorgonian communities of two contrasting environments from oceanic Caribbean atolls. Bull. Mar. Sci. 61, 61–72. Sánchez, J.A., Diaz, J.M., Zea, S., 1998. Octocoral and black coral distribution patterns on the barrier reef-complex of Providencia Island, Southwestern Caribbean. Carib. J. Sci. 34, 250–264. Sánchez, J.A., Lasker, H.R., Taylor, D.J., 2003a. Phylogenetic analyses among octocorals (Cnidaria): Mitochondrial and nuclear DNA sequences (lsu-rRNA, 16S and ssu-rRNA, 18S) support two conver- 527 gent clades of branching gorgonians. Mol. Phylogenet. Evol. 29, 31–42. Sánchez, J.A., McFadden, C.S., France, S.C., Lasker, H.R., 2003b. Molecular phylogenetic analyses of shallow-water Caribbean octocorals. Mar. Biol. 142, 975–987. Schmidt, H.A., Strimmer, K., Vingron, M., von Haeseler, A., 2002. TREEPUZZLE: Maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18, 502–504. Shimodaira, H., 2002. An approximately unbiased test of phylogenetic tree selection. Syst. Biol. 51, 492–508. Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16, 1114–1116. Shimodaira, H., Hasegawa, M., 2001. CONSEL: for assessing the conWdence of phylogenetic tree selection. Bioinformatics 17, 1246–1247. SwoVord, D.L., 2002. PAUP*: Phylogenetic Analysis Using Parsimony (* and other methods), Version 4. Sinauer Associates, Sunderland, MA. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The ClustalX windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nuclic. Acids Res. 24, 4876–4882. Tursch, B., Tursch, A., 1982. The soft coral community on a sheltered reef quadrat at Laing Island (Papua New Guinea). Mar. Biol. 68, 321–332. Tyler, P., 2003. Ecosystems of the Deep Oceans. Elsevier, New York. Veron, J.E.N., Odorico, D.M., Chen, C.A., Miller, D.J., 1996. Reassessing evolutionary relationships of scleractinian corals. Coral Reefs 15, 1–9. Watling, L., Auster, P.J., 2005. Distribution of deep-water Alcyonacea oV the Northeast Coast of the United States. In: Freiwald, A., Roberts, J.M. (Eds.), Cold-Water Corals and Ecosystems. Springer, Berlin, Heidelberg, pp. 279–296. Williams, G.C., 1994. Biotic diversity, biogeography, and phylogeny of pennatulacean octocorals associated with coral reefs in the IndoPaciWc. Proceedings of the 7th International Coral Reef Symposium Guam 1992 (2), 729–735. Williams, G.C., 1997. Preliminary assessment of the phylogeny of Pennatulacea (Anthozoa: Octocorallia), with a reevaluation of Ediacaran frond-like fossils, and a synopsis of the history of evolutionary thought regarding sea pens. Proceedings of the 6th International Conference Coelenterate Biology 1995, 497–509. Williams, G.C., Cairns, S.D., 2005. Systematic list of valid octocoral genera. <http://www.calacademy.org/research/izg/OCTOCLASS.htm/>. Wirshing, H.H., Messing, C.G., Douady, C.J., Reed, J., Stanhope, M.J., Shivji, M.S., 2005. Molecular evidence for multiple lineages in the gorgonian family Plexauridae (Anthozoa: Octocorallia). Mar. Biol. 147, 497–508. Won, J.H., Rho, B.J., Song, J.I., 2001. A phylogenetic study of the Anthozoa (phylum Cnidaria) based on morphological and molecular characters. Coral Reefs 20, 39–50.