SABELLARIIDAE

Capa, María, Hutchings, Pat & Peart, Rachael, 2012, Systematic revision of Sabellariidae (Polychaeta) and their relationships with other polychaetes using morphological and DNA sequence data, Zoological Journal of the Linnean Society 164 (2), pp. 245-284 : 255-256

ANALYSES WITHIN SABELLARIIDAE

Maximum-parsimony analyses of 32 sabellariid species, with all the 12 genera represented by at least the type species, and 31 characters, of which 29 were parsimony-informative, resulted in 249 mostparsimonious trees (TL 73, CI 0.55, RI 0.86, Fig. 2 ). The trees were rooted with Spionidae as the outgroup, according to previous analyses ( Fig. 1D ). The polytomy at the base of the consensus tree reflects the large amount of homoplasy in the dataset for resolving bifurcating branching pattern ( Fig. 2A ), with only few clades being outlined, Bathysabellaria (JK 70), Phalacrostemma (JK 87), Mariansabellaria (JK <50), Tetreres (JK 78), and a clade (JK <50) containing a paraphyletic Lygdamis and Idanthyrsus (JK 52) ( Fig. 2 ). Sabellariinae and Lydaminae are not recovered as monophyletic. A dataset containing only the informative characters for the type species of the 12 genera resulted in a similar topology, indicating that the specific features are not the only ones responsible for the overall homoplasy.

Implied weighting with concavity of k = 3 recovered three most-parsimonious trees (TL 76, CI 0.55, RI 0.86; Fig. 3A ), the relationship within Idanthyrsus being the only difference between topologies. The monophyly of most genera are suggested except for Gesaia and Sabellaria . Two main clades are outlined and they do not fully concur with the sabellariid subfamilies. Clade A, supported by having a long operculum (lobes longer than wide), contains the genera Idanthyrsus , Lygdamis , Tetreres , Mariansabellaria , and Gesaia ; Clade B is supported by the absence of neurochaetae on segment 1 (on both sides of the building organ) and contains Phalacrostemma , Bathysabellaria , Paraidanthyrsus , Gunnarea , Sabellaria , Phragmatopoma , and Neosabellaria ( Fig. 3A ). Within Clade A, a sister-group relationship between Lygdamis and Idanthyrsus is supported by the presence of an oblique distal end of the operculum, compound tentacular filaments, flat blades of outer paleae, and inner paleae arranged in a straight line on the inner margin of the opercular lobes, from dorsum to ventrum. Clade AI containing Gesaia , Mariansabellaria , and Tetreres is the sister-group to Clade AII ( Lygdamis and Idanthyrsus ) supported by the presence of only capillary chaetae in the neuropodia of parathoracic segments. Mariansabellaria and Tetreres are closely related based on the presence of very long palps, a median organ that appears to be absent or inconspicuous, and the absence of thoracic branchiae (in segment 2). The basal group of Clade B is Phalacrostemma , sister-group to all other taxa, and Phragmatopoma and Neosabellaria are recovered as a derived clade and supported by several plesiomorphies. Paraidanthyrsus and Gunnarea , two monotypic genera, are found to be closely related and sharing two plesiomorphic features but both species present several morphological differences, including different paleae, level of fusion of the opercular lobes, length of palps, and presence or absence of nuchal spines ( Fig. 3A ).

Similar analyses with weighting concavities of k = 4–6 recovered one identical most-parsimonious tree (TL 77, CI 55, RI 0.86, Fig. 2C ) where Clade B is also recovered but the basal nodes of Sabellariidae differ from previous hypotheses ( Figs 3B , 4 ). Mariansabellaria and Tetreres form a basal clade, sister to a clade with the rest of the terminals. Within this large clade Gesaia and two clades are recovered, Clade B and another containing Idanthyrsus and Lygdamis (Clade AII). The monophyly of all genera is assessed except for Gesaia , Sabellaria , and Lygdamis ( Figs 3B , 4 ).

Synapomorphies defining each genus are documented in the diagnoses of the taxonomic account below and two particular cases, Gesaia and Sabellaria , are discussed.

Figures

Figure 1. Trees resulting from parsimony analyses of Sabellariidae and previously related taxa (including members of Sabellida, Terebellida and Spionida). A, strict consensus after analyses based on 99 morphological features with jackknife support values. B, first of 25 most-parsimonious trees (TL 177, CI 0.58, RI 0.74) after analyses of morphological data with unambiguous changes marked on the topology. Numbers under nodes indicate jackknife values; black dots: synapomorphies, white dots: homoplastic character states. C, shortest tree (TL 9215, CI 0.54, RI 0.36) resulting from analysis of partial 18S, 28S and EF-1a sequences with jackknife support values. D, strict consensus of two most-parsimonious trees (TL 9434, CI 0.54 RI 0.37) of the combined dataset, with jackknife support values.

Figure 2. Strict consensus of 429 most-parsimonious trees after maximum-parsimony analysis of morphological data of members of Sabellariidae rooted with Spionidae (TL 73, CI 0.55, RI 0.86). Numbers under nodes indicate jackknife support values.

Figure 3. Trees resulting from parsimony analyses of morphological data of members of Sabellariidae and rooted with Spionidae, implementing implied weighting. Unambiguous changes are marked on the topology; black dots: synapomorphies, white dots: homoplastic character states. A, strict consensus of three most-parsimonious trees (constant of concavity k = 3); B, most-parsimonious tree (constant of concavity k = 4–6).

Figure 4. Schematic representation of sabellariid relationships based on maximum-parsimony analyses of the morphological data (constant of concavity k = 4–6) and stylized drawing of opercula from top view, modified from Kirtley (1994). See text for further details.