Phylogeny of Kinorhyncha, based on analyses of molecular data, supplemented with information from morphology

The kinorhynch interrelationships were analyzed using molecular sequence data, supplemented with information from morphology. The molecular taxon sampling included 18S rRNA data (18S) from a diverse set of taxa, and 28S rRNA data (28S) from a more restricted taxon sample. The 18S rRNA taxon sample included in total 61 OTU, representing at least 54 distinct species (some OTU may have been conspecific), and covering all kinorhynch genera, except Fissuroderes, Polacanthoderes, Cateria and Neocentrophyes. Included were also representatives of two, yet undescribed genera (description of one genus is currently in review). The 28S rRNA data set included 26 OTU, representing at least 22 distinct species, and covering 15 kinorhynch genera, including one of the undescribed ones. Sequence data from the two molecular loci were analyzed separately and combined (analyses of combined data sets included only taxa for which information from both loci were available), using maximum likelihood (ML) and Bayesian inference (BI). Subsequently, it was attempted to establish the phylogenetic positions of the four genera for which molecular sequence data were unavailable, by character tracing of morphological information from a data matrix that was assigned to a tree topology defined by ML analysis of 18S rRNA. Analyses of 18S rRNA and 28S rRNA data in combination or separately produced very similar tree topologies, and also the analytical approaches, ML or BI, produced very congruent results. All analyses formed two clades with some correspondence to the two commonly recognized orders, Cyclorhagida and Homalorhagida. Differences from traditional classification included the inclusion of Dracoderes into Homalorhagida. Furthermore, the two yet undescribed genera appeared surprisingly to be homalorhagid ingroup taxa, even though they show clear morphological affinities to the otherwise cyclorhagid taxon Cateria. Most analyses supported Campyloderes as the most basal cyclorhagid clade (only BI analysis of 18S rRNA data was unable to recover this position, and placed Campyloderes in a polytomy with Homalorhagida and other cyclorhagids). The analyses supported a monophyletic group consisting of all remaining cyclorhagids. These formed two clades, one including all echinoderid species and the other all taxa (except Campyloderes) with midterminal spine, i.e., species of Centroderes, Condyloderes, Wollunquaderes, Semnoderes, Sphenoderes, Antygomonas, Tubulideres, Triodontoderes, and Zelinkaderes. The other main group included in all analyses species of the homalorhagid genera, of Dracoderes and of the two undescribed genera in all analyses. Internal relationship of the last group differed between the analyses though. ML analysis of 18S, rRNA and ML and BI analyses of the combined loci supported Dracoderes as the most basal clade, whereas BI analyses of 18S rRNA supported a clade with Paracentrophyes and the two undescribed genera as most basal. All analyses with 18S rRNA data included supported Pycnophyes as a paraphyletic group that with nested nesting Kinorhynchus inside it. Only analyses of 28S rRNA alone recovered Pycnophyes and Kinorhynchus as monophyletic. Character tracing of morphological data from a matrix with 33 characters (2 additive, 31 non-additive, representing 86 character states) on a tree topology based on ML analyses of the combined molecular datasets, allowed us to recover the positions of representatives for the four genera from which molecular data were unavailable. Species of Fissuroderes and Polacanthoderes could easily be assigned to the clade with all other echinoderid taxa. With species of Fissuroderes and Polacanthoderes included, the clade is still supported by several morphological characters, inclusive presence of trichoscalid plates, restriction of middorsal spines to segments 4 to 8, and gender determined presence of lateral accessory terminal spines. Species of Cateria possess several character states that are autapomorphic for the genus, but still, characters such as the bifurcated primary scalids and the absence of placids, suggest a close relationship between Cateria and the new genera. As a result, relocation of Cateria from Cyclorhagida to Homalorhagida would have to be considered. Species of Neocentrophyes were difficult to position because the exact topology of basal Homalorhagida appears incongruent across the different analyses. However, the commonly assumed position near Paracentrophyes , and therefore also near Cateria and the two undescribed genera, still seems to be plausible. In summary, analyses of the most comprehensive kinorhynch data set analyzed so far, confirm certain parts of traditional kinorhynch classification, but they also demonstrate that several aspects need to be reconsidered and revised completely