This study is the most extensive molecular study of the gastropod families Trochidae and Calliostomatidae published to date, in terms of both numbers of taxa and of gene sequences. As a result of Bayesian phylogenetic analyses of molecular sequence data from one nuclear gene and three mitochondrial genes, we propose dramatic changes to Trochidae family systematics, present the first molecular phylogeny for Calliostomatidae and include the first published sequence data for the enigmatic subfamily Thysanodontinae. Our phylogeny demonstrates that within the family Trochidae there is strong support for three subfamilies new to traditional classifications: Alcyninae subfam. nov., Fossarininae and Chrysostomatinae subfam. nov. As proposed, Alcyninae consists only of the nominotypical genus Alcyna, which is sister to all other trochids. The subfamily Fossarininae, as defined here, includes Fossarina, Broderipia, Synaptocochlea and "Roya"eximia and probably also Clydonochilus and Minopa. The subfamily Chrysostomatinae comprises the genera Chrysostoma and Chlorodiloma. Additional molecular support is also obtained for recently redefined Trochinae, Monodontinae, and Cantharidinae and for the traditionally recognised subfamilies Umboniinae and Stomatellinae. The subfamily Lirulariinae is not supported by the molecular data, but rather is incorporated into Umboniinae. We also demonstrate that the current concept of the subfamily Margaritinae (previously a trochid subfamily, but recently and provisionally assigned to Turbinidae) is not monophyletic. We provide preliminary evidence that whereas Margarella rosea (previously a member of Margaritinae) belongs in the trochid subfamily Cantharidinae, its presumptive congener M. antarctica is not a trochid, but instead clusters with the thysanodontine genus Carinastele. Based on the phylogenetic placement of C. kristelleae, we agree with previous proposals based on morphological data that Thysanodontinae are more closely related to Calliostomatidae than Trochidae. Both Calliostoma and Carinastele are carnivorous and if a sister relationship can be confirmed between Carinastele and Margarella antarctica it might mean that carnivory evolved twice in Trochoidea. The direction of dietary changes was not investigated in this study, but mapping diet onto the phylogeny suggests that true herbivory is predominantly a derived character. The new classification system also means that five trochid subfamilies are predominantly associated with hard substrata, one with soft substrata (Umboniinae) and two with algae and seagrass (Alcyninae and Cantharidinae). There has been a shift back to hard substrata in one umboniine clade. Two of three clades within Calliostomatidae were predominantly associated with hard substrata, but one Japanese clade is associated with sand. The finding of three new, unidentified species from very deep water means that Trochidae, like Calliostomatidae, now includes species found at bathyal depths. More deep-water species may be found as increased sampling leads to the discovery of new species. Crown Copyright 2009. Published by Elsevier Inc. All rights reserved.
