Cylindrobulla and Ascobulla in the western Atlantic (Gastropoda, Opisthobranchia, Sacoglossa): Systematic review, description of a new species, and phylogenetic reanalysis

Two often-confused western Atlantic shelled opisthobranchs - Cylindrobulla beauii P. Fischer, 1857, and Ascobulla ulla (Marcus & Marcus, 1970) - are redescribed and compared, via a full literature review; scanning electron microscopy of shells and radulae, and serial histological sectioning of soft anatomy. A third sympatric species, C. gigas, from Florida, the Caribbean and Bermuda, is described as new, and also provides the first larval development data for the genus. The overlapping classificatory histories of the genera Cylindrobulla and Ascobulla are summarized and discussed with reference to the Cephalaspidea, Sacoglossa (= Ascoglossa), Anaspidea, and Diaphanoidea. Two of the most-recent studies (Jensen 1996a, Jensen 1996b; Mikkelsen 1996) have resulted in conflicting conclusions about the relationships and classification of Cylindrobulla. In the author's earlier phylogenetic analysis (Mikkelsen 1996), Cylindrobulla formed a monophyletic group with the shelled Sacoglossa with synapomorphies in digestive, reproductive, pallial, and nervous system characters. Jensen's two-part study (Jensen 1996a, Jensen 1996b) resulted in removal of Cylindrobulla to a new opisthobranch order, Cylindrobullacea. These two divergent cladistically produced classifications were tested by critically re-evaluating Jensen's dataset, especially with regard to Cylindrobulla, and re-running the analysis. Thirteen of Jensen's 52 characters were recoded for Cylindrobulla according to new morphological data and the new species described here. Other coding changes were required to update anatomical data for other taxa, add supplementary characters and taxa, more clearly define the outgroup, and to critically re-evaluate the cladistic criteria behind several characters. The resultant reanalysis produced a shorter tree using Fewer a priori assumptions, but nevertheless preserved the basic tree topology, including three monophyletic sacoglossan clades, presented by Jensen (1996b). Cylindrobulla, however, rejoined the shelled sacoglossans (Oxynoacea), supported by two synapomorphies. This analysis therefore reconfirmed Cylindrobulla's proper placement in the Sacoglossa, rather than in the traditional Cephalaspidea or segregated in its own monogeneric opisthobranch order. (C) 1998 The Norwegian Academy of Science and Letters.