Taxon sampling, character sampling and systematics: how gradist presuppositions created additional ganglia in gastropod euthyneuran taxa

In gastropods, the pentaganglionate condition of the nervous visceral loop has been accepted as a general character and one of the most important synapomorphies of the Euthyneura. The review of published data on 50 generic taxa shows an extreme confusion in both terminology and application of the homology concept to the two parietal ganglia observed in some Euthyneura, in addition to the three observed in most gastropods. A parsimonious re‐interpretation of the data leads to the conclusion that the occurrence of five visceral ganglia is ascertained in only six genus‐group taxa, and therefore cannot be accepted as a general character of Euthyneura. We propose that in phylogenetic analysis, taxonomic sampling should be determined according to the variability of characters and independently of the rank of the taxa taken into account, until every taxon in the data matrix is monomorphic for every character, and until all the taxa in which the characters occur are represented in the data matrix. We propose to use the term ‘domain of definition’ of a character for such a taxonomic sample. Implementing the domain of definition of characters in construction of data matrices would avoid the abuse of generalities leading to circularity in evolutionary interpretations of classification, which in the case of Euthyneura result from an a priori gradist interpretation of the evolution of the nervous system.