Are the sexual, somatic and genetic characters enough to solve nomenclatural problems in lumbricid taxonomy?

Traditionally, phylogenetic relationships within Lumbricidae have been investigated by using morphological characters which show high variability between individuals. Furthermore, the lack of agreement in their ranking as diagnostic characters for taxonomic and phylogenetic purposes has led to situations in which the same species receives different names in different parts of the world or may be included in different genera depending on the classification system proposed. Although the recent use of molecular tools in metazoan phylogeny has questioned our traditional understanding of animal classification, these techniques have not been yet fully exploited in earthworm evolutionary biology. Here we construct molecular phylogenies to contrast them with the traditional morphological classification within the Lumbricidae. We have put special emphasis on clarifying the phylogenetic relationships of those species commonly found in the European Atlantic area, trying to highlight some of the nomenclatural problems associated with certain species and to provide a better understanding of this group for earthworm ecologists. We obtained DNA sequences for two fragments of the mitochondrial COI and 16S genes for thirty-six earthworm taxa belonging to the family Lumbricidae and seven species belonging to six outgroup families according to Sims and Gerard (1999)'s terminology (Criodrilidae, Glossoscolecidae, Megascolecidae, Acanthodrilidae, Eudrilidae and Hormogastridae). We interpreted the results in the light of the available information on genital, somatic and genetic characters published over the last 100 years and performed a detailed anatomical study of certain species for which there are contradictory descriptions. In agreement with previous results, our study concludes that the 16S rDNA and COI sequence fragments have a limited discriminatory value above the genus level. In general, the relationships suggested by 16S were more in agreement with morphology than COI but the two genes proved to be useful at the shallowest taxonomic levels. Accordingly, both Lumbricus and Octolasion genera were shown to form distinctive clades despite the distinct geographical distribution of some of their members (e.g. Lumbricus friendi versus Lumbricus terrestris). In contrast, in the case of the heterogeneous Dendrobaena, Aporrectodea and the so-called catch-all genus Allolobophora, the phylogenetic resolution was very limited. In addition, both markers seem to be suitable in revealing ecological similarities within species complexes. This was particularly true in the case of the Octolasion species and those included in the Ap. caliginosa complex, which suggest that occupying different niches in the soil may act as an isolating mechanism of their populations and could lead to speciation. Importantly, our results provide good support for removing Levinsen's L. eiseni from the Lumbricus genus, although its taxonomic position is still problematic and further studies are required, using slower markers and by including a greater number of the forms of each genus in the analyses. Similarly, the anatomical, genital and genetic characters investigated here do not allow the full clarification of the taxonomical status of Dendrobaena/Eisenia hortensis and D./E. veneta; however, in order to maintain the nomenclature stability, we propose to retain them in Dendrobaena until further analyses are carried out on test type specimens (e.g. paratypes or topotypic material). Furthermore, more research is also needed to solve the great heterogeneity of the genus Dendrobaena. Our review also highlights the need for more collaborative work between alpha taxonomists and molecular phylogeneticists to make any meaningful progress in earthworm classification, so that the material included in the phylogenetic trees is properly identified and the results interpreted adequately. Furthermore, in order to avoid further taxonomical ambiguities, soil ecologists are advised to provide the most accurate identification of the species used in their studies or at least a good description of the specimens used together with the geographical location and the soil characteristics where the material was collected. Preserving some material for further taxonomic and genetic reference is also advised for amending possible mistakes. We finally conclude that anatomical or genetic information alone cannot define any lumbricid genus and that gathering morphological, biological, physiological, ecological and genetic evidence together provides the best tool for solving the current unequal chaos of lumbricid taxonomy and for full elucidation of phylogenetic relationships within the Lumbricidae. (C) 2009 Elsevier Ltd. All rights reserved.