Contributions to Zoology, 84 (3) – 2015Lucio Bonato; Alessandro Minelli; Leandro Drago; Luis Alberto Pereira: The phylogenetic position of Dinogeophilus and a new evolutionary framework for the smallest epimorphic centipedes (Chilopoda: Epimorpha)

To refer to this article use this url:

Molecular evidence

After aligning the sequences of the 16S, 18S and 28S rRNA genes of Dinogeophilus with those of 21 species of Geophilidae and 7 species of Schendylidae (Appendix 3), we obtained a concatenated sequence of 5913 positions. The average genetic distance between species of Geophilidae was found similar to the average distance between species of Schendylidae (Table 2; MW: p = 0.70 for both p-distances and K2P distances), and the average distance between species in different families was found significantly higher than between species in the same family (MW: p = 0.00 for both Geophilidae and Schendylidae, and for both distances). The sequence of Dinogeophilus was found more similar on average to those of Schendylidae than to those of Geophilidae, even though without statistical significance (MW: p = 0.16 for p-distances, p = 0.18 for K2P distances). In both NJ and ME analyses (Fig. 3A), Dinogeophilus clustered together with all Schendylidae, and to the exclusion of all Geophilidae, with high statistical support. This came out from alternative analyses performed on p-distances and K2P distances, the only differences regarding the relative positions of some species within Geophilidae and within Schendylidae.


Table 2. Pairwise genetic distances between species, including Dinogeophilus oligopodus, 21 species of Geophilidae and 7 species of Schendylidae (Appendix 3). Averages are given, together with the range of variation in squared brackets.

For the ML phylogenetic analysis of the concatenated sequences, the GTR+G+I substitution model (selected as the best-fit model under AIC) and the K2P+G+I model (selected under BIC) produced two fully consistent trees (Fig. 3B). Dinogeophilus was found well nested within the Schendylidae, and the entire group of Schendylidae including Dinogeophilus was strongly supported. The species of Schendylidae included in the analysis were found representatives of four moderately to strongly supported clades: (i) Plesioschendyla confossa (New Caledonia) (ii) species of Pectiniunguis and Schendylops (mainly Neotropical schendylines), (iii) species of Hydroschendyla and Schendyla (mainly Palearctic schendylines), and (iv) species of Ballophilus and Ityphilus (ballophilines). Dinogeophilus was found to belong to clade ii, together with schendylines from the Neotropical region.

From the MP of the concatenated sequences, we obtained four equally most parsimonious trees, from 1352 informative positions, with consistency index 0.40 and retention index 0.51. The MP trees were 821 step long and their strict consensus (Fig. 3C) was consistent with the ML tree (Fig. 3B) in recovering a monophyletic Schendylidae including Dinogeophilus, but with a more ambiguous position of the latter either closer to the Neotropical schendylines or the ballophilines.


Fig. 3. Similarity analysis and phylogenetic analyses of the concatenated sequences of 16S, 18S and 28S rRNA genes of Dinogeophilus oligopodus and representative species of Geophilidae and Schendylidae. Bootstrap percentages are not shown when <50%. A. NJ and ME tree (bootstrap percentages: above node, using p-distances; below nodes, using K2P distances). B. ML tree (bootstrap percentages: above node, using the GTR+G+I model; below nodes, using the K2P+G+I model). C. MP consensus tree. Species of Schendylidae are in bold.