Hypotheses from mitochondrial DNA: congruence and conflict between DNA sequences and morphology in Dolichopodinae systematics (Diptera:Dolichopodidae)

Abstract

The molecular phylogeny of the subfamily Dolichopodinae (Diptera : Dolichopodidae) is reconstructed based on 79 species of 7 dolichopodine genera as ingroup, and 10 non-dolichopodine species from different genera as outgroup. A Bayesian analysis based on a mitochondrial DNA dataset consisting of 1702 characters (COI : 810; 12S : 366; 16S : 526) was carried out. Genital and non-genital morphological characters from a hitherto unpublished data matrix (based on 57 Dolichopodidae species) were used to explain and support the lineages hypothesised by our molecular phylogenetic analysis. The monophyly of the subfamily Dolichopodinae, and of the genera Dolichopus and Gymnopternus, was confirmed. The molecular analysis yielded nine species groups in Dolichopus that were proposed in previous studies using COI and Cyt-b. No evidence was found to support a clade including Dolichopus, Ethiromyia, and Gymnopternus. The genus Hercostomus proved polyphyletic with respect to Poecilobothrus, Sybistroma, and Gymnopternus. The following lineages were represented by strongly supported clades: Hercostomus germanus species group, H. vivax species group, H. nigrilamellatus species group, H. plagiatus species group, H. longiventris species group, H. fulvicaudis species group, and Poecilobothrus, Gymnopternus, Tachytrechus and Sybistroma (including Hercostomus nanus and H. parvilamellatus). Two clades that were previously established on the basis of morphology were confirmed in our phylogenetic analysis: (i) Poecilobothrus and the flower-feeding Hercostomus germanus species group, and (ii) the H. longiventris lineage and Sybistroma. In most cases, the groups identified in the molecular analysis could be supported and explained by morphological characters. Species of the Hercostomus germanus species group, Poecilobothrus, the Hercostomus longiventris species group, and a Sybistroma subclade have a similar microhabitat affinity.