On wings of lace: phylogeny and Bayesian divergence time estimates of Neuropterida (Insecta) based on morphological and molecular data

Neuropterida comprise the holometabolan orders Neuroptera (lacewings, antlions and relatives), Megaloptera (alderflies, dobsonflies) and Raphidioptera (snakeflies) as a monophyletic group sister to Coleoptera (beetles). The higher‐level phylogenetic relationships among these groups, as well as the family‐level hierarchy of Neuroptera, have to date proved difficult to reconstruct. We used morphological data and multi‐locus DNA sequence data to infer Neuropterida relationships. Nucleotide sequences were obtained for fragments of two nuclear genes (CAD, 18S rDNA) and two mitochondrial genes (COI, 16S rDNA) for 69 exemplars representing all recently recognized families of Neuropterida as well as outgroup exemplars from Coleoptera. The joint posterior probability of phylogeny and divergence times was estimated using a Bayesian relaxed‐clock inference method to establish a temporal sequence of cladogenesis for the group over geological time. Megaloptera were found to be paraphyletic with respect to the rest of Neuropterida, calling into question the validity of the ordinal status for Megaloptera as presently defined. Ordinal relationships were weakly supported, and monophyly of Megaloptera was not recovered in any total‐evidence analysis; Corydalidae were frequently recovered as sister to Raphidioptera. Only in relaxed‐clock inferences were Raphidioptera and a paraphyletic Megaloptera recovered with strong support as a monophyletic group sister to Neuroptera. A monophyletic Neuroptera diverged from a common Raphidioptera + ‘Megaloptera’ ancestor during the Late Carboniferous. Contrary to some previous hypotheses, Coniopterygidae, not Nevrorthidae, were recovered as sister to the rest of Neuroptera, with Nevrorthidae recovered with Osmylidae and Sisyridae. The monophyly of the universally recognized Myrmeleontiformia was confirmed, with an origin in the mid‐Triassic, but a monophyletic Hemerobiiformia was not recovered in any analysis. Dilaridae were not closely related to the clade comprising Mantispidae and Berothidae, and diverged earlier than proposed previously. The phylogenetic status and taxonomic composition of Polystoechotidae and Ithonidae are in need of re‐evaluation, as Oliarces Carpenter (presently Ithonidae) was placed well within the present circumscription of Polystoechotidae.