Phylogenetic relevance of the genital sclerites of Neuropterida (Insecta: Holometabola)

Abstract Segment 9 of male Raphidioptera, comprising tergite, sternite, gonocoxites, gonostyli and gonapophyses, is a benchmark for homologies in the male and female terminalia of the three Neuropterida orders Raphidioptera, Megaloptera and Neuroptera. The segments relating to genitalia are 9, 10 and 11 in males and 7, 8 and 9 in females. Results from holomorphological and recent molecular cladistic analyses of Neuropterida agree in supporting the sister‐group relationships between: (1) the Raphidioptera and the clade Megaloptera + Neuroptera, and (2) the suborder Nevrorthiformia and all other Neuroptera. The main discrepancy between the results of these studies is the nonmonophyly of the suborder Hemerobiiformia in the molecular analysis. The monophyly of the Megaloptera (which has been repeatedly questioned) is further corroborated by a hitherto overlooked ground pattern autapomorphy: the presence of eversible sacs within the complex of the fused gonocoxites 11 in Corydalidae and Sialidae. The recently discovered paired complex of gonocoxites 10 (parameres) in Nipponeurorthus (Nevrorthidae) indicates that the curious apex of sternite 9 of Nevrorthus and Austroneurorthus is the amalgamation of the sclerites of gonocoxites 10 with sternite 9, interpreted as synapomorphic. In the molecular study, the Nevrorthidae, Sisyridae and Osmylidae branch off in consecutive splitting events, a result that is supported by the analysis of male genital sclerites reported here. Extraordinary parallel apomorphies (e.g. excessive enlargement and modification of gonocoxites 10 ending in a thread‐like ‘penisfilum’) in derived representatives of Coniopterygidae, Berothidae, Rhachiberothidae and Mantispidae corroborate the dilarid clade of the morphological analysis and leads us to hypothesize a sister‐group relationship of the Coniopterygidae with the dilarid clade. A re‐interpretation of the tignum of Chrysopidae as gonocoxites 11 means that the structure previously called the gonarcus represents the fused gonocoxites 9. In Hemerobiidae, the corresponding sclerite is consequently also homologized as fused gonocoxites 9. The enlargement of the lateral wings of the gonocoxites in both families is interpreted as a synapomorphy. Excessive enlargement of gonostyli 11 in the Polystoechotid clade and Myrmeleontiformia supports a sister‐group relationship of these two clades. The occurrence of certain serial homologues of female genitalia structures (gonocoxites and gonapophyses), such as the digitiform processus together with the flat appendices in segment 8 of certain Myrmeleontidae, or the wart‐like processus together with the flat circular sclerites in segment 7 of certain Berothidae, as well as the presence of gonocoxites 8 as pseudosternites in certain Nemopteridae and Coniopterygidae, are probably character reversals. The digitiform processus of tergite 9 (pseudogonocoxites) in Rhachiberothidae and Austroberothella (Berothidae) are either independently developed acquisitions with a function in oviposition, or are homologous sclerites, possibly of epipleurite origin.