Abstract

Larvae of 60 genera representing the following families and subfamilies of Scarabaeoidea were studied and analysed phylogenetically: Lucanidae (Aesalinae, Nicaginae, Syndesinae, Lampriminae, Lucaninae), Passalidae (Passalinae, Aulacocyclinae), Trogidae, Pleocomidae, Geotrupidae (Taurocerastinae, Lethrinae, Geotrupinae), Bolboceratidae, Ceratocanthidae, Hybosoridae, Glaphyridae, Scarabaeidae (Aphodiinae, Scarabaeinae, Melolonthinae, Dynastinae, Cetoniinae). Seventy-eight larval morphological characters were employed in the analysis. Our data confirm that Dascillidae are not closely related to Scarabaeoidea. The monophyly of the superfamily is supported by 20 apomorphic character states, 18 of them unique. Monophyly of the following scarabaeoid clades is supported (with the number of larval synapomorphies followed by the bootstrap value in parentheses): Scarabaeoidea without Passalidae (6/67), Passalidae (9/100), Pleocomidae (11/93), Trogidae (8/93), Glaphyridae (10/96), Lucanidae (9/95), Ceratocanthidae + Hybosoridae (5/74), Scarabaeinae (9/98). The family Ceratocanthidae was found to be paraphyletic with respect to Hybosoridae. Monophyly of the family Scarabaeidae is not supported. The resolution of the basal parts of the strict consensus tree is higher when using Dascillidae + Eulichadidae v. Agyrtidae + Helophoridae as an outgroup, but the differences in topology become insignificant after bootstrapping. It is suggested that larval morphology alone is not an adequate tool to address basal relationships of Scarabaeoidea and a total evidence analysis should be performed.