Pii: S0020-7519(01)00114-x

Recent studies based on molecular data (18S rDNA and partial 28S rDNA) and morphology did not resolve a terminal polytomy within the Polyopisthocotylea. Here, we have used sequences from the full domain D2 of the 28S rDNA for 24 species (18 new sequences) with three phylogenetic methods, maximum parsimony, neighbour-joining and maximum likelihood, to infer the relationships among the Polyopisthocotylea. The analysis of the domain D2 of the 28S rDNA has been performed on two data sets. The ®rst one, complete, included the Polystomatidae as the outgroup in order to infer general relationships, and the second one, reduced, excluded the Polystomatidae and the polyopisthocotylean parasites of chondrichthyans, but used the Mazocraeidae as the outgroup in order to resolve the relationships between the terminal groups. The topology found, sustained by high bootstrap and decay index value, is: (outgroup (Chimaericolidae (Mazocraeidae (Gastrocotylinea, other Polyopisthocotylea)))). The polyopisthocotylean parasites of chondrichthyans are the sister-group of the polyopisthocotylean parasites of teleosts. In the latter, the Mazocraeidae, essentially parasites of Clupeidae, have a basal position. The polytomy between Gastrocotylinea, Discocotylinea and Microcotylinea is partially resolved in this study for the ®rst time: the Gastrocotylinea are the sister-group of an unresolved group including the Microcotylinea, Discocotylinea and Plectanocotylidae. Inclusion of the Plectanocotylidae in the suborder Mazocraeinea is rejected. Monophyly of the Microcotylinea and Plectanocotylidae is con®rmed, but monophyly of the Discocotylinea is questioned by the exclusion of Diplozoon. q 2001 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved.