Problems arise when relationships of the early members of new classes of organisms are being discussed. Our concern about this problem arises from the recent work on Diabolepis, which some workers have regarded as the sister group of the Dipnoi. The main morphological work on this genus by Chang (1995), has set out the main features of the skull of the genus. She has analyzed the relationships without cladistic analysis, but other workers have cladistically analyzed the data. In this paper we criticise many of the observations on morphology discussed by Chang, and contest many of the synapomorphies used by workers who do the cladistic analysis. Broad conclusions concerning the development of sarcopterygian fishes and more detailed results such as the evolution of the dental plates of dipnoans, have been falsely drawn because of this.