Progress in understanding the higher classification of butterflies has not kept pace with increase in the number of described species. Important points of uncertainty or contention include, apart from ranking problems, monophyly of Papilionoidea plus Hesperioidea, their relationship with other Lepidoptera in general and the Hedyloidea in particular, the question of the sister group of the Pieridae (either Papilionidae, or Lycaenidae + Nymphalidae), and the division of families into subfamilies. Traditional groupings are discussed and compared with the results of a cladistic analysis using 103 characters and 74 species (59 butterflies and 15 moths). The cladistic analysis supports a number of currently held views about butterfly classification, such as monophyly of five major family groupings (Hesperiidae, Papilionidae, Pieridae, Lycaenidae and Nymphalidae) and suggests sister group relationships between Papilionoidea and Hesperioidea, and Pieridae and (Lycaenidae + Nymphalidae). Most traditional subfamilies, however, are not supported on the basis of the data set used but the Riodininae, which always appeared as a monophyletic, subordinate group within the Lycaenidae, are a notable exception. Further, the analysis suggests that, contrary to traditional ideas, the Parnassiinae, not Baroniinae, are sister to the remainder of the Papilionidae, Pseudopontiinae are internal to (Pierinae + Coliadinae), Dismorphiinae are sister to all other Pieridae, and that Liptena, Poritia and Miletus represent the closest relatives of the Riodininae. The data set is not well suited for an assessment of the position of the butterflies amongst other Lepidoptera. Nevertheless, of the moths used, Macrosoma (Hedylidae, Hedyloidea) and Urania (Uraniidae, Geometroidea) appear to be the closest relatives of the butterflies. With regard to the higher classification of the butterflies many problems thus remain, and several ways to tackle these are discussed. The need for some form of international co-operation between fieldworkers, comparative morphologists and molecular systematists is stressed.