Systematic Entomology (1998) 23, 387–398 End-products of behaviour versus behavioural characters: a phylogenetic investigation of pupal cocoon construction and form in some North American black flies (Diptera: Simuliidae) ALISON E. STUART and FIONA F. HUNTER Department of Biological Sciences, Brock University, Ontario, Canada Abstract. Cocoon spinning was analysed, using video recording and playback, in eighteen Nearctic black fly species, comparing nine Simulium species, six Eusimulium species, Stegopterna mutata, Cnephia dacotensis and Prosimulium mixtum. Fourteen behavioural characters were revealed that produced twenty-two equally parsimonious trees (CI = 0.93, RI = 0.96). Another tree was constructed on the basis of five characters relating to the cocoon structure (end-product characters). The goal of the study was to determine whether characters relating to behavioural components of black fly cocoon spinning or those based on end-products of the behaviour are superior for revealing phylogenetic relationships. This was accomplished by comparing both data sets to a phylogeny constructed with the use of cytological and morphological characters. If taxa are grouped according to end-products (the cocoons) there are some spurious groupings. The behavioural analysis only required one extra step to duplicate the morphological and cytological tree. In the case of black flies, it is more informative to use characters resulting from the analysis of the cocoon spinning behaviour than cocoon morphology. Introduction As the number of phylogenetic analyses grows, it is becoming clear that a wide range of traits possess phylogenetic information. However, behavioural characteristics are one suite of traits used infrequently in phylogenetic reconstruction (Brooks & McLennan, 1991). Some authors believe that behavioural characters are too labile and therefore not phylogenetically informative (Atz, 1970; Aronson, 1981; Carpenter, 1988). These arguments are becoming dated because of the recent publication of a number of experimental and theoretical analyses of behavioural characters. Many experimental analyses have looked at various aspects of reproductive behaviour in a wide range of groups including, for example, the sticklebacks (McLennan et al., 1988; McLennan, 1993), manakins (Prum, 1990), seabirds (Paterson et al., 1995), Pelecaniformes (Kennedy et al., 1996) and newts (Arntzen & Sparreboom, 1989). There have also been some analyses of non-reproductive behaviours, including web Correspondence: A. E. Stuart c/o Entomology Department, Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario, Canada M5S 2C6. E-mail: astuart@zoo.utoronto.ca © 1998 Blackwell Science Ltd 387 building by spiders (Eberhard, 1982; Coddington, 1986) and nest building by wasps (Wenzel, 1993). Several analyses that combine behavioural characters with other characters have also been performed (see de Queiroz & Wimberger, 1993, for a thorough list). Two theoretical papers of importance in this debate are de Queiroz & Wimberger (1993) and Wenzel (1992). Both give extensive lists of behavioural analyses and conclude that behavioural characters are no more homoplastic than morphological characters and, therefore, should be treated in a similar fashion. Proctor (1996) also addresses the issue of homoplasy and behaviour, acknowledging that different behaviours may evolve at different rates, but she points out that this is also the case with morphological and molecular characters. Some behaviours result in a well defined end-product (e.g. web, nest, cocoon). Therefore, characters can be found either in the morphology of the end-product or from the observation of the behavioural components required to construct the end- product. End-products are easier to collect (as either still photographs or specimens), transport and analyse than even video-taped behavioural components. Although time con- suming, it can be argued that the analysis of the components of behaviour gives a more thorough understanding of the