A serological investigation of 48 predominantly Australian species of Acacia subgenus Phyllodineae was undertaken to assess the established classifications of this group. The study confirmed a previous finding (based on less extensive serological data, as well as on non-serological data) that section Botrycephalae is closely related to certain racemose species of section Phyllodineae, e.g. A. podalyriifolia. However, the serological data suggest a number of patterns of relationship which were unexpected and which are at Variance with established classifications. The following are the most significant new findings revealed by this study. We were unable to demonstrate a clear division between uninerved species of section Phyllodineae and the plurinerved species belonging to sections Juliflorae and Plurinerves. Section Phyllodineae appears to comprise two reasonably coherent subgroups (i.e. racemose species and non-racemose species), and there is no sign of a close relationship between these two. Indeed, species of the racemose group are quite distinct from all other groups in the study; this group appears to consist of three subgroups centred on A. ligulata, A. pyrifolia and A. podalyriaolia. The non-racemose species of section Phyllodineae shaw affinities to certain species currently included in sections Plurinerves and Juliflorae. While section Plurinerves appears to be very coherent, doubt is cast on section Juliflorae as a natural group. One subgroup of species within section Juliflorae has its strongest connections with non-racemose species of section Phyllodineae, and the other with species of section Plurinerves. if the relationships as indicated by this serological study are valid, then a major re-appraisal of the classification of subgenus Phyllodineae is needed. Copyright (C) 1996 Elsevier Science Ltd
