Abstract —Gondwanan biogeography, particularly the relationships between southern South America, New Zealand, Australia, New Guinea and New Caledonia, has been much studied.Nothofagusis often used as the “test taxon”, and many papers have been directed at usingNothofagusto explain Gondwanan biogeography. Cladistic biogeographers, working on plant material, have generally failed to find congruence among taxa expected from the southern Pacific disjunctions. New morphological and molecular data on the phytogeny ofNothofagushave re‐opened the issue, and we analysed these data to construct a new hypothesis of the biogeography of the genus. We assembled all plant taxa for which we could find reasonably robust phylogenetic hypotheses, and sought a parsimonious biogeographical pattern common to all. Two analyses, based on different assumptions, produced the same general areacladogram. We use the general area‐cladogram, in conjunction with the fossil record ofNothofagusto construct a historical scenario for the evolution of the genus. This scenario indicates extensive extinction, but also suggests that Australia has a more recent relationship to New Zealand than to southern South America. This is not congruent with the current geological theories, nor with the patterns evident from insect biogeography. We suggest that concordant dispersal is an unlikely explanation for this pattern, and propose that the solution might be found in alternative geological h