Do phytogeographic patterns reveal biomes or biotic regions?

We present the largest comparative biogeographical analysis that has complete coverage of Australia's geography (20 phytogeographical subregions), using the most complete published molecular phylogenies to date of large Australian plant clades (Acacia, Banksia and the eucalypts). Two distinct sets of areas within the Australian flora were recovered, using distributional data from the Australasian Virtual Herbarium (AVH) and the Atlas of Living Australia (ALA): younger Temperate, Eremaean and Monsoonal biomes, and older southwest + west, southeast and northern historical biogeographical regions. The analyses showed that by partitioning the data into two sets, using either a Majority or a Frequency method to select taxon distributions, two equally valid results were found. The dataset that used a Frequency method discovered general area cladograms that resolved patterns of the Australian biomes, whereas if widespread taxa (Majority method, with >50% of occurrences outside a single subregion) were removed the analysis then recovered historical biogeographical regions. The study highlights the need for caution when processing taxon distributions prior to analysis as, in the case of the history of Australian phytogeography, the validity of both biomes and historical areas have been called into question.