Aim: Biodiversity studies typically use species, or more recently phylogenetic diversity (PD), as their analysis unit and produce a single map of observed diversity. However, observed biodiversity is not necessarily an indicator of significant biodiversity and therefore should not be used alone. By applying a small number of additional metrics to PD, with associated statistical tests, we can determine whether more or less of the phylogeny occurs in an area, whether branch lengths in an area are longer or shorter, and whether more long or short-branched endemism occurs in an area, than expected under a null model. Location: Australian continent. Methods: We used a phylogeny sampling 90% of Australia's angiosperm genera, and 3.4 million georeferenced plant specimens downloaded from Australia's Virtual Herbarium (AVH), to calculate PD, relative phylogenetic diversity (RPD) and relative phylogenetic endemism (RPE). Categorical analysis of neo- and palaeo-endemism (CANAPE) and randomization tests were performed to determine statistical significance. Results: We identify several combinations of significant PD and endemism across the continent that are not seen using observed diversity patterns alone. Joint interpretation of these combinations complements the previous interpretations of Australia's plant evolutionary history. Of conservation concern, only 42% of the significant endemism cells found here overlap with existing nature reserves. Main conclusions: These spatial phylogenetic methods are feasible to apply to a whole flora at the continental scale. Observed richness or PD is inadequate to fully understand the patterns of biodiversity. The combination of statistical tests applied here can be used to better explain biodiversity patterns and the evolutionary and ecological processes that have created them. The spatial phylogenetic methods used in this paper can be also be used to identify conservation priorities at any geographical scale or taxonomic level.