Aim - Our aims were (1) to compare observed, estimated and predicted patterns of species richness using the Australian native Asteraceae as an example, (2) to identify candidates for hotspots of diversity for the study group, and (3) to examine the distortion of our perception of the spatial distribution of species richness through uneven or misdirected sampling efforts. Location - Australia. Methods - Based on data from Australia's Virtual Herbarium, we calculated and visualized observed species richness, the Chaol estimate of richness, the C index of collecting completeness, and an estimate of richness derived from environmental niche modelling for grid cells at a resolution of 1o. The 20 cells with the highest diversity values were used to define hotspots of diversity. Results - Uneven collecting activity results in misleading diversity patterns for the family Asteraceae. While observed species richness is much higher in central Australia than in other parts of the arid interior, this is an artefact resulting from the area being a hotspot of collecting activity. The mountain ranges of southeastern Australia and Tasmania are candidates for unbiased hotspots of species richness. Main conclusions - Vast areas of the Australian interior are insufficiently sampled on a local scale, although most of them can be expected to be relatively species poor. Some areas in the south-east and south-west of the continent remain undersampled relative to their high species richness. Observed species numbers, estimators and environmental niche-modelling all have their unique advantages and disadvantages for the inference of patterns of diversity.