-Six hundred and eightv-two quad rats were located ,.lCross the Western Australian wheatbelt and adjacent regions to cover as much of the geographical, edaphic and geomorphological variation of the terrestrial plant communities as possible. The study area covered 230,000 km' or 70% of the South West Botanical Province, one of the world's 25 biodiversity hotspots. It included all or part of the six biogeographical regions centred on the wheatbelt. The native vegetation in the study area is highly fragmented with 74'1., having been cleared for agriculture. Clearing has not been uniform; more extensive areas of bushland remain along the southern and western margins.A total of 2609 taxa of vascular plants in 103 families was recorded from the quad rats. The frequencv p,lttern followed a lognormal distribution. At a scale of 400 m', there \vas little difference in species richness between woodland, mallee and shrubl,lIld formations but those on duplex soils had lower species richness than those on deep sand, Iaterite and granik soils.Over 60';:, of taxa were recorded in fewer than five quad rats. These uncommon taxa were not randomlv distributed but concentrated at the periphery of the study area, particularly in the west and south. At least eight previously unrecognized taxa were collected for the first time. Further collections of 15 taxa listed as Declared Rare under the Wildlife Conservation Act, and 161 taxa on the Department of C'onservation and Land Management's priority flora list were made.'There was little congruence in phytogeographic patterning between lifeforms or taxonomic levels. This implies that analyses of only some components of the flora, or analyses at higher taxonomic levels will not elucidate the full phytogeographic pattern. Phytogeographic patterning in the quad rat classification, which included the 1022 most common species, was strong Iv related to climatic gradients and edaphic factors.Twenty-six assemblages were defined from the species classification, four were later combined with the total number of assL'mblages analyzed reduced to 23. Of these 20 formed natural groups that were consistent with the known distribution and habitat preferences of the component taxa. Species richness in these assemblages was modeled as a surrogate for composition using climatic and soil parameters, Explanatory power of the models ranged from 43,2 to 84.2%.'rill' present biogeogr,lphical regions and subregions defined for the studv area received little support from the ,malyses of either the quad rat classification or the assemblage patterning, !\ vailable structural vegetation mapping also showed little correlation with composition. Species pattL'rning across the studv area was primarily gradational along a northe,lst-southwL'st rainfall gradiL'nt and a north-south temperature gr,ldient, as well as a widespn'ad diffuse pattern associakd with specific grimite ,md saline associ a tions.