Field and controlled environment studies were undertaken to determine the seasonal variation in vesicular-arbuscular mycorrhizal (VAM) propagules under different types of vegetation in the mulga (Acacia aneura) shrublands of south-western Queensland and how inoculation with VAM affects the growth and response to phosphorus fertilizers of several grasses common to this region. A most probable number (MPN) technique was used to estimate the number of viable VAM propagules under mulga shrubland and native pasture. There was a pronounced rise in MPN at a native pasture site from June to November, peaking at 1 propagule/g soil, before declining between November and May to non detectable concentrations. In the mulga shrubland site, MPN remained very low (less than 0.2 propagules/g soil) throughout the study. Pasture species in the field were screened for VAM infection in both winter and summer. Only three of the 14 species sampled were infected with VAM in winter (June) after a long drought. In contrast, 10 of the 14 species sampled in December possessed VAM-infected roots. A pot experiment was conducted to examine the influence of VAM inoculation on growth and the response to phosphate fertilizer of several important pasture grasses in pastures derived from mulga shrublands. The treatments comprised a factorial combination of inoculation or non-inoculation with VAM, four grass species (Aristida armata, Cenchms ciliaris cv. USA, Digitaria ammophilla, and Thyridolepis mitchelliana), and four rates of P fertilizer designed to range over very deficient to non-limiting for each species. All species except D. ammophilla produced growth responses to VAM inoculation. Though C. ciliaris and T. mitchelliana responded to VAM inoculation only in soil unamended with P fertilizer, A. amata showed growth responses across all P rates examined, suggesting some factor other than P was limiting this species. The growth response of the grasses to VAM inoculation was poorly correlated with the percentage of root infected with VAM. D. ammophilla had the highest levels of root infection (32%) despite producing no growth response to VAM. In contrast, T. mitchelliana had less than 2% of the root infected with VAM. All species had thin fibrous roots with long (0.35-0.47 mm), frequent root hairs. The differential growth responses to VAM inoculation of A. armata compared with the other grasses may provide a management strategy to control Aristida ingress into pastures established from mulga shrublands.