Arbuscular mycorrhizal (AM) fungi occur in all agricultural soils but it is not easy to assess the contribution they make to plant growth under field conditions. Several approaches have been used to investigate this, including the comparison of plant growth in the presence or absence of naturally occurring AM fungi following soil fumigation or application of fungicides. However, treatments such as these may change soil characteristics other than factors directly involving AM fungi and lead to difficulties in identifying the reason for changes in plant growth. In a glasshouse experiment, we assessed the contribution of indigenous AM fungi to growth of subterranean clover in undisturbed cores of soil from two agricultural field sites (a cropped agricultural field at South Carrabin and a low input pasture at Westdale). We used the approach of estimating the benefit of AM fungi by comparing the curvature coefficients (C) of the Mitscherlich equation for subterranean clover grown in untreated field soil, in field soil into which inoculum of Glomus invermaium was added and in soil fumigated with methyl bromide. It was only possible to estimate the benefit of mycorrhizas using this approach for one soil (Westdale) because it was the only soil for which a Mitscherlich response to the application of a range of P levels was obtained. The mycorrhizal benefit (C of mycorrhizal vs. non-mycorrhizal plants or C of inoculated vs. uninoculated plants) of the indigenous fungi corresponded with a requirement for phosphate by plants that were colonised by AM fungi already present in the soil equivalent to half that required by non-mycorrhizal plants. This benefit was independent of the plant-available P in the soil. There was no additional benefit of inoculation on plant growth other than that due to increased P uptake. Indigenous AM fungi were present in both soils and colonised a high proportion of roots in both soils. There was a higher diversity of morphotypes of mycorrhizal fungi in roots of plants grown in the Westdale soil than in the South Carrabin soil that had a history of high phosphate fertilizer use in the field. Inoculation with G. invermaium did not increase the level of colonisation of roots by mycorrhizal fungi in either soil, but it replaced approximately 20% of the root length colonised by the indigenous fungi in Westdale soil at all levels of applied P. The proportion of colonised root length replaced by G. invermaium in South Carrabin soil varied with the level of application of P to the soil; it was higher at intermediate levels of recently added soil P.