We assessed whether 6-year-old applications of phosphorus (P) fertilizer to a previously unfertilized field site resulted in the selection of populations of vesicular-arbuscular (VA) mycorrhizal fungi that were more ‘tolerant’ of added P than the original VA mycorrhizal population at the site. In 1977, four rates of P (0, 49, 97 and 200 kg P/ha) were applied to a field site at Mt Barker, Western Australia. In 1983, either no P or a rate of P sufficient for maximum plant growth (352 kg P/ha) was applied to each of these plots and the formation of spores and colonization of roots by VA mycorrhizal fungi were examined in the following 3 years. Residual P from fertilizers applied in 1977 increased both the percentage of root length colonized and the length of root colonized by VA mycorrhizal fungi in 1983, 1984 and 1985. Colonization by ‘medium hyphae’ Glomus spp., Acaulospora laevis Gerd. and Trappe and fine endophyte increased in response to the 1977 applications of P. By contrast, colonization by Scutellospora calospora (Nicol. and Gerd.) Walker and Sanders decreased with the 1977 applications of P. Application of an adequate rate of P to the field plots in 1983 generally decreased the development of VA mycorrhizal infection in plots, to a greater extent where larger rates of P had previously been applied in 1977. We attributed this effect to higher initial levels of mycorrhizal colonization in the plots which received larger rates of P in 1977. The 1977 applications of P are unlikely to have resulted in the selection of strains of VA mycorrhizal fungi that are less tolerant of added P than the strains present in the unfertilized plots. There was a common relationship between VA mycorrhizal colonization and the residual value of the P applications which provided indirect evidence that there was no change in the P-tolerance of the indigenous VA mycorrhizal population in response to P applied in 1977. Interpretation of the effects of the 1977 and 1983 applications of P on VA mycorrhizal colonization was compounded by the effects these applications of P had on the botanical composition of the pasture and also on the inoculum potential in each plot. Spores of A. laevis and S. calospora were recovered from each field plot. The number of spores of A. laevis increased in response to P applied in 1977 and generally decreased in response to P applied in 1983. These effects could be directly related to the effects the 1977 and 1983 applications of P had on the length of root colonized by A. laevis in the preceding growing season.