A pot experiment was designed to evaluate the interactive effects of multiple microbial inoculation treatments and rock phosphate (RP) application on N and P acquisition by alfalfa plants using 15N and 32P isotopes. The microbial inocula consisted of a wild type (WT) Rhizobium meliloti strain, its genetically modified (GM) derivative, which had an enhanced competitiveness, the arbuscular mycorrhizal (AM) fungus Glomus mosseae (Nicol. and Gerd.) Gerd. and Trappe, and a phosphate-solubilizing rhizobacterium (Enterobacter sp.). Inoculated micro-organisms became established in the root tissues and/or in the rhizosphere soil of alfalfa plants (Medicago sativa L.). The GM Rhizobium strain did not interfere with AM formation. Inoculated phosphate-solubilizing rhizobacteria established in the alfalfa rhizosphere, but the level of establishment was lower where the natural population of phosphate-solubilizing bacteria was stimulated by AM inoculation and RP application. The stimulation of these indigenous bacteria was also greater in the rhizosphere of alfalfa nodulated by the GM Rhizobium. Improvements in N and P accumulation in alfalfa corroborate beneficial effects of the improved GM Rhizobium on AM performance, in RP-amended plants. Inoculation with Enterobacter did not improve the AM effect on N or P accumulation in the RP-added soil, but it did in the non RP-amended controls. Measurements of the 15N[ratio ]14N ratio in plant shoots indicated enhanced N2 fixation rates in Rhizobium-inoculated AM-plants, over that achieved by the same Rhizobium strain in non-mycorrhizal plants. Regardless of the Rhizobium strain and of whether or not RP was added, AM-inoculated plants showed a lower specific activity (32P[ratio ]31P) than did their comparable non-mycorrhizal controls, suggesting that the plant was using otherwise unavailable P sources. The phosphate-solubilizing, AM-associated, microbiota could in fact release phosphate ions, either from the added RP or from the indigenous ‘less-available’ phosphate. Deficiency in Ca concentration in soil solution in the neutral test soil might benefit P solubilization. The proportion of plant P derived either from the labelled soil P (labile P pool) or from RP was similar for AM inoculated and non-mycorrhizal controls (without Enterobacter inoculation) for each Rhizobium strain, but the total P uptake, regardless of the P source, was far higher in AM-plants. Enterobacter inoculation seems to improve the use of RP in the rhizosphere of non-mycorrhizal plants inoculated with the WT Rhizobium.