This study was undertaken in an attempt to assess the involvement of a mycorrhizal fungus in N-transfer from legume to grass in a field-grown intercrop. Maize (Zea mays L., cv Pioneer 3859) and nodulating soybean (Glycine max (L.) Merrill, cv. Maple Arrow) inoculated with Bradyrhizobium, or maize and a non-nodulating soybean (Evans x L66-70) were grown in a clay soil, previously fumigated with methyl bromide. The plants were either fertilized with 0%, 50% or 100% of the recommended P-fertilizer level, as controls, or were inoculated with the endomycorrhizal fungus Glomus intraradices Schenck and Smith. The N-15-dilution method was used to assess N-transfer from soybean to corn. Non-mycorrhizal maize and soybean plants did not respond to P fertilization. However, non-P-fertilized plant dry mass and P concentration increased by over 100% following inoculation with the mycorrhizal fungus. In the presence of the endophyte, K concentrations of maize and soybean were increased, and the concentrations of N and Zn in soybean were decreased. According to the Diagnosis and Recommendation Integrated System (DRIS), inoculation with G. intraradices changed the order of nutrient requirement of maize, by lessening the P deficiency at the expense of the plant N status, and improved the balance of nutrients in both maize and soybean. The N-15-dilution method gave no evidence of N transfer between soybean and maize plants, even in the mycorrhizal plots, where the fungus increased the legume N2-fixation rate by 55%. However, in a four-member association (grass and legume plants, diazotrophic and mycorrhizal endophytes), the reliability of the N-15-dilution method appears questionable.