Information on the amount and proportion of alfalfa (Medicago sativa L.) N derived from symbiotic N2 fixation is required in order to exploit this source of N in alfalfa-weeping lovegrass (Eragrostis curvula Schrad Nees) pastures in semiarid Argentina. A field experiment was conducted to determine N2 fixation by a legume grown alone or in two different associations with weeping lovegrass, using pure or associated lovegrass as the reference crop. On average, alfalfa derived over 80% of its N from fixation during the seeding year, equivalent to about 140 kg N ha-1. Compared to N2 fixation in the pure alfalfa sward, the percentage of N derived from atmospheric N2 in alfalfa increased significantly in the associated pastures and even more so by increasing the seeding ratio of both crops sown in the same row. The atom % 15N excess in associated grass was sometimes slightly lower than that in pure lovegrass, suggesting any possible release of N from alfalfa and subsequent uptake by grass to have been small. The highest proportion of grass N that could have derived from the legume was estimated to be 23% in the second harvest, equivalent to only 6 kg N ha-1. Differences in atom % 15N excess of pure and associated grass reference plants did not significantly affect the estimated percentage of N derived from atmospheric N2 in alfalfa.