This method is based on the small difference in 15N abundance which frequently occurs between N derived from N2-fixation and N derived from other sources. The requirement of the method is that this difference be significant. Whether this requirement is met is site specific and must be empirically established at each site of interest. Advantages and disadvantages of this method are compared with those of more conventional methods. Sources of error, including heterogeneity of 15N abundance of non-atmospheric N sources are considered. Tests of the method, under both greenhouse and field conditions, are described. Estimates based on this method compare favourably with other methods for field evaluation of N2-fixation, provided that the site and the sampling strategy are appropriate for application of the method. Applications of the method in several ecosystems are described.