It has been shown that the diagnosis of the availability of nitrogen in natural pastures lacks accuracy whenever the proportion of legumes becomes important. Such a diagnosis, when based on the method of the nitrogen nutrition index, overestimates the availability of nitrogen in the environment when legumes are present, since the analysis takes into account the amount of nitrogen of atmospheric origin that is contained in their biomass. Two trials, situated in two different environments, show a linear relationship between the excess in the estimation and the proportion of legumes in the pasture. Thus, if legumes constitute approximately 10 and 30% of the pasture biomass, the nitrogen indices are overestimated by 7 and 21 points respectively. The linear relationship existing between the contribution of the legumes to the total biomass and the overestimation of the indices gives the possibility of correcting these where the proportion of legumes has been quantified. It makes it also possible to have but a single measurement of nitrogen instead of two after the sorting of the two constituents of the biomass.