Numerous studies have been performed to study the effect of N fertilization on cereal-legume intercrops, and their results are sometimes conflicting. Our objective was to do a meta-analysis on cereal-legume intercrops testing the effects of N fertilization on land equivalent ration (LER; partial and total LER), yield ratio, and proportion of legume in the mixture of crop grains. This analysis was based on 17 published studies reporting the results of experiments performed in 15 countries on six species of cereals and 10 species of legumes. Experiments were generally based on replacement (50-50, i.e., in the intercrop, each species is sown at half the sowing rate used for the sole crop) or full substitutive (100-100, i.e., the sowing rate for each crop in the intercrop is identical to that for sole crops) designs. Nitrogen fertilization rates ranged from 0 to 180 kg N ha -1. The effect of N fertilization and its inter-study variability were analyzed with mixed-effect statistical models, including study as a random effect. Results showed that N fertilization had non-significant effects on average LER and average yield ratio but that the inter-study variability of these effects was large. Nitrogen fertilization was found to significantly decrease the grain proportion of the legume in the mixture and the partial LER of the legume in studies based on C 3 cereal intercrops. The database used for the meta-analysis is freely available (http://www6.versailles-grignon.inra.fr/agronomie/Meta-analysis-in-agron omy/Datasets/Dataset-Intercrop).
