Biological N2-fixation by three tropical forage legumes and its transfer to Brachiaria humidicola in mixed swards

A study was conducted (a) to determine the biological N2-fixation (BNF) by three tropical pasture legumes (Centrosema pubescens, Stylosanthes hamata and Pueraria phaseoloides) grown in monoculture and associated with the grass Brachiaria humidicola; and (b) to assess the potential transfer of fixed N from the legume to the companion grass. To calculate the proportion of N in the legume derived from air, the 15N isotope dilution technique was used. The fertilizer (ammonium sulfate 10% enriched with 15N) was applied in seven split doses during the growing season. Aboveground biomass of the grass and the legumes in monocrop and in mixture were determined twice a year, after 3 months of regrowth, and samples were analyzed for total N and 15N. Over the two evaluation periods, the association B. humidicola/C. pubescens was the most stable, with a proportion of 20–30% of the legume in the mixture. As the sole crop, B. humidicola had the highest production of shoot biomass (972 g m−2) among all treatments during the first regrowth period (middle of the rainy season). Total shoot biomass in the associations ranged from 574 (B. humidicola/C. pubescens) to 807 g m−2 (B. humidicola/S. hamata). Dry matter production for the second evaluation (end of the rainy season) showed the same trend, but lower yields were obtained due to soil drying conditions. The N content in the shoot biomass was always higher in the legumes than in grass. However, B. humidicola in monoculture showed the highest total N accumulation (18.1 g m−2) due to its higher growth rate. Legumes in monocrop had a very significant proportion of N derived from BNF (47–69%), with C. pubescens as the better N2-fixing legume (51–69%). A higher contribution from BNF also was observed in the legumes when mixed with the grass (57–76%). No clear transfer of N from the legume to the companion grass was observed, since the 15N in B. humidicola was always higher in the mixed swards. This grass appeared to have a different pattern of N uptake, obtaining a very significant amount of unlabeled N from outside the marked plots.