Summary Effects of three forage harvesting regimes—total removal of foliage and branches once (T-12) or twice a year (T-6) and 50% removal every 2 months (P-2)—on growth and biological dinitrogen fixation of Gliricidia sepium (Jacq.) Walp

removal of foliage and branches once (T-12) or twice a year (T-6) and 50% removal every 2 months (P-2)—on growth and biological dinitrogen fixation of Gliricidia sepium (Jacq.) Walp were studied under subhumid tropical conditions in Guadeloupe, French Antilles. Gliricidia sepium was grown in association with the perennial C4 grass Dichantium aristatum (Poir) C.E. Hubbard in a two-storied fodder production system. The medium-term effects of pruning on N2 fixation were assessed by the N natural abundance method. Gmelina arborea Roxb. was used as the non-fixing reference. The trees in the T-12 regime followed the natural phenological cycle, and flowering and podfilling at the beginning of the dry season reduced both foliage and nodule biomass. The T-6 regime impeded flowering, and only a few flowers, on older branches, were produced in the P-2 regime. In trees in the T-12, T-6, and P-2 regimes, fixed N comprised 54–87, 54–92, and 60–87%, respectively, of the total N in aboveground biomass, depending on sampling date. Total annual accumulation of N in harvestable aboveground biomass was highest in trees in the T-6 regime at 313 kg ha−1, of which 204 kg ha−1 of N was fixed from the atmosphere. In all treatments, about 70% of the N exported per year from the plot in the fodder harvest came from N2 fixation. Thus, N2 fixation makes an important contribution to the N economy of the G. sepium–D. aristatum forage production system, and greatly reduces the need for fertilizer application.