Transfer of N fixed by a legume tree to the associated grass in a tropical silvopastoral system

Below-ground transfer of nitrogen (N) fixed by legume trees to associated non-N2-fixing crops has received little attention in agroforestry, although the importance of below-ground interactions is shown in other ecosystems. We used 15N natural abundance to estimate N transfer from the legume tree Gliricidia sepium (Jacq.) Kunth ex Walp. to C4 grass Dichanthium aristatum (Poir.) C.E. Hubb. in a silvopastoral system, where N was recycled exclusively by below-ground processes and N2 fixation by G. sepium was the sole N input to the system. Finding a suitable reference plant, a grass without contact with tree roots or litter, was problematic because tree roots invaded adjacent grass monocrop plots and soil isotopic signature in soil below distant grass monocrops differed significantly from the agroforestry plots. Thus, we used grass cultivated under greenhouse conditions in pots filled with agroforestry soil as the reference. A model of soil 15N fractionation during N mineralization was developed for testing the reliability of that estimate. Experimental and theoretical results indicated that 9 months after greenhouse transplanting, the percentage of fixed N in the grass decreased from 35% to <1%, due to N export in cut grass and dilution of fixed N with N taken up from the soil. The effect of soil 15N fractionation on the estimate of the reference value was negligible. This indicates that potted grass is a suitable reference N transfer studies using 15N natural abundance. About one third of N in field-grown grass was of atmospheric origin in agroforestry plots and in adjacent D. aristatum grassland invaded by G. sepium roots. The concentration of fixed N was correlated with fine root density of G. sepium but not with soil isotopic signature. This suggests a direct N transfer from trees to grass, e.g. via root exudates or common mycorrhizal networks.