USE OF RESIDUAL FERTILIZER N-15 IN SOIL FOR ISOTOPE-DILUTION ESTIMATES OF N-2 FIXATION BY GRAIN LEGUMES

Estimates of the proportion of nitrogen (N) derived from the atmospher e (pN(atm)) by chickpea and lentil in the alternate phase of a cereal- legume 2-year rotation, for each of 3 seasons (1993, 1994, and 1995) i n northern Syria, were obtained from isotope dilution methodology usin g residual fertiliser N-15 in the soil (IDres). The N-15 had been immo bilised, during the year antecedent to the legume, from N-15-enriched fertiliser which had been applied at sowing to wheat in the cereal pha se of the rotation at 30 kg N/ha. For lentil in 1994, and for chickpea in 1993 and 1994, the IDres estimates of pN(atm) were compared with t hose obtained by using the classical N-15 isotope dilution method (ID) where N-15-enriched fertiliser (either 30 or 10 kg N/ha) was added at sowing to both the legume and non-fixing reference crops. Estimates o f pN(atm) for 1994 from the 2 methodological approaches were significa ntly (P < 0.01) different for lentil, with ID resulting in a higher es timate than IDres (0.92 v. 0.85). For chickpea in the same season (199 4) the IDres estimate was significantly higher than the ID estimate (0 .88 v. 0.78) at 30 kg N/ha because the N fertiliser inhibited biologic al N fixation (BNF). However, using a lower fertiliser rate (10 kg N/h a) for ID the estimate of pN(atm) obtained for chickpea in 1994 was 0. 91, which was slightly higher than the IDres estimate. Proportional re liance on BNF was estimated to be greater in spring than at harvest fo r both lentil and chickpea. The estimates of pN(atm) obtained at harve st were greatest (>0.82) for both crops in 1994 and less, but similar, for both crops (0.64-0.79) in the other 2 seasons (1993 and 1995). Al though substantial amounts of residual fertiliser N were present in th e soil, only a small proportion of the original fertiliser N added (<5 %) was utilised by plant uptake plus any losses in the residual year, indicative of a slow remineralisation rate for the immobilised labelle d N. Nevertheless, the crops in the residual year were sufficiently en riched to allow for estimation of pN(atm). The N-15 abundance, at harv est, of wheat shoots from the N-15 IDres method was similar to that of the soil nitrate and ammonium pools, suggesting that plant N uptake t hrough the season had been from an N pool of reasonably constant enric hment. This was in contrast to wheat receiving N-15-labelled fertilise r at sowing, where the shoots at harvest had a higher N-15 abundance t han the plant-available N pool, indicating a declining N-15 enrichment of plant-available N in the soil through the season. Furthermore, var iability in the N-15 abundance of plant-available N with soil depth wa s also demonstrated to be greater where the N-15 IDres method was used , for ammonium N at least. These differences in N-15 enrichment patter ns of the plant-available N pool for the 2 methods resulted in signifi cantly different estimates for pN(atm) of lentil in 1994 but for all t he other comparisons there were no major differences between estimates obtained using either ID or IDres.