Influence of pool substitution on the interpretation of fertilizer experiments with 15 N

SUMMARY The uptake of labelled and unlabelled N by wheat was measured in pot and field experiments with 15 N‐labelled fertilizer. Soils from two sites on the same series were used in the pot experiment; one had been bare‐fallowed for 22 years and contained 1.6% organic C, the other had been under grass for many years and contained 3.8% organic C. Fertilizer N increased the uptake of unlabelled soil N in both soils, i.e. there was a positive ‘added nitrogen interaction’ (ANI). There was no ANI in the field experiment. A simulation model is used to show how positive ANIs can arise as a result of ‘pool substitution’—labelled inorganic fertilizer N standing proxy for unlabelled inorganic soil N that would otherwise have been immobilized. In the low‐organic fallow soil, pool substitution accounted for the whole of the observed ANI and fertilizer N did not enhance either gross or net mineralization of soil N. Pool substitution also operated in the high organic grassland soil, but here net mineralization of soil N increased with increasing additions of fertilizer, giving rise to a ‘real’ ANI in addition to the larger ‘apparent’ ANI caused by pool substitution. This increase in net mineralization is probably caused by a decrease in immobilization of N as fertilizer N additions increase, not by an increase in gross mineralization of soil N. For pool substitution to operate, fertilizer N and soil inorganic N must occupy the same pool. This occurred in the pot experiment but not in the field experiment, where fertilizer and soil inorganic N remained separate and there was no ANI. When pool substitution occurs, fertilizer use efficiency is predictably lower as measured by the isotopic method than as measured by the conventional non‐isotopic procedure.