MEASUREMENT AND SIMULATION OF THE EFFECTS OF N-FERTILIZER ON GROWTH, PLANT COMPOSITION AND DISTRIBUTION OF SOIL MINERAL-N IN NATIONWIDE ONION EXPERIMENTS

To aid the development of simulation models for N-response, N-fertilizer experiments with onions (Allium cepa L.) were carried out on 5 different sites. In each experiment, there was little loss of fertilizer-N in soil during the period between application and rapid crop growth and little loss of mineral N by leaching at any time. Even so, a substantial proportion of the N applied as fertilizer could not be accounted for in the crop and soil at harvest; the sum of soil mineral-N plus crop N (excluding fibrous roots) was always linearly related to N rate applied over the entire range (0-300 kg N ha-1) and the gradient was always approximately the same, 0.64, irrespective of soil type or the amount of nitrate remaining in soil at harvest. Evidence was obtained that the phenomena resulted from roots retaining N and inducing immobilization at a rate proportional to soil nitrate concentration and that the proportionality constant was similar on all sites. Throughout plant growth there was little luxury consumption of N and the critical %N was related to plant mass by an equation previously deduced for other C3 crops (Plant and Soil 85, 163); plant nitrate concentration in the early stages increased with soil mineral-N (0-30 cm) to a maximum which varied from site to site but the nitrate concentration in the mature crop was always negligible. Plant yield in the early stages of growth generally declined with increase in fertilizer-N, despite the crops having been planted as sets and no more than 150 kg N ha-1 broadcast at one time; but at maturity, yield always increased asymptotically with increase in fertilizer-N. Mineralization rates were approximately the same in the first as in the second half of each experiment. At harvest, residual soil mineral-N in the upper 30, 60 and 90 cm of soil increased with increase in fertilizer-N even when crop demand for N exceeded supply. At harvest in every experiment, the ratio of crop dry weight in the absence of added N to the maximum obtained was approximately equal to the ratio of plant %N (with no fertilizer) to critical %N. The various phenomena concerning yields, plant-N contents, and values of soil mineral-N at harvest were quite well simulated by a slightly modified version of a previously published model (Fert. Res. 18, 153) with few site-dependent inputs.