YIELD AND PROTEIN RESPONSES TO NITROGEN, AND NITROGEN-FERTILIZER REQUIREMENTS OF GRAIN-SORGHUM, IN RELATION TO SOIL NITRATE LEVELS

Sorghum fertiliser experiments at 40 sites on the north-western slopes and plains of New South Wales demonstrated that many soils are severe ly deficient in nitrogen (N), but most yield responses to fertiliser N occurred on sites in the southern part of the region. Grain yields re sponded to fertiliser in fewer than half of the experiments but protei n concentrations responded in about 75%.There were 4 distinct types of protein response curve, and the type of curve was related to the degr ee of N deficiency. In the most deficient experiments (mean protein 6. 1% or less), response curves were convex to the z-axis or linear; at i ntermediate deficiency (mean protein 7.2%), response curves were sigmo id; and at low deficiency (mean protein 9.7%), response curves were Mi tscherlich. Yield response never occurred where grain protein was >10% . Maximum grain yield responses and amounts of fertiliser N for maximu m profit, estimated by fitting the Mitscherlich equation to response c urves, were significantly correlated with soil nitrate N levels at var ious depths in the southern experiments, but not in the northern exper iments. This difference in N responses appeared to be caused by lower rainfall and higher soil N in most of the northern experiments. Nitrat e-N levels in soils sampled to 15 or 30 cm depth were better correlate d with yield response (r > 0.81) and fertiliser requirement (r > 0.72) than N levels to deeper depths. There was little or no fertiliser N r ecovery in the grain in the northern experiments but substantial recov ery in the south where it was generally greater than recovery by wheat in earlier experiments in the same region. Fertiliser requirement in relation to soil nitrate-N levels was lower than that of these wheat e xperiments. This was attributed to mid-spring soil sampling for sorghu m which underestimates the soil N available to the sorghum.