Wheat fertilizer experiments at 58 sites on the north-western slopes and plains of New South Wales clearly demonstrated a widespread and severe deficiency of nitrogen on many soils. The frequency (70%) and magnitude of responses to nitrogen were much greater than previously recorded. Nitrogen fertilizer required to achieve near-maximum yields was also much greater, with more than half the experiments requiring more than 30 kg N/ha and 23 experiments requiring more than 60 kg/ha. Deficiency of nitrogen for grain protein was almost universal with only two experiments failing to respond to nitrogen fertilizer. The yield response curves for all except three experiments were well fitted by the exponential (Mitscherlich) equation, but the majority of protein response curves were convex to the X axis, or linear, so that maximum protein concentrations could not be estimated. There were four distinct types of protein response curves, and their occurrence seemed to be related to the degree of nitrogen deficiency. Where nitrogen was most deficient (mean protein <10.5%), response curves were convex or linear; at intermediate deficiency (mean protein 11.7%), response curves were sigmoid, and at low deficiency (mean protein 13.4%), curves were exponential. Yield response rarely occurred where grain protein was greater than 12%. In 10 experiments with convex or sigmoid curves, the first increment of fertilizer depressed protein levels, due to the dilution effect of a large yield response. Increasing amounts of phosphorus fertilizer increased the response to nitrogen in nine experiments and in most of these the response curvature was correspondingly decreased, especially at the highest rate of phosphate. These interactions showed that nitrogen was the primary limiting factor in most of these experiments. P fertilizer tended to depress protein concentrations, especially in the absence of N fertilizer, but it had no consistent effect on protein response to N. Because of the dominance of convex protein response curves, much higher levels of fertilizer N were required to give maximum protein responses than were required to give maximum incremental yield responses. It was usually uneconomic therefore to use fertilizer solely to maximize protein increases.

In order to better understand crop growth (dry matter accumulation, DM), crop evapotranspiration (Et), and their interrelationships, we studied dryland crops of wheat (Triticum aestivum L. cv. Timgalen) sown at various dates and seeding densities in each of three years at Tamworth, N.S.W. Soil water stress was minimal before anthesis in each year, but in two years substantial stress arose before maturity. DM was increased consistently by increased plant density, and decreased at anthesis and maturity by later sowing. Crop growth rates determined over 2-week intervals around anthesis ranged from 3 to 20 g m-2 d-1, representing a range in efficiency of utilization of intercepted total solar radiation of 0.48 to 2.35%, variation which was adequately explained (R2 = 0.80) by ontogeny (days from anthesis) and Et / Ep ratio (Ep = class A pan evaporation). Et at anthesis, but not at maturity, was increased slightly by higher seedmg density; crop Et was not consistently affected by sowing date. Et / Ep over 2-week periods around anthesis was related to leaf area index, and to a lesser extent to available soil water and Ep (R2 = 0.58). For the period from the first sowing date in June or July until the middle of October, the relationship of total Et to DM production was linear and close each year, but the slope varied from 6.2 g m-2 mm-1 (cold dry year) to 14.0 g m-2 mm-1 (wet year). This variation could be attributed to annual variation in the soil evaporation component of Et, and in the ratio of DM to crop transpiration (= transpiration efficiency, TE). For 2-week periods around anthesis, TE ranged from 2.9 to 5.4 g m-2 mm-1 and was inversely related to Ep (R2 = 0.56). Provided soil evaporation can be allowed for, since it ranged from 18 to 41% of crop Et from sowing to maturity, it is argued that the crop transpirationtranspiration efficiency approach is particularly useful for analysing the growth and water use of dryland wheat.

Nitrogen uptake by wheat from both soil and fertilizer, and the efficiencies of fertilizer N (up to 116 kg/ha) for increasing yield and protein, were measured in 53 wheat fertilizer experiments during 1985-89 on the north-western slopes and plains of New South Wales. There was a highly significant (r2> 0.70) and common relationship between N uptake in unfertilized wheat (tops and grain) and soil nitrate to 90 cm depth for 4 of the 5 years of the study. A different but significant relationship occurred in 1988 when heavy rainfall before sampling leached some of the soil N beyond the sampled depth but within the rooting zone. The uptake and recovery of fertilizer N were lower in 1989, when in-crop rainfall was much lower than in the other 4 years. However, there was greater transfer of N from the herbage to the grain than in the wetter years. With increasing increments of fertilizer N. there was a much larger average decline in agronomic efficiency than in the recovery of fertilizer N or in physiological efficiency. Consistent with this, the average protein efficiency of fertilizer N tended to increase with increasing increments in every year except 1989. Although the highest increment of fertilizer N was always the least efficient for increasing grain yield, it exceeded the level required for profitability (8 kg grain/kg fertilizer N) in 20% of experiments. In experiments in which agronomic efficiency of the highest fertilizer increment was too low for profitability, there were at least 10 experiments in which the protein response was probably sufficient to make the highest increment profitable. The agronomic, protein and physiological efficiencies of fertilizer N in at least 10% of these experiments were higher than previously recorded in Australia and are comparable with the highest values recorded for wheat in other regions of the world.

The effect of the narrow-leafed lupin (Lupinus angustifolius L.) on growth and production of subsequent cereal crops was studied at three sites on the acidic, N-infertile soils of the Pilliga Scrub region of northern New South Wales.Beneficial effects of lupins on dry matter production were evident in either one (Kamala, Florida A) or two subsequent crops of wheat (Florida B). In the absence of fertilizer N, wheat following lupins outyielded wheat following wheat by an average of 57%. At Florida B, the second wheat crop following lupins outyielded the wheat only plots by 35%. The response to lupins was established early in vegetative growth and was essentially maintained. Effects of lupins on grain yields, evident at all sites in the first crop, were maintained at the two Florida sites for a second crop. Increases in the absence of fertilizer-N were between 0.29 and 1.22 t ha-1 (55 and 145%) (crop 1) and as great as 0.39 t ha-1 (38%) in year 2. Barley responded similarly with increases of 127 and 47% in years 1 and 2, respectively. Amounts of fertilizer required to raise the yields of wheat following wheat to those of wheat following lupins ranged between 40 and 80 kg ha-1. Effects of lupins on protein concentration of cereal grains were inconsistent; largest effects were achieved through N fertilization. Both the incidence and severity of root diseases at Florida A were reduced in the rotation plots; the effect persisted into the third wheat crop following lupins. Disease control was not a factor in the lupin effect at Kamala. Nitrogen budgets for the unfertilized wheat-wheat and lupinwheat sequences at Kamala and Florida A indicated that lupin cropping (i) resulted in potential net gains of soil N of 128 (Kamala) and 29 kg ha-1 (Florida A), and (ii) increased N yields of the following wheat crops by 37 (Kamala) and 20 kg ha-1 (Florida A).

Factors responsible for a decline in wheat yields in some paddocks in the Gunnedah (northern New South Wales) area were investigated in a series of field experiments. Yield decline is not apparent on soils cultivated for less than 10 years, and barley and sorghum crops are not affected. Plant nutrition apparently is not a major factor in the decline, because, although there was a small response to zinc application in 1 experiment, wheat yields were not significantly increased by the application of N, P, K, Cu, Mg, B, Mn, Mo or S in a paddock where the yield of wheat was half that of the surrounding barley crop. Fumigation of the soil with methyl bromide increased wheat cv. Timgalen yield by 78% to equal that of barley cv. Clipper in a second experiment. Large numbers of root lesion nematodes (Pratylenchus thornei) were associated with unfumigated wheat plots, with smaller numbers present on barley plots. In a third experiment, the grain yields of all of the 9 cereal genotypes tested were increased when sown with 4 kg/ha of aldicarb. The relative yield increase associated with aldicarb showed good correlation with the reputed tolerance of the genotypes to P. thornei. Pratylenchus thornei occurred to a depth of 120 cm. Aldicarb did not affect P. thornei numbers below 15 cm, while methyl bromide controlled P. thornei to depths of 30-90 cm. Aldicarb virtually eliminated P. thornei from the surface 10 cm of soil, but had no effect on the incidence of the fungal diseases common root rot (Bipolaris sorokiniana) or crown rot (Fusarium graminearum). Soil fumigation with methyl bromide controlled B. sorokiniana and F. graminearum as well as P. thornei and gave yields 9% higher than did treatment with aldicarb. In another experiment, ethylene dibromide (2.7-10.8 L/ha) had no consistent effect on the yield of wheat or numbers of P. thornei while terbufos (0.6 and 1.2 kg/ha) gave a small reduction in P. thornei numbers and a small increase in yield. These data indicate that P. thornei is at least partially responsible for wheat yield decline in parts of northern New South Wales.