Soil mineral N and soil water content at planting, biomass accumulation, yield and grain quality parameters (hectolitermass and protein percentage) were measured on an unfertilized and recommended-N-application treatment during two consecutive growing seasons (1997-1998). The trials were planted in a fallow-wheat-wheat cropping system at three representative localities in the summer rainfall region of South Africa. High levels of available soil water and mineral N were measured following the fallow period preceding the start of the trials in 1997. For example, soil water content was 81.7%, 69.6%, and 78.2% of DUL at Bethlehem, Kroonstad and Petrusburg respectively. Although comparable total soil profile water contents to 1997 were measured in 1998 at all three sites, the cultivation zone (0-400 mm) had a substantially lower soil water content. This was due to erratic rainfall distribution during the fallow period, which prevented effective soil cultivation management, subsequent soil water conservation and residue decomposition. Undecomposed residue in the cultivation layer at planting appeared to affect availability of soil mineral N to the growing crop. At planting in 1998, undecomposed crop residue amounted to 53.6% at Bethlehem, 32.5% at Kroonstad and 46.9% at Petrusburg of that added at harvest in 1997. Soil mineral N was lower at planting in 1998 compared to 1997 due to decomposing residue (C:N ratio of above 73) in the cultivation zone immobilizing soil mineral N. This reduced initial growth, N accumulation, yield, and grain protein percentage without additional fertilizer N. Distribution of soil mineral N showed notable amounts in the 600-1200 mm soil layers, with limited changes over the trial period. This was linked to low root exploration of these soil layers (10-15% of total root distribution). The ratios of soil mineral NH(4+):N0(3)- for the different soil layers indicated similar values over the trial period. Climatic data for the localities indicated differences in the amount and distribution of rainfall and temperatures during the study period, which influenced crop development, yield and grain protein percentage. At Bethlehem above average in-season rainfall was measured during 1997, at Kroonstad average rainfall and at Petrusburg below average in-season rainfall. Response to applied N at the localities varied in magnitude during 1997. Nitrogen application significantly increased N concentrations of plant components, N uptake, yield and grain protein percentage, although values for all these parameters were lower in 1998 than in 1997. Indeed higher yields were produced in 1997 (mean=1.838 t ha(-1)) compared to 1998 (mean=0.980 t ha(-1)). A significant yield response to applied N was measured at the two higher yielding localities in both cropping years, but there was no significant response at the lower yielding locality. The limiting factors appeared to be the availability of soil water and residual soil mineral N. From the calculated response functions, the variables soil water content at planting, soil mineral N content at planting, in-season rainfall, and added fertilizer N explained the bulk of the variations in grain protein percentage, plant N uptake, and yields. It was concluded that the present fertilizer N recommendation system for dryland wheat production, which is based on fertilizer response curves for specific yield potentials, should be augmented by using initial soil mineral N and water contents in the profile measured prior to planting.

Thesis (M.Sc.Agric.)-University of Natal, Pietermaritzburg, 2002.