|dc.description.abstract||Two field experiments were conducted on different soil types and locations in Canterbury during the 1981/82 and 1982/83 growing seasons. Genetic variability in grain yield and flour protein content (protein content) of wheat (Triticum aestivum L.) was evaluated. The influence of several agronomic practices on the yield and protein relationship was studied. The cultivar x environment interactions for grain yield, protein content and other related characters and their stability estimates were also examined.
Eight wheat cultivars (Ruru, Hilgendorf, Tiritea, 2633,01, Wakanui, Bounty, Atlas 66 and Kopara) of varying grain yield and protein content were used in both experiments. In the first experiment, the wheat cultivars were grown in six environments achieved by varying site, soil fertility and sowing date. In the second experiment, another eight environments were created with four levels of nitrogen (N) fertiliser (0, 100, 200 and 400 kg N ha⁻¹) and two plant densities (125 and 500 plants m⁻²). Each N level was split equally between Zadoks growth stages 22, 30 and 39.
Grain yields differed between cultivars. Heavy grain yields were mainly associated with large harvest indices (HI) and (to a lesser extent) with many ears m⁻², grains spikelet⁻¹, grains ear⁻¹, spikelets ear⁻¹ and a large biological yield at maturity.
There were significant differences between cultivars in protein content in experiment I, however, after adjusting their protein contents to a common HI using covariance analysis, the differences were small. There were also differences between cultivars in protein content in experiment II. However, protein contents of the eight cultivars in this experiment could not be adjusted to a common HI. This was because the regression slopes of protein contents on HIs differed between cultivars. This was due mainly to the cultivar x N interaction for HI but not for protein content.
Protein content increased with each increment of N application. The response of grain yield to the N application was, however, hyperbolic. The 100 kg N ha⁻¹ level was, in general, the optimum level to increase both grain yield and protein content. The grain N yield and N uptake increased with each increment of N level. The N harvest index (NHI), however, decreased with increasing N level. Biological yield at maturity increased with increasing N application but the 'response of HI to N level' was hyperbolic.
Higher plant density reduced grain yield despite an increase in number of ears m⁻² as this was compensated by a greater decline in grain number. Increased plant density reduced grain N yield and N uptake mainly through reduction in grain yield and biological yield respectively.
There was a hyperbolic response of grain yield to N uptake. On the other hand, protein content increased linearly with increasing N. uptake. For every g m⁻² increase in N uptake there was a .3 percentage point increase in protein content.
N uptake was positively associated with grain yield at the phenotypic level but negatively related at the genotypic level. N uptake was positively correlated with protein content at both the phenotypic and genotypic levels. Also, grain N yield (hence protein yield m⁻²) was positively correlated with grain yield and protein content at both the phenotypic and genotypic levels.
Grain yield and NHI were positively related but the correlation between NHI and protein content was weak at both the phenotypic and genotypic levels. The stability estimates across environments under study showed that NHI was a stable character in most cultivars.
There were cultivar x environment interactions for grain yield, biological yield, HI and other characters related to N uptake and distribution. However, the cultivar x environment interaction for protein content was negligible.
The broad sense heritability for grain yield was low (ranging from 12.0 to 32.7%). However, the heritabilty for protein content was high (ranging from 75.7 to 81.3%). The broad sense heritabilities for grain N Yield, N uptake and NHI were relatively low (ranging from undetectable to 35.5%, from undetectable to 4.4% and from 12.8 to 49.7% respectively). The heritability for HI was moderately high (ranging from 49 to 57%) but the heritability for biological yield was low (ranging from undetectable to.8.5%).||en