The associations between yield, nitrogen and soluble-carbohydrate concentration within S. 23 perennial ryegrass were investigated at three levels of nitrogen application. Controlled growth room conditions were used.
The simple correlation between yield and nitrogen concentration was negative at the low rate of nitrogen application, not present at the intermediate rate of nitrogen and positive though not significant at the high rate of nitrogen. Yield and soluble-carbohydrate concentration were only correlated (positively) at the low nitrogen rate.
At low nitrogen there were ‘efficient’ genotypes with relatively high yield and low nitrogen concentration. ‘Inefficient’ genotypes had relatively low yields and high nitrogen concentration. The majority of genotypes were neither markedly efficient or inefficient. The efficient genotypes at low nitrogen maintained their yield advantage at higher nitrogen levels with average nitrogen concentration and high numbers of tillers. Inefficient genotypes remained relatively stable in yield, numbers of tillers and nitrogen concentration.
Partial correlation indicated an underlying high degree of dependence between yield, nitrogen and soluble-carbohydrate concentration at all nitrogen levels. The association of yield and nitrogen concentration showed a similar trend over nitrogen levels to that obtained using simple correlation. Yield and soluble-carbohydrate concentration were positively related, and nitrogen and soluble-carbohydrate concentration inversely related, when the effects of the remaining attribute were eliminated.
Data extracted from Lee et al. (1977) confirmed that yield and nitrogen concentration for varieties under field conditions varied with level of nitrogen application in a similar manner to genotypes in the controlled growth room. The variety (Melle) could be characterized as being particularly efficient.
Both sets of results indicate that selection for high yield of nitrogen in herbage can best be achieved through selection for yield alone.