The effect on yield distribution of early generation selection for quality

Abstract

In simulation of a wheat-breeding situation, random F3 lines derived from seven wheat crosses were quality tested, and decisions to retain or discard them were made on the individual test results for flourmilling extraction, flour protein content and four separate measures of protein quality. Decisions were also made using regimes of tests that selected entries for high flour extraction and their balance of grain hardness, protein content and quality optimum for different wheat end-uses (bread, cakes, pastries, etc.). In the F4 generation all entries were grown in replicated yield trials. The effect of early generation selection for quality on the subsequent yield distribution was assessed by comparing the F4 yield distributions of the entire population with those for the designated quality test. Independent culling for flour-milling extraction and protein quality did not alter the nature of the subsequent yield distributions. However, independent culling in the F3 for flour protein content resulted in the loss of more lines of above average yield than below average yield in the F4 generation. Furthermore, the significant negative correlation between F3 flour protein content and F4 grain yield indicates that overemphasis on selection for high protein content would reduce the rate of yield improvement in the following generation. Application of the early generation testing regimes resulted in up to 66% of the lines being classified as having undesirable quality without the nature of the subsequent yield distribution being affected. The results of this study have application to the planning of wheat-breeding programmes that aim to produce high-yielding, good-quality wheats.