Towards a breeding strategy for yield improvement in selected lines of hexaploid triticale (X. triticosecale Wittmack)

Abstract

To establish a breeding strategy for grain yield improvement, 11 hexaploid triticale lines and 24 of their F₁ hybrids were studied in four environments (August vs. October sowing /22 vs. 67 plants m⁻²). Two cultivars and a third generation selection were used as testers. Statistical analysis of genotype x environment interaction, line x tester mating design and principal components of correlations were used to evaluate 20 quantitative characters.

Sufficient genetic variability was found between the lines for their inclusion in a breeding programme. Aranui, Juanillo, Karere, Lasko, Salvo and Swans were well adapted and had stable high yield in all environments. Presto was suitable for favourable environments only. Genotype x environment interaction was important in all characters except biological yield, flag leaf area, flag leaf length and plant height and influenced GCA in most characters including grain yield and SCA in 8 characters, as indicated by line x tester analysis. GCA was more stable than SCA for all characters except plant height (female parent) and 1000 seed weight and spike length (male parent).

Most characters were controlled by additive gene action. An exception was 1000 seed weight. Lasko, Presto and Salvo were good general combiners for grain yield but not for early maturity and erect morphotype. Aranui, Juanillo, Karere, Swans and Till were good combiners for these characters.

Narrow sense heritability was high (≥0.50) for harvest index, number of grains per spike and number of tillers per plant. Moderate heritability (0.30 - 0.50) was found for 10 characters including grain yield per plant, which should therefore respond to intense direct selection.

Neither SCA effects nor F₁ performance alone were good predictors of grain yield heterosis. High F₁ performance could not predict high SCA effects. For developing hybrid varieties, simultaneous consideration of these three parameters is suggested. Grain yield heterosis effects were environment dependent and ranged from -70.0% to 51.8% over the better parent. Heterosis over the highest yielding line was found for Presto x Salvo, Presto x Pulis, Lasko x Salvo, Lasko x Fulis and Presto x Aranui in some environments, suggesting the feasibility of hybrid varieties.

Grain yield per plant was determined by two principal components. Principal component and genotypic correlation analysis indicated that days to heading, days to anthesis and harvest index are suitable for indirect selection for grain yield. However, harvest index and number of grains per spike gave the greatest expected correlated response in yield. As grain yield had the highest expected genetic advance, direct selection is suggested.

Simple breeding methods such as pedigree, bulk, single seed descent or modified breeding methods which can exploit the fixable genetic variance should give satisfactory results.