The predictive value of parental, F₁ and early generation hill-plot testing for yield among rapidly advanced hexaploid triticales

Abstract

Rapid generation turnover techniques were used to identify parental, single cross or backcross populations with yield potential in triticale. Both complete (complete rye genome) and substituted (2D/2R substitution) triticale karyotypes were crossed. Temperature and light controlled greenhouses, immature seed germination, embryo culture and off season nurseries were used to advance three generations in 11 months. Due to seed limitations in the early generations, hill-plots were used to measure total dry matter, grain yield and harvest index. Mid-parent, F2 and F3 hill-plot yield was equivalent to or better than dry matter and harvest index for the prediction of F4 hill-plot yield. Both hill-plot yield (r = 0.54) and dry matter (r = 0-51) were significantly correlated (P < 0.01) with large plot grain yield at the same site in the same year. The best predictors of F4 hill-plot yield were mid-parent, F2 and F3 mean hill-plot yield with phenotypic correlations of 0- 26¦ 0.17, 0.3510 -03 and 0.44¦ 0.06 respectively. The F1 (r = -0.02¦0.12) provided no association with F4 hill-plot yield, reflecting the high average heterosis for yield (22%) observed in this material. Mid-parent, F2, F3 and F4 dry matter and yield means were significantly lower in inter-karyotypic crosses than intra-karyotypic crosses; however, the relationship between generations appeared to be independent of karyotype. A response to visual selection of individual spikes from hill-plots was obtained at two sites for dry matter and grain yield.