Inheritance and selection in two New Zealand wheat crosses

Abstract

The objective of this study was to evaluate the inheritance of yield per plant and yield components, their responses to single plant selection in F₂, and line selection in F₃ of the two crosses, Hilgendorf x Oroua and Kopara x Oroua.

Parents, F₁, F₂ and first generation backcrosses grown in spaced planting were used in the inheritance studies. Scaling tests (Mather and Jinks, 1971), joint scaling test (Cavalli, 1952), estimates of parameters on an epistatic model (Jinks and Jones, 1958) and heritability estimates in narrow sense (Warner, 1952) were computed. Additive gene effects made a significant contribution to the genetic control of spikelet number per ear and yield per plant in Hilgendorf x Oroua, grain number per ear, weight per grain and ear number per plant in Kopara x Oroua, and grain number per spikelet in both crosses. Grain number per spikelet was shown to be under relatively simple additive genetic control. Dominance and non-allelic gene interaction effects appeared to be important in the inheritance of the remaining characters. As estimated by the method of Warner (1952), the narrow sense heritabilities of ear number arid yield per plant were not detectable and known to be low. Other yield components showed a range of heritability from moderate to high.

The spaced plant F₂ populations of the two crosses formed the base populations for selection. Single plants were selected in the F₂ for grain number per spikelet, grain number per ear, weight per grain and yield per plant based on their highest score. The observed responses (in the spaced plant F₃) due to selection for the three yield components were either close to or more than two-thirds of the responses predicted from F₂. The values of the components of yield per plant were significantly increased as a result of direct selection. The yield per plant showed a lack of response to direct selection. Selection for grain number per spikelet and grain number per ear led to significant positive responses in yield per plant. The values of F₂ - F₃ correlations for these yield components were comparatively higher than those for weight per grain and yield per plant. Selection for weight per grain did not change yield per plant in Kopara x Oroua but gave a small response in yield per plant in Hilgendorf x Oroua. Poor correlations were observed between F₂ and F₃ for weight per grain and yield per plant.

F₃ lines descended from F₂ single plant selections were selected on the basis of high mean performance (in spaced plant F₃) for yield per plant and the three yield components. The populations of each cross, and the low and high densities within a cross, were grown separately. The responses to F₃ line selection for the yield components were positive. The yield per plant did not show any response to direct selection. The realised heritabilities of the yield components were higher than those estimated for yield per plant. The yield per plant had a low realised heritability. The dissimilarities in realised heritabilities for the same character at low and high densities were presumed to be due to different abilities of the populations to react to the environments.

F₃ line selection for grain number per spikelet gave positive correlated responses in most yield components and yield per plant at both densities in Hilgendorf x Oroua and at high density in Kopara x Oroua.

Although grain number per ear and weight per grain were compensating for each other, selection for grain number per ear in Kopara x Oroua brought about positive correlated responses in yield components (except weight per grain) and yield per plant at high density. Selection for weight per grain in both crosses led to significant increases in yield per square metre at low density, mainly due to high ear number per square metre.

Though the responses in yield per square metre were variable, these selected populations exceeded the highest yielding parental cultivar of the respective cross for yield. The lack of significant positive correlations between the performances of populations at low and high densities for ear number and yield suggested high genotype-environment interaction effects.