Yield gains in wheat by three methods of selection

Thumbnail Image
Seitzer, J. F.
Journal Title
Journal ISSN
Volume Title
Three crosses of spring wheat (Triticum aestivum L.), differing in degree of heterosis were chosen to compare the efficiency of three methods of selection for identifying high yielding late generation lines. The three methods compared were (1) a pedigree method where visual selection was practiced in F3, (2) an early generation yield test where F3 plots were compared to adjacent controls, and (3) an early generation yield test wherein replicated tests with hill plots were used to evaluate the yield potential of F3 families. The efficiency of these methods was evaluated in F5. One hundred and eighty lines derived from 45, 6 and 3 selected F3 lines in methods (1), (2) and (3) respectively, were grown at each of 2 locations. Comparisons among crosses were made for mean yield, variances among F3 families and among F5 lines within F3 families, actual line yield and line yield relative to the control variety. The methods did not differ significantly with regard to mean yield, variances and line yield. Significant differences between methods were obtained in two crosses when line yields were compared to the control. In cross I, all three methods of selection were effective in recovering lines outyielding the control, but the control plot method retained the most. In cross II none of the methods was successful in identifying lines which exceeded the control. In cross III the hill plot method isolated significantly more lines than the pedigree method and was slightly better than the control plot method of selection. Both crosses I and III were more heterotic than cross II. It was concluded that early testing may have an advantage when dealing with crosses of lower yield potential.