Quantitative genetics of rice II. A comparison of the efficiency of four breeding methods

Abstract

A comparison of the efficiency of four breeding methods, the pedigree, modified pedigree and bulk methods, and single seed descent, with a pair of crosses whose potential was assessed in a previous paper, showed that, on average, the F 6 products of the three methods in which selection had been practiced outscored those produced by single seed descent for most characters for which improvement was sought. However, though the best lines produced by these methods achieved the desired targets for every character considered singly, none were significantly better than the best single seed descent line for any character in either cross. These results indicate that while it has been possible to practice effective negative selection during the course of inbreeding, by culling plants of poor performance, attempts to carry out positive selection of plants of superior performance were less successful. An audit of the costs incurred in producing the most promising F 6 lines that were extracted by these methods from each cross revealed that the single seed descent lines were the cheapest and those of the pedigree and modified pedigree methods the most expensive. These results indicate, therefore, that single seed descent is at least as effective as the other methods, is less costly and, where three generations can be raised per annum, is more rapid. Comparison of the observed proportions of superior single seed descent F 6 lines, with those predicted from information obtained from the F 3 generation of these crosses showed that for the majority of characters, the former were close to the latter. Single seed descent, therefore, has the further advantage that it is possible to predict the potential of a cross on information obtained from the early generations of the pedigree.