关键词: 双列杂交, 配合力, 估算方法

Abstract: Combing ability including the general combining ability(GCA) and the special combining ability(SCA), which was proposed by Sprague and Tatum(1942), can be expressed as the following statistical model, x_ij=μ+g_i+g_j+s_ij+e_ij. SCA is the difference between the observed value x_ij and the sum of μ+g_i+g_j if the error item e_ij were not thought about. Another method of estimating SCA, firstly proposed in a paper "A preliminary report on the study of quantitative traits in corn inbred lines" written by Prof. Yang Yunkui in 1965, is defined as the difference between x_ij and the mean parents value(MP), viz. s'_ij=x_ij-(1)/(2)(x_i.+x_.j)=s_ij+(1)/(2)(g_i+g_j). The two methods of estimating SCA differs from each other because the origins denote differently. GCA was enlarged and SCA was dwindled as the genetic hypothesis of Sprague's definition of combining ability and the origin were improper, and the hybrid superiority was expressed truly. On the other hand, the Yang's definition of SCA, which was consistent to hybrid superiority, possessed unambiguous foundation of genetic theory. Moreover, the most used method of calculating GCA and SCA is the one proposed by Griffing(1956) which adopted the least squares. By using the theories and methods of quantitative genetics, three methods and their relationship were analyzed and compared to understand the different results from method 3 and 4 of Griffing's diallel cross trials. To these trails, the results of GCA and SCA may be inter-transformed and inter-validated among three methods. If the object is evaluating the value of a series parents, three methods can yield the same conclusion. Otherwise, if forecasting hybrid superiority is the mainly object, the Yang's method of estimating SCA can be the most reliable.

Key words: diallel cross trial, combining ability, method of estimation