A testcross procedure for selecting exotic strains to improve pure-line cultivars in predominantly self-fertilizing species

Methods for identifying germplasm carrying alleles with the potential to improve a particular singlecross hybrid have been proposed and discussed in recent years. There is a need for similar methods to be used in breeding crops for which pure-line cultivars, rather than hybrids, are the goal, The objective of this research was to develop a method to identify germplasm lines with the potential to contribute favorable alleles not present in a specified pure line or set of pure lines. Given a set of adapted pure lines (A(1), A(2),..., A(m)) to be improved and a set of germplasm lines (P-1, P-2,..., P-f), the procedure consists of producing all f x m possible hybrids and evaluating them along with the parents. The testcross statistic T-ij is defined by T-ij = gamma(F-ij-A(j))+(1-gamma) (F-ij-P-i), where A(j), P-i, and F-ij represent the performance of the j(th) adapted line, the i(th) germplasm line, and their hybrid, respectively. The statistic T-i= (1/M)Sigma(T-ij) is the mean value of T-ij over all adapted parents A(j). If gamma = (1/2)(1+d'), where d' = the mean degree of dominance, then T-ij measures the potential for alleles from P-i to improve A(j), and (T-i) over bar measures the potential for alleles from P-i to improve the set A(1),A(2),..., A(m). Use of data on soybean and peanut hybrids published by other researchers suggests that the value assumed for d' has little effect on the P-i chosen. The ability of the T-ij and (T-i) over bar statistics to identify germplasm strains carrying rare favorable alleles should be assessed in empirical studies.