RETENTION OF GENETICALLY SUPERIOR LINES DURING EARLY-GENERATION TESTCROSSING OF MAIZE

Selection of partially inbred maize (Zea mays L.) lines, based on early-generation testcross performance, involves some risk of losing lines that would eventually perform well in testcrosses at homozygosity. The objectives of this study were to calculate the probabilities of retaining genetically superior lines during early testing and examine selection intensities appropriate for different testing generations and levels of heritability. Expected genetic and phenotypic correlations among testcrosses at different selfing generations were used to calculate the conditional probabilities that (i) the testcross genetic value of a homozygous (S(x)) line is in the upper alpha(x) of the distribution, given that the testcross phenotypic value of its ancestral S(n) line is in the upper alpha(n) of the distribution; and (ii) the testcross phenotypic value of an S(n) line is in the upper alpha(n) of the distribution, given that the testcross genetic value of its descendant S(x) line is in the upper alpha(x) of the distribution. Low heritability values severely decrease the effectiveness of early testing. If heritability is low, selfing for two or three generations prior to testcrossing may be desirable to increase the likelihood of retaining lines that would perform well at homozygosity. Larger proportions of lines need to be retained at earlier than at later selfing generations to prevent excessive loss of genetically superior lines. The conditional probabilities given in this paper may help breeders choose a combination of stage of testcrossing and selection intensity that would allow a balance between number of lines retained for further testing and risk of losing genetically superior lines.