A GENETIC STUDY OF COTTONSEED OIL CONTENT ASSOCIATED WITH GLANDED AND GLANDLESS STRAINS (DIALLELS, COMBINING ABILITY, GOSSYPOL, INHERITANCE, MATERNAL EFFECT)

Twelve cotton lines with different seed-oil contents, six glanded and six glandless, were crossed in a full diallel mating system. All progenies were evaluated for seed-oil percentage (SO), seed index (SI, mg/seed), and seed-oil index (SOI, mg/seed), to obtain information on the inheritance of these traits.;Data were analyzed according to the procedures of Hayman (1954a, 1954b, 1957, 1958) and Griffing (1956), and by utilizing generation means of individual crosses. Results indicated that maternal effects were not significant for all traits, but reciprocal effects were significant for SI and SOI. Additive effects were the most prevalent in the analyses. However, the analysis of F(,2) data detected significant amounts of epistasis for all traits, indicating that a more complex model than that of Hayman (1954b) is needed to explain gene action in the population used in this study. The genetic analysis by generation means also detected mostly additive effects, but also epistasis of additive type, and to a lesser extent, epistasis of dominant type. The highest additive components were found in the glandless crosses involving parents of high and low oil content. Analysis of gene effects by generation means was, therefore, more adequate to describe gene action for SO, SI, and SOI, than was the Hayman's procedure.;The effect of glandless genotypes on SO was significant and positive. Analysis of segregating F(,2) families for glandless and glanded phenotypes indicated that SO was higher for glandless phenotypes than for glanded ones. However, this observation was firmly established only in crosses involving parents with high oil content. In this regard, the highest glandless oil parent would be an effective source of high oil content in breeding programs.;The experiments were carried out in the 1983 and 1984 seasons, at the Agronomy Field of Texas A&M University, adopting a randomized complete block design with four blocks. However, four intermediate parents for SO, and the reciprocals were omitted in the second year. All statistical analyses were performed considering the genotypes as fixed, and blocks and block interactions as random effects. The F(,2) generations were used from crosses within glanded and within glandless parents in the second year. Moreover, F(,2) families derived from four different crosses among glanded and glandless parents were separated into four phenotypic glanded/glandless classes to verify potential genetic relationships between those four classes and SO.