Use of combining ability estimates to identify the genetic potential of selected winter parental lines (Triticum aestivum L.em Thell)

Graduation date: 1989A major factor limiting the efficiency of plant improvement programs is\ud the lack of knowledge as to which parental combinations to make when working\ud with quantitatively inherited traits. The primary objective of this study was to\ud provide information regarding the use of combining ability analysis to predict\ud which cross combinations would result in the greater frequency of desirable\ud segregates.\ud Five winter wheat parents were used to develop a 5x5 diallel cross. Over\ud a two year period parents, Fl, F2, F3, and F4 generations involving top cross\ud populations were evaluated. Information was collected on an individual plant\ud basis for seven traits: plant height, tiller number, spike length, spikelets per spike,\ud biological yield, grain yield, and harvest index. Griffing's (1956) model 1,\ud method 2 was used to estimate GCA and SCA in Fl and later generations.\ud Narrow-sense heritability estimates were computed using regression in standard\ud units across generations.\ud Significant GCA estimates were obtained for most traits in all generations\ud with the exception of tiller number and grain yield in the Fl during 1986-87. Specific combining ability estimates were also found to be important for all traits,\ud except spikelets per spike in the Fl and F3 generations. High GCA to SCA\ud ratios indicated a preponderance of additive gene action. Exceptions to the trend\ud were grain yield and tiller number in the Fl during 1986-87. Low GCA and\ud narrow-sense heritability estimates suggest that selection for these two traits\ud would have to be delayed until later generations.\ud When considered over generations, both GCA and SCA estimates were\ud found to be significant. Significant SCA over generations indicated the presence\ud of additive x additive epistasis. The GCA effects contributed by a particular\ud parent were consistent over all generations indicating that F2 and F3 generations\ud can be used effectively for combining ability analysis. However, the results did\ud suggest that GCA estimates were more influenced by the environment than SCA.\ud For the traits measured and in the populations used in this study, GCA\ud effects did predict the best parental combinations for obtaining the highest\ud percentage of desirable progeny in segregating populations