BREAD WHEAT - F1-HYBRID PERFORMANCE AND PARENTAL DIVERSITY ESTIMATES USING MOLECULAR MARKERS

In wheat, the possibility of introducing F-1 seed into practical agric ulture has been greatly enhanced by the discovery of effective chemica l hybridising agents (CHAs). Although some technical and economic prob lems concerning the use of CHAs for large-scale production of F-1 seed remain to be solved, a first group of F-1 hybrids has been submitted for registration in several European countries i.e., France, England a nd Italy. Combining ability for grain yield and several agronomic and quality traits was studied in an eight-parent diallel cross. Highly si gnificant combining ability effects were observed for all the traits w hile specific combining ability effects were statistically significant for grain yield, plant height, heading time and Chopin alveograph par ameter P. The level of genetic diversity between parents as estimated using molecular markers is considered a tool for predicting the hybrid performance and heterosis of crosses. To explore this possibility, RF LP and RAPD markers were used to predict the performance of hybrids ob tained from diallel and top crosses. The performance of the hybrids wa s determined in replicated plot trials sown at normal seed density in several locations. Coefficient of parentage (rp), based on pedigree in formation for all the pairwise combinations of the parents ranged from 0.01 to 0.34. The parents were assayed for random amplified polymorph ic DNA (RAPD) with 87 primers which generated 304 polymorphic bands. G enetic similarity between parents, estimated on the basis of common ba nds using the Jaccard's similarity coefficient (J), ranged from 0.25 t o 0.57. Correlation between parental diversity and hybrid performance was generally weak. A positive trend is observed in the yield potentia l of the hybrids produced in Italy in the last 10 years. In fact among the first set of hybrids produced by random crossing of the available cultivars, none produced 10% more than the checks whereas the last ge neration of hybrids includes combinations yielding 15% more than the b est standards. Our results clearly indicate the need to develop specif ic strategies in order to identify and/or to select parental lines wit h a high level of general combining ability (GCA) and specific combini ng ability (SCA). The information regarding the genetic diversity of t he parental lines do not appear helpful for predicting F-1 performance .