Many techniques have been proposed to increase the probability of obtaining superior segregating populations, and major efforts have been invested by researchers worldwide. In this sense, the present work had as objective to verify the efficiency of using parental average, F1 heterosis and heterobeltiosis and the morphological and molecular genetic distance in the prediction of grain yield in oat. Grain yield and its primary components, were measured in F2 and F3 generations originated from the diallel cross involving five parents. Data were obtained from two experiments carried out in the Centro Agropecuário da Palma, in Capão do Leão County, RS, Brazil, in 2002 and 2003. The results indicate that the performance F2 and F3 populations can not be adequately predicted on the basis of the performance of F1 generation. Grain yield of F2 and F3 populations could not be predicted by the genetic distance calculated by morphology or molecular markers. Crosses of high-yielding genotypes did not always yield offsprings with positive transgressive segregation.
White oat; heterosis; genetic distance; plant breeding
