-

In 1996, seven local and alien rice cultivars and lines were crossed. In 1997 parents and F 1 progenies were sowed, F 1 progenies being backcrossed with their parents. Also F 1 generation was selfed to obtain seed for F2 generations.In 1998 six genration including parents, F 1, F2, BC 1, and BC2were sown in a split plot design with 3 replications. Characteristics such as grain yield, plant height, tiller per plant,
panicale length, full and empty grainper panicle were Ineasured and evaluated. The Inean squares for family as well as for generation within fmnily were statistically significant for all the traits studied., indicating therefore, high genetic potential and variances amongst the sample Inaterials studied. Generation mean analysis as described by Mather and Jinks revealed that for most families, the traits (grain yield, plant height, paniclelength) were genetically controlled by additive and dominant gene effects,although nonallelic gene interactions were observed in some cases. The degree gf dominance in most of the fatllilies indicated the predominance of dominant and overdominant gene effects for full or eillpty grain per panicle.
Estimated narrow sense heritability based on the Ineans for 12 families were as
follows: grain yield 46%, plant height 37%, tiller per plant 39%, panicle length
38%, -filled grains per plant 41 % and untilled grains per plant 49%. So , selection of
eflte lines in segregation generations regarding additive gene effects Inay be more or less successful. In cases where the dolninance effect of genes, concerning the evaluated characteristics, were significant (tiller per plant, filled and unfilled grain per panicle), a success full selection, for these traits, my not be possible.