Variability in the third and fourth generation of selfed progenies of regenerants derived from an indica and japonica rice (Oryza sativa L.)

Six characters in R3 and five in R4 generation showed significant variations in UPL Ri 7. Panicle length (among anthers and within regenerants/A5), culm width (among anthers), leaf length (within DH lines/R1/A1), and percent fertility (within DH lines R1/A2) differences were noted in R3 generation while panicle length (within regenerants/A5), culm number (within DH lines/R2/A1), culm width (among anthers and within DH lines R1/A3), leaf length (within DH lines/R1/A1), and percent fertility (among anthers) showed variability in R4 generation. In Taipei 309, significant variations were observed in four quantitative traits each in R3 and R4 generation. Plant height (within regenerants/A2 and within DH lines/R1/A5), panicle length (within regenerants/A3 and with DH lines/R1/A4), culm width (within DH lines/R1/A4), and percent fertility (among anthers) showed differences in R3 generation. Variations were also recorded in plant height (within regenerants/A2), panicle length (within regenerants/A3), culm width (within DH lines/R1/A4), and percent fertility (within DH lines/R1/A2) in R4 generation. Results in both cultivars indicated that mutations affecting these traits occurred during the process of anther culture and these alterations could still be observed in vivo in R3 and R4 generations. Three out of eight instances in UPL Ri 7 and three out of seven in Taipei 309 showed consistent significant differences across 96 and 97 dry seasons, respectively. Pairwise mean comparison in each case showed that changed characters tended to maintain their performance relative to others across generations, indicating that some variant quantitative characters were stably inherited. Variant traits that did not perform with the same fidelity could be attributed to environmental fluctuations. Some laboratory and field conditions were applied to both subspecies of rice. Significant variations were detected in five UPL Ri 7 characters across generations. In Taipei 309, four traits showed variability, patitioning of the total variance in each cultivar showed that more significant differences were registered by UPL Ri 7 than Taipei 309. Results confirmed intergenotypic differences in terms of gametoclonal variation. In general, this experiment further corroborates previous investigations that indicate that anther culture system currently used in generating doubled haploid lines at PhilRice, UPLB Laboratory is mutagenic. The level of gametoclonal varition detected here, however, was not as extensive as reported by previous authors in some cultivars and in one instance in Taipei 309. Furthermore, visual observation of DH lines failed to notice any segregation of traits characterized by early generations of conventionally derived lines. In this context, advancement of DH lines to three or more seed generations could be undertaken fully to determine their stability and breeding worth