(go to KOMUGI Home) (go to WIS List) (go to NO.87 Contents)

Results and discussion

Significant mean square due to general combining ability (gca) and specific combining ability (sea) indicated the importance of both additive and non-additive gene effects (Table 1). However, general predictability ratio as suggested by Baker (1978) indicated that additive gene effects were predominant and prediction for crosses to obtain desirable segregants could be made on the basis of gca effects. The parents WH 291 and WH 377 were found to be good general combiners whereas RAJ. 1972 was found to be an average general combiner (Table 2). Other parents showed poor gca effects. Paroda and Joshi (1970) and Gill et al. (1972) argued that crosses involving one of the parents as good general combiner can be expected to throw desirable segregants. Therefore, three crosses WC 29 x WH 291, SGP 14 x RAJ. 1972 and WH 377 x HD 2329 which involved one of the parents as good or average general combiner and had good to poor sea effects were selected for further studies.

The joint scaling test of Cavalli (1952) revealed that the additive-dominance model was inadequate in all three crosses (Table 3). The failure of the model may be due to linkage, or two or many genes that may also involve higher order epistasis. Therefore, the six parameter model in each cross was applied. In general, there was a close agreement between the results of joint scaling test and six parameter model. However, the dominance component estimated through joint scaling test in WC 29 x WH 291 was found to be non-significant in six parameter model. This may be attributed to either sampling error leading to a high standard error of the estimate or internal cancellation of the gene effects in the presence of epistasis. Similar discrepancies between the results of joint scaling test and six parameter model were reported by Tonk (1988). The additive effect was higher in magnitude than dominance effect in WC 29 x WH 291 and SGP 14 x RAJ. 1972. In such cases, intermating and mass selection in early generation followed by single plant selection in later generations could be useful to derive desirable segregants (Bhatt 1972). In WH 377 x HD 2329, dominance effect was higher in magnitude than additive effect. This revealed the possibility of dispersion of genes for field resistance among the parents. In such a situation, transgressive segregants could be expected in later generations. Negative value of dominance effect exhibited to directional dominance of increasing alleles imparting field resistance. Contradictory results were reported by Singh et al. (1988). A perusal of epistatic gene effect indicated the preponderance of additive x additive gene effects in WC 29 x WH 291 whereas in other two crosses it was non-significant. Other types of non-fixable epistatic gene effects were also noted in different crosses. Under such situations, in order to get transgressive segregants and break undesirable linkages intermating followed by selection in early generations has been advocated by Gill et al. (1974) and Singh et al. (1986).

In the present study, a higher frequency of transgressive segregants was observed in the BIPs than F2 and F4 populations in WC 29 x WH 291 and SGP 14 x RAJ. 1972 (Table 4). In WH 377 x HD 2329, frequency of transgressive segregants was higher in F4 than F2 and BIPs. In general more plants were selected in F2 than F4 but majority of them were discarded on progeny testing because they showed either a low mean value or segregation. This was not unexpected because of high level of heterozygosity in the F2 population. Significant differences between predicted and observed frequencies of transgressive segregants in the F2 and BIPs were recorded. This may be due to the failure of the model. Prediction for the Finf generation was found to be more accurate from F3 families than that from generation mean analysis. This was evident from non-significant values of chi2 between observed frequencies in F4 and predicted frequencies in Finf, The present results are in conformity with the findings of Jinks and Pooni (1980) in tobacco. The highest frequency of transgressive segregants in F4 was obtained in WH 377 x HD 2329 and gave good fit to the predicted frequencies in F4.

<--Back | -->Next

(go to KOMUGI Home) (go to WIS List) (go to NO.87 Contents)