(go to KOMUGI Home) (go to WIS List) (go to NO.80 Contents)


Significant deviation of 'b' from zero and the non-significant departure of regression coefficient from unity in respect of days to heading, peduncle length, number of grains per spike, grain weight per spike, 1000 grain weight and grain yield indicated that the aforesaid diallel assumptions were valid for these traits (Table 2). However, rest of the characters showed partial failure of the assumptions but estimates of the population parameters for that traits were still possible (Hayman 1954) though certainly the estimates for such a trait are less reliable than they would have been if all assumptions had been fulfilled. With the fulfillment of most of the assumptions of the diallel analysis fully or partially in the present study, the conclusions drawn are expected to be valid and should form a guideline for improvement in the genetic material studied.

The estimates of components of genetic variance (
Table 2) exhibited that additive component (D) was highly significant for all the characters except days to maturity. The two measures of dominance H1 (dominance effect) and H2(proportion of dominance due to positive and negative effect of genes) were also highly significant for all the traits studied. Thus, it is suggested that additive and non-additive gene effects were equally important for all the characters except days to maturity, for which only non-additive component was important. These findings confirm the results of Jatasra and Paroda (1980), Sharma and Singh (1982), Nanda et al. (1983), Sharma et al. (1986), Dasgupta and Mondal (1988), Solanki et al. (1993) and Singh et al. (1993) obtained through combining ability analysis for the same set of characters.

The estimates of 'F' value which indicated the relative frequency of dominant and recessive alleles in parents were found to be positive and highly significant for days to heading, number of grains per spike, grain weight per spike, 1000 grain weight, harvest index and grain yield indicating an excess of dominant alleles, while 'F' was significant and negative for plant height, which indicated the directional dominance of the decreasing genes. Positive but non- significant 'F' in days to maturity, flag lead area and tiller number gave some indications of the excess of dominant alleles in the parental lines. The environmental component (E) was significant for days to maturity, flag leaf area, grain weight per spike and grain yield. The proportion (H1/D)1/2 representing the degree of dominance was more than unity for all the characters indicating the existence of overdominance.

The present study revealed that both additive and non-additive components of genetic variances were involved in governing the inheritance of yield and yield components, although preponderance of non-additive genetic variance was noted. In such a situation, the most suitable breeding procedure would be one which mops up the additive genetic variance and at the same time maintains heterozygosity. Therefore, it is desirable to practice bi-parental mating and/or recurrent selection, intermating the most desirable segregants, alternately with selection. This would lead to an elevation of the genetic plateau, by accumulating favorable additive genes and simultaneously exploiting the dominance variance. Although it is difficult to produce enough seed in wheat by conventional methods, the 'obligate' cross-fertilization system using male sterility, as proposed by Athwal and Borlaug (1967) can bring about large-scale intermating among selected genotypes, as envisaged under the recurrent improvement programme.

<--Back | -->Next

(go to KOMUGI Home) (go to WIS List) (go to NO.80 Contents)