In the present paper, genetic models for quantitative traits of endosperm in cereal crops (Zhu 1992; Zhu and Weir 1994; Zhu 1996a, b) were used to evaluate genetic main effects and genotype x environment (GE) interaction effects of seed, cytoplasm and maternal plant, and to predict the breeding value of parents and heterosis for cooking quality traits of rice. Nine cytoplasmic male sterile lines (Zhexie 2A and others) as females and five restorer lines (Cezao 2-2 and others) as males were used in an incomplete diallel cross (9 x 5) in two years. These CMS and restorer lines were randomly sampled from a reference population. Seedlings of parents and F₁s with three replications were planted in the field of experimental farm at Zhejiang Agricultural University in 1994 and 1995. The seeds were sown on March 28 for 1994 or April 3 for 1995, and single plants of 31-day seedlings were transplanted at spacing of 20 x 20 cm. There were 24 plants in each plot. Seed samples of parents and F₂s from the F₁ plants were derived at maturity from eight plants in the middle part of the plot. The F₁ seeds which were harvested from female parents were obtained by crossing CMS lines to restorer lines at flowering during the growing season. Quantitative traits of rice cooking quality analyzed were amylose content (AC, %), gel consistency (GC, mm) and alkali spreading score (ASS, grade) which is closely related to gelatinization temperature. These traits were measured with three replications for each sample of parents, F₂s and F₂s.

The cooking quality traits studied were mainly controlled by genetic main effects, but also influenced by GE interaction effects, especially for AC and ASS (Table 1). Among the genetic main effects, seed effects and maternal effects were the main effects for AC and ASS, respectively. Cytoplasmic effects were the main effects for GC. Among the GE interaction effects, AC and ASS traits were mainly affected by maternal interaction effects and GC by seed interaction effects. Additive effects and/or additive interaction effects were the main factors of controlling the performance of rice cooking quality traits except for GC which was affected by dominant interaction effects. Since the genetic effects can be partitioned into genetic main effects and GE interaction effects, the heterosis of cooking quality traits in rice can also be partitioned into genetic heterosis and interaction heterosis. The heterosis of cooking quality showed that these traits were influenced by genetic heterosis and GE interaction heterosis especially for amylose content trait in indica rice (Table 2). Cytoplasmic heterosis (0.05), maternal interaction heterosis in 1994 (-0.03) and in 1995 (0.10) were main heterosis components for AC traits, but the genetic heterosis and interaction heterosis in 1995 were mainly from cytoplasmic heterosis and cytoplasmic interaction heterosis for ASS, respectively. Cytoplasmic heterosis and maternal interaction heterosis in 1994 or in 1995 were the main heterosis components for the performance of GC trait in crosses. It was indicated by the predicted genie effects that P7, P8, P12 and P14 were better than other parents for the genetic main effects in improving rice cooking quality traits of progenies. The predicted GE interaction effects showed that the genetic stability of P12 was expected for improving amylose content in different environments.

References

Zhu, J., 1992. Mixed model approaches for estimating genetic variances and covariances. J. Biomath. 7(1): 1-10.

Zhu, J., 1996a. Analysis methods for seed models with genotype-environment interactions. Chin. J. Genet. 23(1): 56-68.

Zhu.J. 1996b. Statistical Methods for Genetic Models. Agricultural Publish House, Beijing, China.