To develop hybrid rice breeding on a commercial basis, it is essential to have an efficient seed-production system. One realistic way to increase the amount of seed production per unit field is to genetically improve the floral traits in a female parent, which is expected to make them more acceptable to pollens dispersions from male parents. An indica variety, IR24, expresses relatively higher frequency of stigma-exserted glumes compared with that of the typical Japanese cultivars. We here report the initial trial of the identification of the quantitative trait loci (QTL) controlling stigma exsertion that may be a useful indicator of seed production improvement.

Seventy-one lines of the recombinant inbred derived from the cross between Asominori and IR24 (Tsunematsu et al., 1997) were cultivated in the field in natural season. At 10-14 days after first panicle emergence, three panicles from each of the ten plants in each line were taken and served to trait evaluation. The frequency of stigma exsertion was defined as the ratio of the number of glumes which closed palea and lemma but still retained stigma(s) outside to the total number of the flowered glumes. In practical evaluation at the paddy, the frequency was scored according the ten grade system by careful observation. The representative scores of the lines for QTL analysis were calculated by the average of the grades in each line.

The frequency distribution of the score of stigma exsertion (Fig. 1) indicates that both parents have several chromosomal regions increasing the frequency of stigma exsertion. The result of the QTL analysis by qGene 2.26 (Nelson 1997) is summarized in Table 1. Two QTL increasing the frequency at Asominori allele and seven QTL increasing at IR24 allele were estimated with a threshold F value > 5.0. The total QTL explain about 60% of the total phenotypic variance, which suggests that stigma exsertion is the trait which could be genetically improved. The largest three QTL increasing the score at IR24 allele are located in the vicinities of R1468B, R1002 and C1468 on chromosome 3 and contribute to stigma exsertion equally. On the other hand, one prominent QTL increasing the score at the Asominori allele in the vicinity of XNpb331 on chromosome 4 is estimated, even though Asominori itself does not

express high stigma exsertion.

We have consequently developed backcrossed lines which possess some of the combinations of the candidate regions under the Japonica genetic background. A detailed analysis is underway to confirm the accuracy of the gene actions of the QTLs detected here.

Acknowledgment

The authors thank Dr. A. Yoshimura, Kyushu University, for providing the experimental materials.

References

Tsunematsu, H., A. Yoshimura, Y. Harushima, Y. Nagamura, N. Kurata, M. Yano, T. Sasaki and N. Iwata, 1997. RFLP framework map using recombinant inbred lines in rice. Breeding Science 46: 279-284.

Nelson, JC, 1997. QGENE: software for marker-based genomic analysis and breeding. Mol Breed 3: 239-245.