Direct-seeding method on dry paddy soil could make rice farming large-scale cultivation and cost-saving, but it has still some problems on poor seedling establishment caused by low temperature and varying sowing depth. This research was carried out to identify weedy rice genetic resources that may help to poor seedling establishment in direct-seeded rice on dry paddy soil in temperate region. The genetic resources screened in this study were 128 accessions that consist of 92 japonica weedy rices, 24 indica weedy rices, and 12 Korean bred cultivars.The weedy rices on average have very superior abilities to emerge form depth than cultivated rice. The germplasm showed that the coleoptile or/and mesocotyl lengths were positively correlated with emergence rate to a high degree. Among the germplasm, WD-3, a japonica weedy rice, evidenced the highest level of emergence with the longest coleoptile and mesocotyl elongation. The emergence ability of WD-3 depending on the burial depth under low temperature condition was confirmed in both the phytotron and the field conditions. WD-3 showed 100% of emergence rate until a burial deptho f 7 cm in the dry soil in the phytotron, and 76.2 % of very high emergence rate in the total layer of paddy field where the seeds were sown from the surface to 10 cm depth. In the emerged plants in the field, the mesocotyl elongation increased with increasing burial depth in a logarithmic fashion, and the coleoptile extension increased exponentially. The total elongation lengths of the mesocotyl and coleoptile were similar with a plant burial depth, which indicated that they could induce the safe emergence of the main leaf of the seedling from the soil to thesurface. Conclusively, the elongation habit of mesocotyl and coleoptile of WD-3 could be one of the important characteristics to develop direct-seeding cultivars.

This experiment was carried out to investigate the relationship between grain shapes and ecotypes in Korean native weedy rice. The 111 lines of Korean weedy rice were classified into japonica (91 lines) and indica (20 lines) by RAPD analysis, which was reported in 2005 (Chung and Ahn). The characteristics of grain shape such as grain length, width, thickness, 1,000 grain weight, the ratio of length to width (L/W), the ratio of length to thickness (L/T), and the ratio of width to thickness (W/T) were measured in each ecotype, and their statistical distributions were compared by t test. The mentioned 7 characteristics were significantly different between japonica and indica weedy rice groups, but these could not classify weedy rice accessions into two groups without exceptions according to the classification by RAPD analysis. Out of variables concerning grain shape, the L/W was the highest distinctive to separate ecotypes. A critical point of L/W to divide ecotypes was calculated from the mid point between upper critical limit of japonica and lower critical limit of indica in confidence interval at α=0.05. At critical point of L/W = 2.138, decision error was 2.2% in japonica group, and 5.0% in indica group. Conclusively, the classification of ecotypes by L/W of grain made almost (under 5% error) same result with that by RAPD markers in Korean native weedy rice.