Abstract: Three crosses derived from parents with distinctly different panicle traits, including Datoudao × Daichangqing(crossⅠ), Laolaihong × Yanjing 2(crossⅡ) and Daichangqing ×Shanghaiqing(cross Ⅲ), and their derived generations, were evaluated to analyze the genetics of yield-related traits of japonica rice landrace from Taihu Lake Region by the method of major gene plus polygene mixed inheritance model. The optimum model for total grain number per panicle in crossⅠ was one major gene plus additive-dominant polygene mixed inheritance model, while that for crossesⅡ and Ⅲ was one complete dominant major gene plus polygene mixed inheritance model. The optimum models for effective panicle number per plant in crossesⅡand Ⅲ were one major gene plus polygene, while that in cross I was two major genes. The genetic models for 1000-grain weight in crossesⅡandⅠ were two major genes plus polygene mixed inheritance model, and that in cross Ⅲ was one major gene plus polygene. The optimum model for filled grain number per panicle was two major genes model without polygene. It was suggested that segregation analysis of quantitative traits among multiple generations of P 1 , P 2 , F 1 , B 1 , B 2 and F 2 offered a higher accuracy and could identify polygene than that of single generation. Based on the genetic information provided by joint analysis of multiple generations, breeding strategies for different crosses and traits were discussed.

Key words: gene, inheritance model, japonica rice, landrace

摘要： 采用主基因+多基因混合遗传模型，对亲本穗部性状差异较大的3个杂交组合，大头稻/呆长青（组合Ⅰ）、老来红/盐粳2号（组合Ⅱ）和呆长青/上海青（组合Ⅲ）的后裔世代的产量相关性状进行了遗传分离分析，得到了这些性状的最适遗传模型。结果表明：组合Ⅰ每穗总粒数的最适遗传模型为一对主基因+加性显性多基因混合模型，而组合Ⅱ、Ⅲ为一对完全显性主基因+加性显性多基因遗传模型；组合Ⅱ、Ⅲ的单株有效穗数受一对主基因控制，组合Ⅰ则受两对主基因控制；组合Ⅰ、Ⅱ千粒重的遗传模型为两对主基因+多基因模型，组合Ⅲ为一对主基因+多基因遗传模型；每穗实粒数为两对主基因遗传模型。选用P1、P2、F1、B1、B2、F2六世代联合分离分析方法，相比于单个分离世代的分析方法，增加了试验的精确度，保证了分析结果的准确性，并可鉴别多基因的存在。根据试验结果，分析了不同性状、不同组合的育种策略。

关键词: 产量相关性状, 基因, 遗传模型, 粳稻, 地方品种