【摘要】 采用因子分析和逐步回归的方法对佳木斯地区１９８１－１９９４年间大豆灰斑病发生程度和相应年份的相关气象资料进行整理分析，筛选出５个非特殊灾变年份大豆灰斑病流行趋势预测初始数学模型。在８０％置信区间内，模型历史回拟率分别达９２．９％—１００％。对田间试验定点调查所得的大豆不同抗性品种的病情与大豆产量损失的数据资料进行相关分析，结果表明，抗病品种和感病品种的病叶率、病荚率、病粒率与大豆产量损失率间均呈直线正相关，相关系数均达极显著水平。更多还原

【Abstract】 Through analysing the relationship between epidemical tendency of soybean frogeye leaf spot and the relative meteorological information from 1981-1994 in Jiamusi region, Heilongjiang province, by using factorial analysis and stepwise regression, the primary mathematic models of forecast for epidemical tendency of soybean frogeye leaf spot for nonparticular calamity year were established. Their historical fitness are 92.9%-100% respectively. By analysing the data of yield loss and disease, the result showed that there were positive linear correlations, between yield loss of soybean with disease severity of leaf, pot, and grains in cultivars both resistant and susceptible.更多还原