摘要： 通过对花椰菜亲本主要经济性状的多元逐步回归分析, 建立杂交一代单球重和成熟期的预测模型.结果表明, 花椰菜杂交一代单球重预测回归模型为B6-1=0.5010-0.0098B3_-2+0.1772B_-2-0.081B12_-2+0.0760C3_-2+0.0503C5_-2-0.0316C6_-2+0.0377C7_-2+0.193C8_-2+0.0086B13-3-0.0287C6-3-0.0256C1-3-0.0318C17-3+0.0391C21-3, 母本的全株重(B_-2)和父本的成熟期(B13-3)对模型贡献最大, 均达极显著水平;花椰菜杂交一代成熟期预测回归模型为B13-1=-12.3706+0.2367B1_-2-0.1005B7_-2-0.3056B8_-2+0.529B13_-2-1.5787C7_-2+1.9000C13_-2-.8968B-3+0.216B5-3+10.3628B6-3+0.7588B15-3_-2.6718C7-3+11.9065C8-3+1.1356C9-3+1.1995C13-3-1.329C18-3-0.9262C20-3, 母本的成熟期(B13_-2)和父本的显球期(B15-3)是影响杂一代成熟期的最主要因子.经田间验证, 杂交一代单球重预测准确率达80.3%, 杂交一代成熟期预测准确率69.5%, 模型预测值与田间实测值相关系数为:单球重r=0.5931, 成熟期r=0.7237, 均达极显著正相关关系.经对回归模型方差分析, F值均达到0.0001水准, 证明回归方程对样本观测值拟合度较高, 可用做杂一代单球重和成熟期性状预测.

关键词: 花椰菜, 育种, 回归模型, 单球重, 成熟期

Abstract: The predictive models for the single head weight and maturity date in F₁ hybrids were established based on multiple regression analysis for their parents'major economic characters of cauliflower. The results showed that predictive regression model for the single head weight of ₁ hybrid is B6 - 1 = 015010 - 010098 B3 _-2 + 011772 B4 _-2 - 010481 B12 _-2 + 010760 C3 -2 + 010503 C5 _-2 - 010316 C6 _-2 + 010377 C7 _-2 + 011943 C8 _-2 + 010086 B13 _-3 - 010287 C6 _-3 - 010256 C14 _-3 - 010318 C17 _-3 + 010391 C21 _-31The whole plant weight of female parent ( B4 _-2) and maturation period of male parent ( B13 _-3) made great cont ribution, and reached highly significant difference. The predictive regression model for maturity date of F₁ hybrid is B13 - 1 = - 1213706 + 012367 B1 _-2 - 011005 B7 _-2 - 013056 B8 _-2 + 014529 B13 _-2 - 115787 C7 _-2 + 119000 C13 _-2 - 418968 B4 _-3 + 012146 B5 _-3 + 1013628 B6 _-3 + 017588 B15 _-3 _-216718 C7 _-3 + 1119065 C8 _-3 + 111356 C9 _-3 + 111995 C13 _-3 - 113294 C18 _-3 - 019262 C20 _-3. Maturation period of female parent ( B13 _-2) and beginning of heading stage ( B15 _-3) are the main factors that impact on maturity date of₁ hybrid. The results from field experiment had proved that the accuracy for predicting the single head weight and maturity date of ₁ hybrid are up to 80.3% and 69.5% respectively. The correlation coefficient of the single head weight and maturity date between values from predictive model and field experiment are 0. 5931 and0.7237 respectively. Regression models appear greatly significant differences (p = 0. 000 1). The high identical degree of regression equations with observed values, therefore, suggests that these regression models can be used to predict the single head weight and maturity date in ₁ hybrid of cauliflower.

Key words: Cauliflower, Breeding, Regression model, Single head weight, Maturity date

中图分类号: 

• S635.3