油棕鲜果穗产量构成因素的相关性和回归分析

Correlation and regression analysis of fresh fruit bunch yield components in oil palm(Elaeis guineensis)

关键词:  油棕, 鲜果穗产量, 相关性分析, 回归分析, 通径分析

Key words:  Oil palm(Elaeis guineensis), fresh fruit bunch yield, correlation analysis, regression analysis, path analysis

摘要:

为明确油棕鲜果穗产量构成因素之间的关系,利用相关性、回归和通径等分析方法,对58株油棕单株鲜果穗产量与其它性状相关性进行了研究。结果表明:油棕鲜果穗产量与鲜果穗重和每株果穗数的相关性最为密切。另外,在油棕高产品种选育过程中,应着重对平均果穗重、每株果穗数、果实大小和含油量等性状进行选择,并综合权衡这些产量性状之间的相互作用。产量性状变异和相关性分析表明,平均种壳厚度变异程度最大(88.34%),平均果长变异最小(11.99%),鲜果穗产量与鲜果穗重、每株果穗数、平均果重呈极显著正相关关系(0.882**,0.714**,0.439**),与平均果长呈显著正相关关系(0.334*)。多元线性回归和通径分析结果表明,通过比较多元线性回归方程的相关参数,建立了可实用的最优多元线性回归方程x1=-281.511+2.787x4+3.455x5-2.672x6+0.411x7+6.690x8+3.664x9,平均鲜果穗重和每株果穗数对单株鲜果穗产量的直接作用最大(0.670,0.439),其它性状对单株产量的直接作用较小,平均果重和平均果长对单株鲜果穗产量的间接作用最大(0.335,0.322)。该研究结果为油棕高产品种选育提供了依据。

Abstract:

Oil palm is the most important oil crop all over the world, its production is 30% of the world's total vegetable oil, but there is still room for improving the yield. In order to verify relationship among fresh fruit bunch yield components, correlation between fresh fruit bunch yield and other traits were studied by using correlation, regression and path analysis. The results indicated that fresh fruit weight, number of bunch were the most closely related with fresh fruit bunch yield. Moreover, breeder should focused on average weight of fresh fruit bunch, number of fresh fruit bunch in each plant, fruit length, fruit width and oil content of mesocarp, and balance interaction among these characters was related with fresh fruit bunch yield in the high yield oil palm varieties breeding. Analysis of variance and correlation showed that the average shell thickness had the most coefficient of variation(CV=88.34%), average fruit length had the least coefficient of variation(CV=11.99%), fresh fruit bunch yield had extremely significant positive correlation with average fresh fruit bunch weight, number of bunch produced, average fruit weight(r=0.882**, 0.714**, 0.439**), and had significant positive correlation with average fruit length(0.334*). Multiple regression and path analysis showed that practical regression equation x1=-281.511+2.787x4+3.455x5-2.672x6+0.411x7+6.690x8+ 3.664x9 was established by comparing parameters related with multiple linear regression equation. Average fresh fruit bunch weight and number of bunch had maximum direct effects(0.670, 0.439), other characters were relatively less, average fruit weight and average fruit length had maximum indirect effects(0.335, 0.322). The results provide information for the high yield varieties breeding.