摘要
为探讨套作复光后大豆补偿生长能力的评价方法,比较不同大豆品系补偿生长能力差异,本研究以136个大豆品系为材料,在田间设置2个处理,即不遮荫和遮荫+光照恢复。于取消遮阳网当日和取消后30d共测定了14项形态指标,成熟期调查每个大豆品种(系)的单株产量。根据遮荫+光照恢复处理和不遮荫处理计算各品种各单项指标的补偿系数,采用主成分分析、隶属函数法、聚类分析和逐步回归分析方法,对大豆的补偿生长能力进行综合评价。结果表明,主成分分析将14个单项指标转换为4个相互独立的综合指标;通过隶属函数计算补偿能力综合得分(D值),并对其进行聚类分析,将136份大豆品种按照补偿生长能力强弱分为5大类群,其中第Ⅵ类群(10个材料)补偿生长能力较强,产量较高。通过逐步回归建立套作大豆补偿生长能力评价数学模型,D=-1.314+1.110 X_4+0.623 X_7+0.831 X_1+0.642 X_9,R^2=0.990,并筛选出光照恢复后地上部总干重、叶面积、SPAD、中间节粗作为套作大豆补偿生长能力的主要鉴定指标。
To explore the methods on evaluating and analyzing compensation growth ability,and to compare difference among soybean cultivars in light recovery period in relay intercropping,136 soybean cultivars（ lines）were planted in a field experiment under natural light condition and shade ＋ light recovery condition. 14 morphology traits were investigated on 30 days after light recovery,the yield of soybean per plant was investigated at the mature period. Principal component analysis,membership function method,cluster analysis and stepwise regression analysis were used to comprehensively evaluate the compensation growth ability based on compensation coefficient of all indexes calculated from both treatments. Results showed that 4 independent comprehensive components were extracted from 14 single indexes by principal component analysis. 136 soybean cultivars（ lines） were divided into 5groups,the compensation growth ability of the Ⅵ group（ 10 cultivars） showed the strongest compensation growth ability and the highest yield. The mathematical evaluation model for soybean compensation growth ability was established as D =- 1. 314 ＋ 1. 110 X4＋ 0. 623 X7＋ 0. 831 X1＋ 0. 642 X9,（ R^2= 0. 990）. Based on the model,aboveground dry weight,leaf area,SPAD,and middle section diameter after light recovery could be used for iden-tification index of soybean varieties to compensation growth ability.
出处
《中国油料作物学报》
CAS
CSCD
北大核心
2016年第4期443-451,共9页
Chinese Journal of Oil Crop Sciences
基金
国家自然科学基金(31171476)
现代农业产业技术体系建设专项(CARS-04-PS19)
关键词
大豆
套作
补偿生长
主成分分析
隶属函数法
聚类分析
逐步回归
Soybean
Relay strip intercropping
Compensation growth
Principal components analysis
Mem bership function method
Hierarchical cluster analysis
Stepwise regression analysis method
