基于玉米区试的籽粒产量抽样方法研究  

Sampling Method on Grain Yield Evaluation Based on Maize Regional Trials

作　　者：濮绍京[1] 金文林[1] 白琼岩 张连平 陈立军 赵波[1] 钟连全[3] 

机构地区：[1]北京农学院作物遗传育种研究所,北京102206 [2]北京市种子管理站,北京100088 [3]北京市昌平区种子管理站,北京102200

出　　处：《作物学报》2008年第6期991-998,共8页Acta Agronomica Sinica

基　　金：北京市科委项目(D0705006040191);北京市属市管高等学校人才强教计划资助项目(PXM2006_014207_021717)

摘　　要：选用夏播玉米京玉11,在北京市昌平区建立面积1209.60m^2、7200株的研究总体,利用金文林、李淼设计制作的“抽样软件”,对已知总体进行针对普通玉米区域试验籽粒产量的抽样方法研究,重点对抽样行数和样本容量进行了探讨。结果表明,以4个样本容量进行1-6行及10行片状随机抽样,样本平均数方差比较显示,3-6行抽样效果优于1-2和10行。500次抽样样本平均数次数分布显示,样本容量为60、120、240和420时,分别采用3、4、5和6行的片状抽样方法较好。结合精确度与工作量两方面考虑,4-6行抽样优于3行抽样;采取工作量较少、精确度较低的抽样策略时,4-6行抽样采用120样本容量较好;采取精确度较高策略时,4、6行抽样采用420样本容量较好,5行抽样采用480样本容量较好。几个方面结果叠加分析认为,精确度要求较高时,应采取6行420样本容量的抽样方法;精确度要求较低时,应采取4行120样本容量的抽样方法。The current design for maize regional trials on the number of rows in plot and plants of yield estimate mostly depends on experiences instead of sufficient, and it was short of statistical foundation. Sampling method for grain yield based on common regional trials was studied in the known population of 7 200 plants in an area of 1 209.60 m^2 with summer maize cultivar Jingyu 11 in Changping District, Beijing through the ＂Sampling Software＂ designed by Jin and Li, and number of sampling rows （lines） and sample size were approached as key points. Sampling with four sorts of sample sizes of 60, 120, 180, and 240 by random piece sample of 1-6 rows and 10 rows, according to the comparative analysis of variance of sample means, sampling method of 3-6 rows was better than that of 1-2 rows and 10 rows. Frequency distribution of the means of 500 samples showed that for different sample sizes, the sampling should choose a different number of rows, and for the sample size of 60, 120, 240, and 420, 3, 4, 5, and 6 rows were the better sampling number of rows respectively. Considering precision and labor load, sampling 4-6 rows was better than sampling 3 rows. In labor-saving strategy with lower precision, sample size of 120 was suggested for sampling 4-6 rows; while in higher precision strategy, sample size of 420 was better for sampling 4 and 6 rows, and that of 480 for sampling 5 rows. Sampling methods with different number of rows and sample sizes were significantly different in frequency distribution of the sample means and the variance of sample means, and the better sampling methods could be distinguished from the others. Comprehensive analyses of the above results indicate that the best methods are sampling 6 rows with 420 plants for high precise trials, and 4 rows with 120 plants for common trials lower in precision.

关 键 词：玉米 抽样方法 区域试验 籽粒产量 

分 类 号：S513[农业科学—作物学] O212.2[理学—概率论与数理统计]