【作者】 高金锋；

【导师】 冯佰利；

【作者基本信息】 西北农林科技大学 ， 作物栽培学与耕作学， 2008， 硕士

【摘要】 荞麦主要分布于西南、西北等老少边贫地区,是这些地区的主要粮食作物和经济作物。由于荞麦是无限开花结实习性,产量水平较低,加之荞麦生长环境远比水稻、小麦、玉米等大宗作物复杂,受地理、地形、降雨及其时空分布的制约大,荞麦品种产量的基因型×环境互作效应的动态性变化显著,品种表现型受环境因子作用使产量波动很大。因此,加强荞麦新品种的选育和推广对促进荞麦产区经济发展和农民增收具有重要的指导意义。但以往的荞麦新品种鉴定主要以产量作为鉴定的指标,造成荞麦新品种(系)的稳定性差或广泛适应性弱,导致了荞麦新品种的推广和种植面积较小,极大限制了荞麦生产的发展,探索荞麦品种鉴定和评价的方法和标准就成为品种区试鉴定的重要课题。本研究通过对荞麦产量及其构成因素与主要农艺性状间的相关分析及通径分析,探讨影响荞麦产量的因素。同时,从统计理论的角度,采用AMMI模型等5种常用的统计分析方法和模型对2003-2005年国家荞麦区域试验结果进行分析,研究不同基因型荞麦品种的丰产性、稳产性及适应性,得出以下主要结论:1.对荞麦产量及其主要农艺性状进行了相关分析结果表明,单株粒重、千粒重与产量呈显著正相关关系,表明千粒重和单株粒重的增加可显著提高荞麦的产量。通径分析结果表明,千粒重、单株粒重、主茎节数、主茎分枝对产量的直接通径系数为正值,且千粒重和单株粒重可通过多个其它农艺性状对产量形成间接的正面作用来提高荞麦产量,故在荞麦生产过程中应十分注重对千粒重和单株粒重的利用;株高和生育日数对产量的直接通径系数为负值,这可能与荞麦的无限生长习性和后期落粒现象有关。2.Franics和Kannenberg模型以品种平均产量和变异系数作为评价品种丰产性、稳定性及适应性的主要指标。该模型简单明了、直观且计算简便,应用广泛;但该法忽视了品种与环境之间的互作效应。3.高稳系数法以高稳系数( HSC i)为主要评价指标分析品种高产性、稳产性及适应性。该法简单、易操作,实用性广泛;该法只考虑了环境因素和遗传因素,忽略了二者之间的互作效应。4.秩次分析法利用不同环境下获得的资料进行常规的单因素方差分析和多重比较,判断品种间有无显著差异,并计算出各品种在不同环境下的分级值( H1 Mi)及秩次值( H2 Mi),进而计算秩次平均值( H2 i)和秩次均方( S i2)来评价品种丰产性、稳定性及适应性。该法方法简洁,含义明确;但该法在进行秩次分析删除无效试点的过程中,也删除了其真实存在的品种产量水平,有可能导致结果的不确定性。5.Shukla模型以Shukla方差的F测验结果是否显著及Shukla变异系数来评价品种稳定性及适应性。该模型方法简洁,含义明确,统计性能较好,信息量大且较为精确;但该法由于互作值受限于总和为零,Shukla方差相互不独立,同质性检验及两两差异性比较困难。6.AMMI模型以品种在IPCA(主成分特征值)空间至原点的距离Di和AMMI双标图中品种图标距横坐标轴或坐标原点的距离来评价品种稳定性及适应性。该模型将方差分析和主成分分析有机地结合在一起,兼具了这两种方法的优点,较好地解释了基因与环境的互作关系,并借助于AMMI双标图,能直观、具体、有效地鉴别品种的适应性、稳定性和丰产性,且有利于区分各试点的分辨力或判别力。但该法在剔除互作中属于误差的一些成分时,可能也剔除了真实存在的互作成分,对于年份与品种的互作效应不能得到较好的解释。7.在进行荞麦稳定性及适应性分析与评价时,根据不同环境,将几种方法和模型结合起来应用可能对荞麦的评价更具有科学性。一般地,首先用Franics和Kannenberg模型和高稳系数法进行分析,鉴别其丰产性和稳定性;然后在进行Shukla模型对品种稳定性及广泛适应性作出准确判断,寻找出高产而具有广泛适应性的优良品种;若要提高品种在各试点估值的精度,可进一步采用加性主效互作可乘(AMMI)模型进行分析,并辅以双标图和品种-试点均值图,寻找出在较大区域范围内具有高产且稳定性和适应性较好的品种。总之,在对荞麦进行分析中,要坚持在高产前提下的稳定性及适应性,同时要注意理解用各种方法得到的稳定性参数的实际含义及其相互间的区别与联系,要以各品种在各环境中的具体分析为依据作出符合客观实际的高产性、稳定性和适应性的综合评价。8.对2003-2005年国家荞麦品种区域试验资料进行综合分析评价,结果表明,平荞2号稳定性、适应性都比较好;品种六荞1号稳定性、适应性均表现较差。更多还原

【Abstract】 Buckwheat, origined from China, is distributed throughout the country, maily as a staple food in remote and poverty-stricken southwest and northwest China. Buckwheat is featured with infinite flowering and seed-setting character, hence low yield is obtained, moreover, buckwheat is grown in complicated environment compared with rice, wheat, maize and other major crops, and its performance restricted by geography, terrain, rainfall and temporal and spatial distribution, therefore, the interaction between genotype and environment is significantly dynamic, and yield fluctuation is observable, leading to low yield stability and weak local adaptability, which impedes development of buckwheat production.This study tried to find factors restricting buckwheat yield by analysis of correlation between yield and constitution elements and agronomic traits and path analysis. By employing statistical methods such as AMMI and other models, materials adpted in national buckwheat cultivar regional trials in 2003-2005 was slso analyzed to research on the productivity, stability and adatability of different genotypes. It got the following results:1.Correlation analysis between yield and agronomic traits showed that there was postitive relationship between grain weight per plant, 1000-kernels weight and yield, indicating improvement of the two agronomic characters could increase the yield. Path analysis showed direct path coefficient of 1000-kernels weight, grain weight per plant, node numbers of main stem, branch numbers of main stem to yield was positive respectively, and grain weight per plant, 1000-kernels weight, the two characters could take use of indirect positive role of other agronomic characters to improve yield, so it was important to take advantage of these two characters in buckwheat production; direct path coefficient of plant height, growth period to yield was negative respectively, which maybe resulted from the infinite growing habit and seed shattering character at the late stage.2.In Franics and Kannenberg model, average yield and variation coefficient, these two indexes were chiefly used to assess the productivity, stability and adaptability of varieties. The model had the advantage of simpilicity, easy calculation and popularity, but had the disadvantage of neglect of the interaction between variety and environment.3.In high stability coefficient method, HSC i was the mainly assessment index. The method was simple, easy to operate and powerfully practical, but just taking the genetic factor and environment factor into account individually and neglect of the interaction of them.4.In rank analysis method, one-way ANOVA and multiple comparison was conducted by using materials obtained in different environments, showing the significance between varieties, and then H1 Mi and H2 Mi, H2 i and S i2were computed to assess the productivity, stability, and adaptability of varieties. This method was simple and had clear meaning, but deletion of the invalid experimental site meant the neglect of variety yield at this site, possibly resulting in the uncertainty of the results.5.In Shukla model, the significance of F test and Shukla variation coefficient was used for assessment. This model was featured with simpilicity, clear meaning, good statistical performance, abundant information and accuracy, howerever, it was difficult to conduct homogeneity test and difference comparison, since the sum of interaction valueshould be null and the Shukla differences were not mutully independent.6.In AMMI model, Di and the distance between variety label and abscissa axis (or coordinate origin) in AMMI biplot was used to assess the stability and adptability of variety. AMMI Method owned the advantages of both variance analysis and principal components analysis, capable of showing the interaction, furthermore, by means of bioplot, the productivity, stability and adptability of variety could be evaluated directly, completely and effectively, and discriminating ability of each single site could be shown. However, when the elements of interaction error were eliminated, the true elements of interaction might also be eliminated, hence the intraction effect between year and variety could not be showed very well.7.When the stability and adaptability of buckwheat variety was evaluated, it was advised that integration of the above-mentioned methods and models was more desirable. Generally, the Franics and Kannenberg model and high stability coefficient method was primarily used to evaluate the productivity and stability, and then Shukla model was employed to evaluate the stability and adaptility of buckwheat accurately, searching for the desirable variety with high yield and wide adaptility. If the accuracy of variety estimation in each single site needs to be improved, the AMMI model could be adopted, together with bioplot and mean value chart for variety and site, the variety with high-yield, stability and adaptility in a large-scale area could be explored. In short, the stability and adaptility of buckwheat variety should be based on its productivity when analysis was conducted, and the stability indexes obtained using different methods should be understood correctly, and the assessment should be made on the basis of concrete analysis of varietity performance in different environment.8.By analysis of the materials about variety regional test in 2003-2005, it was showed that the control Pingqiao 2 performed better in stability and adaptility, and Liuqiao 1 perfromed poorly.更多还原