论文部分内容阅读
多数品种评价性状都运用区组设计，并运用传统的方差分析法进行；然而，区组设计标准分析时常不能恰当地表明空间变异性。最近在空间统计方面的进展表明有更好的择一性。本研究的基本目的是比较具有２个最邻近调整（ＮＮＡ）方法的完全随机区组（ＲＣＢ）分析；作为另一种消除空间变异性的随机田间程序也给予测定。用于比较的田间资料取自于包括各种不同的适应与不适应基因型３个重复的选种圃，于１９８８－１９８９年期间分种子内布拉斯加州的４个点。根据较小的变异系数和对品种差异区分最大能力，ＮＮＡ方法 点所有选育圃最优化的一种。１２个性状中有９个性状的ＲＣＢ品种平均值与ＮＮＡ品种平均值呈高度相关，表明２种程序都能够确定类似的骨干品系。然而，３个试验估计的品种产量及来自ＲＣＢ和ＮＮＡ程序的品种排序是不同的，因而这些不同品系将被选择。田间随机分析用于ＲＣＢ和ＮＮＡ分析间差异最大的试验；这就产生了类似于ＮＮＡ的结果。本研究结果表明，空间趋势在内布拉斯加是常见的，并协调了区组设计标准分析的准确性与精确度。因此，ＮＮＡ或田间随机分析应该适用于改进育种试验的分析。
Most cultivars evaluate traits using block design and using traditional analysis of variance (ANOVA); however, block design standard analyzes often fail to adequately characterize spatial variability. Recent progress in spatial statistics shows that there is a better choice. The basic purpose of this study was to compare complete randomized block (RCB) analysis with two nearest neighbor adjustment (NNA) methods; as another random field program to eliminate spatial variability. Field data for comparison was taken from three replicate selection plots that included a variety of different adapted and unacceptable genotypes, with seeds distributed within 4 points of Nebraska during the period 1988-1989. According to the smaller coefficient of variation and the ability to differentiate between varieties, the NNA method selects one of the most optimized of all breeding options. The average of RCB cultivars with 9 traits in 12 traits was highly correlated with the average of NNA cultivars, indicating that both programs were able to identify similar backbone lines. However, the yield estimates for the three trials and the order of the breeds from the RCB and NNA programs were different, so these different lines will be selected. A field-based random analysis was used to test for the most significant difference between RCB and NNA analyzes; this produced a NNA-like result. The results of this study show that spatial trends are common in Nebraska and that the accuracy and precision of the block design standard analysis are coordinated. Therefore, NNA or field stochastic analysis should be applied to improve the analysis of breeding trials.