改良拉丁方设计及其在油菜品种试验中的应用

Modified latin square design and it were practised on rape cultivar heterogeneity experiment

针对一些常规试验设计存在的不足，在传统拉丁方设计的基础上提出改良拉丁方设计，这种设计试验次数少，能从纵横两个方向消除土壤差异，试验数据可沿用传统拉丁方设计的统计分析方法，具有较高的试验精确度。

In view of the situation that some defects existed in some conventional experiment designs, a modified latin square design (MLS) was put forward on the basis of conventional latin square design (CLS). The MLS design, in which the incomplete blocks were laid out in latin square, eliminated heterogeneity of soil fertility in two way and could be tested significances of experimental results by using the statistical analysis method of CLS design, with smaller experiment number than CLS design and higher experiment precision than randomized complete block design.