扩大小区面积与增加重复次数,都能降低試驗机誤。但在一个处理总面积相等的情况下,扩大小区面积不如增加重复次数更为合适。面积小的小区以采用方形合适,面积大的小区以采用长方形合适。作者計算了小麦試驗中十一种合适的小区面积与重复的配合方式,其机誤均不超过10％。 Expand the area of ​​the plot and increase the number of repetitions, can reduce the test machine error. However, in a situation where the total area under treatment is equal, it is more appropriate to expand the size of the area than to increase the number of repetitions. Small area of ​​the cell to adopt a square suitable, large area to use rectangular suitable. The authors calculated eleven suitable cell sizes and replicates in a wheat trial with machine errors of no more than 10%.