Figure 1.

Linked to the type of randomization, treatment effects may interfere with nontreatment effects. For instance, in a completely randomized experimental block design, all replicates of treatment A1 may lie in the west of the field, whereas those of A2 lie in the East (A). In this case, wind or water flows from a certain direction might cause treatment differences by nontreatment effects. Such randomization might be done with a more balanced arrangement (B). East–west effects can be controlled by blocking (C) with a restriction for each treatment to appear two times in the east and two times in the west.