ABSTRACT
The purpose of any experiment is to detect even small differences that exist among the treatments under comparison. This is possible only when the experimental units used in the experiment are homogeneous. Experi mental units such as plots (used in agricultural field experiments), animals, and human beings (used in medical research), etc., show variation among themselves. The degree and pattern of variation depend upon several factors such as soil fertility, stand of the crop, moisture content in the soil, tempera ture, sunshine, microorganisms, alkalinity, salinity, etc., over which experi menters have little or no control. Generally, plots adjacent to each other are more alike than plots that are farther away (Fisher, 1966). Detailed informa tion about the nature and variation of soil can be obtained from uniformity trials for different characters. When such plots are used in experiments, the observed treatment differences are not only attributable to real treatment ef fects but also differences in the experimental units (plots). Similar situa tions occur in animal experiments. Animals exhibit variation in age, sex. initial weight, height, health conditions, breed, genetic constitution, and these factors mask the real effect of the treatment.
