In this study the behavior of classical contrasts in the experiments in subdivided plots with main treatments arranged in incomplete balanced blocks was analyzed. It was considered the existence of constant correlation p between subplots of the same plot and independence between subplots of different plots. The variance of the basic contrasts, for this type of experiments, was carried out as in the BIB tests, as well as in the trials in subdivided plots delineated in randomized complete blocks.
