In an earlier paper (?) reference was made to the fact that no piece of land, however carefully selected for experimental purposes, is absolutely uniform. If the area is divided into equal sections and the sections sub-divided into plots of the same size, the yields from similarly treated duplicate plots planted on the same day with selected homogeneous seed will rarely be the same, and not infrequently will be widely divergent. For example, the relative percentage differences in yields?i.e.