ABSTRACT
In the previous chapters, we came across a number of designs that were constructed to eliminate the effects of extraneous factors, while studying the effects of treatments or factors on some responses. One of the chief instruments used to achieve precision in these designs was blocking. A block is represented by the row or column when Latin and Graeco-Latin square designs are employed. In the randomized complete block designs, we assume that blocks always contained enough experimental units to ensure that all treatments or treatment combinations (if factors set at different levels are used) occur at least once in each block. However, in practice, performing the complete block experiment is impossible under certain circumstances. The early part of this chapter aims to explore some of the techniques used in dealing with such situations. The full and partially balanced lattices as well as designs in which nesting of factors occur are studied in the latter part of this chapter.
