The factorial combination of treatments in rotation experiments

The paper deals with points arising in the design of factorial schemes for incorporating fertilizer and other treatments in rotation experiments. These schemes make extensive use of fractional replication and confounding, including split-plot confounding. They are complicated by the need to apply some treatments cumulatively and others with levels changed between plots in successive cycles so as to minimize errors due to constant plot differences. Design keys are used to identify the treatment main effects and rotation differences with particular plot contrasts. The keys exhibit the structure of a design more compactly and clearly than a complete enumeration of identities and confounded contrasts and enable treatment combinations to be generated systematically.