In conducting a replicated varietal yield trial, the experimenter usually is interested in obtaining the highest yielding variety or varieties from a group of varieties (or strains). A useful kind of information under such circumstances would be knowledge concerning the optimum allocation of replicates and varieties in any single experiment. Yates (23) and Perotti (12) have given a formula for determining the best combination of replicates and varieties in selecting the highest yielding variety or varieties from a simple randomized complete block design. For the more complex designs, formulas for determining the optimum allocation of varieties, replicates and locations (or years) would be useful.