Statistical aspects of aquaculture research: one and two season crossover designs

One-and two-season cross-over designs are compared with completely randomized designs for their potential use in pond aquaculture experiments. Cross-over designs allow treatment comparisons to be made within the same pond, removing much of the between-pond variability. Different treatments are applied consecutively to the same pond. The one-season design involves splitting the season into two periods and applying different treatments in each period. The fish growth in each of the periods is compared taking account of both pond and period differences. The two-season design compares the harvest weight or growth over a whole season in consecutive seasons. Each of the designs is explained further. Data from uniformity trials are used to make these comparisons. Trials on the use of buffalo manure supplemented with urea and triple sulphate, which were completed in consecutive seasons, are used to verify the conclusions. The advantages and disadvantages of each design are discussed.