THE PRECISION OF FIELD EXPERIMENTS WITH VEGETABLE CROPS AS INFLUENCED BY PLOT AND BLOCK SIZE AND SHAPE: I. SWEET CORN
In 1956 weights of marketable ears of hybrid sweet corn from three irrigated uniformity trials were recorded to determine the effect of varying plot and block sizes and shapes on yield variability. The coefficients of variation decreased with increased plot size. For block shapes of 1 × 6 plots (1 plot long and 6 plots wide) and block shapes of 3 × 2 plots (3 plots long and 2 plots wide) widening the plots was more effective in reducing variation than lengthening them. However, for block shapes of 6 × 1 plots and block shapes of 2 × 3 plots lengthening the plots was most effective. Estimates of cost were obtained showing the percentage of the total cost that was proportional to the number of plots used and the percentage that was proportional to the total area used per treatment. Optimum plot size for yield and quality studies was found to be one of two basic units (10 ft. by 6 ft., or 20 ft. by 3 ft.) depending upon cost assumed and the value of the regression coefficient. Smith's method for determining optimum size of plot agreed with the coefficients of variation method as to optimum size of plot.