Continuous variable (CVD) and randomized block, split plot (RBSD) designs were used to produce data from which production functions were developed relating corn yield to soil water and nitrogen fertilizer. Data were collected at Logan, Utah in 1972 and Farmington, Utah in 1973.The CVD water and nitrogen treatments, respectively, were sequential or continuous and not randomized. The design is compact but has some statistical uncertainty. The CVD had 7 or 8 soil water (W) levels and 22 or 24 nitrogen (N) levels compared to 4 W and 5 N levels for RBSD. The area used for the RBSD was 3 or 4× that used for the CVD.The data analyzed for the CVD led to the same conclusions as the analysis of the RBSD. The production function (regression equation) at Logan was different from that at Farmington except when comparisons were made on a relative basis.