This study is based on several long-term experiments of crop fertilization at three Iowa locations: Howard County, with Clyde and Cresco soils; Hancock County, with acid and calcareous Webster soils; and Wayne County, with Seymour soil. The trials at these widely separated experimental farms included corn and oats fertilized in a 3-year rotation of corn-oats-meadow. The periods of the trials were: Clyde and Cresco soils, 1945-1960; Webster soils, 1954-1960; Seymour soil, 1949-60. Although meadow was not fertilized, residual nutrients from fertilization of the oats nurse crop were expected to affect hay yield. Applied nutrients included only phosphorus and potassium.

The objectives of the analysis were: (a) to estimate annual production functions for each crop and compare them with average production functions estimated from the several years of data for the same crops, locations and soil types; (b) to analyze the variability or degree of uncertainty involved in such physical and economic relationships as isoquants, isoclines and profit-maximizing nutrient inputs; (c) to estimate weather indexes and their quantitative relationship to fertilizer response; (d) to estimate generalized production functions that incorporate weather, soil nutrients, location and soil into the production function along with the quantities of K and P applied annually.