Date of Award:


Document Type:


Degree Name:

Doctor of Philosophy (PhD)


Plants, Soils, and Climate

Department name when degree awarded

Soil Science and Biometeorology


R. J. Hanks


Soil moisture and nitrogen are often the primary controlling factors of crop production . Production functions are valuable in describing crop responses to such controlling factors. Such functions are al so needed for determining the economic reliability of crop production.

Soil moisture and nitrogen fertilizer were maintained at various levels to determine the response of field corn to these controlling factors. A conventional split plot design, a continuous function experimental design, and a confined barrel plot design were used to obtain data for determining production functions.

The complete factorial, conventional split plot data was successfully used to generate reliable production functions in two locations. The multiple correlation coefficient was approximately 0. 68. The continuous function design consisted of a large number of treatment combinations with only small sequential increments in treatment levels , from plot to plot. This made it possible to eliminate border rows and use a much smaller plot size. The design was tested in two locations, using 10 soil moisture levels and 22 nitrogen fertilizer levels. The multiple correlation coefficients were about 0.33 and 0.49. Using this design, it was necessary to replicate entire blocks.

The production functions from the conventional split plot design and continuous variable plot design compared favorably within locations, but not between locations. Production functions were also determined for relative yields and treatments . The inputs were expressed as fractions relative to both the maximum and optimum levels. The relative production functions were similar for both locations. The production functions were tested for transferability by making adjustments in the input data to account for site specific controlling factors. No definite conclusions were possible although the results suggest that transferability is possible.

Several reduced field designs were also tested using the data obtained. Three different incomplete factorial, split plot designs appear to be suitable for generating production functions. Grouping of the continuous function plots to as few as four soil moisture levels by four fertilizer levels also gave reliable production functions. Several suggestions regarding the use of reduced field designs are presented along with suggested future research needs in this area . The data indicate that the continuous variable design is quite useful for determining production function s. The barrel plots consisted of field corn grown under various treatment combinations on undisturbed soil cores. The data obtained from these barrel plots (210 liters) were too variable to generate a reliable production function.

Included in

Soil Science Commons