Degree Type


Date of Award


Degree Name

Doctor of Philosophy




To determine the value of near infrared reflectance (NIR) measurement of corn nutrient composition to livestock feeders, sequential decision models of beef cattle and swine feeding were developed. Corn composition data from the National Feedstuff Composition Data Bank (NFCDB) were used to derive probability density functions for crude protein, fat, and lysine content. Expected values of profit and other variables were computed by numerical integration. Means of crude protein and lysine of the NFCDB sample were lower and higher, respectively, than National Research Council estimates. These differences caused errors in diet formulations which resulted in losses due to imperfect information of up to 47 per year per head of one-time capacity in feeding yearling steers, up to 41 per year per head of one-time capacity in feeding steer calves and up to 6.50 per year per head of one-time capacity in feeding swine;The difference between the expected values of NIR-generated information and of NFCDB sample means was very small. One-time capacities of approximately 40,000 yearling cattle, 12,500 calves of 2,200 swine were necessary to justify investment in an analyzer based on net present value when NIR-generated information (versus NFCDB sample means) was used to formulate diets to meet recommended nutrient allowances. Minimum one-time capacities declined to 2,550 yearling cattle, 1,700 calves or 1,080 swine when diets were formulated to meet nutrient requirements;NIR-generated information was more valuable to yearling steer feeders than to steer calf feeders on an annual basis, given continuous production. This result was due to yearling steers' requiring fewer days to reach market weight thereby allowing more animals to be fed per year;Losses due to imperfect knowledge in beef cattle feeding were largest when corn had low crude protein content and high fat content. Losses incurred by swine feeders were highest when crude protein content was high and fat content was low;The use of NIR results to reject corn with low crude protein and/or fat content increased the expected value of NIR-generated information by roughly ten percent for both species. This gain in expected value does now, however, account for costs that may be incurred by feeders in finding corn that meets their specified standards.



Digital Repository @ Iowa State University,

Copyright Owner

Steven Roger Meyer



Proquest ID


File Format


File Size

135 pages