Date of Award
Master of Science
Haley M. Jamison
William R. Backus, Robert S. Dotson
Records on 383 bulls, tested at The University of Tennessee Bull Evaluation Station for five testing periods in 1971 through 1975, were studied to determine the degree to which one could predict 140-day weight. Objective measures from birth to weaning and obser-vations prior to testing were used in studying the performance of animals on test. These data indicated that the only significant prediction that can be made is predicting the first 28-day weight from the on test weight. The overall mean for adjusted 205-day weight, on test weight, and backfat at the beginning of test was 551.26 pounds, 662.24 pounds, and 3.59 millimeters, respectively. The correlations between the adjusted 205-day weight and on test weight to the five feeding periods were highly significant (P < .01). The correlation between backfat at the beginning of test for the same periods were highly significant with the exception of the 112 and 140-day weights. Many of the observations in these data were a part-whole rela-tionship since the greatest contributions for predicting the 140-day weight were from on test weight, adjusted 205—day weight and backfat at the beginning of test. In order to remove the sources of environ-mental variation, the analyses were calculated on a within year-breed basis. In the data the order of inclusion of independent variable, on test weight, adjusted 205—day weight and backfat at beginning of test, were based on the readily available data. Weigh periods taken each 28 days during the 140-day full feed periods were used as dependent variables. In the 56-day weights, a decreased percent of the variation was explained, 87.86 percent. The percent of variation explained for 84, 112, and 140-day weights was 84.89, 80.78, and 74.80 percent, respectively. In these data using prediction equations for weight at the five weight periods, y, which represents 28, 56, 84, 112, and 140-day weights, is the best estimator of the population. Various combinations of independent variables could not be accurately used as predictors of 140—day weight. The percent vari-ation explained in average daily gain by various combinations of independent variables revealed that a maximum daily gain of 8.146 percent could be accounted for. The only significant prediction which can be drawn from these data is the ability to predict 28-day weight from on test weight. Accurate predictions cannot be made at any other full feed period.
Law, Earl W., "An analysis of the U.T. bull evaluation test 1971-1975. " Master's Thesis, University of Tennessee, 1975.