Ability of Subjective Linear Scores to Represent Cow Differences in Objective Body Measurements

Abstract

Body measurements, experimental linear descriptive scores, and Holstein descriptive codes on nearly 300 lactating cows in five herds were used 1) to compare the effectiveness of the later two subjective systems at depicting cow differences, and 2) to quantitate the loss of accuracy from reducing the 50 point linear scale to 25, 10, and 5 classes. Indirect selection differentials were calculated for various percentages saved (1 to 40%) when selection was based on linear scores and descriptive codes and compared with direct selection differentials. Indirect selection on linear stature and udder depth scores was nearly as useful as direct selection for stature and distance from rear udder to hock. Selection differential for linear scores for rear udder width were intermediate. For the measurements of chest depth, hip width, and thurl width, linear scores were only slightly better than descriptive scores. Reducing the number of linear classes to 25 had little impact on any of the traits studied. Reducing classes to 5 had a larger impact on explaining cow differences in chest depth (especially at the extreme), hip width, thurl width, and rear udder width. Where comparable, 10 to 25 classes of linear scores explain most of the cow differences that can be measured objectively.