Design and Conduct of Calf Nutrition Studies. 1. One- Vs. Two- and Three-Day Growth Measurements

Open Access
Publisher Website
Google Scholar
Abstract
Effects of 1-, 2-, and 3-day measurements for weight, heart girth, girth at paunch, height at withers, and length from withers to pin bones for 48 calves, raised on a limited whole milk-calf starter system, were studied at biweekly intervals for 13 weeks. Because the components of variance approach for studying different numbers of measurement was found to be questionable, measurements taken on successive days on the same calf can not be safely assumed to be independent; conclusions were drawn from the errors estimated by using the 1-day measurements or the average of 1 and 2 or the average of 3 and making comparisons among these errors by the Pitman-Morgan test for correlated mean squares. A single-day measure- ment was sufficient to determine growth in weight, heart girth, and girth at paunch; whereas, 3-day measurements were significantly better than 1- or 2- day measurements for length from withers to pin bones, but the results were inconclusive for height at withers. Sizes of the coefficients of variation for each growth criteria decreased markedly up to 6 weeks, and thereafter the decreases were relatively small. Editor. Physical measurements are used in nutrition experiments to determine the effect of different rations or nutrients on growth. This effect is assessed by a comparison of the changes with time in measurements, ordinarily the dif- ferences between initial and final measurements, of animals on different treat- ments. This comparison is subject to two kinds of error: (a) between-animal variability in growth, which is present even when animals of the same age, sex, and breed are used, and (b) measurement variability. These constitute the error or residual variability found in the analysis of variance. By taking an infinite number of determinations of a measurement 011 each of the experimental animals, it is theoretically possible to eliminate the measurement variability entirely. If, however, the animal variability is much greater than the measurement vari-