Comparison of weight estimation methods for wild animals and domestic livestock.
Abstract
Body weights and measurements were determined on a number of species of antelope, the zebra and the hyena, 7-158 animals of each species being examined. Using a linear equation to predict body weight from chest girth, the residual standard deviations varied between species from 8 to 33 lb., and the amount of variation in body weight accounted for by variation in chest girth from 42 to 92%. Equations of the form W = a+bG+cG2 (where W = weight, G = heart girth, and a, b and G are constants for the species) only slightly improved prediction. Power law relationships of the form log W = d log G+e were established for wildebeest and Thomson's gazelle, the logarithmic regressions accounting for 94 and 75% resp. of the total variances. Power law relationships transformed back to the original units are tabulated for 5 antelope species; in some species prediction was more precise than with linear regression, and with some it was less so. Prediction equations of the form W = aLG2+b, where L = body length, were computed; these accounted for 46-92% of the variance in body weight, depending on species, and had residual standard deviations of 46-13.5 Ib. The prediction equations were re-computed for 2 antelope species using body weight minus weight of stomach contents, or eviscerated weight, as the dependant variable. This procedure reduced prediction efficiency or gave no improvement. Coefficients of variation for estimated body weight were less than 10% and were lower than those reported for cattle. Published and unpublished data of other authors are quoted to show that wild ungulates have a higher percentage of lean in the carcass than cattle, and a much smaller range in fat percentage.