The measurement of chest girth as an alternative to weight determination in the performance recording of meat sheep

The aim of this study was to assess, for two Italian meat sheep breeds (Appenninica and Merinizzata italiana), the relationship between an easily recorded measurement (girth of chest) and the character used for selection purposes (weight), and to define the most appropriate mathematical methods to infer the second from the first.For the Appenninica 1392 lambs were measured, for the Merinizzata italiana 1559 lambs were measured. The possibility of estimating weight through chest girth (CG) measurement was evaluated, separately for each breed, by taking the most suitable model between those including different kinds of regression effect. The model was chosen in relation to the value of the determination coefficient and the sum of square residuals. The prediction accuracy of the model was assessed by comparing the expected values with the observed ones through a number of statistical tests.A further prediction analysis was carried out using the mean values of the observed weights that fell in each 1 cm class of girth, in order to reduce the error derived by the varying numbers of observations per unit of chest girth.The model including the square regression nested within the sex effect and the flock random effect nested within the sex effect was observed to be the most suitable one to predict the weight from the chest girth; the determination coefficients ranged between 0.944 (Appenninica) and 0.955 (Merinizzata). The prediction parameters were: -10.458+ 0.241 (CG) + 0.004 (CG2) for the Appenninica males; -6.121 + 0.093 (CG) + 0.005 (CG2) for the Appenninica females; -6.325 + 0.189 (CG) + 0.004 (CG2) for the Merinizzata males; -4.676 + 0.078 (CG) + 0.005 (CG2) for the Merinizzata females. The correlation between the observed and expected values was always higher than 0.97. The equations estimated using the mean weights for each girth showed extremely high determination coefficients (≅ 0.99) due to the reduction of variability implied by this method. Choosing between the equations calculated on the entire data set or on the mean weights will only be possible after a period of field tests.