Liveweight estimation from body measures in "Nero di Parma" pigs

In the past, predictive equations have been applied to the calculation of body weight from body measures of several species and breeds. Their employment is an important aspect in herd management, most of all for the outdoor rearing system. In fact, in such conditions it’s difficult to use a balance and, moreover, the “wild” aptitude of animals and the dangerousness of their handling must be considered. The weighing is a stress for animals, while body measuring is much more less invasive, even it involves the capture of the animals. Moreover, body measures can also be taken by image analysis. The aim of the research was to define accurate and easy-to-use equations for the prediction of body weight from body measures in “Nero di Parma” pig, a local genetic type suitable for outdoor rearing system typical of different areas of Northern Italy, particularly of the Emilia-Romagna region. The trial was carried out on the weights of 1023 pigs between 1 and 1585 d of age (532 females, 32 males and 459 castrated males), group reared in 18 herds and fed commercial standard diets given ad libitum and integrated with pasture in the brushwood. During weighing, the hearth girth (HG) was also measured on all the animals by a tape measure, while on 190 pigs the height at withers (HW) and at rump (HR), the length of the body (LB), the fore-rear length of the thigh (LT) and the width of the rump (WR) were also recorded. Data were submitted to stepwise regression analysis, by introducing the independent variables (body measures and age) in the linear, quadratic and cubic form. The ability of the models to fit the raw data was evaluated by means of R2 and SE of the dependent variable. Two equations for live weight (LW) estimation have been calculated: the first one by submitting to analysis all the body measures and age (n = 190) and the second one by using HG (n =1023) only. Equation 1: LW = 9.035 + 0.006696 * HG2 - 0.454 * HR + 0.000120884 * HR3 + 0.00159 * LB2 - 0.172 * HW (SE = 2.39 kg; R2 = 0.995; P<0.001) Equation 2: LW = 3.483 - 0.342 * HG + 0.010998 * HG2 (SE = 1.95 kg; R2 = 0.988; P<0.001) Both equations have shown high R2 values and low SE of the dependent variable. The real and the estimated weights by applying the two prediction equations resulted highly correlated (0.997 and 0.994, respectively for equation 1 e 2). In conclusion, equation 2, though slightly less accurate (lower values of R2), is more reliable (lower values of SE of the dependent variable); this could be due to the higher number of observations used for the calculation. Finally, from a practical point of view equation 2 is much more easy to apply to field conditions, because it only counts in one body measure (HG).