Growth patterns of Italian local chicken populations

Predictions of growth are important factors that contribute to the profitability of an operation in poultry production. Modern commercial hybrids have a higher body growth in comparison to the local purebreds. However a niche market for meat and egg poultry production needs to be established using local purebreds to promote biodiversity. The aim of this study was to model the growth response of male and female chickens belonging to five local Italian populations, a commercial slow-growing hybrid (Berlanda, B), the Padovana pure breed (two plumage varieties: chamois, camosciata - PC, and silver, argentata - PA) and their crosses PCxPA and PCxB. A total of 398 one-day birds were reared until 180 d of age under indoor conditions. The linear and three non linear models (logistic, Gompertz and Richards) were compared to study the growth patterns of these chicken populations. Significant (P<0.01) differences were observed among the genotypes for several curve parameters. In males, PCxB showed the lowest age at inflection point, B showed the highest age and body weight, whereas PA showed the highest age and the lowest weight. In females, the age at the inflection point did not differ among the groups; B showed the highest weight. All the non linear models gave a good fit of male and female data with R2 ranging from 0.992 and 0.999, but the logistic equation had higher value of root mean square error than the Gompertz and the Richards values. Based on residual sum of squares for both sexes, the Richards model was better (P<0.05) than the logistic but not superior to the Gompertz. The logistic equation showed an overestimation of initial body weight for all the groups and sex. For Italian local chicken populations, the Richards model requires a measure of body weight recorded at 90 d or after to obtain a good fit of the asymptotic weight. However, the Gompertz model has the advantage that it requires one fewer parameters than the Richards model.