Comparison of nonlinear models to describe the feather growth and development curve in yellow-feathered chickens

Feathers play a critical role in thermoregulation and directly influence poultry production. Poor feathering adversely affects living appearance and carcass quality, thus reducing profits. However, producers tend to ignore the importance of feather development and do not know the laws of feather growth and development. The objective of this study was to fit growth curves to describe the growth and development of feathers in yellow-feathered broilers during the embryonic and posthatching periods using different nonlinear functions (Gompertz, logistic and Bertalanffy). Feather mass and length were determined during the embryonic development and posthatching stages to identify which growth model most accurately described the feather growth pattern. The results showed that chick embryos began to grow feathers at approximately embryonic (E) day 10, and the feathers grew rapidly from E13 to E17. There was little change from E17 to the day of hatching (DOH). During the embryonic period, the Gompertz function (Y = 798.48e-203 431exp(-0.87t), Akaike's information criterion (AIC) = -0.950 × 103, Bayesian information criterion (BIC) = -0.711 × 103 and mean square error (MSE) = 559.308) provided the best fit for the feather growth curve compared with the other two functions. After hatching, feather mass and length changed little from the DOH to day (D) 14, increased rapidly from D21 to D91 and then grew slowly after D91. The first stage of feather molting occurred from 2 to 3 weeks of age when the down feathers were mostly shed and replaced with juvenile feathers, and the second stage occurred at approximately 13 to 15 weeks of age. The three nonlinear functions could overall fit the feather growth curve well, but the Bertalanffy model (Y = 116.88 × (1-0.86e-0.02t)3, AIC = 1.065 × 105, BIC = 1.077 × 105 and MSE = 11.308) showed the highest degree of fit among the models. Therefore, the Gompertz model exhibited the best goodness of fit for the feather growth curve during the embryonic development, while the Bertalanffy model was the most suitable model due to its accurate ability to predict the growth and development of feathers during the growth period, which is an important commercial characteristic of yellow-feathered chickens.