Abstract:

【Objective】 An experiment was made in order to precisely predict nutrient requirements of individual dairy cattle and study lactating characteristics of milk compositions. 【Method】 Two hundred and eighty-nine dairy cows were divided into 7 groups from 6 000 kg to 10 000 kg with an interval of 1 000 kg according to 305-day milk yield in the second parity. The adjusted month-milk-yields were fitted using 3 equations (Wood, Gompertz and Dijkstra equations). Twenty-one sets of lactation curve models for different milk yield ranges were obtained by SAS nonlinear parameter estimation and mean square error (MSE) analysis and lactation feature indices of each yield range including initial yield (y0), time to peak yield (tm), peak yield (ym) and relative rate of decline at the point halfway between peak yield and the end of lactation (r(th)) were revealed by model parameters. 【Result】 Analyses of the model parameters and feature indices showed that the above three mathematic equations are all able to describe lactation characteristics of Chinese Holstein dairy cows among different milk yield ranges for the second parity well,but their characteristics were different: a,b and c from Wood and Gompertz model changed regularly with the average yield (Yavg) of yield ranges (except for model 14 ); Y0 estimated by Gompertz was the best by means of regularity while Y0 of Wood had limitations, that of Dijkstra was different widely and could not reflect practical situations. Results of tm estimated by Wood was better than that of the other models which the variation was big in the lowest and highest yield ranges. Values of ym and r(th) of 3 models were highly consistent. 【Conclusion】 Above lactation curve models for different lactation ranges could provide reference for the accurate prediction of nutrient requirements and precise feeding for dairy cattle.

Key words: China, Holstein dairy cattle, lactation range, lactation curve, model