The objective of this study was to propose and validate a model to estimate the apparently digestible fraction of crude protein (CP) of cattle diets. The theoretical framework for model development was based on true digestibility coefficient and metabolic fecal contribution. The estimation of both parameters was done by using the Lucas test. The dataset for model building was composed by 39 and 54 means from 10 experiments with lactating cows and 23 experiments with growing and finishing cattle, respectively. For comparisons of animal categories a Dummy variable was used. The dataset for model validation came from two trials with lactating cows and two trials with growing and finishing cattle. The true digestibility coefficient, which was estimated as 0.7845 did not differ among animal categories. The greatest metabolic fecal CP was observed on lactating cows but it was not affected by dietary CP level when expressed as percentage of dietary CP. Thus, for animals fed ad libitum the content of apparently digestible CP (CPad%) in the diet can be estimated by the following equations: CPad% = 0.7845CP% - 0.97 for lactating cows and CPad% = 0.7845CP% - 1.61 for growing and finishing cattle. The validation procedure indicated no differences between observed and estimated values of CPad%. It can be concluded that the proposed models will improve the accuracy of TDN prediction using the chemical composition of the diet.
apparent digestibility; Lucas test; NRC system; summative equations
