Influence of model and marker on digesta kinetic estimates for sheep.

Several models and markers are available for digesta kinetic studies. In this study, kinetic estimates derived from chromium-mordanted hay or pellets were compared to estimates derived from rare earth markers (Yb, Dy, or Er) applied individually to samples. Twelve yearling rams (52 kg) were given ad libitum access to either hay or a commercial pelleted diet in a crossover experiment. Digesta kinetic estimates were obtained both by nonlinear analysis with two age-independent rates (G1G1) or with gamma time dependency in the fast compartment (G2G1 to G4G1) and by linear regression of natural log transformed fecal marker concentrations (LN method of estimate). Model did not influence total tract (P > .21) or ruminal mean retention times (P > .87). Partitioning of total retention time was similar (P > .21) for the LN, G3G1, and G4G1 models, but the G1G1 and G2G1 models did not adequately fit these data. Nonlinear models overestimated, and LN underestimated, fecal DM output by 9% (SEM = 4.7) for the hay diet. All the nonlinear models provided fecal DM output estimates that were within 5% of actual fecal DM output, but the LN model underestimated it by 18% (SEM = 3.3) for the pelleted diet. Ruminal outflow rate was slower (P < .01) and both ruminal and total mean retention time estimates were longer (P < .01) for Cr than for the rare earths. Despite the marker chosen, relative diet effects were similar. The three rare earth markers gave identical results for digesta kinetic estimates, indicating that they are useful for simultaneous study of more than one ingredient or particle, but direct comparison of rare earth and Cr-mordant passage rates is not advisable.