The objective of this study was to evaluate the fitness capacity of a non-linear model in the process parameter estimation of the solid transit kinetics of bovine at pasture, using different sequences of fecal sampling. Five F1 Limousin X Nellore steers, fistulated in esophagus and rumen on Brachiaria decumbens grazing and concentrate supplemented during the rainy season period were used. The experiment was developed in three experimental periods; in a randomized completely blocks design. Chromium mordant, produced from extrusa samples, was used as marker. The double exponential time-dependent model was adjusted to the curves of fecal excretion of marker, and the following sample sequences were used: SEQ 1 - 22 samples (complete sample sequence); SEQ 2 and SEQ 3 - 17 samples (reduction of collection points in the ascendant and descendent phases of the curve, respectively); and SEQ 4 and SEQ 5 - 13 and 10 fecal samples (reduction of the number of samples in the whole profile of the curve). The reduced sequences were produced starting from the omission of specific points of the complete sample sequence (SEQ 1). The comparison among the estimates of the kinetic parameters and fecal excretion; the descriptive analysis of the necessary number of iterations to the convergence of the model and of the determination coefficient; and the evaluation of lack of fitness showed no differences among the sequences. However, the residual analysis pointed improvements in the graphic behavior and runs of signal profile of the residues with the reduction of the number of collections in SEQ 4 and SEQ 5. In function of small efficiency loss, evaluated from the residual variance, observed in SEQ 5, the employment of 13 fecal collections (SEQ 4) for evaluation of the fecal excretion curve of the marker in similar studies was recommended.
fecal excretion; external markers; mathematical models; pulse dose; rate of passage
