AbstractFive mathematical models were compared to select the most satisfactory model to describe digesta kinetics of solids and fluids in the gastrointestinal tract of buffaloes (Mediterranean bulls), cattle (Friesian bulls) and sheep (Delle Langhe rams) given food at maintenance level, according to a Latin-square arrangement for four consecutive periods of 21 days. Chromium mordanted alfalfa hay and cobalt-ethylenediamine tetraacetic acid were used as nonabsorbable markers and were administered through the rumen cannula in a single dose. Four different isonitrogenous diets (N × 6·25 = 140 g/kg dry matter) with different concentrate:forage ratios (12·5:87·5, 25:75, 37·5:62·5, 50:50) were used.Faecal chromium and cobalt concentration curves were fitted with five non-linear models: three gamma (G2, G3, G4) age-dependent one-compartment, one gamma age-dependent/age-independent two-compartment (G2G1) and one multicompartment (MC).Wilcoxon tests on residual sums of squares of the different models for solids showed that MC and G4 gave a better fit than G2G1, G2, G3 for all the data and within the species. The comparison of MC v. G4 did not show any significant difference (P > 0·05) for all the data computed together or within each species. Nevertheless, MC had a higher number of curves with lower residual sums of squares in comparison with G4 and was also able to produce estimates of digesta kinetics in the second compartment.The cobalt excretion curves for fluids, considering all the data, and only within sheep, showed G4 as the best fitting model. When G4 was compared with other models no significant differences were recorded either for cattle: G4 v. G2 (F = 0·6645), G4 v. G2G1 (P = 0·0620) and for buffalo: G4 v. G2 (P = 0·1575), G4 v.G3(P = 0·0796) and G4 v. G2G1 (P = 0·1641).It is concluded that the multicompartment model (MC) and G4 model were the best fits for solids and for fluids respectively.
Improving the predictive ability of the signal-averaged electrocardiogram with a linear logistic model incorporating clinical variables.
