Twenty-four Saanen goats (fifteen multiparous and nine primiparous) during their lactation period were used, from the kidding until the 200th day of lactation; and were distributed in a completely randomized design and in a factorial arrangement (3 diets x 2 parity orders). The diets were composed of either soybean meal (SBM), soybean meal plus inactive dry yeast (SMDY) or inactive dry yeast (DY) used as the protein source. The analyses of milk yield, body weight and dry matter intake (DMI) were made through the Bayesian methodology. For construct the lactation curves were used the Wood's nonlinear model. The inclusion of inactive dry yeast in the diets reduced the rate of yield increase until the lactation peak for primiparous goats, but for multiparous goats the reverse was noted. The SMDY diet reduced the rate of yield decline after peak for primiparous goats, whereas for multiparous goats the inclusion of inactive dry yeast in the diet increased the decline of the rate of yield during this period. It was observed that inactive dry yeast can be used to replace soybean meal in diets used for multiparous Saanen goats.
bayesian methodology; milk productivity; Wood's nonlinear model; Saccharomyces cerevisiae
