Comparações entre equações de predição de consumo de capim-elefante cortado para vacas lactantes

This work aimed to compare intake prediction equations with values obtained by direct methods using chopped elephant-grass offered to crossbreed lactating cows with rumen canulas. The experimental design was a 3 x 3 Latin square (three animals and three cutting ages: 30, 45 and 60 days). The equations used for intake prediction (y) were: (1) y= -1.19 + 0.035(a+b) + 28.5c; (2) y= [%NDF on DM]*[NDF intake]/[(1- a - b)/KP+b/(c+kp)]/24; (3) y= -0.822 + 0.0748(a+b) + 40.7c and (4) equation 2 with values of intake measured directly. The predictions of NDF intake by equations were not different among treatments, instead of the difference among values measured directly: the 30 day-old had lower intake (5.29 kg/day) in relation to 45 (6.57 kg/day) and 60 (7.31 kg/day) day-old grasses. In general, equations overestimated the DM intake in relation to direct measuring (9.0 kg/cow/day), with exception of equation 3 which underestimated the intake (7.7 kg/day). The means of DM intake found by equations 1 and 2 (13.7 and 13.4 kg/cow/day, respectively) were similar between themselves and superior in relation to those found by equation 4 (9.7 kg/cow/day). The intakes measured directly were similar to those found in equation 4 and higher than those found by equation 3. The mean of rumen fill of 7.5 kg was superior to those of 5.2 kg estimated by equation. The prediction equations based on in situ degradability parameters do not supply estimates of DM intake, NDF intake and rumen fill in agreement with values obtained by direct methods.