Prediction of the relative intake potential of grass silage by dairy cows

Abstract

Data based on the mean treatment values from production studies in lactating dairy cows were used to estimate relationships between silage composition and silage dry matter intake (SDMI). Data from 21 studies were used to estimate relationships between D-value (g digestible organic matter in kg dry matter (DM)) and crude protein (CP) content, and SDMI. The silages were harvested at different maturities but using the same ensiling techniques within the study. Relationships between silage fermentation characteristics and SDMI were estimated using data from 47 studies. The silages were harvested at the same time from the same sward but using various additives. When factorial designs were used, silage data within each treatment (subexperiment) were recorded separately. The data sets included 125 D-value and 234 fermentation observations. Relationships between SDMI and silage parameters were analyzed using mixed model regression analyses with experiment as a fixed factor and subexperiment within experiment as a random factor. The ranges for silage CP concentration and D-value were 111 to 238 (S.D. 25.2) and 589 to 756 (S.D. 48.5) g/kg DM, respectively. D-value was a much better predictor of SDMI than CP ( R 2 within experiment 0.71 vs. 0.26). The effect of D-value on SDMI diminished with increasing level of concentrate supplementation (interaction P<0.001). D-value had a greater effect on SDMI when the total DM intake was high. SDMI was negatively correlated with concentrations of ammonia N, lactic acid, individual and total volatile fatty acids and total fermentation acids and positively correlated to the concentration of residual water soluble carbohydrates. Total acid concentration was the best SDMI predictor of individual fermentation parameters ( R 2 within experiment 0.41) followed by lactic acid and ammonia N. Use of quadratic regressions generally increased the variation accounted for by the model. The best multiple regression accounted for 0.51 of the variation in SDMI within experiment. To facilitate interpretation of data, a theoretical model is presented. An interplay between physical load and capacity to use energy determines SDMI of well-fermented silages over the whole range of D-values. With increasing extent of fermentation SDMI is constrained by nutrient imbalance, most probably amino acid to energy ratio at the tissue level, resulting from reduced microbial protein synthesis in the rumen. Low palatability can further constrain silage DM intake. For advisory silage evaluation the following model predicting relative SDMI is proposed: SDMI index=100+0.151×( D-value−690)−0.000531×(TA 2−6400)−4.7650 [Ln(Ammonia N) −Ln(50)], where D-value and total acids (TA) are expressed as g/kg DM and ammonia N as g/kg total N, respectively. Regression coefficients are scaled to a mean SDMI of 10 kg DM/day. The values of 690, 80 and 50 are used as standard D-value (g/kg DM), total acid (g/kg DM) and ammonia N (g/kg N) for high quality restrictively fermented silages. The model parameters were limited to those which are currently available for Finnish farm silage analyses, i.e., by near infra-red reflectance spectra ( D-value) and electrometric titration (fermentation parameters).