Prediction models of silage fermentation products on crop composition under strict anaerobic conditions: A meta-analysis

A meta-analysis was conducted to establish linkages between crop and fermentation variables. Data from well-controlled mini silage studies were used in which no additives had been used and no ingress of air had occurred. The silage set consisted of data on crop chemical composition and epiphytic lactic acid bacteria count, and fermentation products (organic acids, alcohols, and ammonia-N) from 118 silages made from 30 grass, 7 legume, 15 grass and legume mixtures, and 66 whole-crop maize samples. The prediction models for fermentation products on crop variables were obtained by stepwise multiple regression analysis. Perennial forage and maize silages were analyzed separately. The best models were obtained for acetic acid in perennial forage silages, with a coefficient of determination of 0.63, and for lactic acid and ethanol in whole-crop maize silages, with coefficients of determination of 0.84 and 0.61, respectively. Fermentation products of perennial forage and maize silages were best related to dry matter and crude protein contents, respectively. Overall, the prediction equations were weak.