Evaluation of Laboratory Methods for Determining Quality of Corn and Sorghum Silages: I. Biological Methods for Predicting In Vivo Digestibility

While many laboratory procedures have been proposed for estimating feeding quality of perennial grass and legume forages, none have been adequately tested for their potential in predicting quality of corn and sorghum silages. The objective of this phase of a comprehensive study was to assess the value of six biological procedures proposed for predicting forage digestibility of silages of corn (Zea mays L.) and sorghum (Sorghum bicolor (L.) Moench, S. sudanense (Piper) Stapf, and their hybrid). In vivo digestible dry matter (DDM) of 51 corn and sorghum silages (17 in each of 3 years) was determined using sheep in conventional feeding trials. A modified Tilley and Terry (1963) in vitro procedure (T-T DDM), gave the highest correlations with in vivo DDM (r = 0.83 for corn silages and 0.91 for sorghum silages). A “direct acidification” in vitro rumen fermentation procedure resulted in in vitro DDM which was also highly correlated with in vivo DDM (r = 0.50 and 0.91 for corn and sorghum silages, respectively). Within the corn silages, the correlation between direct avidification DDM and in vivo DDM of 18 entries with high grain percentages was only 0.19, while that for seven entries with low grain percentages was 0.93. Apparently, the combination of large amounts of available carbohydrates in the high grain entries and the phosphate buffer led to erratic growth and digestive action by the rumen bacteria. A 6-hour rumen fluid fermentation procedure gave DDM values which were 36 to 38% of the T-T DDM values and were poorly correlated with in vivo DDM. A cellulase-acid pepsin procedure gave DDM values for both corn and sorghum silages that were 56% of the T-T DDM values; however, cellulase-acid pepsin DDM was positively correlated with in vivo DDM of sorghum silages (r = 0.72) and negatively correlated with in vivo DDM of corn silages (r = −0.42). Two procedures that involved the neutral detergent extraction (Goering and Van Soest, 1970) and a 48-hour rumen fluid fermentation gave in vitro DDM values which were highly correlated with in vivo DDM (r = 0.73 to 0.90). A multiple correlation analysis revealed that little improvement of the simple correlation between specific laboratory procedures and in vivo DDM was attained by combining procedures. We developed simple regression equations to predict in vivo DDM from T-T DDM and direct acidification DDM for corn and sorghum silages.