Reliability of in vivo, in vitro, in silico, and near infrared estimates of pure stand alfalfa hay quality: Component degradability and metabolizability of energy

Nine samples of pure stand alfalfa hay, representative of the compositional variability found in Mediterranean and Intermountain West climates, were collected throughout California and western Nevada during the 2008 growing season. Samples were evaluated in vivo (at either maintenance or ad libitum intakes by lambs), in vitro, in silico, and by near infrared spectrophotometry to determine characteristics of DM and fiber degradation as well as energy metabolizability. Also determined, from these and other routine analyses, were common indices of alfalfa quality including relative feed quality, relative feed value, in vivo TDN and ME. Significant differences (P < 0.05) in observed ME (Mcal/kg) were noted for alfalfa hays depending on source; quality indices were compared with observed ME. None of the indices evaluated were valid predictors of observed ME (P < 0.050), with the exception of near infrared predicted ME at either level of intake. Users of quality indices may expect less than optimal prediction of animal performance from use of indices or calculations based on indices failing to accurately predict ME. Results from this study indicate that ME estimated from near infrared spectrophotometry of pure stand alfalfa hay may improve characterization of alfalfa hay quality in that, depending on the feeding standard, ME input explains the vast majority of variation in animal output. However, given the limited numbers of alfalfa hay samples evaluated in this study, more studies are required.