Abstract

Metabolizable energy (ME) is the most limiting nutrient for milk production in pastoral systems. In Chile, using four wethers and grass cut on six measurement dates, Garrido and Mann (1981) developed an equation to predict grass ME concentration based on content of in vitro digestible organic matter in DM. The application of the Garrido and Mann (1981) equation has been generalized in Chile for all feed ingredients, diets and ruminant species. We hypothesized that ME prediction from this equation has limitations for its use in lactating grass-fed dairy cows. The objective of this study was to validate the ME prediction equation by Garrido and Mann (1981) for lactating grass-fed dairy cows in the autumn and spring seasons. The study was conducted at INIA Remehue (Osorno, Chile) using two sets of eight Holstein Friesian cows, with each group used twice in 21-d experimental periods in autumn and spring, respectively. Fresh grass-only diets with a predominance of perennial ryegrass (Lolium perenne L.) were supplied through soiling. In the last 5 d of each period, grass ME concentration was measured by subtracting energy lost in feces, urine and methane from ingested gross energy. A weak linear relationship was obtained between the ME determined in trials and ME predicted from the Garrido and Mann (1981) equation (R-2 = 0.266; P = 0.003). The ME prediction equation currently in use is inadequate for lactating grass-fed dairy cows. Therefore, new and improved equations for estimating grass ME specific to dairy cows are needed.