PREDICTION OF ALFALFA CHEMICAL-COMPOSITION FROM MATURITY AND PLANT MORPHOLOGY

To make informed harvest management decisions, producers need information on the chemical composition of alfalfa (Medicago sativa L.) forage. Our objective was to develop a fast and simple method for predicting forage chemical composition (crude protein, neutral- and acid-detergent fiber, and acid-detergent lignin) using morphological characters. Five alfalfa cultivars were evaluated across a range of environments, growth cycles, and harvest maturities. The 15 morphological characters examined for association with chemical composition included number of nodes per plant, plant height, mean stage count (MSC), mean stage weight (MSW), and alternative stage of development descriptions. Simple regression analyses based on node number and plant height provided better prediction equations than when based on MSC or MSW. Multiple regression equations based on two and three morphological characters provided higher R2 and lower root mean square error (RMSE) values than did linear and quadratic models based on a single morphological character. Multiple regression of the two most easily measured characters, height of the tallest stem (MAXHT) and maturity stage of the most mature stem (MAX), provided equations with RMSE values only slightly greater than those of the best two-factor multiple regression models based on characters that are more difficult to measure. Thus, multiple regression techniques should improve the accuracy of predicting forage chemical composition, and the use of easily determined characteristics such as MAXHT and MAX may provide the best compromise between prediction accuracy and ease of use.