ABSTRACT - This experiment was conducted in an elephant grass pasture from EMBRAPA/CNPGL to compare methods for estimating forage availability, as well as to evaluate the accuracy of three observers in the evaluations. The cutting, plant height, ground cover and height x ground cover methods, multiple regression analysis and the visual comparative yield were compared with the total productions in respective half of evaluated area. The visual comparative yield method always provided the highest coefficients of determination of estimating equations, followed by the multiple regression analysis with height and cover plant. However, the plant height and ground cover methods showed the worst results, showing the low viability of estimating elephant grass availability, by means of isolated traits. Although the visual comparative yield method provided the highest coefficients, the multiple regression analysis method showed to be viable, meanly by the high practicability and the good results. Inexperienced observers had always the worst results, meanly in the first evaluations and in grazing areas.
indirect methods; pastures; Pennisetum purpureum; dry matter yield
