Evaluation of the grass-tanzania ("Panicum maximum") using multivariate analyses

Abstract

The experiment was conducted with objective of evaluating the tanzania grass, using mult-trait procedures: discriminant function analysis and principal components analysis. The variables were: dry matter of leaf (DML), dry matter of the plant (DMP), nitrogen (N), phosphorous (P), magnesium (Mg), calcium (Ca) and potassium (K) in the aerial part and in the root of the grass, during the rain period. The plants were fertilized with four doses of N as bovine manure (0, 60, 120, 180 kg/ha). Main components analysis was capable to reduce the 15 original variables to 6 principal components, which the most important variables were calcium at root, potassium at leaf, potassium at plant, magnesium at leaf, phosphorous at leaf and phosphorous at root. The discriminant analysis had difficulties in classifying the fertilizer dose, especially related to the levels 60, 120 e 180 kg/ha.