EFFECT OF WOOOY ANO HERBACEOUS LEGUMES ON THE GROWTH ANO NUTRIENT CONTENT OF TWO TROPICAL GRASS SPECIES ? -Development of Resumen M. Domínguez l , D. Kass 2 , M. Ibrahim 2 , F. Jiménez 2 I Former graduate student, CA TIE, Turrialba, Costa Rica. Agraforestry Systems Unit, Watershed and Agraforestry Systems Area. CA TIE, Turrialba 7170, Costa Rica. Se realizó un experimento factorial (2 X 2 X 5) con dos especies de gramineas, Paniculll maximum Jacq. (Var. CIAT 16061) y Brachiaria humidico/a Rendle (Var. CIAT 679) en monocultivo o en asocio con Centrosema macrocarpum (Benth). Había cinc( niveles de adición de mantillo: ninguno; Erythrina poeppigiana (Walp. ) O.F. Cook (138 g maceta· l ) para dar 150 kg N ha· l ; E. poeppigiana (276 g maceta- I ) para dar 300 kg N ha ol ; Acacia mangium Will. (86 g maceta ol ) para dar 150 kg N ha ol yA. mangium (173 g maceta- I ) para dar 300 kg N ha ol • El mejor crecimiento de gramíneas con mayor contenido de nutrientes se obtuvo con el nivel más alto de E. poeppigiana y asocio con C. macrocarpulIl. Sin embargo, el mantillo de A. mangium también tenía efecto significativo sobre el control. El efecto de C. macrocarpum fue siempre positivo. Palabras claves: Acacia mangium. Brachiaria humidicola, Centrosema lIJacrocarpum Ervthrina poeppigiana, N supply, nutrient cycling, nutrient use efficiency. Panicum maximum, tropical pastures Introduction It should be possible to increase both the grawth, N content and nutritional value of tropical pastures by associating them with either herbaceous legumes or N-fixing-trees. A range of different strategies are possible, utilizing either herbaceous or arboreal legumes in association with different grass species. The present research was carried out to see to what extent grawth and nutrient content of two grass species, widely used in the humid tropics, could be improved by the addition of prunings of N fixing trees and/or by association with herbaceous legumes. Efficiency of nutrient absorption has been used to estimate how much of a nutrient applied in different forms is utilized by the crap to which it is applied and how much is taken up by other sinks within and outside of the system (Novoa and Loomis, 1981, MolI et al., 1982) Methodology A cOlllpletely randolllized experiment was set up in a greenhouse in Turrialba, Costa Rica (9 0 53' N, 83°38' W, altitude of 603 m). Temperatures in the greenhouse varied between 20.3 and 32 oC during the experilllent. The design was a 2 X 2 X 5 factorial with two grass species, Palli culIJ maxinzum Jacq. (Var. CIAT 16061) and Brachiaria humidicola Rendle (Val'. CIAT 679), grawn alone or in association with Centrosema macrocarpum Benth. There were five levels of mulch addition: none; Erythrina poeppigiana (Walp.) O.F. Cook (138 g por l ) to give 150 kg N ha ol ; E. poeppigialla (276 g por l ) to give 300 kg N ha ol ; Acacia mangium Will. (86 g pot 0 1 ) to give 150 kg N ha ol , and A. mangium (173 g por l ) to give 300 kg N ha ol . The E. poeppigialla contained 260 mg g"1 dry malter with an N content 252