DECOMPOSITION OF LEAF-LITTER FROM TROPICAL FORAGE GRASSES AND LEGUMES

In tropical pastures which are generally underutilized, nutrients are returned to the soil mainly via plant litter. The release and recyclin g of nutrients from litter is known to be a function of climate and li tter composition. However, little is known about the decomposition of tropical forage species in pastures. We compared rates of above ground litter decomposition and changes in nutrient contents of six legume a nd four grass species in litter bags over the wet and dry season of a well-drained isohyperthermic savanna of Colombia. Decomposition consta nts and litter half-lives were estimated by fitting a single exponenti al model to the data. Litter production was measured monthly in pure s tands using matched quadrats. Of the six legume species studied, rates of decomposition of organic matter (OM) were fastest in Stylosanthes capitata and Arachis pintoi and slowest in Desmodium ovalifolium. The remaining legumes, Centrosema acutifolium, Pueraria phaseoloides and S tylosanthes guianensis, decomposed at rates similar to the grasses And ropogon gayanus, Brachiaria decumbens, B. dictyoneura and B. humidicol a. Rates of litter decomposition decreased and litter half-lives incre ased during the onset of the dry season. Rates of release of nutrients (N, P, K, Ca and Mg) generally followed a pattern similar to that of OM, although the absolute amounts of N, K and Ca released were greater in the legumes than in the grasses due to higher initial concentratio ns in legumes. Immobilization of N occurred only in litter with high C :N ratios (> 109) and high lignin:N ratios (>25). Of the chemical vari ables studied the lignin: N and (lignin + polyphenol): N ratios were l inearly correlated with the loss of OM from grass and legume litter (r = 0.66). Losses of N were best correlated with the % lignin + (C: N r atio) (r = 0.67) but lignin: N, (lignin + polyphenol): N and C-to-N ra tios also gave similar correlations (r = 0.61-0.67). The linear relati onship between the lignin-to-N ratio and the loss of OM was improved w hen rainfall was included in a multiple regression (r = 0.82). The use of polyphenols as an indicator of litter decomposition is discussed b ut not recommended using the current simple methodologies available. T he rates of N released were estimated from the litter decay equations and the mean monthly rates of fitter production from pure stands of ea ch species and were related to the N requirements of a tropical pastur e producing 0.85-1.7 kg DM m-2 yr-1. The estimates indicated a slow an d inadequate rate of N release from grasses and a wide variation in ra te of release amongst the legumes and hence a variable, species-depend ent, ability to supply a substantial proportion of a pasture's require ments from leaf litter decomposition.