Pms. Ashton et al., ROLE OF LEGUMES IN RELEASE OF SUCCESSIONALLY ARRESTED GRASSLANDS IN THE CENTRAL HILLS OF SRI-LANKA, Restoration ecology, 5(1), 1997, pp. 36-43
Most of the world's forest has been cleared, cultivated, and then ofte
n abandoned. In many instances these areas have changed to successiona
lly arrested grasslands, shrublands, or fernlands maintained by freque
nt fires and high herbivore populations, Many studies have shown that
various herbaceous, nitrogen-fixing legumes can protect soil surfaces,
retain soil moisture, improve soil fertility, and retard ground fires
. Our objective was to ascertain if some of these species can potentia
lly inhibit herbivory and satisfactorily establish in these arrested g
rassland areas to serve as sites for reforestation. We evaluated the p
otential far four species of nitrogen-fixing legumes (Calapogonium muc
umoides, Centrosema pubescens, Desmodium ovalifolium, and Pueraria pha
seoloides) to establish on exposed soil within successionally arrested
grasslands of Panicum maximum and Cymbopogon nardus in the central hi
lls of Sri Lanka. Four different sites within rectangular grassland ar
ea were cleared of graminoids and sown with seed of each legume. Half
of each clearing was protected from browsing rabbits and procupines, a
nd have was not protected. After 6 months, certain plots were destruct
ively sampled to determine dry biomass gain for each species and treat
ment. Analyses of variance were performed to test for differences amon
g sites, treatments and species. All three factors revealed difference
s, in dicating that species must be matched to site. On site with high
amounts of herbivory, D. ovalifolium had the greatest dry biomass gai
n after 6 months of growth, possibly because of its relatively low nit
rogen and moisture content. Where herbivory was absent, P.phaseoloides
and C. muconoides had the greatest dry biomass gain. Dry biomass gain
of all four legume ground covers was low on sites with lowest pH and
nutrient concentrations. Under conditions of low relative fertility an
d low pH, establishment of the tested legumes failed. Though soil mois
ture availability was not measured, we speculate that these low fertil
ity sites were also prone to drought. Findings support the site-specif
ic establishment of legume species for purposes of reforestation and w
atershed protection in central Sri Lanka. This work is applicable to o
ther regions particularly dominated by successionally are rested grass
lands with similar circumstances in other parts of south and southeast
Asia.