Sri Lankan rice farmers rarely practise green manuring and they depend mostly on chemical fertilizers for the nutrient requirement of their crops. With the removal of government subsidy on fertilizers since 1990, they are now faced with the dilemma of meeting the cost of production to sustain previous yield targets. Therefore the present study devotes to evaluate alternative cheaper nutrient sources like green manures for rice. Results of recent work in Sri Lanka have shown that Sesbania speciosa Taub. ex Engl.is a promising green manure for lowland rice in the dry zone. It was evaluated to adopt a suitable green manuring system for rice. Its year-round biomass production, nutrient accumulation during the vegetative growth and nitrogen fixation were monitored. Field trials were conducted to investigate its ability to suppress weed growth when grown during fallow periods and its contribution to soil fertility and effect on rice yield. Highest biomass of around 100 g dry matter per plant, was achieved when S. speciosa was seeded in August and harvested at flowering in November. When this was planted during a fallow period at a spacing of 0.25 x 0.25 m (16 plants m 2), percentage weed reduction was as much as 40%. Its N 2 fixation activity was similar to that of S. sesban (L.) Merr.. Nitrogen concentration in S. speciosa increased until 70 days after establishment and then decreased towards flowering. Phosphorus and K accumulation was also effective during this period. Therefore suitable time for soil-incorporation of S. speciosa is around 70 days after establishment. S. speciosa-manured and chemically fertilized rice crops were comparable in terms of grain yield, indicating the effective nutrient supplying potential of this green manure. Straw biomass was significantly high under S. speciosa, because of the high N-allocation to vegetative growth of rice under this. This evaluation shows that S. speciosa not only supplies N, but is also an effective and complete supplier of other nutrients for rice. A scheme to introduce this into rice production system is proposed, where S. speciosa is seeded/broadcast two weeks before crop harvest (around January) in Maha (major) season as well as Yala (minor) season (around June) so that its biomass can be incorporated into soil during land preparation of subsequent seasons.

Changes in N, P, K, Ca and Mg in soil and rice plants were investigated during a cropping season following a long fallow period in a system of traditional cultivation practised for several centuries, under a village tank irrigation system. Soil, N, P, K, Ca and Mg were not found to be deficient for rice production throughout the season. Flooding did not produce toxic levels of Fe and Na and soil pH remained at 6.4 during the season. The average grain yield (3.5 t/ha) without any addition of chemical fertilizer was almost the same as that from fields under major irrigation systems where fertilizer application (less than the recommended level) was common. The sustainability of soil fertility under the traditional system of rice cultivation appears to be dependent upon long fallow periods. The natural build-up of soil fertility during a three-year fallow was evidently adequate to support a good growth of the crop which produced a yield comparable to that obtained in chemically fertilized, more intensively cropped rice fields under major irrigation systems.