Replenishment of Organic Matter for Improving San Sai Soil Series Productivity in Rice-Soybean Cropping System   Piratcha Wassananukul Prachaya Dhanyadee and Preedee Deeraksa   ABSTRACT         The study on replenishment of organic matter for improving soil productivity in rice-soybean cropping system was conducted during May 1992 - 1995 at the Multiple Cropping Center, Chiang Mai University. The experiment was factorial in randomized complete block design with 3 replications. The treatments were composed of two varieties of rice, RD 7 and KDML 105. The five nitrogen sources were no fertilizer, urea 20 kg/rai, soybean residues at 200 kg/rai, soybean residues at 200 kg/rai in combination with urea 10 kg/rai and Sesbania rostrata.       The results from the study showed that average 3 year yield revealed significant difference among treatments with different N sources and between varieties, the incorporation of Sesbania rostrata and rice yield in both RD 7 and KDML 105 produced the highest yield in 3 years average yield 785 kg/rai and 691 kg/rai, respectively. While the treatments from the application of urea 20 kg/rai, soybean residues 200 kg/rai and soybean residues 200 kg/rai in association with urea 12 kg/rai were not significantly different, but obviously different from the treatment of no fertilizer application, which furnished the lowest yield of RD 7 and KDML 105 down to 567 and 529 kg/rai respectively. Soil chemical analysis revealed that the total nitrogen in soil tended to increase in all treatments after the experiment. The increase of total nitrogen in soil due to the addition of nitrogen sources as Sesbania rostrata, soybean residues, soybean residues plus urea, urea and no-N including weed residues in all treatments. Also soybean could fix nitrogen from the air. On the contrary the organic carbon in soil in all treatment tended to decrease after the experiment. The residue effect of soybean grown after rice were markedly shown in the third year. The application of 20 kg/rai in rice field gave the lowest soybean yield with average 202 kg/rai.       Economic evaluation of treatment showed that. Sesbania rostrata gave the highest net return over total variable cost in both RD7 and KDML 105 with net return over variable cost 1,274.50 bath/rai and 1,291.00 bath/rai, respectively. However, the MRR of all N-sourceswere higher than 40%. These 4 N-sources can be used in rice - soybean system with exceptable economic return. From these results, it revealed that Sesbania rostrata and soybean residues can be successfully established for rice in rice-soybean system. These can replace urea or inorganic fertilizer which have to be used.         Keywords : Sesbania rostrata, weed, panicle