The Bangladesh indigenous rice (Oryza sativaL. ssp. indica) variety "Boterswar" was grown hydroponically (48) in glasshouse at the University of Ryukyus for 55 days. At this stage, the average tiller per hill was 16, and the plants developed an extensive and strong root system and obtained an average of 143g of fresh biomass per hill. After harvesting the rice plants were stored at -20°C until use. Seeds of L. sativumL. were purchased from the Green Field Project (Kumamoto, Japan) and seeds of E. crus-galli L. were collected from the rice field of the Okinawa Agricultural Research Centre, Nago, Okinawa, Japan. Because of its known germination behavior, L. sativum was used as a model test plant for the bioassay (55), and E. crus-galli, which has developed resistance to many herbicides, is considered one of the worst weeds in rice production in 61 countries, including Bangladesh. Aqueous methanol extraction Extracts were prepared using the method described by Salamet al. (50) for isolating allelochemicals. A total of 3 kg of fresh rice plants (roots, stems and leaves) were blended and extracted with 15L of 80% (v/v) aqueous methanol for 48 h. The extract was filtered through one layer of filter paper (no. 2; Toyo Roshi Kaisha Ltd., Tokyo, Japan), and the filtrate was extracted again with the same volume of methanol for another 48 h and filtered then, both filtrates were stirred and concentrated at 400C in-vacuo to prepare the aqueous concentrate (100 mL). Plant extract bio assay Rice plants (100g fresh weight) were extracted and concentrated as described above for bioassay experiment. An aliquot of the aqueous concentrate (1, 3, 10, 50 and 100 mg fresh weight [FW] equivalent extract per mL final assay concentration) was evaporated on an evaporator at 400C until dry. Then, the dried sample was dissolved in coldmethanol (0.2 mL) placed on a sheet of filter paper (no.2) in a 3 cm Petri-dish, desiccated in a draft chamber and then soaked in 0.8mLof 0.05% (v/v) an aqueous solution of Tween 20 (polyoxyethylene sorbitan monolaurate, Nacalai, Tesque, Inc., Kyoto, Japan) as a surfactant. For the control treatment, methanol (0.2 mL) was added to a sheet of filter paper in the Petri-dish and evaporated, as described above. Ten seeds of L. sativum or E. crus-galli were placedon the filter paper and then incubated at 250C in a dark incubator. Germination was assessed every 12 h with a magnifying glass by counting the germinating seeds, searching for the rupture of the seed coats and the emergence of a radicle ≥1 mm (42), until no further seeds germinated (48 h for L. sativum and 72 h for E. crus-galli). The germination (%) was determined for the control (without extracts) according to methods by Salamet al. (50). For the seedling growth bioassay, tenuni form germinated seedling so fL. sativum and E. crus-galli were placed in the Petri-dishes and then incubated using the aforementioned procedure. The root and shoot lengths of the test species were determined after 48h of incubation. The growth inhibition (%) was calculated with respect to control (without extracts) seedlings. Purification of active substances in the ethyl acetate fraction According to Salamet al. (50), the aqueous concentrate was adjusted to pH 7.0 with 1 M phosphate buffer, separated five times against the same volume of ethyl acetate to obtain aqueous and ethyl acetatefractions. The biological activity of the aqueous and ethyl acetate fractions was determined by germination and seedling growth bioassays using L. sativum and E. crus-galli. The activeethyl acetate fraction was evaporated until dryness after standing with added anhydrous Na2 SO4 over night and then chromatographed on a column of silica gel (70 g, silica gel 60N, 70-230 mesh; ASTM, Kan to Chemical Co., Inc., Tokyo, Japan), eluting with a step wise gradient of ethyl acetate (10% per step, v/v; 150 mLper step) and methanol (300 mL) inn-hexane, affording 11 fractions. The biological activity of the collected fractions was determined using the L. sativum germination bioassay according to the above procedure, and complete inhibition was found in fractions obtained by elution with 70-80% ethyl acetate in n-hexane. After evaporation, the concentrate was filtered through a column of Sephadex LH-20 (60 g,GE Healthcare Bio-Sciences AB SE-75184 Uppsala, Sweden), eluting with 20, 40, 60, and 80% (v/v) aqueous methanol (150 mL per step) and methanol (300 mL). Themost active fraction was eluted with 60% aqueous methanol and subsequently evaporated until dryness. The concentrate was dissolved in 20%(v/v) aqueous methanol (2mL) and loaded onto a reverse-phase C18 Sep-Pak cartridges (Waters Corporation, Milford, Massachusetts, USA) and purified with20, 40, 60, 80% (v/v) aqueous methanol and methanol (15 mL per step). The most active fraction was eluted with 20% aqueous methanol and evaporated until dryness. The concent rate was finally purified by C18 reversed-phase HPLC (COSMOSIL 5 C18-AR-II; Nacalai Tesque, Inc., Kyoto, Japan), elutingat a flow rate of 3 mL/min with 50% aqueous methanol and detecting at 220 nm. Complete inhibition was detected for four peaks that eluted at 16.0, 19.0, 20.0 and 24.0 min as colorless substances. Mass Spectrometry with electro-spray ionization (ESI-MS) analysis was carried out on a Watersmass specttrometer. NMR spectra were measured in CDCl3 on BrukerNMR spectrometers (500 MHz for 1H and 125 MHz for 13C). Allchemical shifts were reported relative to tetramethylsilane (TMS). Optical rotation was measured in chloroform ona JASCO P-1010 polarimeter. Bioassay of the isolated compounds The isolated compounds were dissolved in methanol to prepare the concentrations of 1, 3, 5, 10, 30, 50, 100, 300, 500 and 1000 μM for each compoundand0.1, 0.3, 1, 3, 5, 10, 30, 50, 100 and 300 μM for a mixture of the compounds at a ratio of 1:1:1:1. The biological activity against E. crus-galli seedlings was examined using the above procedure. Statistical analysis To compare the results,the bioassays were carried out twice using a completely randomized design with three replicates. Significant differences between the treatments and controls were analyzed by Fisher's Protected Least Significant Difference test for each L. sativum and E. crus-gallispecies.The Type I error was set at 0.01 for all the statistical comparisons. The I50 (concentration of approximate 50% inhibition of the growth rate) value in the assay was analyzed from the regression equation of the concentration curves.