Aqueous leachates from fragmented lalang (I. cylindrica (L) Raeuschel var. major,/ dead leaves in situ, caused delayed germination in biossays of all six species of pasture legumes tested, while radical and shoot growth were inhibited in two of the test species. Phytotoxicity was dependent on the quantity of plant residues available in the proximity of germinating phytometer species. Leachates from rhizome and root fragments exhibited strong inhibitory activity even at the 5% (w/v) level while leaf leachates were inhibitory at the 10% level and higher. In pot experiments with soil containing living lalang tillers or incorporated with rhizome or root fragments, phytotoxicity development in the soil extracts was evident at four weeks of growth or decay. In soils with decaying roots or rhizomes there was a sub-sequent loss in activity over the eight week incubation period Extracts of soils with living tillers showed loss of activity at six weeks of growth followed by a significant stimulatory response at eight weeks. The use of two controls (control-soil extract and distilled water control) enabled detection of inherent stimulatory activity in the control soil Comparison with both controls to eliminate such influences showed only living tillers, and decaying rhizomes and roots as major potential sources of inhibitory substances. The common soil organism, Bacillus cereus (Frankland & Frankland), Alcaligenes faecalis Castellani & Chalmers} and a Trichoderma sp. were found to be associated with the decay of rhizomes resulting in an initial stimulatory growth reponse in the phytometer species followed by inhibition and subsequent loss in activity. The nature of these organisms suggest that the release of phy to toxins through decay of rhizomes as a common, natural phenomenon in most soils. The dead leaves accumulating as litter in mature lalang vegetation may serve as an additional source of water teachable allelochemicah if sufficient quantities were available in contact with the soil.