A procedure was developed to isolate and purify a toxin(s) from the sclerotia of Phymatotrichum omnivorum. Sclerotial exudate was obtained by soaking oven-dried sclerotia in sterile deionized water in the dark. After reducing to a small volume by SpeedVac, the exudate was treated with ammonium sulfate (60% and then 90% saturation) to precipitate proteins and centrifuged at 10,000 rpm for 30 minutes. The pellets, solubilized in a small amount of water, and the aqueous phase, reduced to a small volume, were passed separately through a Sephadex G-50 column and eluted with deionized water. Fractions were obtained by a fraction collector and aliquots were taken to determine carbohydrate and protein contents and aliquots of each peak were taken for toxicity assay on cotton suspension cells by the tetrazolium viability assay. Results indicate that the aqueous phase, but not the pellet, killed cotton cells and that the toxic fractions contained predominantly carbohydrates with a small amount of protein, even after deproteination of the exudate. The toxic fractions, pooled and reduced to a small volume, were further purified on a DEAE-cellulose column by eluting sequentially with deionized water, 0.05M and 0.5M potassium phosphate buffer pH 6.5, and 0.5M NaOH to obtain fractions Fl, F2, F3, and F4, respectively. In all four fractions, brown pigments were eluted and became invaluable visual markers for purification. Fl, F2, and F4 contained substantial amounts of carbohydrates, but all four fractions contained trace amounts of proteins. In speculation, F3 could represent a chromogenic moiety of the toxin(s) which was hydrolyzed inadvertently, or not, during the purification process. However, only F4 was toxic when tested against cotton suspension cells. To determine the effect of heat on the toxin, the exudate was autoclaved and purified. The unautoclaved exudate showed similar fractionation profiles to that of the autoclaved exudates. Toxicity properties was not lost in autoclaving. Thus, the toxin or toxins from the sclerotia of the phytopathogen appear(s) to be (a) heat-stable oligosaccharide(s) with the presence of some peptides or amino acids, and small enough to pass through a 10,000 molecular weight cutoff (MWCO) membrane, but large enough not to pass through a 500 MWCO filter membrane.