As has been shown in previous reports from this laboratory (10, 25, 75), toxic effects resembling traumatic shock have been observed following the intravenous injection of fluids exuding from-dog muscles after prolonged ischemia.The fact that 28 per cent of these fluids were highly toxic, whereas the remainder seemed harmless, strongly suggested that their toxic effects were not due to the presence of intracellular substances lost as the result of cell damage (70). For the same reason, it appeared doubtful that the occasional toxic effects were caused by breakdown products of muscle cell constituents resulting from ischemia, since the experimental conditions were all approximately identical. Attention has been directed, therefore, to the one constituent of these fluids which was obviously highly variable, namely, the bacterial flora.Another bit of indirect evidence pointed in this direction. On several occasions, dog muscle tissue was excised and minced with sterile precautions and incubated in sterile dog plasma at 370 C., for 5 hours. The plasma was then separated by centrifugation and injected intravenously into recipient dogs. Such material was invariably toxic, rapidly producing a shock-like state and death, and it was always infected with micro-organisms which on direct smear were seen to be large gram-positive rods. However, if a mixture of blended dog muscle and plasma were sterilized by passage through a Seitz filter before incubation, of the fluid to be tested were mixed with 4.5 ml. of beef broth, and 3 successive 1: 10 dilutions in broth were made. Two blood agar plates were then divided into quadrants and a loopful (approximately 1I4.) of each dilution was streaked onto 1 quadrant of each plate. One of these was incubated aerobically and the other in an atmosphere of 5 per cent CO2 in nitrogen. After 36 hours of incubation at 37°C., differential colony counts were made and identification of the bacterial species present was undertaken. Colony counts were multiplied by the appropriate factors to give numbers of viable bacteria per ml. of original fluid, and these were checked qualitatively with observations on direct smears. Clostridia were isolated and cultivated from various fluids.In an attempt to reproduce at will the picture of toxicity following fluid injection, 2 strains of Clostridia perfringens thus isolated were used in 3 experiments. In these experiments, at the time of muscle ligation, 2 to 5 ml. of a broth culture of these organisms were injected into each of the 2 muscle groups; after the period of occlusion of the blood supply the ligatures were released as usual and fluid collected. These have been termed "reinforced" fluids. Bacterial counts were made on the first fluid obtained from each leg, on a final sample drain-856