Studies of penicillin action in vitro have indicated that, under given experimental conditions, the antibiotic is bactericidal rather than bacteriostatic (1-5), and it has been further demonstrated that it exerts its bactericidal effect on susceptible microorganisms only when they are actively metabolizing, a state indicated usually by rapid growth (6-9). Although there is also evidence that when it is administered in the initial phase of experimental infections due to susceptible organisms penicillin is bactericidal (10, 11), elucidation of its effect in vivo has been complicated by the participation of host cellular defenses which, under certain circumstances, may in themselves be extremely efficient (12 tact with the hardened basal portion. Two ml. of plain agar were added as a cover to complete the disk, and after the agar had hardened, the aluminum ring was removed.Preparation of inoculum. In the experiments herein described, a single strain of group A streptococcus, S23 Burbank, was employed. The organism was obtained originally from Dr. Maclyn McCarty of the Rockefeller Institute for Medical Research; identical stock samples were prepared by concentrating cultures of the organism, grown in beef infusion broth with 10 per cent sheep serum and 0.5 per cent dextrose (BIB-SD), and storing 0.2 ml.aliquots at -720 C. For the various studies described, 5 ml. of BIB-SD at pH 7.8 were inoculated with 2 loopfuls of rapidly thawed stock culture and incubated at 370 C for 16 hours, after which 0.5 ml. was transferred to 5 ml. of fresh BIB-SD and incubated for 4 hours. Tenfold dilutions were made in tryptose phosphate broth, and 1.0 ml. of the 105 dilution was added to 9.0 ml. of melted agar at 400 C and thoroughly mixed. The 0.2 ml. portion of this mixture, employed in the disk, contained 350 to 400 chains.Implantation of disks. New Zealand rabbits, obtained from a single breeder, and free of demonstrable infection were used. The animals weighed 2.5 to 3.0 Kg., and were maintained under standard laboratory conditions. Preoperative preparation included careful depilatation of the abdomen over a rectangular area approximately 12 X 8 cm.; this area was then scrubbed vigorously with hexachloroprene soap (pHisoderm@). Laparotomy was performed under local anesthesia with 1 per cent procaine hydrochloride; aseptic technique was employed. A 6 to 8 cn. midline incision was made, and the disks placed so that contact between their surfaces was prevented. It was possible to insert as many as six disks on either side of the peritoneal cavity without difficulty. After implantation was completed, the peritoneum and muscle layers were closed with a continuous silk suture and the skin with Michele clips.Disks were removed through the same incision, and ground separately, without sand, in sterile mortars. In order to bring the final volume to 8.0 ml., 4.8 ml. of tryptose phosphate broth were added, and poured blood agar plates were made with 1 ml. aliquots of suitable tenfold dilutions of the ground disk suspensions for determination ...