An important property of certain bacterial parasites, notably those responsible for tuberculosis, leprosy, and brucellosis, is their ability to survive and even to replicate intracellularly within phagocytic cells. For the most part, research on host parasite interactions at the cellular level has been carried out by using leukocytes or suspensions of macrophages isolated from the peritoneal cavity or the bronchial tree. Although these investigations have yielded valuable information, they tell little about how the fixed phagocytes of liver and spleen interact with the bacteria that they remove from the bloodstream. The functional characteristics of cells in the mononuclear phagocyte system (MPS) may vary depending upon localization of the cells in the body. Since the outcome of an infection may depend upon clearance of bacteria by fixed macrophages, investigations of these cells in vivo will further the understanding of infectious diseases. Certain characteristics of brucellae make them well suited for in vivo studies to determine how phagocytic cells deal with ingested bacteria. Brucellae grow slowly, plate easily, and have little endotoxic activity, and some strains are not damaged by serum factors. In a previous report it was demonstrated that most of the brucellae injected intravenously into guinea pigs were removed from the blood by liver and spleen, and that the bacteria in the liver were localized within the vacuolar apparatus of phagocytic cells. The present investigation describes the fate of brucellae within this subcellular system of liver phagocytes. Cells of B. melitensis strain 16 M were labeled with 32P. When injected into normal guinea pigs, labeled, viable bacteria were taken up and inactivated in liver and spleen during the 60 min after infection. Both uptake and inactivation increased if brucellae were coated with antibrucella antibody. Neither viability nor radioactivity were lost when labeled brucellae were incubated for 60 min in vitro with normal guinea pig blood, liver homogenates, or in defined medium. Incubation for 12 hr with antibrucella rabbit immunoglobulin G similarly was innocuous. Livers were removed from infected animals at various times up to 60 min after injection and were separated into subcellular fractions. The numbers of total (determined by radioactivity measurements) and viable brucellae as well as the acid phosphatase activity in the various fractions were determined. Total bacteria and acid phosphatase activity were progressively transferred from the mitochondrial plus light mitochondrial (M + L) fraction to the nuclear (N) fraction. Viability of brucellae declined more rapidly in the N fraction than in other fractions. Examination of M + L fractions by isopycnic centrifugation showed a decrease in viability of both free brucellae and those in particles. The results indicated the formation of bacteria containing heterolysosomes which progressively increased in size and in which brucellae were inactivated. The antibrucella activity of phagocytes of guinea pig liver in vivo appeared to be greater than that of peritoneal macrophages from immune rabbits or of bovine leukocytes studied in vitro.