An experimental model for empyema thoracis in the Duncan-Harley guinea pig is introduced. Empyema thoracis development and early death (less than 14 days after bacterial inoculation) were noted after various concentrations and species were inoculated into the pleural space with a piece of umbilical tape, which was used as a cofactor. The effect of concomitant hemothorax was also tested. Group I (N = 90) had intrapleural inoculation of umbilical tape and various concentrations (10(4), 10(6), 10(8) organisms/ml) of various bacterial species, which included Staphylococcus aureus (N = 30), Escherichia coli (N = 30), and Bacteroides fragilis (N = 30). Group II (N = 90) had intrapleural inoculation of umbilical tape, 1 ml of autologous blood, and the same varying concentrations and species of bacteria as Group I. The observation period was 14 days, during which time early deaths were noted. Fifty-eight percent of the staphylococcal group of animals, 37% of the E. coli group of animals, and none of the B. fragilis group of animals developed empyema. Animals with empyema developed significant weight loss (p less than 0.05) and roentgenographic evidence of empyema, which was supported by postmortem pleural reaction and pneumonia scores (p less than 0.05). Higher concentrations of inoculated bacteria produced a higher incidence of empyema in the S. aureus and E. coli groups (p less than 0.05), but concomitant hemothorax did not increase the already high incidence of empyema and early death in the E. coli group. Empyema caused by B. fragilis did not develop, even with cofactors of umbilical tape and blood. Anaerobic infections in this model may require the presence of other aerobic or facultative organisms, the presence of necrotic lung, prior malnutrition, or a combination thereof.