The study of experimental infections has afforded the opportunity to examine the response to antimicrobial therapy in the complex milieu of living animals. This review summarizes the information gained from studies of combination antibiotic therapy for experimental endocarditis, meningitis, and other infections in which host defenses are unable to pal ticipate in effecting cure. The studies cited suggest that antibiotic combinations that yield rapid microbial killing are more effective in the treatment of these selected infections than are those combinations that do not give enhanced killing. In addition, the relevance of various methods of demonstrating antimicrobial synergy is examined, and in light of the results of treatment of experimental infections, suggestions for the applications of these methods to the evaluation of the efficacy of antibiotic combinations on infections in humans are given.