Anaerobic bacteria are recognized with increasing frequency as etiologic agents in a variety of infectious diseases. The pathogenic significance of these microbes is supported by experimental studies with animals, chemotherapeutic trials, and definition of specific virulence factors by microbial analysis. In vitro sensitivity tests show that many clinically significant anaerobic species are resistant to penicillin; susceptibility to an expanding array of alternative antimicrobial agents is quite variable. Nevertheless, clinical trials suggest that several antibiotic regimens are equally effective. These studies provide guidelines for empiric selection of agents, although in vitro studies or the unique properties of some drugs must be considered in occasional cases. Several commonly accepted tenets held in the past may now be challenged: some abscesses do not require surgical drainage; plasmids conferring resistance to clindamycin in Bacteroides fragilis appear prevalent in some institutions; clindamycin may be the preferred agent for treatment of serious anaerobic pulmonary infections; and metronidazole is the only new agent advocated for anaerobic infections that offers potential advantages over prior agents.