The anaerobic broth disk (AnBD) method of Wilkins and Thiel and a new modification, designated the thioglycolate broth disk method, were compared with an agar dilution technique. The thioglycolate broth disk method was incubated aerobically (AeTBD) or anaerobically (AnTBD). One hundred anaerobic bacteria representing 15 species were tested with clindamycin, chloramphenicol, erythromycin, penicillin, and tetracycline. Agreement of results by the two methods with minimal inhibitory concentration determinations were: AnBD, 95.2%; AnTBD, 91.5%; AeTBD, 94.5%. With clindamycin, chloramphenicol, and penicillin, the agreement of the AeTBD and agar dilution results was 100%, 100%, and 95%, respectively. Using the AeTBD method, only 1.1% of all tests gave false susceptible readings, whereas 4.4% gave false resistant readings. All susceptibility testing errors occurred with tetracycline, erythromycin, and, to a lesser extent, penicillin. For each method, the changes in designation of bacteria as being susceptible or resistant to an antibiotic between trials primarily involved strains with minimal inhibitory concentrations which were +/- one dilution of the respective breakpoint value. The same situation was true for most bacteria that yielded false resistant readings within each trial. False resistant readings with tetracycline were determined to be unrelated to excess cation content of test media. These results reaffirm the reliability of the AnBD method and indicate that the AeTBD modification is equally reliable. The greater convenience and lower cost of the AeTBD method should make possible more widespread performance of susceptibility testing for anaerobic bacteria in hospital laboratories.