Three methods for rapidly determining the susceptibility of Mycobacterium Tuberculosis isolates to isoniazid, rifampin, ethambutol, streptomycin, and para-aminosalicylic acid were evaluated in a large-scale, blind study. Two of the methods measured evolution of CO2 from radio-labeled substrate (14CO2), and one method measured incorporation of 3H-uracil into ribonucleic acid. Rapid indirect drug-susceptibility test results for nearly 300 isolates were compared with those obtained using a standard modified proportion technique. The 3H-uracil uptake method proved to be unacceptable. Over-all, the results obtained using the 14CO2 methods and the standard method were similar. In general, there was greater agreement between the 14CO2 and proportion techniques with drug-susceptible strains than with drug-resistant strains. Among drug-resistant strains, both 14CO2 methods were more reliable for determining resistance to rifampin than to other drugs. This study demonstrates that large-scale, blind evaluations of new laboratory procedures are valuable. Our results indicate that methods relying o the enzymatic release of 14CO2 should be further refined and evaluated.