A human tumor clonogenic assay (HTCA) has been used to evaluate standard and experimental anticancer drugs with respect to their inhibition of clonogenicity of both fresh human cancers and human tumor cell lines. By comparing the inhibitory effect on tumor colony-forming unit (TCFU) growth of 1-h and continuous drug exposures in the HTCA we were able to identify and separate schedule-dependent (e.g., bleomycin, vinblastine, and etoposide) and schedule-independent drugs (e.g., actinomycin D, adriamycin, bisantrene, and cis-platinum). Vinblastine, bleomycin, and etoposide, which are known to have 'cell cycle-specific' characteristics, caused exponential reduction in tumor colony formation when given by continuous exposure, whereas when given with a short exposure each of these drugs caused plateau-type dose-response curves. For comparison of the relative efficacy of the two dosing schedules, a ratio (1-h versus continuous exposures) was calculated of the drug concentrations which reduced growth of TCFU to 50% of the control values (ID50) for fresh human tumors and human tumor cell lines. For fresh tumors, ID50 ratios for adriamycin, actinomycin D, and bisantrene ranged between 2 and 60 (median 14), whereas the ID50 ratios for bleomycin, vinblastine, and etoposide ranged between 100 and 3,000 (median 600). The fact that actinomycin D, adriamycin, and bisantrene (a new anthracene-type drug) had similarly shaped dose-response curves and very low ID50 ratios suggests that the cytotoxicity of these compounds may not be schedule-dependent. On the other hand, the steep dose-survival curves we observed after continuous drug exposure and the high ID50 ratios of bleomycin, vinblastine, and etoposide suggest that these drugs may possess schedule-dependent cytotoxicity characteristics. Before final conclusions are drawn concerning a drug's schedule dependency it is essential to evaluate its in vitro stability and protein-binding characteristics. Finally, it must be emphasized that unlike the results obtained with 1-h exposure studies, the in vitro continuous exposure schedules have yet to be shown to be predictive of clinical response for any agent or tumor type.