An advantage of established tumor cell lines compared to fresh human tumor specimens used in sensitivity assessments is the possibility of repeated experiments. Ultimately a database of sensitivity profiles on a panel of cell lines can be made and the sensitivity to new drugs compared with historical data. A prerequisite of this strategy is a minimal interexperimental variation. The sensitivity of eight human small cell lung cancer cell lines to adriamycin, daunomycin, aclacinomycin A, and mitoxantrone was tested in the clonogenic assay. A covariation in the sensitivity to the drugs emphasized the importance of simultaneous drug testing on the same batch of cells. On one cell line (NCI-N592) the interexperimental variation was further evaluated and a significant correlation was found between preexposure culture conditions, size of S-phase, and sensitivity to adriamycin, daunomycin, and mitoxantrone. Rigorous standardization of the growth conditions prior to clonogenic assay reduced the variation in the sensitivity to adriamycin from a factor of five to only 10-15%. It is concluded that simultaneous experiments on the same batch of cells in drug comparisons should be used if possible. Specification and standardization of culture conditions are necessary in the comparison of drugs tested in different experiments. Inclusion of the same reference drug in all experiments may further increase the validity of comparisons in different experiments.