Continuous exposure to inhibitory concentrations of methotrexate produces distinct rates of steady-state growth of murine leukemia L1210 and human leukemia CCRF-CEM cells in culture. Addition of thymidine to the medium produces reversal (6 to 40%) of this steady-state growth rate inhibition. This study utilized combinations of methotrexate and thymidine for an evaluation of the accompanying relationship between steady-state growth rate and changes in the ribo- and deoxyribonucleoside triphosphate pools. In L1210 cells exposed to methotrexate alone, the deoxythymidine 5'-phosphate (dTTP) pools decreased, whereas deoxyadenosine 5'-triphosphate, deoxyguanosine 5'-triphosphate, and deoxycytidine 5'-triphosphate (dCTP) remained relatively constant up to 70% inhibition of growth rate, with dCTP at a constant 112% of controls. The corresponding ribonucleoside triphosphates decreased only slightly. With the combination of methotrexate and thymidine resulting in up to 40% inhibition of growth rate, there was also a decrease in the dTTP pool while the other deoxyribonucleoside triphosphates remained relatively constant, and the corresponding ribonucleoside triphosphates again decreased only slightly. The dCTP pool was reduced to a constant 42% of control comparable to that produced by thymidine alone. With greater than 40% (with thymidine) or 70% (without thymidine) inhibition of growth rate, all pools decreased, but only dTTP was substantially reduced in proportion to the growth rate inhibition caused by methotrexate. The dTTP pool became depleted in spite of the presence of exogenous thymidine. Evaluation of CCRF-CEM cells indicated that inhibition of growth rate and nucleotide pool perturbations by methotrexate were similar to those observed in L1210 cells. However, in the presence of thymidine, inhibition of growth rate appeared related to decreased pools of dCTP, deoxyadenosine 5'-triphosphate, and deoxyguanosine 5'-triphosphate, rather than dTTP as was observed for L1210 cells. Hence, mammalian cells were capable of responding in a differential fashion to pharmacological perturbations, and this capacity may play a role in determining therapeutic selectivity. Since the ribonucleoside triphosphate decreases were slight and relatively uniform during methotrexate-induced perturbations, the deoxyribonucleoside triphosphate pools appear to be more directly related to inhibition of growth rate. The results are consistent with the concept that slight imbalances in the deoxyribonucleoside triphosphate pools dramatically inhibit DNA synthesis, as mediated through their interaction with DNA polymerase.