High dose, multiple alkylating agent chemotherapy is being employed in conjunction with autologous marrow transplantation in the clinic. We have investigated the scheduling of several alkylating drugs in an effort to optimize their antitumor effects. In vitro modeling of "continuous" (up to 72 h) versus "bolus" (1 h) exposure in MCF-7 cells showed that for N,N',N"-triethylenethiophosphoramide (thiotEPA), cis-diamminedichloroplatinum(II) (CDDP), 4-hydroperoxycyclophosphamide, carboplatin, and L-phenylalanine mustard (L-PAM) "continuous" exposure yielded essentially the same killing kinetics as "bolus" exposure. For N,N'-bis(2-chloroethyl)-N-nitrosourea (BCNU), however, even with fresh drug additions every 30 min, "bolus" exposure produced superior cytotoxicity. In vivo modeling of "continuous" (three i.p. injections over 9 h) versus "bolus" (single dose) administration of the alkylating agents cyclophosphamide, BCNU, thiotEPA, melphalan, CDDP, and carboplatin was conducted in mice bearing EMT6 tumors, and tumor cell killing as measured by tumor cell survival in vitro was compared with killing of bone marrow (CFU-GM) measured in culture as a representative sensitive normal tissue. With cyclophosphamide there was a considerable increase in the therapeutic index (killing of tumor cells/killing of CFU-GMs) when the same total dose of drug was administered in multiple injections versus a single injection. For BCNU and thiotEPA, smaller increases in therapeutic index were observed. With L-PAM and CDDP, some advantage to multiple versus single dose administration was observed, and for carboplatin a decrease in the therapeutic index was seen. In conclusion, for all six alkylating agents examined, the multiple dose schedule was at least as effective against the tumor as the single dose schedule at all dose levels.