The cytotoxic efficacies of several substituted anthraquinones, ametantrone, dihydroxyanthracenedione, dihydroxyanthracenedione diacetate, mitoxantrone, bisantrene, and doxorubicin, were evaluated on an established human colon adenocarcinoma cell line by the method of inhibition of colony formation. The concentration-dependent survival curve following treatment for 1 hr was biphasic exponential for all agents. At concentrations below 1 microgram/ml, mitoxantrone was about twice as active as both hydroxyl-substituted anthracenediones and doxorubicin, about 14 times more efficacious than ametantrone, and about 22 times more powerful than bisantrene. At higher concentrations, these differences in efficacy became even more pronounced. Treatment in stationary phase decreased the lethal efficacy of doxorubicin but not that of the other agents. No recovery of potentially lethal or sublethal damage was noted for any agent, but for anthracenedione derivatives, there was a small but statistically significant increase in cell kill during fractionated exposure. Continuous treatment with mitoxantrone or bisantrene resulted in marked degrees of cell killing, reaching 99.95 and 99.5%, respectively, after 24 hr. For doxorubicin, cell kill efficacy declined after 4 hr. Mitoxantrone was 10-fold more active on cells in G2 phase than on those in mid- to late-S phase. Sensitivity in G1 phase was intermediate. Thus, mitoxantrone appears as the most active compound while bisantrene and ametantrone are the least active agents. The cytotoxic efficacy of bisantrene increases during prolonged continuous exposure, while that of mitoxantrone increases in fractionated administration. These characteristics could be exploited in clinical strategies designed to improve the performance of these agents.