The antimalarial action of methylene blue (MB) was first noted by Paul Ehrlich in the late 19th century. Although it has only sporadically been adopted as a serviceable drug, the resolution of its antimalarial action seems warranted, as it is currently used for the treatment of various methemoglobinemias. In this work we have used MB, and its analogues Azures A (AZA), B (AZB), C (AZC), and thionin (TH), as well as the oxazine Celestine blue (CB) and azine Phenosaphranin (PS). All MB analogues inhibit the growth of various strains of Plasmodium falciparum in culture with IC50s in the 2 x 10(-9)-1 x 10(-7) M range, with the rank order MB approximately AZA > AZB > AZC > TH > PS > CB. The IC50s for a mammalian cell line were in the 3 x 10(-6)-4 x 10(-5) M range, and the rank order was TH approximately AZB > AZA approximately PS > AZC approximately CB > MB. As MB could affect cell growth through the oxidation of NADPH, we tested the action of the various compounds on the hexose-monophosphate shunt activity. Appreciable activation of the shunt was observed at 1 x 10(-5) M in both cell types, thus accounting for inhibition of growth of mammalian cells but not of parasites. All compounds were found to complex with heme in a rank order similar to their antimalarial effect. It is therefore suggested that MB and its congeners act by preventing the polymerization of heme, which is produced during the digestion of host cell cytosol in the parasite food vacuole, into hemozoin. In this respect, these compounds seem to act similarly to the 4-aminoquinoline antimalarials. All compounds effectively suppressed the growth of P. vinckei petteri in vivo with IC50 in the 1.2-5.2 mg/kg range, and MB and AZB suppressed P. yoelii nigeriensis in the 9-11 mg/kg range (i.e. at doses similar to those of chloroquine). The potential toxicity of these compounds may restrict their clinical use, but their impressive antimalarial activities suggest that the phenothiazine structure could serve as a lead compound for further drug development.