Washed human erythrocytes were incubated with titanium dioxide (TiO2) particles at 37 degrees C for 1 hr and hemolysis was determined by the percentage of hemoglobin released (optical density at 540 nm; OD540) from the cells. Effects of TiO2 on OD540 were corrected and dose-response curves were analyzed by the Hill plot. Judging from the estimated dose to cause 50% hemolysis, the anatase form of micron-scale (<5000 nm) particles was 73 and 11 times more potent than the amorphous (<50 nm) and rutile (<5000 nm) forms, respectively, whereas it was 1.3 times more potent than the nano-scale (<25 nm) anatase particles. Plasma abolished the hemolysis due to anatase and rutile forms. Thus, hemolytic effects of TiO2 can be greatly different depending on the polymorph but not on the primary size (nano- or micron-scale) of particles. TiO2-induced hemolysis is unlikely to occur in vivo because of the presence of plasma.