We measured the rate of generation of reactive oxygen species (ROS) [hydroxyl radicals (•OH) and hydrogen peroxide (H2O2)] catalyzed by ambient particulate matter (PM) in the dithiothreitol (DTT) assay. To understand the mechanism of ROS generation, we tested several redox-active substances, such as 9,10-phenanthrenequinone (PQ), 5-hydroxy-1,4-naphthoquinone (5H-1,4NQ), 1,2-naphthoquinone (1,2-NQ), 1,4-naphthoquinone (1,4-NQ), copper(II), manganese(II), and iron (II and III). Both pure compounds and their mixtures show different patterns in DTT oxidation versus ROS generation. The quinones, known to oxidize DTT in the efficiency order of PQ > 5H-1,4NQ > 1,2-NQ > 1,4-NQ, show a different efficiency order (5H-1,4NQ > 1,2-NQ ≈ PQ > 1,4-NQ) in the ROS generation. Cu(II), a dominant metal in DTT oxidation, contributes almost negligibly to the ROS generation. Fe is mostly inactive in DTT oxidation, but shows synergistic effect in •OH formation in the presence of other quinones (mixture/sum > 1.5). Ten ambient PM samples collected from an urban site were analyzed, and although DTT oxidation was significantly correlated with H2O2 generation (Pearson's r = 0.91), no correlation was observed between DTT oxidation and •OH formation. Our results show that measuring both DTT consumption and ROS generation in the DTT assay is important to incorporate the synergistic contribution from different aerosol components and to provide a more inclusive picture of the ROS activity of ambient PM.