In the past decade, a MnO2-impregnated cartridge technique has been widely used to extract 234Th from seawater. One of the inherent assumptions associated with this technique is that all Th species in the dissolved phase are subject to extraction by the MnO2 cartridges. In this study, an intercalibration between the cartridge technique and a small-volume MnO2 coprecipitation technique was carried out to test this assumption. We demonstrated that the collection efficiency for 234Th could be substantially overestimated by the MnO2 cartridge technique. This may be the result of organic complexation of a significant portion of 234Th in seawater, causing this fraction of Th to pass through the MnO2 cartridges. The overestimate in collection efficiency may explain the deep-water 234Th deficit observed in some oceanographic settings (e.g., the Gulf of Mexico, the Middle Atlantic Bight, and the Gulf of Maine). Sensitivity tests show that using the cartridge technique can yield 234Th-based particulate organic carbon export rates that are overestimated by factors of up to 10. Furthermore, the frequent observed disagreements between Th fluxes recorded by shallow sediment traps and estimated using the cartridge method may be ascribed, at least partially, to this methodological issue.