SEDIMENT traps are widely used to measure the vertical flux of particulate matter in the oceans. In the upper ocean, sediment traps have been used to determine the extent to which CO2 fixed by primary producers is exported as particulate organic carbon 1-3. In addition, the observed decrease of particle flux with depth has been used to predict regeneration rates of organic matter and associated elements 3. Over seasonal or annual timescales, the import of limiting nutrients into the upper ocean (new production) should be balanced by particle export 4,5. Given the importance of accurately determining the sinking particle flux, it has been suggested that Th-234 might be used to 'calibrate' shallow-trap fluxes 6. Here I present a re-evaluation of existing Th-234 data which indicates that trap-derived and model-derived Th-234 particle fluxes can differ by a factor of +/- 3-10, suggesting that shallow traps may not provide an accurate measure of particle fluxes.