Suspended particulate matter was collected by sediment traps deployed in the Sargasso Sea (Site 82), the north equatorial Atlantic (Site E), the north equatorial Pacific (Site p), and the Panama Basin (STIE Site). Additional samples of suspended particles were obtained by in situ filtration at Site F., at the STIE Site, and in the GuatemalaBasin. Concentrations of dissolved Tn and Pa were determined by extraction onto manganese dioxide adsorbers at Site P, at a second site in the Sargasso Sea (Site D), at the STIE Site and in the Guatemala Basin. Sediment samples were obtained from cores taken near Sites E and P.Resul ts have shown uneauiv~cally that suspended particulate matter in the open ocean preferentially scavenges rh relative to Pa. This behavior could not Mave been predicted from the kno\m physical chemistry of Th and Pa. Dissolved 230Thl231pa activity ratios \.¡ere 3-5 at Sites P and D Particulate 230Th/23lpa ratios _,,!ere negatively correlated \i7ith the concentra tion of suspended particles. However, variable scavenging rates, as indicated by variable particle concentration, do not completely control the ratio at which 'Th and Pa are scavenged from solution. Major biogenic and inorganic components of trapped material were found in approximately the same proportions in the STIF. samples and in samples from Sites E and S2' Lower 230rh/23lpa ratios found in the SrIE samples must therefore result from subtle changes in the chemical properties of the particles. Consideration of 230Thl23lpa ratios in several depositional environments indicates that no single factor controls the ratio at which Th and Pa are adsorbed from seawater.Fluxes of 210'th and 2jlpa were less than their rates of production' in the overlying w'ater column in every trap at Sites S2, E,-3-and P. In the Panama .Basin, fluxes measured with the same traps ..¡ere greater than or tqual to their rates of production. These resul ts are a strong indication that even exti:emely reactive elements such as Th and Pa are redistributed within the oceans. Redistribution occurs because variable scavenging rates in different environments set up horizontal concentration gradients. Horizontal mixiug processes produce a net horizonUll transport of Tl: and Pa from areas of 10",1 scavenging -i-ates to areas of high scavengi,ng rates. Protactinium is redistrihuted to a gn:ater extent than Th. Fluxes of 230'1h can be used to set 1.ot"re1. 1 imits for liorizoiital transport of Pa even when absolute trapping efficiencies of the sediment traps are not known. Less than 50% of the Pa produced d t the open ocean sites is removed from the ""rater col uwn by scavenging to settling particles The remainder is removed by horizontal transport to other environments.At Sites E and P, 230'1h/23lpa ratios were identical in the deepest sediment tn1p sample and in surface sediments. Ho¡,rever, 230Th/232Tli and 23Lp~/232ih ratios were 2.5tirnes higher in trapped particlE!S than in surface sediments. The 2301'0/232'10 ratios "..ere 5.5 times higli.sr in particles filtered...