Po-210, Pb-210, Ra-228 and Ra-226 were measured in surface sea waters along the 1989-1990 global traverse of the oceans using the new R.V. Hakuho-Maru. Where the traverse intersects other expedition routes, the data are generally confirmatory. In the high-productivity regimes like the Red Sea, and the Arabian Sea Po-210 is removed from the mixed layer at much faster rates than Pb-210. This fractionation occurs during scavenging presumably because Po-210 is, strongly sorbed by organic particles, whereas Pb-210 is more likely associated with inorganic detritus. The Po-210/Pb-210 activity ratios leaving the mixed layer by particulate transport can be estimated from the steady state balance of Pb-210 and Po-210 in, the surface waters for different oceanic regions, and are compared with those in the literature obtained by sediment-trap experiments. Although this comparison appears to merge, there exist some inconsistencies: which may be attributable to one of the following possibilities: (1) the model-derived atmospheric Pb-210 flux is overestimated for the North Pacific and the North Atlantic, or (2) the sediment-trap data do not represent the real Po-210/Pb-210 ratio in the vertical particulate flux due to some experimental artifacts, such as incomplete trapping and size fractionation. Further careful studies of sediment trapping including seasonal variation are needed to resolve this issue. Our Ra data confirmed that strong sources for Ra-228 exist in the Bay of:Bengal and the Southeast Asian continental shelf zone, whereas the Mediterranean and the Red Sea, though they are surrounded by land, are not effective sources of Ra-228 in the surface water. (C) 1998 Elsevier Science Ltd. All rights reserved.