On the other hand, fractionation is usually much less from large explosions at or near the surface cf the sea. In these cases, the condensed particles are sea-water salts and water. Condensation is. late because cooling to 100°C or less is required for condensation, and even then the small size of the droplets allows escape of the radioactive gases, Much less variation then occurs in the composition of the radioactive fallout as a function of distance from these explosions near water. Thus, the most frequent form of testing at Eniwetok Atoll by means of surface barges in the lagoon just off islands tended not to result in depletion of particular fission products in the fallout on the local Eniwetok Islands. In particular, cesium-137 and ' strontium-90 are expected to be depleted much less in these local fallouts from tecte gwar ar under water ccmpared to tesis at cle itaud surface. An example of this reduction in fractionation for barge tests is provided by samples of the radioactive cloud from the Tewa test at Bikini Atoll.- Tewa was 5 megatons, and the sample analyzed was taken at 16,000 meters: at 2.2 hours after the event (Nathans, 1970). Specific Activity for 8.2 microns Radionuclide Specific Activity for 41 microns strontium-90 11 promethium-147 13 uranium 5 Leaving aside the dependence of fractionation upon particle size, an analysis has been made of fractionation for all particle sizes. (Freiling, 1961). ; These results also show relatively little strontium-90 and cesium-137 in the particles, consistent with the arguments above. BEST AVAILABLE COPY '