684 ERIKSSON AND BOLIN the surface of the earth by precipitation since the tritium concentration in atmospheric moisture decreases downward. The importance of turbulence for the vertical transfer thereby is increased. A few remarks should be given here on the difference in vertical transfer of a radioactive gas such astritiated water and radioactivity in the form of particles such as "Sr. In the case of *°Sr, evaporation from and condensation onto falling droplets has no direct effect on the transfer of radioactivity by precipitation, which indicates that turbulent transfer and deposition may be more important for tritium than for sr, Furthermore, the transfer to a smooth ocean surface (more than 95% of the ocean surface is smooth) is due to molecular diffusion in the boundary layer next to the water surface. Also in this respect there is a marked difference between transfer of a gas and transfer of particles. Very small particles behave like air molecules and large particles (1,4) settle through the thin molecular boundary layer rather quickly, whereas the transfer of intermediate particles (0.01 to 1 ») is considerably reduced. It is well known that a considerable part of the radioactivity found in particles is in those of submicron size. We have so far restricted the discussion to vertical moisture transfer only and have disregarded the importance of horizontal flux. This may be justified as a first rough approximation over areas where local evaporation and vertical transfer exceed the large-scale hori- zontal convergence of moisture. For any study of the circulation of water in nature, the horizontal transfer processes are fundamental; and in this context it is of interest to study the geographical variations of the isotopic composition of precipitation. Dansgaard’ has shown how the 'O-content of precipitation systematically decreases toward the poles and inland over the large continents. He has deduced @ vlver relation between the '°0 content of precipitation and the mean air tem- perature. In a Rayleigh process the 180 content of the condensate depends on the isotopic composition of the vapor when it is evaporating from the sea surface and the fraction of the vapor that remains at the instant of precipitation.’ Since the amount of water remaining in the atmosphere at condensation depends on the condensation temperature and since this temperature on the average may be some fixed number of degrees lower than the mean temperature at the earth’s surface, such a relation between '%0 content and temperature may be expected. We note, however, that this is valid only poleward from about 55°N, where less than 10% of all oceans are found and where evaporation is rather limited owing to low temperatures. Only in these parts of the world can the successive removal of water vapor, as we approach the excessively cold arctic and antarctic regions, be approximately considered as a Rayleigh process; elsewhere the exchange between the atmosphere and the sea is of dominating importance, constantly main- taining the ‘89 content rather close to that of sea water. As was shown

Select target paragraph3