particular, measurements of Sr®® in the stratosphere at high latitudes at the end of the polar night might confirm the hypothesis that the seasonal increase is due to the subsidence of cold air in those regions. REFERENCES OE ARAPwN yp mairanddsmz . . . . G. G. J. F. Stewart, R. N. Crooks, E. M. R. Fisher, AERE HP/R 2017. Osmond, A. G. Pratchett, J. B. Warricker, AERE C/R 2165. Megaw, R. C. Chadwich, AERE HP/M 114. Libby, Proc. Nat. Acad. Sci. 42, 365, (1956). . J. Bryant, A. C. Chamberlain, A. Morgan, and G. S. Spicer, AERE HP/R 2353. . M. B. Dobson, Proc. Roy. Soc. London, A., 236, 187, (1956). . W. Brewer, Quart. J. Roy. Meteorol. Soc., London, 75, 351, (1949). . W. Kellogg, J. Meteor., 9, 446, (1952). . Miller, Petermans Geographische Mitteilungen, 95, 1, (1951). Appendix 1 THE RELATION BETWEEN THE CONCENTRATION OF Sr*° IN THE LOWER ATMOSPHERE AND THE SPECIFIC Sr®* CONTENT OF RAIN WATER An important assumption madein this report is that the specific Sr*° activity of rain water is proportional to the concentration of Sr® in the lower atmosphereand, in particular, that the observed seasonal variation of the former (Fig. 3) reflects a similar variation in the latter. The data given in Sec. 5.2 show that over a long averaging period the specific activity of rain water is fairly constant over the U. K. and is relatively independent of the amountof rain which falls. It is reasonable to deduce from this that the assumed proportionality between the specific contents of air and rain water is valid for long periods of sampling. A more detailed proof can be obtained by examining the factors which could affect this proportionality, namely dry deposition and the dependenceof the specific activity of rain water on the size of showers (Sec. 4). Dry deposition is generally a small factor under conditions of average rainfall and it has been calculated from published data! that the ratio of dry deposition to average rain deposition for dust of stratospheric origin (steep gradient in the atmosphere) is only 0.025, using a dry deposition velocity of 0.07 cm/sec (Sec. 4). Thuslittle error is in- troduced into the sampling system under average conditions, but the factor increases in inverse proportion to the rainfall. It has been calculated from the rainfall records that the effect of dry deposition on the 1955 peak of Fig. 3 is negligible but that the 1956 peak should be reduced relatively by 7%. The statistics of the individual showers falling during the sampling periods have been examined and relative corrections made for the corresponding washout efficiencies in accordance with Fig. 1. Again, the correction to the 1955 peak in Fig. 3 is small while the maximum to minimum ratio for the 1956 peak is reduced by 8%. The deduction therefore is that the 1955 peak reflects the air concentrations faithfully and that the 1956 peak does so to within 15%. 245 Keir