units, after correction for insulation leakage. The outstanding features of tliis Table 

 are the high mean values at 5 p.m. (time of ISHtli meridian) and 11 p.m. 



The mean values are given in Table L' The Table is unfortunately incomplete 

 owing to the omission of days dunn- \\liieh observations could not be recorded, but 

 the variations from day to day are clearly of the same magnitude as the mean hourly 

 variations. 



On consideration of the values, there seems little doubt the observed variations 

 from hour to hour must be due to imperfections in the experimental apparatus, but 

 there is some evidence that the mean daily values are subject to other variations, 

 which we would be inclined to say were due. to radioactive substances derived from 1 IK- 

 TABLE 2. 



air and deposited on the earth. No connection, however, has been traced with the 

 meteorological conditions prevailing during this period, though, somewhat strangely. 

 the mean hourly value associated with northerly winds (193-5) and southerly winds 

 > 15 miles per hour (192-5) is greater than that associated with calms and southerly 

 winds less than I 5 miles per hour (185-0). The difference may, however, be accidental. 

 and it is known that the mean radioactive content of the air is slightly lower at high 

 wind velocities, while the mean value is not greatly different from that observed over 

 the ocean.* 



If such a variation can be infem-d. a question at mice arises as to ll -i'jiu of 



the radiations causing the variation, in view of the small amount ol rock exposed on 

 the Antarctic Continent, the small radioactive content of the air and the time which 

 must elapse durin.ii the transfer of products derived from other parts of the earth's 



* '.M.'trnn.lo-y.' vol. I. ].. .",20. 



45 



