The variations observed from time to time were, however, considerable, and there is 

 little doubt that the discrepancies must be ascribed to imperfections in the experimental 

 arrangements. The mean temperature in the cave during this period was 25 C., 

 and it may well be that the voltage of the dry cells used to furnish the high potential 

 was subject to variations. 



Since the magnitude of the variations (about 100 per cent.) lay far outside the 

 limits of observational error, an opportunity was made in the succeeding winter to 

 investigate this point more closely. For this purpose a hut was constructed of petrol 

 cases lashed together and furnished with a canvas roof, and arrangements made to 



TABLE 1. 



take an hourly photographic record of the variations. Except for the difficulty in 

 maintaining the clock of the barograph drum at low temperatures and in keeping out 

 drift snow, the arrangement worked satisfactorily until the hut was blown down in 

 a gale, but a record was obtained covering the greater part of the period May 20th 

 to June 9th, 1912. The variations from hour to hour were very large, the largest 

 value observed (in arbitrary units) being 270 and the smallest 119. In view of these 

 variations, it is hardly to be expected that any evidence of a regular daily variation 

 could be traced during such a short period and this is confirmed by a scrutiny of Table 1, 

 which give the mean hourly values of the natural ionisation for this period in arbitrary 



II 



