312 



ATMOSPHERIC ELECTRICITY. 



snow itself. If this were found to be negative, the conclusions would be proved beyond doubt. 

 The need to make such an investigation was realised while we were in the Antarctic, and I 

 had devised the method to be used, but my unexpected recall at the end of the first year 

 prevented the experiments being actually made. It is to be hoped that the simple experi- 

 ments will be made by the first observer who has the opportunity. 



Yearly variation of the potential gradient. 



It is customary always to exclude periods of disturbed weather when obtaining the 

 normal potential gradient. This practice has been followed in the following discussion, 

 with the further limitation that all periods when the wind velocity was greater than 5 miles 

 an hour have also been excluded. This was necessary in order to rid the observations of the 

 effect of drift. The procedure adopted was to tabulate the values of the mean hourly poten- 

 tial gradient during periods in which the wind velocity was to 5 miles an hour. To rid 

 the observations still further of the effect of recent or approaching high wind and drift, the 

 first and last hours of each period were also rejected. Thus every value used in determining 

 the normal potential gradient was not only obtained during periods in which the wind was 

 5 miles an hour or less, but also at least one complete hour separated it from past or 

 approaching bad weather. Needless to say all hours during which snow fell during calm 

 weather were also excluded. 



Table 157 contains the mean potential gradient for each month * and the number of 

 hours used in the determination. 



Table 157. 

 Mean monthly potential gradient at Cape Evans. Volts per metre. 



These values are plotted on figure 91, from which it will be seen that the yearly variation 

 is fairly regular with a minimum in May and a maximum in December. 



In other words the potential in the summer is distinctly higher than in the winter. 

 A similar variation of the potential gradient has been found wherever observations have been 

 made in the Antarctic. These are the Belgica Expedition, 71° S., 87° W. ; Hut Point, 78° S., 

 38 E. ; Port Charcot, 65° S., 64° W. ; and Petermann Island, 65° S., 64° W. It can therefore 

 be accepted without question that the potential gradient in the Antarctic is higher in the 

 summer than the winter. This is a most important result, for it is the exact opposite of the 

 conditions in the northern hemisphere, where numberless observations have shown that the 

 potential is higher in the winter than the sununer. The question at once arises are the 

 observations made in the Antarctic typical of the conditions over the whole of the southern 



* I very much regret that the values of yearly and daily potential gradient given in my paper 'Chief Results of 

 the Meteorological Observations made on Captain Scott's Antarctic Expedition ' published in the Quarterly Journal 

 of the Royal Meteorological Society , Volume XL, page 221, 1914, are not quite accurate. This was due to the care, 

 lessness of one of my clerks who made many mistakes in computing the tables of potential gradient. The er - 

 are not important and cannot affect any conclusion whicli may have been d^a^^^^ from the published values. 



errors 



