629 



obtained at both levels (observations disturbed liy drift, &<., being excluded)) we get the following iv.-uli- 

 for the high-level potential at each of the hours mentioned, and its ratio to the low-level potential : 



There is evidence of a daily variation at the higher level in agreement with that shown at the lower 

 level. 



Again, at the higher level as at the lower level, the mean potential at each of the hours is less in 

 January than in December. For the subsequent months the data are very scanty, but so far as they go 

 they show agreement between the variations at the two levels. 



6. ATMOSPHERIC POTENTIAL AND TIIK AUKOKA. 



The occurrence of auroral displays was without any obvious effect on the electrical field. 

 On several occasions during an aurora, additional observations of potential were made with an electro- 

 meter resting on the ship's rail. The actual values of the potential thus observed are, of course, of little 



3 Y 



