ATMOSPHEEIC ELECTEICITY IN HIGH LATITUDES. 69 



be seen that all four curves are practically parallel* and are similar in shape to that of 

 the ionization. This shows at once that both the curves of the dissipation and 

 ionization are correct, and that there is a real difference in the yearly course of 

 the two, and also that there is a closer relation between potential gradient and 

 dissipation than between potential gradient and ionization. 



APRIL I MAY | JUNE | JULY | AUGUST [SEPTEMBER 



OCTOBER | NOVEMBERIDECEMKRlJANUARY [FEBRUARY] 



Fig. 2. 



The value of the ratio !+/!_ shows a very distinct yearly period with a maximum 

 in the winter and a minimum during the summer. Later it will be seen that very 

 probably this ratio depends largely on the potential gradient, so that its yearly 

 period might be expected on account of the yearly variations in the potential 

 gradient. 



Daily Variations. 



Potential Gradient. The daily course of the potential gradient varies greatly 

 according to the season of the year. For this reason five curves of the daily course 



* The lowness of the two curves for wind strengths 0-1 and 1-2 during the first part of the winter is 

 due to the fact that, owing to the darkness at both the morning and evening observations then, it was 

 impossible to see if the smoke of the village was drifting towards my place of observation or not. Nor 

 was I quite aware then of the fact, which I found later, that with no wind the smoke of the village 

 extended in an almost invisible haze over the whole valley, out of which it could not get. This smoke 

 effect, of course, only acted when there was insufficient wind to drive the smoke away, and its effect is 

 not at all visible on the two curves with wind strength greater than two, i.e., a steady breeze. 



