ATMOSPHERIC ELECTRICITY IN HIGH LATITUDES. 



found are slightly less than they would have been if observation could have been 

 made in all winds ; but the final results are very little affected. Observations also 

 could not be made during rain, owing to trouble with the insulation. Except when 

 it was impossible to observe, owing to these two causes, measurements of the 

 dissipation were made three times each day, between 7.30 and 9.30 in the morning, 

 between 12 and 2 midday, and between 6 and 8 in the evening. 



In expressing the dissipation, ELSTER and GEITEL'S example has been followed, 

 i.e., the dissipation is expressed as the percentage of charge lost by a charged body in 

 a minute. Thus a+ = TOO per cent, means that 1 per cent, of the positive charge on 

 any body will be dissipated in a minute ; similarly, a_ expresses the dissipation of a 

 negative charge. The ratio a_/a+ is written q. There are two methods of obtaining 



the mean value of this ratio, either - 2 ( ) or - ; in most cases these two are very 



n \a+/ Sa+ 



nearly equal. In this paper q is always obtained by the latter method. 



lonization. EBERT'S instrument for measuring the ionization was used at the same 

 time as the dissipation instrument. It was often possible, however, to use the Ebert 

 apparatus on days when the wind made -it impossible to use the dissipation apparatus ; 

 but, on the other hand, the insulation of the Ebert instrument would often fail, owing 

 to high humidity of the atmosphere, when satisfactory measurements of the dissipation 

 could be obtained. EBERT'S instrument also could not be used when the temperature 

 fell below 20 C., for then the oil in the air turbine froze and prevented the clock- 

 work running freely. EBERT'S method of expressing the ionization has been followed ; 

 the positive ionization is expressed as the number of electrostatic units of free 

 positive ions in a cubic metre of air ; similarly for negative ionization. The symbols 

 used to denote positive and negative ionization are 1+ and I_ respectively. The ratio 

 of positive ionization to negative, i.e., I+/I-, is written r, and the mean is obtained by 

 the process 2I+/SI-. 



RESULTS OF THE OBSERVATIONS. 



Yearly Variations. 



Potential Gradient. Table I. gives the monthly values of the potential gradient. 



TABLE I. Potential Gradient. 



VOL. CCV. A. 



