Variation of Atmospheric Electrical Quantities. 291 



Eclipse Observations , August 21, 1914. 

 Instrumental readings were taken every hour from 4 A.M. 

 to 10 a.m., and then every half-hour to £ P.M. The duration 

 of the eclipse was from 10 a.m. to 1 p.m., with a maximum 

 phase of 0*7 for Eskdalemuir. The meteorological con- 

 ditions were not ideal, us in the early morning there was 

 a wet mist, which made observation very difficult. About 

 8-9 a.m. the air became drier, but it was still dull, the sun 

 being only visible for short periods, and then dimly through 

 thin Stratus and Strato-Cumulus. At 2 p.m. and 2.40 p.m. 

 a little rain fell, but not sufficient to prevent observation. 

 The diurnal variation observed is quite different from that 

 obtained on ordinary days. In order to compare the results 

 more conveniently, they have been represented graphically 

 in figs. 2 and 3. 



Figr. 2. — Eskdalemuir, 1914. 



per c.c.)xlO 



140 

 120 

 100 

 80 

 60 

 40 



Noon 

 10 II 12 13 



16 17 18 19 20 



(-Charge 20 



perc.c)X/O y 

 in e. m. u. 



Ratio 6 

 if Charges 4 



2 

 200 

 160 

 120 



Conductivity XIQ 40 

 =- in e.rn. u. 





























I 1 















Hc/i f 



seFc 



nod 







































Et 











N. _ 



^ - 















E + 

























' — 















-- 



~,~ 





•^ 









/ 







E- 





















^ 













L^ 















\ 



- 



~^<; 





























—-■ 































■i 



























1 































—7*J— 



."' 







- 



•*- 





















— y i 



























■ 1 



' 















































































* 













\~ 



, — 



















'/ "1 















- 







































































August mean diurnal variation (smoothed). 



Eclipse day variation (Unsmoothed) 



As the normal variation for August is only obtained from 

 6 or 7 days, it appeared desirable to smooth the values given 



in Table IV. by means of the formula cc + 2b + \ The 



