Manchester Memoirs, Vol. xlviii. (1904), No. \%. 3 



the observed diminution of potential. A correction for 

 the sh'ght leak due to defective insulation having been 

 applied, the ionic density is calculated. Ebert found that 

 if we could separate all the positive and negative ions in 

 the air, we should obtain under ordinary conditions 

 between one half and one electrostatic unit of each kind 

 in a cubic metre of air. 



My addition to Ebert's apparatus consists of a tube 6", 

 5 cms. wide and 60 cms. long, which can be pushed over 

 Ebert's tube, so as to form a sufficiently air-tight joint. 

 Two side tubes H and K are open at the bottom but may 

 be closed by means of a cap. Inside one of the caps I 

 place a small tube containing 5 mgms. of radium bromide 

 enclosed in a short lead cylinder so that nearly all the 

 radiation is upward. Observation shews that the direct 

 effect of the radium on the electroscope is too small to 

 affect the result. If the air sucked through by A passes over 

 the radium in H, the strong ionization causes the potential 

 of the electroscope to diminish by 100 volts, in about 15 

 seconds. If the radium is placed in K, the same diminution 

 of potential requires the current to be maintained for about 

 half a minute. The difference in the observed times for a 

 given drop of potential measures the rate at which re- 

 combination has taken place while the air traverses the 

 distance HK which, in the apparatus used, was 40 cms. 



The rate at which positive and negative ions combine 

 is proportional to the products of the ionic densities of 

 each kind, so that if there are n ions per cc. of each kind 

 present, and represents the time 

 dn „ 



a is a constant which has been determined for dry dust-free 

 air, and which with the help of this apparatus may be 

 determined for ordinary air, by a method already employed 

 by Rutherford, Townsend and others. 



