THE lONISATION OF THE AIR. 267 



For practical use, Gerdien's expression is incorrect. The 

 I 

 expressions log^ r„, which is, of course, a theoretical expression 



''i 

 for the capacity of a simple concentric cylinder condenser, can- 

 not be used in this instance. The point is important ; for results 

 obtained with the help of Gerdien's formula have been found to 

 differ by as much as 40 per cent, from results given by the 

 amended formula : 



where A' is the meastn"ed capacity of the system made up of (i) 

 the outer cylinder; (2) the inner cylinder: and (3) the thin 

 support to the inner cylinder which connects it to the electroscope, 

 and is necessarily itself exposed in part to the air current. 



The Ebert form of apparatus is much the same in jirinciple 

 as the Gerdien. but is more self-contained. The concentric 

 cylinders are vertical. The radius of the outer is of the order 

 of 2 cms. instead of 8, and the inner cylinder is fitted directly 

 to the leaf system of the electroscope. Air is again drawn 

 betw^een the cylinders by a turbine actuated by clockwork, the 

 number of c.c. passed per second being automatically recorded 

 by a calibrated andlemometer. Various observations allow of 

 the quantities //. 7'. and iicz' being separately calculated. 



Observations with the Ebert apparatus are tedious, and 

 unless the ionisation is abnormally great, cannot be conducted 

 quickly. This is an important disadvantage, as the state of 

 ionisation of the air, except under peculiarly favourable condi- 

 tions, varies continuous!}-. With the Gerdien apparatus, on the 

 other hand, the necessary observations for the calculation of the 

 conductivity can be made in a few minutes. 



Both the Gerdien and the Ebert types of apparatus have 

 come into general use, and, rather unfortunately, both are often 

 accepted as standard instruments. 



The writer has made observations in the open air at the 

 South African College with both instruments. The observations 

 were made in the most favourable weather conditions. The 

 instruments were set up near one another, and measurements 

 with each were made simultaneously. 



Values of the conductivity iicz' as given by either instrument 

 were found to be reasonal)ly consistent. But there was never 

 any agreement between any one value as given by the Gerdien 

 and the corresponding value as given by the Ebert. Thus, the 

 conductivity measured by the Gerdien was found to be con- 

 sistently of the order of lO"^, the conductivity contributed by the 

 positive ions being generally in excess of that contributed by the 

 negative ions. On the other hand, the conductivit}' measured 



