Atmospheric-Electric Observations, 1915-16 389 



The fans are then stopped, and the first leakage tests are made simultaneously for the two 

 instruments. The fans are then started again, the conductivity apparatus is charged to its 

 appropriate potential, and the ion-counter fiber is released from earth. The time of 

 passage of the right-hand fiber of the conductivity apparatus across its first fixed mark is 

 then noted, and when the ion-counter fiber has reached its first mark the meter is read. 

 When the ion-counter fiber has gotten to its second fixed mark the meter is again read, and, 

 the fixed range for the conductivity apparatus having been previously chosen so that the 

 fiber of this instrument is, by this time, approaching its second mark, the time is noted 

 when that mark is reached. The conductivity apparatus is then recharged, the ion-counter 

 is earthed, and the operation is repeated, and so on. In this way the observations of the 

 ionic content are carried on simultaneously with those for the conductivity and so provide a 

 means for determining the corresponding specific velocity, which is recorded on the same 

 form as the conductivity and ionic content. 



PENETRATING RADIATION. 



For the measurement of this element a copper vessel of about 27 Uters capacity is 

 employed ; it is provided with a central rod which is insulated from the vessel by an amber 

 plug and connected to the fiber of a single-fiber electroscope. A potential of about 150 

 volts is apphed to the vessel and the ions of the corresponding sign produced therein are 

 driven to the central rod, so that on releasing the fiber of the electroscope from earth, it 

 commences to move at a rate determined by the rate of production of the ions in the 

 vessel. The principle of noting the time taken by the fiber in moving over a fixed range is 

 adopted here as in the other instruments and the observations are recorded as in the 

 example given on page 399. The quantity sought is, of course, the number of pairs of ions 

 produced per cubic centimeter per second in the closed vessel. 



The insulating plug which supports the rod is divided into two parts by means of an 

 earthed guard-ring, so that, since the apparatus is not subjected to air-currents from out- 

 side, leakage may be taken as neghgible. Indeed, leakage may be entirely eUminated by 

 starting observations with the fiber charged in such a sense that it crosses the zero during 

 the observations, for it may then be arranged that the fiber-readings which are chosen as 

 the bases of the measurements lie at equal distances on each side of the zero. The copper 

 vessel, which is hennetically sealed, was thoroughly cleaned on the inside before it was 

 installed. 



In order to avoid movements of the fiber of the electroscope resulting from variations 

 of the potential appUed to the large vessel, an additional attachment has recently been 

 incorporated. This attaclmnent comprises a cylindrical piece, which is in electrical con- 

 nection with the central system, and is surrounded by an insulated hollow cyUnder which 

 does not touch it. Although the volume of the attaclunent is small, the capacity of the 

 portion inclosed by the hollow cylinder is comparable with, and may be made nearly equal 

 to the capacity of the portion of the rod surrounded by the large vessel. The two ends of a 

 megohm are connected respectively to the large vessel and to the outer cyUnder of the 

 attachment, the mid-point of the megohm being connected to the case of the electroscope. 

 A battery of 300 volts is also connected to the two ends of the megohm. It will be obvious 

 that under these conditions, and when the capacities above referred to are adjusted to 

 equaUty, fluctuations of the battery potential may take place without affecting the potential 

 of the insulated system. The result is the same as if the battery did not fluctuate. In 

 the laboratory, Swann has used this method with success for producing the equivalent of 

 a battery constant to 1 part in 10,000,000 or more. 



