DEPARTMENT OF TERRESTRIAL MAGNETISM. 303 



in the wall. The air surrounding the exposed face of the insulator near the 

 outer end of this tube is kept constantly above the dew-point by a small 

 electric heating-coil. 



Automatic records of the base-line and of the insulation are obtained several 

 times daily. The method used for these controls may be described as follows: 

 Just outside the end of the outer insulating support for the collector rod, but 

 still within the supporting tube referred to above, a rigid sulphur insulator is 

 introduced in the collector rod; this insulator is bridged across by a mercury- 

 contact device, the contact being broken several times daily by means of an 

 electromagnet so as to insulate the last section of the collector rod (including 

 the collector) from the remainder of the system. This condition is maintained 

 about 12 minutes, the electrometer system being automatically grounded for 

 one minute at the middle of the interval; thus the photographic record, before 

 the earth contact is made, will indicate the state of insulation of the main 

 system on the assumption that the special insulator is functioning properly, 

 while the record obtained after the earthing will indicate to what extent the 

 exposed insulators may be defective. This procedure eliminates the uncer- 

 tainty usually existing as to whether the recorded minima relate to the poten- 

 tial gradient or to the humidity and other conditions affecting the insulation. 



The conductivity apparatus is a duplex system consisting of separate units 

 for photographically recording the positive and negative conductivity. Each 

 of these units is a modification of the Gerdien conductivity apparatus sim- 

 ilar in general to that described by Swann, 1 but including various features 

 and improvements resulting from several years' experience with the original 

 apparatus. In this form the records are obtained by means of a quadrant 

 electrometer and accessories. 



One pair of quadrants of the electrometer is connected to the insulated 

 central cylinder and the other to the case of the electrometer, which is main- 

 tained at a potential of the order of 100 volts. The two pairs of quadrants 

 are permanently connected through a high-resistance radioactive cell of the 

 type developed by Swann and the author {Terr. Mag., vol. 22, pp. 1-21, 1917), 

 and the apparatus is readily adjusted for a linear relation between elec- 

 trometer deflection and conductivity. Provisions are made also for frequent 

 calibrations (see p. 302 and J. Opt. Soc. Amer. and Rev. Sci. Instr., vol. 6, pp. 

 852-858) and for elimination of spurious leak-effects by methods similar to 

 those described originally by Swann. 2 To eliminate effects arising from radio- 

 active materials in the atmosphere, which are deposited on the wall of the 

 air-flow tube or on the insulated central cylinder, the air is caused to pass 

 first through an auxiliary condenser of large capacity; the auxiliary condenser 

 is ordinarily earthed, but once each hour is charged to a potential several times 

 as high as that on the main cylinder. 



The outer member of the main pair of concentric cylinders of each unit 

 forms the middle section of a vertical air-flow tube extending through both 

 the roof and the floor. Air is taken in through a suitable hood above the 

 roof and discharged into a space between the raised floor and the ground. 

 There is thus little chance of air being drawn a second time through the 

 apparatus, and the maintenance of satisfactory insulation is greatly facili- 

 tated by the fact that dust and spray from the air-stream are prevented from 

 depositing on the insulators of electrometers and accessory apparatus. 

 Especial care has been taken in the design to secure accessibility and 



« See Annual Report of the Director of the Department of Terrestrial Magnetism, Year Book 

 of the Carnegie Institution of Washington for 1917, p. 278. 



J See Annual Report of the Director of the Department of Terrestrial Magnetism, Year Book 

 of the Carnegie Institution of Washington for 1917, p. 279. 



