PHYSICS — BARNETT. 209 



main portion of the rotor A, and coaxial with it, extend from the bearings BB 

 almost into contact with the rotor, and form electric screens. Except for the 

 openings shown in the figure, the space within the water-jacket is closed, with 

 the lumen bearings and brass plates. A slow stream of dry, dust- free air enters 

 this chamber through the inlet U. A resistance box RS, with total resist- 

 ance R -{- S, always high and constant, and battery T with commutator 5, 

 produce the voltage V^ (along R) used for calibration purposes. Battery 

 and electrometer connections with J and B are made by amberoid keys, i, 2, 

 3, 4, of special design, and an amberoid commutator not shown in diagram. 



The dielectric constant K of the insulator F is determined with the con- 

 nections shown in the lower part of the figure. The condenser is a 0.02 

 M. F. mica condenser, and key 6 an ordinary Kempe key. Two operations 

 are necessary: 



(i) Initially R is given a considerable value, Rj^; i, 2, 3 are closed, H 

 down, 4, 5 open, and the lever of 6 up. Then, in succession, 5 is closed 

 for 30 seconds, H raised, 5 opened, i opened, 4 closed, the lever of 6 lowered, 

 3 opened, H lowered, and the electrometer scale reading taken. The process 

 is repeated with the battery reversed, and the double deflection, d^, obtained. 



(2) Initially R is given a much smaller value, R2; i, 2, 3 are closed, 4, 5 

 open, H down, and the lever of 6 down. Then, in succession, the lever of 

 6 is moved to "insulate," 5 closed, the lever of 6 moved down and back to 

 "insulate" as quickly as possible, 5 opened, i opened, 4 closed, 3 opened, the 

 lever of 6 moved down, and the scale reading taken. The process is repeated 

 with the battery reversed, and the double deflection, d^, obtained. From d^^, dz, 

 Ry, i?2, the capacity of the standard condenser, and the dimensions of F, the 

 dielectric constant is readily determined for charging time 30 seconds, its 

 duration in the principal experiment. 



The mean magnetic flux <^ across a section of F perpendicular to its 

 length for a current of i ampere in the coil, the corresponding flux </>' for a 

 similar section of the thin conductor surrounding F, the mean voltage V 

 from the surface of the bronze cylinder to the earth, for one revolution per 

 second and a current of i ampere in the coil, the voltage Vq, the coil current 

 /, and the period of revolution T of the cylinder, are obtained by well-known 

 methods. 



The constants of the coil, etc., being determined, an experiment is per- 

 formed as follows, in two parts : 



(i) Calibration: Initially i, 2, 3 are closed, H down, 4, 5 open. Then, 

 in succession, 5 is closed for 30 seconds, H is raised and 5 opened imme- 

 diately, I is opened 20 seconds or 30 seconds after H is raised (an increase 

 of time producing no eflfect) , 4 is closed, 3 opened, H lowered, and the elec- 

 trometer scale-reading taken. The process is repeated with the battery 

 reversed. The double deflection d^, corresponding to the voltage V^ along 

 R, is obtained. 



