ROTATING A DIMM-CTRIC IN A MAGNETIC FIKLD. 127 



The speed counter was supported on a steel rod, so that it could be slipped up and 

 down into position, and oil was supplied to the screw thread through a pipe leading 

 to the oil reservoir. The numher registered hy the revolution counter in al>out 

 2 minutes was usually taken when determining the speed. It was found that the 

 speed olwerved agreed very nearly with that deduced from the speed of the motor 

 and size of the pulleys, and also that the speed remained very constant for long 

 intervals of time. In fact, no appreciable variation of the speed ever occurred 

 during any of the experiments, the results of which are recorded in this paper. 



The motor and solenoid were run from two different sets of cells, so that varying the 

 current in the solenoid did not affect the voltage on the motor. This latter voltage 

 was always indicated hy a Weston voltmeter, so that if any change in it occurred, it 

 was at once known that something was wrong. The current through the solenoid 

 was measured by a very good Weston ammeter, which read from to 20 amperes, 

 and it could be reversed by means of a mercury-cup commutator. 



The motor and ^the apparatus described were screwed down to a large wooden 

 board, which was supported on a stone slab in the Cavendish Laboratory. Heavy 

 weights were placed on the board, and the apparatus was very steady, even when 

 running at 200 revolutions per second. 



It remains now to descril)e the quadrant electrometer and its connections, and the 

 arrangement used for determining its sensibility. The quadrant electrometer used 

 was of the Dolezalek type.* It was placed on a separate table standing on the 

 floor, and was not affected at all by the vibrations set up by the rotating cylinder. 

 The quadrants were supported by " amberoid " pillars, which were found to insulate 

 very well. The needle was suspended by a quartz fibre, and was charged by 

 touching it with a wire connected to one pole of a battery of secondary cells, the 

 other pole of which was connected to the case and quadrants of the electrometer. 

 A. concave mirror 7 millims. in diameter was attached to the needle, and the 

 deflection of the needle was observed by means of the image of an incandescent 

 lamp filament, formed by this mirror on a millim. scale, distant 2 metres from the 

 electrometer. The image was $ millim. wide and perfectly sharply defined. Its 

 position could be easily read to millim. The loss of charge by the needle through 

 the surrounding air and the quartz fibre diminished the sensibility of the electro- 

 meter for potential difference about 50 j)er cent, in 24 hours. The needle was 

 usually charged up to a jwtential of 40 volts, at which potential the sensibility for 

 quantity of electricity was a maximum. 



During the course of a series of measurements the sensibility of the electrometer, 

 for quantity of electricity, was measured from time to time, but the variations were 

 inappreciable, owing to the very slow rate at which the sensibility of the electro- 

 meter, for quantity, varied with the jxitential of the needle when the sensibility foi 

 quantity was near its maximum value. The torsion of the quartz fibre used was just 

 * DOI.KX.AI.KK, 'Zeite. fur Instrument.,' December, 1901. 



