THE CONSTITUTION OF TIIK I.I.KCTRIC SPARK. 191 



shown in fig. 4. The hall-hearing, AB, consists of the hub and axle of a bicycle 

 pedal, the hub being supported by two sets of three strong twisted cords, stretched 

 and fixed as shown in fig. 3, which gives a back view of the apparatus. The same 

 figure also shows a narrow vertical tube for oiling the bearing. The moving parti 

 are surrounded by a circular cast-iron box, which can be covered both in front and at 

 the back. The light forming the image of the spectrum passes through a rectangular 

 aperture cut into the front cover. Fig. 4 (Plate 8) gives a front view with the smaller 

 disc detached and placed in front of the cover. Fig. 5 (Plate 8) shows the whole appa- 

 ratus ready for experiment. The driving power was a half h.-p. electric motor capable 

 of revolving 33 times a second, and carrying a disc with three grooves. The disc was 

 provided with two pulleys of equal size, so that at high speeds it might be driven from 

 two motors at opposite sides, thus avoiding the side pull The motor and spinning 

 wheel were tightly clamped to the table of a lathe, in such a way that the apparatus 

 could be dismounted and put together again, with all parts occupying the same 

 position. This was of importance, the method of focussing adopted rendering the 

 ivnioval of the disc necessary during part of the operation. The diameters of the 

 two steel discs were 33 and 22 -2 centims., the photographs being taken in the 

 annular space of 10'8 centims., left free when the smaller disc was placed on the 

 larger one. 



In all the experiments the driving cord was passed round the second pulley of 

 the motor, which had a diameter of 217 centims., while the groove of the pulley 

 on the spinning-wheel was cut to 2 inches, or 5'1 centims. The speed of the motor 

 was measured by an indicator of ELLIOTT BROTHERS, which was tested and found 

 correct. The ratio of the angular velocities of motor and disc might be obtained 

 approximately by calculation from the diameter of the pulleys, .allowance being made 

 for the thickness of the cord, or we might turn the motor slowly by hand, counting 

 the number of turns of the disc at the same time. For our purpose this would have 

 been sufficient, as the other uncertainties of the experiments do not at present allow 

 a very great accuracy, but in order to avoid any doubt, and on account of the interest 

 which attaches to the amount of slipping which takes place at high speeds, an 

 independent stroboscopic method was employed to determine the ratio of the angular 

 velocities. We were surprised to find no measurable slip, except at the highest 

 speeds, such as were never used by us in our experiments. 



In our experiments the disc revolved generally about 120 times a second, giving a 

 linear velocity of between 90 and 100 metres per second for that part of the film on 

 \\hich the photograph was taken. When the cord was passed over the largest 

 pulley of the motor we could spin the disc over 160 times a second ; this was tried 

 and the speeds were tested, but no photographs were taken, as the smaller velocities 

 gave us sufficiently good results. 



Our first successful experiments were made with a single-plate Voss machine, 

 the discharge being taken from four Leyden jars in the circuit, but the beauty 



