Short-P eriod Electrometer. 243 



A sufficiently short exposure to give clear images of the spots 

 was obtained by rapidly moving across the rays a card with 

 a slit on it about 3 mm. wide. Fig. 4 shows one of the 

 photographs with the wave-curves and the spots. Each spot 

 was generally accompanied by two fainter images due to 

 reflexion at the two surfaces of the rotating mirror, but these 

 did not cause any difficulty in determining the positions of 

 the principal images. Three or four wave-curves were taken 

 on each plate, this being raised vertically in its own plane 

 between the exposures, and finally a series of about fifteen 

 expos ares for the stationary spots. 



The measurements were made on the negative by means of 

 a Beck travelling microscope. First the distance of each 

 electrometer spot from one side of the plate and the horizontal 

 distance between it and the corresponding fork spot were 

 measured. By interpolation the distance of the spots apart 

 for any position of one of them could be calculated. Hence 

 the portion of the tuning-fork curve corresponding to any part 

 of the electrometer curve, and also the ratio of the average 

 velocities of the two spots in any interval, can be found. 



In measuring the wave-length of the electrical wave on a 

 negative, this was first placed on the microscope-stage so that 

 the horizontal cross-wire was just below the lowest points of 

 the curve. The length occupied by a number of waves 

 (generally 6 or 8) was then measured, the position of each 

 terminal zero-point being determined by adjusting the 

 microscope until the vertical cross-wire coincided with the 

 point of the wave at the shortest distance from the horizontal 

 cross-wire. Each curve was measured a number of times 

 and on different days. In a good negative in which the 

 amplitude of the waves measured (and consequently the 

 curvature at the zero-points) w T as not very small, the individual 

 values found for the wave-length did not differ from the mean 

 by more than one part in 800. With flatter curves, however, 

 as were obtained with the largest condenser, greater deviations 

 from the mean value occurred, amounting sometimes to 1 

 part in 200. 



After the electrical wave was measured the corresponding 

 portion of the fork-wave was found, and the length of the 

 whole number of fork-waves most nearly coinciding with this 

 determined by the same method. The range of the readings 

 for a fork curve was about the same as for a well-defined 

 electrical curve on a good plate if the fork was sounding 

 loudly during the exposure. 



If X, A/, r, v f , are the wave-lengths of the electrical and 

 fork waves and the velocities of the spots across the plate, 



