﻿648 Miss M. 0. Saltmarsh on the 



pass, and this gave the time of complete darkness, BC. 

 The time DE, during which the illumination was uniform, 

 was obtained in the same manner; and the total time of 

 illumination was given by CF. 



The observations were made as follows : — The lights were 

 adjusted at equal heights from the bench and at a distance 

 of about 190 cm. apart ; the photometer was set between 

 them at the same height in the position of equal illumination, 

 and the distances from the lamps measured to -^ mm.; the 

 sides of the spot were reversed, and the photometer was set 

 again. This was repeated independently three times for the 

 same distance apart of the lights, and the means of the dis- 

 tances a and b of the lamps from the photometer were taken 

 for one side of the grease spot, and the means a' and V for 

 the other side were taken. 



The ratio -=. = ~ was calculated. The light I x was then 



moved in 10 cm. and the process repeated, giving another 



value for ~ • The mean of the two was taken. 

 I2 

 For the whole set of observations the mean difference of 

 a value from the corresponding mean value was *57 per cent. 

 This represents the average error of the observations. If 



the brightness, when the disk is rotating, is - of what it is 



tc 



when the photometer is illuminated by the full light from 

 the lamp, the observation will give a value of k. 



If the residual visual impression does not alter the apparent 

 brightness of the surface of the photometer, k will also re- 

 present the fraction of the whole light which is transmitted 

 through the rotating screen. If it does, k will not be the 

 same as this fraction, and the illumination will appear 

 flickering when the residual visual impression only partly 

 overlaps the next stimulus. Thus flickering would disappear 

 for a certain speed of opening and closing the shutter, and 

 would reappear again for a higher speed. 



In the case of equal-sized shutters and apertures, the ratio 

 of the duration of full brightness to total darkness is 1. 

 With shutters which are smaller than the apertures the 

 ratio is greater than l,and there would be more overlapping; 

 while with apertures smaller than the shutters the ratio is 

 less than 1, and there would be less overlapping. 



In the tables I T is the intensity of the unscreened light and 

 I 2 that of the screened. I x was about 18*7 candle-power, as 

 compared with a 10 candle-power pentane standard. 



