1407 
apertures P, and P, on either side of a horizontal line through the 
middle of the scale. In the position drawn in the figure the light 
can leave through P,, because the opening VU, in the moving screen 
allows this. When the screen rests against 6,, P, is covered, and 
O, comes in the position, in which it is possible that light is emitted 
through P,. 
As was explained in II § 11 an image of the slit arrangement is 
projected on the rotating disc , which is provided with radial slits. 
With a weakly magnifying telescope the image projected on the 
rotating dise is observed with intervals of 0,001 sec. In this time 
the beam moves about 1 cm. at the velocity used in the neigh- 
bourhood of 10 M/,... The observer sees the coloured scale S in the 
field of vision, and then the “star” P,, or rather the after-image of 
the light emitted by this star at a former transmission. The place 
of the star on the scale can be read accurately down to $ mm., 
and the distance from P, to S being about 50 mm., the velocity 
can be determined certainly accurately down to 1°/,. All the changes 
in the velocity of the beam are immediately visible, and the velocity 
corresponding to every photo taken can at once be noted down. 
It is necessary that the “star” moves at the same level at which 
the axis of the rotating disc has been placed, for else a small cor- 
rection must still be applied to the velocity. 
Thanks are due to Mr. W. M. Kok, assistant at the Physical 
Laboratory, for his valuable help in the execution of the arrange- 
meni for the velocity measurement. 
8. Results. The extreme values of the velocity which were directly 
measured in our experiments, were 918 and 994 cm/sec. 
There were made two series of measurements, which were dis- 
tinguished by the way in which the velocity was found. In the 
first series, A, the method of communication II, in the second, B, 
that described above in § 7, was followed. All the results for the 
effect were reduced to a velocity of 1000 cm./sec. 
Series A. 
When the measurements of the 34 separate photos obtained on 
11 plates are combined, the effect is found to be 0.247 + 0,006. 
When first the observations on each plate are combined, and the 
mean is taken of the results of 11 plates, the effect is found to be 
0,247 + 0,009. 
Series B. 
gives for the effect derived from 49 observations divided over 
13 plates 0,238 + 0,006, for the effect derived from the mean of 
the results of the 13 plates 0,240 + 0,008. 
91 
Proceedings Royal Acad. Amsterdam. Vol. XXIII. 
