520 
After substitution in (8) we find for the part of A that depends 
on w: 
? 
nl ap 1 n dv 
2n elv ec vidn 
; ; 5 22 c 
or after introduction of n= DE I= 
u 
by the aid of which (9) of $ 9 follows immediately. 
11. Direct determination of the velocity of the beam. 
In I § 3 the method has been indicated by which the maximum 
velocity was determined. We have, however, also measured it 
directly by the following method, which was, not applied until the 
experiments with glass were undertaken, but which is described 
here, because it has confirmed the velocity determination on the 
supposition of a fly-wheel revolving with a constant angular velocity. 
To the beam BB is attached a black sereen with two slits 
S, and S,, across which threads are stretched for accurate refer- 
B 
IL ji § 0 I R 
NS 1 
I: | | OS 
u | ! | | 
| EE | 
B | | 
i 17 a 20 ab /f- a zó- : 16. 3 
ence. Consider only the slit S,, which moves in the field of light of 
the condenser £,. By the aid of the achromatic lens Z, a sharp image 
of S, is projected on the plane F of the circular plate of a light inter- 
ruptor used for a large galvanometer. This latter apparatus, to which 
our attention was drawn by Mr. WerTHeEIM SALOMONSON, and which 
was put at our disposal by him, consists of an electromotor with a 
centrifugal speed indicator. One of the five axes of the apparatus 
revolves 25 times per second. The aluminium plate 2 (diameter 50 cm) 
has 40 slits in its circumference, each about 1 mm. wide, so 
that 1000 flashes arise per second. The distance between two slits 
ta 
