6 Gustaf W. Elmcn 



opposite ends, where contact was made by a couple of brass 

 clamps which also served to hold the plates in position. Copper 

 wires were soldered to these clamps and taken out through holes 

 bored in the end plates of the tube. The electrodes were sep- 

 arated in the same way as before. They rested without support 

 in the tube containing the liquid. Fig. 3 shows a vertical and an 

 end view of the electrodes. The formation of AS when the 

 electrodes were in contact with CSn Avas slow, and they could be 

 used for some time without resilverinsr. 



1_Z DJ 



rig.3 



The liquid used was CS.^. The ordinary commercial CS^ was 

 found to have a great number of small particles in it which 

 would vibrate between the electrodes when the charge was put 

 on, so that it was found very difficult to maintain a high differ- 

 ence of potential between them. It was therefore found neces- 

 sary to clarify the liquid by filtering it through a porous cup 

 which was fastened into the neck of a glass flask by means of 

 plaster of Paris. Immediately below the neck was a hole in the 

 flask into which a glass tube was cemented. This tube was con- 

 nected to a vacuum pump, and thus the liquid was forced through 

 the porous cup. The filtering of the CS^ did not, however, seem 

 to appreciably change the electro-optic constant of the liquid. 

 Table I gives the manner in which each recorded value of B 

 in the following tables was obtained. For voltages below 200 

 volts per millimeter distance between electrodes, each value of B 

 is the mean of twenty observations instead of nine. Tables II 

 and III give mean values obtained with the brass electrodes with 

 a=.2545 cm. and a— .184 cm. Table IV gives values obtained 

 with the silvered electrodes. The values between the three sets 

 agree fairly well. Table V gives values obtained for different 

 colors. For the same value of P for different wave-lengths. 6 

 remains approximately the same, and in calculating the value of 

 B it was only necessary to multiply the value of B, obtained from 



