BLACKWELL. DISPERSION IN ELECTRIC DOUBLE REFRACTION. 657 



Optical Parts. 



The principal parts of the illuminating system are an arc lamp, short 

 focus objective, and Nicol prism. The arc lamp is entirely enclosed in 

 a sheet-iron case into which cool air from another room is forced by a 

 fan. This precaution is taken against fire, so that if the carbon bisul- 

 phide should by any chance leak, the heavy vapor could not creep into 

 the lantern and explode. The arc-circuit switch is opened under oil. 

 Since the vapor of carbon bisulphide ignites very much below red heat 

 a water cell is placed in front of the condensing lens of the lantern, so 

 that the slit which is at the focus cannot become hot enough to ignite 

 the inflammable vapor in case of a leak. A half-inch microscope objec- 

 tive sends a bright, thin, very little divergent beam from the slit 

 through the polarizing Nicol and up the tube. 



The Babinet's compensator is fastened to the face of the spectroscope 

 slit, behind which, in the collimating tube, is the analyzing Nicol. A 

 direct-vision prism-train receives the light, and a camera with long 

 focus achromatic objective takes the place of the observing telescope. 



j}wEz|fl — I — <s> » mf 



Figure 5. Arrangement of Apparatus. 



A half-inch objective is used to send a strong beam of polarized light through 

 the carbon bisulphide. 



The tube containing the electrodes and liquid is mounted sloping, in 

 order that air bubbles may not lodge under the spacer pins between the 

 bars during filling and cause spark discharge later. The arrangement of 

 apparatus is shown in Figure 5. 



The amount of double refraction due to ether drift, if there is such an 

 effect, is less than lO" 8 ^ according to Lord Kayleigh,^ and the values 

 of the electric double refraction obtained in the present experiments are 

 not less than 70 • 10" V» = (/* e — /*„), so that they cannot be affected 

 by more than 1.5 per cent. But if the effect of ether drift should turn 

 out not to be negligible the position of the tube should be known. It 

 is nearly in the magnetic meridian, and slopes down to the north about 

 23° from horizontal. 



1C Phil. Mag., 4, 678-683 (Dec, 1002). 



VOL. XLI. — 42 



