Observations on various Liquids. 245 



(10) When nitrobenzol is traversed by an intense electric 

 current, it exerts a purely birefringent action on transmitted 

 light. The action is similar to that of a positive uniaxal plate 

 with axis parallel to the line of discharge. 



58. In conclusion I shall give several reasons for preferring 

 liquid dielectrics to solids in electro-optic experiment. 



Liquids are free from temper and irregular strains ; and, 

 if we except some such bodies as oil of turpentine, they 

 are of themselves quite inactive in the polariscope. Solids, on 

 the contrary, such as moderately thick plates of glass, are 

 almost in every case irregularly active of themselves when 

 inserted between two crossed Nicols. This difference is of 

 much consequence when the work is delicate. 



The optical effects of electric force are more complex in 

 solids than in liquids, being partly due to electric force at 

 the point or points viewed, and partly due to mechanical re- 

 actions of distant and differently strained parts of the body. 

 I think it extremely probable that the remarkable phenomena 

 lately observed by Mr. Gordon in plates of glass were due to 

 these mechanical reactions. 



Liquid dielectrics are not permanently damaged in any 

 degree when they are traversed by disruptive discharge, while 

 solids in such a case are rendered useless. And to see the 

 practical importance of this difference, it should be noticed, 

 first, that the preparation of a solid for accurate examination 

 in electro-optic experiment is always a matter of some expense 

 and trouble, and, secondly, that we cannot submit any dielectric 

 thoroughly to such examination without running the risk of 

 electric discharge through its mass. 



When a plate cell has been carefully constructed once 

 for all, any number of liquids may be examined in it succes- 

 sively, each liquid (if only clean enough) adapting itself 

 perfectly to all the conditions of the experiment. Solid 

 dielectrics, on the contrary, have to be drilled and mounted 

 individually ; and this (as already said) is always a matter 

 of expense and trouble. 



The plate cell affords a geometrically constant field for any 

 number of dielectrics in succession, which is a matter of 

 capital importance in all comparative observations and mea- 

 surements. 



For the reasons now indicated, and notwithstanding the 

 strength of glass and some other solids as dielectrics, I think 

 it probable that liquid dielectrics will supersede solids 

 altogether in this line of experiment. 



Phil. Mag. S. 5. Vol. 8. No. 48. Sept. 1870. S 



