Observations on various Liquids. 95 



the difficulty of cleaning it, I found it not a good liquid to 

 work with. After many rinsings of the cell, the electrified 

 balls were still connected by visible chains of particles ; and 

 although the larger particles fell to the bottom of the cell in a 

 little time, and the desired effect came out very clearly, still 

 the liquid was hardly ever purely transparent, but generally 

 somewhat misty or faintly speckled. With this one drawback, 

 xylol acted perfectly well in the electro-optic experiment (9). 

 The optical effect was clearly stronger than that of benzol. 

 The extinction -bands were well developed against a tension of 

 two pounds. The light restored by electric action was always 

 extinguished perfectly by horizontal compression of glass. 



17. Cumol (C 9 H 12 ). — This compound is somewhat viscous, 

 and not nearly so volatile as the preceding liquids. The only 

 sample of it that I have worked with is of a faint yellowish 

 colour, but purely transparent. When tested in the usual 

 way, it acts as a good insulator (10). The optical effect of 

 electric force is of the same kind in cumol as in the former 

 liquids, and is equally regular and pure, being always neu- 

 tralized perfectly by horizontal compression or vertical tension 

 of glass. This liquid is likely to hold an important place in 

 electro-optics : it is in all respects very easily managed as a 

 dielectric ; and after CS 2 , which it follows indeed at a large 

 interval, it gives an optical effect more intense and of longer 

 range than any other liquid yet examined. Against a tension 

 of four pounds in the fixed compensator, the extinction-bands 

 are developed almost as finely in cumol as in CS 2 . 



When the electric action and the compensating strain are 

 intense, the bands appear to assume a peculiar form in cumol. 

 Returning to the experiment described in (12), where the two 

 compensators were applied in combination with an intense 

 electric action on CS 2 , it will be remembered that the effect 

 of a strong compression of the hand compensator was to bring 

 the bands in towards the axis of the field, where they finally 

 coincided in one axal band. In cumol the effect takes a dif- 

 ferent form. As the bands approach the axis, they become 

 largely inclined to each other, converging from the outer 

 parts of the surface of the inductric ball towards the intersec- 

 tion of the axis of the field with the surface of the opposite 

 ball. A contrary form of effect was observed in xylol, where 

 the bands diverged from the axal part of the surface of the 

 inductric ball towards the outer parts of the opposite ball. 

 Other liquids gave traces of similar variations, but none so 

 distinctly as the two that I have mentioned. Although I have 

 not made a particular study of these phenomena, but have 

 merely noticed them carefully in passing, I cannot believe 



