Experiments on various Liquids. 161 



perfected by careful regulation of temperature and by the use 

 of a neutralizing plate. One turn of the machine gave a strong 

 restoration in the polariscope; and the effect was purely nega- 

 tive, being neutralized by tension of the hand-compensator in 

 a direction parallel to the lines of force, and regularly strength- 

 ened by compression. Good results of the same kind were 

 obtained with half a turn of the plate, and even with a quarter 

 turn. TVhen the electric force was intense, there was some 

 disturbance produced by movements in the liquid. When the 

 temperature was raised very high, the regular effect disap- 

 peared, probably from failure of the insulation; and when the 

 temperature was lowered to the point of solidification, and 

 particles began to appear in the liquid, the optical effect 

 became irregular and very weak, and (on the whole) appa- 

 rently positive. But when these exceptional temperatures 

 were avoided, the action of this alcohol was seen to be con- 

 stantly, strongly, and purely negative. All the following 

 members of the series were examined in the plate cell, and at 

 ordinary temperatures. 



Capryl alcohol (n==8) gave no spark in the test for insula- 

 tion, and was inactive accordingly when tried as a noncon- 

 ductor. Tried as a conductor, with the jar and two or three 

 turns of the plate, it gave a brilliant effect, which was found to 

 be purely negative, being strengthened by compression of the 

 hand-compensator in a direction parallel to the lines of force, 

 and neutralized by tension, weakened in fact from a high 

 intensity down to perfect extinction. Tried then without the 

 jar, by discharge at every turn or half turn of the plate, it 

 gave a clear and pretty strong effect, always purely negative 

 by both tests. 



Amyl alcohol (ft = 5) is weaker than the last, both in insu- 

 lating and in electro-optic power; for when tried as a con- 

 ductor, without the jar, it gave a mere suspicion of effect. 

 Tried then with the jar, and two or three turns of the plate, it 

 gave a pretty strong effect, which was found to be purely ne- 

 gative by one test, being strengthened by compression parallel 

 to lines of force, but not sensibly affected by tension. The 

 failure of the second test, weakening by tension, was evidently 

 due to the great abruptness of the electro-optic action, the 

 momentary darkening of a steady flame being less perceptible 

 than a momentary brightening. If any proof of the truth of 

 this explanation were required, it could be supplied by a simple 

 form of experiment, which I may describe further on. 



Butyl alcohol (»=4) is weaker than the last. Tried with 

 the jar, it gave a very good restoration by discharge at every 

 tenth turn of the plate. Five turns, powerful formerly with 



