Double Refraction in Liquids. 489 



The last solution contains the same amount of gelatine as 

 IV., and was used at the same temperature. It was, however, 

 boiled and cooled several times. This treatment increased the 

 error due to depolarization materially, so that no reading 

 could be made with a velocity of more than 25 revolutions 

 per second. It also appears to have increased the absolute 

 amount of the retardation, but not to have essentially altered 

 the appearance of the curve. 



A solution containing ten grams in a litre was allowed to 

 stand five hours before use. When one of the cylinders was 

 turned through a small angle, say thirty degrees, the light was 

 depolarized, the whole field becoming bright and remaining 

 so. If the cylinder was now turned in the opposite direction, 

 the field first grew dark as the zero-point was passed and then 

 again bright as the motion continued in that direction. If 

 the rotation continued beyond this small angle the substance 

 of the structure seemed to break down, for the field did not 

 remain bright but required continued motion to keep it in 

 that state, but a very small velocity was sufficient ; a high 

 speed produced great depolarization, and a left-handed 

 rotation, instead of the previously observed right-handed, was 

 plainly evident. 



A study of the data given in the curves and tables for 

 gelatine leads to the following conclusions : — For small velo- 

 cities the double refraction increases with the speed of the 

 rotation, though not in proportion to it. 



This increase continues up to a certain point, where an 

 elastic limit see ins to be reached. Beyond this point, the 

 amount of double refraction decreases, and finally changes 

 sign as the speed is increased. With very dilute solutions, 

 one gram in a litre, for example, this breaking down takes 

 place at so small a speed that it cannot be observed, and 

 the decrease of the double refraction with the speed is very 

 slow. Hence it was not found to show this reversal of effect. 

 On the other hand, in a solution containing ten grams to the 

 litre, the depolarization is so great as to prevent observations. 

 The rate of decrease of double refraction is less in dilute 

 solutions than in stronger ones. With one gram to the 

 litre, the amount of double refraction is independent of the 

 speed of rotation. This falling off is less abrupt at higher 

 temperatures, as shown by solution VII., containing four 

 grams to the litre. At 28° the amount of double refraction 

 is independent of the speed. In the solution containing 

 twenty grams to the litre, at 27°, the double refraction 

 decreases very slowly. At 34°, a solution containing three 

 grams to the litre gave a retardation of '000098 X for a speed 



