Hydrometers of Total Immersion. 509 



The probable error in t on account of 



B is +-0106, 



C ±-0198, 



B' ±-0080, 



C ±-0004, 



A +-0000, 



A' ±-0000. 



The total probable error in t is +'024° C. 

 Hence the method will give the point of absolute density 

 correct to about -^° C. with about 11 experiments. 



Determination of the Specific Gravity of Quartz. 



The hydrometer designed for solids was made by Messrs. C. E. 

 Miiller, in June 1898, of normal Jena glass. It was sent to 

 me accompanied by another of almost exactly the same shape, 

 size, and weight, which was used as a counterpoise. 



Its weight was 44* 13594 grammes, and its volume at 

 18°"591 was 136"58407 cub. centim. 



If P grammes of mercury, weighed in vacuo, are required 

 to bring the hydrometer to the same specific gravity as water 

 at t° ; and if the sum of W grammes of quartz and T l grammes 

 of mercury are required to bring the hydrometer to the same 

 specific gravity as water at t°, then 



■ Q _i lj M • 



In which p = tlie specific gravity at t° of water, 



M= „ „ ,, of mercury, 



and Q= ,, „ ,, of quartz. 



The mercury used was purified by allowing it to flow many 

 times from a glass tube drawn out to a capillary point, through 

 a long column of a weak solution of ferric chloride, and after- 

 wards through distilled water. It was then redistilled 

 in vacuo. 



Mercury to nearly the required quantity was poured into 

 a small weighing-bottle, and, while the latter was on the pan 

 of the balance, the necessary additions were made from a 

 20 c. c. dropping-bottle with the neck closed to a capillary 



