184 PROCEEDINGS OF THE AMERICAN ACADEMY 



evaporated to a convenient bulk in platinum crucibles of 100 cm.^ 

 capacity, amyl alcohol was added, the water expelled by boiling, and 

 the heatinor continued for some minutes after the thermometer in the 

 liquid indicated 132° C, the boiling point of the alcohol employed. 

 The liquid was then decanted with care and the residue dried at a 

 temperature below its melting point and weighed. "When the chlo- 

 rides are precipitated in the manner described, the deposit generally 

 adheres so closely, and such particles as do remain loose settle so well, 

 that the supernatant liquid may be decanted to the end without appre- 

 ciable transportation of the insoluble residue. For the sake of per- 

 fect security, however, in this part of the manipulation the decanted 

 liquid was filtered under gentle pressure upon asbestos, with the aid of 

 the device which I have previously described for such purposes,* and, 

 after gentle heating, the increase in weight of the felt and the con- 

 taining perforated crucible added to the weight of the residual salt. 

 In no case did this increase exceed a few tenths of a milligram, and 

 often could not be detected. 



As a source of heat, a bath in which the sand of the sand-bath is 

 replaced by smooth asbestos board is a convenience, or a piece of 

 asbestos board simply, about 30 cm. square, supported by a broad 

 tripod and heated under the middle by a Bunseu burner, answers 

 equally well to secure every gradation of heat without danger of 

 igniting the evaporated alcohol. 



As a control upon the results obtained by weighing the residue as 

 described, the filtrate was evaporated in a large platinum crucible, and 

 the residue thus left gently heated and weighed. Though the evapo- 

 ration be conducted with extreme care, the residue is almost sure to 

 show some blackening, due to the carbonization of matter carried by 

 the alcohol, which will not disappear entirely without the application 

 of a degree of heat which the salts cannot bear without danger of 

 volatilization. The weight of the residue from the amyl alcohol itself 

 is small, — one portion of 50 cm.^ yielding 0.0003 grm. and its mate 

 0.0007 grm., — so that the data obtained by the evaporation of the 

 filtered alcohol of the experiments, if not quite so trustworthy as the 

 former testimony, may nevertheless serve the purpose of a very close 

 control. Both sets of data are given in the following table. 



* These Proceedings, Vol. XIII. p. 342. 



