126 RESEARCHES UPON ATOMIC WEIGHTS. 



crucible, or nephelometrically by comparison with dilute standard solutions 

 of silver. 



Before evaporation, the supernatant solution was carefully filtered through 

 the Neubauer crucible and the precipitate of silver iodide was washed by de- 

 cantation several times with pure water to remove any adsorbed silver nitrate. 

 That washing was able to remove any adsorbed silver nitrate was shown by 

 treating a very carefully washed precipitate of silver iodide resulting from one 

 of the analyses with a solution of silver containing a few tenths of a milligram 

 of silver in 6 liters. After thorough agitation with the precipitate the solution 

 was filtered and evaporated as in the analyses, after washing of the precipi- 

 tate. Practically all this silver was found in the evaporated solution. Although 

 the silver determined in this way includes the silver iodide dissolved in the 

 mother-liquor, the error introduced is very small, the solubility of silver iodide 

 being probably at least as low as 0.0000035 gni- P^^ liter.^ The platinum boat 

 was not appreciably changed in weight. 



DETERMINATION OF MOISTURE IN IODINE PENTOXIDE. 



The determination of the water content of the iodine pentoxide when dried 

 under the conditions used in the analyses for iodine, was fully as difficult as 

 the iodine determination, and it was in preliminary attempts to determine the 

 water content that evidence was obtained which led to the adoption of the 

 above precautions for reducing the water content to a constant low value. The 

 method of operation consisted in brief in completely decomposing the pent- 

 oxide into iodine and oxygen by heating in a current of dry air, and, after re- 

 moving the iodine as far as possible by condensation and finally by a layer of hot 

 silver, collecting the water in a weighed phosphorus pentoxide tube. Great 

 pains were taken that the iodine pentoxide should be prepared for the water 

 determinations exactly as for iodine determinations. In order to avoid absorp- 

 tion of water between the heating and the decomposition, the decomposition of 

 the pentoxide took place immediately upon the completion of the heating with- 

 out interruption of the experiment, the weighing of the pentoxide, which need 

 not be very accurate on account of the low per cent of water, having taken place 

 before the long heating at 240°. 



The iodic acid used in the water determinations was not made from either 

 purified iodine or purified nitric acid, and most of it was prepared in glass ves- 

 sels. However, it was very carefully purified by repeated crystallization from 

 aqueous solution in platinum vessels with centrifugal drainage of the crystals 

 in each case, and gave every outward evidence of great purity. Not only were 

 the final crystals pure white, but the final mother-liquor was colorless. The 

 crystals gave a perfectly clear solution in water and left no appreciable residue 

 when decomposed in the platinum boat. 



1 Kohlrausch, Zeit. physik. Chem., 50, 355 (1904). 



