BAXTER. — A REVISION OF THE ATOMIC WEIGHT OF IODINE. 425 



tated with an excess of ammonium iodide. This precipitate of silver 

 iodide, if over 0.0005 gram in weight, was collected on a small weighed 

 Gooch crucible, and the filtrate, together with the original filtrate and 

 wash waters, was passed through a small filter repeatedly until clear. 

 If the precipitate weighed less than the above quantity, it was collected 

 wholly upon a filter. These filters were of course ignited and treated 

 as before. The weight of the silver iodide and asbestos was used in 

 computing the results, the weight of electrolyzed silver serving merely 

 as a check upon the weight of silver iodide. 



In order to drive off the last traces of moisture from the silver iodide 

 it was fused in a porcelain crucible. The bulk of the precipitate, freed 

 as completely as possible from asbestos, was transferred to a clean 

 crucible, which was then weighed with its cover. The crucible and 

 cover were placed inside a large porcelain crucible, and were heated until 

 fusion took place. A temperature much above the melting point of 

 silver iodide was avoided, since this substance is distinctly volatile at 

 higher temperatures. The loss in weight was then determined. The 

 fused salt when cold was light yellow in color, with no trace of darken- 

 ing, showing that no appreciable loss of iodine had taken place. 



The density of pure silver iodide was found by displacement of water. 

 Three determinations were made with material which had been fused in a 

 porcelain crucible and cooled by pouring the fused mass upon a cold tile. 

 The solidified salt was broken into small fragments and was introduced 

 into a weighed pycnometer for solids, which was then reweighed. Water 

 sufficient to cover the salt was introduced into the pycnometer and the 

 system was placed in a vacuum desiccator which was kept exhausted until 

 it seemed probable that all air had been extracted from the solid. Then 

 the pycnometer was immersed in a bath at 25° C. and exactly filled with 

 water. Finally the system was weighed. Two more determinations 

 were carried out with material which had been cast into sticks by pour- 

 ing the fused salt into a porcelain boat. These sticks were used in an 

 elongated form of pycnometer for solids.* The average of the last two 

 experiments is undoubtedly more accurate than that of the first three, 

 since it was extremely difficult to pump out the air from the more 

 porous material used in the first experiments. 



The vacuum correction for silver iodide, when weighed with brass 

 weights, calculated from this value for its specific gravity, is -|-0. 000071 

 gram for every apparent gram of salt. A vacuum correction of 

 —0.000031 gram was applied to every apparent gram of silver. 



* Baxter and Ilines, Amer. Cliem. Jour., 31, 222. 



