A REVISION OF THE ATOMIC WEIGHT OF IODINE. 107 



RATIOS BETWEEN IODINE, SILVER, AND SILVER 



IODIDE. 



For this purpose iodine was purified as before by thrice converting it into 

 hydriodic acid with hydrogen sulphide, and then heating the hydriodic acid 

 with a sUght excess of potassium permanganate which had been freed from 

 chlorine by crystallization. The final product of iodine was distilled once with 

 steam, freed from water by suction upon a porcelain Gooch crucible, and dried 

 as far as possible in a desiccator over concentrated sulphuric acid. It was then 

 subUmed once from porcelain boats in a current of pure dry air in a hard-glass 

 tube, and then a second time from the hard-glass tube into the weighing-tube, 

 which was constructed as described on page 99 of this paper. This weighing- 

 tube remained constant in weight within a few hundredths of a milligram in each 

 experiment, and lost in weight in all only o.i mg. in the eight final analyses. 



Next, the iodine was dissolved in sulphurous acid, and was then precipitated 

 by adding a solution of a slight excess of pure silver to the solution of hydriodic 

 acid in a precipitating flask. The silver iodide clotted together very rapidly in 

 the presence of the nitric acid and excess of silver, and offered none of the diffi- 

 culties met when the precipitation is carried out with an excess of iodide, 

 for it may be washed almost indefinitely with pure cold water without showing 

 the least tendency to pass into colloidal solution. The silver iodide was collected 

 upon a Gooch crucible, dried, and weighed. The loss on fusion was determined, 

 and the asbestos shreds in the filtrate were collected. Twelve determinations 

 carried out in this way with solutions varying from tenth to thirtieth normal gave 

 results from 126.87 to 126.92. The variations were finally traced to the carrying 

 down of silver nitrate by the silver iodide; for the more concentrated the solu- 

 tions during the precipitation and the greater the excess of silver employed, the 

 lower was the observed atomic weight of iodine. The occlusion of silver nitrate 

 by silver halides is well known in the cases of silver bromide and silver chloride, 

 but in these cases it is possible to wash out the occluded salts by oft-repeated 

 treatment with water. The difficulty in removing the silver nitrate in the case 

 of silver iodide is doubtless due to the lesser solubility of this haUde. 



Similar observations have been made also by Kothner and Aeuer,^ who found 

 that with fifth normal solutions the carrying down of silver nitrate is very con- 

 siderable, and that an excess of iodide converts this occluded silver nitrate into 

 silver iodide only very slowly if at all. They found, however, that, if the pre- 

 cipitated silver iodide is washed with ammonia, the greater part of the occluded 

 matter may be eliminated, owing possibly to the slight solubiUty of silver 

 iodide in ammonia. On account of this newly discovered tendency of silver 

 iodide, it seemed probable that the results of the synthesis of silver iodide from 

 known weights of silver in the early part of this research were somewhat too low; 

 for although the precipitations took place in ammoniacal solution and an excess 



1 Liebig'si4«w., 337, 123 (1904). 



