[ARCHIBALD] ATOMIC WEIGHT OF POTASSIUM 49 



the atomic weight of magnesium. Chemically pure silver nitrate was 

 dissolved in enough water to give a solution containing about five per 

 cent of silver. Hydrochloric acid was then added until all the silver 

 had been thrown down as chloride. This precipitate was thoroughly 

 v^ashed with water and then reduced with an alkaline solution of moist 

 sugar; again thoroughly washed, dissolved in nitric acid and the nbovp 

 process repeated. After being reduced the second time, the silver was 

 fused on sugar charcoal before the blow-pipe, into pellets weighing about 

 20 grams each. One of these was dissolved in nitric acid, the excess 

 of acid driven off and the silver nitrate made up to a solution of about 

 25 per cent. In this solution the remaining silver was electrolysed., 

 using a very weak current. Under these conditions the silver separates 

 out in crystals of considerable size. These crystals were collected, 

 washed and fused on a boat of pure lime (made by igniting a mixture 

 of lime and anhydrous calcium nitrate) before the blowpipe. Every 

 precaution was taken to ensure the complete expulsion of any oxygen 

 which may have been absorbed by the fused silver. 



These pellets of pure silver were now cut up into pieces of the 

 proper size with a clean steel chisel, treated first with hydrochloric acid, 

 then with ammonia, and finally with a little nitric acid. They were 

 kept under distilled water until used. 



The water used was distilled and condensed in block tin vessels. 

 It left no residue upon evaporation, and was free from ammonium com- 

 pounds. Samples were always tested in the nephelometer ^ Ijefore 

 using for the presence of chlorides. 



The different acids used were redistilled, the first and last portions 

 of the distillate being discarded in each ease. The nitric acid which 

 was to be used to dissolve the silver was always tested in the nephelo- 

 meter for the presence of chlorine ions. 



Method of Analysis. 



In the case of the analysis given above (by Richards and Archi- 

 bald) ^ the potassium chloride before being weighed had been ground 

 very fine and heated in a current of dry pure nitrogen to a temperature 

 very little below the point of fusion: being subsequently bottled in an 

 atmosphere of dry air. It is probable that the error due to the reten- 

 tion of any water or mother liquor in the fine crystals at this high 

 temperature would be very small, more especially as the salt had been 

 precipitated from a hydrochloric acid solution, and the mother liquor 

 contained about the same per cent of chlorine as the salt. In the 



^ Richards. Proc. Amer. Acad., 1894, 30, 385. 

 Sec. III., VJ04. 4. 



