78 



RESEARCHES UPON ATOMIC WEIGHTS. 



Fig. 4. — Solid aluminum drying oven. 



ature within a very few degrees by means of a Bunsen flame. We are indebted 

 to Dr. Arthur Stahler of the University of Berhn for the suggestion of this 

 method of heating. 



In order to purify and dry the air it was passed through a tower filled with 

 beads moistened with dilute silver-nitrate solution, through a tower filled 

 with small lumps of solid potassium hydroxide, then through 3 towers filled 



with beads moistened with 

 concentrated sulphuric acid, 

 and finally through a tube 

 filled with resublimed phos- 

 phorus pentoxide. 



The apparatus was con- 

 structed wholly of glass, with 

 ground joints, and was simi- 

 lar to that shown in fig. i, 

 page 8. 



After being heated, the 

 quartz tube was transferred 

 to a desiccator and was allowed to come to the temperature of the balance 

 case before being weighed. The quartz tube was then placed upon hard- 

 glass supports, in a horizontal position, one end being slipped into a larger 

 tube, through which could be passed a current of either dry hydrochloric- 

 acid gas or dry air. The other end of the quartz tube slipped into one of 

 the arms of a large U-tube filled with glass pearls, which served to con- 

 dense any silver-chloride vapor, which might escape from the quartz tube. 

 The other arm of the U-tube was connected with the flue of a hood, the 

 suction thus caused being sufficient to prevent the escape of gaseous arsenic 

 compounds from the apparatus. The quartz tube was protected from dust 

 by a covering of sheet mica. 



The usual method of procedure was as follows: The quartz tube containing 

 the silver arsenate being in place, a current of hydrochloric-acid gas was passed 

 through the tube, and the tube was slowly revolved with pincers tipped with 

 platinum wire in order that the salt might be thoroughly exposed to the action 

 of the acid. Neglect to do this at the commencement of the reaction always 

 resulted in the caking of the salt in the tube, thereby rendering the action of 

 the acid less rapid. The hydrochloric acid was dried by passing through three 

 towers containing beads moistened with concentrated sulphuric acid. The ap- 

 paratus for generating and purifying the acid was constructed wholly of glass, 

 and was similar to that shown in fig. i, page 8. 



In the earlier experiments the salt was gently heated from the commence- 

 ment of the reaction. To all outward appearance it was entirely converted into 

 silver chloride in a few hours. Upon fusion, however, it presented a very 

 cloudy appearance, owing to the presence of arsenic compounds, which could 



