M. Stas on the Determination of Atomic Weights. 201 



Iodine. 



Stas . . = 126-857 



Marignac . . . . • = 126-840 



Marignac = 126-847 



126-848 

 These experiments of Stas, made with the most extreme care, 

 agree in a surprising manner with, the numbers obtained by 

 Marignac by simpler methods. 



HI. 



The last part of the paper contains new determinations of 

 the atomic weights of nitrogen, potassium, sodium, and lithium. 

 They were effected by the conversion of the chlorides into nitrates. 



The glass vessels used in this, as in the earlier parts of the in- 

 vestigation, were made from a special glass. Bohemian glass 

 alone resists, in the most powerful manner, the action of concen- 

 trated acids. By experiments at a glass manufactory, the author 

 was convinced that a lime-soda glass possessed the requisite 

 properties, provided it contained an adequate excess of silicic 

 acid. The following mixture was used : 



Silica 77-0 



Potash ..... 7-7 



Soda 50 



Lime 10-3 



1000 



in which equal atoms of Ca, K and Na are present, and which con- 

 sisted of the purest materials. On evaporating in a flask made 

 from this glass a pound of the most concentrated nitric acid, the 

 flask, after washing out the small residue, was only found to have 

 lost a milligramme in weight. Melting nitre caused no loss of 

 weight. Nitre, twice evaporated with concentrated nitric acid 

 and fused, diminished the weight of the flask by 2 '5 milligrammes. 



Pure nitric acid was prepared from commercial (spec. grav.= 

 1*5) by boiling and by two rectifications. The chlorides were 

 heated in a platinum crucible, and while still hot filled into hot 

 glass tubes provided with glass stoppers. It was established by 

 special experiments that, for the complete change of the chlorides, 

 three parts of nitric acid were required for the conversion of 

 IK CI, four parts for the conversion of INa CI, and 5*5 for one 

 part of Li CI. 



The weighed salt was introduced into a long-necked flask, 

 which, by means of a glass stopper, was connected with a 

 bulb-tube that dipped into a flask with water. When the liquid 

 was heated to 40-50°, decomposition took place without effer- 



