50 ROYAL SOCIETY OF CANADA 



present case, however, it was found possible to fuse the chloride and 

 bottle it under the same favourable conditions mentioned above, so that 

 even this error must have been entirely eliminated. The apparatus 

 necessary for preparing and drying the nitrogen and, when necessary, 

 delivering a stream of dry air, was quite similar to that described in 

 the paper on the Atomic Weight of Caesium.^ Air saturated with 

 ammonia was j^assed over heated copper, and the resulting nitrogen 

 dried by passing through sulphuric acid, a tower filled with solid 

 potassium hydroxide, then over resublimed phosphorus pentoxide, into 

 the Eichards' bottling apparatus, where the fusion took place. 



The following figure will give an idea of this bottling arrange- 

 ment. 



A — Tube leading from phosphorus pentoxide ; B — weighing bottle : DD — hard 

 Fig. 1. — Bottling Apparatus, Horizontal Section. 

 glass tube ; E — platinum boat containing fused potassium chloride ; F — tube con- 

 taining phosphorus pentoxide to prevent any moist air from difEusiug inward. 



The fusion of the salt takes place in the platinum boat E, while 

 a current of dry nitrogen is passing through the apparatus. After the 

 salt has cooled the nitrogen is swept out with a stream of dry air; the 

 tube F is then removed, a long glass rod inserted and the platinum boat 

 and contents pushed over into the weighing- bottle B. A slight twist 

 given to the apparatus allows the stopper C to fall out of its nitch. It 

 i.-^ then pushed home, the apparatus disconnected at H, and the weighing 

 bottle and contents transferred to a desiccator. All the weighings are 

 carried out by substitution. In addition to the weighing bottle contain- 

 ing the platinum boat, another one of almost thp same size and weight 

 is prepared, and in this is placed a quantity of platinum equal in weight 

 to the platinum boat. This tube and its platinum conte]its is* then used 

 as a tare when weighing the other tul)e ^^lus the platinum boat. Under 

 these conditions no correction need of course be applied to the platinum 

 boat and bottle for the buoyancy of the air, but only to the salt which 

 is being weighed. 



The bottles were allowed to stand in a desiccator for two or three 

 hours before weighing. The salt was then dissolved in about two-tliirds 

 of a litre of water, contained in an Erlenmeyer flask fitted with a ground 

 glass stopper. iThese solutions were always neutral and quite free from 

 any insoluble residue. The platinum boat showed no signs of having 

 been attacked by the fused salt. 



* Log. cit. 



