21.- 



AND RATIO OF THEIR ATOMIC WEIGH IK. 33 



-OXYOEN ]!Y SECOND JIETIIOD. STANDARD OF PRESSURE ANM) TKM I'Ki; A TURE. 



The globe a, Fig. 11, whose capacity was about sixteeu litres, contained the 

 hydrogen which was used to reduce to the standard temperature and pressui'e the 



Fu: 



-('.lobe containing liyflrogen for a standard of comparison. 



oxygen which was to be weighed. It was supported in the copper cylinder b />, 

 having a water-tight cover provided with a tube e. This globe was connected by 

 a somewhat flexible glass tube d to a difEerential manometer, of which part is 

 shown in more detail in Fig. Ua. The tubes, of 24 millimetres internal diame- 

 ter, contain platinum points which are in nearly the same horizontal plane. It is not 

 necessary that they be at exactly the same level, but it is important that they pre- 

 serve an invariable relation to each other. The gauge was therefore firmly secured 

 by brass brackets, a, Fig. 12, to a brass tube which was solidly attached to a brick 

 wall. The tube and the gauge touched nothing else during the whole series of 

 experiments, except that they were partly immersed in water. The branches of the 

 manometer were some eight3'-five centimetres long ; a tube b was fused into the 

 bend and provided with a stopcock c, by which the volume of the mercury in the 

 manometer could be regulated. At d, Fig. 11, was a tube leading to the Toepler 

 pumi. and to a supply of pure dry hydrogen. When the globe had been exhausted 



