36 DENSITIES OF OXYOEN AND HYDROGEN 



tlie Toepler air }miii|). Coimecliuii is uuule by fusion, as in the previous series of 

 experinients. Tlie stojK'ock is o[ieiie(l hy the handle 7, Fig. 13, and when the 

 exliaustion has been completed, the stopcock is closed in the same way, the handle 

 is withdrawn, the ground joint is removed, and the desiccator is put in place under 

 the balance. The counterpoise is hung on one pan of the balance and the globe 

 on the other ami then released from the clamp, when it is ready to be weighed. 



23. O.WGKN UV SECOND MKrilOIi. DlIVINc; TIIK All! IN TIIK DKSICC.VroRS. 



It would be difficult to keep di'y the aii- in desiccators as large as these, 

 especially when open to the air during the weighings, unless the air is dried iu 

 some other apparatus and introduced into them. A current of dry air was accord- 

 ingly brought into the desiccator through a tube for the purpose, h. Fig. 13, which 

 delivered it at the bottom of the desiccator in a fanlike horizontal current. The air 

 was foi'ced by a hydraulic blower through two carboys containing pure sulphuric 

 acitl. While the globe was at the balance, this cui-rent passed through the a.xis of 

 the mechanism for weighing the globes by reversal, which was used in the second 

 series of determinations, and is described at page 42. The cui-rent was stopped 

 a few minutes before each weighing, but suffered little other inteiruption. When 

 the globes were filling or exhausting, one opening of the desiccator was closed and 

 the other was loosely stopped witli cottoii-wtiol ; an occasional introduction d' dr\ 

 air was then thought sufficient. 



24. OXYGEN BY SECOND MKIIIOD. FILLING GLOBES WTIH OXYGEN. 



When it was desired to fill a globe with oxygen, the desiccator containing it 

 had the cover a, Fig. 13, put in pl.uc. The ground Joint of the globe was cemented 

 into its corresponding piece, and to the latter was fused a tube long enough to pro- 

 ject above the tube 1; Fig. 15, which was then screwed into [(osition. The desic- 

 cator was then placed in the tank a a, near the globe of hydrogen and the differential 

 manometer. The tube h was then fused to the tubes leading to the air puini), the 

 manometer, and the ap]iaratus [nv producing oxygen. The filling with oxygen was 

 conducted in all respects as in the first series of determinations, the only difference 

 being that the key of the stopcock of the globe could not be reached by the hand 

 but by the handle g, Fig. 13. 



While the globe was filling, beams d d, Fig. If), were secuicd in position to 

 hold down the desiccator containiiii' the ^lobe and also the cylinder containiiiir the 

 globe serving as a standard of pressure and tempeiature. The tank was then filled 

 with water, which was well stirred by a current of aii-. AVheu the pressure of the 



