THE ATMOSPHERE. 279 



closed end is afterwards graduated into hundredths of a cubic 

 inch, and the tube is bent in the middle, like a syphon, as repre- 

 sented by a in the figure. By a little dexterity, a portion of 

 the gaseous mixture to be exploded is transferred to the sealed 

 limb of the instrument, at the water or mercurial trough, and 

 the measure noted with the liquid at the same height in both 

 limbs. The mouth of the open limb may then be closed by a 

 cork, which can be fixed down by soft copper wire. A chain 

 being now hung to the one platinum wire, the other is pre- 

 sented to the prime conductor of an electric machine, or to the 

 knob of a charged Leyden phial b, so as to take a spark 

 through the mixture, which is thereby exploded. The risk of 

 the tube being broken by the explosion, which is very consi- 

 derable in the ordinary form of the eudiometer, is completely 

 avoided in this instrument by the compression of the air re- 

 tained by the cork in the open limb, this air acting as a recoil 

 spring upon the occurrence of the explosion in the other limb. 

 3. The combustion of the mixed gases may be determined 

 without explosion by means of a little pellet of spongy 

 platinum, and the experiment can then be conducted over 

 mercury in an ordinary graduated tube. 4. Another 

 exact method of removing oxygen from air, recommended by 

 Gay-Lussac, is the introduction into the air of slips of copper 

 moistened with hydrochloric acid, which absorb oxygen with 

 great avidity. 5. Lastly, a method lately practised by Saus- 

 sure ; in which the air is deprived of its oxygen by agitating it 

 with a small quantity of water and metallic lead in thin turnings, 

 which becomes white hydrated oxide of lead. All these methods 

 give accurate results when conducted with proper precautions. 

 The conclusion which they have led to is, that the proportion 

 of oxygen in 100 volumes of dry and pure air is not subject to 

 variation, and lies between 20.8 and 21 volumes.* It is gene- 

 rally assumed as 21 volumes, which gives the proportions 



ATMOSPHERIC AIR BY WEIGHT. 



Oxygen 23.1 



Nitrogen 7^.9 



100.0 



* Saussure, An. de Ch.ct de Ph. t. fi2, p. 219. 



