Prof. Hare's improved Eudiometrical Apparatus, 133 



exhaustion be carried to 400 degrees, and let 300 measures of 

 the mixture enter, so as to depress the mercury in the gauge 

 to 100 on the scale. An explosion being effected, the mercury 

 in the gauge will rise at first to about 215 degrees, and after 

 the gas shall have regained its previous temperature, will be 

 found somewhat above 220 degrees. 



Of course a deficit will have ensued of more than 120 

 parts, of which one-third, or a little more than 40 parts, will 

 be the quantity of oxygen in 200 parts of the air, subjected to 

 analysis. 



In order to ascertain the influence of temperature, a ther- 

 mometer is placed in the receiver, the state of which is noted 

 before and after explosion ; and the deficit is estimated either 

 by allowing for the difference produced by the temperature, 

 or awaiting the refrigeration until the mercury in the ther- 

 mometer be at the same height as before the explosion. 



From this account of the barometer-gauge eudiometer, and 

 those previously given of the sliding-rod instruments*, it must 

 be evident that I have contrived three methods of analysing 

 the atmosphere, or other mixtures containing oxygen or hy- 

 drogen gas. 



In the barometer-gauge instrument, the deficit is known by 

 its effect upon the mercury in the gauge-tube ; in one of the 

 sliding-rod instruments, the deficit is compensated by water, 

 and the quantity of this liquid which enters for this purpose, 

 is known by the portion of the sliding rod which remains with- 

 out, after excluding the residual gas. In the instrument with 

 the sliding rod and gauge, the deficit is compensated by intro- 

 ducing the rod, the gauge enabling us to know when it has 

 been introduced sufficiently; while the graduation shows the 

 ratio of the gaseous matter condensed, to the quantity confined. 



When the diversity of these methods is considered, it is 

 pleasing to observe but little difference in the results obtained 

 by them. 



A great number of experiments performed by means of the 

 barometer-gauge eudiometer, or those of the sliding-rod con- 

 struction, over water and over mercury, gave 20 -^^ as the 

 quantity of oxygen in 100 parts of the air. In twenty expe- 

 riments the greatest discordancy did not amount to £jjfc$th 

 part in 100 measures of air. 



In lieu of the glass receiver a strong metallic vessel may be 

 used, as for instance, one of the iron bottles employed to con- 

 tain mercury. The igniting wire may be placed so as to be 

 visible in a very stout glass tube projecting from the bottle. 



* See Phil. Mag. vol. liv. p, 209. 



But 



