BY DR. BURDON-S ANDERSON. 211 



laboratory of the Jardin des Plantes, in Paris. The analysis 

 is made in a circular mercurial trough, in the centre of which 

 is a well sixteen inches deep, and large enough to contain 

 about 12 Ibs. of mercury. The gas having been transferred 

 from the tube in which it is collected from the pump, to a 

 eudiometer, the latter is plunged into the mercury, in order 

 that its contained air may acquire the temperature of the 

 metal. It is then raised with the aid of a wooden tube-holder 

 until the level of the mercury inside is the same as that out- 

 side. The quantit}' of gas having been measured, a fragment 

 of caustic potash is introduced, which rapidly dissolves in the 

 few drops of water which always float on the surface of the 

 mercury. The column of mercury is then gently agitated by 

 alternately raising and lowering the eudiometer, which, after 

 the completion of absorption, is again plunged into the mercury. 

 The gas having been again measured, about a centimetre of 

 strong solution of pyrogallic acid is introduced with the aid 

 of a pipette with a bent beak. The agitation is repeated and 

 continued for some time. As soon as the absorption of the 

 oxygen appears to be complete, the tube is transferred to a 

 basin containing water, into which the mercury with the P3 r ro- 

 gallate of potash is allowed to fall. The residue, consisting of 

 nitrogen, is read over water. The results obtained by this rough- 

 and-ready method must necessarily be erroneous, not only be- 

 cause the measurements are inaccurate, but because the absorp- 

 tions must always be incomplete. If, however (as in certain 

 pathological inquiries), it is more important that the analy- 

 ses should be numerous than that they should be exact, it may 

 be available. For class illustrations of the general nature of 

 the blood gases, it is completely adapted. 



For more exact purposes the process of gas analysis has 

 been during the last few years much shortened by Frankland, 

 Russell, and others. With a view to the analysis of the gases 

 of drinking Water, Frankland has introduced an apparatus of 

 great simplicity (see Fig. 200), the working of which will be 

 readily understood by the diagram. It consists of two parts, 

 viz., a laboratory tube (&), in which the gas to be analyzed is 

 first received, and a measuring apparatus to which it can be 

 transferred from the laboratory, in order that its volume ma}- 

 be determined before and after each absorption. The measur- 

 ing apparatus consists of two tubes (a, 6), fixed vertically side 

 by side in a stand, surrounded by a chamber containing 

 water (??). They communicate below both with each other 

 and (by the long flexible tube) with a mereurj'-holder (<), like 

 that of Alvergniat's pump. One of them can be brought into 

 communication by the arm (g) with the laboratory tube ; the 

 other (b) is open at the top. A scale of millimetres is en- 

 graved on it, the zero of which is opposite o. A corresponding 



