BY DR. BURDON-SANDERSON. 215 



In the preceding example such variations of temperature 

 and barometric pressure as may occur during the analysis are 

 disregarded. The readings are taken immediately after the 

 absorption of the carbonic acid gas ; as the time occupied in 

 the analysis up to this point is very short, the error arising 

 from the variations in question is inconsiderable. As regards 

 the absorption of oxygen, the error might be of more conse- 

 quence, were it not that the residue of nitrogen is so small. 

 As it is, it can be easily shown that it would require a differ- 

 ence of pressure amounting to three millimetres, and a dif- 

 ference of a degree of temperature, to make an error of one- 

 hundredth of a percentage in the result as regards nitrogen or 

 oxygen. Within these limits, therefore, the errors arising from 

 this source may be regarded as trivial. 



Although determinations of oxygen made by absorption 

 with hydrate of potash and pyrogallic acid are not entirely 

 free from objection on the score of accuracy, the results ob- 

 tained by the method above described are quite accurate enough 

 for most of the purposes of physiological research, for the small 

 errors are practically inappreciable, as compared with the varia- 

 tions in the proportion of oxygen contained in the blood to be 

 analyzed, produced by what might be regarded as very trifling 

 differences in the mode of collecting it. If it is desired to have 

 recourse to explosion with hydrogen, the best methods for the 

 purpose are those of Dr. W. Russell, and of Frankland, and 

 Ward. The following short description of the latter will be 

 readily understood from what has preceded. The apparatus 

 (Fig. 201) consists of two parts, corresponding to the labora- 

 tory-tube and measuring-tube of the instrument previously de- 

 scribed. The measuring-tube communicates, as in that instru- 

 ment, with a second tube (the one most to the right in the 

 figure) containing a column of mercury, by the height of which 

 the pressure to which the gas to be measured is subjected, can 

 be estimated. The chief difference is that, whereas in the for- 

 mer more simple instrument the pressure-tube is open at the 

 top, so that if air is contained in the measuring-tube, and the 

 stopcock by which it communicates with the laboratory-tube 

 is closed, the difference between the heights of the two columns 

 indicates the difference between the tension of the gas in the 

 measuring-tube and that of the atmosphere; in the instrument 

 now before us the tube is closed, and constitutes a barometer, 

 so that the difference expresses the actual tension of the gas in 

 inches of mercury. In the horizontal channel, by which the 

 measuring-tube and barometer communicate at the bottom, is 

 a three-way stopcock (not shown in the figure), by which they 

 may be brought into communication either with a vertical 

 escape-tube, the end of which dips into a receptacle containing 

 mercury several feet below, or with a tube open at the top (the 



