METHODS OF INVESTIGATION. 7 
to be impossible te secure complete absorption of the carbon dioxide 
and at the same time maintain adequate véntilation. 
In 1862 v. Pettenkofer * introduced the important improve- 
ment of diverting a known aliquot of both the ingoing and outcom- 
ing air for analysis. The results of these analyses, calculated upon 
the whole volume of air used, show the amounts of carbon dioxide 
and other gases added by the subject of the experiment. 
The Pettenkofer apparatus has the advantage of placing the 
subject under unquestionably normal conditions as to purity of 
air, of maintaining a practically uniform temperature and mixture 
of gases throughout the apparatus, and of dispensing with the ex- 
treme care necessary in the Regnault apparatus to prevent gaseous 
diffusion between the air outside and that inside the apparatus. 
Its great drawback is that it does not in practice permit the deter- 
mination of the amount of oxygen consumed.t To this is to be 
added the magnification of experimental errors involved in com- 
puting the results obtained by the analysis of small samples upon ‘ 
the whole volume of air used. 
. Despite these drawbacks, however, the Pettenkofer apparatus 
in various forms has been widely used, especially in experiments 
upon domestic animals, and has shown itself capable of yielding 
very accurate results within its scope. Laulanié,t by largely re- 
ducing the rate of ventilation, has been able to make determinations 
of the oxygen consumed which he regards as satisfactory, while 
_ Haldane § has constructed an apparatus for small animals, in which 
the entire air current is passed over absorbents before entering 
and after leaving the apparatus, which also permits of a satisfac- 
tory indirect determination of the oxygen consumed. Sondén and 
Tigerstedt || have also constructed a modified Pettenkofer respira- 
* Ann. Chem. Pharm., Suppl. Bd. II, p.1. See also Atwater, U. S. Dep. 
Agr., Office of Experiment Stations, Bull. 21, p. 106. ; 
+ Such a determination is theoretically possible from a comparison of the 
oxygen content of ingoing and outcoming air, but the delicacy of the measure- 
ments and analyses required is so great as to render the method impracti- 
cable, while the determination by difference concentrates all the errors in 
this one quantity. 
t Archives de Physiologie, 1895, p. 619. 
§ Jour. Physiol., 13, 419. 
| Skand. Arch. Physiol., 6, 1. 
