526 Experiments 



It really seems very difficult at first glance to attribute death to 

 a lack of oxygen in experiments where 12, 15, or 17 per cent of 

 it remained in the air. But this difficulty disappears after suffi- 

 cient reflection. 



Indeed, we know that when a bird dies at normal pressure in 

 confined air, this death is due (for the most part, at least; we shall 

 give a longer explanation later) to a lack of oxygen, or to speak 

 more exactly, to the too weak proportion, and more exactly still, 

 to the too weak tension of this gas in the ambient atmosphere. 

 This tension can be expressed at normal pressure precisely by the 

 figure which indicates the percentage. One may say, for example, 

 that at one atmosphere the tension of the oxygen of ordinary air 

 is 20.9; and likewise that the tension of the oxygen of confined air 

 which has become lethal varies somewhere between 3 and 4 

 per cent. 



According to this, the tension of oxygen at a pressure lower 

 than one atmosphere will evidently be represented by a number 

 obtained by multiplying the percentage by the proportion of this 

 pressure to normal pressure, both expressed for greater simplicity 

 in centimeters of mercury. Thus the pressure of the oxygen of 

 ordinary air at 30 cm. of pressure will be represented by the 



30 

 number 20.9 x — = 8.2. 

 76 



Applying now this simple calculation, whose formula is 



2 xP 



-, to all the figures listed in our table, we reach the results 



76 



given in Column 10. We see that at whatever pressure our birds 

 were placed, their death came when the pressure of the oxygen 

 was lowered to values varying between 3.0 and 4.3, which are 

 precisely the values with which air becomes irrespirable at normal 

 pressure. The table shows that, even at very low pressures, we 

 find figures (numbers 28, 30, and 33) which indicate the most com- 

 plete exhaustion and the weakest tension, when we have taken 

 sufficient precautions, to which we shall refer in a moment. 



The differences between the results of the analyses at different 

 pressures are exactly of the same order as those which separate 

 the results obtained at the same pressure. This stands out in the 

 most striking manner in the graph in Figure 18, which expresses 

 the results in Column 10. Thus at the same pressure of 24 cm., 



