634 Experiments 



but it is stranger to see that the carbonic acid diminished equally, 

 although a considerable quantity remained in the air of the bag, 

 which should have increased the proportion of this gas. 



If we wish to compare, not the trend, which quite evidently is 

 the same, but the exact value of the variation of the blood gases in 

 diminished pressure on the one hand, and in asphyxia on the other, 

 we need only take the graph in Figure 31 and add to it the average 

 results of the last experiments, making the same calculation of the 

 numbers to bring the original value of the oxygen to 20, and that 

 of the carbonic acid to 40. 



This has been done in Figure 39. 



On the ordinates are plotted, as usual, the numbers relating to 

 the percentages of the gases extracted from the blood. 



The abscissae measure both the percentage of ambient oxygen 

 and the barometric pressure. So 20.9 corresponds to 76 centimeters; 



20.9 



a half-atmosphere, 36 centimeters, corresponds to , etc. . . .; 



2 



that will permit us to see whether or not there is agreement be- 

 tween the results of the two kinds of experiments. 



The points relating to diminutions of pressure are marked, as 

 we have already said, by little circles connected by dashes and dots 



— . — .o. — . — . — o. The dotted lines express the average of 



the two experiments of simple asphyxia which I have just reported. 



Now in regard to the oxygen, we notice at once the remarkable 

 agreement existing between the two curves; only for the rarefied 

 air, as I have already indicated, I could not go below a pressure of 

 17 centimeters, corresponding to about 4.7 per cent of oxygen. Here 

 is the first point gained. 



For the carbonic acid, the agreement is less perfect. But we 

 must note first that there remained in the air on the way to ex- 

 haustion a certain quantity of carbonic acid, without which the 

 dotted graph would certainly have dropped more than it has. 

 Furthermore, the irregularities between the averages connected by 

 the line — o. — .o are very great for the carbonic acid, as is shown 

 by the little isolated circles, which correspond to each of the ex- 

 periments. It is therefore probable that from a very great number 

 of experiments we should obtain averages which would be closer 

 together; but it seemed to me unimportant to continue this investi- 

 gation. 



The great interest lay in showing that, even at normal pressure, 

 if the proportion of oxygen in the respirable air is low, we find in 



