506 Experiments 



permitted this comparison, that in going across lots instead of 

 following the highway, I should find out something useful and 

 strange. The reader may judge whether my hope was deceived; 

 I merely wish to clear myself in advance of the charge of lacking 

 logic which might appear well founded, without daring never- 

 theless, in spite of my great desire to do so, to declare that this 

 indirect, and as it were oblique approach should be in many cases 

 adopted as the general method of research. 



Now that this first question was given, I had to consider it from 

 all its points of view, and they are numerous. I could first consider 

 animals of the same species dying in closed receptacles under 

 pressures higher or lower than the normal barometric pressure. 1 

 could next compare to each other animals of different species in 

 similar barometric conditions. Finally, I must examine the action, 

 under different pressures, of respirable media the chemical com- 

 position of which differed from that of atmospheric air, because 

 this last consideration, applied to the theory of asphyxia, had given 

 Claude Bernard data of great interest. 



I therefore adopted these various points of view, and I shall 

 give an account successively of the results which experimentation 

 gave me. I shall begin with the study of ordinary air and end with 

 that of air of different composition, and in both cases I shall take 

 up first decreased, then increased pressure. Each part of my 

 research will furnish the text of individual discussions; but it is 

 clear that general conclusions can be drawn only from their 

 simultaneous study, since the different data will so complete each 

 other and will be so intermeshed, so to speak, as to lead to a general 

 result which even now I can state in this rather paradoxical form: 

 pressure in its greatest variations, for example, from 10 centimeters 

 of mercury to 20 atmospheres, when these variations are made 

 with sufficient slowness, acts on living beings not as a direct 

 physical agent, but as a chemical agent changing the proportions 

 of oxygen contained in the blood, and causing either asphyxia, 

 when there is not enough of it, or toxic symptoms when there is 

 too much. It is upon the demonstration of this truth that all the 

 experimental data, the details of which I shall now give, converge. 



