738 Experiments 



as shown in Column 7 of the table. We find rather large differences 

 there. Whereas, for example, in Experiment CCLXXXI, in which 

 the oxygen proportion rose from 14.9 to 32.5, although seized by 

 violent convulsions, the animal survived, the dog in Experiment 

 CCLXXXVII died in 20 minutes, without having in its blood more 

 than 30.1% of oxygen, the initial proportion being 17.2. All the 

 results show that it would be impossible to fix exactly either the 

 absolute quantity of oxygen with which convulsions and death 

 occur or its proportional increase. Yet whenever the animal died, 

 the quantity of oxygen always exceeded 30 volumes per 100 vol- 

 umes of blood. 



The average increase is, we see, very slight, since it oscillates 

 between a third or a half above what exists normally. 



If, in order to examine their general course with more profit, 

 we express by graphs, in our usual manner, the results contained 

 in Column 7, we get Figure 60. The many variations which we 

 noted are shown here very clearly. 



But if we slide all these lines up vertically, making the origin 

 of each the number 20, and if we take the average of the different 

 points corresponding to about the same pressure, we get definitely 

 a line of remarkable regularity, that is, a straight line. 



So, in the living animal, we find confirmed the experiments in 

 vitro included in Subchapter V of Chapter II: from one atmosphere 

 on, there is added to the biood only dissolved oxygen. 



It is a fact worth noting that convulsions may appear when the 

 blood has an oxygen content appearing sometimes in healthy ani- 

 mals, which they may almost reach after rapid respiration. We see 

 first then that it is not the proportion oi oxygen contained in the 

 blood which is of itself dangerous; we see next that the increase 

 of this proportion, even to a high degree, does not constitute the 

 danger. This increase must be permanent, must be the result, not 

 of a better saturation of the corpuscles as an effect of more com- 

 plete aeration, a saturation which the reducing action of the 

 tissues soon restores to the normal degree, but of a saturation due 

 to the fact that the tissues themselves are saturated with oxygen 

 and in equilibrium with the blood. 



That is why the convulsions occur only after the compression 

 has lasted some time. The tissues must be impregnated with 

 oxygen in addition to what the blood, loaded with it in the lungs, 

 brings them and incessantly gives over to them. 



At the beginning of these experiments I asked myself whether 

 the blood was not directly altered by the excess of oxygen, and did 



