METHODS 97 



bacteridia, we obtain the result of the following function 

 (Bacteridium x Bouillon) and not that of the function (Bac- 

 teridium x Sheep). The fact that this filtered liquid is never- 

 theless harmful to the sheep proves that the two functions 

 do not differ greatly from each other. We have already 

 seen that the functions (Bacteridium x Mouse), (Bacteridium 

 x Guinea-pig) and (Bacteridium x Sheep) are all of the same 

 order. The bacteridium has not specific actions well char- 

 acterized. 



It is quite different with another microbe which is also 

 well-known the rouget of the hog. The function (Rouget x 

 Babbit) antagonizes the function (Rouget x Pigeon), since 

 the rouget's habituation to killing the rabbit diminishes 

 in virulence with the pigeon. 



Whatever their importance in scientific medicine, such 

 facts of transportability of functional results in dead liquids 

 have far greater importance in biology. 



Suppose that to each function clearly defined in a living 

 species there should correspond a specific functional activity 

 transportable in a dead colloid. We should then be able to 

 preserve in labelled jars every possible function of a living 

 being in the form of colloids able to perform, at least in part, 

 each of these specific functions. That would be a real 

 analysis of the vital activity of the species studied. Such 

 an analysis might be more or less complete, but it would be 

 indeed real. 



One of the first results obtained along this line was the 

 fabrication of alcohol in a sugared must by means of dead 

 colloids furnished by yeasts. This occasioned the assertion 

 that such dead colloids are partially alive which, at least in 

 this form, seems absurd. What is true is that we can 

 decompose a greater or less part of the vital activity of a 

 living being, and perhaps all its activity, into partial activi- 



7 



