100 



IMMUNITY AND ANAPHYLAXIS 



to appreciate this because it is the smallest granules of 

 antigens, which, by traversing the membranes of cells, 

 provoke the condition of congestion x and dilatation of the 

 cellular membrane which then becomes permeable for more 

 voluminous granules. 



From the chemical point of view, we know that each mole- 

 cule of the chlorhydrate of dioxydiamino-arsenobenzene may 

 theoretically combine with two molecules of silver chloride 

 but when a dilute solution of silver chloride in potassium 

 cyanide is added drop by drop to a dilute solution (1 to 500) 

 of arsenobenzene, we find that each drop of silver chloride is 

 precipitated en masse without forming a globule and is then 

 redissolved. If the liquid is carefully shaken, the dissolu- 

 tion of each drop which follows becomes more difficult. 

 When we reach the proportion of almost one molecule of 

 silver chloride to one molecule of arsenobenzene, dissolution 

 no longer takes place. 



An insoluble precipitate is also obtained by adding quickly 

 one molecule of a concentrated solution of silver cyanide to 

 twenty molecules of arsenobenzene, although there is still 

 in the liquid enough of the latter to fix and hold in solution 

 nineteen other molecules of silver cyanide. The difference 

 in these two cases may be explained by assuming that in the 

 first case by adding slowly into dilute media and by shaking 

 the mixture the silver chloride is more or less uniformly 

 distributed among all the molecules of arsenobenzene while 

 in the latter case each molecule of arsenobenzene which 

 comes into contact with silver chloride fixes two molecules 

 of it and there is thus formed an insoluble compound. 



The granule of arsenobenzene may thus fix a salt for which 

 it has a certain affinity in quantity equal to the sums of 

 the affinities of the molecules which compose it, but it is 

 evident that it can also fix less of them and it is thus that 

 we can explain the "phenomenon of surcharge" and the 

 "phenomenon of least saturation" already described, in 

 studying the properties of mixtures with their antibodies 

 in vitro and in vivo. 



For the arsenobenzenes we know exactly the chemical 

 equivalents; for biologic antigens we do not know them but 



