that the precipitate formed with small quantities of antigen 

 must he composed mainly of precipitin, and. from his 

 equation of equilibrium 



A - P x B - P _ R 



it follows that his large precipitates must contain propor- 

 tions of antigen and antiserum similar to those contained 

 in his small precipitates. Thus a small precipitate of 2 

 units (P = 2) contains 2 equivalents of A (antigen) and 2 

 equivalents of D (precipitin ; and a large precipitate of 64 

 units (P=64) contains 64 equivalents of A and 64 equival- 

 ents of B. If, as Arrhenius states, the small precipitate 

 is composed mainly of precipitin, then, on lus own h • 

 the large precipitate is in the same ratio composed of pre- 

 cipitin. Ai rhi'i ius <'2.'>, also points out that t he precipitates 

 deviating complement in the " deviation test of Moreschi" 

 must be derived mainly from precipitin — a fact which we 

 fully recognise (24). 



The experiments of Von Dungern (25), P. T. Miiller (26, 

 27), Fleischmann and Michaelis (28) have led them to the 

 view that a diminution of both components of the precipitate 

 (antigen and antibody) occurs in the interaction; and that, 

 in favourable circumstances, both components may be 

 completely removed from the solution. The precipitate is 

 therefore composed of an insoluble compound of both com- 

 ponents, neither of which acts as a ferment on the other. 

 If a considerable mass of precipitin reacts with a small 

 mass of "preeipitablesubstaiM'r, - ' the precipitate is formed 

 of mucli precipitin and iittle *■ pteeipit ahle substance." If 



Von Dungern (21 



with ascending quantities of antigen. 



