by experiment. It has long been known that the precipil In 

 reaction is one of extreme delicacy, that recognisable pre- 

 cipitates may be obtained with minute traces of the antigen 

 (homologous protein) in solution. We have been able to go 

 a step further, and to show that recognisable precipitates 

 can be obtained from a mixture of antiserum and antigen, 

 although the total amount of protein in the antigen solution 

 is too small to be revealed by any chemical precipitant of 

 protein (11, 12). Obviously the visible precipitate can not 

 have come from the invisible amount of antigen. It must 

 therefore have been derived from the antiserum. 



At a later stage we obtained corresponding results with 

 quantities of antigen and of precipitate that were large 

 enough to be measured by weight (13). Thus 1 mg. hen 

 egg-white (antigen) interacting with 52 cc. of antisera for 

 hen egg-white yielded a precipitate weighing 25'9 mg., and 

 the superfluid still contained antigen in solution. Obviously 

 again a precipitate weighing more than 25 mg. can not 

 have been derived from an amount of antigen weighing 

 only 1 mg. and still recognisable in the superfluid. The 

 antiserum must therefore be the main source of the "pre- 

 cipitable substance." 1 



Further study of the reaction by weighing the precipi- 

 tates (14, 15) has shown that the weight of the precipitate 

 is directly proportional to the amount of antiserum, and 

 independent of the quantity of antigen, provided that the 

 antigen is in sufficient quantity to yield the full precipitate 

 from the amount of antiserum employed. With this proviso, 

 if, in a series of tubes, 1 cc, 2 cc, 3 cc, etc. antiserum 

 interact with 100 mg. antigen in a volume of 50 cc. saline 

 solution, the dried precipitates will show the ratio in their 

 weights of 1, 2, 3, etc; and the superfluids removed from 

 1 Moll (32) and Eodet (33) h 



