300 Drs. D. A. Welsh and H. G. Chapman. On the [Apr. 30, 



insufficient to yield any visible precipitate when acted on by trichloracetic 

 acid. Salicyl sulphonic acid, hydroferrocyanic acid, picric acid, mercuric 

 chloride, and heat also failed to demonstrate the presence of 0-000005 gramme 

 dried serum in - 5 c.c. saline. Similar observations have already been 

 recorded, notably by Linossier and Lemoine, as indicating the delicacy of the 

 precipitin test for minute traces of homologous proteid, but we are not aware 

 that the inference that the antiserum must be the main source of the substance 

 that goes to form the precipitate has previously been made. 



It is also noteworthy that the maximum precipitate is given by an amount 

 of homologous proteid which, as will next be shown, is adequate to neutralise 

 all (or practically all) the precipitin, and that no further increase of the 

 homologous proteid will augment that precipitate. In view of the phenomena 

 that we have yet to describe, we should therefore regard a precipitate of about 

 1*5 mm. as representing the precipitable content of 0'01 gramme of this 

 particular antiserum. And, roughly speaking, since a precipitate of l - 5 mm. 

 is given by trichloracetic acid with 0-00005 serum, we may also say that the 

 weight of "precipitable " substance in - 01 gramme of this antiserum may be 

 represented by a figure of the same order of magnitude as 0-00005 gramme, 

 in other words, that one part of this antiserum contains 0'OOx. part of 

 "precipitable" substance. 



If after 48 hours one-fifth part of the clear superfluid from each precipitin 

 reaction be transferred to each of two clean tubes, A and B, and if one series, 

 A, be tested for residual precipitin by the addition of 0-0001 gramme dried 

 horse serum in - 5 c.c. saline, while the other series, B, is searched for 

 residual horse serum by addition of 0'01 gramme dried antiserum in - 5 c.c. 

 saline, the following secondary interactions occur, the precipitates being read 

 in 48 hours, as shown in Table II : — 



Table II. — Secondary Interactions. Experiment I. 



No. 



Precipitates in A series 

 ( + horse serum). 



Precipitates in B series 

 + antiserum) . 





mm. 



mm. 



1 



Not tested 



Not tested. 



2 



None 



2-5 



3 



None 



1 -5 



4 



0-3 



0-5 



5 



0-5 



0-3 



6 



0-5 



0-3 



7 



0-5 



3 



8 



-5 



None 



These results show that the so-called precipitin was completely neutralised 

 in Tubes Nos. 2 and 3, partially neutralised in Tube No. 4, and not 



