1906.] Main Source of " Precipitable " Substance, etc. 301 



noticeably diminished in Tubes Nos. 5, 6, and 7, when the precipita are 

 compared with that given in the control Tube No. 8. The homologous 

 proteid, on the contrary, was not exhausted in any tube, and in spite of the 

 facts that only one-fifth of the amount originally present was taken, and 

 that it had already reacted with antiserum to yield definite precipita, it was 

 still capable of eliciting from a second supply of 0"01 gramme antiserum 

 precipitates not less than those obtained in the primary interactions. An 

 apparent exception is Tube No. 4, in which occurred a deposit much smaller 

 than that primarily given. But the smaller precipitate in this instance 

 reinforces our argument, since it is in all probability attributable not to 

 exhaustion of the homologous proteid, but to the fact that the amount of 

 horse serum (0-00005 gramme) originally present was just insufficient to 

 neutralise - 01 gramme antiserum, and that the amount actually present in 

 the secondary reaction was 0*00001 gramme (already once acted on), making 

 the interacting quantities more nearly comparable to those in the primary 

 Tube No. 5. 



In Tubes Nos. 4, 5, 6, and 7 uncombined precipitin coexisted with uncom- 

 bined homologous proteid in the clear superfluids, an observation repeatedly 

 made by Linossier and Lemoine, Eisenberg, Ascoli, and others. 



Assuming for the moment that our two main conclusions were correct, 

 (1) that the precipitum is derived mainly from the antiserum, and (2) that 

 the homologous proteid is relatively not exhausted in any interaction, we 

 were naturally led to investigate the effect of adding to minute quantities 

 of the homologous proteid amounts of precipitin antiserum larger than those 

 used in Experiment I. The results were confirmatory of our hypotheses. 



Experiment II. — Taking an amount of dried horse serum (0'000005 gramme), 

 sufficiently minute to escape detection by trichloracetic acid and other 

 proteid precipitants (cf. Tube No. 5, Experiment I), we allowed it to interact 

 with increasing amounts (0'01 gramme, 0"05 gramme, 04 gramme) of dried 

 horse antiserum, the salt solution in each tube being made up to 14 c.c. 

 The results are summarised in Table III. 



Table III.— Experiment II. 



No. 



Amount of dried 

 horse serum in 

 grammes. 



Amount of dried 

 antiserum in 

 grammes. 



Precipita in tubes 

 to which antiserum 

 was added. 



Precipita obtainable 

 by ordinary proteid 

 precipitants. 









mm. 





1 



-000005 



o-oi 



0-3 



Inappreciable 



2 



-000005 



0-05 



1 





3 



-000005 



o-i 



2 



33 



