86 AI.AX ARTHUR BOYDEX. 



3. The Effect of Antigen Concentration on the Tiler of Antisera. 



Table II. shows the result of titrating the same antiserum 

 (P4i) with two homologous standard antigens of different 

 strengths. All tubes were unknown in this experiment. 



The results given in Table II. show that the titer of an anti- 

 serum varies directly with the concentration of the antigen used. 

 Indeed it has long been known in a general way that such was 

 the case. It has also been known that various antisera differ in 

 titer when tested with the same homologous antigen. There are 

 therefore two chief factors in this reaction, (i) the concentration 

 of antigen, (2) the concentration of antibody in the antiserum. 

 Now of these two the latter is but partly controllable. For we 

 may inject various amounts of protein into animals but we 

 must take whatever titer of antibody they yield. On the other 

 hand the concentration of antigen can be definitely controlled 

 and this should be made the standard of comparison for all 

 workers in this field. For this reason only standard antigen 

 solutions have been used in this investigation. 



The results of the nitrogen determinations on the standards 

 used are given in Table III. 



From the data given in Table III. it may be seen: (i) that the 

 blood sera used in general possessed about the same protein 

 concentration, but that the dog serum was enough lower than 

 the average to affect definitely the titer of its reactions, (2) that 

 the amount of non-protein nitrogen in the fresh blood sera used 

 was generally so small as to be well within the range of error of 

 the titration, (3) that the amount of non-protein nitrogen 

 generally increased in the standard solutions which were some- 

 kept for several months. The values for non-protein 

 nitrogen in these older standards sometimes approached a level 

 high enough to become of consequence in the comparative 

 titration of heterologous antigens. 



Probably in fresh sera the error due to calculation of protein 

 from total nitrogen would seldom exceed 10 per cent. Total 

 nitrogen determinations alone would then give figures sufficiently 



curate for comparative studies It is hoped that the users of 

 the precipitin reaction in the study of related proteins or in 

 other quant itive studies will feel the necessity of getting com- 



