Antigenic Specificity 215 



which precipitate appears. Thus, in a heterophile reaction, two antigens 

 diffusing towards each other through a field of antibody will form two lines 

 of precipitate that become confluent on meeting. Although free antigen mig- 

 rates past the diffusion front of the other, no precipitate will form because 

 all the antibody is already combined in what is a region of antigen excess. 

 In the case of cross reactions a different situation exists. As Landsteiner (7) 

 and many others have demonstrated, where the antigenic determinants differ 

 from one another, the heterologous antigen will precipitate only part of the 

 antibody, leaving enough antibody in solution to form precipitate when homo- 

 logous antigen is added. Thus, the peptide glycyl glycyl glycyl glycyl glycine 

 (heterologous antigen) fonns a precipitate with antiserum to glycyl glycyl 

 leucvl glycyl glycine (homologous antigen). If this precipitate is removed and 

 the antiserum is then mixed with the homologous antigen a precipitate will 

 appear. The antibody that reacts with the heterologous antigen is probably 

 'imperfect', i.e. its specificity is less than that of the uncombined or 'avid' 

 antibody. In our system one would expect, then, that the advancing homologous 

 antigen would find uncombined antibody behind the front of the heterologous 

 antigen and form a line of precipitate there. We hope that we will be able to 

 grade the intensity of this reaction through some aberration in the band of 

 precipitate and thus be able to classify the cross reactions as to relative strengths. 



V. RESULTS 



As of this writing, we have obtained some preliminary data which we present 

 here as an indication of the limits or power of our test system. 



We obtained antisera to whole body extracts of the following organisms: 



Grasshopper Melanoplus differentialis phylum Arthropoda 



Frog Rana pipiens phylum Chordata 



Bacterium Escherichia coli sub-division Bacteria 



Horse mussel Modiolus modiolus phylum MoUusca 



Sea pen Ptilosarcus quadrangularis phylum Coelenterata 



Giant star fish Pisaster giganteus phylum Echinodermata 



In addition we tested extracts for which we had no antisera. These were: 



Baker's yeast Saccharomyces cerevisiae sub-division Fungi 



Earthworm Lumbricoides terrestris phylum Annelida 



The results of testing each of the six antisera against each of the eight 

 antigen extracts are shown in Table I. We observed from four to ten homo- 

 logous reactions per test and two heterologous reactions. Since we have no 

 data on the number of homologous reactions for yeast and earthworm extracts 

 we assigned each of them a value of 7.3, which is the mean of the homologous 

 reactions for the other six tests. Then, we have six tests, in each of which m^ 

 was (on average) 7.3; w,, being the sum of antigens in all other complexes, 

 was 7 X 7.3 = 51 ; the average number of heterophile reactions was 2/6 = 0.33. 

 Then, if reciprocal tests are considered as independent, we obtain: 



- 7.3x51 ,,_ 

 ^ 033 ^^^ 



loggTV^ 10 bits 



