ANALYSIS OF ANTIGEN-ANTIBODY COMPLEXES 205 



neutralization of antigen by antibody, a failure perhaps due to a ready dissociability of 

 these particular antigen-antibody complexes (Youraans and Colwell 1940, ColweU and 

 Youmans 1941). The nature of the antigen may also affect the equivalence point to a 

 marked degree. Thus, Youmans and Colwell (1943), working with rabbit and guinea-pig 

 antisera to whelk hsemocyanins, report that, in contrast to other antigens, the constant- 

 antibody optimal ratio was weU in the region of antibody excess, and bore no relation 

 to the antibody content of the sera. 



Such a wide range of equivalent combining proportions argues either hetero- 

 geneity of the reagents, or multiple combining proportions of homogeneous anti- 

 body particles with homogeneous antigen particles. In many of the systems 

 studied, it is probable that the antibody and antigen particles concerned vary 

 either in kind or in combining capacity, but in others the evidence for homogeneity 

 of the reagents is sufficient to permit deductions about the combining proportions 

 of antigen and antibody. 



Analysis of Antigen-Antibody Complexes. 



Antibody is a globulin (see pp. 242-246) and the chemical analysis of antigen- 

 antibody precipitates yields Uttle information about combining proportions where 

 the antigen is itself a protein of similar gross composition. Heidelberger and 

 Kendall (1929) took advantage of the nitrogen-free purified polysaccharide, obtained 

 from a Type III pneumococcus, to make a quantitative study of specific precipi- 

 tates. The polysaccharide precipitates with a Type III pneumococcus antiserum, 

 and micro-Kjeldahl estimations of the nitrogen in the precipitates, give a measure 

 of the antibody in combination. In the zone of antibody excess, and in the equiva- 

 lence zones where no antigen is detectable in the supernatant fluid, the precipitate 

 contains all the antigen. They found antigen-antibody ratios ranging from 1 : 125 

 with the lowest concentration of antigen used to 1 : 69 at the beginning of the 

 antigen excess zone, and attributed the varying proportions of antigen and anti- 

 body to the presence in the precipitates of mixtures of three different compounds, 

 each formed by union of antigen and antibody in constant proportions. That is, 

 symbolizing antibody and antigen by A and G respectively, they postulated a 

 series of compounds AGi, AGj, AG3, etc. 



Working with an azo-protein antigen (see pp. 253, 254) whose concentration in the 

 precipitate could be determined colorimetricaUy, Heidelberger, Sia and Kendall (1930) 

 found the antigen-antibody ratios in precipitates from mixtures in antibody excess, equiva- 

 lence zone and antigen excess were 1 : 14-1, 1 : 6-7 and 1 : 3. With haemoglobin as antigen, 

 Breinl and Haurowitz (1930) found the precipitate might contain from 6 per cent, to 

 24 per cent, of antigen according to the proportions present in the reacting mixture. 



With azo-proteins and iodo-proteins as antigens, Marrack and Smith (19316) found 

 that the proportions of antigens in the precipitate at the limit of antigen excess was about 

 1-5 times as great as at the equivalence point. It is clear that though antigen and antibody 

 unite to give optimal flocculation only when mixed in one proportion, yet they can unite 

 in a series of varying proportions. Marrack and Smith (1931a, b) and Marrack (1934, 

 1938) would represent the compounds formed by the symbols A„tG„, where A and G 

 represent antibody and antigen respectively, and m and 7i may take any integral values 

 within a certain range. 



Heidelberger and Kendall (1935a, 6, c) adopted the conception of union in 

 varying proportions from a quantitative analysis of other precipitin systems. 

 Their results are illustrated in Fig. 32, which shows the relation between the amount 

 of antigen added to a constant amount of antibody in a constant reacting volume, 



