The overlapping reactions commonly observed in serological tests are ascribed by most authors to multiple antibodies formed as a result of the presence of several substances or distinct specific groups in the immunizing antigen. Yet studies on azoprotein immune sera have shown that antibodies corresponding to a particular compound regularly react with other substances which are sufficiently related in chemical structure (1, 2). Thus cross reactions in general can be understood on this principle alone without postulating a multiplicity of antibodies. However, this explanation does not cover the observations made when immune sera are partially exhausted by heterologous antigens.It is unnecessary to cite examples of the well known and widely used fact that immune sera for bacteria and blood cells after treatment with a heterologous antigen (sensitive to its action) still react with the homologous antigen but no longer with that used for exhaustion. This method does not succeed as readily and regularly with precipitin sera for proteins (3). Yet in a number of experiments such sera could be made highly specific by fractional precipitation with heterologous antigen which reveals the presence of several antibodies. For instance, Hooker and Boyd (4) found that a precipitin for chicken serum acting also on duck serum, after removal of the precipitate formed on addition of the latter, still precipitated chicken serum intensely, but no longer duck serum. The assumption made here and in similar instances that the sera contain several antibodies directed toward different determinant groups in the protein has not been proved by chemical evidence. In order to investigate this question experiments were carried out with precipitins for artificially conjugated antigens 325 on

In previous work 1 the conclusion was arrived at that serological specificity depends on the chemical constitution of the reacting groups in general and probably also on their spatial configuration. The latter assumption had been made on account of the reactions observed with antigens containing aromatic compounds substituted in various positions of the benzene nucleus. The experiments dealt with in the present paper were carried out in order to secure conclusive evidence for this conception. Thus we endeavored to determine whether optical isomers can be differentiated by means of serum reactions.The substances chosen for this purpose were l-and d-phenyl (para-aminobenzoylamino) acetic acid. These compounds had been used by Ingersoll and Adams * in an investigation on the affinity to fibers, of optically isomeric dyes. By diazotization and coupling to dimethylaniline they obtained two isomeric dyes which were absorbed by wool in different amounts; when B-naphtol was used in place of dimethylaniline the dyes were absorbed equally well.The preparation of the phenyl (para-am~nobenzoylamino) acetic acid was carried out in the main according to the directions of Ingersoll and Adams. Phenylaminocyanide made by the interaction of benzaldehyde, sodium cyanide, and ammonium chloride in water solution, was hydrolyzed to phenylaminoacetic acid3 The d-/-phenylaminoacetic acid was resolved into the levo and dextro product by means of fractional crystallization of the d-camphor sulfonate and /-camphor sulfonate respectively. In the recrystallization of the/-amino acid d-camphor x A review of the subject can be found in The chemical aspects of immunity, by

Whereas the usual serological reactions involve the use of highmolecular antigens of unknown constitution, a method has been described in a previous paper (1) which permits of the application of serological reactions to compounds of simple chemical composition. The method is based upon the possibility of partially synthesizing antigens (2, 3), by attaching to proteins chemical substances of simple structure as can be done by coupling with diazo compounds. Immune sera produced by injecting such "synthetic antigens" exhibit a specificity depending on the simple substances forming a part of the complex, particularly if the test antigen used contains a protein different from that used for immunization.In ordinary precipitin tests, it has long been noticed that the reactions can be inhibited by addition of an excess of the antigen. Applying this observation to artificial complex antigens, it has been shown that the precipifin reactions of these substances are often inhibited specifically by the addition of a sufficient quantity of the same or a similar chemical substance to that used in building up the antigenic complex (1). These observations have since been confirmed by Klopstock and Selter (4), Avery and Goebel (5), and by ourselves.The inhibition reactions with substances of known chemical structure can be used for the study of serological specificity in general, and in this regard they have certain advantages over the precipitin or complement fixation tests made with the full artificial antigens. Thus, in the first place it is not necessary to use substances which can be combined with proteins and, consequently, a greater variety of compounds can readily be subjected to the test. Furthermore an advantage may arise from the circumstance that the possible

The question of serological cross reactions has recently been commented upon by other authors and ourselves (1-5). To recapitulate briefly, an immune serum may exhibit cross reactions by virtue of an antibody able to combine with substances more or less closely related to the homologous antigen in chemical structure, or it may contain multiple antibodies, differing in specificity, some of which cross react with certain heterologous antigens. The appearance of several antibodies after immunization with a particular antigenic material may depend upon the presence in the latter of different antigenic molecules, or upon the existence, in a single molecule, of more than one determinant group; moreover, as has been shown in our studies on azoproteins (4), multiple antibodies varying somewhat in specificity may be produced in response to one determinant structure in cases where the antigen does not contain divers chemical groupings that in part are shared by the reacting heterologous antigens. To illustrate the latter, an immune serum produced by injecting an azoprotein made from m-aminobenzenesulfonic acid was found to contain several antibody fractions reacting differently towards antigens made from o-aminobenzenesulfonic, m-aminobenzoic and m-aminophenylarsenic acids. Similar results were found in other instances and may'well be expected to apply to natural antigens and their antibodies.One can assume therefore that, by and large, in using any immune serum one deals not with a single antibody but with a mixture of somewhat different immune bodies (cf.

No abstract