44 DISCUSSION 



reveal isoantibodies which were not apparent in dextran-haemagglutin- 

 ation or haemolytic tests. But if, as Prof. Medawar's work clearly in- 

 dicates, the antigens show chemical heterogeneity, there is all the more 

 reason to beHeve that the sermn antibodies induced will also be hetero- 

 geneous in their specificities. I think we may give too much weight to 

 the red cell antibodies and their titres under these circumstances. 



