Immunological methods have been used for analysis ever since the demonstration of ‘optimal proportions’ in the early 1920s, at a time when playing with such things was still a ‘hobby for gentlemen’. One may say nevertheless that quantitative immunology dated from this discovery, which led to a crucially important clinical advance, namely the definition of ‘units’ of toxins and antitoxins. This made possible rational immunotherapy and immunoprophylaxis in diphtheria and other diseases. Based on this understanding, in vivo tests using groups of animals allowed estimations down to the microgram level of either reagent; and the tests were good in that the activities measured (toxicity of toxin, protective power of antisera) were those relevant to clinical use. Developments in the 1930s gave a clearer picture of the nature of determinant groups of epitopes. The immunogenicity of chemically combined haptens was described and determinant groups were shown to have molecular sizes of around 200–1000 daltons. After the development of isotopic labelling the complete set-up was available for very precise immunoassays of high, medium or low molecular weight substances by competition methods. The very great advances of the 1950s in knowledge of antibody structure and cellular immunology were only marginally relevant to immunoassays. In view of the passage of 30 years it is disappointing that the assay of small molecules, such as drugs and neurotransmitters, has not developed more rapidly. However methods for estimating peptide and other hormones of intermediate molecular weight have proceeded in step with their discovery.