The sensitivities of immunoassays relying on conventional radioisotopic labels (i.e. radioimmunoassay (RIA) and immunoradiometric assay (IRMA)) permit the measurement of analyte concentrations above ca 10⁷ molecules/ml. This limitation primarily derives, in the case of ‘competitive’ or ‘limited reagent’ assays, from the manipulation errors arising in the system combined with the physicochemical characteristics of the particular antibody used; however, in the case of ‘non‐competitive’ systems, the specific activity of the label may play a more important constraining role. It is theoretically demonstrable that the development of assay techniques yielding detection limits significantly lower than 10⁷ molecules/ml depends on:

• 1
  the adoption of ‘non‐competitive’ assays designs;
• 2
  the use of labels of higher specific activity than radioisotopes;
• 3
  highly efficient discrimination between the products of the immunological reactions involved.

Chemiluminescent and fluorescent substances are capable of yielding higher specific activities than commonly used radioisotopes when used as direct reagent labels in this context, and both thus provide a basis for the development of ‘ultra‐sensitive’, non‐competitive, immunoassay methodologies. Enzymes catalysing chemiluminescent reactions or yielding fluorescent reaction products can likewise be used as labels yielding high effective specific activities and hence enhanced assay sensitivities.

A particular advantage of fluorescent labels (albeit one not necessarily confined to them) lies in the possibility they offer of revealing immunological reactions localized in ‘microspots’ distributed on an inert solid support. This opens the way to the development of an entirely new generatio of ‘ambient analyte’ microspot immunoassays perrnitting the simultaneous measurement of tens or even hundreds of different analytes in the same small sample, using (for example) laser scanning techniques. Early experience suggests that microspot assays with sensitivities surpassing that of isotopically based methodologies can readily be developed.