Practical application of the idiotype-anti-idiotype reaction to hapten immunoassays has been demonstrated with cotinine as an example. The assay relies on the ability of cotinine, a major nicotine metabolite, to inhibit binding between a monoclonal anti-cotinine antibody (the idiotype) and a second monoclonal antibody (the anti-idiotype) specific for the antigen combining region on the idiotype. A solid phase enzyme-linked immunoadsorbent assay (ELISA) format was adopted in which fluid phase anti-cotinine and cotinine present either as a standard or in a test sample were incubated in microtiter plate wells coated with F(ab')2 fragments of the anti-idiotype. Horseradish peroxidase-labeled protein A and o-phenylenediamine were used to detect idiotype-anti-idiotype binding. Under optimal assay conditions, 0.9 ng cotinine inhibited immune binding by 50% and as little as 0.04 ng could be detected. In contrast, nearly 70 times more trans-3'-hydroxycotinine, a major urinary metabolite, and over 1000-fold more nicotine were required for 50% inhibition. Several other metabolites and structurally related compounds also were poor competitors. Assay reliability was good over a range of cotinine concentrations from 5 to 500 ng/ml saliva with intraassay coefficients of variation between 6 and 10% and interassay values between 6 and 13%. Also, there was a strong correlation (R2 = 0.994) between the cotinine levels found in saliva from 35 cigarette smokers with the idiotype-anti-idiotype assay and a cotinine-anti-cotinine ELISA. Because only monoclonal antibodies and antigen are required, the idiotype-anti-idiotype immunoassay offers a high degree of standardization without the need to prepare labeled hapten derivatives or macromolecular conjugates for solid phase assays.