The cAMP response to TSH stimulation in dog thyroid cells in monolayer culture was adapted as a means to assay TSH bioactivity. Of a variety of polypeptide hormones examined, only TSH and LH stimulated thyroid cell cAMP generation, but stimulation by LH probably represented contamination with TSH. Serum was found to be a potent, noncompetitive inhibitor of the thyroid cell cAMP response to TSH stimulation. The inhibitor(s) present in serum was nondialyzable and, on Sephadex G-200 gel filtration, eluted over a wide range between the void volume and the albumin peak. Because of this inhibitory effect of serum, partial purification of TSH from serum was necessary, and was achieved using Sephacryl S-200 gel filtration. Columns were calibrated with human TSH as measured by radioimmunoassay and bovine TSH as measured by bioassay. Intraassay variation of the thyroid cell cAMP response to a 100 microU/ml TSH standard was 9.2%. TSH standards were added to human serum following which the TSH was extracted by gel filtration and assayed for bioactivity. With TSH standards of 100 microU/ml intra-assay variation was 19.9%; and with TSH standards of 25 microU/ml, interassay variation was 34.5%. TSH bioactivity was demonstrable in serum from 21 of 25 patients with primary hypothyroidism, 4 of 14 normal subjects and none of 3 patients on exogenous thyroxine replacement. Although a positive correlation was observed between serum TSH bioactivity and immunoactivity, numerous individual samples displayed a dichotomy between the two measurements. This study provides new evidence that immunoassayable TSH in the serum of patients with primary hypothyroidism is not necessarily synonymous with TSH bioactivity.