In mammals, thyrotropin, or thyroid-stimulating hormone (TSH), assay is used for the diagnosis of primary hypothyroidism. Hypothyroidism is the most common type of thyroid disorder in cattle. The aim of this study was to develop and validate, under physiologic and pathologic conditions, a radioimmunoassay (RIA) for bovine TSH (bTSH). Double RIA was performed with purified bTSH and specific bovine antiserum. Laboratory validation included research of minimal detection limit, accuracy, and reproducibility. The physiologic validation included a thyrotropin-releasing hormone (TRH) challenge performed on euthyroid cows and a follow-up of bTSH concentration over a 24-hour period. Furthermore, bTSH concentration was assayed in a large population of healthy dairy and beef cows to define reference interval. The pathologic validation was made by assaying bTSH and thyroid hormones on healthy and goitrous newborn calves. The minimum detection limit (MDL) for bTSH assay was 1.3 microU/ml. The recovery was 101% to 106%. The intra- and interassay coefficients of variation (CVs) ranged from 5% to 11% and 11% to 15%, respectively. The RIA covered the whole range of physiologic bTSH values, as shown by bTSH values induced by TRH-challenge. A pulsatile secretion of bTSH was observed, accompanied by a diurnal variation with lower night values than day values. Reference intervals of bTSH ranged from 1.3 to 13.0 microU/ml for beef and dairy breeds. Finally, bTSH easily discriminated goitrous newborn calves from healthy ones, leading to the definition of a cutoff value of 35 microU/ml. The bTSH assay positively reacted to physiologic and pathologic conditions. The accuracy and precision of the RIA were satisfying.