Human choriogonadotropin (hCG) and free hCG-beta values for 934 serum samples from patients with seminomatous or nonseminomatous testicular cancer were measured by highly specific immunoradiometric assays (IRMAS). In non-seminoma samples, hCG and hCG-beta were highly correlated (r = 0.82, P less than 0.001). Of 112 "marker-positive" seminoma samples, only 46 (41.1%) showed both increased hCG and hCG-beta. In 39 cases (34.8%) only hCG-beta and in 27 cases (24.1%) only dimer-hCG was increased. This makes the determination of hCG and hCG-beta, either by two assays or by a single hCG (+hCG-beta) assay, most reliable in these patients. For all samples, hCG (+hCG-beta) was measured by a polyclonal RIA and a monoclonal IRMA, which differed in their cross-reactivities with hCG-beta (234% and 720%, respectively). The hCG (+hCG-beta) IRMA, as a result of its higher hCG-beta cross-reactivity, was superior to the hCG (+hCG-beta) RIA in detecting slightly increased hCG-beta. Additionally, 11 widely used commercial hCG kits were tested for their hCG-beta cross-reactivities and showed values between less than 3% and 264%.