61 coded sera from patients with primary biliary cirrhosis (24) and various other hepatic and nonhepatic diseases (37) were tested for hepatitis B antigen (H.B. Ag.) by electron microscopy and radioimmunoassay. Investigators were not aware of the diagnosis of any patient. Particles, morphologically indistinguishable from H.B. Ag., were observed in 40 sera, including 23 which were negative for H.B. Ag. by radioimmunoassay. Assuming that radioimmunoassay is both a sensitive and reliable method for detecting H.B. Ag., there were 23 false positives and 17 true positives by electron microscopy. Of 24 patients with primary biliary cirrhosis, 5 were positive for H.B. Ag. by both radioimmunoassay and electron microscopy. There were 8 false positives by electron microscopy. It was concluded that there are particles in sedimented serum that are morphologically similar but unrelated to H.B. Ag. Therefore electron microscopy alone is not a reliable way to detect H.B. Ag. in serum. There was no evidence to suggest that H.B. Ag. is a cause of primary biliary cirrhosis.