Ninety-eight patients with clinically localised Hodgkin's disease underwent laparotomy and splenectomy to determine the extent of microscopic spread. In 68 patients the procedure was carried out for untreated disease apparently confined above the diaphragm. Abdominal disease cannot be confidently excluded on the basis of non-invasive investigation at presentation. Clinical assessment of splenic disease was unreliable unless gross splenomegaly was present. Pedal lymphography was accurate in assessing para-aortic and iliac disease but of no value in assessing other intra-abdominal lymph node involvement, including that of the mesenteric lymph node. Trephine bone marrow biopsy findings were normal in all patients before surgery, and only one patient was found to have diseased bone marrow by Stryker-saw biopsy at operation. Liver disease was identified at operation in nine patients, some of whom were asymptomatic with clinically undetectable splenic and nodal disease. Detailed clinical staging failed to detect disease in one-third of patients who underwent laparotomy. These studies show that if radiotherapy is to remain the treatment of choice for disease truly localised to lymph nodes a detailed staging procedure, including laparotomy and splenectomy, remains essential. The value of this potentially curative treatment is considerably diminished in the patient who has been inadequately staged.