There were 425 consecutive patients treated for Hodgkin's disease at this Medical Center from 1943 to 1983. Of these, 255 patients underwent a staging laparotomy and had complete preoperative clinical records. Overall, 35% had a change in stage (24% were upstaged, 11% downstaged). Twenty-nine per cent of clinical stage I patients were upstaged; 31% of stage II patients were upstaged, while less than 1% were downstaged; and four per cent of stage III patients were upstaged while 44% were downstaged. The diagnostic laparotomy yielded involvement in the spleen in 71% of patients with abdominal involvement, in the periaortic lymph nodes in 41%, in the liver in 11%, and the bone marrow in seven per cent. Only 12% of the 135 patients with negative laparotomies subsequently relapsed in the abdomen after a mean follow-up of 4.8 years. A multifactorial analysis was performed to identify dominant factors predicting the risk for abdominal disease. The factors best predicting abdominal involvement in stage I and II patients were: antecedent symptoms (greater than or equal to 2, 1, 0; p less than 0.00001), histological type [nodular sclerosing (NS) less than lymphocyte-predominant (LP) less than mixed cellularity (MC) less than lymphocyte-depleted (LD); p = 0.0009], and sex (females less than males, p = 0.01). The clinical stage (I vs. II), the site of lymphoma presentation, and the age and race of the patient did not have significant predictive value for the risk of abdominal disease after the other factors were accounted for. A mathematical model was derived for identifying dominant prognostic factors for predicting the risk of abdominal disease in an individual patient setting. The lowest risk patients were asymptomatic females with NS histology (6%) or LP histology (8%), while the highest risk patients were men with multiple symptoms and either MC histology (85%) or LD histology (93%). This information can be useful in making clinical decisions in Hodgkin's lymphoma patients, especially those at an increased risk for surgery.