Summary: By using the method of clonal analysis the evidence to prove that Hemophagocytic syn drome (HPS) is reactive or malignant was investigated to probe into the pathogenesis of HPS and its relations with clinical prognosis. The macrophages abnormally proliferated in bone marrow were iso lated. Electrophoresis analysis was made after DNA extraction, enzyme restriction of human ardro gen receptor (HUMARA) genetic locus, and PCR amplification. In the 9 specimens, clonal prolifer ation was found in 2 cases and nonclonal proliferation in 7. Among the 7 cases of nonclonal prolifera tion, 3 were voluntarily discharged without clinical outcome, 2 cases fully recovered after 2-3 week treatment of large dose gamma globulin intravenous drip and hormone therapy, 1 case died at the 43th day after the hormone and anti-infection therapy, and one case was found to have granular leukoblast in peripheral blood after 3 weeks and diagnosed as having M2a after bone puncture. For the two patients with clonal proliferation, one obtained remission after chemotherapy and the other was died after 32 days without chemotherapy. It was concluded that there do exist clonal or malig nant proliferation in HPS, so not every case is reactive.