MRLfl mice spontaneously develop a hindlimb arthropathy, as well as a number of immunologic abnormalities, including circulating rheumatoid factors. Although previous studies have suggested that this arthropathy is primarily an inflammatory process, we performed a comprehensive histomorphologic study which indicated that inflammation is a late manifestation of MRL/I arthritis. The pathologic changes that occur in the joints of these mice can be divided into 3 stages. The first stage develops between the ages of 7 and 13 weeks and consists of synovial cell proliferation in the joint ,recesses. The second stage is characterized by continued proliferation of synovial cells which take on an appearance similar to that of transformed mesenchyma1 cells. The earliest destructive changes occur in the second stage and include marginal erosions, followed soon after by progressive destruction of articular and meniscal cartilage. The final stage is characterized by a diminution of synovial cell proliferation, extensive cartilage destruction, formation of scar tissue and fibrocartilage, and a very moderate infiltration of the synovial stroma by mononuclear and polymorphonuclear inflammatory cells. Throughout the disease progression there is a striking dissociation between inflammatory cell infiltration or exudation and tissue destruction. The histomorphologic similarities between human rheumatoid synovitis and the arthritis of MRL/I mice, as well as the presence of rheumatoid factors, make this mouse strain an excellent model for studying human rheumatoid arthritis.Historically, the joint destruction which occurs in rheumatoid arthritis (RA) has been attributed to complex cellular interactions among acute and chronic inflammatory cells, proliferating synovial lining cells, chondrocytes, macrophages and other monocytic modulatory cells, and to the production of a variety of humoral factors such as proteolytic enzymes and soluble mediators (1-3). However, the elucidation of the precise mechanisms of destruction has been ham-' pered by the lack of an appropriate spontaneous animal model. Until recently, the only available animal models have been based on the induction of inflammatory synovitis by a variety of exogenous agents (4-7). However, the development of the MRLA mouse strain by Murphy and Roths (8) has addressed this problem. These mice, already useful as a model for SLE (9), spontaneously develop inflammatory arthritis o$ the hindlimbs, possess serum IgG and IgM rheumhtoid factors (lO,ll), and develop antibodies to types'I and I1 collagen (12), thus exhibiting the combined morphologic and immunologic similarities to the hu,man disease that experimentally-induced arthritides lack