SYNOPSISThe commonly accepted mechanism of rheumatoid C1-C2 dislocation is the incompetence of the transverse ligament of C1 vertebra. The presence of rheumatoid granulation tissue between the anterior arch of C1 vertebra and the odontoid process is an important factor. This might explain why the dislocation becomes irreducible when the disease progresses. In a 55 year old man, transoral removal of the rheumatoid granulation tissue made the previously irreducible C1-C2 dislocation reducible. Stabilization of the reduced C1-C2 articulation was accomplished at a second operation by posterior fusion of the occiput to C1, C2, and C3 vertebrae.Although dislocation of the first or the second cervical vertebra ('C1-C2 dislocation') is a wellrecognized complication of rheumatoid arthritis (Conlon et al., 1966), the exact mechanism producing the dislocation is not yet fully understood.The most common type of C1-C2 dislocation in rheumatoid arthritis is the forward dislocation of C1 on C2 vertebra. The diagnosis is made when, in radiographs of the lateral cervical spine, the distance between the anterior arch of C1 vertebra and the odontoid process is greater than 2-5 mm in females and 3 mm in males (Conlon et al., 1966).During the early stages of the disease, the distance between the anterior arch of C1 vertebra and the odontoid process varies with head position, being greatest with the head flexed and minimal or normal with the head extended. and increased vascularity ... were lax and resulted in extreme mobility of the atlantoodontoid joints'. This was the cause of a fatal dislocation in a 61 year old white female with a 20 year history of rheumatoid arthritis. In a 53 year old female who died of rheumatoid C1-C2 dislocation, Webb et al. (1968) found the transverse ligament of C1 vertebra destroyed to such a degree that it could not be identified.In the presence of a lax or ruptured transverse ligament, C1 vertebral body slips forward on C2 vertebra in flexion, but the displacement of the odontoid process remains limited. The odontoid process does not reach the posterior arch of C1. Werne (1957) proposed that the alar ligaments limit the extent of backward dislocation of the odontoid process when the transverse ligament is destroyed. Were this hypothesis correct, the forward dislocation of C1 on C2 vertebra should be reducible with the head placed in extension. However, frequently the reduction remains incomplete and the gap between the anterior arch of C1 vertebra and the odontoid process with the head in extension remains excessive, although less than when the head is flexed. We postulate that this failure to reduce the dislocation might be explained by the presence of rheumatoid granulation tissue between the anterior arch of C1 vertebra and the odontoid process. In a 65