Os odontoideum was first described in the late 1880s and still remains a mystery in many respects. The genesis of os odontoideum is thought to be prior bone injury to the odontoid, but a developmental cause probably also exists. The spectrum of presentation is striking and ranges from patients who are asymptomatic or have only neck pain to those with acute quadriplegia, chronic myelopathy, or even sudden death. By definition, the presence of an os odontoideum renders the C1-2 region unstable, even under physiological loads in some patients. The consequences of this instability are exemplified by numerous cases in the literature in which a patient with os odontoideum has suffered a spinal cord injury after minor trauma. Although there is little debate that patients with os odontoideum and clinical or radiographic evidence of neurological injury or spinal cord compression should undergo surgery, the dispute continues regarding the care of asymptomatic patients whose os odontoideum is discovered incidentally. The authors' clinical experience leads them to believe that certain subgroups of asymptomatic patients should be strongly considered for surgery. These subgroups include those who are young, have anatomy favorable for surgical intervention, and show evidence of instability on flexion-extension cervical spine x-rays. This recommendation is bolstered by the fact that surgical fusion of the C1-2 region has evolved greatly and can now be done with considerable safety and success. When atlantoaxial instrumentation is used, fusion rates for os odontoideum should approach 100%.