IN 1952, Brain et al1described a form of cervical myeloradiculopathy which they attributed to compression by vertebral osteophytes and designated cervical spondylosis. Mounting evidence now implicates etiologic factor(s) in addition to, or instead of, compression of neural structures because: (1) levels of dysfunction often do not correspond to the location of vertebral osteophytes2-5; (2) evidence of spinal cord compression by ventral osteophytes is often slight or absent at laminectomy2,3,6-8; and (3) myelographic findings do not always correlate with neurologic involvement.8-10 On the assumption that spondylotic myelopathy could be caused by circulatory impairment of the cervical spinal cord,1,3,5,78,11,12-20we attempted to reproduce this disorder in dogs by interrupting a series of extramedullary arteries. (The arteries supplying the canine cervical cord had been investigated earlier.21) The purposes of this report are to describe our experimental results and to relate them to observations of other investigators.