1. Five cases of scoliosis with paraplegia are reported, and thirty-six comparable cases from the literature are reviewed. These forty-one cases have been studied with the object of determining the etiology of scoliosis, the reason why cord compression sometimes develops, and the results of conservative and operative treatment of such compression of the cord.

2. The cause of paraplegia is nearly always compression of the spinal cord by the dura, which, in severe scoliosis, is under longitudinal tension because of its firm attachment to the foramen magnum above and the sacrum below. Such tension, resisting displacement of the spinal cord from the straight line, may be shown to cause incomplete spinal block even when there is no paralysis.

3. When paralysis occurs it usually develops during the years of most rapid growth, the tight dura being unable to accommodate itself to the rate of growth of the spinal column; cord compression is probably increased by narrowing of the dural sac by rotational displacement.

4. The most striking results have been secured by laminectomy with section of the dura and sometimes division of dentate ligaments and tight nerve roots. After such division there is evidence of release of compression: the cord herniates through the dural slit; and spinal pulsation returns.

5. It is important to control bleeding in order to avoid post-operative compression by blood clot; and to prevent leakage of cerebro-spinal fluid through the arachnoid.

6. It is unwise to perform spinal fusion at the same time as decompression because it increases the danger of haematoma formation. Moreover the improvement gained by decompression is maintained even if no fusion of the spine is performed.

7. Conservative treatment of scoliosis with paraplegia should not be continued for long periods unless there is evidence of early and progressive improvement because prolonged compression causes irreversible changes in the cord.

8. In three cases, paraplegia was not due to dural compression: one turned out later to be a case of syringomyelia; one, reported by Heyman, was due to the pressure of a bone spur; and one, reported in this series, was due to a congenital tight band of developmental origin which might have caused the scoliosis as well as the paralysis, and in which, after resection of the band, recovery from the paralysis was complete.