The literature for evidence of autoimmunity in multiple sclerosis (MS) is analysed critically. In contrast to the accepted theory, the human counterpart of the animal model experimental autoimmune demyelinating disease, experimental allergic encephalomyelitis (EAE), is not MS but a different demyelinating disorder, i.e. acute disseminated encephalomyelitis and acute haemorrhagic leucoencephalitis. Extrapolation of EAE research to MS has been guided largely by faith and a blind acceptance rather than sound, scientific rationale. No specific or sensitive immunological test exists that is diagnostic of MS despite the extensive application of modern technology. Immunosuppression has failed to have any consistent effect on prognosis or disease progression. The available data on MS immunotherapy are conflicting, at times contradictory and are based on findings in animals with EAE. They show predominantly a 30% effect in relapsing/remitting MS which suggests powerful placebo effect. Critical analysis of the epidemiological data shows no association with any specific autoimmune diseases, but does suggest that geographic factors and age at development posit an early onset possibly dependent on environmental influences. Certain neurological diseases are, however, found in association with MS, namely hypertrophic peripheral neuropathy, neurofibromatosis-1, cerebral glioma, glioblastoma multiforme and certain familial forms of narcolepsy. These share a common genetic influence possibly from genes on chromosome 17 affecting cell proliferation. A significant number of these disorders are of neural crest origin, the classical example being abnormalities of the Schwann cell. These and other data allow us to propose that MS is a developmental neural crest disorder, i.e. a cristopathy, implicating glial cell dysfunction with diffuse blood-brain barrier breakdown. The data on transcription factor SOX10 mutations in animals may explain these bizarre clinical associations with MS and the phenotypic variability of such alterations (Cossais et al. 2010). Research directed to the area of neural crest associations is likely to be rewarding.