Multiple sclerosis (MS) epidemiology suggests that different factors are involved in the clinical expression of the disease. Alimentary cofactors have already been considered, but mainly theoretically. We have studied the relationship between MS prevalence and dairy product consumption in 27 countries and 29 populations all over the world, with Spearman's correlation test. A good correlation between liquid cow milk and MS prevalence (rho = 0.836) was found; this correlation was highly significant (p < 0.001). A low but still significant correlation was obtained with cream or butter consumption (rho = 0.619 and rho = 0.504, respectively). No correlation was found for cheese. These results suggest that liquid cow milk could contain factor(s) - no longer present in the processed milk - influencing the clinical appearance of MS. The possible role of some dairy by-products is discussed in the light of a multifactorial etiology of MS.