The uneven distribution of multiple sclerosis (MS) across populations can be attributed to differences in genes and the environment and their interaction. Prevalence and incidence surveys could be affected by inaccuracy of diagnosis and ascertainment, and prevalence also depends on survival. These sources of error might play a part in the geographical and temporal variations. Our literature search and meta-regression analyses indicated an almost universal increase in prevalence and incidence of MS over time; they challenge the well accepted theory of a latitudinal gradient of incidence of MS in Europe and North America, while this gradient is still apparent for Australia and New Zealand; and suggest a general, although not ubiquitous, increase in incidence of MS in females. The latter observation should prompt epidemiological studies to focus on changes in lifestyle in females. New insights into gene-environment and gene-gene interactions complicate interpretations of demographic epidemiology and have made obsolete the idea of simple causative associations between genes or the environment and MS.