Studies in Canada have provided strong evidence that environmental factors act at a population level to influence the unusual geographical distribution of multiple sclerosis (MS). However, the available data accommodate more than one type of environmental effect. Migration studies show that changes to early environment can greatly affect risk, and there are recent indications that risk can be altered in situ. The rising incidence rates of MS in Canada implied by longitudinal increases in sex ratio place this effect in temporal context and narrow the candidates for mediating the effect of environment. Similarly, geographical patterns in Australia imply that modifiable environmental factors hold the key to preventing some 80% of cases. Genetic epidemiology provides overwhelming evidence that genetic background has an important complementary role. If genetic factors are held constant, the environment sets the disease threshold. Although these could be independent additive risk factors, it seems more likely that susceptibility is mediated by direct interactions between the environment and genes.