An epidemiological survey of multiple sclerosis (MS) in three Australian cities, Perth, Newcastle and Hobart, was undertaken with its prevalence day being the national census day on June 30, 1981, exactly twenty years after a previous survey of the same cities. The relationship between increasing prevalence and increasing south latitude found in the 1961 survey was confirmed in this present study. Prevalence rates had increased significantly over the twenty years between the studies. Over the same time period incidence rates had also increased in Newcastle and Hobart but had remained essentially stable in Perth although these changes were not significant. The rise in prevalence was due to a combination of factors of differing importance in each city. These factors included better case ascertainment, increased recognition of the less severely disabled patient, increased survival time and differential immigration of a population at a higher risk of developing MS than the indigenous population. Finally, analysis of MS prevalence rates amongst migrant populations in Perth and Hobart suggested that either the risk of acquisition of MS may extend over a wider age range than is generally accepted or that environmental factors prevalent in the former city have modified disease expression there.