We have evaluated genetic and environmental influences in multiple sclerosis (MS) by comparing age of onset in 99 sibling pairs concordant for the disease. We used three methods of analysis: (1) comparison of mean differences in age of onset and year of onset, (2) linear regression of differences in age or year of onset vs difference in ages, and (3) intraclass correlation of age of onset which is also used for monozygotic twins concordant for MS. Comparison of the mean differences in age of onset or year of onset is found to be inappropriate and potentially misleading. No significant results were found in linear regression of the age of onset or year of onset vs differences in ages, although a trend towards onset at the same age is present. However, nontwin siblings show a significant intraclass correlation for age of onset (P less than 0.01) as is seen in genetic disorders. A stronger intraclass correlation in age of onset in concordant monozygotic twins vs concordant sibling pairs further suggests that age of onset is partly under genetic control, assuming common exposure to an environmental agent. The results give little support for common exposure to an environmental trigger in concordant MS sibling pairs. They are consistent with a mixture of random independent exposures and common exposures leading to the development of the disease, with the former predominating.