Evaluation of presumed "multiple sclerosis families" and comparison with recently reported families has led us to the following observations: (1) Seven of our original fourteen presumptive multiple sclerosis families had to be eliminated after personal clinical evaluation of family members failed to confirm the diagnosis in a second close relative. (2) No segregation of HLA type was noted between affected and unaffected individuals in our seven bona fide multiple sclerosis families, and no consistent segregation was noted in the twenty-eight families reported elsewhere. This supports other genetic evidence that there is not a single, major gene mapping in the HLA complex which predisposes to multiple sclerosis. (3) The DW2 antigen was increased in frequency among affected members of our families, and the A3 B7 haplotype was more frequent among affected members of other families reported. But unaffected members also tended to have an increased frequency of these same antigens. (4) No relationship was noted between HLA type and antimeasles antibody titer within our families.