Segregation of HL-A haplotypes was analyzed in 10 families in which there were at least two cases of multiple sclerosis. In nine families, multiple sclerosis was associated with only one parental HL-A haplotype. Specific HL-A determinants associated with multiple sclerosis differed among the families, suggesting that another histocompatibility-linked factor, possibly a gene determining susceptibility (or lack of resistance) played an etiologic role. Lod score analysis based on nine families suggested a close association between such a gene (labeled MSS) and the HL-A gene complex. However, when all 10 available families were analyzed, the association approached but did not reach statistical significance. Thus, the HL-A haplotype segregation did not prove that a histocompatibility-linked gene is related to the cause of multiple sclerosis, but study of additional multiplex families is certainly warranted. Other factors, possibly genetic (although not HL-A-linked), environmental, or the two together, may be required for multiple sclerosis to become clinically apparent.