Assessments of genetically determined variations in the T-cell antigen receptor in multiple sclerosis (MS) have yielded conflicting results. We used three restriction fragment length polymorphisms (RFLPs) and a polymorphic microsatellite repeat as markers for the T-cell receptor (TCR) beta locus (7q32-35) in multiplex MS families. Affected sibling-pair analysis of the RFLP data failed to show evidence for linkage (127 families) whereas analysis of the microsatellite data (86 families) provided weak evidence for linkage with a maximum lod score of 0.98 (p < 0.05). We repeated the analysis in those families (n = 53) in which the affected sibling pairs were concordant for the HLA haplotype DR15/DQ6. This altered the proportion of affected siblings sharing 0, 1, 2 RFLP haplotypes from 0.24, 0.50, and 0.26 (p = NS) before stratification to 0.16, 0.41, and 0.43 (p < 0.05) in the DR15/DQ6 positive pairs alone; for the microsatellite data, sharing altered from 0.16, 0.50, and 0.34 (p < 0.05) in all pairs to 0.07, 0.49, and 0.44 (p < 0.01) in the DR15/DQ6 concordant siblings.