Two recent lines of evidence raise the possibility that instability in germ-line or somatic cells arises by a common mechanism that involves defective mismatch repair. Mutations in mismatch-repair proteins are known to cause instability in hereditary nonpolyposis colorectal cancer, instability that is physically similar to germ-line instability observed in Huntington disease (HD). Furthermore, both germ-line and somatic-cell instability are likely to be mitotic defects, the former occurring early in embryogenesis. To test the hypothesis that defective repair is a common prerequisite for instability, we have utilized two disease groups that represent different instability "conditions." Germ-line instability within simple tandem repeats (STR) at 10 loci in 29 HD families were compared with somatic instability at the same loci in 26 colon cancer (CC) patients with identified or suspected defects in mismatch-repair enzymes. HD is known to be caused by expansion within the CAG repeat of the locus, but the extent or pattern of STR instability outside this region has not been examined systematically. We find a distinctly different pattern of STR mutation in the two disease groups, suggesting different mechanisms. Instability in HD is generally confined to a single locus, whereas instability is widespread for the same loci in CC. Our data do not support a causative role for defective mismatch-repair enzymes in instability associated with HD; rather, our data are consistent with a model in which DNA structure may inhibit normal mismatch repair at the expansion site.