This study of allelic association using three intra- and two extragenic markers within 150 kb of the Huntington disease (HD) mutation has provided evidence for linkage disequilibrium for four of five markers. Haplotype analysis of 67 HD families using markers in strong linkage disequilibrium with HD identified two haplotypes underlying 77.6% of HD chromosomes. Normal chromosomes with these two haplotypes had a mean number of CAG repeats significantly larger than and an altered distribution of CAG repeats compared with other normal chromosomes. Furthermore, haplotype analysis of five new mutation families reveals that HD has arisen on these same two chromosomal haplotypes. These findings suggest that HD arises more frequently on chromosomes with specific DNA haplotypes and higher CAG repeat lengths. We then studied CAG and CCG repeat lengths in the HD gene on 896 control chromosomes from different ancestries to determine whether the markedly reduced frequency of HD in Finland, Japan, China and African Blacks is associated with an altered frequency of DNA haplotypes and subsequently lower CAG lengths on control chromosomes compared to populations of Western European descent. The results show a highly significant inverse relationship between CAG and CCG repeat lengths. In populations with lowered prevalence rates of HD, CAG repeat lengths are smaller and the distribution of CCG alleles is markedly different from Western European populations. These findings suggest that, in addition to European emigration, new mutations make a contribution to geographical variation of prevalence rates and is consistent with a multistep model of HD developing from normal chromosomes with higher CAG repeat lengths.