Genetically isolated populations, such as Newfoundland, have contributed greatly to the identification of disease-causing genes. A linkage disequilibrium (LD) study involving six Newfoundland families predicted a critical interval for Bardet-Biedl syndrome 1 (BBS1) (Young et al. in Am J Hum Genet 65:1680-1687, 1999), but the subsequent identification of BBS1 revealed that it lies outside this region. This suggested that either there is another gene responsible for BBS in these families or the Newfoundland population may not be ideal for LD studies. We screened these six Newfoundland families for mutations in BBS1 and found that affected individuals in five of them were homozygous for the same M390R mutation. There was no evidence for any BBS1 mutation in the affected individual in the sixth family. Therefore, one of the criteria for LD mapping was not met; namely, there should be a single disease-causing allele in the population. Haplotype analysis of unaffected individuals from south-west Newfoundland and English BBS1 patients homozygous for M390R, revealed that a second criterion for LD mapping was violated. The M390R mutation occurred in a common haplotype and both of these chromosomes, the ancestral wild-type and disease-causing haplotypes, were introduced to Newfoundland and spread by a founder effect. Moreover, it was found that disease-associated alleles occurred at relatively high frequencies in normal haplotypes and this probably accounted for the incorrect prediction in the previous LD study. Knowing the amount of genetic variation and its distribution in the Newfoundland population would be useful to maximize its potential for mapping hereditary disorders.