The DNA of 147 patients of European origin clinically diagnosed with idiopathic hemochromatosis and 193 controls was examined for mutations of the HLA-H gene at nt 845 and nt 187. One hundred twenty-one (82.3%) of the hemochromatosis patients were homozygous and 10 (6.8%) heterozygous for the 845A (C282Y) mutation. All of the homozygous patients were also homozygous for nt 187C, and all 845A heterozygotes had at least one copy of 187C. Thus, the nt 845 and nt 187 mutations were in complete linkage disequilibrium; nt 187 was a C on all chromosomes with the 845A mutation. Eight of the 10 heterozygotes for 845A were heterozygous for 187G(H63D). The excess of heterozygotes at both nt 187 and nt 845 suggested either the presence of as yet undiscovered mutations existing in trans with 845A and in linkage disequilibrium with 187G, or that the 187G itself is a deleterious mutation, which in concert with the 845A can give rise to hemochromatosis. None of the 193 normal controls were homozygous for 845A and 29/193 (15%) were heterozygous for 845A. Although 47/193 (24.3%) of normal controls were heterozygous for the 187G mutation only two of these carried the 845A mutation. If the 187G mutation complemented the 845A mutation with high penetrance in causing hemochromatosis, then the population frequency of the two genes would require that a high proportion of patients with hemochromatosis be heterozygous for 845A and 187G. Instead, the frequency of homozygotes for the 845A mutation was much higher than that of the 845A/187G genotype. Based on our data, the penetrance of the 845A/187G genotype is only 1.5% and based on the data of Feder et al. only 0.5%. In contrast, the penetrance of the homozygous 845A/845A genotype seems to be very high. Thus, screening for this genotype should be very useful.