Behçet's disease has been known to be strongly associated with a particular HLA-B allele, B51. To address the possibility that the HLA-C gene, which is closely linked to HLA-B but has been poorly defined for allo-antigen specificity by the serologic method is involved in the susceptibility to Behçet's disease, HLA-C genotyping was performed for 90 Japanese Behçet's disease patients by the PCR-SSP method. The frequencies of HLA-Cw*14 and -Cw*15 were significantly higher in the patient with Behçet's disease as compared to the controls (48.9% vs. 24.0%, p = 0.0005, and 17.8% vs. 7.3%, p = 0.0434, respectively). On the other hand, the frequencies of HLA-Cw*0304 and -Cw*01 were significantly decreased in the patient group as compared to the control group (7.8% vs. 25.0%, p = 0.0027, and 23.3% vs. 37.5%, p = 0.0398, respectively). The significantly higher HLA-Cw*14 and -Cw*15 alleles may tightly correlate with the B51 antigen, and hence may have increased as a result of a linkage disequilibrium with B51. Accordingly, the HLA-C allele frequencies were compared for the B51-positive or -negative patients and controls, but there was no HLA-C allele showing a significant difference between these patient and control groups. Conversely, analysis of the HLA-B allelic distribution in association with HLA-Cw*14 revealed that in the healthy controls, B44 and B51 were present at the frequencies of 57.1% and 35.7% of the HLC-Cw*14-positive individuals, respectively. In contrast, in the Cw*14-positive patients the frequency of B44 was merely 14.0% (p = 0.0001) and that of B51 was significantly high, amounting to 82.0% (p = 0.0001). These facts suggest that the pathogenic gene of Behçet's disease is not the HLA-C gene (HLA-Cw*14 and/or HLA-Cw*15) but the HLA-B gene (HLA-B51) itself or a non-HLA gene residing in the centromeric side of the HLA-B gene rather than in the telomeric side around the HLA-C gene. This finding supports our previous mapping result, which located the susceptible gene between the TNF and HLA-B genes.