Ultraviolet radiation (UVR) may protect against prostate cancer via a mechanism involving vitamin D. Thus, the vitamin D receptor (VDR) gene is a susceptibility candidate, though published data are discrepant. We studied the association of prostate cancer risk with five VDR single nucleotide polymorphisms (SNPs): G/A(1229) (SNP 1), A/G(3944) (SNP 2), T/C(30875) (SNP 3), C/T(48200) (SNP 4) and C/T(65013) (SNP 5), in 430 cancer and 310 benign prostatic hypertrophy (BPH) patients. The SNP 2 GG genotype frequency was lower in cancer than BPH patients (odds ratio = 0.63, 95% CI = 0.41-0.98, p = 0.039). SNPs 1 and 2, and SNPs 4 and 5, were in linkage disequilibrium. Two copies of haplotypes comprising SNPs 1-2, G-G (odds ratio = 0.63, p = 0.039), SNPs 2-3 G-C (odds ratio = 0.45, p = 0.008) and SNPs 1-2-3 G-G-C (odds ratio = 0.44, p = 0.006), but not SNPs 1-3, G-C (odds ratio = 0.81, p = 0.34), were associated with reduced risk (reference, no copies of the haplotypes). These associations were observed after stratification of subjects by extent of UVR exposure. These data show that SNP 2 GG genotype mediates prostate cancer risk, complementing studies reporting this allele is protective in malignant melanoma pathogenesis. They further suggest that published associations of risk with SNP 1 may result from linkage disequilibrium with SNP 2.