Since the initial report of linkage of autosomal dominant retinitis pigmentosa (adRP) to the long arm of chromosome 3, several mutations in the gene encoding rhodopsin, which also maps to 3q, have been reported in adRP pedigrees. However, there has been some discussion as to the possibility of a second adRP locus on 3q. This suggestion has important diagnostic and research implications and must raise doubts about the usefulness of linked markers for reliable diagnosis of RP patients. In order to address this issue we have performed an admixture test (A-test) on 10 D3S47-linked adRP pedigrees and have found a likelihood ratio of heterogeneity versus homogeneity of 4.90. We performed a second A-test, combining the data from all families with known rhodopsin mutations. In this test we obtained a reduced likelihood ratio of heterogeneity versus homogeneity, of 1.0. On the basis of these statistical analyses we have found no significant support for two adRP loci on chromosome 3q. Furthermore, using 40 CEPH families, we have localized the rhodopsin gene to the D3S47-D3S20 interval, with a maximum lod score (Zm) of 20 and have found that the order qter-D3S47-rhodopsin-D3S20-cen is significantly more likely than any other order. In addition, we have mapped (Zm = 30) the microsatellite marker D3S621 relative to other loci in this region of the genome.