Objective The effects of the CYP2D6*17 and *29 alleles on substrate specificity and enzyme activity were studied by correlating CYP2D6 genotype to phenotype with 4 probe drugs (codeine, debrisoquine, dextromethorphan, metoprolol) in black Tanzanians and white Swedes. Methods The black Tanzanian subjects represented the following 6 genotypic groups: A, (CYP2D6*1 or *2)/(*1 or *2) (n = 13); B, CYP2D6*17/*17 (n = 5); C, CYP2D6*29/*29 (n = 4); D, CYP2D6*1/*17 (n = 5); E, CYP2D6*5/*17 (n = 4); and F, various genotypes (n = 4). The white subjects were from 4 groups, as follows: A, (CYP2D6*1 or *2)/(*1 or *2) (n = 7); B, (CYP2D6*1 or *2)/(*3, *4, or *5) (n = 7); C, homozygous for defect alleles (n = 7); and D, duplicated CYP2D6 gene (n = 2). Results The metabolic ratios of the 4 probe drugs correlated significantly (rs = 0.69–0.92; P < .001) in both populations. Tanzanian subjects homozygous for the CYP2D6*17 allele were slower metabolizers when debrisoquine or dextromethorphan was used as the probe drug than when codeine or metoprolol was used, showing a different substrate specificity of CYP2D6.17 than of CYP2D6.1 and CYP2D6.2. This was confirmed with analysis of covariance of the different metabolic ratios for a subgroup of subjects carrying only the CYP2D6*17 mutated allele (n = 9) compared with all other subjects (n = 44). The metabolic ratios of dextromethorphan and metoprolol differed significantly among Tanzanian subjects homozygous for the CYP2D6*29 allele compared with those with CYP2D6*1 or *2 alleles. Conclusion We found differences in the disposition of 4 CYP2D6 probe drugs in black Tanzanians compared with Swedes. The differences were caused by the presence of CYP2D6.17 and CYP2D6.29. The results show that CYP2D6.17 exhibits altered substrate specificity compared with CYP2D6.1 and CYP2D6.2. Clinical Pharmacology & Therapeutics (2002) 71, 77–88; doi: 10.1067/mcp.2002.120239