The gene for Friedreich ataxia (FA), a severe recessive neurodegenerative disease, has previously been shown to be tightly linked to the polymorphic markers D9S15 and D9S5 on human chromosome 9. In addition, the observation of linkage disequilibrium suggested that D9S15 is within 1 centimorgan (cM) of the disease locus, FRDA. Although D9S5 did not show recombination with FRDA, its localization was less precise (0-5 cM) due to its lower informativeness. We have now identified additional polymorphisms at both marker loci. Two cosmids spanning 50 kilobases around D9S5 were isolated, and a probe derived from one of them detects an informative three-allele polymorphism. We have found a highly polymorphic microsatellite sequence at D9S15 which is rapidly typed by the DNA polymerase chain reaction. The polymorphism information contents at the D9S5 and D9S15 loci have been increased from 0.14 to 0.60 and from 0.33 to 0.74, respectively. With the additional polymorphisms the lod (log10 odds ratio) score for the D9S15-FRDA linkage is now 48.10 at recombination fraction theta = 0.005 and for D9S5-FRDA, the lod score is 27.87 at theta = 0.00. We have identified a recombinant between D9S15 and FRDA. However, due to the family structure, it will be of limited usefulness for more precise localization of FRDA. The linkage disequilibrium previously observed between D9S15 and FRDA is strengthened by analysis of the haplotypes using the microsatellite polymorphism, while weaker but significant disequilibrium is found between D9S5 and FRDA. Extended haplotypes that encompass D9S5 and D9S15 show a strikingly different distribution between chromosomes that carry the FA mutation and normal chromosomes. This suggests that both marker loci are less than 1 cM from the FRDA gene and that a small number of mutations account for the majority of FA cases in the French population studied. D9S5 and D9S15 are thus excellent start points to isolate the disease gene.