The apparent associations between fragile X mutations and nearby microsatellites may reflect both founder effects and microsatellite instability. To gain further insight into their relative contributions, we typed a sample of 56 unrelated control and 37 fragile X chromosomes from an eastern Finnish population for FMR1 CGG repeat lengths, AGG interspersion patterns, DXS548, FRAXAC1, FRAXE and a new polymorphic locus, Alu-L. In the controls, the most common FMR1 allele was 30 repeats with a range of 20 to 47 and a calculated heterozygosity of 88%. A strong founder effect was observed for locus DXS548 with 95% of fragile X chromosomes having the 21 CA repeat (196 bp) allele compared to 17% of controls, while none of the fragile X but 69% of controls had the 20 repeat allele. Although the FRAXAC1 locus is much closer than DXS548 to FMR1 (7 kb vs. 150 kb), there was no significant difference between fragile X and control FRAXAC1 allele distributions. The FRAXE repeat, located 600 kb distal to FMR1, was found to show strong linkage disequilibrium as well. A newly defined polymorphism, Alu-L, located at approximately 40 kb distal to the FMR1 repeat, showed very low polymorphism in the Finnish samples. Analysis of the combined loci DXS548-FRAXAC1-FRAXE showed three founder haplotypes. Haplotype 21-19-16 was found on 27 (75%) of fragile X chromosomes but on none of controls. Three (8.4%) fragile X chromosomes had haplotypes 21-19-15, 21-19-20, and 21-19-25 differing from the common fragile X haplotype only in FRAXE. These could have arisen by recombination or from mutations of FRAXE. A second haplotype 21-18-17 was found in four (11.1%) fragile X chromosomes but only one (1.9%) control. This may represent a more recent founder mutation. A third haplotype 25-21-15, seen in two fragile X chromosomes (5.6%) and one (1.9%) control, was even less common and thus may represent an even more recent mutation or admixture of immigrant types. Analysis of the AGG interspersions within the FMR1 CGG repeat showed that 7/8 premutation chromosomes lacked an AGG whereas all controls had at least one AGG. This supports the hypothesis that the mutation of AGG to CGG leads to repeat instability and mutational expansion.