A multipoint test of heterogeneity on published data from 57 families with the fragile X syndrome has been undertaken. The hypothesis being tested was that there are two loci coding for fragile X expression, mutations at either of which can produce the phenotype. No predivision of the families was undertaken, as the test used an admixture parameter. Maximum likelihoods of the hypothesis have been calculated and compared with those produced on assuming a single locus for fragile X. The data do not suggest that there are two such loci within the interval between probes 52a and St14. In particular, the large kindred published by Camerino et al. (1983) does not supply convincing evidence of heterogeneity under this test. It is argued that the observed heterogeneity between factor IX and fragile X must have another explanation. There is some evidence for a second locus for fragile X outside the interval noted above; this locus being most probably proximal to these probes. The majority of the data suggesting this result comes from a family published by Davies et al. (1985).