There have been claims that a gene on the X chromosome may contribute to susceptibility to schizophrenia. Crow (1988) initially proposed that such a gene might lie in the pseudoautosomal region, but when evidence that weakened this hypothesis accumulated, he proposed that a susceptibility locus might be present elsewhere on the sex chromosomes instead. DeLisi et al. (1994) found a small nonsignificant positive lod score between the marker DXS7 and schizophrenia, but other failed to replicate this finding. Another study reported by Crow and DeLisi's group was also weakly positive for this marker (Dann et al., 1997). This locus was then investigated in a collaborative study by Laval et al. (1997), which produced a nonparametric lod score of 2.44. Using a sample of 17 pedigrees from Britain and Iceland, we have also tested the hypothesis of linkage between DXS7 and schizophrenia. The 17 families were selected from a larger sample on the basis of an absence of male-to-male transmission for schizophrenia. These families were originally selected for having multiple cases of schizophrenia within them and for having no cases of bipolar affective disorder. We genotyped subjects for a marker at DXS7 and performed classical lod score and model-free linkage analysis using broad and narrow definitions of affection with schizophrenia. We found strongly negative lod scores and no evidence for linkage using model-free analysis. Therefore, this study does not support the hypothesis of linkage of schizophrenia to DXS7, and the evidence for a susceptibility locus on this part of the X chromosome is weakened.