Inconsistent findings in X linkage studies of manic-depressive illness (MDI) have been ascribed to the presence of phenotypic uncertainties (incomplete penetrance), considerable variation in form and severity of MDI, and the likely presence of phenocopies (or false positives). In order to address some of these issues, previous X linkage data with colour blindness, glucose-6-phosphate dehydrogenase deficiency, and blood coagulation factor IX (F9) markers were reanalysed using a narrow and a broad definition of MDI. Our results confirm the X-linked hypothesis for MDI genetic transmission when controlling for diagnostic variation. The lod score (log of odds ratio) is reduced for a more conservative definition of the disease, but nevertheless remains significant. However, conclusive linkage between the MDI gene and the F9 gene in the Xq27 region is not maintained in our series. Our findings emphasize the need to reanalyse previous genetic data with more sophisticated diagnostic and statistical techniques.