Genetic linkage studies have opened new vistas for behavioral and psychiatric genetics. However, phenotypic diversity and diagnostic uncertainties can lead to spurious linkage findings. A method of analysis is proposed that takes these factors into account. When applied to manic‐depressive disease, the results indicate that previous evidence for a major gene localized on the distal long arm of the X‐chromosome cannot be ascribed to phenotypic uncertainties and misclassifications, i.e., a type I error. Although the lod score (the logarithm of odds) favoring linkage is reduced with the more restrictive clinical definitions of the phenotype, it remains significant nonetheless. Thus, the linkage finding is robust over a range of phenotypic patterns and presumed phenocopy frequencies. The results also suggest that the X‐linked phenotype is a particularly severe form of manic depression characterized by early onset, high familial prevalence of the bipolar form, and high recurrence rate of major depression. These findings may have important implications for the design and interpretation of genetic linkage studies and for refining diagnostic techniques in mental disorders. Copyright © 1990, Wiley Blackwell. All rights reserved