Sir-In the November 1994 issue of Nature Genetics, Collod and colleagues1 claim to provide definitive evidence for the assignment of a second locus for Marfan syndrome (MFS) to chromosome region 3p24. 2-p25. This conclusion is based upon genetic linkage study of a single large kindred, but may have serious consequences for molecular genetic testing and genetic counselling of many families with this disorder. Since the validity of linkage studies and resulting gene assignments is completely dependent upon the reliability of phenotype classification, we wish to challenge the conclusion of this study on the basis that (i) the clinical presentation of this family fails to meet the diagnostic criteria for MFS and (ii) the assignment of'affected'status by these authors is arbitrary. The fact that evidence for linkage to markers on chromosome 3p was found does not validate the phenotype assignments nor does it prove that a single mutant locus is segregating in this family. Spurious linkage can be generated by chance when a large number of markers are tested and by invoking reduced penetrance for a disorder known for its complete penetrance.

A consideration of the clinical diagnosis of MFS is central to this discussion. In 1986, an international consortium of investigators consolidated decades of clinical experience into diagnostic criteria for MFS, the so-called'Berlin nosology'2• In the absence of an unequivocally affected first-degree relative, involvement of the skeleton and at least two other organ systems is required. In addition, one of the'major'manifestations (lens dislocation, aortic dilatation or dissection, dural ectasia) must be present. With an unequivocally affected first-degree relative, one must only see involvement of two systems; a major manifestation, while preferred, is not required. It is also stressed that the phenotype within the family must be considered when deciding whether to give diagnostic significance to a specific feature seen in any given individual. If a physical finding is common in the general population