Twenty-three, mostly nuclear, families were ascertained through a JME proband. The affected status of relatives of the probands was assigned by 4 different clinical criteria, and separate analyses were done assuming an autosomal dominant model with 90% penetrance and an autosomal recessive model with full penetrance. A linear age-of-onset correction with maximum penetrance at age 20 years was incorporated into the analyses. The maximum lod score obtained was 3.11 at theta-(m) = 0.001, theta-(f) = 0.20, assuming autosomal dominant inheritance and using the second definition of the disease phenotype. There was strong support for linkage using the other phenotype definitions and the autosomal dominant model, although the lod scores did not exceed 3.0. There was also support for linkage of a JME locus to this region under the autosomal recessive model, although the results varied depending upon the definition of the disease phenotype. There was no significant evidence for linkage heterogeneity.