An allele of the behavioural mutant pawn-B96 has been reported as a typical recessive gene but was found to show a peculiar inheritance. When the F2 progeny from crosses between the wild-type and pwB96 were obtained by autogamy, the 1[ratio ]1 phenotypic segregation ratio was observed as expected. However, two-thirds of the wild-type progeny in the F2 were thought to be heterozygotes because they became mixed progeny of wild-type and pawn clones in successive autogamies. Four marker genes showed the expected segregation ratio and stable phenotypes in these crossings. This result and the results of crossings using segregants from the above crosses indicated that parental pwB96 is a tetrasomy of the chromosome carrying the pwB gene. To determine the cause of chromosomal duplication in the mutant, the stability of the chromosome carrying the pwB locus was examined by genetic analyses. The disomy of both pwB and wild-type and the tetrasomy of pwB showed genotypes that were relatively stable during several autogamous generations. However, in clones initially pure for the tetrasomy of wild-type, disomic cells appeared within a few autogamous generations. The difference between the stabilities of the tetrasomy of pwB96 and that of the wild-type might be due partly to differences between the growth rate of tetrasomy and disomy in pwB96 and the wild-type, but mostly the result of an unknown contribution of the chromosome carrying the pwB96 allele to the tetrasomic composition.