We have previously reported a large Israeli-Arab pedigree with sensorineural deafness possibly determined simultaneously by two loci--one mitochondrial, and one autosomal recessive. This was analyzed by extending classic segregation analysis methods to the many nuclear families derived from the maternal line pedigree. Here we expand this pedigree and extend our analysis by using the regressive models for segregation analysis on the entire pedigree. The corresponding REGD computer program was utilized and the marrying-in males' and paternal line members' affection statuses were assigned as unknown to accommodate the exclusive maternal transmission pattern. For the autosomal locus, a simple autosomal recessive (q = 0.52) model with a nearly complete penetrance (0.93) was found to be the best-fitting model. Equally importantly, we were also able to use the power of the regressive models to test the hypothesis of mitochondrial heteroplasmy as an alternative for the proposed autosomal locus. We found no evidence for the heteroplasmy hypothesis as an explanation for the incomplete maternal transmission of deafness in this pedigree. Thus, even if the mitochondrial mutation occurred in a heteroplasmic distribution in the family members, this could not explain the familial aggregation in this pedigree, and an autosomal recessive locus is still required. These results provide further support for the concept that the sensorineural deafness occurring in this large Israeli-Arab pedigree results from simultaneous involvement of two genes at two different loci, one mitochondrial and likely homoplasmic, and the other autosomal and recessive.