To describe a kindred with a dominantly inherited neurologic disorder manifested either as uncomplicated spastic paraplegia or ataxia, spastic paraplegia, and mental retardation. Neurologic examinations and molecular genetic analysis (exclusion of known SCA and HSP genes and loci; and trinucleotide repeat expansion detection [RED]) were performed in six affected and four unaffected subjects in this family. MRI, electromyography (EMG), and nerve conduction studies were performed in three affected subjects. The phenotype of this dominantly inherited syndrome varied in succeeding generations. Pure spastic paraplegia was present in the earliest generation; subsequent generations had ataxia and mental retardation. MRI showed marked atrophy of the spinal cord in all patients and cerebellar atrophy in those with ataxia. Laboratory analysis showed that the disorder was not caused by mutations in genes that cause SCA-1, SCA-2, SCA-3, SCA-6, SCA-7, SCA-8, and SCA-12; not linked to other known loci for autosomal dominant ataxia (SCA-4, SCA-5, SCA-10, SCA-11, SCA-13, SCA-14, and SCA-16); and not linked to known loci for autosomal dominant hereditary spastic paraplegia (HSP) (SPG-3, SPG-4, SPG-6, SPG-8, SPG-9, SPG-10, SPG-12, and SPG-13) or autosomal recessive HSP SPG-7. Analysis of intergenerational differences in age at onset of symptoms suggests genetic anticipation. Using RED, the authors did not detect expanded CAG, CCT, TGG, or CGT repeats that segregate with the disease. The authors describe an unusual, dominantly inherited neurologic disorder in which the phenotype (pure spastic paraplegia or spastic ataxia with variable mental retardation) differed in subsequent generations. The molecular explanation for apparent genetic anticipation does not appear to involve trinucleotide repeat expansion.