Studies on HLA antigens were conducted in several patient populations with the following findings: (a) All 135 Japanese narcoleptic patients, eight of whom were considered to have "symptomatic" narcolepsy, were found to be HLA-DR2 and HLA-DQw1 positive. All 17 members of a subgroup of the original population were also found to be HLA-Dw2-positive. It was concluded that HLA-DR2 is a prerequisite for the development of narcolepsy and that the diagnosis of narcolepsy can be excluded if HLA-DR2 or HLA-Dw2 is negative. The distinction between idiopathic and symptomatic narcolepsy needs to be reconsidered. (b) Haplotype studies in three families with narcoleptic members enabled detection of children at high risk for narcolepsy. (c) Of the 54 patients with disorders of excessive daytime sleepiness other than narcolepsy, those with essential hypersomnia (EHS) had a higher frequency of HLA-DR2; the others had a lower frequency. The DR2-positive EHS group could include members with an incomplete form of narcolepsy; the DR2-negative EHS group had disorders essentially different from narcolepsy, although both positive and negative groups manifested hypnagogic hallucinations, sleep paralysis, and sleep onset REMs. Two further studies were conducted in subgroups of the original narcoleptic population studied. In a subgroup of 30 patients who underwent lymphocyte subset studies, no T-cell abnormalities were detected; it is unlikely that an autoimmune mechanism is involved in the development of narcolepsy. In a subgroup of 33 narcoleptic patients, Southern's blot analysis of DNA using a DQ beta probe revealed three specific restriction fragments. Further studies are necessary to locate the DNA locus that carries the susceptibility gene for narcolepsy.