IN THE EIGHTH edition of his textbook, Kraepelin¹ first noted that patients with late-onset variants of dementia praecox had atypical features and lacked the deterioration of personality typical in younger patients. Bleuler² concluded that such late-onset cases of "schizophrenia" were rare, noting that only 15% of hospitalized patients had onset of symptoms after age 40 years, and a mere 5% after age 60 years. Assuming a lifetime prevalence of schizophrenia near 1%, the latter figure implies that the proportion of elderly people who will develop this condition is about 0.05%. Roth and Kay^3,4 applied the term "late paraphrenia" to these cases, and noted the lack of a family history of schizophrenia (compared with younger-onset cases) and a preponderance of women. In their classic textbook, Slater and Roth⁵(p580) asserted that such illness accounted "for some89 percent of all first admissions (among women) over the age of 65." The relevant admission rates at the time (200-400 cases per 100 000) suggested a population prevalence among elderly women of 0.016% to 0.036% (this was presumably lower among men).⁵(p534) Thus, for decades it has been taught that isolated paranoid disorders among elderly people are rare.