Between 1896, when Kraepelin published his first formulation of dementia praecox (DP), and 1917, when the American Medico-Psychological Association issued the first official American psychiatric nosology that contained DP and manic-depressive insanity (MDI)--Kraepelin's key categories--psychiatric nosology in the United States underwent a transformation. I describe and contextualize historically this process using Thomas Clouston, a Scottish Psychiatrist and widely-read textbook author, as a representative pre-Kraepelinian diagnostician. Clouston used three major diagnostic categories based on symptomatic presentation--mania, melancholia and paranoia--all derived from the beginnings of modern psychiatry in the early nineteenth century. He observed that these categories contained good-outcome cases and those progressing to 'secondary dementia'. Kraepelin designed his categories of DP and MDI to reflect putative distinct disease processes reflected in their course and outcome. Although Clouston and Kraepelin each saw similar patients, their nosologies started from different first principles: symptomatic presentation versus presumed etiology. Driven largely by social forces with American psychiatry, Kraepelin's system spread throughout the United States in the succeeding decades replacing older diagnostic approaches typified by Clouston's. In 1896, American psychiatry was demoralized as the idyllic asylums had become overcrowded, isolated scientific backwaters. Kraepelin's nosology was derived from and was championed by individuals working in high-status research-based university psychiatric clinics. It brought excitement, the promise of subsequent research breakthroughs and the high prestige then associated with German biomedicine. Scientific research comparing the older and Kraepelinian diagnostic systems played little role in this transition. Using empirical methods to guide changes in our diagnostic system is a recent development in the history of psychiatry.