From about 1930 to the late 1960s a definition of addiction dominated in the United States that made opiate-style abstinence reactions essential, and distinguished sharply between true addiction and merely psychological drug habituation. This definition was so narrow that it left all stimulants out of the addictive category, and it was not uncontested. By looking at the postwar efforts of one of the chief architects of this definition, pharmacologist Maurice Seevers, to defend his conception of addiction in both scientific and policy realms, I demonstrate the contingency of this construction of addiction upon methodological commitments and industrial interests related to amphetamines and tobacco.