In 1948, when seeking estimates of costs of alternative weapon systems, the potentially embarrassing error of relating costs to rates of output while ignoring another relevant variable, quantity of items produced, was made obvious by access to the airframe production data analyzed in this paper. By 1949 the present paper had been completed for the RAND Corporation, but reliance on "military classified" data and sources prevented open publication at that time. Although the sources and data have been declassified for several years, it seems appropriate to publish the paper in its original form now that the phenomenon is being incorporated in formal economic theory.