In spite of the importance accorded the steam engine during nineteenth-century industrialization, little is known about its rate of diffusion and the determinants thereof in the United States. The primary purpose of this paper is to enhance our knowledge about the spread of this technology. New evidence on steam power use in 1820, 1850, and 1860, combined with published census data from 1870, permits quantitative estimates of the regional variations in timing, pace, and extent of usage before 1900. Second, we advance reasonable conjectures for the regional differences that appear. Although lack of evidence precludes a definitive delineation of causality, with simulation techniques we are able to use the limited evidence available on costs to reconcile, albeit imperfectly, the historical pattern with economic-theoretic predictions regarding the process of innovation.