This research investigates innovative behavior in the U.S. steel industry under the assumption that basic oxygen and large electric furnaces are competing technologies. The empirical model of innovation is based on recent theoretical research. The coefficients are estimated by a "seemingly unrelated" Tobit approach. The research finds innovation is strongly influenced by a demonstration effect. Adoption is also influenced by potential cost savings and technical progress. There is no evidence that large U.S. steel producers modernized more slowly than their smaller rivals. The large electric furnace becomes sufficiently attractive by 1980 that firms choose not to adopt basic oxygen.