A three equation model of innovation in the chemical industry has been developed. The model includes the following three relationships: (1) a technical knowledge production function with technical knowledge as a function of research and development input, (2) a production function with output as a function of conventional inputs (capital, labor, and materials) and a 'stock' of knowledge, and (3) a derived demand equation for innovation inputs. The model was tested using data on 48 publicly owned firms covering a 20 year period 1960-1980. Financial data were taken from the firms' 10-K annual reports, and patent data, used as measures of innovation, were obtained from the U.S. Patent Office. The resulting regression estimates (1) supported the hypothesis that all firms have the same elasticity of patenting with respect to their R & D stock, (2) yielded estimates of elasticity of output relative to R & D stock about twice as large as an earlier study, and (3) supported a model of R & D intensity as a function of expected growth rate. The report suggests additional refinements and testing of the model and methods for utilizing it in assessing the impact of the Toxic Substances Control Act.