In this paper, we use the locally flexible translog functional form to investigate the demand for energy and interfuel substitution in the United States and to provide a comparison of our results with most of the existing empirical energy demand literature. Motivated by the widespread practice of ignoring theoretical regularity, we follow Barnett's (2002) suggestions and estimate the model subject to theoretical regularity, using methods developed by Diewert and Wales (1987) and Ryan and Wales (2000), in an attempt to produce inference consistent with neoclassical microeconomic theory. Moreover, we use the most recent data, published by the U.S. Energy Information Administration (EIA), and in addition to investigating interfuel substitution possibilities in total U.S. energy demand, we follow Serletis et al. (2009) and also examine interfuel substitution possibilities in energy demand by sector. Moreover, we test for weak separability, with the objective of discovering the structure of the functional form in total energy demand as well as energy demand by sector.