Abstract
The Allen/Uzawa elasticity of substitution, still commonly used in the empirical studies to account for "net" substitutability, has been subject to a number of theoretical criticisms. In this paper we reconsider the Allen/Uzawa elasticity of substitution as a measure of "gross" substitutability, and show that to this relevant purpose it can be meaningfully interpreted in conjunction with the returns-to-scale properties of the technology.