We consider the strain energy recently proposed, on a phenomenological basis by Alan Gent to take into account limiting chain extensibility in rubber-like materials. We show that this model gives the simplest rational approximation for the reduced tensile force associated with uniaxial extension that satisfies the usual basic assumptions of continuum mechanics. Then by examination of the classical Treloar data on uniaxial extension of rubber, we explain why the Gent model cannot give good predictions for small and moderate strains. We propose some modifications and find a particular one which is able with a minimum number of phenomenological coefficients to give a very good fit to uniaxial data over the full range of deformations.