Nerve membranes consist of an approximately equal mixture of lipids and proteins. The propagation of nerve pulses is usually described with the ionic hypothesis, also known as the Hodgkin-Huxley model. This model assumes that proteins alone enable nerves to conduct signals due to the ability of various ion channel proteins to transport selectively sodium and potassium ions. While the ionic hypothesis describes electrical aspects of the action potential, it does not provide a theoretical framework for understanding other experimentally observed phenomena associated with nerve pulse propagation. This fact has led to a revised view of the action potential based on the laws of thermodynamics and the assumption that membrane lipids play a fundamental role in the propagation of nerve pulses. In general terms, we describe how pulses propagating in nerve membranes resemble propagating sound waves. We explain how the language of thermodynamics enables us to account for a number of phenomena not addressed by the ionic hypothesis. These include a thermodynamic explanation of the effect of anesthetics, the induction of action potentials by local nerve cooling, the physical expansion of nerves during pulse propagation, reversible heat production and the absence of net heat release during the action potential. We describe how these measurable features of a propagating nerve pulse, as well as the observed voltage change that accompanies an action potential, represent different aspects of a single phenomenon that can be predicted and explained by thermodynamics. We suggest that the proteins and lipids of the nerve membrane naturally constitute a single ensemble with thermodynamic properties appropriate for the description of a broad range of phenomena associated with a propagating nerve pulse.