The literature on aqueous thermogalvanic cells has been reviewed. Wherever possible, the power conversion efficiency, phi(r), relative to that of a Carnot engine operating between the same temperatures and the figure of merit, Z, of the cell have been extracted from the literature data in order to assess the cell’s possible use for solar energy conversion. The determination of phi(r) and Z in such a cell requires the temperature dependence of its open-circuit potential difference and the knowledge of the current delivery characteristics of the cell. The current delivery characteristics depend on the effects of activation overpotential, ohmic overpotential, and mass transport overpotential on the cell’s current. The determination of the cell performance is hindered because a number of researchers apply an external potential to their cells or use forced electrolyte stirring, both of which introduce unknown amounts of energy into the energy balance for the cell. The best performance found for an aqueous thermogalvanic cell which does not have such external energy inputs, is phi(r) = 0.50% and Z = 0.58 X 10(-4) K-1. Estimates of the maximum phi(r) likely to be obtained from an aqueous thermogalvanic cell suggest that it would be difficult to obtain values in excess of phi(r) = 1.2%, which corresponds to Z = 1.5 X 10(-4) K-1. These values are somewhat lower than the typical values for metal and semiconductor thermocouples. The main cause of the low efficiencies of aqueous thermogalvanic cells is the presence of high concentrations of water molecules which conduct heat from the hot to the cold electrode but which are not themselves charge carriers. It is shown that the replacement of aqueous electrolytes by molten salts would not provide a sufficiently large increase in phi(r) to justify the very high temperatures required for such systems.