A suggested application of a thermoelectric generator is to exploit the natural temperature difference between the air and the soil to generate small amounts of electrical energy. Since the conversion efficiency of even the best thermoelectric generators available is very low, the performance of the heat sinks providing the heat flow is critical. By providing a constant heat input to various heat sinks, field tests of their thermal conductances in soil and in air were performed. A prototype device without a thermoelectric generator was constructed, buried, and monitored to experimentally measure the heat flow achievable in such a system. Theoretical considerations for design and selection of improved heat sinks are also presented. In particular, the method of shape factor analysis is used to give rough estimates and upper bounds for the thermal conductance of a passive heat sink buried in soil.