The tectonics and volcanism of the terrestrial planets are controlled by the loss of heat from the planetary interior. On the Earth, about 70% of the heat flow through the mantle is attributed to the subduction of cold lithosphere. In order to understand the tectonic and volcanic processes on Venus it is necessary to understand how heat is transported through its mantle. In this paper, three alternative end-member hypotheses are considered. The first is the steady state loss of heat through the mantle to the surface in analogy to the Earth. However, without plate tectonics and subduction on Venus, a steady state requires either a very high plume flux or very rapid rates of lithospheric delamination. The required plume flux would be equivalent to about 80 plumes with the strength of the Hawaiian plume. The required delamination flux implies a 50% delamination of the entire Venus lithosphere every 10 m.y. Neither appears possible, so that it is concluded that Venus cannot transport heat through its mantle to its surface on a steady state basis. The second hypothesis is that there has been a strong upward concentration of the heat-producing elements into the crust of Venus; the heat generated is then lost by conduction. Surface measurements of the concentrations of the heat-producing elements place constraints on this model. If everything is favorable this hypothesis might be marginally acceptable, but it is considered to be highly unlikely. The third hypothesis is that heat is lost by episodic global subduction events followed by long periods of surface quiescence. The near-random distribution of craters suggests that the last subduction event occurred about 500 Ma. This model implies a thick thermal lithosphere (≈300 km) at the present time, which is consistent with a variety of surface observations. Lava lakes on the Earth are considered as analogies to plate tectonics; they also exhibit episodic subduction events. ¿ American Geophysical Union 1995