Large volcanic swells on Venus are believed to be a manifestation of mantle upwelling, or hotspots. The study of these regions provides important information on the interior of the planet. Numerical experiments are carried out to examine the interaction of mantle plumes with the thermal lithosphere and a layer of depleted mantle, a product of pressure-release melting, using an axisymmetric finite element and finite difference code that incorporates temperature-dependent viscosity and pressure-release melting. The lithosphere is defined as a high-viscosity lid; plumes are initiated and maintained by a prescribed temperature at the base of the computational domain. The topographic uplift, the geoid-to-topography ratio, and the volume of pressure-release melt are compared to estimated values for possible hotspots on Venus to constrain the properties of plumes and the lithosphere. The effects of lithospheric thickness, depleted layer viscosity and thickness, mantle temperature, and plume temperature and duration on the surface observables are predicted. Models with a thermal lithospheric thickness of approximately 100--150 km are consistent with observations, assuming a mantle temperature of 1300 ¿C, a maximum plume temperature of approximately 1500 ¿C, and mantle plume durations of 150--250 m.y. To be consistent with estimated volumes of volcanics at Venusian hotspots, a significantly thinner thermal lithosphere requires a much cooler mantle; a thicker lithosphere requires a hotter mantle. Models with a depleted layer thickness of 100--250 km, in addition to a 100-km thick thermal lithosphere, predict the range of parameters found on Venus. Interpretation of large volcanic swells on Venus based on the evolutionary sequence predicted here implies that Beta, Atla, Western Eistla, and Imdr Regiones overlie either active or recently active mantle plumes. The geoid-to-topography ratio at Beta Regio is much larger than values found at other hotspots and may indicate that it is in an early stage of evolution or that the chemical or thermal boundary layers are significantly thicker. Bell, Dione, and Themis Regiones appear to represent very late stage, possibly now extinct, hotspots. An estimate of the total plume buoyancy flux for Venus is much lower than the value obtained for Earth. The small areal distribution of volcanism is consistent with a low level of ongoing resurfacing at a few active hotspots. ¿ American Geophysical Union 1996