Among the most enigmatic features on Venus are the more than 500 quasi-circular structures collectively referred to as coronae. Associated volcanism and tectonism suggest that coronae result from small, transient mantle plumes, although only about 20% have clearly associated positive gravity anomalies indicative of current activity. Coronae are concentrated within a region defined by the dynamically active highland areas of Beta, Atla, and Themis Regiones (BAT). Coronae avoid the lowest or highest topography and geoid. Subsets of the corona population indicate that the spatial distribution of coronae has evolved from a previously less concentrated global distribution to the current BAT concentration. Recent laboratory experiments on thermal convection at high Rayleigh number, using fluids with temperature-dependent viscosity, suggest a straightforward interpretation of these relationships. Small-scale upwellings from a hot lower thermal boundary layer tend to be excluded by longer-wavelength downwelling structures from the cold upper boundary layer. The relative paucity of coronae in the broad lowland plains regions is consistent with the inference of mantle downwellings beneath the plains. Coronae are largely excluded from compensated highlands, perhaps because the thickened crust is not easily disrupted in these regions. Laboratory experiments indicate that small, transient, boundary layer instabilities can coexist with, and be focused toward, large-scale upwellings, possibly explaining the spatial concentration of coronae in the BAT region. Thus coronae provide important insights into interacting scales of convective flow in Venus's mantle and into the evolution of Venus's mantle dynamics. Direct tests of our model are possible through geological mapping.