Voronoi diagrams are widely used to represent geographical distributions of information, but they are not readily stacked in a hierarchical fashion. We propose a simple mechanism whereby each index Voronoi cell contains the generators of several Voronoi cells in the next lower level. This allows various processes of indexing, paging, visualization and generalization to be performed on various types of data. While one-level Voronoi indexes have been used before, and hierarchies of the dual Delaunay triangulation have been used for fast point location, we believe that this is the first time that the advantages of their integration have been demonstrated.