Three-dimensional virtual worlds promise to complement the Web with 3D models of virtual places that are fanciful or that model and mirror the real world. Marketplace evolution is one way to wait and see if and how this will come about. Another approach is to identify current limitations of virtual worlds, deconstruct and study their architectures, and consider how to evolve them to realize their promises. This special issue presents articles that explore architectural issues related to virtual world evolution. The articles cover topics like integration of 3D virtual world viewers with Web browsers, federation of virtual world grids, extensibility of virtual world platforms, adding knowledge-based services, accuracy of virtual worlds, standards, and serious applications built on top of virtual world platforms to support collaboration and training.