A mobile ad hoc network is usually assumed to be homogeneous, where each mobile node shares the same radio capacity. However, a homogeneous ad hoc network suffers from poor scalability. Recent research has demonstrated its performance bottleneck through both theoretical analysis and simulation experiments and testbed measurements. This is further exacerbated by heavy routing overhead of ad hoc routing protocols when the network size is large. In this paper, we present a design methodology to build a hierarchical large-scale ad hoc network using different types of radio capabilities at different layers. In such a structure, nodes are first dynamically grouped into multi-hop clusters. Each group elects a cluster-head to be a backbone node (BN). Then higher-level links are established to connect the BNs into a backbone network. Following this method recursively, a multilevel hierarchical network can be established. Three critical issues are addressed in this paper. We first analyze the optimal number of BNs for a layer in theory. Then, we propose a stable and light overhead clustering scheme to deploy the BNs. Finally landmark ad hoc routing (LANMAR) is extended to operate the physical hierarchy efficiently. We show that the hierarchical LANMAR can incorporate and efficiently utilize backbone links to reach remote destinations (thus reducing the hop distance). Simulation results using GloMoSim confirm that our proposed schemes achieve good performance.