We propose Mote Herding, a new system architecture for large scale, heterogeneous sensor networks. Mote herding uses a mix of many 8-bit sensor nodes (motes) and fewer but more powerful 32-bit sensor nodes (microservers). Mote herding groups motes into flocks that are connected via a multihop network to a microserver acting as a shepherd. Shepherds exploit their greater communications and compute power to form an overlay network, with many flocks joining to form a herd. By keeping each flock small and utilizing several shepherds, the herd can support many nodes with better latency, reliability, and energy efficiency than homogeneous architectures.