We study the benefits of multicast routing in the performance of wireless ad hoc networks. In particular we show that if a node wishes to communicate with n distinct destinations, multicast can reduce the overall network load by a factor , when used instead of unicast. One of the implications of this scaling property consists in a significant increase of the total capacity of the network for data delivery. Hence, we show that the aggregate multicast capacity of wireless ad hoc networks is larger than the unicast capacity, when the group size n is small compared to the total number of nodes in the network. We discuss how these information theoretic results can be taken into consideration in the operation of a multicast protocol for wireless mesh networks using Multicast Overlay Spanning Trees (MOST). We perform simulations of the MOST protocol to compare with the theoretical results, and we present a fully working implementation for real network environments.