Energy is a scarce resource in ad hoc wireless networks and it is of paramount importance to use it e±ciently when establishing com- munication patterns. In this work we study algorithms for computing energy-e±cient multicast trees in ad hoc wireless networks. Such algo- rithms either start with an empty solution which is gradually augmented to a multicast tree (augmentation algorithms) or take as input an initial multicast tree and `walk' on di®erent multicast trees for a ¯nite number of steps until some acceptable decrease in energy consumption is achieved (local search algorithms). We mainly focus on augmentation algorithms and in particular we have implemented a long list of existing such algorithms in the literature and new ones. We experimentally compare all these algorithms on random geometric instances of the problem and obtain results in terms of the energy e±ciency of the solutions obtained. Additional results concerning the running time of our implementations are also presented. We also ex- plore how much the solutions obtained by augmentation algorithms can be improved by local search algorithms. Our results show that one of our new algorithms and its variations achieve the most energy-e±cient solu- tions while being very fast. Our investigations shed some light to those properties of geometric instances of the problem which make augmenta- tion algorithms perform well.