We focus on the problem of energy-efficient convergecast in sensor networks. This problem identifies the energy-latency tradeoff during convergecast. Whenever a group of sensors communicate an event of interest, the latency involved in delivering such messages to the base station should be minimized. Since the sensors are constrained by limited power and are mostly idle, it is important that the sensors conserve energy. We show how time division multiple access (TDMA) can be effectively used to provide energy-efficient convergecast. This solution allows the sensors to save energy when the network is idle and to switch to active mode when the network observes an event. Furthermore, for a typical application where the event probability is less than 10 15%, our solution improves the network lifetime by approximately 3 fold.