Data gathering is one of the most important services provided by wireless sensor networks (WSNs). Since the predominant traffic pattern in data gathering services is many-to-one communication, it is critical to understand the limitations of many-to-one information flows and devise efficient data aggregation protocols to support prolonged operations in WSNs. In this paper, we provide a theoretical characterization of data aggregation processes under different communication modalities in WSNs. We demonstrate that data aggregation rates of Theta(log(n)/n) and Theta(1) are optimal when operating in fading environments with power path-loss exponents that satisfy 2 < alpha < 4 and alpha > 4, respectively. Furthermore, the optimal rate can be achieved using a generalization of cooperative beam-forming called cooperative time-reversal communication. In contrast, the non-cooperative multihop relay strategies widely adopted in literature are shown to be suboptimal in the low-to-medium attenuation regime (for 2 < alpha < 4).