In this work, the problem of receive antenna se- lection (AS) is considered, in a multiple antenna communication system having a single radio frequency (RF) chain at the receiver. The AS is performed on a per-packet basis, and AS decisions are based on noisy estimates of the channel gains obtained using pilot symbols embedded in the data packet for coherent demodulation, along with the receiver’s knowledge of the time correlation of the channel. The problem is posed as a partially observable Markov decision process (POMDP) with the goal of minimizing the average PER. We explore the optimality of a myopic policy by comparing its performance with that of the POMDP solution, and show that the myopic policy is optimal under certain conditions. As the POMDP approach requires the channel gains to be quantized to a finite set of states, we also propose two heuristic AS schemes that use the continuous-valued received pilot symbols to make AS decisions, and thereby offer comparable or better performance than the POMDP approach. Unlike previous work, the schemes proposed here for AS do not require a lengthy AS training phase to precede each data packet. The performance improvement offered by the POMDP solution and the proposed heuristic solutions relative to existing AS training-based approaches, is illustrated using Monte Carlo simulations.