In this paper we solve analytically the problem of distributing optimally a set of t slots between two readers in RFID dense environments. In these environments only one reader can transmit simultaneously, otherwise they would interfere each other, and tag identification would not be possible. This problem is addressed considering both static and dynamic frame length readers. The goal was to maximize the expectation on the number of tags successfully identified. Results demonstrate how the optimal time distribution outperforms trivial assignments, and how the results depend on the underlying reading algorithm of the reader.