In this paper, the problem of controlling bilateral teleoperation systems in the presence of time-varying communication delays and dynamic uncertainties is studied. By exchanging only position signals between the master and slave robots, an adaptive controller is proposed to ensure stability if the control gains are contingent to the communication delays. Without exact knowledge of the time-varying delays, the developed control algorithm is able to ensure tracking errors between the master and slave robots to the origin in free motion. Moreover, the closed-loop control system with passive external forces from the human operator and remote environment is also studied. The proposed teleoperation system and control algorithm are validated using numerical simulations on 2-DOF robotic manipulators for both the master and slave robots.