The effects of available visual information in the periphery (enlargement of the functional visual field) on performance in navigation were evaluated in an experimental setup searching for a “compromise” between desktop and head-immersion situations. A Fixed Vision condition (fixed display) and two Mobile Vision conditions (head-tracking with a visual field of variable width) were compared in six virtual environments of different complexity and in four successive sessions. First, a global improvement in performance throughout the sessions revealed a gradual integration of the properties of the simulation device. Second, performance was higher in the Mobile Vision conditions, as shown by the smoothness of the subjects' paths (sharp curves could be negotiated without stopping), indicating the importance of a wide functional visual field. In conclusion, the need to design realistic and functionaly efficient human-machine interfaces for navigation is discussed.