Strategies for finding one's way through an unfamiliar environment may be helped by computer generated 2D maps, 3D virtual environments, or other navigation aids. The relative effectiveness of 2D and 3D virtual navigation aids was investigated. The wayfinding experiments (navigation tests) were conducted in a large, park-like environment. 24 participants (12 men, 12 women; age range = 22-50 years; M=32, SD = 7.4) were divided into three groups of four individuals per gender, who 1) explored a computer generated 2D map of the given route prior to navigation, 2) received a silent guided tour by means of an interactive 3D virtual representation, or 3) acquired direct experience of the real-world route through a silent guided tour where they were accompanied by a human individual who had expert knowledge of all the routes in the park. Participants from the different preparation groups then had to find the same route again on their own. 12 observers (six men and six women) were given a "simple" route with only one critical turn, and the other 12 a "complex" route with six critical turns. Navigation performances were compared with those of three experts who were highly familiar with all the routes of the park. Those among the naive participants who had benefitted from a direct experience (guided tour) prior to navigation, all found their way again on the simple and complex routes. Those who had explored the interactive 3D virtual environment were all unable to find their way on the complex route. The relative scale representation in the virtual 3D environment may have given incorrect impressions of relative distances between salient objects along the itinerary, rendering important landmark information useless.