The hypothesis that the relative locations of the landmarks in a town is not accurately remembered before the paths between them have been learned was refuted in the present study showing that the reverse may in fact be true. In two of the three experiments carried out, undergraduates were found to give accurate estimates of the relative directions to different landmarks in an average‐sized Swedish town (Umeå with about 80 000 residents), even if the subjects had been living there no more than two to four months. However, for some reason which could not be disentangled relative distance estimates turned out to be less accurate. In the third experiment, undergraduates who had lived for two months in the town by contrast hardly remembered which places among designated ones they would pass by the most direct path, whereas those who had been one year longer in the town were much better although not perfect. Finally, no sex differences were observed and there were no important differences depending on whether memory was tested for the central area of the town or for the whole town including the outskir