The category adjustment (CA) model predicts that memory for items in a spatial category is biased towards the centre of the category (i.e. the category prototype) and away from the category boundaries (Huttenlocher, Hedges, & Duncan, 1991). These predictions are supported by studies that assess memory for dot locations within a circle. The current study tests whether memory for places in a large scale environment is similarly biased. Forty-three residents of Southampton (a UK city) were asked to locate 22 places in Southampton on a map. No significant bias was found towards the geographical centre. A second analysis, which tested whether remembered locations where biased towards the commercial centre, was significant. However, the results were not as predicted: places were remembered further away from, rather than closer to, the commercial centre. These findings suggest that the CA model needs to be refined to account for location memory in large-scale places.