Digital maps of geographic areas are increasingly common in many types of workplace, in education and in the public domain. Their interactivity and visual features, and the complexity of geographic(al) information systems (GIS) which create, edit and manipulate them, create special cognitive demands on the end-user which are not present in traditional cartographic maps or in most human-computer interaction (HCI). This thesis reviews cross-disciplinary literature regarding cognitive aspects of viewing and interacting with digital maps. Data from an observational study of GIS use, including real-time recordings of normal workplace activities, was analysed using various approaches to examine the interactive and visual aspects of people's work. The implications for cartographic, psychological and HeI aspects of GIS are discussed, in the context of the actual tasks people perform with them (rather than the computationally advanced analyses assumed by most literature). The second phase of the research examined the spatial knowledge attained and used during this interaction. The relevance of specific concepts in cognitive psychology, and of factors that create individual differences in cognition, are discussed in depth, alongside work in environmental and educational psychology, cartography and geography. A controlled experiment examined the degree to which task characteristics induce a different spatial model or reference frame when viewing a digital map. It was shown that even novice users can switch between considering the map as an abstract geometric display or as a geographical representation, without affecting performance. However, tasks forcing subjects to focus entirely on the geometry rather than the geography did affect performance in a surprise post-test photograph identification task. Map users' mental model or reference frame is apparently affected by these task constraints; this has implications for GIS design and practice as well as for understanding spatial cognition The study also considered the role of expertise and other individual difference factors, although conclusions were limited by sample size. Further research issues are highlighted, particularly regarding the knowledge structures and spatial language used in interpreting digital maps.