When thinking about socio-cultural factors influencing settlement location choice, the accessibility of places in the landscape is a potentially important variable to take into account. How to define accessibility in such a way that it might be used as a variable that can be used for analysis is however still very much open to debate. Most published research considering landscape accessibility limits it to the ease with which humans can reach a certain location. So-called hiking equations are often applied to obtain cost surfaces of accessibility at the level of the individual pixel, and accumulative cost surfaces are then applied to find the travel time or energy expenditure needed to reach a single destination from all points in the area studied. By adding up these accumulated cost surfaces for each and every pixel, a map of differential accessibility of the landscape can be obtained (total path costs; Llobera 2000), a concept similar to the creation of Shimbel-matrices in network analysis. This accessibility can also be analyzed for different travel times (short/medium/long distance).However, these methods do not provide much information on the possible foci of movement in the landscape. We argue that to this purpose some additional steps are needed. We depart from the creation of multiple least cost paths from and to random locations (Verhagen in press). This will result in structures of least cost paths that resemble networks with different weights attached to the edges, but that do not have any real 'nodes', similar to the kind of structures that are analysed in space syntax or other forms of spatial network analysis. We can then assume that edges with a high weight will have been most attractive to travel, and thus may have been more attractive to settlement as well. However, the presence of settlement itself must undoubtedly have influenced travel intensity, and a combination of landscape based accessibility and network analysis techniques would seem necessary to get a grip of their combined influence. An exploratory network analysis was therefore performed to examine a range of structural features of the junctions in the landscape that were identified by the least cost paths. The relative prominence of these nodes as indicated through different network measures was then compared with their relative position to features in the landscape and with their relative position on the least cost paths.Unfortunately, satisfying software solutions to combine the raster-GIS cost surface techniques with node-and-edge-based network analysis software are missing at the moment. In this paper we want to present an exploration into possible solutions, and show some examples of what might be done by combining both analysis techniques, focusing on settlement location analysis and predictive modelling.ReferencesLLOBERA, M., 2000. Understanding movement: a pilot model towards the sociology of movement In Lock, G. (ed.) Beyond the Map. Archaeology and Spatial Technologies, pp: 65-84. IOS Press/Ohmsha, Amsterdam.VERHAGEN, P., in press. On the road to nowhere? Least cost paths, accessibility and the predictive modelling perspective. Proceedings of CAA2010, Granada.