This paper discusses the results of an investigation into ways in which the safety risks of travel on road and rail interact with each other in Great Britain, other than through physical contact such as at level crossings. The two main foci of the paper are: (1) an analysis of the 'whole journey' risks of journeys for which the national rail system is the main mode, but which also include stages by other transport modes to provide access to the railway system: and (2) an analysis of the effect on safety risk of inter-modal transfers between rail and road. On (1), walking to and from stations was estimated to account on average for 65% of the overall door-to-door risk of being killed on rail journeys: the rail system itself accounts for 21% of the risk, and other access modes account for the remaining 14%. The average distance walked to and from stations is 0.9 km per rail journey, and this walking accounts for 5% of all walking nationally. On (2), it was found that increasing rail fares to fund railway safety measures may lead passengers to switch from rail to car, but for most sensible rail safety measures, the additional risks from such diversions are small compared with the intended rail safety benefits. However, for high-cost rail safety measures funded by passengers, the additional risks from diversions may be of the same order as the intended safety benefits. The last section of the paper explores the effects of variations in the casualty rates of rail users as pedestrians and car users. because their road risks may be different from those of all road users. Such variations could alter the detailed conclusions of the paper, but the scale of such effects appears to be modest. (c) 2008 Elsevier Ltd. All rights reserved.