Long-term, large-scale empirical studies on the simultaneous development of transport infrastructure and the built environment are scarce. This paper provides a long-term study of the development of the railway network and its impact on the built-up area—and vice versa—using the case study of the Randstad in the Netherlands between 1850 and 2010. The analysis is both qualitative and quantitative. We describe the shares of the built-up area in concentric buffers of 1-kilometer intervals from railway stations and estimate binomial logit models to predict the likelihood of new stations being built based on the amount of the preceding and subsequent built-up area and the likelihood that a new station might have encouraged further growth. Results show that during the early days stations followed existing urbanization patterns. But as time went by, new stations were more likely to be located in undeveloped areas and less likely to be located within the established built-up areas, which were already serviced by existing stations. Moreover, they prompted further growth, increasing the likelihood of more urbanization in their vicinity.