The U.S. 1998 Highway Capacity Manual contains a delay model for evaluating signalized intersections. This model, consisting of three delay components, is a revised version of the model presented in Chapter 9 of the 1994 Highway Capacity Manual. There are ambiguities about what type of delays the 1998 model really estimates, and the ability of this model to yield accurate estimates for non-random flows is uncertain. To provide a better understanding of the model, this paper discusses the theoretical basis of the model and uses computer simulation to identify some limitations of the model. Only intersections controlled with pretimed signals are considered. The analysis reveals that the model has serious structural flaws that can lead to large overestimates of delays.