The application of railway signals on a rail layout is determined by various factors, principally the location of points of potential conflict, as well as the speed and frequency of trains and the movements they require to make.

Before discussing the application of signals, it is useful to highlight some situations where signals are not required:

Signals exist primarily to pass instructions and information to drivers of passing trains. The driver interprets the signal's indication and acts accordingly. The most important indication is 'danger', which means 'stop'. Not every signal is equipped with a 'danger' aspect.

Signals are provided where required to protect items of infrastructure where conflict may arise, including:

A signal cannot display a 'proceed' aspect unless the infrastructure item(s) that it protects are in the correct position for the passage of a train or, in the case of flat junctions and gauntlet track, no other signal is cleared for a conflicting move. This is enforced by interlocking.

Running lines (as distinct from sidings) are divided into sections. Under normal circumstances, only one train may occupy any section at a time. A signal is provided at the start of every section, which may only display a 'proceed' aspect when the section ahead is completely empty, at least as far as the next signal.

The length of the sections, and hence the distance between signals, determines the railway's capacity. A railway with short sections can accommodate more traffic than one with long sections.

The provision of signals is dependent on the intended use of the layout and the movements that are expected to take place. It is not necessary to provide a signal for every conceivable movement that the layout physically permits. If, during perturbed working, a movement has to be made in a direction for which no signal is provided, this can be done under special instruction. The same applies during signal failure.