You can see that by using references, you get a lot more modeling flexibility. However, since there are no visual representations of references in the model diagrams, the dependency structure is hidden 窶彿nside窶 the edges. Thus, if you use a lot of references in your models, the diagrams may eventually become very difficult to interpret.
To overcome such problems, it is often a good idea to include 窶彡omments窶 in your model indicating the hidden dependencies.