In the cases we have seen so far, a random variable has a single distribution, possibly depending on parameters. While, in general, a random variable can have only one distribution, it is often easier to model a situation by thinking of things in a hierarchy. This allows a complicated process to be modeled by a sequence of relatively simple models placed in a hierarchy. Dealing with the hierarchy is no more difficult than dealing with conditional and marginal distributions. Mixture distributions are distributions that arise from a hierarchical structure.