Let a be a finite measure on the line. Let F be chosen from the Dirichlet prior D( a). Consider the random variable X = ∫ xF( dx). We show this has a density and give a formula for this in terms of a. This provides numerous examples for a theory of Cifarelli and Regazzini. We also give connections between this work, the Markov moment problem, and recent work of Kerov which gives an elegant way of choosing X given a. This paper gives examples and pointers to the literature.