A theory is developed for incorporating prior probabilities into multiple hypotheses testing. Its potential aaplications are illustrated. It is shown that the theory is particularly useful when one wishes to incorporate partial prior information in testing a parameter of interest in the presence...