Rejoinder to discussion of ``The asymptotic distribution of P-values in composite'' null models.