1: \begin{abstract}

2: This is the second part of the series on constructing uniformly consistent

3: estimators of the proportion of false null hypotheses via solutions to

4: Lebesgue-Stieltjes integral equations. We consider estimating the proportion

5: of random variables for two types of composite null hypotheses: (i) their

6: means or medians belonging to a non-empty, bounded interval; (ii) their means

7: or medians belonging to an unbounded interval that is not the whole real line.

8: For each type of composite null hypotheses, uniform consistent estimators of

9: the proportion of false null hypotheses are constructed respectively for

10: random variable that follow Type I location-shift family of distributions and

11: for random variables whose distributions form continuous natural exponential

12: families with separable moments. Further, uniformly consistent estimators of

13: the proportion induced by a function of bounded variation on a non-empty,

14: bounded interval are provided for the two types of random variables mentioned

15: earlier. For each proposed estimator, its uniform consistency class and speed

16: of convergence are provided under independence. \medskip\newline%

17: \textit{Keywords}: Bessel functions; Composite null hypothesis; Dirichlet

18: integral; Lebesgue-Stieltjes integral equations; natural exponential family;

19: proportion of false null hypotheses\medskip\newline\textit{MSC 2010 subject

20: classifications}: Primary 35C05, 60E05; Secondary 60F10.

21: 

22: \end{abstract}

23: