1: \begin{abstract} 

2: Let $p_1 \ge p_2 \ge \dots$ be the prime factors of a

3: random integer chosen uniformly from $1$ to $n$, and let

4: $$ 

5: \frac{\log p_1}{\log n}, \frac{\log p_2}{\log n}, \dots

6: $$

7: be the sequence of scaled log factors.

8: Billingsley's Theorem (1972), in its modern formulation, asserts that

9: the limiting  

10: process, as $n \to \infty$, is the Poisson-Dirichlet process with parameter $\theta =1$.

11:  

12: In this paper we give a new proof, inspired by the 1993 proof by

13: Donnelly and Grimmett, and extend the result to factorizations of

14: elements of normed arithmetic semigroups satisfying certain growth

15: conditions, for which the limiting Poisson-Dirichlet process need not

16: have $\theta =1$.  We also establish Poisson-Dirichlet limits, with

17: $\theta \ne 1$, for ordinary integers conditional on the number of

18: prime factors deviating from the usual value $\log \log n$.

19: 

20: At the core of our argument is a purely probabilistic lemma

21: giving a new criterion for convergence in distribution to a

22: Poisson-Dirichlet process, from which the number-theoretic

23: applications follow as straightforward corollaries. The lemma uses

24: ingredients similar to those employed by Donnelly and Grimmett, but

25: reorganized so as to allow subsequent number theory input to be

26: processed as rapidly as possible.

27: 

28: A by-product of this work is a new characterization of

29: Poisson-Dirichlet processes in terms of multi-intensities.

30: \end{abstract}

31: