1: \begin{abstract} Using elementary results from Fourier analysis,

2: we provide an alternate proof of a necessary and sufficient

3: condition for the sum of $M$ independent continuous random variables

4: modulo $1$ to converge to the uniform distribution in $L^1([0,1])$,

5: and discuss generalizations to discrete random variables. A

6: consequence is that if $X_1, \dots, X_M$ are independent continuous

7: random variables with densities $f_1, \dots, f_M$, for any base $B$

8: as $M \to \infty$ for many choices of the densities the distribution

9: of the digits of $X_1 \cdots X_M$ converges to Benford's law base

10: $B$. The rate of convergence can be quantified in terms of the

11: Fourier coefficients of the densities, and provides an explanation

12: for the prevalence of Benford behavior in many diverse systems. To

13: highlight the difference in behavior between identically and

14: non-identically distributed random variables, we construct a

15: sequence of densities $\{f_i\}$ with the following properties: (1)

16: for each $i$, if every $X_k$ is independently chosen with density

17: $f_i$ then the sum converges to the uniform distribution; (2) if the

18: $X_k$'s are independent but non-identical, with $X_k$ having

19: distribution $f_k$, then the sum does not converge to the uniform

20: distribution.

21: \end{abstract}

22: