MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $z_1,\ldots,z_n$ be complex numbers of modulus one. Does it exist an increasing sequence $k_j\in\mathbb{N}$ such that $\lim_{j\to\infty}z_i^{k_j}=1$ for all i?

share|cite|improve this question
up vote 7 down vote accepted

Yes. Let $z=(z_1,...,z_n)$, a point on the torus $(S^1)^n$. Since the torus is compact, the sequence $z^m$ has a convergent subsequence $z^{m_1}$, $z^{m_2}$,..., and we may choose it so that $k_j=m_{j+1}-m_j$ are nonnegative and increasing. Then ${\rm lim} z_i^{k_j}=1$ for all $i$.

share|cite|improve this answer
Pavel, you can also delete your comments :P – Mariano Suárez-Alvarez Feb 3 '10 at 18:31
Ah, thanks! I did. – Pavel Etingof Feb 3 '10 at 18:33


A standard lemma is that some element of $\{\alpha, 2\alpha, ..., m\alpha\}$ is within $1/(m+1)$ of 0 mod 1, since otherwise, there would have to be two multiples of $\alpha$ between some $k/(m+1)$ and $(k+1)/(m+1)$ which means their difference would be close to 1.

The same pigeonhole argument works on $(S^1)^n$. Consider the first $m^n+1$ multiples of $(\alpha_1,...\alpha_n)$. Two must be in the same $n$-dimensional box $[k_1/m,(k_1+1)/m]\times...[k_n/m,(k_n+1)/m]$ which means their difference is a multiple $t_m\times(\alpha_1,..,\alpha_n)$ within $1/m$ of 0 on each coordinate.

If you want more details and better approximations, then there are some multidimensional versions of simple continued fractions which might work, but this suffices to show that a sequence of integers $\{t_m\}$ exists so that $z_i^{t_m}$ is within $2\pi/m$ of 1.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.