# The limit points of $\{\sin (n+n^2)\mid n\in\mathbb N\}$

We can prove that $\exists \{c_i\}_{i=1}^{\infty} \subseteq \mathbb N$ and $c_1<c_2<c_3<...$ such that $\lim_{i\rightarrow \infty}\sin (c_i^{\alpha})=c$, $\forall c\in [-1,1]$,where $\alpha$ is a positive rational number.

But what about the limit points of { $\sin(n+n^2)\mid n\in N\}$? Is it also the closed interval $[-1,1]$?

Furthermore, what about $\sin \bigg(\sum_{i=1}^ma_in^{i}\bigg)$, where all coefficients $a_i$ are rational numbers?

• You may want to write $n+n^2=\frac{(2n+1)^2}4-\frac14$; not sure though if it helps... – punctured dusk Aug 16 '15 at 10:26

After being puzzled by this question for most of the day I received a tip that this can be proved by using van der Corput's theorem (which I never heard about before), it clearly works for the $n^2+n$, but should also work for all polynomials with rational coefficients* according to my source (I guess it's done by consecutive applications).
• As a matter of fact it's not necessary for the coefficients to be rational, a sufficient condition is that at least one coefficient is not a $q\pi$ for some rational $q$.