I am given the following fourier-series:

$$\sum_{n=1}^{\infty} \frac{(-1)^n}{n}\,\sin(nt)$$

I want to figure out if the series converges uniformly or pointwise on $\Bbb R$, and if converges pointwise for $t=\frac{\pi}{2}$:

$$\sum_{n=1}^{\infty} \frac{(-1)^n}{n}\,\sin\left(\frac{\pi}{2}n\right)$$

Can I use the Weierstrass-M-Test here?

$$\left\lvert\frac{(-1)^n}{n}\sin\left(\frac{\pi}{2}n\right)\right\rvert \le \left\lvert\frac{1}{n}\right\rvert$$

and since the the series $\sum_{n=0}^{\infty}\frac{1}{n}$ diverges my series is not uniformly convergent.

  • $\begingroup$ Please take care to several points. (1) your definition of $f_n$ doesn't make sense (with your current definition the series is constant) and (2) you can't speak of uniform or pointwise convergence at a point. $\endgroup$ Jun 7, 2015 at 13:38
  • $\begingroup$ That's weird. It explicitly states on our problem sheet "investigate the behavior of $\sum_{n=1}^{\infty}\frac{(-1)^n}{n}\sin(nt)$ when $t=\frac{\pi}{2}$. Does the series converge uniformly or pointwise on $\Bbb R$". Are you saying this question doesn't make any sense? $\endgroup$
    – qmd
    Jun 7, 2015 at 13:43
  • $\begingroup$ Hint to explain (1) by Jean-Pierre: look for $n$s on both sides of an identity. $\endgroup$
    – Did
    Jun 7, 2015 at 13:45
  • $\begingroup$ @Did I am not sure what you mean. $\endgroup$
    – qmd
    Jun 7, 2015 at 13:49
  • $\begingroup$ Sorry but I do not "mean" anything, rather, I say that you should look for $n$ on both sides of the identity Jean-Pierre pointed you at. Do it. $\endgroup$
    – Did
    Jun 7, 2015 at 13:52

1 Answer 1


Your series is the Fourier series of the function $f(x)=-\frac{x}{2}$ over $(-\pi,\pi)$, extended by periodicity. The proof is straighforward, you just have to compute $$ \frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\sin(nx)\,dx $$ through integration by parts. Then the situation is the following:

$\hspace1in$ enter image description here

with pointwise convergence for every point of $I=(-\pi,\pi)$.

In virtue of Gibbs' phenomenon, the convergence on $I$ is not uniform.

  • $\begingroup$ Thankyou very much. Is there another way to show convergence is not uniform on $I$? I am not familiar with the Gibb'S phenomenon. $\endgroup$
    – qmd
    Jun 7, 2015 at 14:07
  • 1
    $\begingroup$ @Suh: A simple way is to notice that $f(x)$ is not continuous over $\mathbb{R}$, while every partial sum of your series is. $\endgroup$ Jun 7, 2015 at 14:08
  • $\begingroup$ While to prove pointwise convergence you may use Carleson's theorem or just the fact that for any $x\in I$ your series is the imaginary part of $-\log(1+e^{ix})$, plus Dirichlet's test. $\endgroup$ Jun 7, 2015 at 14:13
  • 1
    $\begingroup$ Thanks again for your help. $\endgroup$
    – qmd
    Jun 7, 2015 at 14:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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