I'm supposed to determine whether this sum diverges or converges and if it converges then find its value: $$ \sum_{n=2}^\infty \frac{1}{n^2-1}. $$

Using the comparison test I eventually showed that this converges. But I can't figure out how to show what this sum converges to. The only sums we actually found values for in my notes are geometric series which this clearly isn't.

I saw that I could use partial fraction decomposition to represent the terms as $$\frac{1/2}{n-1}- \frac{1/2}{n+1} $$ but that just gets me $\infty - \infty$, so this isn't the way to do it.

I'm not sure how to find the value of this sum. I don't need the full solution but a hint would be appreciated. Thanks. :)


Use partial fraction decomposition to represent the sum and notice that it is a telescoping series.

  • $\begingroup$ Got it. Thanks. :) $\endgroup$ – user307485 Jan 23 '16 at 22:47

Since $$\frac{1}{n^2-1}=\frac{1}{(n-1)(n+1)}=\frac12\left(\frac{1}{n-1}-\frac{1}{n+1}\right)$$ as you already have, then write for $N\in \mathbb N$
\begin{align}\sum_{n=2}^{N}\left(\frac{1}{n-1}-\frac{1}{n+1}\right)&=\left(1-\frac13\right)+\left(\frac12-\frac14\right)+\left(\frac13-\frac15\right)+\dots+\left(\frac1{N-1}-\frac1{N+1}\right)\\[0.2cm]&=1+\frac12+\left(\frac13-\frac13\right)+\dots+\left(\frac1{N-1}-\frac1{N-1}\right)-\frac1N-\frac1{N+1}\\[0.2cm]&=\frac32-\frac1N-\frac1{N+1}\end{align} In other words, this sum telescopes. Now let $N\to \infty$ (and of course do not forget 1/2 in front of the sum) to conclude that $$\sum_{n=2}^{+\infty}\frac{1}{n^2-1}=\lim_{N\to+\infty} \frac12\sum_{n=2}^N\left(\frac{1}{n-1}-\frac1{n+1}\right)=\frac12\lim_{N\to+\infty}\left(\frac32-\frac1N-\frac1{N+1}\right)=\frac34$$

  • $\begingroup$ OP says he has this. $\endgroup$ – Ross Millikan Jan 23 '16 at 22:47
  • $\begingroup$ @RossMillikan Yes, sorry. I edited my answer. $\endgroup$ – Jimmy R. Jan 23 '16 at 22:53
  • 1
    $\begingroup$ @JimmyR. You didn't expand the telescoping sum correctly. The answer should be 3/4. $\endgroup$ – user307485 Jan 23 '16 at 22:55
  • $\begingroup$ @user307485 Yes, you are right. $\endgroup$ – Jimmy R. Jan 23 '16 at 22:57
  • $\begingroup$ I think the partial sum is actually $\dfrac32-\dfrac1N-\dfrac1{N+1}$. $\endgroup$ – Akiva Weinberger Jan 23 '16 at 23:12

hint: $\dfrac{1}{n^2-1} = \dfrac{1}{2}\cdot \left(\dfrac{1}{n-1}-\dfrac{1}{n+1}\right) = \dfrac{1}{2}\left(\dfrac{1}{n-1}-\dfrac{1}{n}\right) + \dfrac{1}{2}\left(\dfrac{1}{n}-\dfrac{1}{n+1}\right)$. From this you see that there are $2$ sums you calculate, and using telescoping the first sum is $\dfrac{1}{2}$, and the second is $\dfrac{1}{4}$, thus the answer is $\dfrac{3}{4}$ as claimed.

  • $\begingroup$ OP says he has proved convergence. $\endgroup$ – Ross Millikan Jan 23 '16 at 22:46

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.