Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I have the following alternating series that I would like to determine whether it is absolutely convergent, conditionally convergent, or divergent:

$$ \sum\limits^{\infty}_{n=1} \frac{1+2(-1)^n}{n} $$

I have applied some tests and I find it reasonable to conclude that it is divergent.

As a sum of two series: $$ \sum\limits^{\infty}_{n=1} \frac{1}{n} + \sum\limits^{\infty}_{n=1} \frac{2(-1)^n}{n} $$

I believe a convergent series when added to a divergent series, results in a divergent series. If this isn't a fact then I would still be left to say that it is inconclusive.

Using the Alternating Series Test, with: $$ a_n = \frac{1+2(-1)^n}{n} $$ although this isn't of 'proper form' $ \sum\limits^{\infty}_{n=1} (-1)^n a_n $ the limit of $a_n $ does approach zero as $ n \rightarrow \infty $. As for monotonically decreasing, the limit of the ratio of absolute terms is divergent for $ n $ even and inconclusive for $ n $ odd, which has me concluding divergent by The Ratio Test as well as not monotonically decreasing, where:

$$ \lim\limits_{n \rightarrow \infty} \left\lvert \frac{2(-1)^n + 1}{2(-1)^n - 1} \frac{n}{n+1} \right\rvert $$

Am I on the right track here? Am I making any really improper assumptions? Was there a better way to go about with the proof?


share|cite|improve this question
+1 for showing you work. – Arturo Magidin Jun 17 '12 at 4:23
up vote 3 down vote accepted

To use the alternating series test, you need to rewrite the series first; you cannot use it directly with the $a_n$ you are given.

Define $b_n$ as follows: when $n$ is even, $b_n = \frac{3}{n}$; when $n$ is odd, $b_n = \frac{1}{n}$. Then $(-1)^nb_n = a_n$, so you can apply the alternating test to your original series by writing the series as $\sum (-1)^n b_n$.

Unfortunately, the alternating series test does not apply: remember that you need not only that $\lim\limits_{n\to\infty} b_n=0$, but also that $b_{n+1}\lt b_n$ for all sufficiently large $n$. However, for $n$ odd, we have $b_n\lt b_{n+1}$, since $\frac{1}{n}\lt \frac{3}{n+1}$ if and only if $n+1\lt 3n$, which holds for all $n$. So the alternating series test cannot be used.

The Ratio Test is likewise inconclusive: the limit does not exist. To see this, note that for $n$ odd, we have $$\left|\frac{a_{n+1}}{a_n}\right| = \frac{\quad\frac{3}{n+1}\quad}{\frac{1}{n}} = \frac{3n}{n+1}$$ which converges to $3$ as $n\to\infty$; whereas for $n$ even, we have $$\left|\frac{a_{n+1}}{a_n}\right| = \frac{\quad\frac{1}{n+1}\quad}{\frac{3}{n}} = \frac{n}{3(n+1)}$$ which converges to $\frac{1}{3}$. Since the sequence of odd terms and the sequence of even terms of $|\frac{a_{n+1}}{a_n}|$ converge to different things, $$\lim_{n\to\infty}\left|\frac{a_{n+1}}{a_n}\right|$$ does not exist. So the Ratio Test is inconclusive.

As to your first method, it's fine: indeed, the sum of a divergent series and a convergent series must diverge. To see this, note that if $\sum r_n$ and $\sum(r_n+s_n)$ both converge, then so does $\sum ((r_n+s_n)-r_n) = \sum s_n$, since the difference of convergent series converges. Thus, if any two of $\sum r_n$, $\sum s_n$, and $\sum (r_n+s_n)$ converge, then so does the third, and in particular if any one of the series diverges, then at least two must diverge.

share|cite|improve this answer
I think you meant "...and a convergent series mut diverge" , at the beginning of the 4th parraph. – DonAntonio Jun 17 '12 at 4:15
@DonAntonio: <BillyCrystal>I hate it when that happens...</BillyCrystal> thank you. – Arturo Magidin Jun 17 '12 at 4:17

If we pair the terms with 2n and 2n+1, we get, using Arturo Magidin's notation, $$b_{2n}-b_{2n+1} = \frac{3}{2n} - \frac{1}{2n+1} =\frac{3(2n+1)-2n}{2n(2n+1)} = \frac{4n+3}{2n(2n+1)} > \frac{1}{n}, $$ so the sum of the first 2n+1 terms is greater than $\sum_{k=1}^n \frac{1}{k} $ which diverges.

In general, when working with an alternating series. I find it useful to pair each even term with the following odd term.

This can be readily generalized to the case where the signs of the terms of a series to be summed follow a repeating pattern.

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.