Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

So I am trying to show the following: $$\sum_{n \leq x} \frac{\mu(n)}{n} \log{\frac{x}{n}}=O(1) $$ so I tried partial summation as following:

Let $A(x)=\sum_{n \leq x} \frac{\mu(n)}{n}$, then we have $$\sum_{n \leq x} \frac{\mu(n)}{n} \log{\frac{x}{n}}= \int_1^{x} \frac{A(t)\mathrm dt}{t},$$ and $A(t)$ is clearly very small, or $o(1)$ for all $t \in [1,x]$. My question is how to go from here to conclude that the error term is $O(1)$?

share|cite|improve this question
For what it's worth, I showed $\sum_{n \leq x} \frac{\mu(n)}{n} \log{\frac{x}{n}}=O(1)$, using generalized Möbius inversion, as part of my answer to a previous question of yours. – Mike Spivey Sep 29 '11 at 3:13
If you go the partial summation route, you will need an estimate of $A(t)$ that's more precise than $o(1)$ to conclude the $O(1)$ result you want at the end. For example, if $A(t) = \frac{1}{1+ \log t}$, then $$\int_1^{x} \frac{A(t) dt}{t} = \left.\log (1+\log t)\right|_1^x =\log (1+\log x),$$ which is not $O(1)$. – Mike Spivey Sep 29 '11 at 3:30
So I think in your answer, there is a error:$$\sum_{n \leq x} \frac{1}{n}=\log{x}+\gamma+O(\frac{1}{x})$$, thus when use Generalized Mobius inversion, we have $$\sum_{n \leq x} \frac{\mu(n)}{n} \log{\frac{x}{n}}=-\sum_{n \leq x} \frac{\mu(n)}{n}\gamma-\sum_{n \leq x} \mu(n) O(\frac{1}{x})+1=-o(1)-\sum_{n \leq x} \mu(n) O(\frac{1}{x})+1$$, However, the middle part doesn't seems to be O(1). – Rob Sep 29 '11 at 5:56
I've corrected the mistake in my other answer. – Mike Spivey Sep 29 '11 at 13:13
Sure. Anything that is $o(1)$ is also $O(1)$, and so you have $O(1) + O(1) + O(1) = O(1)$. – Mike Spivey Sep 30 '11 at 0:16

As you mentioned, we have $$\sum_{n\leq x}\frac{\mu(n)}{n} \log\frac{x}{n}=\int_1^x \left(\sum_{m\leq t}\frac{\mu(m)}{m}\right) \frac{dt}{t}.$$

You need a particular bound on the inner sum. For instance, since $\sum_{m\leq t}\frac{\mu(m)}{m}\ll (\log t)^{-2}$, the result follows.

Here is cool trick however. Notice that $$\sum_{nd\leq x}\frac{\mu(n)}{nd}=\sum_{k\leq x}\frac{1}{k}\sum_{d|k}\mu(d)=1.$$ Now, write the first sum as $$\sum_{n\leq x}\frac{\mu(n)}{n}\sum_{k\leq\frac{x}{n}}\frac{1}{k}=\sum_{n\leq x}\frac{\mu(n)}{n}\left(\log\left(\frac{x}{n}\right)+\gamma+O\left(\frac{n}{x}\right)\right).$$

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.