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

It is well known that $\sum_{p \text{ prime}} \frac{1}{p}$ diverges, and in fact - it behaves like log of the harmonic series: $$ \sum_{p \le x} \frac{1}{p} = \log \log x + O(1). $$ It is also well known that $\sum\limits_{p \text{ prime}} \frac{1}{p^2}$ converges. What is known about the rate? Letting $C = \sum\limits_{p \text{ prime}} \frac{1}{p^2}$, what can be said about $C - \sum\limits_{p \le x} \frac{1}{p^2}$?

I am reading an article (a survey of Artin's Primitive Root Conjecture - which follows from the GRH). I am trying to understand what are the condition on functions $f_1\le f_2$ tending to infinity in order that $$ \sum_{f_1(x) \le p \le f_2(x)} \frac{1}{p^2} = O\left(\frac{1}{\log x}\right). $$ Of course I can take $f_1(x) = \log\log\log x$, $f_2(x) = \log\log x$, but I want the general conditions.

share|cite|improve this question
up vote 6 down vote accepted

Lets rearrange the sum $C-\sum_{p\leq x}\frac{1}{p^{2}}=\sum_{p>x}\frac{1}{p^{2}}$. Using integration by parts this is $$\sum_{p>x}\frac{1}{p^{2}}=\int_{x}^{\infty}\frac{1}{t^{2}}d\left(\pi(t)\right)=\frac{\pi(t)}{t^{2}}\biggr|_x^\infty+2\int_{x}^{\infty}\frac{\pi(t)}{t^{3}}dt.$$ Using the prime number theorem, that is the asymptotic for $\pi(x),$ you can deduce that the quantity on the right hand side is $\sim\frac{1}{x\log x},$ which is your rate of convergence.

Just worth noting, this is exactly what we would expect. The tail of the sum over all integers has size $\frac{1}{x}$, that is $\sum_{n>x} \frac{1}{n^2}\sim \frac{1}{x}$ and the primes occur with density $\frac{1}{\log n}$ around $n$, so we would expect the tail of the sum to be of size $\frac{1}{x\log x}$. Partial summation/integration allows to prove this.

Edit: Replaced $\asymp$ with $\sim$, since as pointed out by Greg Martin in the comments, the Prime Number Theorem is strong enough to yield this/

share|cite|improve this answer
Using the prime number theorem actually gives an asymptotic formula for the tail of the series, not just the order of magnitude, right? – Greg Martin Jan 22 '12 at 10:29

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.