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

$\pi(x)$ is the prime counting function (no. of prime within x)

For the interval $(x, x + \delta x]$, $\delta > 0$, what is the smallest integer $x_{0}$ such that for any $x >= x_{0}$, $\pi(x + \delta x) - \pi(x) > 0$ is always true?

For example, Bertrand's Postulate tells us that when $\delta = 1$, the smallest integer to make the above statement true is $x_{0} = 2$.

The following result might help: one paper by Rosser and Schoenfeld gives out two inequalities about $\pi(x)$:

$\frac{x}{\log{x}}(1 + \frac{1}{2\log{x}}) < \pi(x)$, for $x>= 59$,

and $\pi(x) < \frac{x}{\log{x}}(1+ \frac{3}{2\log{x}})$, for $x>1$

share|cite|improve this question
If you haven't already, you might want to look at, especially the section "Further results". – joriki Sep 30 '12 at 19:47
What is the requirement on $\delta$? – Ankur Sep 30 '12 at 20:14
@Ankur, there is no requirement for $\delta$. I'm curious if we can find the smallest integer for any $\delta$. – Lindsay Duran Sep 30 '12 at 23:13
up vote 4 down vote accepted

From Proposition 6.8 on pdf page 8 of DUSART, you may take $$ x_0 = \max \left( 396738, \; e^{\left( \frac{1}{5 \sqrt \delta} \right)} \right). $$ This is not the optimal value of your $x_0 = x_0(\delta)$ but it works.

Note: Dusart's adviser was Guy Robin, whose adviser was Jean-Louis Nicolas. It all fits.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= enter image description here =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

share|cite|improve this answer
This is great. Maybe it would be worth including the x < p < x(...) statement from the paper in your answer? – daniel Oct 1 '12 at 2:06
@daniel, I have a scanner at home, I made a jpeg of the appropriate page. My experience is that pdf's are not good enough for this website, not enough pixels or suchlike. – Will Jagy Oct 1 '12 at 2:48
Looks good to me. I found it very interesting. Hopefully others (including the OP) will too. – daniel Oct 1 '12 at 2:57
@WillJagy, Thank you! This is indeed very helpful! – Lindsay Duran Oct 1 '12 at 21:51

I don't think that the answer above is correct. In fact, such an $ x_{0} $ does exist. It is a consequence of the Prime Number Theorem. One may refer to the Wikipedia article on Bertrand's Postulate: As for the original question, I'm not sure if bounds have been established for the size of $ x_{0} $, except in certain cases. Examples are also given in the Wikipedia article mentioned above. One of the references listed there, an article by Lowell Schoenfeld, gives the following result: For any $ n \geq 2010760 $, there exists a prime between $ n $ and $ \left( 1 + \frac{1}{16597} \right) n $. I believe that $ 2010760 $ is the smallest $ n_{0} \in \mathbb{N} $ corresponding to $ \delta = \frac{1}{16597} $. However, I haven't read the article yet, so please go ahead and read it to help me verify what I've written here.

share|cite|improve this answer
I had misread the question; I read $\delta x$ as an additive constant. You're right that it's probably meant to be $\delta$ times $x$. I've deleted my answer. Note that "the answer above" isn't meaningful on this site, since answers are ordered according to criteria that can a) be selected differently by different voters (active/oldest/votes) and b) change over time. – joriki Sep 30 '12 at 19:41
@daniel, is it possible to find $x_{0}$ for any $\delta$? – Lindsay Duran Sep 30 '12 at 23:17
@LindsayDuran: Sorry, the result in the answer about Schoenfeld is not conditional. There are conditional results in the paper but I think this result is from studying the zeros of the zeta function very carefully. It's really a three-paper series. – daniel Oct 1 '12 at 1:14

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.