This paper by NC Ankeny mentions that " S. Chowla has proved that there exist infinitely many primes $k$ where the first $c_1\log k$ residues $(\bmod k)$ are all quadratic residues".

I recently found a proof of this here.

The proof constructs such primes using this idea:

  1. Consider numbers which satisfies these condition : $n=1\bmod 8$ and $n=1\bmod r_i$ where $r_i$ are first consecutive odd primes less than $y$.
  2. Let $R:= 8\prod r_i $ , clearly $n=1\bmod R$ by Chinese remainder theorem.
  3. By Linniks there exist a prime $p$ in this arithmetic progression $\{ 1+rk_j\}$ such that $p= \mathcal{O}(R^6)$.
  4. Note that for this prime all numbers less than $y$ are quadratic residue, since p $\equiv 1 \bmod 8 \implies \big(\frac{a}{p} \big) = \big( \frac{p}{a} \big) $.
  5. By prime number theorem it is easy to see that y = $O(\log p)$

But this analysis clearly doest rule out the possibility of having least quadratic non-residue $O(\log p)$ . Infact if we write all the constants and work it out than it doesn't even rule out the possibility of least quadratic non residue less than $10\log p$.

I want to know is this just what Chowla proved or there is a better argument for this. Any refrence for the same would be of great help.

  • I found this paper by Graham and Ringrose which proved that there are infinitely many primes $p$ such that the least quadratic nonresidue $n(p)$ satisfies $n(p) \gg \log p \log \log \log p $. But still now my question is does it removes the possibility of proving n(p) = O(\log p \log \log \log p) – xyz Jul 14 '16 at 18:19
  • Under GRH $n(p) = O((\log p)^2)$ and, conjecturally $n(p) = O((\log p)^{1+\epsilon}), \forall \epsilon > 0$. – Felipe Voloch Jul 14 '16 at 18:31
  • @FelipeVoloch can you give any reference for the conjecture $n(p)=O((\log p)^{1+\epsilon})$. – xyz Jul 14 '16 at 19:10
  • There is a somewhat vague discussion in Bach and Sorensen, Math Comp 65 (1996) bottom of pg 1718. I've seen it in other places too, but I don't recall a specific reference. – Felipe Voloch Jul 14 '16 at 20:31
  • Montgomery showed that conditional on GRH there are infinitely many primes $p$ such that $n(p)\gg \log p \log \log p$. – Thomas Bloom Jul 16 '16 at 15:38

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.