While experimenting with random polynomials I've found this conjecture:

A polynomial $f\in\mathbb Z[X]$ of degree $n$ with co-prime coefficients have no fixed prime divisor $p> n$.

A fixed prime divisor is a prime $p$ such that $p|f(m)$ for all $m\in\mathbb Z$.

Is this known? Proved? Or are there counterexamples?

  • $\begingroup$ Where did you find this conjecture? $\endgroup$ Jan 10, 2019 at 21:37
  • $\begingroup$ @SmileyCraft: as I wrote, I've done some experimenting. With my own software Bigz. $\endgroup$
    – Lehs
    Jan 10, 2019 at 21:39
  • $\begingroup$ do you mean pairwise co-prime? If you mean that there should be no prime which divides all the coefficients then it seems to be false. For instance, take $f(x)=x(x+1)(x+2)=x^3+3x^2+2x$. Clearly $2\,|\,f(n)$ for all $n\in \mathbb Z$. $\endgroup$
    – lulu
    Jan 10, 2019 at 21:47
  • 2
    $\begingroup$ Ok, but the $p>n$ case is trivial. No polynomial of degree $n<p$ can vanish for all the residues $\pmod p$ and your polynomial can't reduce to $0\pmod p$ since that would imply that every coefficient was divisible by $p$. $\endgroup$
    – lulu
    Jan 10, 2019 at 21:54
  • 1
    $\begingroup$ And little Fermat implies that $p | x^p-x$ for all primes $p$ $\endgroup$ Jan 10, 2019 at 21:55

1 Answer 1


Since $p$ does not divide each of the coefficients of $f(x)$, $f(x)$ reduces to a non-trivial polynomial $\overline f(x)$ of degree $≤n$ $\pmod p$. But if $p>n$ then $\overline f(x)$ can have at most $n$ roots $\pmod p$, hence it is non-zero on at least one residue $a \pmod p$. But then $f(a)\not \equiv 0 \pmod p$ and we are done.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.