Here I found a question:

Show that every prime not equal to $2$ or $5$ divides infinitely many of the numbers $1,11,111,1111,\dots$ etc.

which is partly solved here Prime numbers divide an element from a set.

From this the following conjecture seems reasonable:

Given any finite set $S=\{q_1,\dots,q_k\}$ of $k$ primes, then any prime $p\notin S$ divides infinitely many of the numbers $a_1,a_2,a_3\dots$, where $a_1=1$ and $a_{n+1}=1+a_n\prod q_i$.

Can this be proved?

  • $\begingroup$ Please don't reapply the conjecture tag to personal conjectures that turn out to be (well-)known results. $\endgroup$ – Bill Dubuque May 6 at 7:15

By induction $\,a_n = \dfrac{q^n-1}{q-1},\ q = \prod q_i.\,$ $a_n$ is a divisibility sequence, i.e. $\,n\mid m\Rightarrow a_n\mid a_m\,$ so $\,p\mid a_n\mid a_{nk}\Rightarrow p\mid a_{nk}\,$ for all $\,k\in\Bbb N$


Define $Q:=\prod q_i$. When $Q=1$ the result is trivial.

Otherwise, it is clear that $a_n = {\underbrace{11\dots1_Q}_{n\ 1\text{'s}}}=1+Q+Q^2+\dots+Q^{n-1} = \dfrac {Q^n - 1}{Q-1}$.

The rest essentially follows from changing $10$ to $Q$. Details below:

Choose your $p\notin S$.

For $p\nmid (Q-1)$, as by Fermat $p\mid (Q^{k(p-1)}-1)$ for all $k\in \mathbb N$, $p \mid a_{k(p-1)}$ by coprimality.

For $p \mid (Q-1)$, observe that $$a_n = 1+Q+Q^2+\dots+Q^{n-1}\equiv1+1+1^2+\dots +1^{n-1}\equiv n \pmod Q$$ so $a_{kp}\equiv kp\equiv 0 \pmod p$ for all $k\in \mathbb N$.


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.