# Is it possible $n(n+1)(n+2)…(n+k)$ is a square?

Let $n,k$ two integers greater than $1$, is it possible that $n(n+1)(n+2)...(n+k)$ is a square $m^2$, with $m$ an integer ?

Thanks in advance.

-
Isn't this reducible via completing the square to Pell's equation in the case where $k=1$? –  Michael Hardy Sep 21 '11 at 18:11
@Michael $n(n+1)$ is never a square for $n > 0$. Because $n$ and $n+1$ are relatively prime, the only way for $n(n+1)$ is a square is to that each of $n$ and $n+1$ is individually a perfect square. It is easy to see that this is impossible unless $n=0$. –  Srivatsan Sep 21 '11 at 18:14
(I'll record another argument for $k=1$. Hopefully someone can see a proof for general $k$. :)) The number $n(n+1)$ is strictly between $n^2$ and $(n+1)^2$, and hence is not a perfect square. –  Srivatsan Sep 21 '11 at 18:20
I see. According to Wikipedia: <b>Pell's equation<b> is any Diophantine equation of the form $$x^2-ny^2=1$$ where $n$ is a nonsquare integer. I missed the "nonsquare" part. –  Michael Hardy Sep 21 '11 at 19:45
cant you just use bertrand postulate an the fact that the largest prime less than n+k divides the product? –  Robert William Hanks Sep 22 '11 at 11:45

## 1 Answer

The answer is no, it can never be a square. This problem was originally solved by Erdos in 1939. The paper can be found here.

Later, in 1975 Erdos and Selfridge improved the result and solved a longstanding conjecture which was first considered by Liouville in the 19th century, by showing that the product of two or more consecutive positive integers is never a perfect power.

-
Thanks for your answer, Eric. –  francis-jamet Sep 21 '11 at 18:33
Here's the link to Erdos 1975 paper: renyi.hu/~p_erdos/1975-46.pdf Some other papers from renyi.hu/~p_erdos/Erdos.html might be of interest too. –  Martin Sleziak Sep 21 '11 at 18:37
@Jason Can you explain your final comment? What convinced you that the problem is difficult for perfect squares too? –  Srivatsan Sep 21 '11 at 18:49
@JasonDeVito: Correction!! I was completely wrong, you had a very good point. The $l=2$ case is much, much easier and was solved 36 years early in 1939 by Erdos! –  Eric Naslund Sep 21 '11 at 18:50
Thanks for the links, Martin and Eric. –  francis-jamet Sep 21 '11 at 19:10