Examining the Tartaglia's triangle, I have observed that all the squares were the trivial cases, that is, $\binom{n^2}1$ or $\binom{n^2}{n^2-1}$.

More formally:

Conjecture: If $\binom nm=k^2$ then $n=k^2$.

Is it known to be true?

I have tried to use the formula

$$\nu_p\left(m!\right)=\sum_{k=1}^\infty\left\lfloor \frac m{p^\alpha}\right\rfloor$$

to prove that the exponents of the factorization of the binomial coefficients are odd, but I realized that this cannot be proved, because the binomial coefficients needn't be square-free: $\binom 63=20$, for example.

Any ideas?

  • 1
    $\begingroup$ For $C(n, 2)= m^2$, you get a Pell equation with infinite solutions. $\endgroup$ – Macavity May 2 '14 at 12:14

It is well known that the equation $$ \binom{n}{k}=m^{\ell} $$ has no integer solutions with $ℓ ≥ 2$ and $4 ≤ k ≤ n − 4$. For $k=3$ and $\ell=2$ we only have the solution $\binom{50}{3}=140^2$, and for $k=\ell=2$ there are more. This is due to Erdős. For details see "Binomial coefficients are (almost) never powers" in the book of proofs by Aigner and Ziegler.

  • 3
    $\begingroup$ Since you were keen to do it right, I inserted the ő for you ;) $\endgroup$ – Hagen von Eitzen May 2 '14 at 12:09
  • $\begingroup$ @HagenvonEitzen: thank you ! $\endgroup$ – Dietrich Burde May 2 '14 at 12:11
  • $\begingroup$ The $k=l=2$ case comes down to Pell's equation as you must have $\frac 12n(n-1)$ a square. An example is ${9 \choose 2}=36$ $\endgroup$ – Ross Millikan May 2 '14 at 13:41

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.