Do you think the following limits are correct?

$\displaystyle\lim_{d\to\infty}\frac{\sum\limits_{k=1}^{d} {\phi(N) \choose k} {d-1 \choose k-1}}{\phi(N)^d}=0$

$\displaystyle\lim_{N\to\infty}\frac{\sum\limits_{k=1}^{d} {\phi(N) \choose k} {d-1 \choose k-1}}{\phi(N)^d}=c$

I plotted the equations and guessed the results according to the graphs but I could not prove them mathematically by myself. Any hints would be appreciated. Graphs are as follows:




  • $\begingroup$ I don't understand your first graph at all. For fixed $N$ and $d$, the resulting expression is a rational number with denominator $\phi(N)$. So, for example, when $N = 7$ it's a rational number with denominator 6. Where are you getting values like $10^{-10}$ from? $\endgroup$
    – JBL
    Commented Oct 14, 2010 at 18:07
  • $\begingroup$ I am sorry but there was a typo, i edited now, the denominator should be $\phi(N)^d$. Sorry again! $\endgroup$
    – Nick
    Commented Oct 14, 2010 at 18:10
  • 3
    $\begingroup$ Also posted in math.stackexchange.com/questions/6803/… $\endgroup$ Commented Oct 14, 2010 at 19:22

1 Answer 1


We have $$\sum_{k = 1}^d \binom{\phi(N)}{k} \binom{d - 1}{k - 1} = \binom{d + \phi(N) - 1}{d}$$ (this is the Vandermonde identity). Thus, with $N$ fixed the numerator of your fraction is polynomial in $d$ and so the first result follows (except for $N = 1, 2$).

Edited to add: Okay, and the second result follows by the same analysis, since $\phi(N) \to \infty$ as $N \to \infty$. In particular, the resulting constant is $\frac{1}{d!}$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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