return to top
source
$$\binom\{n\}\{k\} \cdot k = \binom\{n-1\}\{k-1\} \cdot n$$.
$$\binom\{n\}\{k\} \cdot k \cdot (k-1) = \binom\{n-2\}\{k-2\} \cdot (n-1) \cdot n$$.
If $$p^\{a+b+1\} \mid n$$, then $$p^\{a+1\} \mid \binom\{n\}\{k\}$$ or $$p^\{b+1\} \mid k$$.
If $$p^r \mid n-1$$ and $$k,n-k$$ are not divisible by $$p$$, then $$p^r \mid \binom\{n\}\{k\}$$.
If $$n=2^\{\phi(p^R)\}$$ and $$p>2^\{r-1\}$$, then $$2^r \mid \binom\{n\}\{k\}$$ or $$p^R \mid \binom\{n\}\{k\}$$.
The condition 1<n is necessary to avoid the range of k being vacuous.
1<n
k
If $$p\>2^\{r\-1\}$$, then $$S\(2^\{phi\(p^R\)\}\) ge r$$.
A positive resolution to Erdos problem #379.