Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Show that $$(p-1)!\left(1+\dfrac{1}{2}+\dfrac{1}{3}+\cdots+\dfrac{1}{p-1}\right)\equiv 0\pmod{p^2}.$$

Maybe use this $$\dfrac{1}{k}+\dfrac{1}{p-k}=\dfrac{p}{k(p-k)}$$ and then I can't. Can you help me to prove it?

Thank you.

share|cite|improve this question
Adapting Wilson's theorem mod $p^2$? – PITTALUGA Nov 6 '13 at 12:09
I think the general expression is not an integer; at least for p=3,5. – user99680 Nov 6 '13 at 12:15
It is, since every denominator appears as a factor of $(p-1)!$. – Christoph Nov 6 '13 at 12:16
More is true: – lhf Nov 6 '13 at 12:17
Ah, yes, I was adding incorrectly. – user99680 Nov 6 '13 at 12:18
up vote 5 down vote accepted

The solution below is adapted from Notes on Wolstenholme’s Theorem by Timothy H. Choi.

Let $$ S=(p-1)!\sum_{k=1}^{p-1} \frac1k $$ Using your insight $$ \dfrac{1}{k}+\dfrac{1}{p-k}=\dfrac{p}{k(p-k)} $$ we have $$ 2S=(p-1)!\sum_{k=1}^{p-1} \left(\dfrac{1}{k}+\dfrac{1}{p-k}\right) = p\sum_{k=1}^{p-1} \frac{(p-1)!}{k(p-k)} = pS' $$ Note that $S'$ is an integer. Now $$ \frac{(p-1)!}{k(p-k)} \equiv (k^2)^{-1} \bmod p $$ where the inverse is taken ${}\bmod p$. This is a consequence of Wilson’s Theorem. Hence $$ S'\equiv \sum_{k=1}^{p-1} (k^2)^{-1} \equiv \sum_{k=1}^{p-1} k^2 = \frac{(p-1)p(2(p-1)+1)}{6} \equiv 0 \bmod p $$ This means that $2S\equiv 0 \bmod p^2$ and so $S\equiv 0 \bmod p^2$. (We need $p>3$ twice here.)

share|cite|improve this answer

As others have noted, the congruence is not true for $p=3$, since $$ 2!\left(1+\frac 1 2\right)=2+1=3,$$ which is not divisible by $9$. We can still use what you suggested to prove the congruence holds $\operatorname{mod} p$. Let $p$ be an odd prime, then \begin{align*} (p-1)!\sum_{k=1}^{p-1} \frac 1 k &= (p-1)! \sum_{k=1}^{(p-1)/2} \left(\frac 1 k + \frac 1 {p-k}\right) \\&= (p-1)!\sum_{k=1}^{(p-1)/2} \frac{p}{k(p-k)} = p\sum_{k=1}^{(p-1)/2} \frac{(p-1)!}{k(p-k)}. \end{align*} Since $(p-1)!$ always contains $k$ and $(p-k)$ as a factor, the fractions in the sum are integers and the result is a multiple of $p$.

See lhf's answer for why it is even a multiple of $p^2$ as long as $p>3$.

share|cite|improve this answer
The congruence will hold for any $p>3$ modulo $p^2$. That's Wolstenholme's theorem as lhf pointed out. – EuYu Nov 6 '13 at 12:39
I adapted my post and referencered lhf's answer, thanks for your comment! – Christoph Nov 6 '13 at 12:54

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.