Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

An earlier question links to a paper of Erdos in which he says that it is "well-known" that the Prime Number Theorem is equivalent to

$(\prod_{p\leq n}p)^{1/n} \to e$ as $n\to \infty.$ **

Here is my confusion.

If $~\prod_{p\leq n}p \sim e^n$ or $e^{\log \prod p}= e^{\sum \log p} \sim e^n,$

(the last relation appears in the linked question, but I take responsibility for it) doesn't this imply that

(*) $\lim_{n \to \infty} (\sum_{p\leq n} \log p - n) = 0?$

Of course it's true that $\lim_{n \to \infty} \frac{\sum \log p}{n} =1 $ and I do not think that (*) is true. But I think we do have in general that

$$ e^{f(x)}\sim e^{g(x)} \implies \lim (f(x) - g(x)) = 0,$$ since $\lim \frac{e^f}{e^g}= e^{f-g} = 1 \implies \lim (f-g) = 0.$

Can someone tell me where I have goofed? Thanks!

**If someone could point me to a proof of this I would appreciate it --I don't see it in Apostol or Hardy & Wright).

share|improve this question
add comment

2 Answers 2

up vote 1 down vote accepted

$(\Pi_{ p \leq n} p)^{1/n} \rightarrow e$, as $n \rightarrow \infty$, doesn't imply $\Pi_{p \leq n} p \sim e^n$.

Example: $(2^n n)^{1/n} \rightarrow 2$, as $n\rightarrow \infty$, but $2^n n$ is not equivalent to $2^n$.

Indeed, $n^{1/n}=e^{\frac{\log n}{n}} \rightarrow 1$ because $\frac{\log n}{n} \rightarrow 0$ as $n \rightarrow \infty$

share|improve this answer
I am pretty sure this is right and your counterexample is good. Then in the question I linked to, the OP seems to make this error too? (**) is from Erdos, but the following line appears in the other question. –  daniel Oct 4 '12 at 14:47
I don't know if (*) is true or not. –  francis-jamet Oct 4 '12 at 15:03
add comment

The fact that it's equivalent is rather simple.

Recall $\pi(x)/(x/ln x) \rightarrow 1$, your limit is quivalent to:

$$\sum_{p\leq n} \log p / n \rightarrow 1$$

if you take $x=n \in \mathbb{N}$ then: $$\sum_{p\leq n} \log p /n \leq (\log n) \pi(n) /n \rightarrow 1$$

And to bound it from below, you notice that this fraction is greater than 1.

Not sure how to argue from below, I'll check next time.

share|improve this answer
Thanks for fixing the typo--but I still don't know what question you are answering. The fact the "it's" equivalent...what is "it?" –  daniel Oct 4 '12 at 14:58
add comment

Your Answer


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

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