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

This was an exercise using the first Chebyshev function, $\vartheta(x)= \sum_{p \leq x} \log p.$ The question is simply how to prove (2) below, the rest is my two thoughts on how to proceed. [Edit: trimmed after answer.]

$p_k$ is a number with k prime factors including repetitions. It is not the kth prime. $\pi_k(x)$ is the number of $p_k \leq x.$ "log" is the natural log but easier to read than "$\ln$" I think.

We know that $\sum_{p \leq x} \log p \sim x $ [Hardy and Wright p. 451]. We also know that $$\sum_{n\leq x} \log n \approx \int_2^x \log t\hspace{2mm}dt \approx x \log x - x.$$

So (omitting the indices),

$$\frac{\sum \log p}{\sum \log n} \approx \frac{\sum \log p}{x\log x - x}\sim \frac{x}{x\log x - x }\sim \frac{1}{\log x } \sim \frac{\pi(x)}{x}.$$

So we could ask whether

$$(1).....\hspace{5mm}\frac{\sum \log p_k}{\sum \log n}\hspace{2mm} ?\sim \frac{\pi_k(x)}{ x}.$$

For example: for x = 10000, $\pi_5 = 963,$ so $\frac{\pi_5(x) }{x} = 0.0963$ and $\frac{\sum \log p_5}{\sum \log 10^4} \approx 0.0974$

So from (1) via the prior line we would get $$(2)...\hspace{4mm}\frac{\pi_k(x)}{\sum \log p_k} ? \sim \frac{1}{\log x}.$$

Since $ \sum_k \pi_k(x) = x,$ we would have

$$(3)...\hspace{4mm}\sum_k \sum_{p_{k}(m) \leq x} \log p_{k}(m) \sim x \log x. $$

In case it's not clear: for each k we sum over $m = 1,2,3,... $ with $ p_k(m) \leq x$ and then we sum over k such that $p_k \leq x$ for some m.

So for $$\pi_2(x) \approx \frac{\sum \log p_2}{\log x}$$ I thought to compare it to the generalized PNT... [Edit: example omitted]

Or from (2) we could get that $\frac{\pi_k(x) = m}{\sum_{n=1}^m \log p_k(n)}\sim \frac{1}{\log x},$ so

$$ \frac{(1/m) m }{(1/m) \log \prod p_k(m)}\sim \frac{1}{\log x} $$ and

$$ \sqrt[m]{p_k(1)p_k(2)...p_k(m)} \sim x \hspace{4mm}...?$$

Suggestions/corrections welcome.

share|cite|improve this question
up vote 2 down vote accepted

We have that $$\pi_k(x)\sim\sum_{p_k\leq x}1 \sim \frac{x (\log \log x)^{k-1}}{(k-1)!\log x}\ \ \ \ \ \ \ \ \ \ (A)$$ and $$\vartheta_k(x)\sim\sum_{p_k\leq x}\log (p_k) \sim \frac{x (\log \log x)^{k-1}}{(k-1)!},\ \ \ \ \ \ \ \ \ \ (B)$$ so I can reinterpret your question as how does $(B)$ follow from $(A)$, as that implies your equation $2$. Equivalently, you are asking, if I know $(A)$, can I prove that $$\theta_k(x)\sim \pi_k(x)\log x.$$ The short answer is that this follows directly from partial summation. First, lets write the sum as a Riemann Stieltjes integral $$\sum_{p_k\leq x} \log p_k=\int_1^x \log t d\left(\pi_k(t)\right),$$ and then using integration by parts, this equals $$\pi_k(x)\log x-\int_{1}^x\frac{\pi_k(t)}{t}dt.$$ Now, it is possible to use the asymptotic $(A)$ to bound the second integral from above, and we have reduced your question to an elementary real analysis problem.

You may also want to take a look at the paper: A simple proof of a theorem of Landau by E.M. Wright. This paper gives an elementary proof of $(A)$ using the hyperbola method.

Additionally, here is a note related problem which I wrote at some point. It may interest you.

share|cite|improve this answer

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.