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

I'm trying to understand the proof of PNT by Don Zagier. But his proof is too simplified so I can't understand it. I got stumped at step II: $\zeta(s)-1/(s-1)$ extends holomorphically to $\Re(s)>0$.

The proof is available here:

Can anyone post a detailed proof that $\zeta(s)-1/(s-1)$ extends holomorphically to $\Re(s)>0$?

Or please tell me which book has a detailed proof so I can look it up?

Zagier's proof:
$\displaystyle \zeta(s)-\frac{1}{s-1} = \sum_{n=1}^{\infty} \int_{n}^{n+1} (\frac{1}{n^s}-\frac{1}{x^s}) dx$
$\displaystyle \left|\int_{n}^{n+1} (\frac{1}{n^s}-\frac{1}{x^s}) dx\right| \leq \frac{|s|}{(n^{\Re(s)+1})}$
therefore $\displaystyle \zeta(s)-\frac{1}{s-1}$ extends holomorphically to $\Re(s)>0$

My question: We have a function
(i) $f(s)=\sum_{1}^{\infty} g_n(s) \quad \forall \Re(s)>0$
(ii) $g_n(s) \leq |s|/(n^{\Re(s)+1}) \quad \forall \Re(s)>0$

Is (i) and (ii) the necessary and sufficient condition that $f(s)$ has holomorphic continuation?
Does there exist functions $f$ and $g_n$ that satisfy (i) and (ii), but does not have holomorphic continuation?
Do we need to prove other conditions, for example $g_n$ must be continuous/holomorphic?

share|cite|improve this question
What part exactly you don't understand in that proof? It looks to me pretty straightforward... – DonAntonio Aug 25 '12 at 18:35
@alek: You might notice that I merged the account you used to write the post with the account you wrote the answer/edit. This will let you comment and edit your own question. Best of luck. – mixedmath Aug 26 '12 at 4:45

Start with the defining series, convergent for $\Re(s)>1$: $$ \sum_{n=1}^\infty \frac{1}{n^s} = \sum_{n=1}^\infty \frac{1}{\Gamma(s)} \int_0^\infty t^{s-1} \mathrm{e}^{-n t} \mathrm{d} t = \frac{1}{\Gamma(s)} \int_0^\infty t^{s-1} \frac{\mathrm{e}^{-t}}{1- \mathrm{e}^{-t}} \mathrm{d} t $$ Also notice that, for $\Re(s)>1$, $$ \frac{1}{s-1} = \frac{\Gamma(s-1)}{\Gamma(s)} = \frac{1}{\Gamma(s)} \int_0^\infty t^{s-2} \mathrm{e}^{-t} \mathrm{d} t $$ Now subtracting we have: $$ \zeta(s) - \frac{1}{s-1} = \frac{1}{\Gamma(s)} \int_0^\infty t^{s-1} \mathrm{e}^{-t} \left( \frac{1}{1- \mathrm{e}^{-t}} - \frac{1}{t} \right) \mathrm{d} t $$ The integral on the right-hand side converges now for $\Re(s)>0$, since at the lower integration bound $\frac{1}{1-\exp(-t)} - \frac{1}{t} = \frac{1}{2} + \mathcal{o}(1)$.

share|cite|improve this answer

Trying to address what I think was troubling the OP.

Fix constants $a,b$ such that $0<a<b$. Consider the set $$L(a,b)=\{x+yi\in\mathbb{C}\mid a<x, |x+yi|<b\}.$$ The functions $g_n(z)$ are holomorphic in the set $L(a,b)$. Item $(ii)$ means that for all $z\in L(a,b)$ and all positive integers $n$ we have $$ |g_n(z)|\le\frac{b}{n^{1+a}}. $$ Therefore, by Weierstrass' M-test (note that here the lower bound on the exponent, $1+a$, does not depend on $z$, only on the set $L(a,b)$), the series $f(z)=\sum_{n=1}^\infty g_n(z)$ converges uniformly in the set $L(a,b)$ and therefore gives a holomorphic function on $L(a,b)$.

The right half-plane is the union of the sets $L(a,b)$, so $f(z)$ is holomorphic there.

share|cite|improve this answer
In other words, use the techniques familiar from proving that a converging power series gives a holomorphic function: Shrink the set in a suitable way such that your estimates prove the convergence to be uniform in that smaller set. – Jyrki Lahtonen Aug 26 '12 at 9:50

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.