MathOverflow is a question and answer site for professional mathematicians. 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

Let $p(z,t)=\frac{1}{2\pi}.\frac{1-|z|^2}{|z-t|^2}$ be the Poisson kernel on the open unit disk $\mathbb{D}$, fix $0<\alpha<1$ . Let $a\in \partial\mathbb{D}=S^1$ be fixed. Then my question is :

what is the limit of $\frac{\int_{S^1}|t-a|^{1 + \alpha}.p(z,t)|dt|}{|z-a|}$ as $z \to a, z\in \mathbb{D}$. I am tending to believe that the limit is zero, because of the following reason :

Take any $0<\alpha' < \alpha$, then $t \mapsto |t-a|^{1 + \alpha'}\in C^{1,\alpha'}(S^1)$.Therfore by Kellog's (or by Kellog-Warschawski's) theorem, its harmonic extension,extended by $H(z)= \int_{S^1}|t-a|^{1 + \alpha'}.p(z,t)|dt|$on $\mathbb{D}$ is $C^{1,\alpha'}(\mathbb{D})$, therefore the harmonic extension is Holder continuous, that is :

$\frac{\int_{S^1}|t-a|^{1 + \alpha'}.p(z,t)|dt|}{|z-a|} \leq M \equiv M(\alpha')$ [Note that, at $a\in \partial \mathbb{D}, H(a)=0$].

But then there should be the 'effect' of this "extra" $\alpha - \alpha'$ in the integration, which, heuristically, should make the limit go to zero. But I am not sure how to prove that ? Is it right at least ? Any help will be highly appreciated, thank you !

share|cite|improve this question
The numerator is a harmonic function $u$ with $C^{1,\alpha}$ boundary data, so we do expect $C^{1,\alpha}$ regularity from the boundary; in particular, if we subtract the linear part of $u$ at $a$, we expect that the quotient tends to $0$. However, in the expression you give we haven't subtracted the gradient plane, which in general has a normal component (as in the nice examples of Alexandre below). Since $u$ and its tangential derivative at $a$ are $0$, the quotient will tend to $0$ if we travel along the boundary but will be nonzero if we approach from the interior. – Connor Mooney Oct 3 '12 at 4:43
up vote 3 down vote accepted

The limit is not zero; it does not exist.

FIRST proof. WLOG let $a=1$. Let $-u(z)$ be your Poisson integral, (in the numerator of your formula) it is a negative harmonic function in the disc, continuous in the closed disc, $u(1)=0$, and negative at every other point of the circle. Let $M(r)=max_{|z|=r}u(z),\; 0\leq r\leq 1$. This is a strictly negative, increasing function on $(0,1)$, and $M(1)=1$. It is known that $M(r)$ is convex with respect to the logarithm, that is $$r\frac{dM(r)}{dr}$$ is increasing. This is called Hadamard's Three Circles Theorem. Thus $M'(1)>0$. Now as $u(re^{i\theta})$ is an even function of $\theta$ decreasing on $(0,\pi)$, we conclude that $M(r)=u(r)$. This means that there exists a sequence $z_k\to 1$,such that $z_k<1$, and
$$\frac{u(1)-u(z_k)}{1-z_k}\; \quad\to c>0.$$ So on this sequence your limit $$\frac{-u(z_k)}{|1-z_k|}\quad =c.$$ On the other hand it is clear that your limit is zero on sequences which tend to $1$ "tangentially$, that is very close to the circle. So the limit does not exist.

SECOND proof. WLOG $a=1$. Let $v$ be your Poisson integral in the numerator. Let $w$ be the Poisson integral of the same function but replaced by $0$ on the right half of the circle. Then $0 < w < v$ in the open disc. But $w(z)=0$ on an arc of a circle near $1$, so by the Symmetry Principle, $\partial w/\partial r \neq 0$ at the point $1$. As $w$ is positive and $w(1)=0$, this derivative is negative. So $v(r)>w(r)>c(1-r)$ for some $c>0$.

share|cite|improve this answer
Thanks Prof. Eremenko, that was helpful. I will soon post a related question regarding this which still confuses me. Please don't feel obligated to reply, but if you do, I will appreciate it. – Analysis Now Oct 23 '12 at 15:03

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.