We know that a set may be of zero harmonic measure without its Lebesgue measure being zero (see Armitage and Gardiner, classical potential theory, pg 178).

Now, consider the following problem. Let $ V $ be a bounded open set in $ \mathbb{R}^{m} $, $ m\geq 2 $, and $W$ the interior of the closure of $V$. Let $E$ be a subset of $ \partial V\cap W $ ($\partial$ means boundary). We suppose $E$ has the following property: for every $x\in E$ and every $r>0$ such that the ball $B(x,r)$ is relatively compact in $W$, the boundary $\partial B(x,r)$ of $B(x,r)$ contains at least one point of $E$. My question is: can the harmonic measure $\omega(E,x,V)$ of $E$ be zero?

Intuitively (Brownian motion interpretation of harmonic measure) the answer seems negative. (?)

  • $\begingroup$ Yes. In $R^3$, $V$ is a ball with removed radius, and $E$ is this radius. $\endgroup$ Dec 10, 2019 at 13:29
  • 1
    $\begingroup$ Your assumption does not imply that $E$ is connected. Neither it implies that Lebesgue measure of $E$ is positive. $\endgroup$ Dec 10, 2019 at 14:00
  • $\begingroup$ Yes you are right. I edited the post. Thanks. $\endgroup$
    – M. Rahmat
    Dec 11, 2019 at 4:26

1 Answer 1


It is easy to construct such an example in $R^n$ for $n\geq 3$ (see my comment). In $R^2$ let $V$ be a rectangle with removed segments: $$\{ x+iy:|x|<1,0<y<2\}\backslash \left((0,i]\cup\cup_{n=2}^\infty ((1/n,1/n+i]\cup(-1/n,-1/n+i])\right),$$ and $E=(0,i)$.

Similar example will work in $R^n$ if you want the $n-1$ Lebesgue measure positive. In fact you can even have $n$-dimensional Hausdorff measure positive, if you take something widely curved for $E$ instead of a straight thing.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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