# Are any two open connected subsets of $\Bbb{R}^n$ homeomorphic?

Are any two open, connected subsets of $\Bbb{R}^n$ homeomorphic? This seems intuitively true.

-
–  user53153 Feb 25 '13 at 18:04
Think about donuts and disks. –  Neal Feb 25 '13 at 18:53

1 $\neq$ 0 $\neq$ 8

(Think of these as being open subsets of the plane. The set consists of the area that is colored black. Remove the infinitely thin boundary)

-
I am not sure how to think of these as open subsets of the plane? Certainly not 1... –  Zev Chonoles Feb 25 '13 at 17:58
This is a bit mysterious to me. I see $0$, but $1$ and $8$... –  julien Feb 25 '13 at 17:59
I think the idea is that you should draw the symbols in the plane, and then take small open neighbourhoods of them. –  Matt Pressland Feb 25 '13 at 17:59
The set consists of the area that is colored black. Remove the infinitely thin boundary. –  Rasmus Feb 25 '13 at 18:00
I guess the point would come across better in bold font: 1, 0, 8. –  user53153 Feb 25 '13 at 18:06

Nope ! For example $\rm B(0, 2)$ is not homeomorphic to $\rm B(0,2) \backslash \rm \overline{B(0,1)}$.

-
If the statement was refined to say: any two open, simply-connected subsets of $\mathbb{R}^n$, would the statement then be true? –  user54147 Feb 25 '13 at 18:00
@Lev: It certainly would not be true for $n>2$ because these spaces are simply-connected in that case. –  Zev Chonoles Feb 25 '13 at 18:01
@LevLivnev: in higher dimensions one should then replace simply connected with "n-connected." –  Rasmus Feb 25 '13 at 18:40
@LevLivnev the idea holds when you define higher dimensional analogues to "simply connected". The idea behind these examples is the fact that the fundamental groups of these spaces are non-isomorphic, the fundamental groups $\pi_n$ detect $n$-dimensional holes. I have no idea to justify this without using these concepts. Has anyone other explanation? –  Vinicius M. Feb 25 '13 at 18:45
@ViniciusM. There are contractible open manifolds that are not homeomorphic to Euclidean space, for example: en.wikipedia.org/wiki/Whitehead_manifold Does this answer your question? –  MartianInvader Feb 25 '13 at 23:30
show 1 more comment

When $n>1$, an open annulus $$\{x\in\mathbb{R}^n\mid a<|x|<b\},\quad 0<a<b$$ and an open ball $$\{x\in\mathbb{R}^n\mid |x|<r\},\quad 0<r$$ are open connected subsets of $\mathbb{R}^n$ that are not homeomorphic.

When $n=1$ or $0$, it is true that any open connected subsets of $\mathbb{R}^n$ are homeomorphic. This is because any open connected subset of $\mathbb{R}$ is an open interval, and $\mathbb{R}^0$ is just a point.

-