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 am trying to prove the following statement.

Consider $C_{i} = \{(x,y):x^2+y^2=i^2\},i=1,2$ two circle in $\mathbb{R}^2$ plane, and define local bases as following:

  • $x\in C_{2},\quad \mathcal{B}_x = \{\{x\}\}$, isolated
  • $x\in C_{1}$, say $x=e^{it}$, let $$B(x,n) = \{e^{is}:|s-t|<2^{-n}\}\cup\{2e^{is}:0<|s-t|<2^{-n}\}$$ e.g. two archs on two circles, and $\mathcal{B}_x=\{B(x,n):n\in\mathbb{N}\}.$

Claim: the space is compact and Hausdorff.

Hausdorff is relatively easy to prove, the only difficulty is to prove the case a point $a=e^{ip}\in C_{1}$ and $b=2e^{ip}\in C_{2}$ , how can I construct such two disjoint open sets containing one of each?

To prove compactness, I think it should be prove by contradiction. Or can it be proved by the theorem:

if $f:X\rightarrow Y$ is a continuous and surjective map, where $X$ is compact, then $Y$ is compact too.

Though I have troubles to construct such a map.

Thanks you in advance.

share|cite|improve this question
This looks like the Alexandrov double circle (from Engelking's book), but the description seems off. Are you sure you have the definition of $\mathcal{B}_x = B(x,n)$ right? Don't you want to remove the point $2e^{it}$ from it? Otherwise I don't think $e^{it}$ and $2e^{it}$ have disjoint neighborhoods. To prove compactness, use Alexander's subbase lemma and observe that $C_1$ carries its usual (compact) topology from $\mathbb{R}^2$. – t.b. Apr 29 '12 at 23:15
@t.b. Thanks a lot. I have overlooked the definition. Now it makes sense now. – newbie Apr 29 '12 at 23:26
Accidentally typed my comment in the answer box again :S Can you please change it to read "and that $C_1$ is compact". The whole space is obviously noncompact, what with all the isolated points in $C_2$. The posted excerpt makes this clear too. – rschwieb Apr 30 '12 at 12:40
@rschwieb: you're wrong, the space $C_1 \cup C_2$ is compact. The reason you mention is no reason for compactness to fail. – t.b. Apr 30 '12 at 12:43
@t.b. Hm, I must not be understanding the topology. It just seems nonsensical that $C_2$ can be discrete and the whole space can be compact. Since the excerpt made no claim that the whole thing is compact, I was tempted to believe it wasn't. Is the set $C_1$ together with the sets $\{z\}$ such that $z\in C_2$ not an open cover of the space? – rschwieb Apr 30 '12 at 12:44
up vote 1 down vote accepted

After you fixed the definition of $B(x,n)$, the Hausdorff property should be easy to check. Distinguish four cases: $x,y \in C_1$, $x \in C_1$ and $y \in C_2$ do [or do not] lie on the same ray emanating from the origin and $x,y \in C_2$.

Observe that the circle $C_1 \subset X$ carries its usual topology inherited from $\mathbb{R}^2$.

To check compactness of $X = C_1 \cup C_2$ I would use Alexander's subbase lemma (see also here for a proof), which tells us that it suffices to consider covers $\{U_{i}\}_{i \in I}$ consisting of sets from the subbase $\bigcup_{x \in X} \mathcal{B}_{x}$ of the topology of $X$.

Since $C_1$ has its usual compact topology, we can find $x_{1},\ldots,x_{n} \in C_1$ and $U_{i_k} = B(x_{k},n_k)$ such that $C_1 \subset U_{i_1} \cup \cdots \cup U_{i_n}$. But then all of $C_1 \cup C_2$ is covered, except possibly the points $2x_{k}$. Since we started with a cover of $C_{1} \cup C_{2}$, we can find $V_{k}$ containing $2 x_{k}$ from the cover $\{U_{i}\}_{i \in I}$ we started with. Then $\{U_{i_1},\dots,U_{i_n},V_{1},\dots,V_{n}\}$ covers $C_1 \cup C_2$ and is a subcover of $\{U_{i}\}_{i \in I}$. Thus we conclude from Alexander's lemma that $X$ is compact.

As I mentioned in a comment, this space $X$ is often called Alexandrov's double circle and appears e.g. as example 3.1.26 in Engelking's General topology (here are two screen shots containing the relevant excerpt: part 1 and part 2). It was first published on pages 13–15 of Alexandroff and Urysohn's Mémoire sur les espaces topologique compacts, Verh. Nederl. Akad. Wetensch. Afd. Naturk. Sect. I 14 (1929),1-96.

Its most prominent features are that it is a first countable compact Hausdorff space that contains a discrete subset of cardinality continuum (the circle $C_2$). Thus $X$ can't be second countable and it isn't separable either. Moreover, $C_2$ is open but not a countable union of closed sets, so $X$ is normal but not perfectly normal.

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.