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

The construction of a Mrówka space (a $\Psi$-space) is not clear for me. Because of this I could not see why it is first-countable, locally compact, and Hausdorff.

Coud you give me some help about this space?

share|cite|improve this question
What in particular about the construction is not clear to you? What construction are you following? – Potato Aug 7 '13 at 3:53
I'll add link to the blog about Mrowka space at Dan Ma's Topology Blog. – Martin Sleziak Aug 8 '13 at 8:07
up vote 8 down vote accepted

Let us recap the construction, also to fix notation. We have $\omega$ (the natural numbers) and a collection $\mathcal{A}$ of infinite subsets of $\omega$ that is almost disjoint: for every $A,B \in \mathcal{A}, A \neq B$ we have that $A \cap B$ is finite.

We define a space $X = \omega \cup \{x_A: A \in \mathcal{A} \}$ (so we add to $\omega$ a collection of new points, one for each member of the almost disjoint family).

We define a topology on this by defining a neighbourhood base for each point: each $n \in \omega$ has neighbourhood base $\{ \{n \} \}$; it is an isolated point (like in the usual discrete topology on $\omega$). Every $x_A$ (for a fixed $A \in \mathcal{A}$) has a neighbourhood base consisting of all sets of the form $B(A,F) = \{x_A\} \cup (A \setminus F)$, where $F \subset A$ is finite. We can also define the open sets directly as all sets $O$ such that: $\forall_{A \in \mathcal{A}}: x_A \in O \rightarrow A \setminus O \mbox{ finite.}$

Note that all subsets of $\omega$ are still open (as usual) and that every $B(A,F)$ intersects $X$ only in $\omega$, and does not contain other added points. So the set $X \setminus \omega$ (all extra points $x_A$) forms a closed and discrete subset of $X$. This shows that if $\mathcal{A}$ is uncountable, then $X$ is not countably compact. Also, no $\{x_A\}$ is open by itself, because all $A$ are infinite, so all basic open sets of $x_A$ intersect $\omega$ non-trivially. So $\omega \subset X$ is dense (so $X$ is separable).

To see that $X$ is Hausdorff, we need to separate a pair of distinct points $x \neq y$ from $X$. If $x,y \in \omega$ the sets $\{x\}$ and $\{y\}$ will do, of course and if, say $x \in \omega$ and $y = x_A$ for some $A \in \mathcal{A}$, then $\{x\}$ and $B(A,\{x\})$ are open, pairwise disjoint neighbourhoods of $x$ and $y$ resp. As the final case, if $x = x_A$, $y = x_B$, for $A \neq B, A,B \in \mathcal{A}$, we know that $F = A \cap B$ is finite and then $x \in B(A,F)$, $y \in B(B, F)$ and these sets are clearly disjoint (we removed the possible intersection $F$). So here we (finally) use that $\mathcal{A}$ is almost disjoint.

That $X$ is first countable: trivial at isolated points, and there are only countably many finite subsets of $\omega$, so the other local bases as described above, are all countable already.

And $X$ is locally compact, because every set $A \cup \{x_A \}$ ($=B(A,\emptyset)$), and even every set $B(A,F)$, is homeomorphic to a convergent sequence (enumerate $A$ as $\{x_n: n \in \omega \}$, then $x_n \rightarrow x_A$ as $n \rightarrow \infty$: every neighbourhood of the limit contains a tail of the sequence. And convergent sequences are compact spaces. Also, this shows that $X$ is zero-dimensional (as all basic elements are clopen being compact in a Hausdorff space) and thus completely regular (Tychonoff, or $T_{3\frac{1}{2}}$).

All spaces of this form are often called $\Psi$-like spaces (or Mrówka spaces). If we choose $\mathcal{A}$ to be a maximal (by inclusion) almost disjoint family of subsets (which are always uncountable and can be as large as the continuum, then $X$ can be shown to be pseudocompact (all real-valued continuous functions defined on it are bounded), and as $T_4$ pseudocompact spaces are countably compact, we conclude that for maximal almost disjoint familes $\mathcal{A}$ this $X$ is then a completely regular, non-normal Hausdorff space that shows that pseudocompactness does not necessarily imply countable compactness for non-$T_4$ spaces.

share|cite|improve this answer
thanks so much. If possible, could you give example for added new – user89330 Aug 8 '13 at 3:33
thanks so much. If possible, could you give example for added new points $x_A$, because ı could not imagine. And also, the singletion $\{x_A\}$ is open if the compement of $\omega$ discrete and closed, as you said.If each singletion $\{x_A\}$ is open, the set complement of $\omega$ must be open. So, the complement of $\omega$ can be clopen?If every B(A,F) intersects X only in ω, how it can be said each singletion $\{x_A\}$ is open? I think there was a wrong here. why not $x_A \in (B(A,F) \cap X)$? – user89330 Aug 8 '13 at 3:50
There is no "example" of the added new points; we just add one point to $\omega$ for each $A \in \mathcal{A}$ that wasn't there before. In a way, we could have the subset $A$ itself be the new point, but that would get a bit confusing, IMHO, so I use indices. – Henno Brandsma Aug 8 '13 at 6:46
No, I edited my post to show that $\omega$ is not closed (it is discrete, of course), but dense in $X$. Of course by definition $B(A,F) \cap X = B(A,F)$ and it contains $x_A$, but $B(A,F) \cap \omega \subset A \subset \omega$ and the intersection is infinite, – Henno Brandsma Aug 8 '13 at 6:51
thanks again.It was very useful for me. – user89330 Aug 12 '13 at 8:00

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.