The title says it all. Let $A$ be a path connected $F_\sigma$ subset of a plane (or more generally $\mathbb{R}^n$). Recall that a subset is called $F_\sigma$ if it is a union of a sequence of closed sets.

Is it true that there is a continuous surjection from $[0,1)$ onto $A$? Equivalently, can $A$ be represented as a union of an increasing sequence of Peano continuums?

Note that we cannot drop "path", since $\{y=\sin(\frac{1}{x}),x>0\}\cup \{(0,0)\}$ is a connected $F_\sigma$ subset that cannot be represented as a union of an increasing sequence of continuums.


No, this fails even for compact subsets of $\mathbb R^2$. Namely, let $X=C\times[0,1]\cup[0,1]\times\{0\}$, where $C$ is the Cantor set. It is clearly path connected. $X$ cannot be an image of $[0,1)$, because the image of any interval $[0,a],a<1$ by this map can contain only finitely many points of $C\times\{1\}$ (because of compactness), and hence the image of $[0,1)$ can only contain countably many of them.

It might be of your interest that there is a complete topological classification of spaces which are images of $[0,1)$, namely they are the path-connected spaces which are countable unions of Hahn-Mazurkiewicz spaces (which means they are compact, Hausdorff, connected, locally connected, metrizable spaces), as shown here.

  • $\begingroup$ I can understand that the image of $[0,a]$ can contain finite number of points of $C\times \{1\}$ due to local connectedness. What do you mean by "because of compactness"? Perhaps I am missing somethig $\endgroup$
    – erz
    Jun 16 at 4:55
  • $\begingroup$ @erz There are probably multiple ways to see it. The argument I had in mind is that for every point $(c,1)$ which the curve crosses, there is an open subset of $[0,a]$ on which the curve is contained in $\{c\}\times(0,1]$. Together with the preimage of the complement of $C\times\{1\}$ these form a cover of $[0,a]$ and we can conclude through compactness from there. $\endgroup$
    – Wojowu
    Jun 16 at 10:06
  • $\begingroup$ Thank you! My point was that I couldn't accept your answer since I didn't fully understand it as stated. May I suggest you include this elaboration into the main text? $\endgroup$
    – erz
    Jun 16 at 11:18

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.