I have a question about some hypothesis in a classical theorem about coverings.

The theorem is the following:

Let $B$ be a connected base space, and $X$ and $Y$ be coverings of $B$ and $g, f: X\to Y$ two $B$-morphisms. If there exists $b\in B$ such that $f_{|X_b} = g_{|X_b}$ then $f=g$.

This is easy to prove is $X$ is connected. If you don't assume $X$ connected, then it is also easy to prove if you assume $B$ locally connected, as in that case every connected component of $X$ is open and closed, and therefore maps surjectively onto $B$ and thus, every connected component of $X$ contains a point of the fiber $X_b$.

This is of course a standard assumption.

But the reference I'm looking at (Douady, Algèbres et théorie galoisiennes) does not make any assumption (other than $B$ connected) and simply states that "every connected component of $X$ contains a point in the fiber $X_b$".

And I fail to see how to prove that whithout the local connectedness assumption, am I missing something?


1 Answer 1


This is only a partial answer: It is true if $B$ is path connected.

Let $P$ be a path component of $X$. Note that in general the components of $X$ do not coincide with its path components; but if we prove that $P \cap X_b \ne \emptyset$ for all $b$, then also $C \cap X_b \ne \emptyset$ for all components $C$ and all $b$. Pick $x_0 \in P$ and let $b_0 = p(x_0)$, where $p : X \to B$ denotes our covering projection. Then $x_0 \in C \cap X_{b_0}$. Now let $b \in B$. Choose a path $u$ in $B$ such that $u(0) = b_0$ and $u(1) = b$. Lift $u$ to a path $\tilde u$ in $X$ such that $\tilde u(0) = x_0$. Since $A =\tilde u([0,1])$ is path connected, we get $A \subset P$. But $\tilde u(1) \in A \cap X_b \subset P \cap X_b$.

  • $\begingroup$ Yes, this works too. I'm still wondering if the theorem is false without any additionnal hypothesis on $B$, be it path connected, or locally connected. In any case I feel like the author is going a bit too fast with his claim, but I'm still wondering whether it's true if you only assume $B$ connected, I've failed to find a counter example. $\endgroup$
    – Nissokam
    Feb 2, 2021 at 10:45
  • $\begingroup$ @Nissokam I doubt that it is true without additional assumptions. But it will not be easy to find counterexamples. $\endgroup$
    – Paul Frost
    Feb 2, 2021 at 10:58

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.