F. Warner exercise 5.8 (foundations of Differentiable Manifolds and Lie groups)

48 Views Asked by At

Context

In section 5.6 of Warner's Book the relationship between sheaves (the author calls sheaves to what is usually defined as Étale spaces) and presheaves is discussed. Given a sheaf S, we define its presheaf of sections $\alpha(S)$. And given a presheaf $F$ we define its associated sheaf $\beta(F)$ as the Étale space consisting of all the germs of sections. The objective is to prove that $S$ and $\beta(\alpha(S))$ are canonically isomorphic (as Étale spaces) and the book itself gives the isomorphism: Let $\xi\in \beta(\alpha(S))$ be the germ at $x$ of some section $f$ of $S$ over an open set containing $x$, that is $\xi=f_x$, for $\Gamma(S,U)$. Then $\psi: \xi \mapsto f(x)$ is the candidate to isomorphism. The exercise 5.8 consists on proving the details.

Question

I have proved that the proposed map is well defined, a local homeomorphism and surjective. All that is left to prove that $\psi$ is injective. I have seen that if $\psi(f_x)=\psi(g_y)$ then $x=y$. But I don't know how to deduce that if two sections $f\in \Gamma(S,U)$, $g \in \Gamma(S,U')$ have the same value at $x$ then both must agree in a neighborhood of $x$. I suspect that this would be false in general if we were not talking about sections with values on an Étale, (which has discrete topology in each fiber) but I don't know how to finish the argument.