Let $\mathscr{Z}$ be an arbitrary sheaf on $\mathbb{R}/\mathbb{Z}=X$ (with the quotient topology). Let $\mathscr{F}$ and $\mathscr{G}$ denote the sheaves of continuous functions on $X$ with values in $\mathbb{R}$ and $X$, respectively.
Given a morphism $f\colon \mathscr{G}\to \mathscr{Z}$ such that the composition with the canonical map $j\colon \mathscr{F}\to \mathscr{G}$ is the zero morphism, how can I see that $f$ has to be zero?

It suffices to prove that for each $x\in X$, the map on the stalks $f_x$ is zero. Can we say that for each open $x\in U$ in $X$ and each $g\colon U\to X$, there exists some $x\in V \subset U$ open such that $g\vert V$ factors through the projection $\mathbb{R}\to X$, i.e. there exists some $h\colon V\to \mathbb{R}$ in $\mathscr{F}(V)$ with $j_V(h)=g_V$?


Here is a sketch without many details, just intuition (but it can be made rigourous).

Since $X$ locally looks like $\mathbb R$, the stalk $\mathscr G_x$ can be thought of as functions $\mathbb R \to X$. These are periodic functions. But the stalk $\mathscr F_x$ can be thought of as arbitrary continous functions $\mathbb R \to \mathbb R$.

Hence we have an injection $k:\mathscr G \to \mathscr F$ such that $jk=id_{\mathscr G}$ (because the set of periodic functions inject into the set of all functions).

Hence $f=f \circ id_{\mathscr G}=f \circ j \circ k= 0 \circ k = 0$.


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.