MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

Fix a ground scheme $S$ (a field say). By atlas for an algebraic stack I mean a smooth and surjective morphism $Y \to X$ from a scheme (or algebraic space or affine scheme) $Y$. If the stack $X$ is smooth then all atlases are smooth schemes over $S$.

What about when $X$ is smooth and proper? Can I find a proper atlas?

If in general the answer is no, what about for Deligne-Mumford stacks?

share|cite|improve this question
It's worth nothing that this is true in the differentiable category; the result is in the thesis of Dorette Pronk. – David Carchedi Feb 3 '13 at 22:10
thanks David, it's good to know. – Jacob Bell Feb 4 '13 at 13:30
up vote 9 down vote accepted

The answer is "no". Consider a smooth, proper $1$-dimensional Deligne-Mumford stack over $\mathbb{C}$ which has coarse moduli space $\mathbb{P}^1$ and which has a single "stacky" point (with any nontrivial stabilizer group at that point).

Edit. A more precise definition of the stack is as follows. Let $m$ be an integer, $m>1$. Let $A$ be $\mathbb{A}^2 \setminus \{(0,0)\}$ with coordinates $x$ and $y$. Let $\mathbb{G}_m$ act on $A$ by $t\ast (x,y) = (tx,t^m y)$. The stack $Q_m$ is the quotient stack $[A/\mathbb{G}_m]$. The quotient morphism $A\to Q_m$ is a smooth atlas, implying that $Q_m$ is smooth (since $A$ is smooth). In fact, this morphism is naturally a $\mathbb{G}_m$-torsor; denote the associated invertible sheaf on $Q_m$ by $\mathcal{L}$. It is not hard to see that this is an $m$-torsion invertible sheaf; the $\text{m}^{\text{th}}$ tensor power of $A$ as a $\mathbb{G}_m$-torsor over $Q_m$ is the quotient stack $[(\mathbb{G}_m\times A)/\mathbb{G}_m]$ where the $\mathbb{G}_m$-action is $t\cdot (u,(x,y)) = (t^mu, (tx,t^my))$. This admits a "trivializing" morphism to $\mathbb{G}_m$, $(u,(x,y)) \mapsto u/x^m = u/y$, each defined on the appropriate open $\mathbb{G}_m\times D(X)$ or $\mathbb{G}_m\times D(y)$. The only nontrivial stabilizer group is $\mathbf{\mu}_m$ acting on the $\mathbb{G}_m$-orbit $\{0\}\times \mathbb{G}_m$ inside $A$. So $Q_m$ is a Deligne-Mumford stack. Finally, $Q_m$ is finite over its coarse moduli space $\mathbb{P}^1$, where the $\mathbb{G}_m$-invariant morphism $A\to \mathbb{P}^1$ is just $(x,y) \mapsto [x^m,y]$.

share|cite|improve this answer
I see. Could I ask you how one would go about building such a stack? – Jacob Bell Jul 31 '12 at 12:22
You use the so called "root construction", which is (roughly) a way of creating stacky structure in codimension one. See Charles Cadman, "Using stacks to impose tangency conditions on curves". – Dan Petersen Jul 31 '12 at 12:37
thanks, I'll check it out. – Jacob Bell Jul 31 '12 at 12:42
This example is old. It is trivial to see that there is no finite \'etale atlas which is an algebraic space. It takes a little more work to see there is no proper, smooth atlas which is an algebraic space. – Jason Starr Jul 31 '12 at 13:25
thanks a lot, it will take me a little bit to digest what you wrote (I fear our definitions or trivial may differ). – Jacob Bell Jul 31 '12 at 13:49

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.