Let $R$ be a Riemann surface of genus $g\ge 2$ and $q$ an holomorphic quadratic differential on $R$. Together they determine a semi-translation structure: an atlas on $X$ such that its changes of charts are of the form $z\mapsto \pm z+c$ and a singular flat metric $|q|$ on $X$.

Suppose that $q$ has at least one singularity of odd order: I can't completely grasp what the $-1$ holonomy implies when considering the lengths of simple closed curves on $(X,q)$. In particular, let $c\subset X$ be a closed simple geodesic for $|q|$ which doesn't meet any singular point of $|q|$. Then, is it true that there is always a cylinder in $(X,q)$ foliated by closed geodesics parallel to $c$ (as in the case of translation surfaces with trivial holonomy)?

If no, then is it still always possible to assume the existence of a concatenation of saddle connections of $q$ parallel to $c$ and with the same length?


2 Answers 2


Yes, this is true, and as far as I can tell the argument doesn't seem any different whether the surface is a pure translation surface or a semi-translation surface. The point is that the argument is purely local, and from this argument you can deduce from the given conditions on $c$ that the rotational holonomy around $c$ is trivial.

Here's the argument. If $c$ is a simple closed geodesic disjoint from the singular points, then by compactness of $c$ there exists $t>0$ such that the "$t \times t$ square" centered on each point of $c$ is disjoint from the singular points. Depending on the slope of $c$ with respect to the horizontal and vertical foliations, there exists $s \in [t,t\sqrt{2}]$ such that the union of these squares forms the metric neighborhood of $c$ of radius $r$. This metric neighborhood is isometric to a Euclidean product annulus, namely the product of $c$ with the interval $[-s,+s]$. It follows from this description that the $\mathbb{Z}/2\mathbb{Z}$-rotational holonomy around $c$ is trivial, since that's the only possibility for Euclidean product annuli.


Yes, it is, of course, true. Even though foliations defined by $q$ are not orientable, the genus $2$ surface is orientable, so an $\varepsilon$-neighborhood of a closed geodesic $\gamma$ is just the cylinder $\gamma\times [-\varepsilon, \varepsilon]$.

This statement holds more generally for orientable surfaces with flat metric and conical singularities. If you find a simple closed geodesic (smooth one and disjoint from conical points) its $\varepsilon$-neighborhood is a cylinder.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.