Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

It is well known that if $F\to B$ is a $n$-finite branched covering over an orbifold with cone-points then the orbifold Euler's characteristics are related via $\chi(F)=n(\chi(B)-\sum_i^r\frac{a_i-1}{a_i})$, where $r$ is the number of cones with stabilizer of orders $a_1,...,a_r$ respectively.

Now, I don't know if everyone would like to know what is the corresponding relation when $B$ has reflector intervals or reflector circles, as I'd rather... so I dare to question:

Is there a generalization in this direction?

share|improve this question
Rather than just accepting the "Well-known" identity, why not prove it. Then you can ask, "what happens if I use the same technique in the other situation?" and you should have your answer. –  Ryan Budney Dec 17 '09 at 4:01
Could you explain why you find Greg Kuperberg's answer insufficiently detailed? Does the following observation help? The identity really just says $\chi_o(F)=n\chi_o(B)$, where $\chi_o$ is orbifold Euler characteristic. Now just count vertices, edges and faces, as in the 'usual' proof. –  HJRW Jan 12 at 17:27
there, I found street babbling difficult to follow and retain –  janmarqz Jan 13 at 5:15
@janmarqz, I fear that something has been lost in translation. Calling an honest attempt to answer your question 'street babbling' will not encourage anyone else to help you. –  HJRW Jan 13 at 10:10
for "street babbling" I was referring to the style there on the Greg's link... but now i understand over this two word the meaning that your culture carries... excuse my clumsy english and for being to rude –  janmarqz Jan 13 at 15:59

2 Answers 2

up vote 3 down vote accepted

A detailed write up of what you are looking for is also available in Chapter 13 of Thurston's notes. Specifically, formula 13.3.4 on page 331, has the desired formula $$\chi(O) = \chi(X_O) -\frac{1}{2} \sum_{i=1} ^N (1-\frac{1}{n_i}) -\sum_{j=1}^M(1-\frac{1}{m_j}),$$ where $O$ is the orbifold, $X_O$ is it's underlying space, there are N points fixed locally by dihedral groups of orders $2n_1,...,2n_N$ and $M$ points fixed locally by rotations of orders $m_1,...,m_M$.

However, I believe the question addressed (at least in part) by Hurwitz in the proof of Hurwitz's theorem, since the bound the order of an automorphism group H acting on surface of genus g is 168(g − 1) if H includes orientation reversing symmetries and 84(g-1) in the orientable case. The key observation is that the Gauss-Bonnet extends to orbifolds as well.

Computing the Euler characteristic for the quotient of $\mathbb{H}^2$ by the (full) (2,3,7) triangle group is a good way to see this in action.

share|improve this answer
what about in the presence of reflector arcs and reflector circles? and say: $N=0$ and $M=0$ –  janmarqz Jan 14 at 21:55
where reflector circle is taken à la Scott, in "Geometries of 3-manifolds" –  janmarqz Jan 14 at 23:02
@janmarqz It turns out those don't really contribute unless they are part of a corner point in the language of Thurston. For example, take an annulus $A = S^1\times[-1,1]$. The core of the annulus, $S^1 \times \{0\}$ is the fixed point set of a reflection. Call this quotient $Q$. Note, $\chi(Q)=0$. Also, $X_Q$ is an annulus and so $\chi(X_Q)=0$. Thus, the orbifold locus, a reflector circle, does not contribute to the euler characteristic directly (however it does affect the topology of $X_Q$, and so it implicitly affects $\chi(X_Q)$). There are perhaps other ways to make this observation. –  Neil Hoffman Jan 15 at 1:54
For an annulus where only one of the two boundary circles is a reflector one is the same to $M\ddot{o}/\sim$. Here we have $M\ddot{o}/\sim$ where $(p,t)\sim(p,-t)$ for $0\le t\le 1$ over the Möbius strip $M\ddot{o}$... So, is the same fate? –  janmarqz Jan 15 at 2:10
@janmarqz I suppose I buried some point in the comment. It is a subtle, but reflector circles do not contribute to orbifold Euler characteristic. For the triangulation of an orbifold $Q$, any reflector circle would decompose into reflection arcs and vertices fixed by the dihedral group of order 2. For these vertices, $n_i =1$ and so looking at the formula in the answer above, these points contribute a $0$ to the sum. Thus, we can just ignore them. Again, I would recommend reading Thurston's notes, especially Chapter 13. It is rich with examples and many people have cut their teeth here. –  Neil Hoffman Jan 17 at 11:21


share|improve this answer

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.