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

The diagram below indicates 2 ways to use Fenchel-Nielsen coordinates to parameterize the Teichmüller space of conformal structures on the 4-holed sphere with totally-geodesic boundary, corresponding to the two pants decompositions, as indicated by this diagram:

Here $a,b,c,d,x,y$ are lengths of the indicated geodesics in a hyperbolic instance of the conformal structure, and $\alpha,\beta$ are the twisting parameters for the gluing.

What is the coordinate transformation implied by these decompositions? That is, what is the formula for $y$ and $\beta$ as a function of $a,b,c,d,x,\alpha$?

share|cite|improve this question
Just to be precise, there are infinitely many ways to give FN coordinates for the four-holed torus, one for every essential non-peripheral simple closed curve. – Sam Nead Mar 2 '14 at 14:32
Indeed, thanks for that point, I will edit the question to clarify this. – Jamie Vicary Mar 2 '14 at 14:36
up vote 5 down vote accepted

See the paper "Effects of a change of pants decompositions on their Fenchel-Nielsen coordinates" by Takayuki Okai, published in Kobe J. Math.

You can also find a version where some of the boundary components are cusps in the paper "The behaviour of Fenchel-Nielsen distance under a change of pants decomposition" by Alessandrini, Liu, Papadopoulos, and Su, published in Comm. Anal. Geom. and also available on the ArXiv.

Added by @JamieVicary: here are the equations as given in the Alessandrini, Liu and Papadopoulos paper in the notation of this question, with $a'=\text{cosh}(a/2)$, $b'=\text{cosh}(b/2)$, $c'=\text{cosh}(c/2)$, $d'=\text{cosh}(d/2)$, $x'=\text{cosh}(x/2)$, $y'=\text{cosh}(y/2)$, $\alpha' = \text{cosh}(\alpha)$ and $\beta' = \text{cosh}(\beta)$:

$$\begin{align} x'&=y' ^{-2} ( a'c'+d'b' + y'(a'd' + c'b') + \beta'(y'^2 + 2 a'b'y' + a'^2 + b'^2 - 1] ^{1/2}(y'^2 + 2 c'd'y' + c'^2 + d'^2)^{1/2}) \\ \alpha'&=(a'^2 + c'^2 + 2 a' c' x' + x'^2-1) ^{-1/2} ( d'^2 + b'^2 + 2 d' b' x' + x'^2 - 1) ^{-1/2} ((x'^2 - 1)y' - a'b' - c'd' - x'(a'd' + c'b')) \end{align}$$

share|cite|improve this answer
Hasn't this all been done earlier by Penner, with his Lambda lengths? – Igor Rivin Mar 2 '14 at 17:23
@Igor - As far as I understand them, Penner's lambda lengths require that the surface have punctures, and also a choice of ideal triangulation (instead of a pants decomposition). – Sam Nead Feb 9 at 19:49
I think that if you allow spun triangulations (or whatever they are called), everything works fine... – Igor Rivin Feb 9 at 19:58

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.