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

Hello all

As is probably well-known to most, in the upper halfplane we have a natural action of $SL_2(\mathbb{R})$ through linear fractional transformations, and a $2$-form $\frac{dx dy}{y^2}$ which is invariant under this action.

In the case of the upper half-space $H_3 = \{ (z,t) \ | \ z\in \mathbb{C} \ , \ t>0 \ \}$ we have an analogous situation, only now we have a natural action of all of $SL_2(\mathbb{C})$. The fastest way of defining this action is by identifying a point $(z,t)$ in $H_3$ with the quaternion $q:= z+t j$, and then setting $$ \begin{pmatrix} a&b\cr c& d\cr \end{pmatrix}.(z,t) := (aq+b)(cq+d)^{-1}$$ Now for my question; I want to try to construct an $SL_2(\mathbb{Z}[i])$-invariant two-form on $H_3$ which when restricted to the plane $Im(z)=0$ takes the form $\frac{dxdt}{t^2}$. Do such a thing exist ? How would I go about constructing it? Any thoughts/references would be greatly appreciated.

share|cite|improve this question
You might be able to do this on a congruence subgroup of $PSL(2,\mathbb{Z}[i])$. Here's a reference: –  Ian Agol Jan 17 '11 at 21:07
Sorry, in my previous comment I assumed you were asking for an automorphic 2-form (like a harmonic form). If you just want a 2-form, you can definitely do this on a subgroup in which the totally geodesic surface becomes embedded and non-separating, such as the Whitehead or Borromean rings link complements. But it might be tricky to write down an explicit formula. –  Ian Agol Jan 17 '11 at 21:15

1 Answer 1

up vote 8 down vote accepted

In upper half space, the thing that's analogous to the invariant 2-form (which measures hyperbolic area) is the hyperbolic volume form $1/z^3 dx dy dz$.

If you want to understand 2-forms invariant by a particular discrete group, they are the same thing as 2-forms on the quotient orbifold. In the case of $SL_2$ of the algebraic integers in a quadratic imaginary field such as this, the groups are usually called Bianchi groups: Bianchi worked out fundamental domains for them for the first set of examples, as well as a general technique to find fundamental domains. On Allen Hatcher's website, there's a handy list of pictures of the folded-up fundamental domains in cases where they can be readily drawn.

In this case, the quotient orbifold has underlying space $S^3$ minus a point (for the cusp), with singular locus in the form of the 1-skeleton of a tetrahedron. Since the underlying space is contractible, any closed 2-form is a coboundary --- i.e. there are plenty of 2-forms, closed or otherwise, but there's no obvious reason they should be interesting. In this particular case, the diagonal matrix with entries $i, -i$ rotates 180 degrees about the $z$-axis, acting as an orientation-reversing map on the hyperbolic plane, so it takes the 2-form above to its negative: it's not invariant even before attempting to extend it.

share|cite|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.