MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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

Let $Q$ be the Lorentzian matrix, that is $Q=\mathrm{diag}(I_n,-1)$, where $I_n$ denotes de $n\times n$ identity matrix. Let $\mathcal R$ the set of integer solutions $x\in\mathbb Z^{n+1}$ of $$ Q[x]:=x^tQx=x_1^2+\dots+x_{n}^2-x_{n+1}^2=-1. $$ The set $\Gamma=\mathrm{O}(n,1)\cap \mathrm{M}(n+1,\mathbb Z)$ acts by left multiplication on $\mathcal R$. It's known that there are finitely many $\Gamma$-orbits.

I think (but I'm not completely sure) that $\Gamma\backslash \mathcal R$ has only one class if $2\leq n\leq 7$, and two if $n=8$ (since $(0,0,0,0,0,0,0,0,1)$ and $(1,1,1,1,1,1,1,1,3)$ leaves in different $\Gamma$-orbits).

What happens in the hermitian case?

More precisely, let $\Gamma=\mathrm{U}(n,1)\cap\mathrm M(n+1,\mathcal O)$ with $\mathcal O$ the ring of integer of a quadratic imaginary extension of $\mathbb Q$, and let $\mathcal R$ be the set of "integer" solutions $x\in\mathcal O^{n+1}$ of $$ Q[x]=x^*Qx=|x_1|^2+\dots+|x_{n}|^2-|x_{n+1}|^2=-1. $$

How many $\Gamma$-orbits are there in the set $\mathcal R$?

[Edit: It's enough for $n=2,3$ and some particular ring $\mathcal O$, like $\mathbb Z[\sqrt{-1}]$.]

Thank you in advance.-.

share|cite|improve this question

I don't know the answer to your question (if there is one), but I just noticed a preprint by Matt Stover on the arxiv that I think is relevant to a related question. This preprint says that the number of commensurability classes of arithmetic lattices with a bounded number of ends is finite. Let's try to see how this might be related to your question.

Let ${\bf Q}[x]=|x_0|^2+|x_1|^2+\cdots+ |x_n|^2-|x_{n+1}|^2$. A solution to $Q[x]=-1$ gives a solution to ${\bf Q}[x]=0$ by setting $x_0=1$. Orbits of solutions to ${\bf Q}[x]=0$ correspond to cusps (ends) of the orbifold $\mathbb{CH}^{n+1}/{\bf \Gamma}$, where ${\bf \Gamma} = U({\bf Q};\mathcal{O})$. Thus, Stover's theorem says that for any $k$, there are only finitely many $\mathcal{O}$ and dimensions $n$ such that ${\bf \Gamma}$ has less than $k$ cusps (I think it's fairly standard that for different $\mathcal{O}$ quadratic rings of integers, the lattices $U({\bf Q};\mathcal{O})$ will be non-commensurable).

I would like to say now that this should imply that there are only finitely many $n, \mathcal{O}$ that there are $\leq k$ solutions to $Q[x]=-1$ up to the action of $\Gamma$. The problem is that not every ${\bf \Gamma}$-orbit of solution to ${\bf Q}[x]=0$ necessarily has a representative with $x_0=1$. However, I would still conjecture this to be true, and one might study the techniques in Stover's paper to see if they carry over to your question.

share|cite|improve this answer
Dear Agol, I'm just interested in the number of elements of $\Gamma\backslash \mathcal R$ for $n=2$ or $3$. – emiliocba Dec 28 '11 at 18:23

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.