The context for this question is from page 284 - 287 of Berger's paper: http://pdn.sciencedirect.com/science?_ob=MiamiImageURL&_cid=272332&_user=209810&_pii=S0021869398976785&_check=y&_origin=article&_zone=toolbar&_coverDate=1999--01&view=c&originContentFamily=serial&wchp=dGLbVlS-zSkzS&md5=2dd8e0d714d264cf7c4acdd9ec58ac84&pid=1-s2.0-S0021869398976785-main.pdf

Particularly, in his assumption at the top of page 287, he says that "From now on, assume that our map $\pi_\mathfrak{p}$ surjects onto $\text{PU}_2(\zeta,\mathcal{O}_K/\mathfrak{p})\cong \text{PSL}_2(\mathbb{F}_q)$, that $q$ is odd, and that $(6,k) = 1$, where $k = \sharp\langle -\zeta\rangle$."

I'm guessing that he's assuming the conditions in proposition 2 (from the previous page) to be true, so that $\pi_\mathfrak{p}$ surjects onto $\text{PU}_2(\zeta,\mathcal{O}_K/\mathfrak{p}) = \text{PSU}_2(\mathcal{O}_K/\mathfrak{p})$, and that he's claiming that the latter group is isomorphic to $\text{PSL}_2(\mathbb{F}_q)$.

Is this generally true?

Also, on page 284, where he gives the matrix $H$ for the hermitian form, he claims that $H\in GL_{n-1}(\mathbb{Z}[t,t^{-1}])$, but the matrix he gives obviously does not lie in that group.

Where might I find a good book on unitary matrices over finite fields?

thanks,

- will

`$\mathrm{PSU}_2$`

in the header. From the finite group perspective, there is only one family of type`$A_1$`

simple groups, usually denoted`$\mathrm{PSL}_2$`

. (Though in higher ranks there are different split and non-split simple groups of type`$A_\ell$`

.) – Jim Humphreys Jan 29 '13 at 20:23