next up previous
Next: The picture in the Up: Cutting up Previous: The geometrical problem

Cutting up the fundamental domain

There is an obvious fundamental domain tex2html_wrap_inline1891 for the action of tex2html_wrap_inline1419 on tex2html_wrap_inline1411 . Let tex2html_wrap_inline1897 be the planes in tex2html_wrap_inline1411 that meet the sphere at infinity in the circles tex2html_wrap_inline1429 , let tex2html_wrap_inline1903 be the corresponding open half-spaces that meet the sphere at infinity in the disks tex2html_wrap_inline1431 , and let tex2html_wrap_inline1891 be the complement of tex2html_wrap_inline1909 in tex2html_wrap_inline1411 . (See Figure 3.)

  
Figure 3: Planes in hyperbolic 3-space.

Note that whereas we described F as the exterior of tex2html_wrap_inline1429 , we can best describe tex2html_wrap_inline1891 as the interior of tex2html_wrap_inline1897 .

The manifold tex2html_wrap_inline1409 is obtained from tex2html_wrap_inline1891 by gluing the boundary components tex2html_wrap_inline1897 together in pairs. If we cut tex2html_wrap_inline1409 apart along the n/2 surfaces along which it was glued, we recover tex2html_wrap_inline1891 . By the cutting law, any lower bound for tex2html_wrap_inline1933 is also a lower bound for tex2html_wrap_inline1467 . Hence instead of cutting up tex2html_wrap_inline1409 we will work on cutting up tex2html_wrap_inline1891 . We will forget all about the pairings, and drop the assumption the n is even. Here's what we will prove:

Theorem. Let tex2html_wrap_inline1891 be the manifold (with boundary) corresponding to the finite collection of circles tex2html_wrap_inline1429 in the Riemann sphere. Then

displaymath1889

for some universal constant tex2html_wrap_inline1485 .



Peter Doyle