next up previous
Next: Annexing annuli Up: Cutting up Previous: Taking care of the

Room to grow

Consider a hemisphere tex2html_wrap_inline2275 , associated to the circle C and the disk D. Let the radius of C be a. We assume that tex2html_wrap_inline2275 is relevant, and not a baby, that is, that a > 2h. At height h, the tubes trapped over tex2html_wrap_inline2275 cover an annulus A(h) in the plane of outer radius a and inner radius tex2html_wrap_inline2297 . (See Figure 11.)

  
Figure 11: No room at infinity reprise.

The Euclidean area of A(h) is

This annulus is the image under projection into the plane of the annulus tex2html_wrap_inline2301 consisting the intersection of the plane z=h with the locus of points above tex2html_wrap_inline2275 and inside the cylinder over C. Earlier we determined that the hyperbolic area of tex2html_wrap_inline2301 was on the order of 1. Now we see that it is exactly pi.

In order to insure that the tubes over tex2html_wrap_inline2275 have enough room to grow, we will need to annex around the base of the hemisphere a region B having Euclidean area on the order of that of A (h), that is, we must have

for some universal constant K. To see that this is the right condition, let tex2html_wrap_inline2323 be the portion of the plane z=h that lies over B. The hyperbolic area of tex2html_wrap_inline2329 is

while the hyperbolic area of tex2html_wrap_inline2331 is

Our assumption on the area of B guarantees that the ratio of the latter quantity to the former will always be tex2html_wrap_inline2335 . Thus in going from height h down to height h/2 the hyperbolic area available to the tubes increases by a definite factor, which is the kind of condition we need.



Peter Doyle