lset.f which enumerates limit points by the infinite word
algorithm for general terminal function groups representing a surface
with at least one puncture (so that the group is free).The choice of generators is based on the theory described in
I. Kra, Horocyclic coordinates for Riemann surfaces and moduli spaces. I: Teichmüller and Riemann spaces of Kleinian group, J.A.M.S., v. 3, 1990, 499-578.
PARI notation to describe the
generators
a = [ 1 , 2 ; 0 , 1]
b = [ 1 , 0 ; mu, 1]
This means:
Click to enlarge The blue arrows label certain limit points with their associated infinite word. Keep reading to learn what the notation means. The red curves are flowlines associated to certain elements of the group, flowing from the repelling fixed point to the attractive fixed point.
where each letter is a generator or an inverse and the word in between brackets is repeated without end. Thus, if z is the attractive fixed point of y1 .... ym then the associated limit point is
A capital letter always represents the inverse of the generator represented by the corresponding small case letter.
Click to enlarge
a = [ 1 + mu1*mu2 , 2*mu1 ; mu2/2 , 1]
b = [ 1 , 2 ; 0 , 1]
c = [ 1 , 0 ; 2 , 1]
Click to enlarge Same Picture: Focus on Bottom Half
Same Picture: Invariant domain filled
It should be emphasized that this algorithm draws these sets as
continuous curves, not as unions of circles. To see the curve-drawing
process clearly, one must view the original PostScript file via a
previewer such as ghostview. Click (or shift-click in
Netscape) on the image below to load the corresponding PostScript
file. It is a small piece of the above limit set, namely, all limit
points corresponding to infinite words beginning with
aa. Note: this is a 0.6 Mbyte compressed PostScript file
(.eps.gz).
A piece of the previous limit set
Riley Slice group inside T_1,2 group
Click to enlarge
Click to enlarge
a = [ 1 + mu1 , -mu1^2/2 ; 2 , 1 - mu1 ]
b = [ 1 , 2 ; 0 , 1]
c = [ 1 , 0 ; 2 , 1]
d = [ 1 + 2*mu2 , 2 ; -2*mu2^2 , 1 - 2*mu2 ]
Click to enlarge
Click to enlarge
a = [ 1 -2*mu1 -2*mu1^2*mu2 , 2 + 4*mu1*mu2 + 2*mu1^2*mu2^2 ;
-2*mu1^2, 1+2*mu1+2*mu1^2*mu2 ]
b = [ 1 + 2*mu2 , -2*mu2^2 ; 2 , 1 - 2*mu2 ]
c = [ 1 , 2 ; 0 , 1]
d = [ 1 , 0 ; 2 , 1]
e = [ 1 + 2*mu3 , 2 ; -2*mu3^2 , 1 - 2*mu3 ]
Click to enlarge
Click to enlarge
Click here to see this picture
a = i* [ mu1*mu2*mu3 + mu1 + mu2 , mu1*mu3 + 1 ; mu2*mu3 + 1 , mu3 ]
b = [ 1 + 2*mu1 , -2*mu1^2 ; 2 , 1 - 2*mu1]
c = [ -mu1*mu4 - mu4 - 1 , mu1*mu4 + mu4 - mu1 ; -mu4 , mu4 - 1]
d = [ 1 , 2 ; 0 , 1]
Click to enlarge
Click to enlarge
Click to enlarge At this cusp, the surface has had two curves pinched to points, resulting in a five-times punctured sphere. We see a set of equivalent components on which the group represents this 5-punctured sphere. The stabilizer of any of these components is a quasifuchsian group. We have highlighted in color the limit set of one such subgroup below.
The next picture has the corresponding domain completely colored in.
Based on the above illustrations, you may want to try to figure out generators for this subgroup and a side-pairing relation.
Click here for answer.
It is also instructive to identify fundamental regions on the above illustrations. We have identified limit points which serve as vertices for fundamental regions. You simply have to connect the dots in the right way.