The pictures were created in a project conducted by David Mumford at Harvard, assisted primarily by Curt McMullen and myself. (Curt McMullen has made available many of his papers and pictures here.) The starting date of this project was 1980, at the same time Mandelbrot was lecturing at Harvard on Julia sets and what has now come to be called the Mandelbrot set.
Here is a collection of topics to explore
kleinian
This picture illustrates the limit set of a quasifuchsian group
representing a pair of punctured tori. The group is a free group
generated by two linear fractional transformations which are
very nearly parabolic. This accounts for the explosively large
spirals in the picture.Here's a PostScript version of this picture:
A quasifuchsian group is a quasiconformal deformation of a Fuchsian group. A Fuchsian group is a discrete subgroup of PSL2(R) or more generally any discrete subgroup of PSL2(C) that possesses an invariant circle in the Riemann sphere.
To say G2 is a quasiconformal deformation of G1 means that there is a quasiconformal homeomorphism w of the Riemann sphere onto itself such that
G2 = w-1 G1 wThus, a quasifuchsian group possesses an invariant quasicircle (quasiconformal image of a circle).
The chief significance lies in the subject of uniformization. Koebe's uniformization says that any Riemann surface S of negative Euler characteristic may be realized as H/G for some Fuchsian group G. The lower halfplane then represents the "mirror" image of S. Possibly Lipman Bers' most elegant theorem (there are many choices) is that of Simultaneous Uniformization: for any pair of homeomorphic Riemann surfaces S1, S2 of negative Euler characteristic and finite topological type (i.e. finite genus and finitely many special points) there is a quasifuchsian group G which represents both S1 and S2 in its two invariant domains (the two components of the ordinary set). One main area of research in uniformization is to what extent this can be generalized to collections of Riemann surfaces of various kinds.
This picture is an enlargement of a portion of the previous
limit set. The two points marked with arrows are limit points
corresponding to two infinite words in the free group which are
extremely close in the lexicographical ordering on the free
group. Yet, nonetheless there is an extremely large and visible
portion of the limit set stretching between these two points.
This is the problem of "necks."
This is a picture of the limit set of a free group on
two generators in which two words have become "accidentally
parabolic". It is a limiting case of a quasifuchsian group
representing a pair of punctured tori. If the generators are denoted
X and Y, then the words that are parabolic are
Y and X15 Y.
This is called the 1/15 double cusp.