T H E O R E M .
Let C be a simply connected proper subset of C; pick some point
z0 ; then there exists a unique biholomorphism f : D such that f(z0 ) = 0
and f0 (z0 ) > 0.
R
d
= (); () = e1/2g(z) , g(z) = g(z0 )+
Proof.
:z0 z) ; is a branch of logarithm,
= / : (ez-w0 ) ;
so that z0 (1/zeg(z) = 0); dfefine g(z0 ) by 1/(z0 )eg(z0 ) = 1, f
1, so that f
: D; by scaling and translating 0 .
then |f|
Introduce the family F of functions f : D with the property that f is
holomorphic, injective, and f(0) = 0, f0 (0) > 0; as a remark, f F, holomorphic
and uniformly bounded; thus f is a normal family.
By Mordells Theorem, for
any sequence fn from this family, there exists a subsequence fnk which converges
to a holomorphic function f on any compact subset of .)
F =
6 because 1 F,
where 1 is f(z) = z.
supfF f0 (0) = s > 0 is attained by some f in F.
There exists a sequence of
functions such that fn0 (0) converges to s ; since (fn ) F, there exists a subsequence
that converges on compact subsets of .
Since fnk (0) = 0, we have f(0) = 0; thus fn0 k (0) = s > 0.
Since |fnk (z)| 1, the
limit also satisfies this Condition.
f is one-to-one.
Recall:
Theorem - Assume that you have a sequence of
holomorphic functions fn that converges uniformly to f on compact sets of ;
note that the supremem over all n of the cardinality of {z |fn (z) = w} is
bounded by some number N, then for arbitrary w C, the number of solutions to
the analagous equation for the limit is less than or equal to N as well.
Note that the functions fnk in F are one-to-one; for all w, the numbber of
solutions to fnk (z) = w is 0 or 1.
most one as well; id est.
Thus the number of solutions of f is at
f F.
It remains to prove that f is surjective.
Assume that it is not surjective.
Then there exists w0 in D \ f (); f() = 1 ; Pick an automorphism g1 of the
unit disc (a Mobius transform); then g1 (w0 ) = 0; then 2 = g(1 ) is simply
connected and does not contain 0.
Pick an automorphism g2 Aut(D) with g2 ( g1 (0) = 0.
Define F = g2
g1 to
be a Map from D D; for a suitable choice of an Angle [0, 2), e( i)0 (0) >
1
0, so that we may assume without loss of generality that F0 (0) > 0; thus the
inverse of F,F-1 exists; F- 1 = g- 1 (x.x2 ) g-2 1 is also a function D D.
However, h is N O T an automorphism, because the sequare function is not an
automorphism near the orogin.
BY the Schwarz lemma, |h0 (0)| < 1.
Consider
0 (0) = h0 (0)G0 (0).
G = F f ; then h G = h F f = f ; thus (h G)(0) = f
Note
that G : D satisfies G(0) = 0, G0 (0) > 0 and G one-to-one on ; thus G F and
0 (0) > f0 (0), in contradiction to f0 (0) = sup
0
1/h0 (0)f
fF f (0).
Therefore f () = D.