Complex Analysis Notes

Math 213a — Harvard University

C. McMullen

1 Basic complex analysis; the simply-connected

Riemann surfaces
1. Background in real analysis and basic differential topology (such as
covering spaces and differential forms) is a prerequisite.
2. Relations of complex analysis to other fields include: algebraic geome-
try, complex manifolds, several complex variables, Lie groups and ho-
b geometry (Platonic solids; hyperbolic ge-
mogeneous spaces (C, H, C),
ometry in dimensions two and three), Teichmüller theory, elliptic curves
and algebraic number theory, ζ(s) and prime numbers, dynamics (iter-
ated rational maps).
3. Algebraic origins of complex analysis; solving cubic equations x3 + ax+
b = 0 by Tchirnhaus transformation to make a = 0. P This is
Pdone by
2 2
introducing a new variable y = cx + d such that yi = yi = 0;
even when a and b are real, it may be necessary to choose c complex
(the discriminant of the equation for c is 27b2 + 4a3 .) It is negative
when the cubic has only one real root; this can be checked by looking
at the product of the values of the cubic at its max and min.
4. Elements
√ of complex analysis: C = R[i], z (and related examples for
Q[ 2]). Geometry of multiplication: why is it conformal? (Because
|ab| = |a||b|, so triangles are mapped to similar triangles!) Visualizing
ez = lim(1 + z/n)n . A pie slice centered at −n and with angle π/n is
mapped to the upper semi-circle; in the limit we find exp(πi) = −1.
5. Definition: f (z) is analytic if f ′ (z) exists.
R Note: we do not require
continuity of f ′ ! Cauchy’s theorem: γ f (z)dz = 0. Plausibility:
R n Rb n ′
Rb
γ
z dz = a
γ(t) γ (t)dt = a
(γ(t)n+1 )/(n + 1)′ dt = 0.
Proof 1: f (z)dz is a closed form, when f is holomorphic, assuming
f (z) is smooth. Discussion: f is holomorphic iff idf /dx = df /dy; from

1
 this d(f dz) = 0. Moreover df /dx = df /dz, where d/dz = 1/2(d/dx +
(1/i)d/dy). We have f analytic if and only if df /dz = 0. Then df =
(df /dz)dz + (df /dz)dz and we see d(f dz) = 0 iff df /dz = 0 iff f is
holomorphic.
Proof 2: (Goursat), assuming only complex differentiability.
R
6. Analyticity and power series. The fundamental integral γ dz/z. The
P n
fundamental power series 1/(1 − z) = z . Put these together with
Cauchy’s theorem, Z
1 f (ζ)dζ
f (z) = ,
2πi γ ζ − z
to get a power series.
P
Theorem: f (z) = an z n has a singularity (where it cannot be analyti-
cally continued) on its circle of convergence |z| = R = 1/ lim sup |an |1/n .
Q P
7. Infinite products: (1 + an ) converges
Q if |an | < ∞. The proof is in
two steps: first, show that when (1+|an |) converges,Q the differences in
successive
P partial Q products bound P the differences for (1 + an ). Then,
show |an | ≤ (1 + |an |) ≤ exp( |an |).
Q
Example: evaluation of ζ(2) = (1/(1 − 1/p2 )), where p ranges over
the primes, converges (to π 2 /6).

8. Cauchy’s bound |f (n) (0)| ≤ n!M(R)/Rn . Liouville’s theorem; algebraic

completeness of C. Compactness of bounded functions in the uniform
topology. Parseval’s inequality (which implies Cauchy’s:)
X Z
2 2 1
|an | R = |f (z)|2 dθ ≤ M(R)2 .
2π |Z|=R

9. RMorera’s theorem (converse to Cauchy’s theorem). Definition

P of log(z) =
z n!
1
dζ/ζ. Analytic continuation,
P∞ natural boundaries, R na z . Lau-
n n+1
rent series: f (z) = −∞ an z where an = (1/2πi) C f (z)/z dz.
Classification of isolated singularities; removability of singularities of
bounded functions. Behavior near an essential singularity (Weierstrass-
Casorati): f (U) = C.
P
10. Generating functions and Fn z n , Fn the nth Fibonacci number. A
power series represents a rational function iff its coefficients satisfiy a

2
 recurrence relation. Pisot numbers, the golden ratio, and why are 10:09
and 8:18 such pleasant times.

11. Kronecker’s theorem: one need only check that the determinants of the
matrices ai,i+j , 0 ≤ i, j ≤ n are zero for all n sufficiently large.

12. RResidue theorem andR evaluation of definite

R ∞ a integrals. Three types:
2π ∞
0
R(cos θ, sin θ)dθ, −∞ R(x)dx, and 0 x R(x)dx, 0 < a < 1, R
a rational function.
R
13. Hardy’s paper on sin(x)/x dx.

14. The argument principle: number of zeros - number of poles is equal to

Z
1 f ′ (ζ)
dζ.
2πi ∂Ω f (ζ)

Similary, the weighted sum of g(z) over the zeros and poles is given by
multiplying the integrand by g(ζ).
Winding numbers of the topological nature of the argument principle:
if a continuous f : ∆ → C has nonzero winding number on the circle,
then f has a zero in the disk.

15. Rouché’s theorem: if |g| < |f | on ∂Ω, then f + g and g have the same
number of zeros-poles in Ω. Example: z 2 + 15z + 1 has all zeros of
modulus less than 2, but only one of modulus less than 3/2.

16. Open mapping theorem: if f is nonconstant, then it sends open sets to

open sets. Cor: the maximum principle (|f | achieves its maximum on
the boundary).

17. Invertibility. (a) If f : U → V is injective and analytic, then f −1

is analytic. (b) If f ′ (z) 6= 0 then f is locally injective at z. Formal
inversion of power series.

18. Phragmen -Lindelöf type results: if f is bounded in a strip {a <

Im(z) < b}, and f is continuous on the boundary, then the sup on
the boundary is the sup on the interior.

3
19. Hadamard’s 3-circles theorem: if f is analytic in an annulus, then
log M(r) is a convex function of log r, where M(r) is the sup of |f | over
|z| = r. Proof: a function φ(s) of one real variable is convex if and only
if φ(s) + ar satisfies the maximum principle for any constant a. This
holds for log M(exp(s)) by considering f (z)z a locally.

20. The concept of a Riemann surface; the notion of isomorphism; the three
simply-connected Riemann surface C, C b and H.

21. A nonconstant map between compact surfaces is surjective, by the open

mapping theorem.

22. Theorem: Aut(C) = {az + b}.

23. The complex plane C. The notion of metric ρ(z)|dz|. The automor-
phism group√is solvable. Inducing a metric |dz|/|z| on C∗ . The cone
metric |dz|/| z| giving the quotient by z 2 .

24. The Riemann sphere C b and its automorphisms. Theorem: Aut(C) b =

PSL2 (C). A particularly nice realization of this action is as the projec-
tivization of the linear action on C2 .

25. Möbius transformations: invertible, form a group, act by automor-

b triply-transitive, sends circles to circles. Proof of last: a
phisms of C,
circle x + y 2 + Ax + By + C = 0 is also given by r 2 + r(A cos θ +
2

B sin θ) + C = 0, and it is easy to transform the latter under z 7→ 1/z,

which replaces r by 1/ρ and θ by −α.

26. Classification of Möbius transformations and their trace squared: (a)

identity, 4; (b) parabolic (a single fixed point) 4; (c) elliptic (two fixed
points, derivative of modulus one) [0, 4); (d) hyperbolic (two fixed
points, one attracting and one repelling) C − [0, 4].

27. Stereographic projection preserves circles and angles. Proof for angles:
given an angle on the sphere, construct a pair of circles through the
north pole meeting at that angle. These circles meet in the same angle
at the pole; on the other hand, each circle is the intersection of the
sphere with a plane. These planes meet C in the same angle they meet
a plane tangent to the sphere at the north pole, QED.

4
 b the extended complex plane; the Riemann sphere; the
28. Four views of C:
Riemann surface obtained by gluing together two disks with z 7→ 1/z;
the projective plane for C2 .

29. The spherical metric 2|dz|/(1 + |z|2 ). Derive from the fact “Riemann
circle” and the map x = tan(θ/2), and conformality of stereographic
projection.

30. Some topology of projective spaces: RP2 is the union of a disk and a
Möbius band; the Hopf map S 3 → S 2 .

31. Gauss-Bonnet for spherical triangles: area equals angle defect. Prove
by looking at the three lunes (of Rarea 4θ)R for the three angles of a
triangle. General form: 2πχ(X) = X K + ∂X k.
b →C
32. Dynamical application of Schwarz Lemma: let f : C b be a rational
map. Then the immediate basin of any attracting cycle contains a
critical point.
Cor. The map f has at most 2d − 2 attracting cycles.

33. Theorem: Aut(H) = PSL2 (R). Schwarz Lemma and automorphisms of

the disk. The hyperbolic metric |dz|/ Im(z) on H, and its equivalence
to 2|dz|/(1 − |z|2 ) on ∆.

34. Classification of automorphisms of H2 , according to translation dis-

tance.

35. Hyperbolic geometry: geodesics are circles perpendicular to the circle

at infinity. Euclid’s fifth postulate (given a line and a point not on the
line, there is a unique parallel through the point. Here two lines are
parallel if they are disjoint.)

36. Gauss-Bonnet in hyperbolic geometry. (a) Area of an ideal triangle is

R1 R∞
−1
√
1−x2
(1/y 2)dydx = π. (b) Area A(θ) of a triangle with two ideal
vertices and one external angle θ is additive (A(α + β) = A(α) + A(β))
as a diagram shows. Thus A(α) = α. (c) Finally one can extend
the edges of a general triangle T in a spiral fashion to obtain an ideal
triangle containing T and 3 other triangles, each with 2 ideal vertices.

5
37. Harmonic functions. A real-value function u(z) is harmonic iff u is
locally the real part ofR an analytic function; indeed, harmonic means
d(∗du) = 0, and v = ∗du. Harmonic functions are preserved under
analytic mappings. Examples: electric potential; fluid flow around a
cylinder.
The mean-value principle (u(0) = average over the circle) follows from
Cauchy’s formula, as does the Poisson integral formula (u(p) = visual
average of u).
38. The Schwarz reflection principle: if U = U ∗ , and f is analytic on U ∩H,
continuous and real on the boundary, then f (z) extends f to all of U.
This is easy from Morera’s theorem. A better version only requires
that Im(f ) → 0 at the real axis, and can be formulated in terms of
harmonic functions (cf. Ahlfors):
If v is harmonic on U ∩ H and vanishes on the real axis, then v(z) =
−v(z) extends v to a harmonic function on U. For the proof, use the
Poisson integral to replace v with a harmonic function on any disk
centered on the real axis; the result coincides with v on the boundary
of the disk and on the diameter (where it vanishes by symmetry), so
by the maximum principle it is v.
39. Reflection gives another proof that all automorphisms of the disk ex-
tend to the sphere.
40. Quotient of the cylinder: s(z) = z+1/z gives the quotient isomorphism,
C∗ /(Z/2) ∼
= C. It has critical points at ±1 and critical values at ±2.
Since z 7→ z n commutes with z 7→ 1/z, there are unique polynomial
pn (z) such that s(z n ) = pn (s(z)). Writing z = eiθ , these are essentially
the Chebyshev polynomials; they satisfy 2 cos(nθ) = pn (2 cos θ).
41. Classification of polynomials. Let us say p(z) is equivalent to q(z)
is there are A, B ∈ Aut(C) such that Bp(Az) = q(z). Then every
polynomial is equivalent to one which is monic and centered (the sum of
its roots is zero). Every quadratic polynomial is equivalent to p(z) = z 2 .
42. Solving the cubic. Every cubic polynomial is equivalent to p3 (z) =
z 3 − 3z. But this polynomial arises as a quotient of z 7→ z 3 ; that is,
it satisfies s(z 3 ) = p3 (s(z)). Thus we can solve p3 (z) = w easily: the
solution is s(u), where s(u3) = w.

6
 The mapping p3 (z) (and indeed all the maps pn ) preserve the ellipses
and hyperbolas with foci ±2, since z 7→ z n preserves circles centered
at z = 0 and lines through the origin. This facilitates visualizing these
polynomials as mappings of the plane to itself.

2 Entire and meromorphic functions

1. Weierstrass/Hadamard factorization theory. (A good reference for this
material is Titchmarsh, Theory of Functions). We will examine the
extent to which an entire function is determined by its zeros. We
beginning by showing that any discrete set in C arises as the zeros of
an entire function.

2. Weierstrass factor. Inspired by the fact that (1−z) exp log 1/(1−z) = 1
and that log 1/(1 − z) = z + z 2 /2 + z 3 /3 + . . ., we set
 
z2 zp
Ep (z) = (1 − z) exp z + +···+ .
2 p

By convention E0 (z) = (1 − z).

Theorem: For |z| < 1, we have |Ep (z) − 1| ≤ |z|p+1 .
P
Proof: Writing Ep (z) = P 1+ ak z k , one may check (by computing
′
Ep (z)) that all P
ak < 0, |ak | =P1, and a1 = a2 = . . . = ap = 0. Then
|Ep (z) − 1| = | ak z k | ≤ |z|p+1 |ak | ≤ |z|p+1 . QED
P Q
3. Theorem: If (r/|an |)pn+1 < ∞ for all r > 0, then f (z) = Epn (z/an )
converges to an entire function with zeros exactly at the an .
Cor: Since pn = n works for any an → ∞, we have shown any discrete
set arises as the zeros of an entire function.

4. Blaschke products. Let f : ∆ → ∆ be a proper map of degree d. Then

d 
Y 
iθ z − ai
f (z) = e
1
1 − ai z

where the ai enumerate the zeros of f .

7
5. Jensen’s formula. Let f (z) be holomorphic on the disk of radius R
about the origin. Then the average of log |f (z)| over the circle of radius
R is given by:
X R
log |f (0)| + log .
|z|
f (z)=0; |z|<R

Proof: Suffices to assume R = 1. Clear if f has no zeros, because

log |f (z)| is harmonic. Clear for a Blaschke factor (z − a)/(1 − az). But
the formula is true for f g if it is true for f and g, so we are done.
Cor: Let n(r) be the number of zeros of f inside the circle of radius r.
Then Z R Z 2π
dr 1
n(r) = log |f (Reiθ )|dθ − log |f (0)|.
0 r 2π 0

Remark: We have used the mean value property of harmonic functions.

This holds for any harmonic function u on the disk by writing u =
Re(f ), f holomorphic, and then applying Cauchy’s integral formula for
f (0).
The physical idea of Jensen’s formula is that log |f | is the potential for
a set of unit point charges at the zeros of f .
6. Entire functions of finite order. An entire function f : C → C is of
finite order if there is an A > 0 such that |f (z)| = O(exp |z|A ). The
least such A is the order ρ of f .
Examples:
√ Polynomials have order 0; sin(z), cos(z), exp(z) have order
1; cos( z) has order 1/2; exp(exp(z)) has infinite order.
7. Theorem. Let f (z) be an entire function of finite order with no zeros.
Then f (z) = eQ(z) where Q(z) is a polynomial.
Proof. Since f has no zeros, f (z) = eQ(z) for some entire function
d
Q(z). Since f has finite order, |f (z)| = O(e|z| ) for some d, and thus
Re Q(z) ≤ |z|d + O(1). Thus there is a constant A > 0 such that Q
maps ∆(2R) into the half-plane U(R) = {z : Re z < ARd }, for R > 1.
By the Schwarz Lemma, Q is distance-decreasing from the hyperbolic
metric on ∆(2R) to the hyperbolic metric on U(R). Since ∆(R) ⊂
∆(2R) has bounded hyperbolic diameter, the same is true for Q(∆(R)) ⊂
U(R). Therefore in the Euclidean metric,
diam Q(∆(R)) = O(d(Q(0), ∂U(R))) = O(Rd).

8
 Therefore |Q(z)| = O(|z|d ) for |z| > 1. By Cauchy’s integral formula,
we find Qd+1 (z) = 0 and thus Q is a polynomial (of degree at most d).

8. Number of zeros. By Jensen’s formula, if f has order ρ, then n(r) =

ρ+ǫ
O(r
P ), where n(r) is the zero counting function for f . Corollary:
1/|ai |ρ+ǫ < ∞, where ai enumerates the zeros of f (other than zero
itself).
In other words, ρ(ai ) ≤ ρ(f ), where ρ(ai ) is the
P exponent of convergence
ρ+ǫ
of the zeros of f , i.e. the least ρ such that 1/|ai| < ∞.
9. Definition: a canonical product is an entire function of the form
Y
f (z) = z m Ep (z/ai )
P
where p is the least integer such that |z/ai |p+1 < ∞ for all z.

10. Hadamard’s Factorization Theorem. Let f be an entire function of

order ρ. Then f (z) = P (z) exp Q(z), where P is a canonical product
with the same zeros as f and Q is a polynomial of degree less than or
equal to ρ.
11. To prove Hadamard’s theorem we develop two estimates. First, we
show a canonical product P (z) is an entire function of order ρ = ρ(ai ).
This is the least order possible for the given zeros, by Jensen’s theorem.
Second, we show a canonical product has m(r) ≥ exp(−|z|ρ+ǫ ) for most
radii r, where m(r) is the minimum of |f (z)| on the circle of radius r.
More precisely, for any ǫ > 0, this lower bound holds for all r outside
a set of finite total length. In particular it holds for r arbitrarily large.
Given these facts, we observe that f (z)/P (z) is an entire function
of order ρ with no zeros. Thus Q(z) = log f (z) is an entire func-
tion satisfying Re Q(z) = O(1 + |z|ρ+ǫ ). This implies
P Qn is a poly-
nomial. Indeed, for any entire function f (z) = an z , we have
n
|an |r ≤ max(4A(r), 0) − 2 Re(f (0)), where A(r) = max Re f (z) over
the circle of radius r.
12. Theorem. A canonical product has order ρ = ρ(ai ) and also satisfies
m(r) ≥ exp(−r ρ+ǫ ) for infinitely many r.
Proof. LetPrn = |an | and r = |z|. Recall that p is the least integer
such that (1/rn )p+1 < ∞. Thus p ≤ ρ ≤ p + 1. For convenience we

9
 P
will assume (1/rn )ρ < ∞. (For the general case, just replace ρ with
ρ + ǫ.)
By construction, the Weierstrasss factor
 
u2 up
Ep (u) = (1 − u) exp u + +···+
2 p

satisfies
up+1 up+2
log Ep (u) = + +···
p+1 p+2
for |u| < 1. Thus for ‘u small’, meaning |u| < 1/2, we have

| log Ep (u)| = O(|u|p+1). (2.1)

On the other hand, for ‘u large’, meaning 1/2 ≤ |u|, we have

| log |Ep (u)|/|1 − u|| = O(|u|p). (2.2)

The |1 − u| can be ignored unless u is very close to 1, making the log

very negative.
Q
Combining these estimates for f (z) = Ep (z/an ), we get the upper
bound:
 
X 1 X 1
log |f (z)| = O r p+1 p+1 + r
p
.
rn rnp
r/rn <1/2 r/rn ≥1/2

Now in the second sum, since p ≤ ρ, we have

X 1 X r ρ−p X 1
n ρ−p ρ−p
= ≤ (2r) ρ = O(r ).
rnp r ≤2r
r ρ
n r n
r/rn ≥1/2 n

Similarly in the first sum, since ρ ≤ p + 1, we have

X 1 X 1 1 X 1
ρ−p−1 ρ−p−1
= ρ ≤ (2r) ρ = O(r ).
r/rn <1/2
rnp+1 rn
r
>2r n
p+1−ρ
rn rn

Altogether this gives

log |f (z)| = O(r ρ),

10
 and thus f has order ρ.
The lower bound works the same way, since equations (2.1) and (2.2)
give bounds for |logEp (u)|. All that remains is to prove that
X
log |1 − z/an | = O(r ρ+ǫ )
r/rn >1/2

for infinitely many values of r.

Note that the number of terms in the sum above is N(2r), the number
of n such that rn < 2r. We have
X
N(2r)(2r)−ρ ≤ rn−ρ < ∞,

and thus N(2r) = O(r ρ). Also when log |1 − z/an | is positive it is no
bigger than O(log r) (the constant depending on the smallest an ), and
so the positive terms contribute at most N(r) log r to the sum, which
is O(r ρ+ǫ).
We now have to worry about the terms where z/an is close P to one. To
ρ ρ
this end, we require that |r − rn | > 1/rn for all n. Since 1/rn < ∞,
a set of r of infinite measure remains. Then for any such r, if z/an is
close to one, we still have

|1 − z/an | = |an − z|/|an | ≥ rn−1−ρ ,

and thus log |1 − a/zn | = O(log rn ) = O(log r). Thus the cases where
z/an is close to one also contribute O(N(r) log r) = O(r ρ+ǫ ).

13. Example:
Y z2

sin(πz) = πz 1− 2 .
n6=0
n

Indeed, the right hand side is a canonical product, and sin(πz) has
order one, so the formula is correct up to a factor exp Q(z) where Q(z)
has degree one. But since sin(πz) is odd, we conclude Q has degree
zero, and by checking the derivative at z = 0 of both sides we get
Q = 0.

11
14. Alternative proof. We have
X ∞
π2 1 π2 π4 2 1
2 = 2
+ + z + · · · =
sin (πz) z 3 15 −∞
(z − n)2

because both sides are periodic and tend to zero as | Im z|− > ∞.
Integrating, we obtain
∞
1 X 1 1
π cot πz = + + ,
z 1
z − n z + n

using the fact that both sides are odd to fix the constant of integration.
On the other hand, the logarithmic derivative of f (z) = sin(πz) is
π cot(πz). Using the fact the logarithmic derivative of 1 − z 2 /n2 is the
same as that of (z − n)(z + n), namely (z − n)−1 + (z + n)−1 we deduce
that g ′ /g = f ′ /f where
∞ 
Y 
z2
g(z) = πz 1− 2 .
1
n

Since sin(πz)/(πz) = g(z)/(πz) = 1 at z = 0, we conclude that f = g.

15. The Gamma Function:

exp(−γz) Y  z −1
∞
Γ(z) = 1+ exp(z/n).
z 1
n

Note that this expression contains the reciprocal of the canonical prod-
uct associated to the non-positive integers. The constant γ is chosen
so that Γ(1) = 1; it can be given by:

Xn
γ = lim( 1/k) − log(n + 1).
1

R n+1
This expression is the error in an approximation to 1 dx/x by the
area of n rectangles of base one lying over the graph.

12
 Q
16. Let G(z) = ∞ 1 (1 + z/n)e
−z/n
be the canonical product for the set of
negative integers. Evidentally

sin(πz)
G(z)G(−z) = · (2.3)
πz
Moreover, we find G(z − 1) = z exp(az + b)G(z) since both sides have
the same zeros and are of order one. Taking the logarithmic derivative
shows a = 0 and b = γ; that is,

G(z − 1) = zeγ G(z).

Defining
1
Γ(z) = ,
zeγz G(z)
we obtain Γ(z + 1) = zΓ(z), Γ(0) = Γ(1) = 1, Γ(z) has no zeros, Γ(z)
has simple poles at −1, −2, −3, . . . ; and Γ(n + 1) = n! for n ≥ 0.
Using the fact that

Γ(z)
Γ(z − n) = ,
(z − n) · · · (z − 1)

we find
(−1)n
Res−n (Γ(z)) = ·
n!
17. Sine/Gamma formula: from 2.3 we obtain:
π
= Γ(z)Γ(−z).
z sin(πz)

18. Gauss’s formula:

n!nz
Γ(z) = lim .
n→∞ z(z + 1) · · · (z + n)

This is obtained by examining the product formula directly for zeγz G(z) =
1/Γ(z).
Corollary: |Γ(z)| ≤ Re z when Re z > 0.

13
19. Uniqueness Theorem (Wielandt, 1939): If F (z + 1) = zF (z) for Re z >
0, F (1) = 1 and F (z) is bounded on the strip {Re z ∈ [1, 2]}, then
F (z) = Γ(z).
Proof. The functional equation allows one to extend F (z) to a mero-
morphic function on the whole plane, whose poles and their residues
agree with those of Γ(z). Thus G(z) = F (z) − Γ(z) is entire, G(0) =
G(1) = 0 and G(z + 1) = zG(z). Our boundedness assumptions now
imply that G(z) is bounded in the strip S = {Re z ∈ [0, 1]}. Thus
H(z) = G(z)G(1 − z) is also bounded in S. The functional equation
for G implies H(z + 1) = H(z) (as in the sine formula), and thus H is
a constant, which must be zero.

20. The integral representation: for Re(z) > 0,

Z ∞
dt
Γ(z) = e−t tz .
0 t

In other words, Γ(z) is the Mellin transform of the function e−t on R∗ .

The Mellin transform is an integral against characters χ : R∗ → C∗
(given by χ(t) = tz ), and as such it can be compared to the Fourier
transform (for the group R under addition) and to Gauss sums. Indeed
the Gauss sum X
σ(χ) = χ(n)e2πin/p
(n,p)=1

is the analogue of the Gamma function for the group (Z/p)∗ .

Proof. Apply the uniqueness theorem above.
√
21. Duplication formula: πΓ(2z) = 22z−1 Γ(z)Γ(z + 1/2). Note that both
sides have poles at −n/2, n > 0.

22. Periodic functions: If f : C → X has period λ, then f (z) = F (exp(2πiz/λ),

where F : C∗ → X. Example:
∞
X 1 π2
= .
−∞
(z − n)2 sin2 (πz)
P∞
From this we get 1 1/n2 = π 2 /6.

14
23. The Mittag-Leffler theorem. Let an → ∞ in C be a sequence of distinct
points, and let pn (z) be principal parts centered at an . Then there
exists a memorphic function f (z) with poles exactly at an , and with
the prescribed principal parts.
Proof. Let qn (z) be the polynomial given by the power series for pn (z)
about z = 0,P truncated so |pn (z) − qn (z)| < 2−n when |z| < |an |/2.
Then f (z) = (pn (z) − qn (z)) is the desired function.

The Weierstrass Theorem — that there exists a function g(z) with

prescribed zeros
R — is a corollary, by taking pn (z) = mn /(z − an ) and
g(z) = exp( f (z) dz).

3 Conformal mapping
1. Normal families: any bounded family of analytic functions is normal,
by Arzela-Ascoli.

2. Riemann mapping theorem: given a simply-connected region U ⊂ C,

U 6= C, and a basepoint u ∈ U, there is a unique conformal homeo-
morphism f : (U, u) → (∆, 0) such that f ′ (u) > 0. Proof: let F be the
family of univalent maps (U, u) → (∆, 0). Using a square-root and an
inversion, show F is nonempty. Also F is closed under limits. By the
Schwarz Lemma, |f ′ (u)| has a finite maximum over all f ∈ F . Let f
be a maximizing function. If f is not surjective to the disk, then we
can apply a suitable composition of a square-root and two automor-
phisms of the disk to get a g ∈ F with |g ′(u)| > |f ′(u)|, again using
the Schwarz Lemma. QED.

3. The length-area method. Let f : R(a, b) → Q be a conformal map of a

rectangle to a Jordan region Q ⊂ C, where R(a, b) = [0, a] × [0, b] ⊂ C,
Then there is a horizontal line (0, a) × {y} whose image has length
L2 ≤ (a/b) area(Q). Similarly for vertical lines.
Proof: apply the Cauchy-Schwarz inequality: there exists a horizontal
length whose image has length at most L where
 Z 2 Z  Z 
2 −1 ′ 1 2 ′ 2 a
L = b |f | ≤ 2 1 |f | = area(Q).
b b

15
 Corollary: given any quadrilateral Q, the product of the minimum
distances between opposite sides is a lower bound for area(Q).

4. Theorem: The Riemann map to a Jordan domain extends to a home-

omorphism on the closed disk.
Proof: given a point z ∈ ∂∆, map ∆ to an infinite strip, sending z to
one end. Then there is a sequence of disjoint squares in the strip tending
towards that end. The images of these squares have areas tending to
zero, so there are cross-cuts whose lengths tend to zero as well, by
the length-area inequality. This gives continuity at z. Injectivity is by
contradiction: if the map is not injective, then it is constant on some
interval along the boundary of the disk.

5. Schwarz reflection: a Riemann mapping f : ∆ → U can be analytically

continued past p ∈ S 1 whenever ∂U is a real-analytic arc near p. When
the arc is a straight line, the continuation is easily given by reflection.

6. Uniformization of annuli: any doubly-connected region in the sphere is

conformal isomorphic to C∗ , ∆∗ or A(R) = {z : 1 < |z| < R}. The
map from H to A(R) is z 7→ z α , where α = log(R)/(πi). The deck
transformation is given by z 7→ λz, where λ = exp(2π 2 / log(R)).

7. Uniformization of quadrilaterals: any quadrilateral is equivalent to a

unique rectangle. Sketch of the proof, by continuity.

8. PDE Application: harmonic functions. The function log((1+z)/(1−z))

on the disk, and the harmonic function =0 on top hemisphere, =1 on
the bottom. fluid flow around a cylinder, and the inverse of z + 1/z.

9. The class S of univalent maps f : ∆ → C such that f (0) = 0 and

f ′ (0) = 1. Compactness
P of S. The Bieberbach Conjecture/de Brange
Theorem: f (z) = an z n with |an | ≤ n.
P
10. The areaP theorem: if f (z) = z + bn /zn is univalent on {z : |z| > 1},
2
then n|bn | < 1. The proof is by integrating fdf over the unit

16
 circle and observing that, since |dz|2 = (−i/2)dz dz, the area A of
complement of the image of f is given by:
Z X Z  X 
i i 2 dz
A=− f df = − (1 − n|bn | ) =π 1− n|bn |2 .
2 S1 2 S1 z

Remark: little is known about optimal bounds for |bn | over Σ. The
area theorem gives |bn | = O(1/n1/2) and it is conjectured that |bn | =
O(1/n3/4 ); see [CJ].
2
11. Proof that |a2 | ≤ 2: first, apply
pthe area theorem to3 conclude |a2 −
2
a3 | ≤ 1. Then consider g(z) = f (z ) = z + (a2 /2)z + · for f (z) =
z + a2 z 2 + · · · ∈ S

12. The Koebe 1/4 Theorem: if f ∈ S then f (∆) ⊃ ∆(1/4). Proof: if

w is omitted from the image, then f (z)/(1 − f (z)/w) ∈ S; now apply
|a2 | ≤ 2.

13. Corollary. The hyperbolic metric on a simply-connected region U ⊂ C

satisfies ρ(z) ∈ [1/2, 2] · d(z, ∂U).

14. The distortion lemmas. Given f : ∆ → C, univalent, investigate the

invariant quantity δ(z) = |f ′ (z)|/ρ(z). This measures the derivative
from the hyperbolic metric to the Euclidean metric.
We claim along any hyperbolic path, we have d(log δ)/ds ≤ 2. It suffices
to treat the case z = 0. Then ρ(z) = 2/(1 − |z|2 ) is stationary, while
|f ′′(z)/f ′ (z)| ≤ |2a2 |/|a1 | ≤ 4. But |f ′′/f ′ | measures d/|dz| which is
twice d/ds, since ρ(0) = 2, so we get the desired bound.
It follows that δ(z)/δ(0) ≤ exp(2d(0, r)), r = |z|. But exp(d(0, r)) =
(1 + r)/(1 − r), as can be seen by using i(1 − z)/(1 + z) to map to H.
Thus δ(z)/δ(0) ≤ (1 + r)2 /(1 − r)2 . It follows that for f ∈ S, since
f ′ (0) = 1, we have:

′ |f ′(z)| (1 + r)2 ρ(r) (1 + r)

|f (z)| = ′ ≤ 2
· = ·
|f (0)| (1 − r) ρ(0) (1 − r)3
Similarly we have the reciprocal lower bound
(1 − r)
≤ |f ′ (z)|.
(1 + r)3

17
 We may finally conclude that S is compact, and indeed we can obtain
by integration the upper and lower bounds:
r r
2
≤ |f (z)| ≤ ·
(1 + r) (1 − r)2

15. Riemann surfaces and holomorphic 1-forms. The naturality of the

residue, and of df .

16. The Residue Theorem: the sum of the residues of a meromorphic 1-

form on a compact Riemann surface is zero. Application to df /f , and
thereby to the degree of a meromorphic function.

17. Remarks on 1-forms: a holomorphic 1-form on the sphere is zero, be-

cause it integrates to a global analytic function. Moreover a meromor-
phic 1-form on the sphere always has 2 more poles than zeros.

18. The Schwarz-Christoffel formula. Let f : H → U be the Riemann

mapping to a polygon with vertices pi , i = 1, . . . , n and exterior angles
πµi. Then Z
dζ
f (z) = α Qn µi
dζ + β,
1 (ζ − qi )

where f (qi ) = pi .
Proof: Notice that g(z) = (f (z) − pi )αi extends by Schwarz reflection
across qi , where αi = 1/(1 − µi ). This implies

f (z) = pi + (z − qi )1−µi hi (z),

where hi (z) is holomorphic near qi and hi (qi ) 6= 0, and thus f ′ (z)

behaves like (z − qi )−µi .
Next we compute the nonlinearity N(f ) = f ′′ (z)/f ′ (z)dz = d log f ′ (z).
By Schwarz reflection, it extends to a meromorphic 1-form on the sphere
with simple poles at z = qi and at z = ∞. (Near infinity, f (z) behaves
like 1/z which has nonlinearity −2dz/z and hence residue 2 at infinity.)
Using the local expression above, one finds that N(f ) has residue −µi
at qi . Since a 1-form is determined by its singularities, we have
X −µi dz
N(f ) = ,
(z − qi )

18
 and the
P formula results by integration. Note that the residue theorem
gives πµi = 2π which is also consistent geometrically (the sum of
the exterior angles in π.)
R
19. Examples of Schwarz-Christoffel: log(z) =p dζ/ζ (maps to a bigon
R
with external angles of π); sin−1 (z) = dζ/ 1 − ζ 2 (maps to a triangle
with external angles π/2, π/2 and π.)

20. Regular polygons. The sameR reasoning applies to Riemann maps on

the disk and shows f (z) = dz/(z n − 1)1/n maps the unit disk to a
regular n-gon.

21. Bloch’s Theorem: there exists a universal R > 0 such that for any
f : ∆ → C with |f ′ (0)| = 1, not necessarily univalent, there is an open
set U ⊂ ∆ (perhaps a tiny set near S 1 ) such that f maps U univalently
to a ball of radius r.
It
√ is known that the best value of R, Bloch’s constant, satisfies 0.433 <
3/4 ≤ R < 0.473. The best-known upper bound comes from the Rie-
mann surface branched with order 2 over the vertices of the hexagonal
lattice.

22. Little Picard Theorem. A nonconstant entire function omits at most

one value in the complex plane. (This is sharp as shown by the example
of exp(z).)

23. These apparently unrelated theorems can both be proved using the
same idea. (Cf. [BD] and references therein.)
We first consider Bloch’s theorem. Given f : ∆ → C, let

kf ′ (z)k = kf ′ (z)k∆,C = (1/2)|f ′(z)|(1 − |z|2 )

denote the norm of the derivative from the hyperbolic metric to the
Euclidean metric. By assumption, kf ′ (0)k = 1/2. We can assume
(using f (rz)) that f is smooth on S 1 ; then kf ′ (z)k → 0 as |z| → 1,
and thus sup |kf ′ (z)k is achieved at some p ∈ ∆.
Now replace f with f ◦ r where r ∈ Aut(∆) moves p to zero. We can
also shrink and translate f so f (0) = 0 and kf ′ (0)k = 1; this will only
decrease the size of its unramified disk. Then kf ′ (z)k ≤ kf ′ (0)k = 1,
and thus f |∆(1/2) ranges in a compact family of nonconstant anayltic

19
 functions. Thus the new f has an unramified disk of definite radius;
but then the old f does as well.

b we now let
24. Given g : C → C,
|g ′(z)|
kg ′(z)k∞ = .
(1 + |g(z)|2)

This is the norm of the derivative from the Euclidean metric |dz| to the
spherical metric of curvature 4. Note that g(z) = exp(z) has kg ′k∞ =
1/2; a function with bounded derivative can be rather wild.
Theorem. Let fn : ∆ → C b be a sequence of analytic functions such
′
that kfn (0)k∞ → ∞. There after passing to a subsequence, there is
a nonconstant entire function g : C → C such that g = lim fn ◦ rn .
Moreover, sup kg ′ (z)k∞ < ∞.
Here rn are Möbius transformations such that any compact set K ⊂ C
satisfies rn (K) ⊂ ∆ for all n ≫ 0.
b the norm
Proof. The main idea is to consider, for g : ∆ → C,
|g ′(z)|(1 − |z|2 )
kg ′ (z)k1 =
(1 + |g(z)|2)

from the hyperbolic metric on ∆ to the spherical metric on C,b scaled

for convenience to get rid of factors of 2; and more generally, for g :
∆(R) → C,b the norm

|g ′(z)|(1 − |z/R|2 )
kg ′ (z)kR =
(1 + |g(z)|2)

using the hyperbolic metric ρR = |dz|/(1 − |z/R|2 ) of curvature −4.

We can assume fn is analytic on ∆; then kfn′ k1 → 0 near S 1 so the
maximum is achieved somewhere, and greater than kfn′ (0)k1 . Using
Möbius transformations rn : ∆ → ∆, we can normalize so the maximum
is achieved at z = 0. Then we still have Rn = kfn′ (0)k1 → ∞, but now
kfn (z)k1 ≤ kfn′ (0)k1 .
b Then kg ′ (z)kR ≤ 1 and the
Now let gn (z) = fn (z/Rn ) : ∆(Rn ) → C. n
value 1 is again achieved at z = 0.

20
 Now for any compact set K, we have ρRn → |dz| uniformly on K. Thus
kgn′ (z)k∞ is uniformly bounded on K. Thus by Arzela-Ascolia, we can
extract a uniformly convergent subsequence. This gives the desired
limit g(z).

25. Proof of the Little Picard Theorem. Suppose if f : C → C is noncon-

1/n
stant and omits 0 and 1. Then fn (z) = fn (z) omits more and more
points on the unit circle. We can of rescale in the domain so the spher-
ical derivative satisfies kfn′ (0)k∞ → ∞. Passing to a subsequence and
reparameterizing, we obtain in the limit a nonconstant entire function
that omits the unit circle. This contradicts Liouville’s theorem.

26. Classical Proof of Little Picard: The key fact is that the universal cover
of C − {0, 1} can be identified with the upper halfplane.
To see this, it is useful to start by considering the subgroup Γ0 ⊂
Isom(∆) generated by reflections in the sides of the ideal triangle T
with vertices {1, i, −1}. For example, z 7→ z is one such reflection,
sending T to −T . By considering billiards in T , one can see that its
translates tile the disk and thus T is a fundamental domain for Γ0 .
Thus the quadratilateral F = T ∪ (−T ) is a fundamental domain for
the orientation-preserving subgroup Γ ⊂ Γ0 , and the edges of −T are
glued to the edges of T to give a topological triply-punctured sphere as
quotient.
Now let π : T → H be the Riemann mapping sending T to H and
its vertices to {0, 1, ∞}. Developing in both the domain and range by
Schwarz reflection, we obtain a covering map π : ∆ → C b − {0, 1, ∞}.
Given this fact, we lift an entire function f : C → C − {0, 1} to a map
fe : C → H, which is constant by Liouville’s theorem.

27. Uniformization of planar regions. Once we know that C b − {0, 1, ∞} is

uniformized by the disk, it is straightforward to prove that any planar
region U omitting two points is covered by the disk. To do this, we
e → U, and
consider a basepoint p in the abstract universal cover π : U
let F be the family of all holomorphic maps
e , p) → (∆, 0)
f : (U

21
 that are covering maps to their image. Using the uniformization of the
triply-punctured shere, we have that F is nonempty. It is also a closed,
normal family of functions in O(U e ); and by the classical square-root
trick, it contains a surjective function (which maximizes |f ′ (p)|). By
the theory of covering spaces, this extremal map must be bijective.

28. Great Picard Theorem. Near an essential singularity, a meromorphic

b with at most two exceptions.
function assumes all values on C

29. Proof of Great Picard. Consider a loop γ around the puncture of the
disk. If f sends γ to a contractible loop on the triply-punctured sphere,
then f lifts to a map into the universal cover H, which implies by
Riemann’s removability theorem that f extends holomorphically over
the origin.
Otherwise, by the Schwarz lemma, f (γ) is a homotopy class that can
be represented by an arbitrarily short loop. Thus it corresponds to a
puncture, which we can normalize to be z = 0 (rather than 1 or ∞).
It follows that f is bounded near z = 0 so again the singularity is not
essential.

4 Elliptic curves
1. Theorem: Any discrete subgroup of (C, +) is isomorphic to Zn with
n = 0, 1 or 2. In the last case, the generators are linearly independent
over R.

2. Entry 1: Let Λ ⊂ C be a lattice. How can we describe the Riemann

surface X = C/Λ? Answer will be: K(X) ∼ = C(x, y) where y 2 =
4x3 + ax + b. This will involve the construction of functions x(t), y(t)
on C that sweep out or uniformize the an elliptic curve, and are periodic
under Λ.
Compare the curves x2 + y 2 = 1, x2 − y 2 = 1 and xy = 1, which are
isomorphic to C/2πZ or C/2πiZ, and are uniformized by (cos(t), sin(t)),
(cosh(t), sinh(t)) and (et , e−t ) respectively. Note that all 3 curves have,
over C, two asymptotes, corresponding to the ends of C∗ .

3. Entry 1’: let f be a nonconstant meromorphic function on C, and let

Λ ⊂ C be the group of translations that leave f invariant. If Λ is

22
 nontrivial then either Λ ∼
= Z or Λ is a lattice, and in both cases we can
ask to construct all functions with the given periods.
For Λ = 2πiZ the answer is: f (t) = g(exp(t)), where g is meromorphic
on C∗ . For Λ a lattice we will find f (t) = g(℘(t), ℘′ (t)), where g is a
rational function of two variables.

4. Properties of elliptic (doubly-periodic) functions. Such functions can

be considered as holomorphic maps f : X → C, b where X = C/Λ. By
general easy results on compact Riemann surfaces: An entire elliptic
function is constant. The sum of the residues of f over X is zero.
More interestingP is thePfact that, if f is nonconstant with zeros ai and
poles pi , then ai = pi in the group law on X.
Proof: working in a fundamental parallelogram F = [0, λ1 ] × [0, λ2 ] in
C, we have Z
X X zf ′ (z) dz
−1
ai − pi = (2πi) ·
∂F f (z)
The integrals
R over opposite edges cancel, up to a terms of the form
λi (2πi)−1 e f ′ (z)/f (z) dz. Since the f is periodic, it has an integral
winding number N(e) on each edge, and these terms have the form
N(e)λi ∈ Λ.
We will see that we may construct an elliptic function with given zeros
and poles subject only to this constraint.

5. Construction of elliptic functions. For n ≥ 3,

X 1
ζn (z) =
Λ
(z − λ)n

defines an elliptic function of degree n. For degree two, we adjust the

factors in the sum so they vanish at the origin, and obtain the definition
of the Weierstrass function:
1 X′ 1 1
℘(z) = 2 + 2
− 2
.
z Λ
(z − λ) λ

To see ℘(z) is elliptic, use the fact that ℘(z) = ℘(−z) and ℘′ (z) =
−2ζ3 (z) (which implies ℘(z + λ) = ℘(z) + Aλ ).

23
 Other ways to construct elliptic functions of degree two: if Λ is gener-
ated by (λ1 , λ2 ), first sum over one period to get:
∞
X 1 π2
f1 (z) = = ;
−∞
(z − nλ1 )2 λ21 sin(πz/λ1 )2

then ∞
X
f (z) = f1 (z − nλ2 )
−∞

converges rapidly, and defines an elliptic function of degree two. Simi-

larly, if we write X = C∗ / < z 7→ αz >, then
X αn z
f (z) =
(αn z − 1)2

converges to an elliptic function of degree two.

These functions are not quite canonnical; there is a choice Zλ1 of which
direction in the lattice to sum over first. As a consequence they agree
with ℘(z) only up to a constant. This constant is a multiple of the
important quasimodular form G1 we will discuss below.

6. Laurent expansion and differential equation. Expanding 1/(z − λ)2 in

a Laurent series about z = 0 and summing, we obtain
X ∞
1 1
℘(z) = 2 + 3G2 z 2 + 5G3 z 4 + · · · = 2 + (2n + 1)z 2n Gn+1
z z 1

where X′ 1
Gn = Gn (Λ) = .
Λ
λ2n
By a straightforward calculation (using the fact that an elliptic function
with no pole is constant) we have:

℘′ (z)2 = 4℘(z)3 − g2 ℘(z) − g3 ,

where g2 = 60G2 and g3 = 140G3 .

24
7. Geometry of the Weierstrass map ℘ : X → C. b We have ℘(x) = ℘(y)
iff x = ±y. Thus the critical points of ℘, i.e. the zeros of ℘′ , are the
points of order two: {0, c1 , c2 , c3 }.
Since ℘ has degree two, all its critical points are simple and its critical
values, ∞ and ℘(ci ) = ei are distinct. Here ri are the roots of the cubic
equation p(x) = 4x3 − g2 x − g3 .
The map z 7→ (℘(z), ℘′ (z)) sends X − {0} bijectively to the affine cubic
curve y 2 = p(x). Thus ℘ “uniformizes” this plane curve of genus 1.

8. Real case. Now suppose Λ = Zλ1 ⊕ Zλ2 with λ1 > 0 and λ2 ∈ iR+ .
Then Λ is invariant under both negation and complex conjugation.
Thus we have
℘(z) = ℘(−z) = ℘(z).
It follows that ℘ must be real on the locus R ⊂ X which is fixed
under z 7→ z or z 7→ −z. This locus is covered in C by the vertical
and horizontal lines through the points of (1/2)Λ. In particular it the
rectangle S with vertices the points of order two, {0, c1, c2 , c3 }.
It is now easy to check that R = ℘−1 (R ∪ {∞}), and (using the fact
that ℘(z) = 1/z 2 + · · · ) that ℘ is decreasing along [0, λ1 /2] and hence
maps S homeomorphically to −H.
The inverse map on the upper halfplane, f : H → e1 + S, satisfies
f (℘(z)) = z and thus f ′ (℘(z))℘′ (z) = 1. Consequently:
1
f ′ (℘) = (℘′ )−1/2 = p ,
4℘3 − g2 ℘ − g3

which we recognize as the Schwarz-Christoffel formula for the Riemann

map to a rectangle.

9. Function fields. Given any Riemann surface X, the meromorphic func-

tions on X form a field K(X). Then K is a contravariant functor from
category of Riemann surfaces with non-constant maps to the category
b = C(z).
of fields with extensions. Example: K(C)
Theorem: For X = C/Λ, K(X) = C(x, y)/(y 2 − 4x3 + g2 x + g3 ).
To see that ℘ and ℘′ generate K(X) is easy. Any even function f :
X →C b factors through ℘: f (z) = F (℘(z)), and so lies in C(℘). Any

25
 odd function becomes even when multiplied by ℘′ ; and any function is
a sum of one even and one odd.
To see that the field is exactly that given is also easy. It amounts
to showing that K(X) is of degree exactly two over C(℘), and ℘ is
transcendental over C. The first assertion is obvious, and if the second
fails we would have K(X) = C(℘), which is impossible because ℘ is
even and ℘′ is odd.

10. Addition law on an elliptic curve.

Consider the curve E ⊂ P2 defined by y 2 = 4x3 + g2 x + g3 and param-
eterized by the Weierstrass ℘-function via (x, y) = (℘(z), ℘′ (z)).
Theorem. For any line L, the intersection L ∩ E = {a, b, c} where
a + b + c = 0 on E = C/Λ.
Proof. The line L is simply the zero set of A℘′ + B℘ + C for some
(A, B, C). This function has all its poles at z = 0. Since the sum of
the zeros and poles is zero, its zeros (a, b, c) also sum to zero.
Cor. The map p 7→ −p on E is given by (x, y) 7→ (x, −y).
Proof. Then the line passes through ∞ which is the origin of E, con-
sistent with the equation p + (−p) + 0 = 0.
Cor. The point c = a + b is constructed geometrically by drawing
the line L through (a, b), finding its third point of intersection (−c) =
(x, −y) on E and then negating to get c = (x, y).
Cor. To construct the point 2a, take the line tangent to E at a, find
its other point of intersection and then negate.

11. An elliptic function with given poles and zeros. It is natural to try to
construct an elliptic function by forming the Weierstrass product for
the lattice Λ:
Y′   
z z z2
σ(z) = z 1− exp + .
Λ
λ λ 2λ2

Then (log σ)′′ = −℘(z), from which it follows that σ(z+λ) = σ(z) exp(aλ +
bλ z). From this it is easy to see that
σ(z − a1 ) . . . σ(z − an )
σ(z − p1 ) . . . σ(z − pn )

26
 P P
defines an elliptic function whenever ai = pi , and therefore this
is the only condition imposed on the zeros and poles of an elliptic
function.
12. The moduli space of complex tori, M1 , classifies Riemann surfaces of
genus 1 up to isomorphism.
Since any elliptic curve has the form X = C/Λ, we see that M1 is
the same as the lattice in C modulo the action of C∗ . If a lattice
Λ = Zλ1 ⊕ Zλ2 is given an oriented basis, then it is similar to Z ⊕ Zτ
where τ ∈ H. Thus H is the space of similarity classes of framed
lattices.
Any two oriented bases of Λ ∼
= Z2 are related by an element of SL2 (Z).
That is, if
Z ⊕ Zτ1 = α(Z ⊕ Zτ2 )
with α ∈ C∗ , then we can write
! !
τ1 aτ2 + b
=α
1 cτ2 + d
for some g = ( ac db ) in SL2 (Z), and thus
τ1 = (aτ + 2 + b)/(cτ2 + d).
(The fact that ad − bc = 1 instead of −1 insures that Im τ1 > 0 if
Im τ2 > 0.) Conversely, any two points of H in the same orbit of
SL2 (Z) give similar lattices and hence isomorphic complex tori. This
shows:

Theorem 4.1 The moduli space M1 is naturally isomorphic to the

complex orbifold H/ SL2 (Z).

Now we can always normalize a lattice Λ by C∗ so that its shortest

nonzero vector is z = 1 and its next shortest vector is τ ∈ H. Then
|τ | ≥ 1, and Z + τ ⊂ Λ so | Re τ | ≤ 1/2. Moreover Z ⊕ Zτ = Λ;
otherwise there would be a lattice point v = a + bτ with a, b ∈ [0, 1/2],
but then
|v|2 ≤ 1/4 + |τ |2 /4 ≤ 1/2
contrary to our assumption that the shortest vector in Λ∗ has length
1. This shows:

27
 Theorem 4.2 The region | Re τ | ≤ 1/2, |τ | > 1 in H is a fundamental
domain for SL2 (Z).

13. To study M1 further, we now introduce the space of cross-ratios

Mord b
0,4 = C − {0, 1, ∞}

and its quotient orbifold

M0,4 = Mord
0,4 /S3 .

Here Mord
0,4 is the moduli space of ordered quadruples of distinct points
b up to the action of Aut(C).
on C, b Any such quadruple has a unique
representative of the form (∞, 0, 1, λ), giving a natural coordinate for
this moduli space.
If we reorder the quadruple, the cross-ratio changes, ranging among the
six values

{λ, 1/λ, 1 − λ, 1/(1 − λ), λ/(λ − 1), (λ − 1)/λ}.

(There is a natural action of S4 , but the Klein 4-group Z/2 × Z/2 acts
trivially.) There are 5 points fixed under the action of S3 : the points
−1, 1/2 and 2 each have stabilizer Z/2, and correspond to the vertices
of a square; while the points
√
±ω = 1/2 ± −3/2

have stabilizer Z/3 and correspond to the vertices of a tetrahedron.

The degree 6 rational map
4 (λ2 − λ + 1)3
F (λ) =
27 λ2 (1 − λ)2
is invariant under S3 , and gives a natural bijection
b − {0, 1, ∞})/S3 ∼
F : (C =C
satisfying F (0) = F (1) = F (∞) = ∞, F (ω) = F (ω) = 0, and F (−1) =
F (1/2) = F (2) = 1 (hence the factor for 4/27). We should really think
of the image as M0,4 and in particular remember the orbifold structure:
Z/2 at F = 1 and Z/3 at F = 0.

28
 b Let Z/n act
14. Orbifolds. Here are some other examples of quotients of C.
on Cb by z 7→ exp(2πik/n)z. Extend this to an action of the dihedral
group Dn by adding in z 7→ 1/z. Then the quotient in both cases
b again: the (n, n)-orbifold and the (2, 2, n)-orbifold respectively.
is C
The quotient map can be given by F (z) = z n in the first case and by
f (z) = z n + z −n in the second.
15. The modular function. We now define a map J : M1 → M0,4 by asso-
ciating to any complex torus, the four critical values of the Weierstrass
℘-function. (Note: any degree two map f : X = C/Λ → C b is equiva-
lent, up to automorphisms of domain and range, to the Weierstrass ℘
function, so the associated point in M0,4 is canonically determined by
X.)
More concretely, given τ ∈ H we define the half-integral points (c1 , c2 , c3 ) =
(1/2, τ /2, (1+τ )/2) and the corresponding critical values by ei = ℘(pi ).
Then their cross-ratio (together with the critical value at infinity) is
given by:
e3 − e2
λ(τ ) = ·
e1 − e2
This depends on an ordering of the points of order two. Now SL2 (Z) =
Aut(Λ) acts on the three points of order two through the natural quo-
tient
0 → Γ(2) → SL2 (Z) → SL2 (Z/2) → 0.
In particular, λ is invariant under the subgroup Γ(2) of matrices equiv-
alent to the identity modulo two.
Now any elliptic element in SL2 (Z) has trace −1, 0 or 1, while the trace
of any element in Γ(2) must be even. Moreover, trace zero cannot arise:
b ) ∈ Γ(2) then −a2 − bc = 1 implies −a2 = 1 mod 4 which is
if g = ( ac −a
impossible.
Summing up, we have:

Theorem 4.3 The group Γ(2) is torsion-free, with fundamental do-

main the ideal quadratilateral with vertices {∞, −1, 0, 1}.

To remove the ambiguity of ordering, we now define

4 λ2 − λ + 1)3
J(τ ) = F (λ(τ )) = ·
27 λ2 (1 − λ)2

29
Theorem 4.4 The map
b − {0, 1, ∞})/S3
J : M1 = H/ SL2 (Z) → M0,4 = (C

is a bijection, and a isomorphism of orbifolds.

Proof. The value of J(τ ) determines a quadruple B ⊂ C b which in turn

determines a unique Riemann surface X → C b of degree two, branched
over B, with X ∼= C/Z ⊕ τ Z. Thus J(τ1 ) = J(τ2 ) iff the corresponding
complex tori are isomorphic iff τ1 = g(τ2 ) for some g ∈ SL2 (Z). Thus
J is injective.
To see it is surjective, we first observe that J(τ + 1) = J(τ ). Now if
Im τ = y → ∞, then on the region | Im z| < y/2 we have

1 X′ 1 1 π2
℘(z) = 2 + − + ǫ(z) = + C + ǫ(z),
z (z − n)2 n2 sin2 (πz)

where ǫ(z) → 0 as y → ∞. (In fact we have

X′ π2
ǫ(z) = = O(e−πy ).
n
sin2 (π(z + nτ ))

We really only need that it tends to zero; and we will not need the
exact value of the constant C = −π 2 /3).
Since sin(z) grows rapidly as | Im z| grows, we have

(e1 , e2 , e3 ) = (π 2 + C, C, C) + O(ǫ)

and thus λ(τ ) = (e3 − e2 )/(e1 − e2 ) → 0 and hence J(τ ) → ∞.

Thus J is a proper open map, which implies it is surjective (its image
is open and closed).
For the orbifold assertion, just note that the square torus has J(τ ) = 1
and the hexagonal torus has J(τ ) = 0; symmetries of X are reflected
in symmetries of the 4-tuple (e1 , e2 , e3 , ∞).

30
16. Picard’s theorem revisited. It is now straightforward, by the theory of
covering spaces, to show:

Theorem 4.5 The map

λ : H/Γ(2) → Mord b
2 = C − {0, 1, ∞}

is a bijection, presenting H as the universal cover of the triply-punctured

sphere.

This gives the one line proof of the Little Picard Theorem: ‘consider
λ−1 ◦ f : C → H.’ Cf. [Bol, p.39–40]:

The question recently arose in conversation whether a dis-

sertation of 2 lines could deserve and get a Fellowship... in
mathematics the answer is yes....
(Theorem.) An integral function never 0 or 1 is a constant.
(Proof.) exp{iΩ(f (z))} is a bounded integral function.
...But I can imagine a referee’s report: ‘Exceedingly striking
and a most original idea. But, brilliant as it undoubtedly
is, it seems more odd than important; an isolated result,
unrelated to anything else, and not likely to lead anywhere.’

Here Ω is our λ−1 .

17. Periods. The lattice Λ associated to 4 points can also be covered ex-
plicitly (up to similarity) as follows.
Let f : X = C/Λ → C b be any degree two map, with branch locus
B ⊂ C, b such that f (z) = f (−z). Note that the quadratic differential
2
dz on X = C/Λ is invariant under z 7→ −z. Thus it descends to
b It has poles at the branch
a well-defined differential q(w) dw 2 on C.
points B, and these four points uniquely determine q up to scale.
In the case w = ℘(z), we have dw 2 = ℘′ (z)2 dz 2 and hence

dw 2 dw 2
q(w) dw 2 = = ·
℘′ (z)2 4w 3 − g2 w − g3

31
 Now we have a map π1 (C b − B) → Z/2 defining the covering X. Let K
√
be the kernel of this map. Then q has a single-valued branch along
any [γ] ∈ K, and [γ] lifts to a loop δ ∈ π1 (X). Consequently:
Z Z
√
q = ± dz ∈ Λ.
γ δ

All elements of Λ arise in this way, and thus Λ can be recovered as the
√
periods of q.
18. Modular forms: algebraic perspective. We now wish to classify holo-
morphic k-forms ω = ω(z) dz k on the orbifold
J
= M0,4 ∼
M1 ∼ = C,

which are ‘holomorphic’ at the orbifold and cusp points. The set of all
such ω forms the finite-dimensional vector space Mk , called the space
of modular forms of weight 2k.
By removable singularities, ω can be considered an ordinary rational
b The condition that it is holomorphic at an orbifold point
k-form on C.
then becomes the following.

Theorem 4.6 At an orbifold point of order n, a holomorphic k-form

has a pole, in classical coordinates, of order at most k(1 − 1/n).

Proof. If the orbifold point is z = 0, we want the form z i dz k to pull

back to a holomorphic form under z = w n . The pullback is given by
w ni(nw n−1 dw)k = nw ni+nk−k dw k , which is holomorphic iff ni+nk−k ≥
0 iff −i ≤ k(1 − 1/n).

When z = 0 is a puncture of the Riemann surface it is like an orbifold

point of order infinity, and the result above formally requires a pole
of order at most k for holomorphicity. This fits well with the local
uniformization z = f (t) = exp(2πit). Under this map we have

f ∗ ((dz/z)k ) = (2πi)k (dt)k ,

so a pole of order k is converted into a function which is ‘constant and

nonzero’ as Im t → ∞ in the uniformizing upper halfplane.

32
Definition. A rational form ω(z) dz k is a modular form of weight 2k
on M0,4 ∼
= C if it has poles of order at most k, k/2 and 2k/3 at z = ∞,
z = 1 and z = 0.
We let Mk denote the finite-dimensinoal space of all modular forms of
weight 2k. The total number of poles of ω as a classical form is 2k.
The allowed orders of zeros and poles, and the dimensions of Mk , are
given in Table 1 for small values of k.

k Z/∞ Z/3 Z/2 Total 2k dim Mk

0 0 0 0 0 0 1
1 1 0 32 0 21 1 2 0
2 2 1 31 1 4 4 1
3 3 2 1 21 6 6 1
4 4 2 31 2 8 8 1
5 5 3 31 2 21 10 10 1
6 6 4 3 13 12 2
7 7 4 31 3 21 14 14 1
8 8 5 31 4 17 16 2
9 9 6 4 21 19 18 2
10 10 6 31 5 21 20 2
11 11 7 31 5 21 23 22 2
12 12 8 6 26 24 3

Table 1. Allowed poles and spaces of modular forms.

Examples:
b with just a single pole, so
(a) There is no meromorphic 1-form on C
dim M1 = 0.
(b) The quadratic differential F2 = dz 2 /(z(z − 1)) spans M2 . Note
that this differential has a simple zero, in the orbifold sense, at
z = 0. It is covered by the differential on the triply-punctured
spheres with double poles of equal residue at each puncture.

33
 (c) The cubic differential F3 = dz 3 /(z 2 (z − 1)) spans M3 . This time
there is a zero of order 2 at z = 1.
(d) The products F22 and F3 F2 span M4 and M5 .
(e) The forms F23 and F32 span M6 , which is two-dimensional.
(f) The discriminant
dz 6
D6 = F23 − F32 = 4 ∈ M6
z (z − 1)3
vanishes at infinity, but has no others zeros or poles (in the orbifold
sense).
(g) Ratios of forms of the same weight give all rational functions.
Indeed, we have
z = F23 /D6 ,
and thus any rational function of degree d can be expressed as a
ratio of modular forms of degree k = 6d.
19. Cusp forms. We let Mk0 ⊂ Mk denote the space of cusp forms vanishing
at infinity. Every space Mk , k ≥ 2 contains a form of the type F2iF3j
which is not a cusp form; thus Mk ∼
= C ⊕Mk0 . By inspection, dim Mk0 =
0 for k ≤ 5. On the other hand, any cusp form is divisible by D6 . This
shows:

Theorem 4.7 The map ω 7→ D6 ω gives an isomorphism from Mk to

0
Mk+6 .

Corollary 4.8 We have dim M6n+1 = n, and dim M6n+i = n + 1 for

i = 0, 2, 3, 4, 5.
S
Corollary 4.9 The forms F2 and F3 generate the ring M = Mk .

Corollary 4.10 The forms F2i F3j with 2i + 3j = k form a basis for
Mk .

Proof. By the preceding Corollary these forms span Mk , and the

number of them agrees with dim Mk as computed above.

34
 Corollary 4.11 The ring of modular forms M is isomorphic to the
polynomial ring C[F2 , F3 ].

Proof. The preceding results shows the natural map from the graded
ring C[F2 , F3 ] to M is bijective on each graded piece.

20. Modular forms: analytic perspective. Classically, a modular form of

weight 2k is a holomorphic form f = f (z) dz k on the upper halfplane
which is invariant under the modular group SL2 (Z), and which is ‘holo-
morphic at infinity’.
In terms of the function f (z), the invariance conditions are equivalent
to:
f (z + 1) = f (z) and f (−1/z) = z 2k f (z).
The first condition implies we can write
∞
X
f (z) = an q n ,
−∞

where q(z) = exp(2πiz); and the condition that f is ‘holomorphic at

infinity’ means an = 0 for n < 0. In other words, if we regard f (q) as
a function on H/Z ∼ = ∆∗ , it is holomorphically at q = 0. If f (q) = 0,
then f is a cusp form.
Note that dz = (2πi)−1 dq/q. Thus as a form on ∆∗ , we have f =
(2πi)−k (dq/q)k , and f has a pole of order at most k at q = 0.
Examples.

(a) Any holomorphic invariant f (Λ) of lattices that is homogeneous

of degree −2k and satisfies the right growth conditions defines a
modular form of weight 2k. That is, if f (tΛ) = t2k Λ, then the
function
F (τ ) = f (Z ⊕ τ Z)
satisfies F (τ + 1) = F (τ ) and

F (−1/τ ) = f (Z ⊕ τ −1 )Z) = f (τ −1 (Z ⊕ τ Z) = τ 2k F (τ ).

35
 (b) In particular, for k ≥ 2 the Eisenstein series
X′
Gk (τ ) = (n + mτ )−2k
are modular forms of weight 2k. Evidentally Gk converges to
2Z(2k) as Im τ → ∞, so these are holomorphic at infinity but not
cusp forms.
(c) Similarly for the normalized functions g2 = 60G2 , g3 = 140G3 .
Theorem 4.12 Under the isomorphism M1 ∼ = M0,4 , an analytic mod-
ular form is the same as an algebraic modular form.
Corollary 4.13 The ring of modular forms is isomorphic to C[g2 , g3 ].
21. Values of g2 and g3 . We now note that for τ = i, the zeros of 4℘3 −
g2 ℘−g3 must look like (−1, 0, 1) and thus g3 (i) = 0. Similarly for τ = ω
the zeros are arrayed like the cube roots of unity and hence g3 (ω) = 0.
To determine the values at infinity, we observe that as τ → ∞ we have
G2 (τ ) → 2ζ(4) = π 4 /45
and
G3 (τ ) → 2ζ(6) = π 6 /945.
This gives the values g2 (∞) = 60G2 (∞) = 4π 4 /3 and g3 (∞) = 140G3 (∞) =
(4/27)π 6, and thus
(g23 /g32)(∞) = 27.
22. The cusp form ∆ and the modular function J. By the preceding cal-
culation, the discriminant
∆(τ ) = g23 (τ ) − 27g3(τ )2
is a cusp form of weight 12. We have seen such a form is unique up to
a scalar multiple, and is nonvanishing everywhere except for a simple
zero at infinity.
Theorem 4.14 We have J(τ ) = g23(τ )/∆(τ ).

Proof. First note that this is a ratio of forms of weight 12 and hence
a modular function, i.e. it is invariant under SL2 (Z). Since g2 (∞) 6= 0,
it has a simple pole at infinity, and thus has degree one on M1 . We
also have J(ω) = 0 since g2 (ω) = 0, and J(i) = 1 since g3 (i) = 0.

36
 23. The product expansion: for q = exp(2πiτ ), we have
∞
Y
∆(q) = q (1 − q n )24 .
1

References
[BD] F. Berteloot and J. Duval. Sur l’hyperbolicité de certain
complèmentaires. Enseign. Math. 47(2001), 253–267.

[Bol] B. Bollabas, editor. Littlewood’s Miscellany. Cambridge University

Press, 1986.

[CJ] L. Carleson and P. Jones. On coefficient problems for univalent func-

tions. Duke Math. J. 66(1992), 169–206.

37