Hardy Spaces in One Complex Variable

HARDY SPACES IN ONE COMPLEX VARIABLE
prof. Fulvio Ricci
A.A. 2004-2005
1
2
Contents
Chapter I Hardy spaces on the unit disc

1. Definition and basic properties
2. Harmonic versus holomorphic functions
3. Poisson integrals
4. Functions in hp -spaces and their limits to the boundary
5. Boundary limits of conjugate harmonic functions
6. The Cauchy projection
7. Blaschke products and the F. and M. Riesz theorem
8. Dual spaces
Chapter II Hardy spaces on the half-plane

1. Definitions and basic facts
3. The Fourier transform and the Paley-Wiener theorem
4. The Cauchy projection and the Hilbert transform
5. Transference of H p -functions and applications
Chapter III Pointwise convergence to the boundary and conjugate harmonic func-
tions in hp
1. The Hardy-Littlewood maximal function
2. Poisson integrals and maximal function
3. Pointwise convergence to the boundary
4. Poisson integrals of singular measures
5. Lp -estimates for the conjugate harmonic function
THE UNIT DISC 3
CHAPTER I
HARDY SPACES ON THE UNIT DISC
1. Definition and basic properties
We begin by presenting the main properties of Hardy spaces on the unit disc
D = {z ∈ C : |z| < 1} .
We shall usually prefer to denote by T, rather than ∂D or similar, the boundary

of D. So
(1.1) T = {z ∈ C : |z| = 1} = {eit : t ∈ R/2πZ} .
The reason is that emphasis will be put on the group structure of T. The natural
identification between T and R/2πZ (both algebraic and topological) will be always
assumed. Hence functions defined on T will be identified with functions on R/2πZ,
i.e. with functions on the line, periodic of period 2π.
Integrals on T will be understood with respect to the normalized Lebesgue mea-
1
sure 2π dt. We shall use as alternative notation for an integral on T any of the
following1 :
Z Z π Z Z π
it 1 it 1
f (e ) dt , f (e ) dt , f (t) dt , f (t) dt
T 2π −π T 2π −π
The spaces Lp (T) must be intended w.r. to the normalized Lebesgue measure.
The 2-dimensional Lebesgue measure on C will be denoted by dz. Hence, in the
polar coordinates z = reit ,
dz = r dr dt .
This may cause some confusion in the occasions where we shall use line integrals
in C, Z I
f (z) dz or f (z) dz ,
γ γ
in which case the symbol dz denotes a linear differential form. However, the meaning
of the symbol dz will be revealed in each case by the domain of integration.
1 Tobe precise, in the first two integrals T is identified with the unit circle, in the last two with
R/2πZ. We may switch from one notation to the other, omitting explicit reference to composition
with the map t 7→ eit .
Typeset by AMS-TEX
4 CHAPTER I
Let f (z) be a holomorphic function on D. Given r ∈ [0, 1) and p ≥ 1, define

Z 1/p
it p
(1.2) Mp (f, r) = |f (re )| dt ,
T
and also
(1.3) M∞ (f, r) = max

it
|f (reit )| .
e ∈T
If we set
fr (eit ) = f (reit )
for 0 ≤ r < 1, we can then say that
Mp (f, r) = kfr kp .
Definition2 . Let 1 ≤ p ≤ ∞. We denote by H p (D) the space of holomorphic

functions f on D such that
(1.4) sup Mp (f, r) = kf kH p < ∞ .

0≤r<1
It is easy to check that (1.4) defines a norm. For the implication kf kH p =

0 ⇒ f = 0, one has to observe that any f ∈ H p (D) is continuous on D, so that
kfr kp = 0 ⇒ fr = 0, if r < 1.
It is quite obvious that H ∞ (D) consists of the bounded holomorphic functions
on D.
Proposition 1.1. If 1 ≤ p < q ≤ ∞, then H q (D) ⊂ H p (D), and, for f ∈ H q (D),
kf kH p ≤ kf kH q .
Proof. The inequality kfr kp ≤ kfr kq follows easily from Hölder’s inequality if q <
∞, and from the trivial majorization if q = ∞. Taking suprema in r, the inequality
is preserved.
All these inclusions are proper. Interesting examples in this respect are the
functions
1
fα (z) = ,
(1 − z)α
with3 α > 0. Given p, we want to determine the values of α for which fα is in
H p (D). It is obvious that fα 6∈ H ∞ (D) for α > 0, so we take p < ∞.
The answer is based on the following lemma4 .
2 We define Hardy spaces only for p ≥ 1. The definition makes perfect good sense also for
p < 1, except that in this case (1.4) does not define a norm. H p -spaces with p < 1 have many
intereseting features, that we will not discuss in this course.
3 It is possible to choose a determination of the α-power of 1 − z on D for every α ∈ C,
bacause 1 − z does not vanish on D and D is simply connected. The “principal” determination is
(1 − z)α = eα log(1−z) , with arg(1 − z) ∈ (−π/2, π/2) (observe that <e(1 − z) > 0 for z ∈ D).
4 We write f g for r → 1 to denote that the ratio |f /g| is bounded from above and from
below by positive constants on a neighborhood of 1.

THE UNIT DISC 5
Lemma 1.2. For s > 0 fixed, consider the integral

Z π
1
Is (r) = it s
dt ,
−π |1 − re |
as a function of r ∈ [0, 1). Then, for r → 1,

(1) if s < 1, Is (r) 1 ;
(2) if s = 1, I1 (r) | log(1 − r)| ;
(3) if s > 1, Is (r) (1 − r)−(s−1) .
Proof. Take r > 12 . Considering that the triangle with vertices 1, r and reit is obtuse
in r, and that | sin θ| ≥ π2 |θ| for θ ∈ [−π/2, π/2], we have that, for t ∈ [−π, π],

|1 − reit | > max r|1 − eit |, 1 − r
11
≥ |1 − eit | + 1 − r
2 2
1 t
(1.5) = sin + 1 − r
2 2

1 1
≥ |t| + 1 − r
2 π
1
≥ (|t| + 1 − r) .
2π
We also have
|1 − reit | ≤ |1 − eit | + |eit − reit | = |1 − eit | + 1 − r ≤ |t| + 1 − r ,
so that Z π
1
Is (r) dt ,
−π (|t| + 1 − r)s
for every s > 0. Then,
Z π Z π
1 1
s
dt = 2 s
dt
−π (|t| + 1 − r) 0 (t + 1 − r)

2 log(1 − r + π) − log(1 − r) if s = 1 ,
= 2
1−s
(1 − r + π)1−s − (1 − r)1−s if s 6= 1 .
The conclusion follows easily.

Proposition 1.3. fα ∈ H p (D) if and only if αp < 1.
Proof. Observe that
1 1/p
Mp (fα , r) =
Iαp (r) .
2π
It follows from Lemma 1.2 that Mp (fα , r) is bounded in r if αp < 1, and
lim Mp (fα , r) = +∞
r→1
if αp ≥ 1.
We show now that H p (D) is complete. A couple of preliminary facts have their
own independent importance.
6 CHAPTER I
Lemma 1.4. Let f ∈ H p (D). Then, for every z ∈ D,

kf kHp
|f (z)| ≤ Cp .
(1 − |z|)1/p
Proof. The statement is obvious for p = ∞, so we assume that p is finite.

Take r with |z| < r < 1, and let γr be the circle centered at the origin and radius
r, oriented counterclockwise. Then
I
1 f (w)
f (z) = dw .
2πi γr w − z
Setting w = reit , we have dw = ireit dt, so that
Z π
1 f (reit )
f (z) = dt .
2π −π 1 − zr e−it
If z = |z|eiθ , this becomes
Z π
1 f (reit )
f (z) = |z| i(θ−t)
dt .
2π 1−
−π r e
By Hölder’s inequality,
Z π 1/p0
1 1
|f (z)| ≤ Mp (f, r)
|z| i(θ−t) p0
dt ,
2π 1 −
r e
−π
where p0 is the dual exponent of p.

Making the change of variable θ−t = u and using the periodicity of the integrand,
we find that
Z π Z π |z|
1 1

|z| i(θ−t) p0
dt =
|z| iu p0
du = I p 0 .
−π 1 − e −π 1 − e r
r r
By Lemma 1.2, since p0 > 1,

1 |z| 1/p0
|f (z)| ≤ kf kH pIp0
2π r
|z| − p0
p0 −1
≤ Cp kf kH p 1 −
r
|z| −1/p
= Cp kf kH p 1 − .
r
Letting r → 1, we conclude the proof.
Corollary 1.5. Convergence in H p (D) implies uniform convergence on compact
subsets of D.
Proof. Let K ⊂ D be compact. Then there is r < 1 such that K ⊂ Dr , the closed
disc of radius r centered at the origin. By Lemma 1.4, if fn → f in H p (D),
kfn − f kH p
kfn − f k∞,K = max |fn (z) − f (z)| ≤ Cp ,
z∈K (1 − r)1/p
also tends to zero.
THE UNIT DISC 7
Theorem 1.6. H p (D) is a Banach space.

Proof. Let {fn } be a Cauchy sequence in H p (D). By Lemma 1.4, for every r < 1
{fn } is a Cauchy sequence in C(Dr ), with the uniform norm. By the completeness
of C(Dr ), there is gr ∈ C(Dr ) such that fn → gr uniformly on Dr .
Obviously, if r1 < r2 < 1, gr1 and gr2 coincide on Dr1 . It follows that the various
gr , with r < 1, are all restrictions of a unique function g continuous on D.
We now prove that g is holomorphic in D. By Morera’s theorem, this is true if
and only if for every closed arc γ in D,
I
g(z) dz = 0 .
γ
Let γ be such an arc. Since γ is contained in Dr for some r < 1, fn → g

uniformly on γ. Therefore
I I
g(z) dz = lim fn (z) dz ,
γ n→∞ γ
and each of these integrals is zero because the fn are holomorphic.

We finally prove that fn → g in H p (D). Given ε > 0, let n̄ be such that
kfn − fm kH p < ε for n, m ≥ n̄. Take r < 1. Since fm → g uniformly on the circle
|z| = r, if n ≥ n̄,
Mp (fn − g, r) = lim Mp (fn − fm , r) ≤ lim kfn − fm kH p ≤ ε .

m→∞ m→∞
Since this holds for every r < 1, kfn − gkH p ≤ ε.
2. Harmonic versus holomorphic functions
A C 2 -function u defined on an open set Ω ⊆ Rn is called harmonic on Ω if its

Laplacian ∆u, defined as
Xn
∆u = ∂x2j u
j=1
is identically zero on Ω. A holomorphic function f is harmonic on its domain in

R2 . This follows from the fact that holomorphic functions are C 2 (in fact analytic)
and from the Cauchy-Riemann equation
1
∂¯z f = ∂x f + i∂y f ) = 0 .
2
Differentiating in x, we find that
∂x2 f = −i∂x ∂y f = −∂y2 f .
Since ∆ is real, if u is harmonic, so are ū, <eu and =mu. In particular, anti-
holomorphic functions are also harmonic.
Harmonic functions are characterized by the mean value property. Let S n−1 be
the unit sphere in Rn , and dσ the surface measure on it.
8 CHAPTER I
Definition. A continuous function u satisfies the mean value property on Ω if, for
any x ∈ Ω and any r > 0 such that the closed ball B(x, r) is contained in Ω,
Z
1
(2.1) u(x) = u(x + ry) dσ(y) .
σ(S n−1 ) S n−1
In two dimensions, (2.1) becomes

Z
u(z) = u(z + reit ) dt .
T
Theorem 2.1. Let u be a continuous function on Ω. Then u is harmonic if and

only if the mean value property holds.
Proof. Suppose u is C 2 in Ω. Given x ∈ Ω, consider the function
Z
1
ϕ(r) = u(x + ry) dσ(y) ,
σ(S n−1 ) S n−1
defined for r ∈ [0, r0 ), where r0 is the radius of the largest ball centered at x and
contained in Ω. It is easy to verify that ϕ is continuous, ϕ(0) = u(x) and that
Z
0 1 ∂
ϕ (r) = u(x + ry) dσ(y) .
σ(S n−1 ) S n−1 ∂r
For any r, the last integral can be written as an integral on the boundary of the
ball B(x, r), in terms of the surface measure dσr on it:
Z Z
∂ 1
u(x + ry) dσ(y) = n−1 ν · ∇u(y) dσr (y) ,
S n−1 ∂r r ∂B(x,r)
where ν denotes the outer normal to ∂B(x, r).

By Green’s formula, we obtain that
Z
0 1
ϕ (r) = ∆u(y) dy .
σ(S n−1 )r n−1 B(x,r)
It follows that, if u is harmonic, then ϕ0 = 0, hence ϕ is identically equal to u(x),

which proves the mean value property.
2
Conversely,
R if u is C and satisfies the mean value property, the same argument
shows that B(x,r) ∆u(y) dy = 0 for any ball B(x, r) ⊂ Ω. This implies that u is
harmonic.
It remains to prove that if u is only continuous on Ω and satisfies the mean value
property, then it is C 2 . Take a ball B(x0 , r) ⊂ Ω, and choose a C 2 -function ψ(x) on
Rn , which is radial, non-negative, non-identically zero, and supported on B(0, r/2).
Define
Z
(2.2) v(x) = u(x + y)ψ(y) dy ,
B(0,r/2)
THE UNIT DISC 9
which is well-defined for x ∈ B(x0 , r/2). Since ψ is radial, we can consider its
“profile” ψ0 , defined on the positive half-line so that ψ(x) = ψ0 (|x|). We then
have, integrating in polar coordinates:
Z r
2
Z
v(x) = u(x + ρw)ψ0 (ρ)ρn−1 dσ(w) dρ
0 S n−1
Z r Z
2
n−1
= ψ0 (ρ)ρ u(x + ρw) dσ(w) dρ
0 S n−1
Z r
2
n−1 n−1
= σ(S ) ψ0 (ρ)ρ dρ u(x) .
0
Since the expression in parentheses is positive, we conclude that v is a non-zero

constant multiple of u on B(x0 , r/2). On the other hand, changing variable of
integration in (2.2), we find that, for x ∈ B(x0 , r/2),
Z Z
v(x) = u(y)ψ(y − x) dy = u(y)ψ(y − x) dy
B(x,r/2) B(x0 ,r)
(the last equality takes into account the fact that ψ = 0 in the extra part of
the domain of integration). This last expression shows, by differentiating under
integral sign, that v is C 2 on B(x0 , r/2). The same is then true for u, and, since
x0 is arbitrary, u is C 2 on Ω.
Corollary 2.2. Harmonic functions satisfy the maximum modulus principle: if u
is harmonic on a connected open set Ω, and it attains its maximum modulus at a
point in Ω, then it is constant.
Proof. Suppose that x0 ∈ Ω is such that |u(x0 )| = maxx∈Ω |u(x)|. By replacing u
by eiθ u for an appropriate θ, we can assume that u(x0 ) is real and non-negative.
Let B(x0 , r) ⊂ Ω. Since <eu is harmonic,
u(x0 ) = <eu(x0 )
Z
1
= <eu(x0 + ry) dσ(y) .
σ(S n−1 ) S n−1
Therefore
Z
1
u(x0 ) − <eu(x0 + ry) dσ(y) = 0 ,
σ(S n−1 ) S n−1
with a continuous integrand u(x0 ) − <eu(x0 + ry) ≥ 0. Therefore u(x0 ) = <eu(x0 +

ry) for every y ∈ S n−1 . Since |u(x0 + ry)| ≤ u(x0 ), we conclude that u(x0 + ry) =
u(x0 ).
This proves that u is constant on a neighborhood of x0 . It follows that the set
of points x ∈ Ω where u(x) = u(x0 ) is both closed and open. Since Ω is connected,
this set is all of Ω.
We shall now restrict ourselves to n = 2, and focus our attention on the relations
between harmonic and holomorphic functions. We shall need the following lemma
of Fourier analysis.
10 CHAPTER I
Lemma 2.3. Let f be a C 2 -function on T, and denote by fˆ(n), with n ∈ Z its

Fourier coefficients. Then
n2 |fˆ(n)| ≤ kf kC 2 ,
and the Fourier series of f converges to f uniformly on T.
Proof. Integrating by parts twice,
Z Z
2ˆ 2 −int
n f (n) = n f (t)e dt = − f 00 (t)e−int dt .
T T
Therefore,
n2 |fˆ(n)| ≤ kf 00 k∞ ≤ kf kC 2 .
By the Weierstrass test5 , it follows that the series
X
fˆ(n)eint
n∈Z
is uniformly convergent on T. If g is its sum, integration term by term shows that

ĝ(n) = fˆ(n) for every n. This implies that g = f .
Theorem 2.4. Let u be harmonic on a neighborhood of the closed disc B(z 0 , R) ⊂
C. Then u admits a power series expansion
∞
X ∞
X
n
(2.3) u(z) = a0 + an (z − z0 ) + a−n (z − z0 )n ,
n=1 n=1
uniformly convergent on B(z0 , R). In particular, u is real-analytic, and it is the

sum of a holomorphic and an anti-holomorphic function.
Proof. We can assume for simplicity that z0 = 0 and R = 1, so that B(0, R) = D.
The restriction of u to T is C ∞ , hence its Fourier coefficients an form an absolutely
summable sequence. The function
∞
X ∞
X
it n
v(re ) = an z + a−n z̄ n
n=0 n=1
is uniformly convergent on D and harmonic in the interior. Since v = u on the

boundary, the maximum principle implies that v = u in D.
Corollary 2.5. Let u be harmonic on a connected and simply connected domain
Ω. Then there are holomorphic functions f and g on Ω, unique up to additive
constants, such that u = f + ḡ.
Proof. By Theorem 2.4, any z ∈ Ω has a spherical neighborhood Uz where u = fz +
gz , with fz , gz holomorphic. If Uz ∩ Uz 0 6= ∅, on the intersection fz − fz 0 = gz 0 − gz .
Since the left-hand side is holomorphic and the right-hand side anti-holomorphic,
this implies that
fz − fz 0 = gz 0 − gz = cz,z 0
5 The
Weierstrass test says that if the functions fn satisfy inequalities |fn (x)| ≤ Mn on a set
P P
E, with Mn ≥ 0 and Mn < ∞, then the series fn converges uniformly on E.
THE UNIT DISC 11
is constant. Therefore fz0 = fz0 0 on Uz ∩ Uz 0 , so that all these derivatives can be

unified into a unique holomorphic function h on Ω. Since Ω is simply connected,
h admits a primitive f , holomorphic on all of Ω. Then fz − f = bz is constant on
each Uz . On each Uz ,
u − f = g z + fz − f = g z + bz ,
showing that u − f is anti-holomorphic on Ω. Setting g = u − f we have that
u = f + ḡ. If f1 and g1 are holomorphic and u = f1 + g1 on Ω, we repeat the
argument given before: f − f1 = g1 − g must be constant.
Corollary 2.6. Let u be a real-valued harmonic function on a connected and sim-
ply connected domain Ω. There is a unique, up to additive constants, real-valued
harmonic function ũ on Ω such that f = u + iũ is holomorphic.
Proof. According to Corollary 2.5, we can write u = 12 (f +ḡ), with f, g holomorphic.
Since u is real, =mf = =mg. Then f −g is a real-valued holomorphic function, hence
it is constant. Adjusting, if necessary, f and g by adding appropriate constants, we
can assume that f = g. So u = <ef , and we then take ũ = =mf .
If v is another real-valued harmonic function such that u + iv is holomorphic,
v − ũ = −i(u + iv) + i(u + iũ) is a real-valued holomorphic function. Then it is
constant.
Definition. The function ũ is called a harmonic conjugate of u on Ω. If u = u 1 +
iu2 is harmonic and complex-valued, the harmonic conjugates of u are the functions
f1 + if
ũ = u f1 and u
u2 , with u f2 harmonic conjugates of u1 and u2 respectively.
If Ω = D, one usually normalizes “the” conjugate harmonic function of u by
imposing that ũ(0) = 0. Consider the power series expansion (2.3) of u,
X∞ ∞
X
n
u(z) = a0 + an z + a−n z n
n=1 n=1
The fact that u is real is reflected by the conditions
a0 ∈ R , a−n = an .
Then X
∞
n
u(z) = a0 + 2<e an z .
n=1
We then take X
∞
n
ũ(z) = 2=m an z ,
n=1
in order to have
∞
X
(2.6) u + iũ = a0 + 2 an z n .
n=1
The power series expansion of ũ is therefore
X∞ ∞
X
(2.7) ũ(z) = −i an z n + i a−n z n .
n=1 n=1
By linearity, formulas (2.6) and (2.7) remain valid if u is complex-valued.
12 CHAPTER I
The Dirichlet problem on D consists in assigning a continuous function f on T

and seeking for a function u continuous on D̄ and harmonic in D, which coincides
with f on T. In other words, we want to solve

∆u = 0 in D
(3.1)
u = f on T
in u ∈ C(D̄). The maximum modulus principle implies that a solution, if it exists,

is unique. In fact, one just has to observe that the difference of two solutions would
be continuous on D̄, harmonic in D, and identically zero on T.
We shall prove the existence of the solution and give its expression.
If f (eit ) = eint with n ∈ Z, the solution of (3.1) is easily found as

u(z) = z n if n ≥ 0 ,
u(z) = z̄ |n| if n < 0 .
Suppose next that f ∈ C 2 (T), and let

X
f (eit ) = fˆ(n)eint
n∈Z
be its Fourier series. It is natural to construct

∞
X ∞
X
(3.2) u(z) = fˆ(n)z n + fˆ(−n)z̄ n .
n=0 n=1
It follows from Lemma 2.3 that this series converges uniformly on D̄. Since each
summand is harmonic in D, so is the sum (verify the mean value property). Then
this is the solution of (3.1).
We use some Fourier analysis to derive an integral formula giving u. By (3.2),
the Fourier series of ur (eit ) = u(reit ) is, for r < 1,
X
ur (eit ) = fˆ(n)r |n| eint .
n∈Z
We set
X
(3.3) Pr (eit ) = r |n| eint ,
n∈Z
where the series converges uniformly on T. Then6

Z
it it
ur (e ) = f ∗ Pr (e ) = f (ei(t−u) )Pr (eiu ) du .
T
The functions Pr in (3.3) form the Poisson kernel on T.

6 The following identity can be verified directly, expanding one of the two factors in Fourier
series and integrating term by term. It is however a consequence of the general identity f[∗ g(n) =
fˆ(n)ĝ(n) and of the uniqueness theorem for Fourier series expansions.
THE UNIT DISC 13
Lemma 3.1. The Poisson kernel equals

1 − r2
Pr (eit ) = , r<1.
1 + r 2 − 2r cos t
The function P (reit ) = Pr (eit ) is harmonic in D.
Proof. Consider the geometric series
∞
X ∞
X
n int 1
r e = (reit )n = .
n=0 n=0
1 − reit
Then
X
∞
it n int
Pr (e ) = 2<e r e −1
n=0
1 − re−it
= 2<e −1
|1 − reit |2
1 − rcost
=2 −1
1 + r 2 − 2r cos t
1 − r2
= .
1 + r 2 − 2r cos t
Moreover, the identity P (z) = 2<e(1 − z)−1 − 1 shows that P is harmonic
in D.
We have so proved that, for f ∈ C 2 (T) and r < 1,
u(reit ) = f ∗ Pr (eit )
Z
1 − r2
= f (ei(t−u) ) du
T 1 + r 2 − 2r cos u
Z
iu 1 − r2
= f (e ) du .
T 1 + r 2 − 2r cos(t − u)
Another way of writing the same identity is
Z
(3.4) u(z) = f (eiu )P (e−iu z) du .
T
Observe now that (3.4) makes sense for f continuous and defines a harmonic
function u in D (again, verify the mean value property using the fact that P is
harmonic). It takes some work to verify that u is continuous on D̄. We do this
now, introducing some preliminary more general notions.
Definition. Let {ϕr }r<1 be a family of integrable functions on T. We say that
they form an approximate identity for r → 1 if 7
R
(1) RT |ϕr (t)| dt ≤ C for some constant C and every r;
(2) T ϕr (t) dt = 1 for every r;
(3) for every δ > 0,
Z
lim |ϕr (t)| dt = 0 .
r→1 δ≤|t|≤π
7 Tosimplify the notation, here we think of the ϕr as periodic functions on the line. We then
write ϕr (t) instead of ϕr (eit ). The same for f in the next proposition.
14 CHAPTER I
The notion of “approximate identity” is quite general. We have chosen here

to give the definition with a parameter r tending to 1. The same definition can
be given, with the obvious modifications, with parameters varying on a different
set, for instance for a sequence of functions. Later on, we will find approximate
identities depending on a positive real parameter ε tending to 0.
The name “approximate identity” is justified by the following property.
Proposition 3.2. Let {ϕr } be an approximate identity on T. If f ∈ C(T),
(3.5) lim f ∗ ϕr = f
r→1
uniformly on T. If f ∈ Lp (T), with 1 ≤ p < ∞, the limit (3.5) holds in the Lp -norm.
Proof. Suppose f is continuous. By uniform continuity, given ε > 0, there exists
δ > 0 such that |f (t − u) − f (t)| < ε for |u| < δ. By condition (3), there is r0 < 1
such that for r0 < r < 1, Z
|ϕr (t)| dt < ε .
δ≤|t|≤π
Observe that, by condition (2), we can write

Z
f (t) = f (t)ϕr (u) du .
T
If r0 < r < 1, we have

Z

f ∗ ϕr (t) − f (t) = f (t − u) − f (t) ϕr (u) du

T
Z
1
≤ f (t − u) − f (t)ϕr (u) du
2π |u|<δ
Z
1
+ f (t − u) − f (t)ϕr (u) du
2π δ≤|u|≤π
1 1
< εkϕr k1 + kf k∞ ε
2π π
1
≤ C + 2kf k∞ ε .
2π
This proves the uniform convergence of f ∗ ϕr to f for f continuous.
Suppose now that f ∈ Lp (T), with 1 ≤ p < ∞. Given ε > 0, there is g ∈ C(T)
such that kf − gkp < ε. Then
kf ∗ ϕr − f kp ≤ k(f − g) ∗ ϕr kp + kg ∗ ϕr − gkp + kg − f kp .
We apply Young’s inequality8 to the first summand, to obtain
k(f − g) ∗ ϕr kp ≤ kf − gkp kϕr k1 < Cε .

8 More 1 1
generally, Young’s inequality says that if f ∈ Lp (T), g ∈ Lq (T), and p
+ q
≥ 1, then
1 1 1
f ∗g ∈ Lr (T),where = + − 1. This property is not specific of convolution on T, but it
r p q
holds for convolution on any locally compact group (for instance on R). Another general property
is that, when p and q are conjugate exponents (hence r = ∞), f ∗ g is continuous.
THE UNIT DISC 15
We then observe that

kg ∗ ϕr − gkp ≤ kg ∗ ϕr − gk∞ ,
so that, by the previous part of the proof, there is r0 such that, for r0 < r < 1,
kg ∗ ϕr − gkp < ε. Then,for r0 < r < 1,
kf ∗ ϕr − f kp < (C + 2)ε .
This proves that f ∗ ϕr tends to f in Lp (T).
Observe that the assumption f ∈ L∞ (T) does not imply that f ∗ ϕr → f in the
L∞ -norm. Since each f ∗ ϕr is continuous, this would imply that f is continuous
too.
With minor modifications to the proof, one can verify that the following more
general statement holds.
Corollary 3.3. Assume that the family {ϕr }r<1 satisfies (1), (3) and
Z
(2’) lim ϕr (t) dt = c .
r→1 T
Then
lim f ∗ ϕr = cf
r→1
uniformly on T if f ∈ C(T), and in the Lp -norm if f ∈ Lp (T), with 1 ≤ p < ∞.

Proposition 3.4. The Poisson kernel {Pr } is an approximate identity for r → 1.
Proof. Since the Pr are positive functions, (1) will be a consequence of (2). Using
the power series expansion (3.3), we have, since the convergence is uniform,
Z X Z
|n|
Pr (t) dt = r eint dt = 1 ,
T n∈Z T
because all the terms with n 6= 0 give integral zero.

We then prove (3). Using (1.5), we have, for t ∈ [−π, π],
1 − r2 2 1 − r2 1−r
(3.6) Pr (t) = it 2
≤ π 2
< 2π 2 .
|1 − re | (|t| + 1 − r) (|t| + 1 − r)2
So, given δ > 0,
Z Z
2 1−r
Pr (t) dt < 2π 2
dt
δ≤|t|≤π |t|≥δ (|t| + 1 − r)
Z
2 1
= 2π du ,
δ
|u|≥ 1−r (|u| + 1)2
where we have made the change of variable t = (1 − r)u. Since (|u| + 1)−2 is
integrable on the line,
Z
1
lim du = 0 ,
r→1 |u|≥ δ (|u| + 1)2
1−r
and this proves (3).

16 CHAPTER I
Theorem 3.5. For f ∈ C(T) the unique solution to the Dirichlet problem (3.1) is
u(reit ) = f ∗ Pr (eit ).
Proof. It follows from the uniform convergence of ur to f and the maximum prin-
ciple.
The same approach that we have used to obtain the Poisson integral u of a
continuous function f can be used to give a formula for its conjugate harmonic
function ũ. Using the Fourier series expansion of f , (3.2) and (2.7), we find that
∞
X ∞
X
ũ(z) = −i fˆ(n)z n + i fˆ(−n)z̄ n .
n=1 n=1
Then
X
(3.7) ũr (eit ) = −1 fˆ(n)sgn n r |n| eint = f ∗ P̃r (eit ) ,
n6=0
where
X
P̃r (eit ) = −i sgn n r |n| eint
n6=0
∞
X
= 2=m r n eint
(3.8) n=0
1
= 2=m
1 − reit
2r sin t
= .
1 + r 2 − 2r cos t
The functions P̃r form the conjugate Poisson kernel.
4. Functions in hp -spaces and their limits to the boundary
If u is harmonic in D, we define the means Mp (u, r) according to (1.2) and (1.3).

We also define hp (D) as the space of harmonic functions u such that
kukhp = sup Mp (u, r) < ∞ .

0≤r<1
The statements of Proposition 1.1, Lemma 1.4, Corollary 1.5 and Theorem 1.6
remain valid with H p replaced by hp . Only Lemma 1.4 requires a modification in
the proof, which goes as follows.
Lemma 4.1. Let u ∈ hp (D). Then, for every z ∈ D,
kukhp
|u(z)| ≤ Cp .
(1 − |z|)1/p
THE UNIT DISC 17
Proof. Take ρ so that |z| = r < ρ < 1, and set v(z) = u(ρz). Then v is harmonic
in D and continuous on D̄. By Theorem 3.5,
(4.1) ur = vr/ρ = v1 ∗ Pr/ρ = ur ∗ Pr/ρ .
By Young’s inequality,
|u(z)| ≤ kur k∞ ≤ kuρ kp kPr/ρ kp0 ≤ kukhp kPr/ρ kp0 ,
where p0 is the dual exponent of p. We then estimate the Lq -norm of Pr for any q.
We know that kPr k1 = 1. Moreover
1 − r2 1 − r2 1+r
kPr k∞ = max = = .
|t|≤π 1 + r 2 − 2r cos t (1 − r)2 1−r
If 1 < q < ∞,
Z q1
kPr kq = Pr (eit )q dt
T
q−1 1
≤ kPr k∞q kPr k1q
1 + r q−1
q
≤ .
1−r
Majorizing 1 + r with 2, we then have

0 0
kPr kq ≤ 21/q (1 − r)−1/q .
Therefore
kukhp
|u(z)| ≤ Cp 1/p ,
1 − ρr
for every ρ ∈ (r, 1). Letting ρ tend to 1, we have the conclusion.

Lemma 4.2. If u is harmonic in D, Mp (u, r) is a non-decreasing function of r.
In particular, u ∈ hp (D) if and only if
lim Mp (u, r) < ∞ ,

r→1
and kukhp equals this limit.

Proof. If r 0 < r < 1,
ur0 = vr0 /r = v1 ∗ Pr0 /r = ur ∗ Pr0 /r ,
by (4.1). By Young’s inequality9 , for any p ∈ [1, ∞],
kur0 kp ≤ kur kp kPr0 /r k1 = kur kp .

9 For p = ∞, one can simply invoke the maximum principle.
18 CHAPTER I
Lemma 4.2 applies in particular to holomorphic functions and H p -spaces.

A simple way to construct hp -functions, for any p ∈ [1, ∞], consists in taking
f ∈ Lp (T) and constructing the Poisson integral
u(reit ) = f ∗ Pr (eit ) .
It follows from Young’s inequality that kukhp ≤ kf kp .

For p = 1, the above construction can be extended to more general objects
defined on the boundary. Take a regular Borel measure µ on T – we write µ ∈
M (T) – and define
Z
0
it it
u(re ) = µ ∗ Pr (e ) = Pr (ei(t−t ) ) dµ(t0 ) .
T
Since the convolution of a finite measure and an L1 -function is in L1 , ur ∈ L1 (T)

for r < 1, and
(4.2) kur k1 ≤ kµkM kPr k1 ≤ kµkM .
The usual verification of the mean value property shows that u is harmonic in
D, so that u ∈ h1 (D) and kukh1 ≤ kµkM .
The question we will discuss now is if every hp -function can be obtained in this
way, i.e. if every hp -function is the Poisson integral of an Lp -function on T if
1 < p ≤ ∞, and of a regular Borel measure10 if p = 1.
Theorem 4.3. Consider the operator P mapping f (understood as either a func-
tion or a Borel measure on T) into the harmonic function Pf on D given by
(4.3) (Pf )r = f ∗ Pr .
Then P maps Lp (T) isometrically onto hp (D) for 1 < p ≤ ∞, and it maps M (T)
isometrically onto h1 (D).
The limit limr→1 (Pf )r exists in Lp if and only if one of the following holds
(1) f ∈ Lp (T) and 1 ≤ p < ∞;
(2) p = ∞ and f ∈ C(T).
In each of these cases, (Pf )r → f in the Lp -norm.
For general elements f of M (T) or L∞ (T), (Pf )r tends to f in the corresponding
weak*-topology 11 .
Proof. Suppose first that 1 < p < ∞. Take f ∈ Lp (T) and set u = Pf . By
Propositions 3.2 and 3.4, ur → f in the Lp -norm. In particular,
(4.4) lim Mp (u, r) = kf kp ,

r→1
so that kPf khp = kf kp . We prove next that P : Lp (T) → hp (D) is onto. Take
u ∈ hp (D) and let {rj }j∈N be a sequence of radii tending to 1. Since kurj k ≤ kukhp ,
10 We
regard L1 (T) as a subspace of M (T), identifying the function f with the measure µ such
that dµ(t) = f (t) dt.
11 We recall that M (T) is the dual space of C(T) and L∞ (T) is the dual space of L1 (T). The
weak*-topology on the dual space X 0 of a Banach space X is the weakest topology induced by
the elements of X as linear functionals on X 0 .
THE UNIT DISC 19
it follows from the Banach-Alaoglu theorem12 that some subsequence {urjk } has a
limit f ∈ Lp (T) in the weak* topology of Lp (T). Set v = Pf and take r < 1. Since
0
Pr ∈ Lp (T), and by (4.3),
v(reit ) = f ∗ Pr (eit )
Z
0
= f (eit )Pr (t − t0 ) dt0
T
Z
0
= lim urjk (eit )Pr (t − t0 ) dt0
(4.5) k→∞ T
= lim urjk ∗ Pr (eit )
k→∞
= lim u(rrjk eit )
k→∞
it
= u(re )
For p = 1, we know by (4.2) that P : M (T) → h1 (D) is continuous with norm

at most 1. In order to prove that this map is onto, we can repeat the argument
given above, using the fact that M (T) is the dual space of C(T), according to the
Riesz representation theorem. As before, we take u ∈ h1 (D), a sequence {rj } of
radii tending to 1, and we regard {urj } as a bounded sequence in M (T). From it,
we can extract a subsequence {urjk } having a weak* limit µ ∈ M (T). The proof
that u = Pµ follows the same lines as in (4.5). Moreover, If g ∈ C(T),
Z Z

g(t) dµ(t) = lim g(t)ur (t) dt
k→∞ jk
T
ZT
≤ lim inf |g(t)||urjk (t)| dt
k→∞ T
≤ kgk∞ lim inf kurjk k1
k→∞
= kgk∞ kukh1 .
Therefore Z

kµkM = sup g(t) dµ(t) ≤ kukh1 .
kgk∞ =1 T
Putting this together with (4.2), we have that kPµkh1 = kµkM .

Consider now p = ∞, and let f ∈ L∞ (T). Then u = Pf satisfies kur k∞ ≤
kf k∞ kPr k1 , so that u ∈ h∞ (D) and kukh∞ ≤ kf k∞ . Using the fact that L∞ (T)
is the dual space of L1 (T), the same proof given above for p = 1, shows that
P : L∞ (T) → h∞ (D) is onto and isometric.
We pass now to the second part of the statement.
By Propositions 3.2 and 3.4, limr→1 (Pf )r exists in the Lp -norm and is equal to
f in cases (1) and (2). Consider p = 1 and take µ ∈ M (T). In order to prove that
(Pµ)r → µ as r → 1 in the weak* topology, we take g ∈ C(T). Using Fubini’s
12 The Banach-Alaoglu theorem says that, if X is a separable Banach space and X 0 is its dual
space, the closed balls in X 0 are sequentially compact in the weak* topology.
20 CHAPTER I
theorem and the parity of Pr (see Lemma 3.1),

Z Z
g(t)(Pµ)r (t) dt = g(t)(µ ∗ Pr )(t) dt
T T
Z Z
= g(t)Pr (t − t0 ) dµ(t0 ) dt
ZT TZ
= g(t)Pr (t − t ) dt dµ(t0 )
0
ZT T
= (g ∗ Pr )(t0 ) dµ(t0 ) .
T
Since g is continuous, g ∗ Pr → g uniformly. Hence

Z Z
lim g(t)(Pµ)r (t) dt = g(t0 ) dµ(t0 ) ,
r→1 T T
showing that µ is the weak* limit of (Pµ)r . If the (Pµ)r have a strong limit
as r → 1, this limit is in L1 (T), and at the same time it must coincide with µ,
because the norm topology is stronger than the weak* topology. This implies that
µ ∈ L1 (T).
The same argument, replacing g ∈ C(T) with g ∈ L1 (T), shows that, if f ∈
∞
L (T), then (Pf )r → f in the weak* topology, and that the convergence is in
norm if and only if f is continuous.
Observe that the inclusions hp (D) ⊂ hq (D) if q < p also give other consequences
for convergence to the boundary. For instance, if u is bounded on D, i.e. u ∈
h∞ (D), and f is its boundary function, then ur → f in any Lq -norm for q < ∞.
Corollary 4.4. For r < 1, let u(r) (z) = u(rz). If u ∈ hp (D), with 1 < p < ∞,
u(r) → u in the hp -norm as r → 1. The same is true for p = 1, provided u] ∈ L1 (T)
and for p = ∞, provided u] is continuous. For 1 0, take
r < 1 so that ku] − ur kp < ε. Since each half of the Taylor series of u,
∞
X ∞
X
u(z) = an z n + a−n z̄ n ,
n=0 n=1
converges to u uniformly on compact sets, there is N such that the sum sN (z) of
the two partial sums of order N satisfies |u(z) − sN (z)| < ε for |z| = r. Then
ku] − (uN )r kp < 2ε.
5. Boundary limits of conjugate harmonic functions
A natural question concerns conjugate harmonic functions: given u ∈ hp (D),

is its conjugate harmonic function ũ also in hp (D)? At this stage we only give
the answer for p = 2 (which is positive) and for p = 1, ∞ (which is negative).
THE UNIT DISC 21
The (positive) answer for the other values of p will be given in Chapter III. In
the last part of this section we discuss under which assumptions the congiugate
harmonic function of a function u continuous up to the boundary can be extended
continuously to the boundary.
The case p = 2 is special in the sense that h2 (D), being isometric with L2 (T), is
a Hilbert space. To be precise, given u, v ∈ h2 (D), there are f, g ∈ L2 (T) such that
u = Pf , v = Pg. If we set
hu, vih2 = hf, giL2 ,
this inner product induces the h2 -norm13 . It is also clear that P transforms or-
thonormal bases of L2 (T) into orthonormal bases of h2 (D). In particular, from the
basis {eint }n∈Z of L2 (T) we derive the orthonormal basis {z n }n∈N ∪ {z̄ n }n≥1 of
h2 (D).
Therefore, if X X
u= an z n + a−n z̄ n ,
n≥0 n≥1
then X
kuk2h2 = |an |2 .
n∈Z
It follows from (2.7) that

X
(5.1) kũk2h2 = |an |2 ≤ kuk2h2 .
n6=0
Then ũ is also in h2 (D). We have so proved the following statement.

Proposition 5.1. The conjugate function operator T : u 7→ ũ maps h2 (D) con-
tinuously into itself.
Consider now p = 1, and take
1 − r2
P (reit ) = ∈ h1 (D) .
1 + r 2 − 2r cos t
Its conjugate harmonic function is
2r sin t
P̃ (reit ) = .
1 + r 2 − 2r cos t
13 It follows from Lemma 4.2 and the polarization identity
1`
kx + yk2 + kx + iyk2 − kx − yk2 − kx − iyk2
´
hx, yi =
4
holding on any complex Hilbert space, that
hu, vih2 = lim hur , vr iL2 .

r→1
22 CHAPTER I
In fact P and P̃ are real, P̃ (0) = 0, and
2
F (z) = P (z) + iP̃ (z) = −1
1−z
is holomorphic. If P̃ were in h1 (D), it would follow that F ∈ H 1 (D), but this is in

contrast with Proposition 1.3.
This example is interesting because it emphasizes one negative property of the

conjugate Poisson kernel: the functions P̃r do not form an approximate identity.
More precisely, the fact that P̃ 6∈ h1 (D) means that
(5.2) lim kP̃r k1 = +∞ .

r→1
Passing to the case p = ∞, the following example shows that the harmonic
conjugate of a bounded function need not be bounded.
Consider an open vertical strip S = {z : a < <ez < b} in the complex plane.
By the Riemann mapping theorem, there is a conformal map ϕ from D onto S.
Adding, if necessary, an imaginary constant to ϕ, we can assume that ϕ(0) ∈ R.
If u = <eϕ, it follows that a < u(z) < b, so that u ∈ h∞ (D). But its harmonic
conjugate ũ is =mϕ, which is not bounded14 .
One may ask if the conclusion ũ ∈ h∞ (D) holds if u is assumed to be continuous

on D̄. The answer is negative again, but instead of giving a counterexample, we
use a functional analytic argument.
Lemma 5.2. Let X and Y be two Banach spaces of harmonic functions in D.
Assume that convergence in each of these spaces implies uniform convergence on
compact subsets of D, and that the conjugate function operator T maps functions
in X into functions in Y . Then T : X → Y is continuous.
Proof. By
the closed graph theorem, it is sufficient to prove that if a sequence
(un , T un ) converges to (u, v) ∈ X × Y , then v = T u. Since T maps real-valued
functions into real-valued functions, we can assume that the un are all real-valued.
The same is therefore true for u and v.
On every compact set, un + iT un converges uniformly to u + iv. Hence u + iv
is holomorphic. Since vn (0) = 0 for every n, also v(0) = 0. We conclude that
v = ũ.
Proposition 5.3. There exist functions u, harmonic in D and continuous on D̄,
whose harmonic conjugate ũ is not bounded in D.
Proof. Denote by h∞ ∞
c (D) the closed subspace of h (D) consisting of those harmonic
functions in D which admit a continuous extension to D̄. Assume that T maps
h∞ ∞
c (D) into h (D). By Lemma 5.2, this map would be continuous, hence there
would exists a constant C > 0 such that
kũkh∞ ≤ Ckukh∞
14 Another
example is u(z) = =m log(1 − z), which is bounded, but whose harmonic conjugate,
−<e log(1 − z) = − log |1 − z|, is not.
THE UNIT DISC 23
for every u ∈ h∞
c (D). Given r < 1, consider the linear functional on C(T)
Z
ψr (f ) = f ∗ P̃r (1) = f (t)P̃r (−t) dt .
T
If u is the Poisson integral of f , we know from Theorem 4.3 that kukh∞ = kf k∞ .

Moreover, ψr (f ) = ũ(r). We would then have
|ψr (f )| ≤ kũkh∞ ≤ Ckf k∞ .
This would imply that

Z

kP̃r k1 = sup f (t)P̃r (−t) dt

f ∈C(T) , kf k∞ ≤1 T
= sup |ψr (f )|
f ∈C(T) , kf k∞ ≤1
≤C ,
in contrast with (5.2).

On the other hand, if we impose a little more regularity on the behaviour of u
on the boundary, we obtain a positive result.
Proposition 5.4. Suppose that u is harmonic in D, continuous on D̄, and α-
Lipschitz 15 on T for some α > 0. Then ũ extends continuously to D̄.
Proof. We shall prove that ũr has a uniform limit ũ1 for r → 1. It is a simple
exercise to prove that this condition is in fact equivalent to the existence of a
continuous extension of ũ on D̄.
Let f be the restriction of u to T, so that ũr = f ∗ P̃r . Consider
Z Z
0 0 0 2r sin t0
f ∗ P̃r (t) = f (t − t )P̃r (t ) dt = f (t − t0 ) dt0 .
T T 1 + r 2 − 2r cos t0
R
Since P̃r is odd in t0 , T
P̃r (t0 ) dt0 = 0. We then have
Z

(5.3) lim f ∗ P̃r (t) = lim f (t − t0 ) − f (t) P̃r (t0 ) dt0 .
r→1 r→1 T
We prove that this limit equals the expression formally obtained by taking the
pointwise limit of the integrand, i.e.
Z π
1 sin t0
˜
f (t) = f (t − t0 ) − f (t) dt0
2π −π 1 − cos t 0
(5.4) Z π
1 t0
= f (t − t0 ) − f (t) cot dt0 .
2π −π 2
15 A function f is α-Lipschitz (or α-Hölder) on a metric space X if there is a constant C such

that |f (x) − f (y)| ≤ Cd(x, y)α for every x, y ∈ X.
24 CHAPTER I
Observe that this integral is absolutely convergent, due to the fact that f is
α-Lipschitz. In fact,
t0 C
(5.5) 0
f (t − t ) − f (t) cot ≤ 0 1−α ,
2 |t |
which is integrable on [−π, π].
Using (1.5), we find that
2r| sin t0 | 2|t0 | 2π 2
(5.6) |P̃r (t0 )| = ≤ π 2
≤ .
|1 − reit0 |2 (|t0 | + 1 − r)2 |t0 |
Hence
C
(5.7) f (t − t0 ) − f (t)P̃r (t0 ) ≤ ,
|t0 |1−α
and dominated convergence can be applied to show that limr→1 f ∗ P̃r (t) = f˜(t)
pointwise. Moreover,
Z π
1 0
kf˜ − f ∗ P̃r k∞ ≤ max f (t − t0 ) − f (t) cot t − P̃r (t0 ) dt0
t∈[−π,π] 2π −π 2
Z π
C t0
≤ |t0 |α cot − P̃r (t0 ) dt0 .
2π −π 2
We can apply dominated convergence again, because the integrand tends to zero
pointwise and
t0 t0 C

cot − P̃r (t0 ) ≤ cot + P̃r (t0 ) ≤ 0 .
2 2 |t |
Therefore,
lim kf˜ − f ∗ P̃r k∞ = 0 .
r→1
This result admits the following local version.

Proposition 5.5. Suppose that f ∈ L1 (T) is α-Lipschitz on an open interval I ⊂ T
for some α > 0. Then f˜ in (5.4) is well defined on I and limr→1 f ∗P̃r = f˜ uniformly
on compact subsets of I. In particular, the harmonic conjugate ũ of u = Pf admits
a continuous extension to D ∪ I.
Proof. For fixed t ∈ I, the integral in (5.4) is absolutely convergent, because the
0
estimate (5.5) holds for t − t0 ∈ I, and cot t2 is bounded for t − t0 6∈ I.
Let J be a compact subinterval of I. Without loss of generality, we can assume
that J = [−δ, δ] with δ < π2 and that the double interval J 0 = [−2δ, 2δ] is also
contained in I. For t ∈ J , we then have
Z π
1 t 0
f˜(t) − P̃r ∗ f (t) = f (t − t ) − f (t) cot − P̃r (t ) dt0
0 0

2π −π 2
Z π

1 t−s
= f (s) − f (t) cot − P̃r (t − s) ds
2π −π 2
Z
1 t−s

≤
f (s) − f (t) cot − P̃r (t − s) ds
2π J0 2
Z
t−s

+ f (s) − f (t) cot − P̃r (t − s) ds
[−π,π]\J 0 2
1
= I1 (t, r) + I2 (t, r) .
2π
THE UNIT DISC 25
We now want to prove that

lim sup f˜(t) − P̃r ∗ f (t) = 0 .
r→1 t∈J
In the integral I1 (t, r), both t and s are in I, so that

Z
t−s
|I1 (t, r)| ≤ |t − s|α cot − P̃r (t − s) ds
2
ZJ
0
t0

= |t0 |α cot − P̃r (t0 ) dt0
t−t0 ∈J 0 2
Z 2π
t0
≤ |t0 |α cot − P̃r (t0 ) dt0 ,
−2π 2
which is independent of t. By dominated convergence, as in the proof of Proposition
5.4, this last quantity tends to zero as r → 1, so that
lim sup I1 (t, r) = 0 .

r→1 t∈I 0
Passing to I2 (t, r), let M = maxt∈J |f (t)|. Then

Z

I2 (t, r) = f (s) − f (t) cot t − s − P̃r (t − s) ds
[−π,π]\J 0 2
Z
t−s
≤ |f (s)| + M cot − P̃r (t − s) ds .
[−π,π]\J 0 2
Observe that, if t ∈ J and s 6∈ J 0 , |t − s| > δ. Therefore

t−s sin(t − s) 2r sin(t − s)

cot − P̃r (t − s) = − 2

2 1 − cos(t − s) 1 + r − 2r cos(t − s)

| sin(t − s)| 2r 1 − cos(t − s)
(5.8) = 1 −
1 − cos(t − s) 1 + r 2 − 2r cos(t − s)
t − s (1 − r)2

= cot
2 1 + r 2 − 2r cos(t − s)

The quantity cot t−s
2
/ 1 + r 2 − 2r cos(t − s) remains bounded for |t − s| > δ.
Given ε > 0, we can then find r0 < 1 such that, for r > r0 ,
t−s

cot − P̃r (t − s) < ε
2
for every t, s with |t − s| > δ.
Therefore, if r > r0 ,
I2 (t, r) ≤ ε2π kf k1 + M .
This shows that
lim sup I2 (t, r) = 0 ,
r→1 t∈I 0
concluding the proof.

26 CHAPTER I
6. The Cauchy projection
The Hardy space H p (D) is a closed subspace of hp (D). It is clear from Section 3
that the Poisson integral of a function (or measure) f on T is holomorphic in D if
and only if fˆ(n) = 0 for n < 0. If we define
Lp+ (T) = {f ∈ Lp (T) : fˆ(n) = 0 for n < 0} ,
and similarly
M+ (T) = {µ ∈ M (T) : µ̂(n) = 0 for n < 0} ,
the statement of Theorem 4.3 can be repeated word by word, replacing hp with
H p , Lp with Lp+ , and M with M+ .
In the case p = 2 we are working with Hilbert spaces, and we want to describe
the orthogonal projection from h2 (D) to H 2 (D), that we shall denote by C (for
Cauchy).
Lemma 6.1. For u ∈ h2 (D), let ũ be its harmonic conjugate. Then
1 1
(6.1) Cu = (u + iũ) + u(0) .
2 2
Denoting by u] ∈ L2 (T) the boundary function of u, i.e. u] = limr→1 ur , then
Z
u] (eit )
(6.2) Cu(z) = −it
dt .
T 1 − ze
Proof. Since {z n }n≥0 ∪ {z̄ n }n≥1 is an orthonormal basis of h2 (D), and {z n }n≥0
spans H 2 (D), if we write
∞
X ∞
X
n
u(z) = an z + a−n z̄ n ,
n=0 n=1
it follows that
∞
X
Cu(z) = an z n .
n=0
Then (6.1) follows from (2.6). Consider now the Fourier series of u] ,
X
u] (eit ) = an eint ,
n∈Z
with convergence in the L2 (T)-norm. Since

∞
X
it
(Cu)r (e ) = an r n eint ,
n=0
one obtains (Cu)r from u] by multiplying each Fourier coefficient ub] (n) = an by r n
if n ≥ 0 and by 0 if n < 0. Consider therefore the function
∞
X
it 1
(6.3) Cr (e ) = r n eint = ,
n=0
1 − reit
THE UNIT DISC 27
\r (n) = ub] (n)C

called the Cauchy kernel. Then (Cu) cr (n) for every n ∈ Z, so that
Cu(reiθ ) = (Cu)r (eiθ )

= u] ∗ Cr (eiθ )
Z
u] (eit )
= i(θ−t)
dt ,
T 1 − re
giving (6.2).
Observe that (6.2) can be rewritten as a contour integral,
I
1 u] (w)
(Cu)(z) = dw ,
2πi ∂D w − z
an expression resembling the ordinary Cauchy integral formula. This is compatible
with the fact that if u is already in H 2 (D) (i.e. it is holomorphic), we then have
the identity, for |z| < r < 1,
I
1 u(w)
u(z) = dw
2πi γr w − z
Z π
1 ur (eit ) it
= re dt
2π −π reit − z
Z
ur (eit )
= z −it dt
T 1 − re
where γr is the circle of radius r oriented counterclockwise. Letting r → 1, one

obtains (6.2), since we are assuming that Cu = u.
The orthogonal projection of h2 (D) onto H 2 (D) corresponds, passing to bound-

ary values, to the orthogonal projection C ] of L2 (T) onto L2+ (T), via the following
commutative diagram:
C]
L2 (T) −→ L2+ (T)
↓P ↓P
C
h (D) −→ H 2 (D)
2
So
C ] f = lim (CPf )r ,
r→1
2
where the limit is meant in the L -norm.
We want to give an expression of C ] that does not involve the harmonic extension
to the interior. Because of Lemma 6.1, this problem can be reduced to that of
finding a more direct formula for the operator
H : L2 (T) −→ L2 (T) ,
mapping f into
Hf = lim P̃r ∗ f ,
r→1
2
again in the L -norm. This operator is bounded, by Proposition 5.1, and
Hf (eit ) = lim f ∗ P̃r (eit )
r→1
(6.4) Z
0 2r sin t0
= lim f (ei(t−t ) ) dt0 .
r→1 T 1 + r 2 − 2r cos t0
28 CHAPTER I
Proposition 6.2. If f ∈ L2 (T),

Z
it 1 0 t0 0
(6.5) Hf (e ) = lim f (ei(t−t ) ) cot dt ,
ε→0 2π ε<|t0 |<π 2
where the limit is in the L2 -norm.

Proof. For 0 < ε < 1, define
Z
it 1 0 t0 0
Hε f (e ) = f (ei(t−t ) ) cot dt .
2π ε<|t0 |<π 2
If we prove that
(6.6) lim kf ∗ P̃1−ε − Hε f k2 = 0 ,

ε→0
this will imply the assertion, by (6.4).

We can write
Z
1 0
f ∗ P̃1−ε (t) − Hε f (t) = f (ei(t−t ) )P̃1−ε (t0 ) dt0
2π |t0 |<ε
Z
i(t−t0 ) 0 t0 0
+ f (e ) P̃1−ε (t ) − cot dt
ε<|t0 |<π 2
= f ∗ ϕε (t) ,
where
0 P̃1−ε (t0 ) if |t0 | < ε
ϕε (t ) = 0
P̃1−ε (t0 ) − cot t2 if ε ≤ |t0 | ≤ π .
By (5.8),
t0 t0 ε2
0
P̃1−ε (t ) − cot = cot
2 2 1 + (1 − ε)2 − 2(1 − ε) cos t0
t0 ε2

= cot 2
2 ε + 4(1 − ε) sin2 t2
0
t0 ε2

≤ cot 2 .
2 ε + π42 (1 − ε)t0 2
We are interested in this quantity for ε ≤ |t0 | ≤ π, where we can use the inequality
t0 1 π π

cot ≤ t 0 ≤ 0 ≤ .
2 | sin 2 | |t | ε
4
We can also assume that ε is small enough so that π 2 (1 − ε) > 1. For ε and t0
subject to these restrictions,
t0 πε
0
cot − P̃1−ε (t ) ≤ 2 .
2 ε + t0 2
THE UNIT DISC 29
On the other hand, if |t0 | < ε,
2(1 − ε)| sin t0 |

P̃1−ε (t0 ) =
1 + (1 − ε)2 − 2(1 − ε) cos t0
2ε
≤ 2 .
ε + t0 2
Putting these two inequalities together, we have
πε
|ϕε (t0 )| ≤ ,
ε2 + t0 2
for t0 ∈ [−π, π]. We can then apply Corollary 3.3 with c = 0. We leave the
verification of conditions (1) and (3) to the reader, and observe that ϕε has integral
0 because it is odd on [−π, π]. This gives (6.6).
We cannot eliminate the limit in (6.5) and write the integral over all of T,
because the integrand would not be absolutely convergent in general. For the same
reason, we cannot replace the integrals outside of the simmetric intervals [−ε, ε]
with integrals outside of arbitrary intervals containing 0, such as, for instance,
[−ε, 2ε], because the result would not be the same.
This matter can be nicely formalized in the language of distribution theory.

Observe that
Z
t
(6.7) lim f (t) cot dt
ε→0 ε<|t|<π 2
does not exist, or is infinite, for a generic continuous function f on T (take for
sgn t
instance f (t) = log |t|
near t = 0).
If, however, f is a C 1 -function16 , we have, using the oddness of cot 2t ,
Z Z
t t
lim f (t) cot dt = lim f (t) − f (0) cot dt
ε→0 ε<|t|<π 2 ε→0 ε<|t|<π 2
Z π
t
= f (t) − f (0) cot dt ,
−π 2
where this last integral is absolutely convergent. Moreover the linear functional
Z
t
(6.8) f 7−→ Φ(f ) = lim f (t) cot dt
ε→0 ε<|t|<π 2
is continuous w.r. to the norm
kf kC 1 = kf k∞ + kf 0 k∞
16 fα-Lipschitz would be enough, but distribution theory is focused on spaces of C k - or C ∞ -

functions.
30 CHAPTER I
on C 1 (T). In fact
Z
π t
Φ(f ) = f (t) − f (0) cot dt
2
−π
Z π

t
≤ f (t) − f (0) cot dt
−π 2
Z π
t
≤ kf 0 k∞ |t| cot dt
−π 2
≤ Ckf kC 1 .
This property is expressed by saying that Φ is a distribution of order 1 on T.

According to a standard terminology, a distribution such as Φ, involving limits of
integrals that would not be abolutely convergent otherwise, is called a principal
value distribution. One writes
Z Z
1 t t
lim f (t) cot dt = p.v. f (t) cot dt .
ε→0 2π ε<|t|<π 2 T 2
The convolution Φ ∗ f is defined for f ∈ C 1 (T) as

Z
t0 0
Φ ∗ f (t) = Φ f (t − ·) = p.v. f (t − t0 ) cot dt ,
T 2
and one can easily prove that Φ ∗ f ∈ C(T).

One also writes
t t
p.v. cot ∗f , or p.v. cot ∗f
2 2
for Φ ∗ f .
As a consequence of Proposition 6.2, we have the following extension result.
Corollary 6.3. For f ∈ C 1 (T), Hf = Φ ∗ f . Therefore the convolution operator

f 7→ Φ ∗ f extends in a unique way to a continuous operator from L 2 (T) to itself.
Moreover
1
C ] f = f + iHf + fˆ(0) .
2
7. Blaschke products and the F. and M. Riesz theorem
Working with holomorphic functions, one comment must be made on products

of H p -functions17 .
17 We
do not discuss this issue for hp -functions because, in general, the product of two harmonic
functions is not harmonic.
THE UNIT DISC 31
Lemma 7.1. Suppose f ∈ H p (D) and g ∈ H q (D), with 1 ≤ p, q ≤ ∞ and p1 + q1 =

1
s
≤ 1. Then f g ∈ H s (D), kf gkH s ≤ kf kH p kgkH q . If p, q > 1, denoting by f ]
and g ] the boundary functions of f, g respectively, then (f g)] is also a function and
(f g)] = f ] g ] .
Proof. Suppose first that p and q are both finite. By Hölder’s inequality,
Ms (f g, r) ≤ Mp (f, r)Mq (g, r)
for every r < 1. Therefore f g ∈ H s (D) and kf gkH s ≤ kf kH p kgkH q . Moreover,
k(f g)r − f ] g ] ks = kfr gr − f ] g ] ks
≤ k(fr − f ] )gr + (gr − g ] )f ] ks
≤ k(fr − f ] )gr ks + k(gr − g ] )f ] ks
≤ kfr − f ] kp kgr kq + kgr − g ] kq kf ] kp
≤ kfr − f ] kp kgkH q + kgr − g ] kq kf kH q .
This shows that (f g)r tends to f ] g ] in Ls (D), so that (f g)] = f ] g ] .

If one or both exponents are ∞, the identity (f g)] = f ] g ] follows from the
above and the inclusion H ∞ (D) ⊂ H t (D) for t < ∞. Once this is established, the
inequality of norms is obvious.
As we see, this result implies (when q = p0 ) that an H 1 -function which is the
0
product of an H p - and an H p -function is the Poisson integral of an L1 -funtion on
the torus. On the other hand, the picture given by Lemma 7.1 is not complete,
because it does not say what happens on the boundary when one of the two ex-
ponents, say p, is equal to 1. In that case we are forced to assume that q = ∞,
but stil we do not know if the product f ] g ] of a Borel measure times a bounded
measurable function makes sense.
All these problems are washed out by the following theorem.
Theorem 7.2 (F. and M. Riesz). Let µ be a Borel measure in M+ (T). Then µ
is absolutely continuous with respect to Lebesgue measure.
In other words, M+ (T) = L1+ (T).
Corollary 7.3. If f ∈ H 1 (D), the limit limr→1 fr exists in the L1 -norm. Holo-
morphic polynomials are dense in H 1 (D).
Proof. The first part is an immediate consequence of Theorems 7.2 and 4.3. For
the second part, proceed as in the proof of Corollary 4.4.
There are several proofs of the brothers Riesz theorem. The one we give involves
an important notion in H p -theory, the Blaschke products.
For α ∈ D, consider the function
z−α
(7.1) ψα = .
ᾱz − 1
One easily verifies that ψα is holomorphic on C\{1/ᾱ} (hence on a neighborhood
of D), injective, and that on T
it |eit − α|
|ψα (e )| = =1,
|ᾱ − e−it |
32 CHAPTER I
hence18 |ψα (z)| < 1 for z ∈ D.

Suppose now that f ∈ H p (D) is not identically zero, and that f (α) = 0 for some
α ∈ D. Then the function
f (z)
g(z) =
ψα (z)
has a removable singularity at z = α. Hence g is holomorphic in D.
Lemma 7.4. g ∈ H p (D) and kgkH p = kf kH p .
Proof. We clearly have
Mp (f, r) = Mp (gψα , r) ≤ Mp (g, r) .
Given ε > 0, it is possible to find r0 > |α| such that if |z| > r0 , then |ψα (z)| >
1 − ε. Then, if r > r0 ,
Mp (f, r) = Mp (gψα , r) ≥ (1 − ε)Mp (g, r) .
Passing to the limit as r → 1 and using the arbitrarity of ε, we have the conclu-
sion.
This suggests a way to “remove zeroes” from a function f ∈ H p (D) which is
not identically 0, without modifying its norm. If the zeroes of f in D are a finite
number, α0 , α1 , . . . , αn (counted with their multiplicities), it is sufficient to repeat
the above construction n + 1 times to obtain a factorization of f (z) as g(z)B(z),
with
n
Y Yn
z − αj
(7.2) B(z) = ψαj (z) = .
j=0 j=0
ᾱ j z − 1
Then g has no zeroes in D and kgkH p = kf kH p .

If the zeroes of f in D are infinitely many, they form a countable set without
accumulation points in D itself. Ordering these zeroes by increasing modulus, and
counting them with their multiplicities, we obtain a sequence {αj }, with |αj | ≤
|αj+1 |, and limj→∞ |αj | = 1.
This is true for any non-identically zero holomorphic function in D. Something
more can be said about distribution of zeroes of H p -functions.
Lemma 7.5. Let f ∈ H p (D) be not identically zero, and let f (z) = az k h(z), with
k ≥ 0 and h(0) = 1. Then
X
1 − |αj | ≤ log khkH p = log kf kH p − log |a| .
j : αj 6=0
Proof. The last equality being obvious, we can assume that k = 0 and a = 1, i.e.
that f = h.
We initially suppose that f is holomorphic on a larger disc DR = {z : |z| < R}
with R > 1, and also that f does not vanish for |z| = 1. We then have only finitely
18 More
precisely, ψα is a conformal map of D onto itself (i.e. a Möbius transformation). It can
be shown that any Möbius transformation has the form eiθ ψα for some α ∈ D and some eiθ ∈ T.
THE UNIT DISC 33
many zeroes α0 , . . . , αn in D, and we can decompose f as f = gB, with B given by

(7.2). Observe that B is defined in a larger disc D 0 = D1+δ and non-zero in D 0 \ D,
so that g is holomorphic in D 0 .
Since g has no zero in D 0 , log g admits a holomorphic determination19 on D 0 .
Then log |g| = <e log g is harmonic in D 0 . By the mean value theorem,
Z
log |g(0)| = log |g(eit )| dt
ZT Z
= log |f (e )| dt − log |B(eit )| dt
it
ZT T
= log |f (eit )| dt .
T
But
n
X
log |g(0)| = log |f (0)| − log ψαj (0)
j=0
n
X
=− log |αj | ,
j=0
so that
n
X Z
(7.3) − log |αj | = log |f (eit )| dt .
j=0 T
Passing to a general f ∈ H p (D), with f (0) = 1, we apply (7.3) to fr (z) = f (rz),

with r < 1 and r 6= |αj | for every j. The zeroes of fr in D are the α0j = αj /r, with
|αj | < r. Therefore,
X Z
|αj |
− log = log |f (reit )| dt .
r T
j:|αj |<r
Observe now that, for 0 < x < 1, log x < x − 1. Therefore
|αj | |αj |
1− < − log ,
r r
and Z
X |αj |
1− < log |f (reit )| dt .
r T
j:|αj |<r
The left-hand side is an increasing function of r, so that, by the Beppo-Levi

theorem,
X |αj | X
∞

lim 1− = 1 − |αj | .
r→1 r j=0
j:|αj |<r
19 This g0
is true since D 0 is simply connected. In fact g
is holomorphic in D, and it admits a
primitive h. Then (eh /g)0 = 0, so that eh = cg. Adding an appropriate constant to h, we obtain
c = 1.
34 CHAPTER I
This limit is then controlled by the supremum over r on the right-hand side, i.e.
∞
X Z

1 − |αj | ≤ sup log |f (reit )| dt .
r<1 T
j=0
It remains to prove that this supremum is finite if f ∈ H p (D). It is sufficient to

take p = 1, by the inclusion properties of Hardy spaces. By Jensen’s inequality 20 ,
if r 6= rj for every j,
Z Z p1
it it p
exp log |f (re )| dt = exp log |f (re )| dt
T T
Z p1
it p
≤ |f (re )| dt
T
≤ kf kH p ,
so that
∞
X
1 − |αj | ≤ log kf kH p .
j=0
This concludes the proof.

In principle, we would like to decompose an H p -function f (z) with infinitely
many zeroes {αj } as the product g(z)B(z), with the finite product (7.2) replaced
by the analogous infinite product. The requirements would be
(1) that the partial products converge uniformly on compact subsets of D,
(2) that the limit function vanish only at the zeroes of f .
These requirements already show that some modification to (7.2) must be made.
Suppose that an infinite product of complex numbers aj converges to a non-zero
value, i.e.
∞
Y Y n
aj = lim aj = lim An = A 6= 0 .
n→∞ n→∞
j=0 j=0
Then necessarily aj 6= 0 for every j and
Aj
lim aj = lim =1.
j→∞ j→∞ Aj−1
Consider therefore a point z ∈ D different from the αj , i.e. such that f (z) 6= 0.
In general, we cannot expect that limj→∞ ψαj (z) = 1. However, this difficulty is
bypassed if we by multiply each ψαj by a contant factor of modulus 1.
So, for α ∈ D, define
ᾱ ᾱ z − α
(7.4) ψ̃α (z) = ψα (z) = if α 6= 0, and ψ̃0 (z) = z .
|α| |α| ᾱz − 1
20 Let
m be a probability measure on a set X, and let f : X → R be measurable. If ϕ : R → R
is convex, then
“Z ” Z ` ´
ϕ f (x) dm(x) ≤ ϕ f (x) dm(x) .
X X
Here it is applied with ϕ(t) = et .

THE UNIT DISC 35
For αj 6= 0,
ᾱj z − |αj |2 − |αj |ᾱj z + |αj |
ψ̃αj (z) − 1 =
|αj |(ᾱj z − 1)
1 − |αj | ᾱj z + |αj |
= .
|αj | ᾱj z − 1
Since
|ᾱj z − 1| ≥ 1 − |αj ||z| > 1 − |z| ,
we have

(7.5) ψ̃α (z) − 1 < 2 1 − |αj | .
j
1 − |z|
It follows that
(7.6) lim ψ̃αj (z) = 1 .

j→∞
We then construct the Blaschke product

∞
Y
(7.7) B(z) = ψ̃αj (z) .
j=0
Proposition
P∞ 7.6.
Let {αj } be a sequence of complex numbers αj ∈ D, such that
j=0 1 − |α j | < ∞. Then the Blaschke product (7.7) converges unconditionally 21
and uniformly on compact sets to a function B(z) ∈ H ∞ (D) vanishing only at the
points αj . Moreover, |B(eit )| = 1 almost everywhere on T.
Qn
Proof. Let Bn (z) = j=0 ψ̃αj (z) be the partial products of (7.4). We discuss
unconditional convergence first.
If z is one of the αj , then Bn (z) = 0 for n large, and this holds independently of
ordering.
If z is not one of the αj , it follows from (7.6) that for j ≥ j0 = j0 (z), the points
ψ̃αj (z) are contained in the disc D̃ centered at 1 and radius 1/2. Let log z be the
principal determination of the logarithm in D̃, i.e. such that log 1 = 0. Then, if
n > j0 ,
n
Y
Bn (z) = Bj0 −1 (z) elog ψ̃αj (z)
j=j0
X
n
= Bj0 −1 (z) exp log ψ̃αj (z) .
j=j0
We can then say that the Blaschke product is unconditionally convergent at z

to a non-zero limit if and only if the series
∞
X
log ψ̃αj (z)
j=j0
21 This means independently of reorderings of the terms.

36 CHAPTER I
log w
is unconditionally (i.e. absolutely) convergent. Since w−1 is holomorphic in D̃ and
continuous on D̃, | log w| ≤ C|w − 1| for w ∈ D̃. By (7.5),
∞
X ∞
X

log ψ̃αj (z) ≤ C 1 − ψ̃αj (z)
j=j0 j=j0
(7.8) ∞
C X
≤ 1 − |αj | .
1 − |z|
j=j0
We prove next uniform convergence on compact sets. Given a compact subset

K of D, there is r < 1 such that |z| ≤ r for z ∈ K. We take now j0 = j0 (K) large
enough so that |αj | > r for j ≥ j0 , and ψ̃αj (z) ∈ D̃ for j ≥ j0 and z ∈ K. This can
be done by (7.5).
We verify the Cauchy condition for the Bn in the uniform norm on K. If z ∈ K
and j0 < n < m,
m
Y
Bn (z) − Bm (z) = Bn (z)1 − ψ̃ (z)
α j
j=n+1
X
m

≤ 1 − exp log ψ̃αj (z) .
j=n+1
By (7.8),
m m
X X
log ψ̃ (z) ≤ log ψ̃α (z)
αj j
j=n+1 j=n+1
m
C 1+r X
≤ 1 − |αj | .
r 1 − r j=n+1
Given ε, 0 < ε < 1, we can then take n0 large enough so that, if n0 ≤ n < m,
m
X
log ψ̃ (z) <ε,
α j
j=n+1
for all z ∈ K. Let a > 0 be such that |1 − ew | < a|w| for |w| < 1. Then
X
m

Bn (z) − Bm (z) ≤ 1 − exp log ψ̃αj (z)

j=n+1
X
m
< a log ψ̃αj (z)
j=n+1
< aε ,
for all z ∈ K.
Finally, we prove that |B(eit )| = 1 a.e. on T. Since |B(z)| < 1 for z ∈ D,
kBkH ∞ ≤ 1, so that |B(eit )| ≤ 1 a.e. on T.
THE UNIT DISC 37
Let
Y∞
B(z)
Rn (z) = = ψ̃α (z) .
Bn (z) j=n+1 j
Then Rn is also a Blaschke product holomorphic in D, and
(7.9) |B(z)| ≤ |Rn (z)| ≤ 1 .
On the boundary we have the identity
B(eit ) = Bn (eit )Rn (eit ) ,
by Lemma 7.1. Moreover, |Bn (eit )| = 1 for every t, so that |B(eit )| = |Rn (eit )|.
Observe that
Z Z
it
|Rn (0)| = M1 (Rn , 0) ≤ |Rn (e )| dt = |B(eit )| dt .
T T
But
∞
Y P∞
log |αj |
lim Rn (0) = lim |αj | = lim e j=n+1 =1.
n→∞ n→∞ n→∞
j=n+1
Therefore, Z
|B(eit )| dt ≥ 1 ,
T
and, combining this with (7.9), we conclude that |B(eit )| = 1 a.e.

Theorem 7.7. Let f ∈ H p (D), not identically zero, and let B be its Blaschke
f
product. Then g = B ∈ H p (D) and kgkH p = kf kH p .
Proof. With Bn denoting the partial product of B. By Lemma 7.4, kf /Bn kH p =
kf kH p . Fix r < 1. Since the functions |f /Bn | converge monotonically to |f /B|, it
follows that, if 1 > r 6= |αj | for every j, then
Mp (f /B, r) = lim Mp (f /Bn , r) ≤ kf kH p .

n→∞
Therefore, g ∈ H p (D) and kgkH p ≤ kf kH p . On the other hand, since |f (z)| ≤

|g(z)| for every z ∈ D, we trivially have kf kH p ≤ kgkH p .
An immediate consequence is the following factorization theorem, a kind of in-
verse of Lemma 7.1.
Corollary 7.8. Let f ∈ H s (D), not identically zero, with 1 ≤ s ≤ ∞, and let
p, q ∈ [1, ∞] be such that p1 + q1 = 1s . Then there exist g ∈ H p (D) and h ∈ H q (D)
such that f = gh, and kgkpH p = khkqH q = kf ksH s .
f
Proof. Let B be the Blaschke product of f and ϕ = B . Then ϕ ∈ H s (D) and has
no zeroes in D. Let s s
g(z) = ϕ(z) p = e p log ϕ(z) .
s
Then |g(z)| = |ϕ(z)| p , so that
Z Z
p it p
Mp (g, r) = |g(re )| dt = |ϕ(reit )|s dt = Ms (ϕ, r)s .
T T
38 CHAPTER I
Therefore g ∈ H p (D) and kgkpH p = kϕksH s = kf ksH s .

s s s
If we set h(z) = ϕ(z) q B(z), then g(z)h(z) = ϕ(z) p + q B(z) = f (z), and moreover
Mq (h/B, r)q = Ms (ϕ, r)s ,
as before. Therefore,
khkqH q = kh/BkqH q = kϕksH s = kf ksH s .

We can prove now the F. and M. Riesz theorem.
Proof of Theorem 7.2. Let f (z) be the Poisson integral of µ. Since µ̂(n) = 0 for
n < 0, then
∞
X
f (z) = µ̂(n)z n
n=0
is holomorphic. Therefore f ∈ H 1 (D). We can assume that µ 6= 0, so that f is not

identically zero. We write f = gh, with g, h ∈ H 2 (D), according to Corollary 7.8.
By Lemma 7.1, f has a boundary function
f (eit ) = g(eit )h(eit ) ∈ L1 (T) .
It follows that dµ(t) = f (eit ) dt, i.e. µ is absolutely continuous w.r. to the Lebesgue
measure.
8. Dual spaces
The Cauchy projection

∞
X ∞
X ∞
X
n n
C : u(z) = bn z + b−n z̄ 7−→ Cu(z) = bn z n
n=0 n=1 n=0
is well defined as a linear map from the space of all harmonic functions on D, with
values in the space of homorphic functions on D. It follows from (6.1) that C maps
hp (D) into H p (D) if and only if the conjugate function operator maps hp (D) into
itself.
Therefore we know from Section 5 that this does not happen if p = 1 or ∞, and
we still have to see that it happens instead if 1 < p < ∞. The scope of this section
is to remark that this issue is relevant for determining the dual space of H p (D).
Consider the inner product in h2 (D):
Z Z X
] it ] it
(8.1) hu, vih2 = u (e )v (e ) dt = lim u(reit )v(reit ) dt = a n bn ,
T r→1 T n∈Z
if u] , v ] are the boundary functions and an , bn their Fourier coefficients (i.e. the
Taylor coefficients of u and v respectively). By the general theory of Hilbert spaces,
the map
v 7−→ ϕv (u) = hu, vih2
THE UNIT DISC 39
establishes a conjugate-linear isometric isomorphism between h2 (D) and its dual

space h2 (D)∗ .
In the same way, restricting to H 2 (D),
g 7−→ ϕg (f ) = hf, giH 2
establishes a conjugate-linear isometric isomorphism between H 2 (D) and its dual
space H 2 (D)∗ .
0
Let us see first how (8.1) can be extended to u ∈ hp (D) and v ∈ hp (D), where
p0 is the conjugate exponent of p.
0
Lemma 8.1. Let u ∈ hp (D) and v ∈ hp (D). If p = 1, assume that either u] ∈
L1 (T) or v ] ∈ C(T) (and symmetrically if p = ∞). Then the limit
Z
B(u, v) = lim u(reit )v(reit ) dt
r→1 T
R ] it
exists and it equals T u (e )v ] (eit ) dt.
Proof. Suppose first that 1 < p < ∞. Then, denoting by k kp the norm in Lp (T),
Z Z

u(re )v(reit ) dt − u (e )v ] (eit ) dt
it ] it

T T
Z Z

≤ it ] it it
u(re ) − u (e ) v(re ) dt + u (e ) v(re ) − v (e ) dt
] it it ] it
T T
] ] ]
≤ kur − u kp kvr k + ku kp kvr − v k
p0 p0
≤ kur − u] kp kvkhp0 + kukhp kvr − v ] kp0 .

By Theorem 4.3, kur − u] kp → 0, and kvr − v ] kp0 → 0, and this proves the
statement.
If p = 1 (and p = ∞ is the same), we have
Z Z

u(reit )v(reit ) dt− u] (eit )v ] (eit ) dt

T T
Z

≤ kur − u k1 kvr k∞ +
]
vr (e ) − v (e ) u] (eit ) dt .
it ] it
T
In this case, the conclusion follows from the fact that kur − u] k1 → 0, and that
vr → v ] in the weak* topology of L∞ (T), again by Theorem 4.3.
Corollary 8.2. The sesquilinear form B in Lemma 8.1 is continuous on hp (D) ×
0 0
hp (D), precisely, if u ∈ hp (D), v ∈ hp (D), then

B(u, v) ≤ kukhp kvk p0 .
h
P
If an , bn are their respective Taylor coefficients, then the series n∈Z an bn is
Abel summable, and X
lim an bn r |n| = B(u, v) .
r→1
n∈Z
Proof. The first statement follows from Lemma 8.1. For the second, it is sufficient
to observe that Z
X √ √
|n|
a n bn r = u( reit )v( reit ) dt .
n∈Z T
In terms of linear functionals on H p (D), the following is an immediate conse-

quence.
40 CHAPTER I
0
Proposition 8.3. If g ∈ H p (D), the linear functional
ϕg (f ) = B(f, g)
is continuous on H p (D), and kϕg kH p (D)∗ ≤ kgkH p0 . The map
Φ : g 7→ ϕg
0
is a norm-decreasing conjugate-linear immersion of H p (D) into H p (D)∗ .
Proof. The only fact that does not follows P
directly from Corollary 8.2 is the injec-
∞ 0
tivity of Φ. But if ϕg = 0 for some g(z) = n=0 bn z n ∈ H p (D), then
bn = B(z n , g) = 0 ,
for every n. Therefore g = 0.

It is then natural to ask if Φ is surjective onto H p (D)∗ .
For p < ∞ there is an interesting connection with the boundedness of the Cauchy
projection.
0
Theorem 8.4. Suppose p < ∞. Then Φ : H p (D) → H p (D)∗ is surjective if and
0 0
only if the Cauchy projection is bounded from hp (D) to H p (D).
Proof. By Lemma 5.2,
(8.3) kgkH p0 ≤ Akϕg kH p (D)∗ ≤ AkgkH p0 ,
for some constant A > 0.

0
Take u ∈ hp (D), and denote u] its boundary function. The linear functional
ψ(f ) = B(f, u) is continuous on H p (D), because
|ψ(f )| ≤ kf ] kp ku] kp0 = kukhp0 kf kH p .

0
In particular, kψkH p (D)∗ ≤ kukhp0 . Being Φ surjective, there is g ∈ H p (D) such
that ψ = ϕg , and
kgkH p0 ≤ AkψkH p (D)∗ ≤ Akukhp0 ,
by (8.3). We show that g = Cu, by observing that, if
∞
X ∞
X
n
u(z) = an z + a−n z̄ n ,
n=0 n=1
and
∞
X
g(z) = bn z n ,
n=0
then ψ(z n ) gives ub] (n) = an for n ≥ 0, but at the same time it equals ϕg (z n ) = bn .
Therefore,
kCukH p0 ≤ Akukhp0 .
THE UNIT DISC 41
0 0
Conversely, suppose that the Cauchy projection maps hp (D) into H p (D), and
let ψ be a continuous linear functional on H p (D). Define
ψ ] : Lp+ (T) −→ C
as ψ ] (f ] ) = ψ(f ). By the Hahn-Banach theorem, ψ ] admits a continuous extension

0
to all of Lp (T), with the same norm. So there is h ∈ Lp (T) such that
khkp0 = kψ ] kLp+ (T)∗ = kψkH p (D)∗ ,
and Z
ψ(f ) = f ] (eit )h(eit ) dt ,
T
p
for every f ∈ H (D).
0
If u ∈ hp (D) is the Poisson integral of h, so that h = u] ,
ψ(f ) = B(f, u) .
Let
∞
X ∞
X
u(z) = bn z n + b−n z̄ n ,
n=0 n=1
and
∞
X
f (z) = an z n
n=0
be the series expansions of u and f . By Corollary 8.2,

∞
X
ψ(f ) = lim an bn r n = B(f, Cu) .
r→1
n=0
0
Hence ψ = Φ(Cu), and Cu ∈ H p (D).
As we have anticipated, we shall see in Chapter III that C is bounded from hp (D)
to H p (D) for 1 < p < ∞, so that, for these values of p, Φ provides an identification
0
between H p (D)∗ and H p (D).
Since C does not map bounded harmonic functions into bounded holomorphic
functions (see Proposition 5.3), not all the continuous linear functionals on H 1 (D)
can be represented as ϕg for some g ∈ H ∞ (D).
Consider, for example, v(z) = arg(1 − z) = =m log(1 − z) ∈ h∞ (D). Then
ϕv is bounded on H 1 (D), and, for f ∈ H 1 (D), B(f, v) = B(f, g) with g(z) =
1 ∞
2i log(1 − z) 6∈ H (D).
More general examples arise from the following Hardy’s inequality.
P∞
Theorem 8.5. Let f (z) = n=0 an z n be in H 1 (D). Then
X∞
|an |
≤ kf kH 1 .
n=0
n + 1
42 CHAPTER I
Proof. We assume first that f (0) = a0 = 0, and prove in this case the stronger
inequality
X∞
|an |
≤ kf kH 1 .
n=1
n
The function Z t
g(t) = f ] (eis ) ds
0
R
is continuous on the real line, and, since T f ] (eit ) dt = f (0) = 0, it is periodic of
period 2π. Hence it defines a continuous function on T, whose derivative is f ] , and
whose Fourier coefficients are
Z 2π Z t
1
ĝ(0) = f ] (eis ) ds dt
2π 0 0
Z 2π Z 2π
1 ] is
= f (e ) dt ds
2π 0 s
Z 2π
1
= f ] (eis )(2π − s) ds
2π 0
Z 2π
1
=− sf ] (eis ) ds ,
2π 0
and
Z 2π Z t
1 −int
ĝ(n) = e f ] (eis ) ds dt
2π 0 0
Z 2π Z 2π
1 ] is
= f (e ) e−int dt ds
2π 0 s
Z 2π
1 1 −ins
= f ] (eis ) (e − 1) ds
2π 0 in
fb] (n)
= ,
in
if n 6= 0. Hence,  an
 if n > 0
ĝ(n) = in

0 if n < 0 .
It follows that the function
X∞
an n
F (z) = iĝ(0) + z ,
n=1
n
which is a holomorphic primitive of f (z)/z in D, coincides with the Poisson integral

of ig, hence it is continuous on D̄.
In particular, by Theorem 4.3,
X∞
an n
iĝ(0) + lim r = F (1) = 0 .
r→1
n=1
n
THE UNIT DISC 43
Assume first that an ≥ 0 for every n > 0. By monotone convergence,

X∞ X∞
an an n
= lim r
n=1
n r→1
n=1
n
Z 2π
i
= sf ] (eis ) ds
2π 0
Z 2π
1
≤ s|f ] (eis )| ds
2π 0
≤ kf kH 1 .
We remove now the assumption that an ≥ 0, and write f = f1 f2 , with f1 , f2 ∈

H (D), and kf1 k2H 2 = kf2 k2H 2 = kf kH 1 , according to Corollary 7.8. Since f (0) = 0,
2
one of the two factors must vanish at 0, say f1 (0) = 0.

If
X∞ ∞
X
n
f1 (z) = bn z , f2 (z) = cn z n ,
n=1 n=0
let
∞
X ∞
X
n
g1 (z) = |bn |z , g2 (z) = |cn |z n .
n=1 n=0
Then kg1 kH 2 = kf1 kH 2 , kg2 kH 2 = kf2 kH 2 . Hence g = g1 g2 ∈ H 1 (D) and

kgkH 1 ≤ kg1 kH 2 kg2 kH 2 = kf kH 1 .
Observe that X X

|an | = bj ck ≤ |bj ||ck | = ãn ,
j+k=n j+k=n
which is the n-th Taylor coefficient of g. Therefore,

X∞ ∞
|an | X ãn
≤ ≤ kgkH 1 ≤ kf kH 1 .
n=1
n n=1
n
If f (0) 6= 0, we just replace f by zf , observing that kzf kH 1 = kf kH 1 .

P∞
Corollario 8.6. Let g(z) = n=0 bn z n with bn = O(n). Then ϕg (f ) = B(f, g) is
a continuous linear functional on H 1 (D).
P∞
Proof. If f (z) = n=0 an z n ∈ H 1 (D), it follows from Hardy’s inequality that
∞
X ∞
X
n
B(f, g) = lim a n bn r = a n bn ,
r→1
n=1 n=1
by dominated convergence. Therefore,

B(f, g) ≤ kf kH 1 sup |bn | .
n∈N n + 1
To conclude, the case p = ∞ must be treated separately. As one can reasonably

expect, the dual space of H ∞ (D) is strictly bigger than Φ H 1 (D) . This can be
shown in many ways.
44 CHAPTER I
One argument is based on the following remark. Let HC(D) the space of holo-
morphic functions that admit a continuous extension to D̄. Then HC(D) is a
closed subspace of H ∞ (D), and proper, because it does not contain the infinite
Blaschke products. By the Hahn-Banach theorem, there is a non-zero continuous
linear functional ϕ on H ∞ (D) which vanishes identically on HC(D). Suppose that
ϕ(f ) = B(f, g) for some g ∈ H 1 (D), and let
∞
X
g(z) = an z n .
n=0
Since z n ∈ HC(D) for n ≥ 0,
0 = B(z n , g) = an ,
so that g = 0. But this contradicts the fact that ϕ is not identically zero.
This argument motivates a new question: can Φ be surjective from H 1 (D) onto
the dual space of HC(D)? The answer is negative again. It is possible to modify
the first part of the proof of Theorem 8.4 to show that, if this is the case, then the
Cauchy projection maps h1 (D) into H 1 (D), which we know to be false.
THE HALF-PLANE 45
CHAPTER II
HARDY SPACES ON THE HALF-PLANE
1. Definitions and basic facts
Let now D+ be the upper half-plane in C, D+ = {x + iy : y > 0}.
Definition. If f is holomorphic (resp. harmonic) in D+ , we say that f ∈ H p (D+ )

(resp. f ∈ hp (D+ )), for 1 ≤ p ≤ ∞, if for every y > 0 fy (x) = f (x + iy) is in
Lp (R), and
(1.1) sup kfy kp < ∞ .

y>0
We keep the notation Mp (f, r) for kfr kp , and kf kH p (resp. kf khp ) for the supre-
mum in (1.1).
Lemma 1.1. If f ∈ hp (D+ ), and z = x + iy ∈ D+ , then
kf khp
|f (z)| ≤ Cp 1 .
yp
Proof. We assume that p < ∞, the other case being trivial.

For every r < y,
Z π
1
f (z) = f (z + reit ) dt ,
2π −π
by the mean value property. Integrating in polar coordinates around z, we then

have
Z Z yZ π
1 1
f (w) dw = f (z + reit ) dt r dr
|B(z, y)| B(z,y) πy 2 0 −π
Z y
(1.2) 2
= 2 f (z) r dr
y 0
= f (z) .
Therefore, using Hölder’s inequality and the inclusion B(z, y) ⊂ Sy = {u + iv :
Typeset by AMS-TEX
46 CHAPTER II
0 < v < 2y},

Z
1
|f (z)| ≤ |f (w)| dw
πy 2 B(z,y)
Z p1
1
≤ |f (w)|p dw
πy 2 B(z,y)
Z p1
1 p
≤ |f (w)| dw
πy 2 Sy
Z p1
1 p
= |f (u + iv)| du dv
πy 2 Sy
Z 2y p1
1 p
≤ Mp (f, v) dv
πy 2 0
Z 2y p1
1
≤ kf khp dv
πy 2 0
2 p1
= kf khp .
πy
As for the corresponding spaces on the unit disc, we then have the following
immediate consequence.
Theorem 1.2. H p (D+ ) and hp (D+ ) are Banach spaces, and convergence in their
norm implies uniform convergence on compact sets.
The information given by Lemma 1.1 concerns the behaviour of an hp -function f
both for y → 0 and for y → ∞. It does not say anything, however, on the behaviour
of f (z) when z tends to infinity within a horizontal strip. The best one can say is
the following.
Lemma 1.3. For 0 < a < b, let Sa,b = {x + iy : a ≤ y ≤ b}. If f ∈ hp (D+ ), with
p < ∞, , then
lim f (z) = 0 .
z→∞, z∈Sa,b
Proof. For z ∈ Sa,b , let Bz the disc centered at z of radius a. By (1.2),

Z Z p1
1 1 p
(1.3) |f (z)| ≤ |f (w)| dw ≤ |f (w)| dw .
πa2 Bz πa2 Bz
Observe that Bz ⊂ S0,b+a , and that

Z Z b+a Z
p
|f (w)| dw = |f (x + iy)|p dx dy ≤ (b + a)kf khp < ∞ .
S0,b+a 0 R
Therefore, given ε > 0, there is M > 0 such that

Z b+a Z
(1.4) |f (x + iy)|p dx dy < πa2 εp .
0 |x|>M
THE HALF-PLANE 47
If |<ez| > M + a, Bz ⊂ {x + iy : |x| > M, 0 < y < b + a}. Putting (1.3) and
(1.4) together, we obtain that |f (z)| < ε.
In contrast with the corresponding spaces on the unit disc, no Hardy space on
D+ is contained in any other, as it can be seen with appropriate examples. For
instance, if α > 0,
1
fα (z) =
(z + i)α
is in H p (D+ ) if and only if pα > 1, and
1
gα (z) =
z α (z + i)2
is in H p (D+ ) if and only if pα < 1.
Before discussing Poisson integrals in D+ , we must recall some facts about con-
volution on R on one side, and about conformal mappings on the other side.
The convolution of two continuous functions f and g with compact support is

defined as
Z
(2.1) f ∗ g(x) = f (x − t)g(t) dt .
R
By a simple change of variables, one verifies that f ∗ g = g ∗ f . It is also easy to

verify that
supp (f ∗ g) ⊆ supp f + supp g = {x + x0 : x ∈ supp f, y ∈ supp g} .
When it will be necessary to distinguish between convolution on T and convolu-

tion on R, we shall write ∗T and ∗R accordingly.
The integral (2.1) may not make sense for more general functions, unless certain
integrability conditions are satisfied22 . One can prove that the integral (2.1) is
convergent for almost every x if f ∈ Lp (R), g ∈ Lq (R), with 1 ≤ p, q ≤ ∞ and
1 1
(2.2) + ≥1.
p q
In this case f ∗ g ∈ Lr (R), with
1 1 1
= + −1 ,
r p q
and the Young inequality
(2.3) kf ∗ gkr ≤ kf kp kgkq

22 Observe that if f (x) = (1 + |x|)−α and g(x) = (1 + |x|)−β , the integral (2.1) is divergent for
every x if α + β ≤ 1. Compare this example with the restrictions on the exponents p and q below.
48 CHAPTER II
holds. When r = ∞ (i.e. p and q are conjugate exponents), f ∗ g is also continuous.

If, in addition, 1 < p, q < ∞, f ∗ g vanishes at infinity.
The convolution of two finite, regular Borel measures µ, ν ∈ M (R) is defined as
the measure µ ∗ ν such that
Z Z Z
(2.4) f (x) d(µ ∗ ν)(x) = f (s + t) dµ(s) dν(t) ,
R R R
for f ∈ C0 (R). One has the inequality

(2.5) kµ ∗ νk1 ≤ kµk1 kνk1 .
In a more subtle way, the convolution µ ∗ f of a measure µ ∈ M (R) with a
function f ∈ Lp (R), 1 ≤ p ≤ ∞, is also well defined, and
(2.6) kµ ∗ f kp ≤ kµk1 kf kp .
Recall now that if A and B are two connected, simply connected, proper open
subsets of C, there exists a conformal mapping23 ϕ from A onto B. For A = D,
the unit disc, and B = D+ , one can explicitely write such mappings. One of them
is
1+z
(2.7) ϕ(z) = i ,
1−z
and it is called the Cayley transform.
Lemma 2.1. The Cayley transform ϕ maps D onto D+ , it is invertible, and
w−i
(2.8) ϕ−1 (w) = .
w+i
Moreover ϕ has a continuous extension to D̄ \ {1}, and
θ
(2.9) ϕ(eiθ ) = − cot ∈ R = ∂D+ .
2
Proof. One has
(1 + z)(1 − z̄) 1 − |z|2 + 2i=mz
ϕ(z) = i = i ,
|1 − z|2 |1 − z|2
so that
1 − |z|2
=mϕ(z) = >0,
|1 − z|2
if z ∈ D . Hence ϕ maps D into D+ .
It is easy to verify that ϕ is injective and that (2.8) gives its inverse function. If
w ∈ D+ , |w −i| < |w +i| by a simple geometric consideration, so that |ϕ−1 (w)| < 1.
This shows that ϕ is onto.
The extension of ϕ to the boundary is easy to derive.
We can use the Cayley transform to trasfer harmonic functions from D to D+

and viceversa.
23 i.e. holomorphic and invertible. The fact we are stating is called the Riemann mapping
theorem.
THE HALF-PLANE 49
Lemma 2.2. Let A, B be open subsets of C. If ϕ : A → B is holomorphic, and u

is harmonic on B, then u ◦ ϕ is harmonic on A.
Proof. We can assume that u is real-valued. Fix z ∈ A, and let U ⊂ B be a
circular neighborhood of w = ϕ(z). By Corollary 2.6 od Chapter I, u|B is the real
part of a holomorphic function f . Then f ◦ ϕ is holomorphic on the neighborhood
V = ϕ−1 (U ) of z, and u ◦ ϕ = <e(f ◦ ϕ) is harmonic on V .
Lemma 2.3. Let u be continuous on D+ , harmonic in D and bounded. Then
Z
1 y
(2.10) u(x + iy) = u(t) dt .
π R (x − t)2 + y 2
Proof. If ϕ is the Cayley transform, the function v(z) = u ◦ ϕ(z) is in h∞ (D). In

addition, v is continuous on D̄ \ {1}. By dominated convergence,
Z Z
iθ
v(e ) dθ = lim v(reiθ ) dθ = v(0) .
T r→1 T
Since v(0) = u(i) and v(eiθ ) = u(− cot 2θ ) by (2.9), the change of variable − cot θ
2 =t
gives
Z 2π Z
1 iθ 1 +∞ 1
u(i) = v(e ) dθ = u(t) 2 dt ,
2π 0 π −∞ t +1
which is (2.10) for x + iy = i.
For a general point x0 + iy0 ∈ D+ , consider the function ũ(z) = u(x0 + y0 z),
which satisfies the same hypotheses as u. Then, with simple changes of variables,
u(x0 + iy0 ) = ũ(i)

Z
1 +∞ 1
= u(x0 + y0 t) 2 dt
π −∞ t +1
Z
1 +∞ y0
= u(x0 + t) 2 dt
π −∞ t + y02
Z
1 +∞ y0
= u(t) dt .
π −∞ (x0 − t)2 + y02
Definition. For y > 0, the function
1 y
Py (x) =
π x + y2
2
is called the Poisson kernel on R.

Formula (2.10) can be written as
(2.11) u y = u0 ∗ Py .
50 CHAPTER II
Corollary 2.4. Let u ∈ hp (D+ ), with 1 ≤ p ≤ ∞. Given 0 < y1 < y2 , then
uy2 = uy1 ∗ Py2 −y1 .
In particular, the Poisson kernel has the semigroup property
(2.12) Py+y 0 = Py ∗ Py 0 .
Proof. Let v(z) = u(z + iy1 ). By Lemma 1.1, v satisfies the assumptions of Lemma
2.3. Therefore
uy2 = vy2 −y1 = v0 ∗ Py2 −y1 = uy1 ∗ Py2 −y1 .
Applying this identity to
u(x + iy) = Py (x) ∈ h1 (D+ )

1
(observe that u is harmonic because it equals − π1 =m x+iy ), we obtain (2.12).
A further important property of the Poisson kernel is that it forms an approx-
imate identity on the real line. By definition, an approximate identity on R (for
ε → 0) is a family of functions {ϕε }ε>0 satisfying
R
(1) R |ϕε (t)|
R dt ≤ C for some constant C and every ε;
(2) limε→0 R ϕε (t) dt = 1 for every r;
(3) for every δ > 0, Z
lim |ϕε (t)| dt = 0 .
ε→0 |t|>δ
Proposition 3.2 of Chapter I extends to approximate identities on R.

R
Lemma 2.5. Take ϕ ∈ L1 (R) such that R ϕ(x) dx = 1. Then the functions

ϕε (x) = 1ε ϕ xε form an approximate identity for ε → 0. In particular, the Poisson
kernels Py form an approximate identity for y → 0.
R
Proof.
R A simple change of variable shows that kϕ ε k 1 = kϕk 1 and ϕ (t) dt =
R ε
R
ϕ(t) dt = 1. This proves (1) and (2). By the same change of variable,
Z Z
|ϕε (t)| dt = |ϕ(t)| dt ,
|t|>δ |t|> εδ
which tends to 0 with ε.

For the Poisson kernel, observe that
1 x
(2.13) Py (x) = P1 ,
y y
and Z
1 1
kP1 k = dx = 1 .
π R x2 +1
We are now in a position to extend to the spaces hp (D) the results proved about
norm or weak* convergence to the boundary. We summarize them in the following
statement.
THE HALF-PLANE 51
Theorem 2.6. If u ∈ hp (D+ ), then Mp (u, y) = kuy kp is decreasing in y, so that

kukhp = lim Mp (u, y) .
y→0
The operator P mapping f (understood as either a function or a Borel measure

on R) into the harmonic function Pf on D+ given by
(Pf )(x + iy) = f ∗ Py (x) ,
maps Lp (R) isometrically onto hp (D+ ) for 1 < p ≤ ∞, and it maps M (R) isomet-
rically onto h1 (D+ ).
The limit limy→0 (Pf )y exists in Lp if and only if one of the following holds
(1) f ∈ Lp (R) and 1 ≤ p < ∞;
(2) p = ∞ and f ∈ C(R).
In each of these cases, (Pf )y → f in the Lp -norm.
For general elements f of M (R) or L∞ (R), (Pf )y tends to f in the corresponding
weak*-topology.
Proof. Once we have proven that Pf is harmonic, it is a matter of adapting the
proof of Theorem 4.3 in Chapter I.
We verify the mean value property for u = Pf assuming that f is a function
(the case f ∈ M (R) is left to the reader). If B(z0 , r) ⊂ D+ , then
Z Z Z
iθ
(4.6) u(z0 + re ) dθ = P (z0 + reiθ − t)f (t) dt dθ .
T T R
We can apply Fubini’s theorem and change the order of integration for the fol-
lowing reason. For fixed θ, write z0 + reiθ = xθ + iyθ . Then
P (z0 + reiθ − t)f (t) = P (xθ + iyθ − t)f (t)
0
is integrable in t, because it is the product of an Lp - and an Lp -function. By (2.13),
Z q1
1 x q Cq
kPy kq = q
P 1 dx = 1 .
R y y y 1− q
if 1 ≤ q < ∞, and similarly for q = ∞.
There is a δ > 0 such that =myθ ≥ δ for all θ. Therefore
Z

P (z0 + reiθ − t)f (t) dt ≤ Cp,δ kf kp
R
uniformly in θ. Therefore the double integral in (4.6) is absolutely convergent.

We apply Fubini’s theorem, and use the fact that the function P (x + iy) is
harmonic in D+ , as observed in the proof of Corollary 2.4. Then
Z Z Z
iθ
u(z0 + re ) dθ = f (t) P (z0 + reiθ − t) dθ dt
T
ZR T
= f (t)P (z0 − t) dt
R
= u(z0 ) .
The following corollary can be seen as a replacement for the (non-existing) in-
clusion relations among the hp (D+ ).
52 CHAPTER II
Corollary 2.7. If 1 < p < ∞, and 1 ≤ q ≤ ∞, then hp (D+ ) ∩ hq (D+ ) is dense in

hp (D+ ). This does not hold if p = 1, ∞ and q 6= p.
Proof. By Theorem 2.6, it is sufficient to prove the corresponding statement for
Lp (R) (1 < p ≤ ∞) and M (R).
If f ∈ Lp (R) with 1 < p < ∞, and 1 ≤ q < p, then fR = f χ[−R,R] is in both
L (R) and Lq (R), and fR → f in the Lp -norm as R → ∞. If p < q ≤ ∞, the same
p
holds for fR = f χ{|f |<R} .

For p = 1 6= q, observe that M (R) ∩ Lq (R) = L1 (R) ∩ Lq (R), so that its closure
in M (R) is only L1 (R).
For p = ∞, observe that the constant function 1 cannot be the uniform limit of
functions in Lq (R) if q < ∞.
We must also mention the behaviour of Mp (u, y) as y → +∞.
Proposition 2.8. If 1 0, there is a continuous function ϕ on R with compact support
such that ku] − ϕkp < ε. Then, for y > 0,
Mp (u, y) = ku] ∗ Py kp
≤ k(u] − ϕ) ∗ Py kp + kϕ ∗ Py kp
< ε + kϕk1 kPy kp
− p10
≤ε+y kϕk1 .
So, if y is large enough, Mp (u, y) < 2ε.
3. The Fourier transform and the Paley-Wiener theorem
We begin this section by recalling some of the basic facts about the Fourier
transform on R.
If f ∈ L1 (R) and ξ ∈ R, one defines the Fourier transform of f at ξ as
Z
(3.1) fˆ(ξ) = f (x)e−iξx dx .
R
The inequality |fˆ(ξ)| ≤ kf k1 is an immediate consequence of the definition. The

function fˆ so defined is continuous on R and vanishes at infinity (this last fact is
known as the Riemann-Lebesgue theorem). Then the linear operator F mapping
f into F f = fˆ is continuous from L1 (R) into C0 (R).
One also defines the Fourier transform of a finite Borel measure µ ∈ M (R) as
Z
µ̂(ξ) = e−iξx dµ(x) .
R
Then |µ̂(ξ)| ≤ kµk1 , µ̂ is continuous, but we can no longer say that it vanishes
at infinity.
THE HALF-PLANE 53
Further properies of F are the following.

(1) if fˇ(x) = f (−x), then fbˇ(ξ) = fˆ(−ξ);
(2) fb̄(ξ) = fˆ(−ξ);
(3) f[ ∗ g(ξ) = fˆ(ξ)ĝ(ξ);
(4) if fε (x) = 1ε f ( xε ), then fbε (ξ) = fˆ(εξ);
(5) if f is absolutely continuous, so that also f 0 ∈ L1 (R), then fb0 (ξ) = iξ fˆ(ξ);
(6) d)(ξ);
if both f and xf are integrable, than fˆ is C 1 and (fˆ)0 (ξ) = −i (xf
(7) if f ∈ L1 (R) ∩ L2 (R), then fˆ ∈ L2 (R) and the Plancherel formula holds:
Z Z
1
(3.2) fˆ(ξ)2 dξ = |f (x)|2 ;
2π R R
√
(8) because of (3.2), F extends to every f ∈ L2 (R), and kfˆk2 = 2πkf k2 ;
(9) since Lp (R) ⊂ L1 (R) + L2 (R), if 1 < p < 2, F is well defined on Lp (R),
fˆ ∈ Lp (R), and the Hausdorff-Young inequality holds:
0
(3.3) kfˆkp0 ≤ kf kp ;
(10) if fˆ is integrable24 , the inversion formula holds:

Z
1
(3.4) f (x) = fˆ(ξ)eixξ dξ .
2π R
Lemma 3.1. The Fourier transform of Py is
cy (ξ) = e−y|ξ| .
P
If u ∈ hp (D+ ), with 1 ≤ p ≤ 2, and u] is its boundary function (or measure),

then
(3.5) cy (ξ) = ub] (ξ)e−y|ξ| .

u
Proof. By (4) above and (2.13),
cy (ξ) = P
P c1 (yξ) ,
and by (3) and (2.12),

\
P c 0 c c 0
y+y 0 (ξ) = P1 (y + y )ξ = P1 (yξ)P1 (y ξ) .
c1 is real and even. Putting these information together

Moreover, by (1) and (2), P
with the fact that Pc1 ∈ C0 (R), we find that P c1 (ξ) = e−a|ξ| for some a > 0.
24 This happens, in particular, if f is differentiable and f 0 ∈ L2 (R).

54 CHAPTER II
c1 ∈ L1 (R), so that we can use the inversion formula in order to

In particular, P
determine a. We are so led to compute
Z Z
1 −a|ξ| ixξ 1 ∞ −aξ
e e dξ = e cos(xξ) dξ
2π R π 0
Z ∞
1 −(a−ix)ξ
= <e e dξ
π 0
1 1
= <e
π a − ix
1 a
= .
π x + a2
2
This must be equal to P1 (x), hence a = 1. The rest of the proof is obvious.
The Fourier transform allows to describe for which functions f ∈ Lp (R), the
Poisson integral u(x + iy) = f ∗ Py (x) is holomorphic. The main result concerns
p = 2, and is called the Paley-Wiener theorem.
Theorem 3.2. A function u ∈ h2 (D+ ) is holomorphic if and only if the Fourier
transform ub] of its boundary function u] ∈ L2 (R) is zero on the negative half-
line. Hence a holomorphic function f on D+ is in H 2 (D+ ) if and only if f is the
Fourier-Laplace transform,
Z ∞
1
(3.6) f (z) = g(ξ)eizξ dξ ,
2π 0
of some g ∈ L2 (R+ ), and in this case kf kH 2 = √1 kgk2 .

2π
Proof. Suppose that u is holomorphic. Calling γa the horizontal curve γa (t) = t+ia,
consider the line integral
Z Z ∞
−iξz
Ia = u(z)e dz = u(x + ia)e−iξ(x+ia) dx = eaξ u
ca (ξ) .
γa −∞
We claim that Ia = Ib if a, b > 0. Suppose a < b, fix M > 0 and let RM be

the rectangle with vertices −M + ia, M + ia, M + ib, −M + ib. Denote by γa,M ,
+ −
γb,M the arcs of γa and γb describing the horizontal edges of RM , and by γM , γM
the arcs describing the vertical edges, oriented upwards. By the Cauchy integral
formula applied to the boundary of RM , we have
Z Z
−iξz
0= u(z)e dz + u(z)e−iξz dz
+
γa,M γM
(3.7) Z Z
−iξz
− u(z)e dz − u(z)e−iξz dz .
−
γb,M γM
Now,
Z Z b
−iξz
u(z)e dz = i u(M + it)e−iξ(M +it) dt ,
+
γM a
THE HALF-PLANE 55
so that
Z Z b

u(z)e −iξz
dz ≤ |u(M + it)|eξt dt
+
γM a
≤ eb|ξ| (b − a) max |u(M + it)| ,

t∈[a,b]
which tends to 0 as M → ∞ by Lemma 1.3. Then, passing to the limit as M → ∞

in (3.7), we obtain that Ia = Ib .
So, by (3.5),
ca (ξ) = e−a(|ξ|−ξ) ub] (ξ)
eaξ u
is independent of a. This forces ub] (ξ) to be a.e. zero for ξ < 0.

Since ub] ∈ L2 (R), then u
cy = e−y|ξ| ub] ∈ L1 (R) for y > 0, and the inversion
formula (3.6) gives
Z ∞ Z ∞
1 1
u(x + iy) = ub] (ξ)e−yξ eixξ dξ = ub] (ξ)ei(x+iy)ξ dξ .
2π 0 2π 0
Conversely, take g ∈ L2 (R+ ) and define f by (3.6). The integral is absolutely

convergent, because |eizξ | = e−ξ=mz ∈ L2 (R+ ). If γ is a closed arc in D+ , then
Z Z Z ∞
1
f (z) dz = g(ξ)eizξ dξ dz .
γ 2π γ 0
If the support of γ is contained in the half-plane =mz ≥ a > 0, then |eizξ | ≤ e−aξ ,
and g(ξ)eizξ is integrable on γ × R+ . We can then change order of integration and
obtain that Z Z ∞ Z
1
f (z) dz = g(ξ) eizξ dz dξ = 0 ,
γ 2π 0 γ
because eizξ is holomorphic in z. Then f is holomorphic by Morera’s theorem.

Observe that fy is the inverse Fourier transform of g(ξ)e−yξ χR+ . Therefore, by
Plancherel’s formula (7),
1
kfy k2 = √ ke−y· gk2 ,
2π
and
1 1
kf kH 2 = √ lim ke−y· gk2 = √ kgk2 .
2π y→0 2π
Corollary 3.3. Let 1 ≤ p ≤ 2, u ∈ hp (D+ ) and u] be its boundary function (or

measure). Then u is holomorphic if and only if ub] is zero on the negative half-line. If
this is the case, then u is the Fourier-Laplace transform (3.6) of g = ub] ∈ Lp (R+ ).
0
Proof. Let v(ε) (z) = u(z + iε). Then v(ε) ∈ hp (D+ ) ∩ h∞ (D+ ), hence in h2 (D+ ).
If we assume that u ∈ H p (D+ ), then v(ε) ∈ H 2 (D+ ). Therefore,
d]
v(ε) cε (ξ) = e−ε|ξ| ub] (ξ)
(ξ) = u
56 CHAPTER II
cy = e−y|·| ub] ∈ L1 (R) for

is zero on the negative half-line. By Hölder’s inequality, u
y > 0, so that the Fourier inversion formula gives (3.6) as before.
Conversely, if u ∈ hp (D+ ) and supp ub] ⊂ [0, +∞), the same is true for v(ε) , which
belongs to h2 (D+ ). Therefore, v(ε) is holomorphic in D+ , i.e. u is holomorphic in
D+ + iε. By the arbitrarity of ε, u is holomorphic in D+ .
4. The Cauchy projection and the Hilbert transform
By the Paley-Wiener theorem and Theorem 2.6, the map which assigns to a
harmonic function u ∈ h2 (D+ ) the Fourier transform ub] is, up to a constant, an
isometry of h2 (D+ ) onto L2 (R), and the image of H 2 (D+ ) is
L2+ = {g ∈ L2 (R) : supp g ⊆ [0, +∞)} ∼ L2 (R+ ) .
Since the orthogonal projection of L2 (R) onto L2+ is obviously the map g 7→
gχR+ , we derive the following recipe for constructing the orthogonal projection
from h2 (D+ ) onto H 2 (D+ ): starting with u ∈ h2 (D+ ), take ub] , multiply it by χR+ ,
and apply (3.6). This gives the Cauchy projection on D+ ,
Z ∞
1
Cu(x + iy) = ub] (ξ)ei(x+iy)ξ dξ
2π 0
(4.1)
= F −1 ub] e−yξ χR+ (x)
= u] ∗ F −1 (e−yξ χR+ )(x) .
Define the Cauchy kernel Cy (x) = F −1 (e−yξ χR+ )(x). Explicitely,

Z ∞
1
Cy (x) = e−yξ eixξ dξ
2π 0
(4.2)
1 1
=− .
2πi x + iy
We can then write (4.1) in the following way.

Theorem 4.1. The orthogonal projection C from h2 (D+ ) onto H 2 (D+ ) is given
by
Z ∞
] 1 u] (t)
(4.3) Cu(z) = u ∗ Cy (x) = dt ,
2πi −∞ t−z
with z = x + iy.
If u is already in H 2 (D+ ), then Cu = u, and (4.3) formally coincides with the
ordinary Cauchy integral formula
Z
1 u] (w)
u(z) = dw ,
2πi γ0 w−z
THE HALF-PLANE 57
where we denote by γ0 the curve describing the real axis with its natural orientation.
Notice however that the curve is not bounded25 and it is not contained in the interior
of the domain where u is holomorphic.
Write now
1 x − iy 1 y i x 1
Cy (x) = − = + = P y (x) + i P̃ y (x) ,
2πi x2 + y 2 2π x2 + y 2 2π x2 + y 2 2
noticing that the real part of 2Cy is the Poisson kernel Py and setting
1 x
(4.4) P̃y (x) = .
π x + y2
2
Therefore (4.3) takes the form
1 i
(4.5) Cu(x + iy) = u(x + iy) + u] ∗ P̃y (x) .
2 2
Observe however that P̃y is not integrable for any y > 0, so that the convolution
u ∗ P̃y is not defined for u ∈ h∞ (D+ ). On the other hand, P̃y is in Lq (R) for any
]
q > 1 and in C0 (R), so that the convolution is well defined for u ∈ hp (D+ ) with
p < ∞.
Proposition 4.2. Let u ∈ hp (D+ ) with p < ∞. Then
ũ(x + iy) = u] ∗ P̃y (x)
is harmonic, u + iũ is holomorphic, and
lim ũ(z) = 0 ,
z→∞ , =mz≥a
for every a > 0. If u is real-valued, the ũ is the only real-valued function satisfying
the above properties.
Proof. We verify the mean value property for ũ. If B(z0 , r) ⊂ D+ , then
Z Z Z
iθ
(4.6) ũ(z0 + re ) dθ = P̃ (z0 + reiθ − t)u] (t) dt dθ .
T T R
We can apply Fubini’s theorem and change the order of integration for the fol-
lowing reason. For fixed θ, write z0 + reiθ = xθ + iyθ . Then
P̃ (z0 + reiθ − t)u] (t) = P̃ (xθ + iyθ − t)u] (t)

0
is integrable in t, because it is the product of an Lp - and an Lp -function. Since
1 x
P̃y (x) = P̃1 ,
y y
25 It becomes a nice closed curve if embedded in the Riemann sphere.
58 CHAPTER II
and P̃1 ∈ Lq (R) for q > 1, we have, as in the proof of Theorem 2.6,
1
kP̃y kp0 ≤ Cp y − p ,
for p < ∞.
The proof continues as for Theorem 2.6.
Observe that in the proof we have implicitely obtained the inequality
1
(4.7) |ũ(x + iy)| ≤ Cp y − p kukhp .
for u ∈ hp (D+ ) and p < ∞.

By (4.5),
ũ = iu − 2iCu ,
so that, if u ∈ h2 (D+ ), also ũ ∈ h2 (D+ ). If ũ] is the boundary function, we have
(4.8) ub̃] (ξ) = iub] (ξ) − 2iχR+ (ξ)ub] (ξ) = −i(sgn ξ)ub] (ξ) ,
so that, by Plancherel’s inequality,
(4.9) kũkh2 = kukh2 .
The operations described in (4.8) take place on the real line, and define a linear
operator H : L2 (R) → L2 (R),

Hf = F −1 − i sgn ξ fˆ(ξ) .
This operator is called the Hilbert transform. Proposition 6.2 in Chapter I has
the following analogue (with a very similar proof, that we omit).
Proposition 4.3. For f ∈ L2 (R),
Z
1 1
Hf (x) = lim f (x − t) dt ,
ε→0 π |t|>ε t
where the limit is in the L2 -norm.
Much of the content of Sections 5 and 6 of Chapter I can be repeated here, with
the unit disc replaced by the upper half-plane.
The Cauchy projection is bounded on hp (D+ ) for p 6= 2 if and only if the same
is true for the conjugate function operator mapping u to ũ. As for the unit disc
(and with the same arguments) this turns out to be false if p = 1. Also for p = ∞,
when the formulas above do not hold, it is not true in general that the conjugate
functions of a bounded harmonic function are bounded26 . This is not even true if
u is bounded and continuous on D+ .
We shall see later that C is bounded from hp (D+ ) onto H p (D+ ) for 1 < p < ∞.
26 Takeu(x + iy) = arccot x

y
= arg z = =m log z. It is bounded, but its harmonic conjugates,
− log |z| + constant, are unbounded.
THE HALF-PLANE 59
5. Transference of H p -functions and applications
We shall now establish simple relations between H p -functions on the unit disc
and the upper half-plane. This will allow us to transfer to D+ many results obtained
in Chapter I for D.
The simplest situation occurs for p = ∞: if ϕ denotes the Cayley transform
(2.7), the map
(5.1) f 7−→ f ◦ ϕ−1 = T∞ f
is an isometry of H ∞ (D) onto H ∞ (D+ ), for the simple reason that f is bounded
if and only if f ◦ ϕ−1 is bounded, and the two ∞-norms are equal27 .
Clearly the same does not hold for p < ∞ (take f = 1). In order to see what
needs to be modified, let us consider the simple case where f is continuous on D̄
and holomorphic in the interior.
Then g = f ◦ ϕ−1 is continuous on D+ , and let us try to compute the Lp -norm
of g|R . By (2.9), with the change of variable x = − cot θ2 we have
Z ∞ p1
−1 p
kg|R kp = |f ◦ ϕ (x)| dx
−∞
Z 2π p1
1 iθ p dθ
= |f (e )| ,
2 0 sin2 θ2
and this differs from the Lp -norm of g by the factor sin−2 θ2 coming from the change
−2i
of variable. This suggest that we must multiply g by a power of (ϕ−1 )0 (z) = (z+i) 2.
p
Precisely, for p < ∞ and f ∈ H (D), we set
1
(5.2) Tp f (z) = 1 2 f ◦ ϕ−1 (z) .
π (z + i)
p p
Proposition 5.1. If 1 < p ≤ ∞, Tp is an isometry of H p (D) onto H p (D+ ). If

p = 1, T1 is an isometry of H 1 (D) onto the subspace H01 (D+ ) of those f ∈ H 1 (D+ )
such that f ] ∈ L1 (R).
At the end of this section we shall prove that in fact H01 (D+ ) = H 1 (D+ ), as a
consequence of the F. and M. Riesz theorem for R.
Proof. The case p = ∞ having been discussed already, we take p < ∞. Assume
first that f is continuous on D̄, bounded and holomorphic in the interior, and let
g = Tp f . Then g is continuous on D+ and
Z
p 1 p 1
Mp (g, y) ≤ kf k∞ 2 2
dx ≤ kf kp∞ ,
π R x + (y + 1)
so that g ∈ H p (D+ ). Let g ] be its boundary function (or measure) on R, in the

sense of Theorem 2.6. If 1 < p < ∞, gy → g ] in the Lp -norm as y → 0, there is a
sequence yn → 0 such that gyn → g ] almost everywhere. Therefore g ] = g|R .
27 This
is true also for bounded harmonic functions, but what we shall do next for H p -functions
with p < ∞ has no analogue for harmonic functions.
60 CHAPTER II
If p = 1 we get to the same conclusion by a different argument. Since gy → g ]

in the weak* topology, for any ϕ ∈ Cc (R),
Z Z Z
]
ϕ(x) dg (x) = lim ϕ(x)g(x + iy) dx = ϕ(x)g(x) dx .
R y→0 R R
Hence dg ] (x) = g(x) dx.

We now have
Z ∞ p1
1 −1 p dx
kg|R kp = |f ◦ ϕ (x)| .
π −∞ |x + i|2
If x = ϕ(eiθ ), then
1 + eiθ 2i
x+i=i + i = ,
1 − eiθ 1 − eiθ
so that
1 2 θ
= sin .
|x + i|2 2
θ dx
Therefore the change of variable x = − cot 2
gives |x+i|2
= 12 dθ and then
kg|R kp = kf|T kp .
We have so proved that Tp is an isometry from a dense subspace of H p (D) (see
Corollaries 4.3 and 7.3 in Chapter I) onto its image, call it V , in H p (D+ ). Therefore
Tp extends to an isometry (also denoted by Tp ) of all H p (D) onto the closure of V
in H p (D+ ). The proof will be finished if we prove that Tp has a dense image in
H p (D+ ) if p > 1 or in H01 (D+ ) if p = 1.
Take g ∈ H p (D+ ), resp. in H01 (D+ ), ε > 0, and define
g(z + iε)
hε (z) = .
(1 − iεz)2
Observe that for z = x+iy ∈ D+ , |1−iεz| ≥ 1+εy ≥ 1, so that the denominator
does not vanish, and moreover
|hε (z)| ≤ |g(z + iε)| .
This implies that hε ∈ H p (D+ ). We prove that limε→0 hε = g in H p (D+ ). In
fact
Z p p1
gε (x) ]
khε − gkH p = 2
− g (x) dx
R (1 − iεx)
Z p1
|gε (x) − g ] (x)|p
≤ dx
R |1 − iεx|2p
Z p p1
] p 1
+ |g (x)| 2
− 1 dx
R (1 − iεx)
Z p1
] p
≤ |gε (x) − g (x)| dx
R
Z p p1
] p 1
+ |g (x)| 2
− 1 dx .
R (1 − iεx)
THE HALF-PLANE 61
As ε → 0, the first
term tends to 0 by Theorem 2.6, and the second by dominated
convergence, since (1−iεx)
1
2 − 1 ≤ 2.
We show that hε is the image under Tp of a function in H p (D). We compute

therefore
1 2
Tp−1 hε (z) = π p ϕ(z) + i p hε ◦ ϕ(z)
1
2i p2 1
=π p
2 g ϕ(z) + iε
1−z 1 − iεϕ(z)
1
2i p2 (1 − z)2
= πp 2 g ϕ(z) + iε
1−z 1 + ε − (1 − ε)z
2
1 2 (1 − z) p0
= π (2i)
p p
2 g ϕ(z) + iε .
1 + ε − (1 − ε)z

The factor g ϕ(z) + iε is bounded because =m ϕ(z) + iε > ε, and the denom-
inator does not vanish on D̄. therefore Tp−1 hε is bounded, hence in H p (D).
This fact will allow us to transfer the factorization results proved in the unit disc
to D+ .
Lemma 5.2. Let 1 ≤ p ≤ ∞. If f ∈ H p (D+ ) is not identically zero and {αj }j∈N
is the sequence of its zeroes (repeated according to their multiplicities), then
X∞
=mαj
(5.3) 2+1
<∞.
j=0
|α j |
Proof. For p > 1, we take g = Tp−1 f ∈ H p (D), and apply Lemma 7.5 in Chapter I.
The zeroes of g are the points
αj − i
βj = ϕ−1 (αj ) = .
αj + i
Then
∞
X ∞
X
2
(1 − |βj | ) (1 − |βj |) < ∞ ,
j=0 j=0
and, if αj = aj + ibj ,
|αj + i|2 − |αj − i|2

1 − |βj |2 =
|αj + i|2
4=mαj
=
|αj + i|2
4bj
= 2
aj + (bj + 1)2
bj

a2j + b2j + 1
=mαj
= .
|αj |2 + 1
62 CHAPTER II

Take now f ∈ H 1 (D+ ). Replacing, if necessary, f (z) by f (1 + δ)z , we can
assume that f (i) 6= 0. Notice that, in doing so, the zeroes change from αj to
(1 + δ)−1 αj , but this change does not affect the convergence of the series (5.3).
For ε > 0, let f ε (z) = f (z + iε). Then f ε ∈ H 1 (D+ ), kf ε kH 1 ≤ kf kH 1 , and
(f ε )] (x) = f (x + iε) = fε (x) ∈ L1 (R). Therefore,
T1−1 f ε ∈ H 1 (D). If ε is small
enough, say ε < ε0 , T1−1 f ε (0) = −4πf (1 + ε)i 6= 0. We then set
1
gε = −1 ε T1−1 f ε .
T1 f (0)
If Eε = {j : =mαj > ε}, the zeroes of f ε in D+ are the αεj = αj − iε with j ∈ Eε ,

and the zeroes of g ε in D are βjε = ϕ−1 (αεj ), j ∈ Eε .
Since g ε (0) = 1, we can apply Lemma 7.5 in Chapter I to obtain that
X Z

1 − |βj | ≤ sup log |g ε (reit )| dt
ε
r<1 T
j∈Eε
≤ log kg ε kH 1 (D)

= log kf ε kH 1 (D+ ) − log 4πf (1 + ε)i
Setting αj = aj + ibj and taking ε < ε0 , we have
=mαεj bj − ε bj − ε
1 − |βjε | ε = 2 2 .
|αj + i| 2 aj + (bj − ε + 1) 2 aj + b2j + 1
Therefore,
X bj − ε
≤ C log kf kH 1 (D+ ) − log 4πf (1 + ε)i .
ε
a2j + b2j + 1
j∈Eε
If ε decreases, the left-hand side increases for each j ∈ N (taking into account
that the sets Eε also increase). Passing to the limit by monotone convergence,
X bj
≤ C log kf k H 1 (D ) − log |4πf (i)| < ∞ .
a2j + b2j + 1 +
j∈N
Blaschke products on D+ are defined as compositions B◦ϕ−1 = T∞ B of Blaschke

products on D with the inverse Cayley transform. If w = ϕ−1 (z) and β = ϕ−1 (α),
then, according to (7.4) in Chapter I, we define
β̄ w − β |α2 + 1| z − α
σα (z) = ψ̃β (w) = = 2
|β| β̄w − 1 α + 1 z − ᾱ
if α 6= i, and
z−i
σi (z) = ϕ−1 (z) = .
z+i
Once we are at this stage, we can simply state the analogues of the results of
Section 7 in Chapter I. They can be transfered directly from D via the maps Tp for
p > 1, or reproved in exactly the same way for p = 1.
THE HALF-PLANE 63
Proposition 5.3. Let {αj } be a sequence of points in D+ satisfying (5.3). The

Blaschke product
Y∞
B(z) = σαj (z)
j=0
converges unconditionally and uniformly on compact subsets of D + to a function

B ∈ H ∞ (D+ ) vanishing in D+ only at the points αj . Moreover, |B ] (x)| = 1 for
almost every x ∈ R.
Theorem 5.4. Let f ∈ H p (D+ ), 1 ≤ p ≤ ∞, be not identically zero, and let B be
f
its Blaschke product. Then g = B ∈ H p (D+ ) and kgkH p = kf kH p .
Corollary 5.5. Let f ∈ H s (D+ ), not identically zero, with 1 ≤ s ≤ ∞, and let
p, q ∈ [1, ∞] be such that p1 + 1q = 1s . Then there exist g ∈ H p (D+ ) and h ∈ H q (D+ )
such that f = gh, and kgkpH p = khkqH q = kf ksH s .
Finally, we can state the F. and M. Riesz theorem for the line.
Theorem 5.6 (F. and M. Riesz). Let µ ∈ M (R) be a measure such that µ̂(ξ) = 0
for ξ ≤ 0. Then µ is absolutely continuous w.r. to Lebesgue measure.
The proof follows from the factorization theorem 5.5 like in the disc.
Corollary 5.7. If f ∈ H 1 (D+ ), then f ] ∈ L1 (R), and therefore limy→0 fy = f ]
in the L1 -norm. The map T1 of Proposition 5.1 is an isometry of H 1 (D) onto
H 1 (D+ ).
64 CHAPTER II
POINTWISE CONVERGENCE 65
CHAPTER III
POINTWISE CONVERGENCE TO THE BOUNDARY
AND CONJUGATE HARMONIC FUNCTIONS IN hp
1. The Hardy-Littlewood maximal function
We attack now a question that has been in the background so far, i.e. the
pointwise behaviour of hp - or H p -functions. In particular, in the cases where f ] is
a function, we wnt to know (e.g. in the disc, to fix the notation) if
lim f (reit ) = f ] (eit )

r→1
for almost every eit ∈ T. If f ∈ H ∞ (D) and f ] is continuous, the answer is positive,
since we know that the functions fr tend to f ] uniformly. Partial answers can be
given easily also in other cases. If f ∈ H p (D) with 1 < p < ∞, then fr → f ]
in norm, and therefore there is a sequence rn → 1 such that frn → f almost
everywhere. But this is much less than what we are looking for.
Sharp answers to our question follow from considerations about maximal func-
tions. This notion makes sense and is useful in many different situations, and it is
worth therefore to discuss their properties in a general context.
Let X be a topological space. A quasi-distance on X is a function d from X × X

to R such that
(1) d(x, y) ≥ 0 for all x, y ∈ X;
(2) d(x, y) = 0 if and only if x = y;
(3) d(x, y) = d(y, x) for all x, y ∈ X;
(4) there is a constant c ≥ 1 such that

(1.1) d(x, z) ≤ c d(x, y) + d(y, z)
for all x, y, z ∈ X;
(5) the balls B(x, r) = {y : d(x, y) < r} form a fundamental neighborhood
system at each x ∈ X.
Let m be a positive regular Borel measure on X. One says that m is doubling
(more precisely, d-doubling) if there is a constant c0 such that

(1.2) m B(x, 2r) ≤ c0 m B(x, r)
for all x ∈ X and r > 0.

Typeset by AMS-TEX
66 CHAPTER III
Definition. A triple (X, d, m), where d is a quasi-distance on X and m is a d-

doubling measure, is called a space of homogeneous type28 .
Examples.
The following are spaces of homogeneous type29 :
0 0
(1) The unit circle T, with the distance d(eit , eit ) = |eit − eit | (or equivalently
with the quotient distance of R/2πZ) and the normalized Lebesgue measure.
(2) Rn , with Euclidean distance and the Lebesgue measure.
(3) R, with the Euclidean distance and the measure dm(x) = |x|α dx, with
α > −n.
(4) Rn , with the Lebesgue measure and the distance
n o
d(x, y) = max |x1 − y1 |1/d1 , . . . , |xn − yn |1/dn ,
where d1 , . . . , dn > 0.
(5) Z, with the distance d(n, m) = |n − m| and the counting measure m(E) =
cardE.
(6) The unit sphere S n−1 ⊂ Rn , with the induced distance from Rn and the
Hausdorff measure.
(7) The unit sphere S 2n−1 ⊂ Cn , with the Hausdorff measure and the distance
d(ζ, ζ 0 ) = |1 − hζ, ζ 0 i|, where hζ, ζ 0 i is the Hermitean inner product in Cn .
Let X be a space of homogeneous type.

Definition. Let f be locally integrable w.r. to m. The function
Z
1
(1.3) M f (x) = sup |f (y)| dy ,
x∈B m(B) B
is called the Hardy-Littlewood maximal function of f and the operator M : f 7−→

M f is called the Hardy-Littlewood maximal operator.
Clearly M is not linear (observe that M f (x) ≥ 0 for every f ), but only sub-linear,
in the sense that
(1.4) M (f + g) ≤ M f + M g , M (λf ) = |λ|M f .
Lemma 1.1. The function M f is lower-semicontinuous, hence measurable.

Proof. Let M (x0 ) > α. Then there is a ball B containing x such that
Z
1
|f (y)| dy > α .
m(B) B
Then M f (x) > α for every x ∈ B.
28 Sometimesthe definition of space of homogeneous type is given by requiring that X be just

a set, and d a function satisfying conditions (1)-(4) only. A topology is then introduced on X by
stating that a set A is open if for every x ∈ A there is a ball B(x, r) ⊂ A. However, this does not
guarantee that the balls are open, not even that they are Borel sets. Therefore, one must impose
that the doubling measure m be defined on a σ-algebra containing both the open sets and the
balls.
29 Some of the following statements would require a proof, that we omit.
Remark. The classical definition of Hardy-Littlewood maximal function (for X =

Rn , with Euclidean distance and Lebesgue measure) is the following:
Z
0 1
(1.5) M f (x) = sup |f (y)| dy ,
r>0 m B(x, r) B(x,r)
i.e. limited to the averages of |f | over balls centered at x. Clearly M 0 f (x) ≤ M f (x);
however M 0 f is not necessarily lower-semicontinuous. The measurability of M 0 f
follows in this case from the fact that the map
Z
1
F (x, r) = |f (y)| dy
m B(x, r) B(x,r)
is continuous in r, so that the sup in (1.5) can be limited to r ∈ Q.

In a general space of homogeneous type, F (x, r) need not be continuous in r,
hence the definition (1.3) is preferable.
We want to discuss boundedness of M on the spaces Lp (X), i.e. inequalities of

the form
kM f kp ≤ Ckf kp .
Notice that it follows from (1.4) that |M f −M g| ≤ M (f −g), hence boundedness
on Lp is equivalent to continuity, as for linear operators.
Obviously M is bounded on L∞ (X). At the other extreme, p = 1, M is not
bounded in general. For instance, in the classical siuation (X = Rn , etc.), taking
f the characteristic function of the unit ball, one has
C
M f (x) ≥ ,
1 + |x|n
so that M f 6∈ L1 .
Nevertheless, the starting point of our proof is that, for p = 1, M satisfies a
weaker form of boundedness. We first recall some properties of the distribution
function, defined for α > 0,

(1.6) δf (α) = m {x : |f (x)| > α}
of an m-measurable function f on X.
Lemma 1.2. Let (X, m) be a measure space. If f ∈ Lp (X, m), 1 ≤ p < ∞, then
the following hold:
(1) the Chebishev inequality
kf k p
p
δf (α) ≤ ;
α
(2) the identity Z ∞

kf kpp =p δf (α)αp−1 dα .
0
68 CHAPTER III
Proof. Let Eα = {x : f (x) > α}. Then

Z kf k p
|f (x)|p p
m(Eα ) ≤ p
dm(x) ≤ .
Eα α α
This gives (1). Suppose

Pn now that f ∈ Lp (X, m) assumes a finite number of
values. Then |f | = j=1 cj χAj , with 0 < c1 < · · · < cn and the Aj pairwise
disjoint. Then X
m(Eα ) = m(Aj ) ,
j:cj >α
and, setting c0 = 0,
Z ∞ n Z
X cj n
X
p−1 p−1
p m(Eα )α dα = p α m(Ak ) dα
0 j=1 cj−1 k=j
n
X n
X
= (cpj − cpj−1 ) m(Ak )
j=1 k=j
n
X
= cpj m(Ak )
j=1
= kf kpp .
For general f , one approximates |f | from below by finite-valued functions.

Definition. Let T be a linear, or sub-linear, operator defined on Lp (X), p < ∞,
and taking values in the space of measurable functions on X. One says that T is
weak-type (p, p) if, for every α > 0
kf k p
p
(1.7) δT f (α) ≤ C .
α
The expression “weak-type” comes from the fact that (1.7) is a weaker condition
than boundedness on Lp , as a consequence of Chebishev’s inequality. In fact, if T
is bounded on Lp (X),
kT f k p kf k p
p p
δT f (α) ≤ ≤ kT kpLp →Lp .
α α
One also says that an operator T is strong type (p, p) if it is bounded on Lp (X).
Going back to the maximal operator M , we shall see that it is weak-type (1, 1).
The proof is based on the following Vitali covering lemma.
Lemma 1.3. Inside a space X of homogeneous type, let {Bj }j∈J be a finite family
of balls covering a measurable set E. There exists a sub-family {B j }j∈J 0 such that
Bj ∩ Bk = ∅ for j, k ∈ J 0 , j 6= k, and
[
m Bj ≥ κm(E) ,
j∈J 0
where κ dipends only on the constants c, c0 in (1.1) and (1.2).

Proof. Start with a ball Bj1 of maximum radius. Inductively, take Bjk+1 with
maximum radius among the balls disjoint from Bj1 ∪· · ·∪Bjk . This procedure stops
after a finite number of steps, precisely when there are no more balls left which are
disjoint from Bj1 ∪ · · · ∪ Bjk . Then set J 0 = {j1 , . . . , jk } and call Bj1 , . . . Bjk the
“selected” balls, and the remaining ones the “excluded” balls.
If B is a ball of radius r, denote by B ∗ the ball with the same center and radius
3c r, where c is the constant in (1.1). Observe that if two balls B, B 0 have non-
2
empty intersection and the radius of B 0 is not larger than the radius of B, then
B 0 ⊆ B ∗ by (1.1).
Let B 0 one of the excluded balls. It necessarily intersects one of the selected
ones. Let `¯ be the smallest integer ` such that B 0 ∩ Bj` 6= ∅. Then the radius Bj`¯
is greater than or equal to the radius of B 0 , so that
B 0 ⊆ Bj∗`¯ .
Therefore [ [
E⊆ Bj ⊆ Bj∗ ,
j∈J j∈J 0
Take ν so that 2ν ≥ 2c. Then

ν
m(Bj∗ ) ≤ m B(x, 2k r) ≤ c0 m(B) ,
1
so that, with κ = c0 ν ,
X X [
m(E) ≤ m(Bj∗ ) = κ−1 m(Bj ) = κ−1 m Bj .
j∈J 0 j∈J 0 j∈J 0
Theorem 1.4. The operator M is weak-type (1, 1).

Proof. Given f ∈ L1 (X) and α > 0, let Eα = {x : M f (x) > α}. By Lemma 1.1,
Eα is open and its measure is the supremum of the measures of its compact subsets.
Let E be a compact subset of Eα . Given x ∈ E, M f (x) > α, so that there is a
ball Bx containing x such that
Z
1
|f (y)| dy > α ,
m(Bx ) Bx
i.e.
Z
1
(1.8) m(Bx ) ≤ |f (y)| dy .
α Bx
Since E is compact, we can extract a finite subcovering {Bxj }j∈J of E from

{Bx }x∈E . By Lemma 1.2, we can further extract a finite family {Bxj }j∈J 0 of
mutually disjoint balls such that
X [
m(Bxj ) ≥ κm Bxj ≥ κm(E) .
j∈J 0 j∈J
70 CHAPTER III
Combining this with (1.8), we have

X Z
−1 κ−1 X kf k1
m(E) ≤ κ m(Bxj ) ≤ |f (y)| dy ≤ κ−1 .
j
α j B x j
α
Taking the supremum over E ⊂ Eα , we obtain that m(Eα ) ≤ κ−1 kf k1 /α.

Combining together the weak-type (1,1) property of M and its boundedness on
L , we can prove that it is bounded on Lp for 1 < p < ∞. What follows is a
∞
special case of the Marcinkiewicz interpolation theorem 30 .

Theorem 1.5. Let T be a linear, or sub-linear, operator which is weak-type (1, 1)
and bounded on L∞ (X, m). Then T is bounded on Lp (X, m) for 1 0, define

f (x) if |f (x)| ≤ α α f (x) if |f (x)| > α
fα (x) = f (x) =
0 if |f (x)| > α , 0 if |f (x)| ≤ α .
Then f α ∈ L∞ (X, m) with kfα k∞ ≤ α, and f α ∈ L1 (X, m). In fact,
Z
α
kf k1 = |f (x)| dm(x)
{|f (x)|>α}
Z
|f (x)|p−1
≤ |f (x)| dm(x)
{|f (x)|>α} αp−1
1
≤ kf kpp .
αp−1
If C∞ = kT kL∞ →L∞ , then kT fα k∞ ≤ C∞ α. Since f = fα + f α ,
|T f (x)| ≤ |T fα (x)| + |T f α (x)| ,
so that
|T f (x)| > 2C∞ α =⇒ |T f α (x)| > C∞ α ,
in other words,
{x : |T f (x)| > 2C∞ α} ⊆ {x : |T f α (x)| > C∞ α} .
We use now (2) in Lemma 1.2 and the weak-type (1,1) of T to obtain
Z ∞
p

kT f kp = p m {x : |T f (x)| > α} αp−1 dα
0
Z ∞
p

= p(2C∞ ) m {x : |T f (x)| > 2C∞ α} αp−1 dα
Z0 ∞

≤ p(2C∞ )p m {x : |T f α (x)| > C∞ α} αp−1 dα
0
Z ∞
p p−1
≤ p2 C∞ C1 kf α k1 αp−2 dα
Z0 ∞ Z
p p−1 p−2
= p2 C∞ C1 α |f (x)| dx dα
0 {x:|f (x)|>α}
Z Z |f (x)|
= p2p C∞
p−1
C1 |f (x)| αp−2 dα dx
X 0
= C 0 kf kpp ,
30 For
its general formulation, see E.M. Stein, G. Weiss, An introduction to Fourier analysis
on Euclidean spaces
having denoted by C1 the weak-type (1,1) constant for T .
Corollary 1.6. If 1 < p ≤ ∞, M is bounded on Lp (X).
2. Poisson integrals and maximal function
The reason for introducing the Hardy-Littlewood maximal function is that it

controls other quantities that intervene in the analysis of the boundary behaviour
of hp - (or H p -) functions. We begin with the upper half-plane D+ , where the
geometric picture is more clear.
For every point x0 ∈ R = ∂D+ , we define different “approach regions” from the
interior of D+ , and for each of them we introduce a maximal operator. The most
natural approach region is the vertical line x = x0 . Correspondingly, for a function
f in some Lp (R), we call vertical maximal function of f the function

(2.1) Mvert f (x) = sup Pf (x + iy) ,
y>0
where P denotes the operator assigning to f its Poisson integral.

Other approach regions are the non-tangential angles. Given a point x0 ∈ R,
consider the open infinite angle Γα (x0 ) inside D+ with vertex in x0 , symmetric w.r.
to the vertical line x = x0 , and with semi-aperture α, 0 < α < π2 . Explicitely,
Γα (x0 ) = {x + iy : |x − x0 | < y tan α} .
For a fixed α ∈ (0, π/2), we define the non-tangential maximal function of f ∈

p
L (R) as
Mnt,α f (x) = sup |Pf (z)| .
z∈Γα (x)
Obviously,
Mvert f (x) ≤ Mnt,α f (x) ≤ Mnt,α0 f (x) ,
if α < α0 .
Lemma 2.1. For every α ∈ (0, π/2) there is a constant Cα > 0 such that, if
f ∈ Lp (R),
Mnt,α f (x) ≤ Cα M f (x) .
In particular, each Mnt,α and Mvert are weak-type (1, 1) and bounded on Lp for
1 π/4, so that a = tan α > 1. Consider the dyadic
a
intervals Ija = [−a2j , a2j ] in the real line, with j ≥ 0. We set Eja = Ija \ Ij−1 for
72 CHAPTER III
j > 0, and E0a = I0a . Then

1 1
P1 (x) =
π 1 + x2
∞
1 1X 1
≤ χE0 (x) +
a
2
χ a (x)
2(j−1) Ej
π π 1 + a 2
j=1
∞
1 4 X −2j
≤ χE0a (x) + 2 2 χEja (x)
π πa j=1
(2.2) 4
∞
4 X −2j
≤ χI0a (x) + 2 χIja (x) − χIj−1
a (x)
π π j=1
∞
4 X −2j
= (2 − 2−2(j+1) )χIja (x)
π j=0
∞
X
=C 2−2j χIja (x) .
j=0
For a general y > 0, (2.2) and the identity

1 x
Py (x) = P1 ,
y y
give
∞
X 2−2j x ∞
X 2−2j
(2.3) Py (x) ≤ C χIja =C χIjay (x) .
y y y
j=0 j=0
It follows that
Z

f ∗ Py (x) ≤ |f (x − t)|Py (t) dt
R
∞
X Z
2−2j
≤C |f (x − t)| dt
(2.4) y Ijay
j=0
∞
X Z x+ay2j
2−2j
= Ca |f (t)| dt .
j=0
y x−ay2j
If x + iy ∈ Γα (x0 ), then x0 ∈ [x − ay2j , x + ay2j ], so that

Z
1
|f (t)| dt ≤ M f (x0 ) .
2ay2j [x−ay2j ,x+ay2j ]
Therefore
∞
X

f ∗ Py (x) ≤ Ca2 2−j M f (x0 ) ,
j=0
and
Mnt,α f (x0 ) ≤ Ca2 M f (x0 ) .
Theorem 2.2. Suppose that u is in hp (D+ ), 1 0
Proof. For p > 1, u∗α ∈ Lp (R) because u] ∈ Lp (R) and by Lemma 2.1.
It remains to discuss the case u ∈ H 1 (D+ ). By Corollary 5.5 in Chapter II, we
1
can factorize u as u = vw, with v, w ∈ H 2 (D+ ) and kvkH 2 = kwkH 2 = kukH
2
1.
Then
u∗α (x) = sup v(z)w(z) ≤ vα∗ (x)wα∗ (x) .
z∈Γα (x)
Therefore,
ku∗α k1 ≤ kvα∗ k2 kwα∗ k2 ≤ C2,α

2 2
kvkH 2 kwkH 2 = C2,α kukH 1 .
We pass now to the unit disc. The substitute for the vertical maximal function
is the radial maximal function,

Mrad f (eit ) = sup Pf (reit ) .
r<1
defined for f ∈ L1 (T).

As non-tangential access regions, we take the Stolz regions Sρ (eit ), with ρ ∈
(0, 1), defines as the open convex envelop of the point eit and the disc with center
0 and radius ρ. Near eit , the Stolz region Sρ (eit ) is the angle pointing towards the
interior of the disc, symmetric w.r. to the radius, and of semi-aperture α = arcsin ρ.
We then define the non-tangential maximal function of f ∈ L1 (T) as
Mnt,ρ f (eit ) = sup |Pf (z)| .

z∈Sρ (eit )
Lemma 2.3. For every ρ ∈ (0, 1) there is a constant Cρ > 0 such that, if f ∈
L1 (T),
Mnt,ρ f (eit ) ≤ Cρ M f (eit ) .
In particular, each Mnt,ρ and Mrad are weak-type (1, 1) and bounded on Lp for
1 < p < ∞.
Proof. We use (3.6) in Chapter I,
1−r
Pr (eit ) ≤ C
t2 + (1 − r)2
with t ∈ [−π, π], and proceed as in the proof of Lemma 1.1, just replacing y by
a(1−r)
1 − r, and stopping the sums in j as soon as the intervals Ij do not intersect
[−π, π].
74 CHAPTER III
Theorem 2.4. Suppose that u is in hp (D), 1 < p ≤ ∞, or in H p (D), 1 ≤ p ≤ ∞.

Then the functions
def
u∗ρ (eit ) = sup u(z) ,
z∈Sρ (eit )
are in Lp (T) and ku∗α kp ≤ Cp,ρ kukhp . The same is true for
def
u∗ (x) = sup u(reit ) .
r<1
3. Pointwise convergence to the boundary
In this section we discuss pointwise convergence to the boundary of harmonic

and holomorphic functions on D or on D+ . In this case we will begin with the unit
disc, where the situation is simplified by the inclusion of all Lp -spaces into L1 .
Our initial problem is the radial convergence of a function u to the boundary,
i.e. the existence of
lim u(reit ) = u] (eit ) ,
r→1
almost everywhere. The results of the previous section, however, induce us to

consider the stronger problem of non-tangential convergence, i.e. the existence of
lim u(z) = u] (eit ) ,

z→eit , z∈Sρ (eit )
for every ρ < 1.

Theorem 3.1. If u = Pf , with f ∈ L1 (T), then u converges to f non-tangent-
ially a.e.
Proof. We can assume that f is real-valued. Consider then the quantity
δu(eit ) = lim sup u(z) − lim inf u(z) ,

z→eit , z∈Sρ (eit ) z→eit , z∈Sρ (eit )
which is non-negative. Obviously,
δu(eit ) ≤ 2Mnt,ρ f (eit ) .
Therefore, using Lemma 2.3 and denoting by |E| the normalized Lebesgue mea-
sure of E ⊂ T, we have, for α > 0,
it
(3.1) {e : δu(eit ) > α} ≤ Cρ kf k1 .
α
Given ε > 0, there is g ∈ C(T) such that kf − gk1 < ε. By Theorem 4.3 in
Chapter I, v = Pg is continuous on D̄, so that
lim sup (u − v)(z) = lim sup u(z) − lim v(z)

z→eit , z∈Sρ (eit ) z→eit , z∈Sρ (eit ) z→eit , z∈Sρ (eit )
= lim sup u(z) − g(eit ) .

In the same way,
lim inf (u − v)(z) = lim inf u(z) − g(eit ) ,

z→eit , z∈Sρ (eit ) z→eit , z∈Sρ (eit )
and therefore
δ(u − v)(eit ) = δu(eit ) .
Applying (3.1) to u − v, we obtain that
it
(3.2) {e : δu(eit ) > α} ≤ Cρ ε ,
α
and this holds for every ε > 0. Hence
it
{e : δu(eit ) > α} = 0 ,
for every α > 0.

Then the set
[ 1
eit : δu(eit ) > = {eit : δu(eit ) > 0}
n
n≥1
has measure zero, i.e.

lim sup u(z) = lim inf u(z) = lim u(z)
z→eit , z∈Sρ (eit ) z→eit , z∈Sρ (eit ) z→eit , z∈Sρ (eit )
almost everywhere. On the other hand, limr→1 kur − f k1 = 0, so that there is a

subsequence rj → 1 such that u(rj eit ) → f (eit ) a.e. We conclude that
lim u(z) = f (eit )

almost everywhere.
Corollary 3.2. If ρ < 1 and u ∈ hp (D), 1 < p ≤ ∞, or in H p (D), 1 ≤ p ≤ ∞,
then for almost every eit ∈ T, limz→eit , z∈Sρ (eit ) u(z) = u] (eit ) for every ρ < 1.
In the proof of Theorem 3.1 we have used in a crucial way the density of con-
tinuous functions in L1 , in order to have (3.2). The same proof can be adapted to
prove the analogous result in the upper half-plane. In this case, we approximate
f ∈ L1 (R) by functions g ∈ C0 (R). This also works for f ∈ Lp (R) as long as p < ∞,
but it breaks down for p = ∞. For this case we use a different argument.
Theorem 3.3. If u = Pf , with f ∈ Lp (R) and 1 ≤ p ≤ ∞, then u converges to f
non-tangentially a.e.
Proof. If f ∈ L1 (R), we proceed as described above. Take now f ∈ L∞ (R). Then
u ∈ h∞ (D+ ). If ϕ is the Cayley transform in (2.7) of Chapter II, then v = u ◦ ϕ ∈
h∞ (D). We know by Corollary 3.2 that v converges to v ] (eit ) non-tangentially a.e.
Fix now x ∈ R and α > 0, let Γ0α (x) = Γα (x) ∩ {z : =mz < 1} and eit =
ϕ (x). Since ϕ0 (eit ) 6= 0, ϕ is a diffeomorphism
−1
of a neighborhood of eit onto a
neighborhood of x. Therefore ϕ−1 Γ0α (x) is contained in a Stolz angle Sρ (eit ).
76 CHAPTER III
This implies that for a.e. x ∈ R,

(3.3) lim u(z) = lim u(z) = lim v(z) = v ] ϕ−1 (x) .
z→x , z∈Γα (x) z→x , z∈Γ0α (x) z→eit , z∈Sρ (eit )
In particular,
lim u(x + iy) = v ] ϕ−1 (x)
y→0
almost everywhere. Since u is bounded, we can apply dominated convergence to

prove that, if g ∈ L1 (R),
Z Z

lim uy (x)g(x) dx = v ] ϕ−1 (x) g(x) dx ,
y→0 R R
i.e. uy → v ] ◦ ϕ−1 in the weak* topology. Hence v ] ◦ ϕ−1 = u] = f , and (3.3) then
says that u converges to f non-tangentially a.e.
At this point, the simplest argument to prove the statement for 1 1 ,
and f1 = f − f∞ .
Corollary 3.4. If α < π/2 and u ∈ hp (D+ ), 1 < p ≤ ∞, or in H p (D+ ), 1 ≤ p ≤
∞, then limz→x , z∈Γα (x) u(z) = u] (x) almost everywhere.
The case p = ∞ in Corollary 3.2 and 3.4 is referred to as “Fatou’s theorem”.
4. Poisson integrals of singular measures
The discussion in the previous section does not say anything about non-tan-
gential limits of general h1 -functions. We complete the picture here, proving that
Corollary 3.4 can be extended to h1 (R) (the same can be done on T, but we omit
the proof). It must be noted however that no maximal function is involved in the
proof.
Every h1 -function on D+ is the Poisson integral of a measure µ ∈ M (R). We
recall the Lebesgue decomposition of µ as
µ = µa + µs ,
where µa is absolutely continuous with respect to Lebesgue measure m (or µa
m), i.e. dµa (x) = h(x) dx with h ∈ L1 (R), and µs is singular with respect to
the Lebesgue measure (or µs ⊥ m). This means that there is a set E such that
m(R \ E) = 0 and |µs |(E) = 0.
The function
ϕ(x) = |µs |(−∞, x)
is non-decreasing, hence differentiable at (Lebesgue-) almost every point, and its
being singular implies that ϕ0 (x) = 0 a.e. This means that at a.e. x ∈ R,
|µs |(x − h, x + h)
(4.1) lim =0.
h→0 h
Lemma 4.1. Let µ be a singular measure on R, and u = Pµ. Then for almost
every x ∈ R,
lim u(z) = 0
z→x , z∈Γα (x)
for every α > 0.

Proof. Take a point where (4.1) holds. We can assume that this point is the origin.
Given ε > 0, take δ such that |µs |(−h, h) < εh for h < δ.
With a = tan α, which we can assume to be greater than 1, take x with |x| < ay.
Then
Z x+ 4δ Z
|u(x + iy)| ≤ Py (x − t) d|µ|(t) + Py (x − t) d|µ|(t)
x− δ4 |t−x|> δ4
Z x+ 4δ δ
≤ Py (x − t) d|µ|(t) + Py kµk1 .
x− δ4 4
2y
Since Py (δ/4) = δ 2 +4y 2
, the last term tends to 0 as y → 0. We must then show
that Z x+ 4δ
lim Py (x − t) d|µ|(t) = 0 .
y→0 , |x|<ay x− δ4
δ
Take y < 4a , so that (x − δ4 , x + 4δ ) ⊂ (− 2δ , δ2 ). Define Ij = (x − 2j y, x + 2j y)
for j ≥ 0 and 2j y < 2δ . Then Ij ⊂ (−δ, δ) for every such j, and (x − 4δ , x + δ4 ) is
covered by the Ij .
Since supt∈Ij \Ij−1 Py (x − t) = Py (2j−1 y) < 22j4 y and (a + 2j )y < δ,
Z x+ δ4
|µ|(I0 ) X 4
Py (x − t) d|µ|(t) ≤ + |µ|(Ij \ Ij−1 )
x− δ4 y 22j y
j≥1

X |µ| − (a + 2j )y, (a + 2j )y
≤C
j
2−2j y
X (a + 2j )y
≤ Cε
j
2−2j y
≤ Ca ε .
Therefore Z x+ 2δ
lim sup Py (x − t) d|µ|(t) ≤ Ca ε .
y→0 , |x|<ay x− δ2
The conclusion follows by the arbitrarity of ε.

Theorem 4.2. Let u = Pµ ∈ h1 (D+ ), and let µ = hm + µs be the Lebesgue
decomposition of µ, with h ∈ L1 (R). Then for a.e. x ∈ R,
lim u(z) = h(x)
z→x , z∈Γα (x)
for every α > 0.

The proof follows easily from Theorem 3.3 and Lemma 4.1.
78 CHAPTER III
5. Lp -estimates for the conjugate harmonic function
We discuss now a crucial point in the theory of Hardy spaces, that has been left
aside in the previous chapters: the fact that for 1 < p < ∞ the conjugate harmonic
function of an hp -function is also in hp .
We begin with the unit disc, and recall that if
∞
X ∞
X
n
u(z) = an z + a−n z̄ n
n=0 n=1
is holomorphic in the unit disc, its harmonic conjugate is

∞
X ∞
X
n
(5.1) ũ(z) = −i an z + i a−n z̄ n .
n=1 n=1
When u is real-valued, ũ is characterized by the properties that it is real-valued,

u + iũ is holomorphic, and ũ(0) = 0.
Lemma 5.1. If u is harmonic and positive on some open set, and p > 0, then
(5.2) ∆(up ) = p(p − 1)up−2 |∇u|2 .
If f is holomorphic and non-zero in some open set, and p > 0, then

(5.3) ∆ |f |p = p2 |f |p−2 |f 0 |2 .
Proof. We have

∂x2 (up ) = ∂x pup−1 ∂x u = p(p − 1)up−2 (∂x u)2 + pup−1 ∂x2 u .
summing with the corresponding formula for ∂y2 (up ), we have (5.2).
As to (5.3), setting the standard notation ∂z = 12 (∂x − i∂y ), ∂z̄ = 21 (∂x + i∂y ),
we have ∆ = 4∂z ∂z̄ . We fix a point z0 in the domain of f , and a determination
p
of log f in a neighborhood of z0 . We then set f p/2 (z) = e 2 log f (z) . Then, in this
neighbourhood of z0 ,
∆|f |p = 4∂z ∂z̄ f p/2 f p/2

= 4(∂z f p/2 )(∂z̄ f p/2 )
= 4|∂z f p/2 |2
= p2 |f |p−2 |f 0 |2 .
Teorema 5.2. If 1 0 in D. Let f = u + iũ. Then
f is holomorphic and non-zero in D.
We apply Green’s formula to the two functions up and |f |p on the disc Dr

centered at the origin and with radius r < 1. We have
Z 2π Z
it p

r ∂r u(re ) dt = p(p − 1) u(z)p−2 |∇u(z)|2 dz ,
0 Dr
(5.4) Z 2π Z
r ∂r |f (reit |p dt = p2 |f (z)|p−2 |f 0 (z)|2 dz .
0 Dr
But
(5.5) |∇u|2 = (∂x u)2 + (∂y u)2 = (∂x u)2 + (∂x ũ)2 = |∂x f |2 = |f 0 |2 .
Since p < 2 and |u| ≤ |f |, we have |f |p−2 ≤ |u|p−2 . Therefore

Z 2π Z
it p 2
r ∂r |f (re | dt ≤ p |u(z)|p−2 |∇u(z)|2 dz
0 Dr
(5.6) Z 2π
p
= r ∂r |u(reit |p dt .
p−1 0
In other words,
∂r Mp (f, r)p ≤ p0 ∂r Mp (u, r)p .
Since Mp (u, 0) = Mp (f, 0) = u(0), we conclude that Mp (f, r)p ≤ p0 Mp (u, r)p for
every r < 1, hence kf kH p ≤ (p0 )1/p kukhp .
Since |ũ| ≤ |f |, we also have
kũkhp ≤ (p0 )1/p kukhp .
If u is a generic function in hp (D), we decompose u] ∈ Lp (T) as the combination

]
u = ϕ1 − ϕ2 + iϕ3 − iϕ4 of four non-negative functions, with the supports of ϕ1
and ϕ2 disjoint, and similarly for ϕ3 and ϕ4 . Then
u = u1 − u2 + iu3 − iu4 ,
where uj = Pϕj is either identically zero (if so is ϕj ), or strictly positive in D, as a

consequence of the fact that the Poisson kernels Pr are strictly positive everywhere.
P4
By construction, j=1 kϕj kp ≤ Cku] kp , so that
4
X
kuj khp ≤ Ckukhp .
j=1
Therefore
4
X
kũkhp ≤ kũj khp ≤ Cp kukhp ,
j=1
which gives the conclusion for 1 < p < 2.

0
Consider now the duality between hp (D) and hp (D) of Section 8 in Chapter I,
Z Z X
it it
B(u, v) = lim u(re )v(re ) dt = u] (eit )v ] (eit ) dt = lim an bn r 2|n| ,
r→1 T T r→1
n∈Z
80 CHAPTER III
if an , bn are the Taylor coefficients of u and v respectively. If 1 < p < ∞, we have
|B(u, v)| ≤ kukhp kvkhp0 , kukhp = sup |B(u, v)| .

kvkhp0 ≤1
By (5.1), the operator mapping u into ũ is skew-hermitian w.r. to B, that is

X
B(ũ, v) = lim (−i sgn n)an bn r 2|n| = −B(u, ṽ) .
r→1
n∈Z
Therefore, if u ∈ hp (D) and 2 < p < ∞,
kũkhp = sup |B(ũ, v)|

kvkhp0 ≤1
= sup |B(u, ṽ)|

kvkhp0 ≤1
≤ kukhp sup kṽkhp0

kvkhp0 ≤1
≤ Cp0 kukhp .
This concludes the proof.

We switch now to the upper half-plane. We take a function u ∈ hp (D+ ) and we
suppose that p < ∞. By Proposition 4.2 in Chapter II, the harmonic conjugate
Z
] 1 t
ũ(x + iy) = u ∗ P̃y (x) = u] (x − t) dt
π R t2 + y 2
is well defined, and, if u is real-valued, it is characterized as the only real-valued

harmonic function such that u + iũ is holomorphic and tends to 0 at infinity in each
half-plane {z : =mz ≥ a} with a > 0.
We know by (4.9) in Chapter II that, for u ∈ h2 (D+ ), ũ ∈ h2 (D+ ) and kũkh2 =
kukh2 . We prove now the analogue of Theorem 5.2.
Theorem 5.3. If 1 < p < ∞, there is a constant Cp such that kũkhp ≤ Cp kukhp
for every u ∈ hp (D+ ).
Following the proof of Theorem 5.2, we want to use Green’s formula to prove
that, if f = u + iũ, then |∂y Mp (f, y)p | ≤ C|∂y Mp (u, y)p |. We shall then apply
Green’s formula over rectangles, and use the fact that the integrals over certain
edges tend to 0 if the edge is moved to infinity. This requires some preliminary
proof.
Lemma 5.4. Suppose u ∈ hp (D+ ) with p < ∞. Then
lim |∇u(x + iy)| = 0 ,

x→∞
uniformly in y ≥ a, for every a > 0. Moreover,
1
Mp (∇u, y) ≤ Mp P|u] |, y ;
y
in particular, limy→∞ Mp (∇u, y) = 0.

Proof. We have
Z Z
]
(5.7) ∇u(x + iy) = ∇ u (t)Py (x − t) dt = u] (t)∇Py (x − t) dt ,
R R
Since Py (x) = − π1 =m x+iy

1
, using the Cauchy-Riemann equations for (x + iy)−1 ,
we have
1 1
(5.8) |∇Py (x)| = 2
= Py (x) .
π|x + iy| y
By (5.7),
Z
1 1
|∇u(x + iy)| ≤ |u] (t)|Py (x − t) dt = P|u] | (x + iy) .
y R y
The first part of the statement then follows from Lemma 1.3 in Chapter II, since
P|u] | ∈ hp (D+ ), and the last part is now obvious.
Proof of Theorem 5.3. As in the proof of Theorem 5.2, we consider first the case
1 < p < 2. We further suppose that u = Pu] , with u] ≥ 0, continuous with compact
support, and non-identically zero. Then u is strictly positive on D+ .
In this hypotheses, we can say right now that ũ is “not far from” being in hp (D+ ).
In fact
− p10
(5.9) Mp (ũ, y) = ku] ∗ P̃y kp ≤ ku] k1 kP̃y kp ≤ Cp y ku] k1 ;
(observe that the same estimate also holds for u:
− p10
(5.9’) Mp (u, y) ≤ ku] k1 kPy kp ≤ Cp y ku] k1 ,
a fact that we shall use later).

Therefore
(5.10) v(ε) (z) = ũ(z + iε) ∈ hp (D+ )
for every ε > 0.

For a, b, R > 0 with a < b, let Qa,b,R be the rectangle [−R, R] × [a, b]. Applying
Green’s formula, the analogues of (5.4) become
Z R Z b
p

− ∂y u(x + ia) dx + ∂x u(R + iy)p dy
−R a
Z R Z b
p

(5.11) + ∂y u(x + ib) dx − ∂x u(−R + iy)p dy
−R a
Z
= p(p − 1) u(z)p−2 |∇u(z)|2 dz ,
Qa,b,R
82 CHAPTER III
and
Z R Z b
p
− ∂y |f (x + ia)| dx + ∂x |f (R + iy)p | dy
−R a
Z R Z b
p
(5.12) + ∂y |f (x + ib)| dx − ∂x |f (−R + iy)p | dy
−R a
Z
= p2 |f (z)|p−2 |f 0 (z)|2 dz .
Qa,b,R
We first fix a and b and let R tend to infinity. By Lemma 5.4 and Lemma 1.3 in
Chapter II,
∂x u(±R + iy)p = pu(±R + iy)p−1 ∂x u(±R + iy)
tends to zero as R → +∞, uniformly in y ∈ [a, b].
To see that the same holds for ∂x |f (−R + iy)p |, observe that
p p p p p
∂x |f p | = <e∂z f 2 f 2 = <ef 2 −1 f 2 f 0 ,
2
so that

(5.13) ∂x |f p | ≤ p |f |p−1 |f 0 | .
2
By (5.5), f 0 (x + iy) tends to zero as x → ±∞ uniformly in y ∈ [a, b]. The same
is true for |f (z)|, as a consequence of (5.10).
We can then let R → +∞ in (5.11) and (5.12), to obtain
Z Z
p

− ∂y u(x + ia) dx + ∂y u(x + ib)p dx
R
(5.14) Z R
= p(p − 1) u(z)p−2 |∇u(z)|2 dz ,
x∈R , y∈[a,b]
and
Z Z
p
− ∂y |f (x + ia)| dx + ∂y |f (x + ib)|p dx
R R
(5.15) Z
2
=p |f (z)|p−2 |f 0 (z)|2 dz .
x∈R , y∈[a,b]
We now make tend b to +∞, and show that the integrals containg b tend to zero.
In (5.14) we use Hölder’s inequality to obtain
Z Z

∂y u(x + ib)p dx ≤ p u(x + ib)p−1 |∂y u(x + ib)| dx

R R
≤ pMp (u, b)p−1 Mp (∂y u, b) ,
which tends to 0 by Lemma 5.5. For the second integral in (5.15), we use the
analogue of (5.13) for ∂y (|f |p ) to obtain that
Z

∂y |f (x + ib)|p dx ≤ p Mp (f, b)p−1 Mp (f 0 , b)
2
R
p p−1
≤ Mp (u, b) + Mp (ũ, b) Mp (∇u, b) .
2
This last quantity tends to zero by (5.10) and Lemma 5.4. So (5.14) and (5.15)
respectively imply that
Z Z
p

− ∂y u(x + ia) dx = p(p − 1) u(z)p−2 |∇u(z)|2 dz
R x∈R , y≥a
(5.16) Z Z
− ∂y |f (x + ia)|p dx = p2 |f (z)|p−2 |f 0 (z)|2 dz .
R x∈R , y≥a

We claim that the two left-hand sides in (5.16) are −∂y Mp (u, y)p | and
p
y=b
−∂y Mp (f, y) | respectively. We postpone this proof, and assume this fact
y=b
to be true for the moment.
Since Mp (u, y) and Mp (f, y) are decreasing in y, the left-hand sides in (5.16) are
in fact positive quantities. Since
lim Mp (u, y)p = lim Mp (f, y)p = 0

y→0 y→0
by (5.9) and (5.9’), the same argument used in the proof of Theorem 5.2 shows that
Mp (f, y)p ≤ p0 Mp (u, y)p
for every y > 0. The rest of the proof will then proceed as in Theorem 5.2, taking
initially u as the Poisson integral of a continuous, complex-valued function u] on R
with compact support. Once we have the estimate
kũkhp ≤ Cp kukhp
for such functions, we use a density argument to extend it to a generic u ∈ hp (D+ ).

We are so left with the proof of our previous claim, concerning
Z p
p
u x + i(b + h) − u(x + ib)p
∂y Mp (u, y) | = lim dx .
y=b h→0 R h
In order to be allowed to move the limit inside the integral, we use dominated
convergence. We shall use vertical maximal functions for this purpose.
Fix ε > 0 and smaller than b/2, and call u(ε) (z), resp. f(ε) (z), the functions
u(z +iε), resp. f (z +iε). Notice that u(ε) , ∂y u(ε) ∈ hp (D+ ) and f(ε) , f(ε)
0
∈ H p (D+ ),
by (5.9), (5.9’) and Lemma 5.4.
We take h > −(b − ε), as we can, in order to stay above the level ε.
By the mean value theorem, for every x there is t = t(x) ∈ (b, b + h) such that
p
u x + i(b + h) − u(x + ib)p
= ∂y u(x + iy)p |y=t = pu(x + it)p−1 ∂y u(x + it) .
h
Therefore,
u x + i(b + h)p − u(x + ib)p p−1

≤ p Mvert u(ε) (x) Mvert (∂y u(ε) )(x) .
h
Applying Hölder’s inequality and Theorem 2.2, we can see that this last function
is integrable. The same applies to f .
84 CHAPTER III
We can now complete some statements left incomplete in the previous chapters.
We give a list of them.
(i) The Cauchy projection is bounded from hp onto H p for 1 < p < ∞, both
in D and in D+ .
(ii) The conjugate function operator on T (see Propsoition 6.2 in Chapter I) and
the Hilbert transform on R (see Proposition 4.3 in Chapter II) are bounded
on Lp for 1 < p < ∞.
(iii) Under the sesquilinear map B in Proposition 8.3 in Chapter I, the dual of
0
H p (D) is identified with H p (D) for 1 < p < ∞. The same can be verified
on D+ .

Hardy Spaces in One Complex Variable

Uploaded by

Copyright:

Available Formats

You might also like

Hardy Spaces in One Complex Variable

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hardy Spaces in One Complex Variable

Uploaded by

Copyright:

Available Formats

HARDY SPACES IN ONE COMPLEX VARIABLE

prof. Fulvio Ricci

Chapter I Hardy spaces on the unit disc

Chapter II Hardy spaces on the half-plane

1. Definition and basic properties

We shall usually prefer to denote by T, rather than ∂D or similar, the boundary

(1.1) T = {z ∈ C : |z| = 1} = {eit : t ∈ R/2πZ} .

Let f (z) be a holomorphic function on D. Given r ∈ [0, 1) and p ≥ 1, define

(1.3) M∞ (f, r) = max

Definition2 . Let 1 ≤ p ≤ ∞. We denote by H p (D) the space of holomorphic

(1.4) sup Mp (f, r) = kf kH p < ∞ .

It is easy to check that (1.4) defines a norm. For the implication kf kH p =

below by positive constants on a neighborhood of 1.

Lemma 1.2. For s > 0 fixed, consider the integral

as a function of r ∈ [0, 1). Then, for r → 1,

|1 − reit | ≤ |1 − eit | + |eit − reit | = |1 − eit | + 1 − r ≤ |t| + 1 − r ,

The conclusion follows easily.

Lemma 1.4. Let f ∈ H p (D). Then, for every z ∈ D,

Proof. The statement is obvious for p = ∞, so we assume that p is finite.

where p0 is the dual exponent of p.

By Lemma 1.2, since p0 > 1,

Theorem 1.6. H p (D) is a Banach space.

Let γ be such an arc. Since γ is contained in Dr for some r < 1, fn → g

and each of these integrals is zero because the fn are holomorphic.

Mp (fn − g, r) = lim Mp (fn − fm , r) ≤ lim kfn − fm kH p ≤ ε .

Since this holds for every r < 1, kfn − gkH p ≤ ε.

2. Harmonic versus holomorphic functions

A C 2 -function u defined on an open set Ω ⊆ Rn is called harmonic on Ω if its

is identically zero on Ω. A holomorphic function f is harmonic on its domain in

∂x2 f = −i∂x ∂y f = −∂y2 f .

In two dimensions, (2.1) becomes

Theorem 2.1. Let u be a continuous function on Ω. Then u is harmonic if and

where ν denotes the outer normal to ∂B(x, r).

It follows that, if u is harmonic, then ϕ0 = 0, hence ϕ is identically equal to u(x),

Since the expression in parentheses is positive, we conclude that v is a non-zero

with a continuous integrand u(x0 ) − <eu(x0 + ry) ≥ 0. Therefore u(x0 ) = <eu(x0 +

Lemma 2.3. Let f be a C 2 -function on T, and denote by fˆ(n), with n ∈ Z its

is uniformly convergent on T. If g is its sum, integration term by term shows that

uniformly convergent on B(z0 , R). In particular, u is real-analytic, and it is the

is uniformly convergent on D and harmonic in the interior. Since v = u on the

is constant. Therefore fz0 = fz0 0 on Uz ∩ Uz 0 , so that all these derivatives can be

The Dirichlet problem on D consists in assigning a continuous function f on T

in u ∈ C(D̄). The maximum modulus principle implies that a solution, if it exists,

If f (eit ) = eint with n ∈ Z, the solution of (3.1) is easily found as

Suppose next that f ∈ C 2 (T), and let

be its Fourier series. It is natural to construct

where the series converges uniformly on T. Then6

The functions Pr in (3.3) form the Poisson kernel on T.

Lemma 3.1. The Poisson kernel equals

The notion of “approximate identity” is quite general. We have chosen here

Observe that, by condition (2), we can write

If r0 < r < 1, we have

We apply Young’s inequality8 to the first summand, to obtain

k(f − g) ∗ ϕr kp ≤ kf − gkp kϕr k1 < Cε .

We then observe that

uniformly on T if f ∈ C(T), and in the Lp -norm if f ∈ Lp (T), with 1 ≤ p < ∞.

because all the terms with n 6= 0 give integral zero.