ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS

1. Differentiation in Rn → Rm
Consider the map F : Rn → Rm where F (x) = (F1 (x), . . . , Fm (x)). The directional
derivative of F in direction v is defined by:
F (a + hv) − F (a)
Dv F (a) := lim
h→0 h
∂F
and similarly the partial derivatives are defined by ∂x j
:= Dej F . The function F is
differentiable at a if ∃ a linear map called the differential (or derivative) DFa : Rn →
Rm such that:
F (a + h) − F (a) − DFa (h)
lim = 0.
h→0 |h|
Of course, differentiability may be checked component-wise in the Fj ’s. There are a few
pleasant expression which relate these various concepts:
(1) Dv F (a) = nj=1 vj Dj F (a) where v = (v1 , . . . , vn )
P
 ∂F1 ∂F1 
∂x1 . . . ∂x n
(2) DF =  . . . . . . . . . 
∂Fm
∂x1 . . . ∂F m
∂xn
n
(3) if f : R → R, then Dv f (a) = ∇f (a) · v
Differentiability Criterion. If the partial derivatives of a function exist and are contin-
uous in a neighbourhood of a point, then the function is differentiable at that point.
Chain Rule. If F : U ⊂ Rn → Rm is differentiable at a ∈ U and G : V ⊂ Rm → Rk is
differentiable at F (a) ∈ V , then G ◦ F is differentiable at a and
D(G ◦ F )a = DGF (a) · DFa .
Constant Functions. If U ⊂ Rn is an open connected set, then the differentiable mapping
F : U → Rm is constant if and only if DFx = 0 for all x ∈ U .
Mean Value Theorem. If f : U ⊂ Rn → R is differentiable and U contains the line
segment [a, b], then ∃ c ∈ [a, b] such that
f (b) − f (a) = f 0 (c)(b − a) = ∇f (c) · (b − a).
(Does not generalize to F : Rn → Rm .)
Equality of Mixed Partials. If f : U ⊂ Rn → R is C 2 , then Di Dj f = Dj Di f .
1
2 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS

Taylor Expansion. Let f : U ⊂ Rn → R be C 2 , then

1
f (x0 + h) = f (x0 ) + Dfx0 (h) + hT Hessx0 (f )h + R2 (x0 , h)
2
R2 (x0 ,h)
where limh→0 |h|2
= 0.

Second Derivative Test for Local Extrema. Let f : U ⊂ Rn → R be C 3 and suppose

that x0 ∈ U is a critical point of f .
(1) If Hessx0 is positive-definite, then x0 is a relative minimum of f .
(2) If Hessx0 is negative-definite, then x0 is a relative maximum.
Lagrange Multipliers. Let f : Rn → R and g1 , . . . , gm : Rn → R be C 1 functions.
If f has a local extremum at x0 ∈ {x : g1 (x) = . . . = gm (x) = 0} and the vectors
∇g1 (x0 ), . . . , ∇gm (x0 ) are linearly independent, then there are real numbers λ1 , . . . , λm
such that
∇f (x0 ) = λ1 ∇g(x0 ) + . . . + λm ∇g(x0 )

2. Vector Calculus
Basic Vector Operators. Let F : R3 → R3 , F = (F1 , F2 , F3 ) be a vector field and let
f : R3 → R be a scalar field.
(1) Gradient: ∇f = ( ∂f ∂f ∂f
∂x , ∂y , ∂z )
(2) Divergence: ∇ · F = ∂F ∂F2
∂x + ∂y + ∂z
1 ∂F3

i j k
∂ ∂ ∂
(3) Curl: ∇ × F (x) = ∂x ∂y ∂z
F1 F2 F3
(4) Gradients are curl-free: ∇ × (∇f ) = 0
(5) Curls are divergence-free: ∇ · (∇ × F ) = 0
Let γ(t) = (x(t), y(t), z(t)) be a C 1 parametrized curve for a ≤ t ≤ b.
Path Integral. If f : γ([a, b]) → R is continuous we define
Z Z b
f ds := f (γ(t))||γ 0 (t)|| dt.
γ a

Line Integral. If F = (F1 , F2 , F3 ) is a continuous vector field on γ([a, b]) we define

Z Z b Z
0
F · ds = F (γ(t)) · γ (t) dt = F1 dx + F2 dy + F3 dz
γ a γ

Fundamental Theorem of Calculus for Line Integrals.

Z
∇f · ds = f (γ(b)) − f (γ(a))
γ
 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS 3

Let Φ : D ⊂ R2 → R3 , Φ(u, v) = (x, y, z) be a C 1 parametrization of a surface. We

define the tangent vectors
∂Φ ∂Φ
Tv := and Tu :=
∂v ∂u
and say that a surface is regular wherever N := Tu × Tv 6= 0. As for path integrals, we
can integrate a scalar f : Φ(D) → R over the surface:
Z ZZ
f dS = f (Φ(u, v))||Tu × Tv || du dv.
Φ D

Surface Integral. If F is a continuous vector field on Φ(D), we define

Z Z
F · dS = F (Φ(u, v)) · (Tu × Tv ) du dv.
Φ D

Green’s Theorem. Let D ⊂ R2

be an oriented 2−manifold-with-boundary and suppose
∂D is positively oriented. If P dx + Q dy be a C 1 differential 1−form in a neighbourhood
of D, then ZZ Z
∂Q ∂P
− dA = P dx + Q dy.
D ∂x ∂y ∂D

Stoke’s Theorem. Let D be an oriented compact 2−manifold-with-boundary in R3 and

let N and T be the positively oriented unit normal and unit tangent vector fields on D and
∂D respectively. If F is a C 1 vector field on an open set containing D, then
ZZ Z
(∇ × F ) · N dA = F · T ds.
D ∂D

Gauss’ Divergence Theorem. Let F be a C 1 vector field defined on a neighbourhood of

the compact oriented smooth n−manifold with boundary V ⊂ Rn . Then
ZZZ ZZ
∇·F = F · N dA
V ∂V
where N is the unit outer normal vector field on the positively oriented ∂V .
Conservative Vector Fields. Let F be a C 1 vector field defined in a contractible region
of R2 or R3 . We say F is conservative if any of the following equivalent conditions hold:
R
(1) γ F · ds = 0 for every closed simple curve γ
(2) F = ∇f for some function f
(3) dF = ∇ × F = 0

3. Integration in Rn → Rm
Change of Variable. Let T : D → T (D) be C 1 −invertible on the interior of T (D), then
Z Z
∂(x, y)
f (x, y) dx dy = f (u(x, y), v(x, y)) du dv.
T (D) D ∂(u, v)
Cylindrical: dx dy dz = r dr dθ dz. Spherical: dx dy dz = ρ2 sin ϕ dρ dθ dϕ.
4 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS

Fubini-Tonelli Theorem. Let f be A × B measurable. If any one of the following three

conditions hold:
Z Z Z  Z Z 
|f (x, y)| d(x, y) < ∞, |f (x, y)|dy dx < ∞, or |f (x, y)|dx dy < ∞
A×B A B B A
then Z Z Z  Z Z 
f (x, y) d(x, y) = f (x, y)dy dx = f (x, y)dx dy.
A×B A B B A
(Tonelli’s Theorem states that for positive functions the iterated integrals of |f (x, y)| con-
verge/diverge together.)
Dominated Convergence Theorem. Let fn be a sequence of measurable Rfunctions con-
Rverging pointwise
R on A to a function f . If ∃ g ≥ 0 such that |fn | ≤ g and A g < ∞ then
f
A n → A f .
Another criterion to guarantee the result of the Dominated Convergence Theorem is
that the sequence fn be uniformly convergent.
Differentiating Under the Integral Sign. Let f : A × J → R be continuous and let
∂
∂t f (x, t) be uniformly continuous on A × J, then
Z Z
∂ ∂
f (x, t) dx = f (x, t) dx.
∂t A A ∂t
Improper Integration. Let f : U ⊂ Rn → R be locally integrable and let Uk be an
approximating sequence for U . We say that f (x) is absolutely integrable if
Z Z
|f | := lim |f | < ∞.
U k→∞ Uk

In this case, the following limit exists and is used as a definition of the improper integral
Z Z
f := lim f.
U k→∞ Uk

Comparison Test for Improper Integrals. Suppose that f and g are locally integrable
on U with 0 ≤ f ≤ g. If g is absolutely integrable on U , then so is f .
Useful test case:
Z ∞ ( (
∞ if p ≤ 1
Z 1 ∞ if p ≥ 1
1 1
dx = 1 & dx = 1
1 xp 1−p if p > 1 0 xp 1−p if p < 1
For functions of a single variable, we can sometimes assign a value even if the function
is not absolutely integrable. If f is continuous on [a, x0 [ ∪ ]x0 , b], we define the Cauchy
Principal Value:
Z b Z x0 − Z b 
PV f (x) dx := lim + f (x) dx.
a →0 a x0 +
 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS 5

4. Fundamental Real Analysis

xn+1
Ratio Test (for sequences). If (xn ) is a sequence of positive real numbers with limn→∞ xn =
L < 1 then xn → 0.
Monotone Convergence Theorem. A monotone sequence of real numbers is convergent
⇐⇒ it is bounded. Moreover, if it is decreasing (increasing) it converges to its infimum
(supremum).
Bolzano-Weierstrass Theorem. A bounded sequence of real numbers has a convergent
subsequence.
Cauchy Convergence Criterion. A sequence of real numbers is convergent if and only
if it is Cauchy.
Contractive Sequence Criterion. A sequence (xn ) of real numbers is contractive if
|xn+2 − xn+1 | ≤ K|xn+1 − xn |
for some 0 < K < 1. Such a sequence is Cauchy and therefore convergent.
Cauchy-Schwarz Inequality. For x, y in an inner-product space:
||hx, yi||2 ≤ hx, xihy, yi = ||x||2 ||y||2 .
Minkowski Inequality. Let 1 ≤ p and ak , bk ∈ C, then
n
!1 n
!1 n
!1
X p X p X p

|ak + bk |p ≤ |ak |p + |bk |p .

k=1 k=1 k=1
1 1
Holder Inequality. If p = 1 and ak , bk ∈ C, then
+ q

n n
!1 n
!1
X X p X q
p p q
|ak · bk | ≤ |ak | · |bk | .
k=1 k=1 k=1

Absolute Convergence Tests. Let an and bn be sequences of positive real numbers.

(1) Limit Comparison Test:
X bn X
an < ∞ and lim sup < ∞ =⇒ bn < ∞
n→∞ an
X bn X
an = ∞ and lim inf > 0 =⇒ bn < ∞.
n→∞ an
(2) Ratio Test:
an+1 X an+1 X
lim sup < 1 =⇒ an < ∞ & lim inf > 1 =⇒ an = ∞
n→∞ an n→∞ an
(3) Root Test:
X X
lim sup (an )1/n < 1 =⇒ an < ∞ & lim sup (an )1/n > 1 =⇒ an = ∞
n→∞ n→∞
6 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS
P∞ n
Alternating Series Test. If an ≥ 0, an+1 ≤ an and limn→∞ an = 0, then n=1 (−1) an
converges.
Integral Test. If f : [c, ∞[→ [0, ∞[ is a positive decreasing continuous function with
f (n) = an for all n ≥ c, then
X∞ Z ∞
an converges ⇐⇒ f (x) dx < ∞.
n=1 c

Continuous Functions. The function f : X → Y between metric spaces X and Y is

continuous at x0 ∈ X if one of the following conditions hold:
(1) ∀ ε > 0 ∃ δ > 0 such that |x − x0 | < δ =⇒ |f (x) − f (x0 )| < 
(2) f (x0 ) ∈ U open =⇒ f −1 (U ) open (for all such U )
(3) xn → x =⇒ f (xn ) → f (x) (for all such sequences xn )
Continuous Functions on Compact Sets. Let X be a compact metric space.
(1) If f : X → R is continuous, f (X) is a bounded set and f attains its supremum and
infimum.
(2) If Y is any metric space and f : X → Y is continuous, then f is uniformly
continuous.
Taylor Expansion. Let f : [a, b] → R be n−times differentiable. If x, x0 ∈ [a, b], then
f 0 (x0 ) f (n) (x0 )
f (x) = f (x0 ) + (x − x0 ) + . . . + (x − x0 )n + Rn (x)
1! n!
Rn (x)
where (x−x0 )n → 0 as x → x0 . There are other forms of the remainder:
(n+1)
(1) If f (n+1) exists in ]a, b[, we may write Rn (x) = f (n+1)!(ξ) (x − x0 )n+1 for some
ξ ∈]x0 , x[.
1 x (n+1)
(2) If f (n+1) is integrable on [a, b], we may write Rn (x) = n! (t)(x − t)n dt.
R
x0 f

Mean Value Theorem. Let f : [a, b] → R be continuous. If f is differentiable in ]a, b[,

then there is some ξ ∈]a, b[ such that
f (b) − f (a) = f 0 (ξ)(b − a).
Integral Mean Value Theorem. If f, g : [a, b] → R where f (x) is continuous and
0 ≤ g(x) is integrable, then there is some ξ ∈ [a, b] such that
Z b Z b
f (x)g(x) dx = f (ξ) g(x) dx.
a a

Intermediate Value Theorem. Let f : [a, b] → R be continuous. If f (a) < k < f (b),
then there is some ξ ∈ [a, b] such that f (ξ) = k.
Darboux’s Theorem. Let f : [a, b] → R be differentiable. If f 0 (a) < k < f 0 (b) then there
is some ξ ∈]a, b[ such that f 0 (ξ) = k.
 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS 7

The Class of Riemann Integrable Functions. Let f, g : [a, b] → R be bounded functions

(an integrable function is necessarily bounded).
(1) (Lebesgue) f (x) is integrable ⇐⇒ it is continuous almost everywhere.
(2) (Composition) f (x) integrable & ϕ(y) continuous =⇒ ϕ ◦ f integrable
(3) (Absolute Value) f (x) integrable ⇐⇒ |f |(x) integrable
(4) (Product) f (x) & g(x) integrable =⇒ f (x) · g(x) integrable
Uniformly Continuous Functions. Let f : A → R be a function.
(1) f is Lipschitz =⇒ f uniformly continuous
(2) f uniformly continuous =⇒ (f (xn )) is Cauchy whenever (xn ) is Cauchy in A
Continuous Extension Theorem. A function f :]a, b[→ R is uniformly continuous
⇐⇒ it admits a continuous extension f : [a, b] → R.
Weierstrass Approximation Theorem. Let f : [a, b] → R be continuous. Given ε > 0,
there is a polynomial pε (x) such that |f (x) − pε (x)| < ε for all x ∈ [a, b].
Banach Fixed Point Theorem. Let X be a complete metric space. If T : X → X
satisfies |T (x) − T (y)| ≤ K|x − y| for some K < 1 (i.e. it is a contraction), then T has a
unique fixed point.
Newton’s Method. Let f : [a, b] → R be twice differentiable and suppose that f (a)f (b) <
0. If |f 0 (x)| ≥ m > 0 and |f 00 (x)| ≤ M for all x ∈ [a, b], then there is a subinterval I ⊂ [a, b]
containing a root of f (r) = 0 such that for any x1 ∈ I the sequence
f (xn )
xn+1 := xn −
f 0 (xn )
converges to r.
Inverse Function Theorem. Let f : [a, b] → R be strictly monotone and continuous. If f
is differentiable at c and f 0 (c) 6= 0 then f −1 is differentiable at f (c) and (f −1 )0 (f (c)) = f 01(c) .
More generally, let F : U ⊂ Rn → Rn be C 1 and let a ∈ U . If DF (a) is invertible, then
F is invertible in a neighbourhood of F (a).
Given G : Rm × Rn → Rn and F : U ⊂ Rm → Rn we say that y = F (x) solves the
equation G(x, y) = 0 in a neighbourhood W if G(x, y) = 0 ⇐⇒ y = F (x) whenever
(x, y) ∈ W .
Implicit Function Theorem. Let G : U ⊂ Rm × Rn → Rn be C 1 in a neighbourhood of
(a, b) with G(a, b) = 0. If ∂G m
∂y (a, b) is invertible, then there is a function F : U ⊂ R → R
n

which solves G(x, y) = 0 in some neighbourhood of (a, b). Moreover, since

d ∂G dx ∂G d ∂G ∂G
0= G(x, F (x)) = + F (x) = (x, y) + (x, y)F 0 (x)
dx ∂x dx ∂y dx ∂x ∂y
h i−1
we can implicitly compute F 0 (x) = − ∂G
∂y (x, y) ∂G
∂x (x, y).
8 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS

5. Analysis on Functions
P
Weierstrass M-test. Let Mk ≥ 0 be a sequence of complex numbers for which Mk < ∞.
If fn : X ⊂ PC → C is a sequence of functions such that |fn (x)| ≤ Mn for all x ∈ X and
n ∈ N, then fn (x) converges absolutely and uniformly on X. In this case we can integrate
and differentiate the series of functions term by term.
Differentiability of Limit Functions. If fn (x) is a sequence of differentiable real-valued
functions defined on an open interval, then
h pointwise unif ormly
i h unif ormly
i
fn −−−−−−→ f & fn0 −−−−−−→ g =⇒ fn −−−−−−→ f & f 0 = g .
1
Counterexample: fn (x) = n sin(nx).
Power Series. Every power series k ak z k has a radius of convergence 0 ≤ ρ ≤ +∞
P

for which k ak z k converges absolutely if |z| < ρ and diverges if |z| > ρ. Moreover, the
P
series converges uniformly on compact subsets.
(1) The radius of convergence is given by the Cauchy-Hadamard Formula
1
ρ= .
lim sup |ak |1/k
(2) Power series can be added and multiplied. In this case the radius of convergence
may shrink to the minimum of the two radii.
(3) If a0 6= 0 we may invert a power series (or compose it with another) formally. In
this case, we just know that the radius of convergence remains strictly positive.
(4) Uniform convergence on compact subdisks allows power series to be differentiated
and integrated term by term without affecting the radius of convergence. This yields
a formula for its terms:
f (k) (0)
ak =
k!
which shows that a power series is uniquely determined by the function it represents.
1
= ∞ k
P
Important Power Series. (1) 1−z k=0 z , |z| < 1.
P∞ k+1
(2) Log(1 − z) = k=0 − zk+1 , |z| < 1
2 3
(3) ez = 1 + z + z2! + z3! + . . . , |z| < ∞
3 5
(4) sin(z) = z − z3! + z5! − . . . , |z| < ∞
2 4
(5) cos(z) = 1 − z2! + z4! − . . . , |z| < ∞
Dini’s Theorem. If fn : A → R is a sequence of continuous functions defined on a
compact metric space, then
pointwise unif ormly
fn −−−−−−→ f & fn (x) ≥ fn+1 (x) =⇒ fn −−−−−−→ f.
Arzela-Ascoli Theorem. Let fn : A → C is a sequence of continuous functions defined on
a compact metric space. If (fn ) is pointwise bounded & equicontinuous, then it is uniformly
bounded and contains a uniformly convergent subsequence.
 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS 9

Fourier Series. Let f (x) be defined in ]−L, L[ and determined outside this interval by
f (x + 2L) = f (x). The Fourier Series Expansion of f (x) is by definition
∞ ∞
a0 X  nπx nπx  X nπx
+ an · cos + bn · sin = cn · ei L
2 L L −∞
n=1

where for any ξ ∈ R we have

1 ξ+2L 1 ξ+2L
Z Z
nπx nπx
an := f (x) cos dx, bn := f (x) sin dx
L ξ L L ξ L
and
Z ξ+2L
1 nπx
cn := f (x)e−i L dx.
2L ξ

Dirichlet Conditions. Let f (x) be defined and single-valued except possibly at a finite
number of points in ] − L, L[ and suppose that f (x + 2L) = f (x).
(1) If f (x) and f 0 (x) are both piecewise continuous, then the Fourier series converges
pointwise to f (x+)+f2
(x−)
.
(2) If f is continuous and f 0 is piecewise continuous, then the Fourier series converges
uniformly to f .
Bessel Inequality and Parseval Identity. If {ϕk }∞ k=0 be an orthonormal system in an
inner-product space V , then:
P∞ 2
P∞ 2 2
(1) k=0 |hf, ϕk i| converges and k=0 |hf, ϕk i| ≤ hf, f i = ||f || for all f ∈ V
∞
(2) {ϕk }∞ 2
P
k=0 is complete ⇐⇒ k=0 |hf, ϕk i| = hf, f i for all f ∈ V.

6. Basic Properties of Analytic Functions

For z 6= 0 we have the following multivalued functions:
(1) log z := log |z| + i arg z = log |z| + i Arg z + 2πmi, m ∈ Z and −π < Arg z ≤ π
(2) z α := eα log z = |z|α eiα arg z
along with their principal branches (other branches specified by choice of m ∈ Z)
(1) Log z := log |z| + i Arg z with branch cut z ∈]−∞,
/ 0]
(2) z α = |z|α eiα Arg z with branch cut z ∈ / [0, ∞[.
1
1
Warning log z n 6= n log z while log z n = n log z as multivalued sets
Phase Change Lemma. Let g(z) be a single valued function, continuous in a neighbour-
hood of z0 . For any branch of (z − z0 )α , the function f (z) = (z − z0 )α g(z) is multiplied by
the phase factor e2πiα when z traverses a complete circle about z0 in the positive direction.
Cauchy-Riemann Equations. The function f = u + iv defined on the domain D is
analytic if and only if the following two conditions hold:
(1) The functions u(x, y) and v(x, y) have continuous first-order partial derivatives.
10 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS

(2) The Cauchy-Riemann equations hold:

∂u ∂v ∂u ∂v
= and =− .
∂x ∂y ∂y ∂x
Analyticity of Limit Functions. Let fk (z) be analytic on D.
unif ormly
(1) If (fk (z)) −−−−−−→ f (z), then f (z) is analytic on D.
normally (m) normally
(2) If (fk (z)) −−−−−−→ f (z), then (fk (z)) −−−−−−→ f (m) (z).
Here a sequence (fk (z)) of analytic functions on D converges normally to the analytic
function f (z) if it converges uniformly to f (z) on every closed disk contained in D.
Proof. Morera’s Theorem. 
Laurent Decomposition. Let f (z) is analytic on the annulus 0 ≤ ρ < |z − z0 | < σ ≤ ∞.
(1) The function can be decomposed as a sum f (z) = fσ (z) + fρ (z) where fσ (z) is
analytic for |z − z0 | < σ and fρ (z) is analytic for |z − z0 | > ρ (including at ∞).
If we normalize the decomposition so that fρ (∞) = 0, then the decomposition is
unique.
(2) Expressing fσ and fρ in terms of power series, we obtain the Laurent series
−1
X ∞
X ∞
X
k k
f (z) = ak (z − z0 ) + ak (z − z0 ) = ak (z − z0 )k , ρ < |z − z0 | < σ
k=−∞ k=0 k=−∞
converging absolutely (and uniformly on compact sub-annuli) with coefficients
Z
1 f (z)
ak = dz.
2πi |z−z0 |=r (z − z0 )k+1
Proof. Uniqueness = Liouville’s Theorem, existence = the Cauchy Integral Formula. 
Classification of Isolated Singularities. Let z0 be an isolated singularity of
∞
X
f (z) = ak (z − z0 )k , 0 < |z − z0 | < σ.
k=−∞
(1) The singularity is removable if ak = 0 for all k < 0.
(2) The singularity is a pole of order N if a−N 6= 0 and ak = 0 for all k < −N .
(3) Otherwise, the singularity is essential.
If z0 = ∞ we apply the above definitions to f (1/z).
Riemann’s Theorem on Removable Singularities.
isolated singularity z0 of f (z) removable ⇐⇒ f (z) bounded in neighbourhood of z0
Detecting Poles. Let z0 be an isolated singularity of f (z).
(1) z0 is a pole ⇐⇒ |f (z)| → ∞ as z → ∞
(2) z0 is a pole of order N ⇐⇒ f (z) = g(z)/(z − z0 )N where g(z) analytic at z0 and
g(z0 ) 6= 0
 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS 11

Casorati-Weierstrass Theorem. If z0 is an essential singularity of f (z), then for every

w0 ∈ C there is a sequence zn → z0 such that f (zn ) → w0 .
Picard’s Theorem. There are two:
(1) Little: The image of an entire & non-constant function misses at most one point
of C.
(2) Big: If an entire function has an essential singularity at z0 , then it assumes all
possible complex values with at most a single exception infinitely often in any neigh-
bourhood of z0 .
Partial Fractions Decompositon. A meromorphic function on C∗ is rational. Every ra-
tional function has a partial fraction decomposition expressing it as the sum of a polynomial
in z and its principal parts at each of its poles in the finite complex plane.
Uniqueness Principle. Let f (z) and g(z) be analytic on D.
(1) If f (z) is not identically zero on D, then the zeroes of f (z) are isolated.
(2) If f (z) = g(z) for all z belonging to a subset D of D containing a non-isolated
point, then f (z) = g(z) on D.
Analytic Continuation. Let f (z) = ak (z − z0 )k for |z − z0 | < ρ. We say f (z) has an
P
analytic continuation along a path γ(t) parametrized by a ≤ t ≤ b if for every t there is
a convergent power series
∞
X
ft (z) := ak (t)(z − γ(t))n , for |z − γ(t)| < ρ(t)
k=0

with fa (z) = f (z) and fs (z) = ft (z) wherever their disks of convergence intersect.
Such an analytic continuation is unique. Further, the coefficients ak (t) and radii of con-
vergence ρ(t) depend continuously on t. The Monodromy Theorem ensures that analytic
continuations along homotopic paths upon which f (z) admits analytic continuations must
coincide.

7. Harmonic Functions and Bounds on Analytic Functions

2 2
A function u(x, y) is harmonic if ∆u := ∂∂xu2 + ∂∂yu2 = 0 and the partial derivatives are
continuous. If f = u+iv is analytic, then u(x, y) and v(x, y) are harmonic as a consequence
of the Cauchy-Riemann equations.
Harmonic Conjugate. Let D be a simply connected domain. If u(x, y) is a real val-
ued harmonic function on D, then there is a unique (up to adding a constant) harmonic
conjugate v(x, y) such that f = u + iv is analytic on D.
Mean Value Property. Let u(z) be harmonic on D. If {|z − z0 | < ρ} ⊂ D, then
Z 2π
1
u(z0 ) = u(z0 + reiθ ) dθ for every 0 < r < ρ.
2π 0
12 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS

It turns out that a continuous function on D has the mean value property if and only if
it is harmonic.
Strict Maximum Principle. Let u(z) be a real valued harmonic function on D with
u(z) ≤ M for all z ∈ D. If u(z0 ) = M for some z0 ∈ D, then u(z) ≡ M .
Proof. Mean value property =⇒ {z : u(z) = M } open. Continuity =⇒ {z : u(z) < M }
is open. Contradicts connectedness of D. 
Maximum Modulus Principle. Let h(z) be a a complex-valued harmonic (analytic)
function on D.
(1) If |h(z)| ≤ M for all z ∈ D and |h(z0 )| = M for some z0 ∈ D, then h(z) is constant
on D.
(2) Suppose further that D is bounded and h(z) extends continuously to ∂D. If |h(z)| ≤
M for all z ∈ ∂D, then |h(z)| ≤ M for all z ∈ D.
Cauchy Estimates. Let f (z) be analytic for |z − z0 | ≤ ρ. If |f (z)| ≤ M |z − z0 | = ρ, then
fore every m ≥ 0 we have the following bound:
m!
|f (m) (z0 )| ≤ M.
ρm
Liouville’s Theorem. A bounded entire function is constant. Here, an entire function
is one who is analytic on the entire complex plane.

8. Complex Integration
Fundamental Theorem of Calculus. As usual, there are two parts:
(1) Let f (z) be continuous on D. If F (z) is a primitive of f (z), then
Z B
f (z) dz = F (B) − F (A)
A

and the integral can be taken over any path in D from A to B.

(2) Let D be simply connected. If f (z) is analytic on D, then a primitive is given by
Z z
F (z) := f (w) dw, for z ∈ D
z0

where z0 is any fixed point and the integral can be taken along any path in D.
Cauchy’s Theorem. Let D be a bounded domain with R piecewise smooth boundary. If f (z)
is analytic on D and extends smoothly to ∂D, then ∂D f (z) dz = 0.

Proof. A C 1 f (z) on D is analytic if and only if the differential f (z) dz is closed + Green’s
Theorem. 
 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS 13

Cauchy Integral Formula. Let D be a bounded domain with piecewise smooth boundary.
If f (z) is an analytic function on D which extends smoothly to ∂D, then for every z0 ∈ D
and m ≥ 0 we have: Z
(m) m! f (z)
f (z0 ) = dz
2πi ∂D (z − z0 )m+1
Morera’s Theorem. Let f (z) be a continuous function on D. If ∂T f (z) dz = 0 for every
R
closed triangle T contained in D, then f (z) is analytic on D.
Residue Theorem. Let D be a bounded domain with piecewise smooth ∂D. If f (z) is
analytic on D ∪∂D except at the isolated singularities z1 , . . . , zm ∈ D, then
Z Xm
f (z) dz = 2πi Res[f (z), zj ].
∂D j=1

If D is an exterior domain with piecewise smooth ∂ D and we let a−1 be the coefficient of
1
z in the Laurent expansion of f (z) convergent for |z| > R, then
Z Xm
f (z) dz = −2πia−1 + 2πi Res[f (z), zj ].
∂D j=1

Equivalently, −a−1 = Res[f (z), ∞] = Res[− z12 f ( z1 ), 0].

Computing Residues. Two tricks:
(1) If f (z) has a pole of order n at z0 , then
1 dn−1
Res[f (z), z0 ] = lim [(z − z0 )n f (z)].
z→z0 (n − 1)! dz n−1
(2) If f (z) and g(z) are analytic at z0 and g(z) has a simple zero at z0 , then
 
f (z) f (z0 )
Res , z0 = 0 .
g(z) g (z0 )
Jordan’s Lemma. If ΓR is the contour z(θ) = Reiθ , 0 ≤ θ ≤ π and a > 0, then
Z
π
|eiaz ||dz| < .
ΓR a
Fractional Residue Theorem. Let z0 be a simple pole of f (z) and let C be an arc of
the circle |z − z0 | = } of angle α, then
Z
lim f (z) dz = αi · Res[f (z), z0 ].
→0 C

Argument Principle. Let D be bounded with piecewise smooth ∂ D. If f (z) is meromor-

phic on D and extends to be analytic on ∂ D with f (z) 6= 0 for all z ∈ ∂ D, then
f 0 (z)
Z Z
d log f (z) = dz = 2πi · [N0 − N∞ ] .
∂D ∂D f (z)
14 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS

Splitting up log f (z) = log |f (z)| + i arg f (z) we obtain the following restatement
Z
d arg f (z) = 2π · [N0 − N∞ ] .
∂D

Rouché’s Theorem. Let D be bounded with piecewise smooth ∂D. If f (z) and h(z) are
analytic on D ∪∂D with |h(z)| < |f (z)| on ∂D, then f (z) and f (z) + h(z) have the same
number of zeroes in D (with multiplicities).
Open Mapping Theorem. A non-constant analytic map on a domain D is open.
Fourier Transform. If f be absolutely integrable on R, its Fourier Transform is well
defined by Z ∞
1
ˆ
f (w) = √ f (x)e−ixw dx.
2π −∞
If fˆ is absolutely integrable and f is continuous at x, then the Fourier Inversion Formula
holds: Z ∞
1
f (x) = √ fˆ(w)eixw dw.
2π −∞
Riemann-Lebesgue Lemma.
Z
1
f ∈ L =⇒ lim f (x)einx dx = 0.
n→∞

9. Complex Mappings
Inverse Function Theorem. If f (z) is analytic on D and f 0 (z0 ) 6= 0, then f has an
analytic inverse f −1 in some neighbourhood of f (z0 ) where (f −1 )0 (f (z)) = f 01(z) .
Conformal Mapping Theorem.
f (z) analytic at z0 & f 0 (z0 ) 6= 0 =⇒ f (z) conformal at z0
 

Schwarz Lemma. Let f (z) be analytic for |z| < 1. If |f (z)| ≤ 1 for all |z| < 1 and
f (0) = 0, then:
(1) |f (z)| ≤ |z| for |z| < 1 where equality at any z0 6= 0 implies that f (z) = eiθ z
(2) |f 0 (0)| ≤ 1 where equality implies that f (z) = eiθ z.
Mobius Transformations. A Mobius Transformation is a function of the form
az + b
f (z) = where ad − bc 6= 0
cz + d
(1) Given distinct z0 , z1 , z2 in C∗ there is a Mobius transformation mapping them onto
any distinct w0 , w1 , w2 ∈ C∗ .
(2) The group of Mobius transformation is generated by dilations, translations and
inversions.
(3) A Mobius transformation maps circles to circles in C∗ .
 ANALYSIS QUALIFYING EXAM EXPANDED SYLLABUS 15

Important Conformal Maps.

H = {=(z) > 0}, D = {|z| < 1} and Sα = {0 < arg z < α}
z−a
(1) Aut D = {eiθ 1−āz : |a| < 1}
(2) z 7→ z−i
z+i maps H conformally onto D
1+w
(3) w 7→ i · 1−w maps D conformally onto H
π
(4) z 7→ z α maps Sα onto H
0 ) (z1 −z2 )
(5) z 7→ (z−z
(z−z2 ) (z1 −z0 ) sends z0 7→ 0, z1 7→ 1 and z2 7→ ∞