§4.7.

Primitive Functions

Although the value of a path integral of a function f from a fixed point

z1 to a fixed point z 2 depends, in general, on the path that is taken,
there are certain functions whose integrals are from z1 to z 2 have
values that are independent of path. (Compare Examples 2 and 3 in Sec.
4.5.) The examples just cited also illustrate the fact that the values of
integrals around closed paths are sometimes, but not always, zero. The
theorem below is useful in determining when integration is independent of
path and, moreover, when an integral around a closed path has value zero.
In proving the theorem, we shall discover an extension of the
fundamental theorem of calculus that simplifies the evaluation of many
path integrals. That extension involves the concept of primitive function,
which is defined as follows.
Definition 4.7.1. For a function f defined on a domain D , if a
function F statisfies F ( z )  f ( z ) for all z in D , then we call
F a primitive function of f .
 Note that a primitive function is necessarily an analytic function.
Note, too, that primitive function of a given function f is unique except
for an additive complex constant. This is because the derivative of the
difference F ( z )  G( z ) of any two such primitive functions F (z )
and G(z ) is identically zero on the domain D ; and, according to
Theorem 2.13.1, an analytic function is constant in a domain D when its
derivative is zero throughout D .
Theorem 4.7.1. Suppose that a function f is continuous on a
domain D . Then the following statements are equivalent:
(i ) f has a primitive function F in D .
(ii ) The integrals of f along any simple paths lying entirely in
D and extending from any fixed point z1 to any fixed point z 2 all
have the same value.
(iii ) The integrals of f around any closed simple paths lying
entirely in D all have value zero.
 Proof. (i )  (ii ) : Let us assume that statement (i ) is true. Let
z1, z2  D and C be any simple path from z1 to z 2 lying in D .
When C is a smooth arc, we let z  z (t )(a  t  b) be the
parametric representation of C , then we know from that
d
F[ z (t )]  F [ z (t )]z (t )  f [ z (t )]z (t ) (a  t  b) .
dt
Because the fundamental theorem of calculus can be extended so as to
apply to complex-value functions of a real variable (Sec. 4.2), it follows
that
b
 f ( z )dz   f [ z (t )]z (t )dt  F[ z (t )] a  F[ z (b)]  F ( z (a)] .
b
C a

Since z (b)  z 2 and z (a)  z1 , we may write

 f ( z )dz  F ( z2 )  F ( z1 )  F ( z ) z 2 .
z
(4.7.1)
C 1
 When C consists of a finite number of smooth arcs Ck (k  1,2,, n) , let
each C k extend from a point ck to a point ck 1 , where c1  z1 , cn 1  z2 . Then
by (4.7.1)
n n

 f ( z )dz    f ( z )dz  [ F (ck 1 )  F (ck )]  F (cn 1 )  F (c1 )  F ( z2 )  F ( z1 ) .

C Ck
k 1 k 1
This shows that the value of the path integral is always, which is evidently
independent of the path C as long as C extends from z1 to z 2 and lies
entirely in D .
(ii )  (iii ) : Let C be any simple closed path lying in D and let z1 and
z 2 denote any two distinct points on C . We divide C into two paths C1 and
C 2 , each with initial point z1 and final point z 2 , such that C  C1  C2 (Fig.
4-14(a)).
 Fig. 4-14(a) Fig. 4-14(b)

Assuming that statement (ii ) is true, one can write

 C1
f ( z )dz  
C2
f ( z )dz , (4.7.2)
that is,

C1
f ( z )dz  
C 2
f ( z )dz  0 . (4.7.3)
Thus, the integral of f around the closed path C  C1  C2 has value
zero.
(iii )  (i ) : We assume that statement (iii ) is true. We let C1
and C 2 denote any two simple paths, lying in D , from a point z1 to a
point z 2 in D . Take a simple path C3 from z1 to a point z 2 such
that C1  C3 and C2  C3 are simple paths lying in D (see Fig.
4-14(b)). From (iii ) , we get that

 C1
f ( z )dz  
 C3
f ( z )dz  0 and  C2
f ( z )dz  
 C3
f ( z )dz  0 .
Thus, equation (4.7.2) holds. Integration is therefore independent of path
in D ; and we can define a function
z
F ( z )   f ( s)ds
z0

on D where the point z 0 is any fixed point in D . The proof of the

theorem is complete once we show that F ( z )  f ( z ) everywhere in
D . Let z be any point in D . Then there is an r  0 such that
z  z  D when | z | r . Then, when 0 | z | r ,
z  z z z  z
F ( z  z )  F ( z )   f (s)ds   f ( s)ds   f (s)ds ,
z0 z0 z

where the path of integration from z to z  z may be selected as a

line segment (Fig. 4-15). Since
z  z
 z
ds  z
we can write
1 z  z
z  z
f ( z)  f ( z )ds ;
and it follows that
F ( z  z )  F ( z ) 1 z  z
z
 f ( z) 
z  z
[ f ( s)  f ( z )]ds .
 Fig. 4-15
But f is continuous at the point z . Hence, for each positive number
 , there exists a positive number   r such that
| f (s)  f ( z) |  whenever | s  z |  . Consequently, when
0 | z |  , we have from Theorem 4.6.1 that
F ( z  z )  F ( z ) 1
 f ( z)   | z |  .
z | z |
This proves that
F ( z  z )  F ( z )
lim  f ( z) ,
z 0 z
that is, F ( z )  f ( z ) . Thus, the function F is a primitive function of
the function f .
This completes the proof.
From the proof of Theorem 4.7.1 we see that the following result is
valid, which gives a formula for computing a path intergral.
Theorem 4.7.2. Suppose that a function f is continuous on a
domain D and has a primitive function F in D . If C is a simple
path from z1 to z 2 lying in D , then

 f ( z )dz  F ( z2 )  F ( z1 )  F ( z ) z 2 ,
z
(4.7.4)
C 1

which is called the Newton-Leibnitz formula.

For example, since (sin z )  cos z (z  C) , we have from (4.7.4)
that
B
 A
cos zdz  sin B  sin A(A, B  C) .
§4.8. Examples

The following examples illustrate Theorem 4.7.1 in Sec. 4.7 and, in

particular, the use of the extension (4.7.1) of the fundamental theorem of
calculus in that section.
Example 1. The continuous function f ( z )  z 2 has a primitive
function F ( z )  z / 3 throughout the plane. Hence, by (4.7.4) we have
3

3 1i
1i z 1 2
0
z 2 dz 
3
 (1  i) 3  (1  i)
3 3
0

for every path form z  0 to z  1  i .

Example 2. The function f ( z )  1 / z 2 , which is continuous
everywhere except at the origin, has a primitive function F ( z )  1 / z
in the domain | z | 0 , consisting of the entire plane with the origin
deleted.
 dz
Consequently, by Theorem 4.7.1, C z 2  0 whenever C is the
positively oriented circle (Fig. 4-16)
z  2e i (     ) (4.8.1)
about the origin.

Fig. 4-16
 Example 3. Let C1 denote the right half
  
z  2e i      (4.8.2)
 2 2
of the circle C in Example 2. The principal branch
log z  ln r  i ( z  rei , r  0,     )
of the logarithmic function serves as a primitive function of the function
1 / z in the evaluation of the integral of 1 / z along C1 (Fig. 4-17):
dz 2 i dz
C1 z  2i z
 
2i
log z  2i

   
  ln 2  i    ln 2  i 
 2  2
 i.
This integral was evaluated in another way in Example 1, Sec. 4.5, where
representation (4.5.2) for the semicircle was used.
 Next, let C 2 denote the left half
 3 
z  2e i     (4.8.3)
 2 2 
of the same circle C and consider the branch
L0 ( z)  ln r  i ( z  rei , r  0,0    2 )

Fig. 4-17 Fig. 4-18

of the logarithmic function (Fig. 4-18). One can write by (4.7.4) that
dz  2 i dz  2i
C2 z 2i z 0 2i  L0 (2i)  L0 (2i)
  L ( z )

 3   
  ln 2  i    ln 2  i 
 2   2
 i.
The value of the integral of 1 / z around the entire circle
C  C1  C2 is thus obtained:
dz dz dz
C z C1 z C2 z  i  i  2i .
 
§4.9. Cauchy Integral Theorem

In Sec. 4.7, we saw that when a continuous function f has a primitive

function in a domain D , the integral of f around any given closed path
C lying entirely in D has value zero. In this section, we present a
theorem, called Cauchy Integral Theorem, giving other conditions on a
function f , which ensure that the value of the integral of f around a
simple closed path is zero. The theorem is central to the theory of
functions of a complex variable; and some extensions of it, involving
certain special types of domains, will be given in Sec. 4.11.
Lemma 4.9.1. Let f be piecewise continuous on a path
C : z  z(t ) (a  t  b) , described in the positive sense
(counterclockwise), then

C
f ( z )dz   udx  vdy  i  vdx  udy ,
C C

where f ( z )  u( x, y)  iv ( x, y) .
 Proof. We may assume that f is continuous on C . According to
the definition of the integral, we have
b
 C
f ( z )dz   f [ z (t )]z (t )dt .
a
(4.9.1)

We write z (t )  x(t )  iy (t ) , then the integrand f [ z (t )]z (t ) in

expression (4.9.1) is the product of the functions
u[ x(t ), y(t )]  iv[ x(t ), y(t )] and x(t )  iy(t )
of the real variable t . Thus
f ( z )dz   u ( x(t ), y (t )) x(t )  v( x(t ), y (t )) y(t )dt
b
 C a

 i  v( x(t ), y (t )) x(t )  u ( x(t ), y (t )) y(t )dt.

b

a
(4.9.2)
In terms of line integrals of real-valued functions of two real variables, we
have

 C
f ( z )dz   udx  vdy  i  vdx  udy .
C C
(4.9.3)
This completes the proof.
 With Lemma 4.9.1, we can prove the following result which will be
generalized as the so-called Cauchy Integral Theorem.
Proposition 4.9.1. If C is a simple closed path and a function f
is analytic on R  ins(C ) and f  is continuous there, then

C
f ( z )dz  0 . (4.9.4)
 Proof. f ( z)  u( x, y)  iv( x, y)  u  iv , then from
Write
Lemma 4.9.1, we get (4.9.3) hold. Since f is is analytic on
R  ins(C ) , the functions u and v are also continuous on R .
Likewise, if the derivative f  of f is continuous on R , so are the
first-order partial derivatives of u and v . Green’s theorem then enables
us to rewrite equation (4.9.3) as

C
f ( z )dz    (v x  u y )dxdy  i   (u x  v y )dxdy .
R R
(4.9.5)

But, in view of the Cauchy-Riemann equations ux  v y , u y  vx , the

integrands of these two double integrals are zero throughout R . So,
(4.9.4) is valid. This completes the proof.
Note that, once it has been established that the value of this integral is
zero, the orientation of C is immaterial. That is, statement (4.9.4) is also
true if C is taken in the clockwise direction, since


C
f ( z )dz  
C
f ( z )dz  0 .
 Example 1. If C is any simple closed path, in either direction, then

C
exp( z 3 )dz  0 .
This is because the function f ( z )  exp( z 3 ) is analytic
everywhere and its derivative f ( z )  3z 2 exp( z 3 ) is continuous
everywhere.
This result was obtained by Cauchy in the early part of the nineteenth
century. Goursat was the first to prove that the condition of continuity on
f  can be omitted. We now state the revised form, known as the Cauchy
Integral Theorem.
Theorem 4.9.1(Cauchy Integral Theorem, CIT). If C is a simple
closed path and a function f is analytic on R  ins(C ) , then

C
f ( z )dz  0 .
The proof is presented in the next section, where, to be specific, we
assume that C is positively oriented. The reader who wishes to accept
this theorem without proof may pass directly to Sec. 4.11.
 Example 2. If C is the unit circle | z | 1 , then

 sin z  e 
cos z dz  0 .
z
C

This is because that the integrand f ( z )  sin z  e z cos z is

analytic on the closed unit disk.
§4.10. Proof of Cauchy Integral Theorem

To complete the proof of the Cauchy integral theorem, we starte with

several lemmas. Recall that a simply connected domain D is a domain
such that every simple closed path within it encloses only points of D .
Roughly speaking, a domain is simple connected if and only if it has no
“holes”.
Lemma 4.10.1. Let D be a simple closed domain and f is
analytic in D , then for every closed polygonal line C  D ,

 C
f ( z )dz  0 . (4.10.1)

Proof. Case 1: Let C  D be the boundary  of a triangle  ,

see Fig. 4-20.
 Fig. 4-20

Put I   C
f ( z )dz . Joining the midpoints of the three sides of 
divides the triangle  into four triangles 1 ,  2 , 3 shown as in the
figure. Clearly,

 C
f ( z )dz  
1
f ( z )dz  
 2
f ( z )dz  
 3
f ( z )dz  
 4
f ( z )dz .
 I
Thus, the modulus of one of the four integrals is larger than or equal to ,
4
say
I
 1
f ( z )dz 
4
.

Similarly, joining the midpoints of the three sides of 1 divides the

triangle 1 into four triangles in which there is at least one triangle, say
 2 , with
I
  2 and  2  1 .
f ( z )dz 
4 2

Continuing this process gives a sequence { n } having the following

properties:
I |  |
  n
f ( z )dz 
4n
,  n 1   n , |  n |
2n
(n  1,2,) .
(4.10.2)
where |  | denotes the length of the curve  .
It is clear that the diameter d ( n ) of  n goes to 0 as n   . Hence,
by the finite covering theorem in calculus, we see that there exists a point
z0  n (n  1,2,) .
Since f is differentiable at z0 , for every   0 there is a
  0 such that
f ( z )  f ( z0 )
 f ( z0 )   (z  N  ( z0 ,  )) .
z  z0
Thus,
f ( z)  f ( z0 )  f ( z0 )( z  z0 )   z  z0 (z  N ( z0 , )) .
(4.10.3)
Since d (n )  0(n  ) , there is an n such that  n  N ( z0 ,  ) .
By (4.10.3), we have
|  |
f ( z )  f ( z0 )  f ( z0 )( z  z0 )   z  z0  n
 (z   n ) .
2
(4.10.4)
From the fact that   n
0dz  
 n
zdz  0 and inequalities (4.10.4),
we see that

  n
f ( z )dz    n
[ f ( z )  f ( z 0 )  f ( z 0 )( z  z 0 )]dz
|  | |  |
n
 n 
2 2
|  |2
 n .
4
Thus, by (4.10.2) we get that I |  |2  . Since  was arbitrary, we
conclude that I  0 .
Case 2: Let C be a closed polygonal line. Put P  ins(C) , then
we can divided P into a finite number of triangles 1 ,  2 ,,  n , as in
Fig. 4-21.
 It is follows from Case 1 that
n

 f ( z )dz    f ( z )dz =0.

C  k
k 1
This completes the proof.

Fig. 4-21
 A convex domain is a domain D such that for any two points
z1, z2  D , the line segment
[ z1, z2 ] : {(1  t ) z1  tz 2 : t [0,1]}
is contained in D .
Clearly, every neighborhood is convex. And the annulus domain
A( z0 , R1, R2 ) : {z : R1 | z  z0 | R2}
is not convex.
Lemma 4.10.2. Let D be a convex domain and f be analytic in
D , then f has a primitive function in D .
Proof. Fixed a point   D and define
1
F ( z )   f (w)dw  ( z   ) f ((1  t )  tz )dt ,
z 0

for all z  D . We get a function F : D  C . We shall prove that

F ( z)  f ( z), z  D . (4.10.5)
Let z, z0  D with z  z0 . By Lemma 4.10.1 we know that

 z0
f (w)dw  
z0 z
f (w)dw  
zz0
f (w)dw  0 .

That is, F ( z )  F ( z0 )   z0 z
f ( w)dw . Thus,

F ( z )  F ( z0 )  ( z  z0 ) f ( z0 )   [ f (w)  f ( z0 )]dw .
z0 z

Fig. 4-22
Since f is continuous at z0 ,   0,   0 such that
| f (w)  f ( z0 ) |  (w  N ( z0 , )) .
Hence, for all z  N ( z0 ,  ) , we have

F ( z )  F ( z0 )  ( z  z0 ) f ( z0 )   z0 z
[ f ( w)  f ( z0 )]dw

  | z-z0 | .
This shows that
F ( z )  F ( z0 )
F ( z0 )  lim  f ( z0 ) .
z  z0 z  z0
Therefore, the function F is a primitive function of f in D . The
proof is completed.
The proof the Cauchy Inetgral Theorem. Suppose that f is
analytic in R  ins(C) , where C is a simple closed path. Thus, there is
an open set G  R such that f is analytic in it. Since the set C is a
bounded closed set in the plane, we can find a finite number of open disks
D1, D2 ,, Dn  G such that
n
(1) C  D k G;
k 1

(2) C  Dk  Dk 1   (k  1,2,, n) , where Dn 1  D1 ,

as in Fig. 4-23.
Take points zk  C  Dk  Dk 1 (k  1,2,, n) in the colockwise
direction, as in Fig. 4-23. Then these points divide C into n simple
paths C1 , C2 ,, Cn such that the initial and final points of Ck are
zk 1 , zk , respectively, where z0  zn . Since each Dk is a convex
domain and f is analytic in it, Lemma 4.10.2 yields that f has a
primitive function Fk in Dk . It follows from Theorem 4.7.1 that

 Ck
f ( z )dz  
z k 1 z z
f ( z )dz (k  1,2,, n) .
 Fig. 4-23

Hence, from Lemma 4.10.1 we see that

 f ( z )dz   f ( z )dz  k 1  f ( z )dz  k 1 

n n
f ( z )dz  0 ,
C C Ck z k 1 z z
since the curve
n
P   zk 1 z k
k 1
is a polygonal line contained G . This completes the proof.
§4.11. Extended Cauchy Integral Theorem

The Cauchy integral theorem can be generalized in the following two

ways. The first one is the following, called the first extended Cauchy
integral theorem.
Theorem 4.11.1(ECIT-1). If a function f is analytic throughout a
simply connected domain D , then

C
f ( z )dz  0 (4.11.1)

for every closed path C lying in D that intersects itself at most a finite
number of times.
Proof. Case 1: Let C be a simple closed path.
In this case, the function f is analytic at each point interior to and
on C ; and the Cauchy integral theorem ensures that equation (4.11.1)
holds.
Case 2: Let C be a closed path that intersects itself a finite number
of times.
 Fig. 4-24

In this case, it consists of a finite number of simple closed paths,

illustrated in Fig. 4-24, where the simple closed paths Ck (k  1,2,3,4)
make up C .
Since the value of the integral around each C k is zero, according to
the Carchy-Goursat theorem, it follows that
4

 f ( z )dz    f ( z )dz  0 .
C Ck
k 1
This completes the proof.
 Note that in the case where C is a closed path that has an infinite
number of self-intersection points, the formula (4.11.1) is also valid. One
method that can sometimes be used to show that the theorem still applies
in this case is illustrated in Exercise 5 below.
The following corollary follows immediately from Theorem 4.11.1.
Corollary 4.11.1. A function f that is analytic throughout a simply
connected domain D must have a primitive function everywhere in D .
Corollary 4.11.1 tells us that entire functions always possess primitive
functions.
The Cauchy integral theorem can also be extended in a way that
involves integrals along the
boundary of a bounded multiply connected domain that is a bounded
domaim which is not simple connected. The following theorem is such an
extension, called the second extended Cauchy integral theorem.
Theorem 4.11.2(ECIT-2). Suppose that
(i ) C is a simple closed path, described in the counterclockwise
direction;
 (ii ) Ck (k  1, 2,, n) are simple closed paths interior to C , all
described in the clockwise direction, that are disjoint and whose interiors
have no points in common (Fig.4-25);
(iii) D is a connected region, where
n
D  ins (C ) \  ins (Ck ) ;
k 1
(iv) A function f is analytic on
n
D  ins (C ) \  ins (Ck ) .
k 1
Then
n

 f ( z )dz    f ( z )dz  0 . (4.11.2)

C Ck
k 1
 Fig. 4-25

Note that, in equation (4.11.2), the direction of each path of

integration is such that the multiply connected domain lies to the left of
that path.
Proof. To specify, let us give the proof for the case where n  2 .
Choose points
Pi  C  C1  C2 (i  1,2,,12)
as in Fig. 4-25. Since the region D is connected, we can find polygonal
lines, say line segments, P1P2 , P4 P5 , P7 P8  D such that
1  P1P2  P2 P3 P4  P4 P5  P5 P6 P7  P7 P8  P8 P12P1
and
2  P1P11P8  P8 P7  P7 P9 P5  P5 P4  P4 P10P2  P2 P1
are all simple closed paths.
Since the function f is analytic in both ins(1 ) and ins(2 ) , it
follows from the Cauchy integral theorem that
 1
f ( z )dz   f ( z )dz  0  0  0 .
2
On the other hand, because that
 P1 P2
f ( z )dz  
P2 P1
f ( z )dz  0 ,

 P4 P5
f ( z )dz  
P5 P4
f ( z )dz  0 ,

 P7 P8
f ( z )dz  
P8 P7
f ( z )dz  0 ,
 P2 P3 P4
f ( z )dz  
P4 P10 P2
f ( z )dz   f ( z )dz ,
C1

P5 P6 P7
f ( z )dz  
P7 P9 P5
f ( z )dz  
C2
f ( z )dz ,

 P8 P12 P1
f ( z )dz  
P1 P11P8
f ( z )dz   f ( z )dz ,
C
we get that
 1
f ( z )dz   f ( z )dz   f ( z )dz  
2 C1 C2
f ( z )dz   f ( z )dz  0 .
C
This completes the proof.
The following corollary is an especially important consequence of
Theorem 4.11.2.
Corollary 4.11.2. Let C1 and C 2 denote positively oriented
simple closed paths, where C 2 is interior to C1 (Fig. 4-26). If a function
f is analytic in the closed region consisting of those paths and all points
between them, then

C1
f ( z )dz  
C2
f ( z )dz . (4.11.3)
 Proof. We use Theorem 4.11.2 to write

 C1
f ( z )dz  
C2
f ( z )dz  0 ;
and we note that this is just a different form of equation (4.11.3).
Corollary 4.11.2 is known as the principle of deformation of paths
since it tells us that if C1 is continuously deformed onto C 2 , always
passing through points at which f is analytic, then the value of the
integral of f over C1 never changes.

Fig. 4-26 Fig. 4-27

 Example. When C is any positively oriented simple closed path
surrounding the origin, Corollary 4.11.2 can be used to show that
dz
C z  2i .
To do this, we need only construct a positively oriented circle C 0 with
center at the origin and radius so small that C 0 lies entirely inside
C (Fig. 4-27). Since
dz
C0 z  2i
and since 1 / z is analytic everywhere except at z  0 , the desired
result follows.
Note that the radius of C 0 could equally well have been so large
that C lies entirely inside C 0 .