70 2 Analytic Functions and Integration

(c) Establish that the velocity field v = (v1 , v2 ) satisfies

v1 = − 2π √1−x 2 on y = 0+ for − 1 < x < 1, and

 

 2π √x 2 − 1
 for x > 1, y = 0
v2 =

 
− √ for x < −1, y = 0
2π x 2 − 1

6. Consider the transformation (see also Problem 5 above) z = 12 (w + w1 ).

Show that T (x, y) = −Im (z), where  = 1/w satisfies Laplace’s
equation and satisfies the following conditions:


T (x, y = 0+ ) = 1 − x2 for |x| ≤ 1


T (x, y = 0− ) = − 1 − x 2 for |x| ≤ 1
T (x, y = 0) = 0 for |x| ≥ 1

and

 1


for y > 0

y + y2 + 1
T (x = 0, y) =

 1

−
for y < 0
−y + y 2 + 1

2.4 Complex Integration

In this section we consider the evaluation of integrals of complex variable
functions along appropriate curves in the complex plane. We shall see that
some of the analysis bears a similarity to that of functions of real variables.
However, for analytic functions, very important new results can be derived,
namely Cauchy’s Theorem (sometimes called the Cauchy–Goursat Theorem).
Complex integration has wide applicability, and we shall describe some of the
applications in this book.
We begin by considering a complex-valued function f of a real variable t on
a fixed interval, a ≤ t ≤ b:

f (t) = u(t) + iv(t) (2.4.1)

where u(t) and v(t) are real valued. The function f (t) is said to be integrable
 2.4 Complex Integration 71

on the interval [a, b] if the functions u and v are integrable. Then

 b  b  b
f (t) dt = u(t) dt + i v(t) dt (2.4.2)
a a a

The usual rules of integration for real functions apply; in particular, from the
fundamental theorems of calculus, we have for continuous functions f (t)
 t
d
f (τ ) dτ = f (t) (2.4.3a)
dt a

and for f  (t) continuous

 b
f  (t) dt = f (b) − f (a) (2.4.3b)
a

Next we extend the notion of complex integration to integration on a curve

in the complex plane. A curve in the complex plane can be described via the
parameterization

z(t) = x(t) + i y(t), a≤t ≤b (2.4.4)

For each given t in [a, b] there is a set of points (x(t), y(t)) that are the image
points of the interval. The image points z(t) are ordered according to increasing
t. The curve is said to be continuous if x(t) and y(t) are continuous functions
of t. Similarly, it is said to be differentiable if x(t) and y(t) are differentiable.
A curve or arc C is simple (sometimes called a Jordan arc) if it does
not intersect itself, that is, z(t1 ) = z(t2 ) if t1 = t2 for t ∈ [a, b], except that
z(b) = z(a) is allowed; in the latter case we say that C is a simple closed
curve (or Jordan curve). Examples are seen in Figure 2.4.1. Note also that a
“figure 8” is an example of a nonsimple closed curve.

(a) Simple, not closed (b) Not simple, not closed (c) Simple, closed
(Jordan Arc) (Jordan Curve)
Fig. 2.4.1. Examples of curves
72 2 Analytic Functions and Integration

Next we shall discuss evaluation of integrals along curves. When the curve
is closed, our convention shall be to take the positive direction to be the one in
which the interior remains to the left of C. Integrals along a closed curve will be
taken along the positive direction unless otherwise specified. The function f (z)
is said to be continuous on C if f (z(t)) is continuous for a ≤ t ≤ b, and f is
said to be piecewise continuous on [a, b] if [a, b] can be broken up into a finite
number of subintervals in which f (z) is continuous. A smooth arc C is one
for which z  (t) is continuous. A contour is an arc consisting of a finite number
of connected smooth arcs; that is, a contour is a piecewise smooth arc. Thus on
a contour C, z(t) is continuous and z  (t) is piecewise continuous. Hereafter we
shall only consider integrals along such contours unless otherwise specified.
Frequently, a simple closed contour is referred to as a Jordan contour.
The contour integral of a piecewise continuous function on a smooth contour
C is defined to be
  b
f (z) dz = f (z(t))z  (t) dt (2.4.5)
C a

where the right-hand side of Eq. (2.4.5) is obtained via the formal substitution
dz = z  (t) dt. In general, Eq. (2.4.5) depends on f (z) and the contour C. Thus
the integral (2.4.5) is really a line integral in the (x, y) plane and is naturally
related to the study of vector calculus in the plane. As mentioned earlier, the
complex variable z = x + i y can be thought of as a two-dimensional vector.
We remark that values of the above integrals are invariant if we redefine the
parameter t appropriately. Namely, if we make the change of variables t → s by
t = T (s) where T (s) maps the interval A ≤ s ≤ B to interval a ≤ t ≤ b, T (s)
is continuously differentiable, and T  (s) > 0 (needed to ensure that t increases
with s), then only the form the integrals take on is modified, but its value is
invariant. The importance of this remark is that one can evaluate integrals by
the most convenient choice of parameterization. Examples discussed later in
this section will serve to illustrate this point.
The usual properties of integration apply. We have
  
[α f (z) + βg(z)] dz = α f (z) dz + β g(z) dz (2.4.6)
C C C

for constants α and β and piecewise continuous functions f and g. The arc C
traversed the opposite direction, that is, from t = b to t = a, is denoted by −C.
We then have
 
f (z) dz = − f (z) dz (2.4.7)
−C C
 2.4 Complex Integration 73
a
because the left-hand side of Eq. (2.4.7) is equivalent to b f (z(t))z  (t)dt.
Similarly, if C consists of n connected contours with endpoints from z 1 to z 2
for C1 , from z 2 to z 3 for C2 , . . ., from z n to z n+1 for Cn , then we have
  n 
f = f
C j=1 Cj

The fundamental theorem of calculus yields the following result.

Theorem 2.4.1 Suppose F(z) is an analytic function and that f (z) = F  (z) is
continuous in a domain D. Then for a contour C lying in D with endpoints z 1
and z 2

f (z) dz = F(z 2 ) − F(z 1 ) (2.4.8)
C

Proof Using the definition of the integral (2.4.5), the chain rule, and assuming
for simplicity that z  (t) is continuous (otherwise add integrals separately over
smooth arcs) we have
   b
f (z) dz = F  (z) dz = F  (z(t))z  (t) dt
C C a
 b
d
= [F(z(t))] dt
a dt
= F(z(b)) − F(z(a))

= F(z 2 ) − F(z 1 )

As a consequence of Theorem 2.4.1, for closed curves we have

 
f (z) dz = F  (z) dz = 0 (2.4.9)
C C

where C denotes a closed contour C (that is, the endpoints are equal).
If the function f (z) satisfies the hypothesis of Theorem 2.4.1, then for all
contours C lying in D beginning at z 1 and ending at z 2 we have Eq. (2.4.8).
Hence the result demonstrates that the integral is independent of path. Indeed,
Figure 2.4.2 illustrates this fact.  
Referring to Figure 2.4.2, we have C1 f dz = C2 f dz because
  
f dz = f dz − f dz = 0 (2.4.10)
C C1 C2

where the closed curve C = C1 − C2 .

74 2 Analytic Functions and Integration

C1
z2

z1
C2
Fig. 2.4.2. Independent paths forming closed curve

(1,1)

C2
C1 C3

(0,0)
(1,0)
Fig. 2.4.3. Contours C1 , C2 , and C3

The hypothesis in Theorem 2.4.1 requires the existence of F(z) such that
f (z) = F  (z). Later in this chapter we shall show this for a large class of
functions f (z).
Sometimes it is convenient to evaluate the complex integral by reducing it
to two real-line integrals in the x, y plane. In the definition (2.4.5) we use
f (z) = u(x, y) + iv(x, y) and dz = d x + idy to obtain
 
f (z) dz = [(u d x − v dy) + i(v d x + u dy)] (2.4.11)
C C

This can be shown, via parameterization, to be equivalent to

 b
f (z(t))z  (t) dt
a

Later in this chapter we shall use Eq. (2.4.11) in order to derive one form of
Cauchy’s Theorem.
In the following examples we illustrate how line integrals may be calculated
in prototypical cases.


Example 2.4.1 Evaluate C z dz for (a) C = C1 , a contour from z = 0 to z = 1
to z = 1 + i; (b) C = C2 , the line from z = 0 to z = 1 + i; and (c) C = C3 ,
the unit circle |z| = 1 (see Figure. 2.4.3).