Complex Functions

2.1 Functions of a Complex Variable

A function/of the complex variable z is a rule that assigns to each value z in a set
D one and only one complex value w. We write
(1) w=/(z)
and call w the image of z under/. The set D is called the domain of definition off
and the set of all images R = {w = f(z): z G D } is called the range off. Just as z
can be expressed by its real and imaginary parts, z = x + iy, we write w = u + /v,
where w and v are the real and imaginary parts of w, respectively. This gives us the
representation
(2) f(x + iy) = u + iv.
Since w and v depend on x and y, they can be considered to be real functions of the
real variables x and y; that is,
(3) u = u(x, y) and v = v(x, y).
Combining equations (1), (2), and (3), it is customary to write a complex function
/ in the form
(4) f(z) = f(x + iy) = u(x, y) + iv(*. y).
Conversely, if w(>, y) and v(x, y) are two given real-valued functions of the real
variables x and y, then equation (4) can be used to define the complex function/

E X A M P L E 2.1 Write/(z) = z4 in the form/(z) = u(x, y) + iv(x, y).

Solution Using the binomial formula, we obtain

f(z) = (x + iy)4 = x4 + 4jc3/y + 6JC 2 (/» 2 + 4*(/y)3 + (iy)4
= (x4 - 6x2y2 + y4) + /(4jc3y - 4xy3).

38
 2.1 Functions of a Complex Variable 39

E X A M P L E 2 . 2 Express/(z) = z Re(z) + z2 + Im(z) in the form of equations

(2) and (3)/(z) = u(x9 y) + iv(x, y).

Solution Using the elementary properties of complex numbers, it follows

that
f(z) = (x - iy)x + (x2 - y2 + i2xy) + y = (2x2 - y2 + y) + /(xy).

These examples show how to find u(x, y) and v(x, y) when a rule for computing
/ i s given. Conversely, if w(x, y) and v(x, y) are given, then the formulas
z+ z % z - z
x = and yy =
2 2/
can be used to find a formula for/involving the variables z and z.

EXAMPLE 2.3 Express/(z) = Ax2 + /4y2 by a formula involving the variables

z and z.

Solution Calculation reveals that

= z2 + 2zz + z2 ~ /(z2 - 2zz + z2)

= ( 1 - /)z2 + (2 + 2/)zz + ( 1 - i)z2.

It may be convenient to use z = re'9 in the expression of a complex function

/ This gives us the representation
(5) /(z) = /(re'fl) = w(r, 6) + /v(r, 6),
where u and v are to be considered as real functions of the real variables r and 6.
Note that the functions u and v defined by equations (4) and (5) are different, since
equation (4) involves Cartesian coordinates and equation (5) involves polar
coordinates.

EXAMPLE 2.4 Express /(z) = z5 + 4z2 — 6 in the polar coordinate form

w(r, 6) + iv(r, 6).

Solution Using equation (1) of Section 1.5, we obtain

/(z) = ^(cos 56 + i sin 56) + 4r2(cos 26 + i sin 26) - 6
= (r5cos 56 + 4r2cos 2 6 - 6 ) + zVsin 56 + 4r2sin 26).
40 Chapter 2 Complex Functions

EXERCISES FOR SECTION 2.1

1. Let/fe) = f(x + iy) = x + y + /(x3y - y2). Find
( a ) / ( - 1 + 3/) (b) /(3/ - 2 )
2. Let/(z) = z2 + 4zz - 5 Re(z) + Im(z). Find
(a) / ( - 3 + 2/) (b) /(2/ - 1)
3. Find/(1 + /) for the following functions.
1
(a) f(z) = z + z~2 + 5 (b) f(z) =
z2 + 1
4. Find/(2/ - 3) for the following functions.
z + 2 - 3/
(a) f(z) = (z + 3)-¾ - 5/)2 (b) f(z) =
z +4- /
5. Let/(z) = z21 ~ 5z7 + 9z 4 . Use polar coordinates to find
(a)/(-1 +/) _ ( b ) / ( l + /73)
6. Express/(z) = z 2 + (2 - 3/)z in the form u + /v.
7. Express/(z) = : in the form u + /v.
z - 1+ i
8. Express/(z) = z5 + zMn the polar coordinate form w(r, 0) + iv(r, 0).
9. Express/(z) = z5 + z 3 in the polar coordinate form w(r, 0) + /v(r, 0).
10. Let/(z) = fix + iy) — evcos y + /evsin y. Find
(a) /(0) (b) /(1) (c) /(m/4)
(d) /(1 + m/4) (e) /(/271/3) (f) /(2 + in)
11. Let/(z) = / ( x + iy) = {111) \n(x2 + y2) + / arctan(yA). Find
(a) /(1) (b) /(1 + /) (c) / ( 7 3 )
(d) / ( 7 3 + /) (e) /(1 + /73) (0 /(3 + 4/)
12. Let/(z) = r 2 cos 20 + /r2sin 20, where z = r^'fi. Find
(a) /(1) (b) f(le^)
(c)/(72e-/3) (d)/(73^ /6 )
13. Let/(z) = In r + /0, where r = | z |, 0 = Arg z. Find
(a) /(1) (b) /(1 + /)
(c)/(-2) (d) / ( - 7 3 + 0
14. A line that carries a charge of <?/2 coulombs per unit length is perpendicular to the z
plane and passes through the point zo. The electric field intensity E(z) at the point z
varies inversely as the distance from zo and is directed along the line from zo to z. Show
that

E(z) - -
z - zo

where k is some constant. ( In Section 10.11 we will see that the answer is in fact

z - zo
15. Suppose that three positively charged rods carry a charge of qll coulombs per unit length
and pass through the three points 0, 1 - /, and 1 + /. Use the result of Exercise 14 and
show that E(z) = 0 at the points z = (2/3) + /(72/3).
16. Suppose that a positively charged rod carrying a charge of qll coulombs per unit length
passes through the point 0 and that positively charged rods carrying a charge of q
coulombs per unit length pass through the points 2 + / and —2 + /. Use the result of
Exercise 14 and show that E(z) = 0 at the points z = ± y + i \ .
 2.2 Transformations and Linear Mappings 41

2.2 Transformations and Linear Mappings

We now take our first look at the geometric interpretation of a complex function. If
D is the domain of definition of the real-valued functions u(x, v) and v(x, v), then
the system of equations
(1) u = u(x, y) and v = v(x, y)
describes a transformation or mapping from D in the xy plane into the uv plane.
Therefore, the function
(2) w =M = u(x, y) + iv(x, y)
can be considered as a mapping or transformation from the set D in the z plane onto
the range R in the w plane. This is illustrated in Figure 2.1.

w =f(z)
*-
u - u{x, y)
v ~ v{x, y)

FIGURE 2.1 The mapping w = f(z).

If A is a subset of the domain of definition D, then the set B = {f(z): z E A}

is called the image of the set A, and/is said to map A onto B. The image of a single
point is a single point, and the image of the entire domain D is the range R. The
mapping w = f(z) is said to be from A into S if the image of A is contained in S.
The inverse image of a point w is the set of all points z in D such that w = f(z). The
inverse image of a point may be one point, several points, or none at all. If the latter
case occurs, then the point w is not in the range off.
The function/is said to be one-to-one if it maps distinct points z\ ^ Zi onto
distinct points f(z\) ^ f(12)- If H> = f(z) maps the set A one-to-one and onto the set
B, then for each w in B there exists exactly one point z in A such that w = f(z).
Then loosely speaking, we can solve the equation w = f(z) by solving for z as a
function of w. That is, the inverse function z ~ g(w) can be found, and the following
equations hold:
(3) g(f(z)) = z for all z in A and
f(g(w)) = w for all vv in B.
Conversely, if w = f(z) and z = g(w) are functions that map A into B and B
into A, respectively, and equations (3) hold, then w = f(z) maps the set A one-to-
one and onto the set B. The one-to-one property is easy to show, for if we have
f(z\) = / f e ) , then g(f(Z])) = g(f(z2)); and using equation (3), we obtain z\ = z2-
42 Chapter 2 Complex Functions

To show t h a t / i s onto, we must show that each point w in B is the image of some
point in A. If w E B, then z = g(w) lies in A andf(g(w)) = n\ and we conclude that
/ i s a one-to-one mapping from A onto B.
We observe that if / is a one-to-one mapping from D onto /? and if A is a
subset of D, t h e n / i s a one-to-one mapping from A onto its image B. One can also
show that if ^ = f(z) is a one-to-one mapping from A onto 5, and w = g(£) is a one-
to-one mapping from S onto B, then the composition mapping w = g(/(z)) is a one-
to-one mapping from A onto B.
It is useful to find the image B of a specified set A under a given mapping
w = f(z). The set A is usually described with an equation or inequality involving x
and y. A chain of equivalent statements can be constructed that lead to a description
of the set B in terms of an equation or an inequality involving u and v.

E X A M P L E 2 . 5 Show that the function/(^) = iz maps the line y = x + 1 onto

the line v = —u— 1.

Solution We can w r i t e / i n the Cartesian form u + iv = f(z) = i(x + ry)

= — y + /jt, and see that the transformation can be given by the equations u = —y
and v = x. We can substitute these into the equation y = x + 1 to obtain — u = v
+ 1, which can be written as v = — u — 1.

We now turn our attention to the investigation of some elementary mappings.

Let B = a + ib denote a fixed complex number. Then the transformation

(4) w = T(z) = z + B = x + a + i(y + b)

is a one-to-one mapping of the z plane onto the w plane and is called a translation.
This transformation can be visualized as a rigid translation whereby the point z is
displaced through the vector a + ib to its new position w = T(z). The inverse
mapping is given by

(5) z = T~[(w) = w-B = u-a + i(v-b)

and shows that T is a one-to-one mapping from the z plane onto the w plane. The
effect of a translation is pictured in Figure 2.2.

w = T(z)
•z+B

u ~x+a
v =y +b

z ~ x 4- iy

FIGURE 2.2 The translation w = T(z) = z + B = x + a + i(y + b).

 2.2 Transformations and Linear Mappings 43

Let a be a fixed real number. Then the transformation

(6) w = R(z) = zeia = reiQeia = reiiH+a)
is a one-to-one mapping of the z plane onto the w plane and is called a rotation. It
can be visualized as a rigid rotation whereby the point z is rotated about the origin
through an angle a to its new position w = R(z). If we use polar coordinates
w — pe'* in the w plane, then the inverse mapping is given by
(7) z = R~l(w) = we~ia = pe^e~ia = pe^- a ) .
This shows that R is a one-to-one mapping of the z plane onto the w plane. The
effect of rotation is pictured in Figure 2.3.

¢ = 6+ a

FIGURE 2.3 The rotation w = R(z) = rel{^

Let K > 0 be a fixed positive real number. Then the transformation

(8) w = S(z) = Kz = Kx + iKy
is a one-to-one mapping of the z plane onto the w plane and is called a magnification.
If K > 1, it has the effect of stretching the distance between points by the factor K.
If K < 1, then it reduces the distance between points by the factor K. The inverse
transformation is given by
1 1 1
(9) z = S'](w) = —w = — u + i—v
K K K
and shows that S is one-to-one mapping from the z plane onto the w plane. The
effect of magnification is shown in Figure 2.4.
44 Chapter 2 Complex Functions

y
A
Ki + Ki
"•!#•' •/••rj^-Hi^':-^'-'-^-^--'-**-'! 'OAA<

w ~ Kz .... i.v''i'.jjfJijj'[Link] ;>vifi'i>X»y''»»'»'»

u ^Kx
v = Ky /+

-h+-x H >-w

FIGURE 2.4 The magnification w = S(z) = Kz = Kx + iKy.

Let A = Ke'a and £ = a + /£>, where /£ > 0 is a positive real number. Then
the transformation
(10) w = W(z) = Az + B
is a one-to-one mapping of the z plane onto the w plane and is called a linear
transformation. It can be considered as the composition of a rotation, a magnifica-
tion, and a translation. It has the effect of rotating the plane through an angle given
by a = Arg A, followed by a magnification by the factor K = | A |, followed by a
translation by the vector B = a + ib. The inverse mapping is given by

(11) z = W-'(viO = — w - —
A A
and shows that W is a one-to-one mapping from the z plane onto the w plane.

E X A M P L E 2 . 6 Show that the 1 inear transformation w = iz + i maps the right

half plane Re(z) > 1 onto the upper half plane Im(w) > 2.

Solution We can write w = f(z) in Cartesian form u + iv = i(x + ry) + i

— ~~y + *'(* + 1) and see that the transformation can be given by the equations
u = —y and v = x + 1. The substitution x = v - 1 can be used in the inequality
Re(z) = x > 1 to see that the image values must satisfy v - 1 > 1 or v > 2, which
is the upper half plane Im(vv) > 2. The effect of the transformation w = f(z) is a
rotation of the plane through the angle a = n/2 followed by a translation by the
vector B = i and is illustrated in Figure 2.5.
 2.2 Transformations and Linear Mappings 45

FIGURE 2.5 The linear transformation w - f(z) = iz + /.

It is easy to see that translations and rotations preserve angles. Since magni-
fications rescale distance by a factor AT, it follows that triangles are mapped onto
similar triangles, and so angles are preserved. Since a linear transformation can be
considered as a composition of a rotation, a magnification, and a translation, it
follows that linear transformations preserve angles. Consequently, any geometric
object is mapped onto an object that is similar to the original object; hence linear
transformations can be called similarity mappings.

E X A M P L E 2 . 7 Show that the image of the open disk |z + 1 + /| < 1 under

the transformation w = (3 — 4/)z + 6 + 2/ is the open disk | w + I — 3/1 < 5.

Solution The inverse transformation is given by

w — 6 — 2/
Z=
3-4/ '
and this substitution can be used to show that the image points must satisfy the
inequality
I w — 6 — 2/ I
+ 1 + / < 1.
I 3 — 4/ I
Multiplying both sides by | 3 - 4/1 = 5 results in
\w - 6 - 2/ + (1 + /)(3 - 4/) | < 5,
which can be simplified to obtain the inequality
\w + 1 - 3/1 < 5.
Hence the disk with center — 1 — / and radius 1 is mapped one-to-one and onto the
disk with center — 1 + 3/ and radius 5 as pictured in Figure 2.6.
46 Chapter 2 Complex Functions

i t i iv. I I 1 I—*-x
•T
FIGURE 2.6 The mapping w = S(z) = (3 - Ai)z + 6 + 2/.

E X A M P L E 2 . 8 Show that the image of the right half plane Re(z) > 1 under
the linear transformation w = ( — 1 + i)z — 2 + 3/ is the half plane v > u + 7.

Solution The inverse transformation is given by

_ w + 2 - 3/ _ ii + 2 + i(v - 3)
Z
" -1+/ -1 + / '
which can be expressed in the component form
-u + v — 5 —u ~ v + 1
x + iy = + /-

The substitution JC = ( - M + V - 5 ) / 2 can be used in the inequality Re(z) = JC > 1

to see that the image points must satisfy ( —w + v — 5)/2 > 1. This can be simplified
to yield the inequality v > u + 7. The mapping is illustrated in Figure 2.7.

w=f(z)

4-4- f V„» t.,,i, \ \ > x

*-*-«

FIGURE 2.7 The mapping w = f(z) = (-1 + i)z - 2 + 3/.

 2.3 The Mappings w = znand iv= zVn 47

EXERCISES FOR SECTION 2.2

1. Let w = (1 - i)z + 1 - 2i.
(a) Find the image of the half-plane Im(z) > 1.
(b) Sketch the mapping, and indicate the points z\ = - 1 + i, Zi = i, and Z3 = 1 + i
and their images wt, vv2, and w%
2. Let w = (2 + /)z - 3 + 4/. Find the image of the line
JC = r, jy = 1 — 2/ for -oo < t < «>.

3. Let w = (3 + 4i)z - 2 + /.
(a) Find the image of the disk \z - 1 | < 1.
(b) Sketch the mapping, and indicate the points z\ = 0, z2 = 1 ~ h and z? = 2 and their
images.
4. Let w — (3 + 4/')z - 2 + /. Find the image of the circle
JC = 1 + cos /, y = 1 + sin t for —n < t < TC.

5. Let w = (2 + i)z — 2/. Find the triangle onto which the triangle with vertices
Z\ = — 2 + /, z2 — - 2 + 2/, and z^ = 2 + / is mapped.
6. Find the linear transformation w = /(z) that maps the points z\ = 2 and zi = —3/ onto
the points w\ = 1 + i and w2 = 1, respectively.
7. Find the linear transformation w = S(z) that maps the circle \z\ = 1 onto the circle
| w — 3 + 2/1 = 5 and satisfies the condition 5 ( - / ) = 3 + 3/.
8. Find the linear transformation w = f(z) that maps the triangle with vertices —4 + 2/,
- 4 + 7/, and 1 + 2/ onto the triangle with vertices 1, 0, and 1 + i.
9. Let S(z) = Kz, where K > 0 is a positive real constant. Show that the equation
\S(z\) - Sizi) I = K\z\ — Z21 holds, and interpret this result geometrically.
10. Give a proof that the image of a circle under a linear transformation is a circle. Hint:
Let the given circle have the parameterization x = JC() + R cos t, y = y0 + R sin t.
11. Prove that the composition of two linear transformations is a linear transformation.
12. Show that a linear transformation that maps the circle | z — z0 | = ^i onto the circle
| w - w{) | = R2 can be expressed in the form
A(w — W[))R\ - (z — Zo) Ri, where \A\ = 1.

2.3 The Mappings w= zn and w= zVn

The function w — f(z) = z2 can be expressed in polar coordinates by
(1) w=f(z) = z2 = r2ei2\
where r > 0 and — 7t < 8 < n. If polar coordinates, w = pe/(t) are used in the w
plane, then mapping (1) can be given by the system of equations
(2) p = r2 and ¢ = 26.
If we consider the wedge-shaped set A = {re' e : r > 0 and — n/4 < 6 < 7C/4},
then the image of A under the mapping f is the right half plane described by the
inequalities p > 0, — n/2 < (j) < TI/2. Since the argument of the product zz is twice
the argument of z, we say t h a t / d o u b l e s angles at the origin. Points that lie on the
ray r > 0, G = a are mapped onto points that lie on the ray p > 0, <> | = 2a.
48 Chapter 2 Complex Functions

If the domain of definition D for/(z) = z2 is restricted to be the set

(3) D = | relB: r > 0 and -~ < 6 < ~ L

then the image of D under the mapping w = z2 consists of all points in the w plane
(except the point w — 0 and all the points that lie along the negative u axis). The
inverse mapping off is
(4) z = f~\w) = wm = p1/2e'^2, where p > 0 and - n < § < n.
] U2
The function f~ (w) = w in equation (4) is called the principal square root func-
tion and shows that f is one-to-one when its domain is restricted by set (3). The
mappings w = z2 and z = w1/2 are illustrated in Figure 2.8.

FIGURE 2.8 The mappings w = z2 and z = w]/2.

Since /(—z) = (-z)2 = z2, we see that the image of the left half plane
Re(z) < 0 under the mapping w = z2 is the w plane slit along the negative u axis as
indicated in Figure 2.9.

FIGURE 2.9 The mapping w z~.

 2.3 The Mappings w = znaxu\ iv= zVn 49

Other useful properties of the mapping w = z2 can be investigated if we use

the Cartesian form
(5) w = f(z) = z2 = x2 - y2 + i2xy
and the resulting system of equations
(6) u = x2 - y2 and v = 2xy.

E X A M P L E 2 . 9 The transformation w = f(z) = z2 maps vertical and horizontal

lines onto parabolas, and this fact is used to find the image of a rectangle. If a > 0,
then the vertical line x = a is mapped onto the parabola given by the equations
72 _ 2
y and v = lay, which can be solved to yield the single equation

(7) u ^ -
4a2'
If b > 0, then the horizontal line y = b is mapped onto the parabola given by the
equations u = x2 — b2 and v = 2xb, which can be solved to yield the single equation

(8) u = -b2 +
4b2
Since quadrant I is mapped onto quadrants I and II by w = z2, we see that the
rectangle 0 < x < a, 0 < y < b is mapped onto the region bounded by the parabolas
(7) and (8) and the u axis. The four vertices 0, a, a + /Z>, and ib are mapped onto
the four points 0, a2, a2 — b2 + ilab, and — b2, respectively, as indicated in Figure
2.10.

lib , a + i b
24TTTT*-

« • x
0.5

FIGURE 2.10 The transformation w = z2

The mapping w = z]/2 can be expressed in polar form,

— -71/2 r\f2ja/2
(9) w = f(z) = z =
50 Chapter 2 Complex Functions

where the domain of definition D f o r / i s restricted to be r > 0, — n < 6 < n. If

polar coordinates w = pe^ are used in the w plane, then mapping (9) can be rep-
resented by the system

(10) p = rm and ¢) =

From equations (10) we see that the argument of the image is half the argument of
z and that the modulus of the image is the square root of the modulus of z. Points
that lie on the ray r > 0, 0 = a are mapped onto the ray p > 0, ty = a/2. The image
of the z plane (with the point z = 0 deleted) consists of the right half plane
Re(vv) > 0 together with the positive v axis, and the mapping is pictured in Figure
2.11.

„ *-X

* v \ . ::':::>*<'/'•

1¾¾¾^ 0 =$ /2
-7T < 0 <7T l;:-..V:.:.;:^::;;:;:;:X::::::,

FIGURE 2.11 The mapping w = zU2

The mapping w = zu2 can be studied through our knowledge about its inverse
mapping z = w2. If we use the Cartesian formula
(11) z = w2 = u2 - v2 + i'2wv,
then the mapping z = w2 is given by the system of equations
(12) = u2 - and 3? = 2uv.

E X A M P L E 2 . 1 0 The transformation w = f(z) = zm maps vertical and hori-

zontal lines onto a portion of a hyperbola, enabling us to find images of half
planes. Let a > 0. Then system (12) can be used to see that the right half plane
given by Re(z) = x > a is mapped onto the region in the right half plane satisfying
u2 - v2 > a and lies to the right of the hyperbola u2 - v2 = a. If b > 0, then system
(12) can be used to see that the upper half plane Im(z) = y > b is mapped onto the
region in quadrant I satisfying 2wv > b and lies above the hyperbola 2uv = b. The
situation is illustrated in Figure 2.12.
 2.3 The Mappings w= z"and w^ zVn 51

1/2
y
JL
9

-£ 1
y= b 4 z - w 1
x- a

' ' *
FIGURE 2.12 The mapping w = z]

Let « be a positive integer and consider the function w = f(z) = zn, which can
be expressed in the polar coordinate form
(13) w = f(z) = 7? = rnein\ where r > 0 and - 71 < 6 < %.
If polar coordinates w = pe'* are used in the w plane, then mapping (13) can be
given by the system of equations
(14) p = r" and 0 = n%.
We see that the image of the ray r > 0, 6 = a is the ray p > 0, <() = na and that
angles at the origin are increased by the factor n. Since the functions cos «6 and
sin «8 are periodic with period 2n/n, we see that/is in general an n-to-one function;
that is, n points in the z plane are mapped onto each point in the w plane (except
w = 0). If the domain of definition D of / i n mapping (13) is restricted to be
-71 A 71 1
(15) D = re/e: r > o, — < e < — ,
rc n )
then the image of D under the mapping w = f(z) = zn consists of all points in the
w plane (except the origin w = 0), and the inverse function is given by
(16) z = f'~l(w) = wVn = p^V*7", where p > 0 and - n< <>f < n.
The function f~l(w) = wUn is called the principal nth root function and shows that
f is one-to-one when it is restricted to be the domain set (15). The mappings
w = zn and z = wl/n are shown in Figure 2.13.
52 Chapter 2 Complex Functions

tft«w ^ U

FIGURE 2.13 The mappings w = z" and z = wl

EXERCISES FOR SECTION 2.3

1. Show that the image of the horizontal line y = 1 under the mapping w = z2 is the
parabola u = v2/4 - 1.
2. Show that the image of the vertical line x — 2 under the mapping w — z2 is the parabola
2
u = 4 - v /16.
3. Find the image of the rectangle 0 < x < 2, 0 < y < 1 under the mapping w = z2. Sketch
the mapping.
4. Find the image of the triangle with vertices 0, 2, and 2 + 2/ under the mapping w = z2.
Sketch the mapping.
5. Show that the infinite strip 1 < x < 2 is mapped onto the region that lies between the
parabolas u = 1 - v2/4 and u = 4 - v2/16 by the mapping w = z2.
6. For what values of z does (z2)1/2 = z hold if the principal value of the square root is to
be used?
7. Sketch the set of points satisfying the following relations,
(a) Re(z2) > 4 (b) Im(z2) > 6
8. Show that the region in the right half plane that lies to the right of the hyperbola
x2 — y2 = 1 is mapped onto the right half plane Re(w) > 1 by the mapping w = z2.
9. Show that the image of the line x — 4 under the mapping w = z l/2 is the right branch
of the hyperbola u2 — v2 = 4.
10. Find the image of the following sets under the mapping w = z l/2 .
(a) {re1": r > 1 and n/3 < 0 < 7i/2}
(b) {re1": l < r < 9 a n d O < 0 < 2TC/3}
(c) {r^'e: r < 4 and -n < 9 < n/2]
11. Find the image of the right half plane Re(z) > 1 under the mapping w = z2 + 2z + 1.
12. Show that the infinite strip 2 < y < 6 is mapped onto the region in the first quadrant
that lies between the hyperbolas uv = 1 and uv — 3 by the mapping w = z l/2 .
13. Find the image of the region in the first quadrant that lies between the hyperbolas
xv = y and xy = 4 under the mapping w = z2.
14. Show that the region in the z plane that lies to the right of the parabola x = 4 — y2/\6
is mapped onto the right half plane Re(w) > 2 by the mapping w = z l/2 . Hint: Use the
inverse mapping z = w2.
 2.4 Limits and Continuity 53

15. Find the image of the following sets under the mapping w = z3.
(a) {re*; 1 < r < 2 and - nIA < 6 < K / 3 }
(b) {re'»: r > 3 and 2TC/3 < 6 < 3n/4]
16. Find the image of the sector r > 2, 7i/4 < 0 < rc/3 under the following mappings,
(a) w = z* (b) w = z 4 (c) w = z6
17. Find the image of the sector r > 0 , —7C<8< 2TC/3 under the following mappings,
(a) w = zm (b) w = zm (c) w = zm
18. Use your knowledge about the complex square root function and explain the fallacy in
the following statement: 1 = 7 ( - 1 ) ( - 1 ) = J{-\)J{-\) = (/)(/) = - 1 .

2.4 Limits and Continuity

Let u = w(x, y) be a real-valued function of the two real variables x and y. We say
that u has the limit u 0 as (JC, y) approaches (JCO, yo) provided that the value of u(x, y)
gets close to the value u0 as (x, y) gets close to (JC0, yo). We write
(1) lim w(x, y) = w0.
(jf,V)-»(-V0,V())

That is, u has the limit w0 as (x, y) approaches (x 0 , yo) if and only if | W(JC, y) — w0 |
can be made arbitrarily small by making both \x — Xo | and \y — y 0 | small. This
is like the definition of limit for functions of one variable, except that there are two
variables instead of one. Since (JC, y) is a point in the xy plane, and the distance
between (x, y) and (JC0, yo) is J(x — XQ)2 + (y — y 0 ) 2 , we can give a precise definition
of limit as follows. To each number e > 0, there corresponds a number 8 > 0 such
that

(2) | u(x, y) - w01 < £, whenever 0 < J(x - x0)2 + (y - y 0 ) 2 < 8.

E X A M P L E 2.11 If u(x, y) = Jt3/C*2 + y2), then

(3) lim w(x, y) = 0.
tr,v)->(0,0)

Solution If x = r cos 6 and y = r sin 0, then

r 3 cos 3 6 .
u(x, y) = 2Q . 2Q = r cos 3 6.
r 2 cos 2 B + r 2 sin 2 6
Since J(x - 0) 2 + (y ~ 0) 2 = r, we see that
| u(x, y) - 01 = r | cos 3 8 | < e, whenever 0 < Jx1 + y 2 = r < £.
Hence for any e > 0, inequality (2) is satisfied for 8 = e; that is, u(x, y) has the
limit wo = 0 as (JC, v) approaches (0, 0).

The value uQ of the limit must not depend on how (x, y) approaches (x 0 , yo).
So it follows that u(x, y) must approach the value w0 when (x, y) approaches (x0, yo)
54 Chapter 2 Complex Functions

along any curve that ends at the point (JC0, yo). Conversely, if we can find two curves
Cx and C2 that end at (x0, Vo) along which u(xy y) approaches the two distinct values
u\ and W2, respectively, then u{x, y) does not have a limit as (x, y) approaches
C*o, Jo)-

E X A M P L E 2 . 1 2 The function u{x, y) = xy/(x2 + y2) does not have a limit as

O, y) approaches (0, 0). If we let (x, y) approach (0, 0) along the x axis, then
W(Q)
r
hm , 0)
u(x, m = r
hm — — = n0.
(.v,0)-K0,0) (.r,0)-»(0,0) X1 + 0^

But if we let (x, 3;) approach (0, 0) along the line y = x, then
WW 1
hm «(x, x) = hm —
Z 1
~ —.
(.vjr)-»(0,0) (.v.v)-»(0,0) X + X 2

Since the two values are different, the value of the limit is dependent on how (x, y)
approaches (0, 0). We conclude that w(x, y) does not have a limit as (x, y) approaches
(0, 0).

Let/(z) be a complex function of the complex variable z that is defined for all
values of z in some neighborhood of zo, except perhaps at the point z0- We say that
/ has the limit w0 as z approaches zo, provided that the value f(z) gets close to the
value wo as z gets close to z0; and we write
(4) lim/U) = w0.

Since the distance between the points z and zo can be expressed by | z — Zo |, we

can give a precise definition of limit (4): For each positive number e > 0, there
exists a 8 > 0 such that
(5) \f(z) - wo I < e, whenever 0 < | z - Zo \ < 8.
Geometrically, this says that for each e-neighborhood | w — wo | < e of the point
w0 there is a deleted 8-neighborhood 0 < | z — Zo | < 8 of Zo such that the image
of each point in the 8-neighborhood, except perhaps z0> li e s in the e-neighborhood
of WQ. The image of the 8-neighborhood does not have to fill up the entire e-neigh-
borhood; but if z approaches z0 along a curve that ends at zo, then w = f(z) ap-
proaches w0. The situation is illustrated in Figure 2.14.
If we consider w = f(z) as a mapping from the z plane into the w plane and
think about the previous geometric interpretation of a limit, then we are led to
conclude that the limit of a function/should be determined by the limits of its real
and imaginary parts u and v. This will also give us a tool for computing limits.
 2.4 Limits and Continuity 55

w=f(z)

FIGURE 2.14 The limit/(z) -» w0 as z -> z0.

Theorem 2.1 Let /(z) = u(x, y) + /v(X y) be a complex function that

is defined in some neighborhood of zo, except perhaps at Zo = x0 + /y0. Then
(6) lim/(z) = w0 = u0 + /v0

//"ana1 o«/y //*

(7) lim u(x, y) = w0 ana1 Urn v(x, y) ~ v0.
(.v,y)->U0,y0) (.v,y)^U 0 ,y 0 )

Proof Let us first assume that statement (6) is true, and show that statement (7)
is true. According to the definition of limit, for each £ > 0, there corresponds a
8 > 0 such that
\f(z) - wo I < £, whenever 0 < | z - z01 < 5.
Since/(z) - w0 = w(x, y) - u0 + /(v(x, y) — v0), we can use equations (2) of Section
1.3 to conclude that
| u(x, y) ~ M0 | < \f(z) ~ wo | and | v(x, y) - v0 | < \f(z) - w0 |.
It now follows that | u(x, y) - u01 < e and | v(x, y) - v01 < e whenever
0 < | z — Zo | < 8 so that statement (7) is true.
Conversely, now let us assume that statement (7) is true. Then for each
e > 0, there exists 8j > 0 and 82 > 0 so that

I u(x, y) - w0 I < — , whenever 0 < | z - Zo \ < 81 and

I v(x, y) - v0 I < — , whenever 0 < | z - Zo | < 82.

Let 8 be chosen to be the minimum of the two values 81 and 82. Then we can use
the triangle inequality
\f(z) - w0 I < I u(x9 y) - wo I + I v(x, y) - v0 |
56 Chapter 2 Complex Functions

to conclude that

\f(z) ~ w0 I < — + ~ = £, whenever 0 < | z - z0 | < 8.

Hence the truth of statement (7) implies the truth of statement (6), and the proof of
the theorem is complete.

For example, lim (z2 - 2z + 1 ) = - 1 . To show this result, we let

z-> 1 f i

f(z) = z - 2z + 1 = x2 - y2 - 2x + 1 + i(2xy - 2y).

2

Computing the limits for u and v, we obtain

lim u(x, )0 = 1 - 1 - 2 + 1 = -1 and

(jr,v)-»(l,l)

lim V(JC, y) = 2 - 2 = 0.
(*,>•)->< 1,1)

So Theorem 2.1 implies that lim f(z) = - 1 .

c-> 1 i /

Limits of complex functions are formally the same as in the case of real func-
tions, and the sum, difference, product, and quotient of functions have limits given
by the sum, difference, product, and quotient of the respective limits. We state this
result as a theorem and leave the proof as an exercise.

Theorem 2.2 Let lim/(z) = A and lim g(z) = B. Then

(8) lim [f(z) ± g(z)] =A±B.

(9) lim/( 2 )*(z) = AB.

(10) lim — = — , where B ^ 0.

^ , g(z) B
Let u(x, y) be a real-valued function of the two real variables x and y. We say
that u is continuous at the point (JC0, v0) if the following three conditions are satisfied:
(11) lim u(x, y) exists.
u,v)->(Ao-yo>

(12) u(x0, y0) exists.

(13) lim u(x, y) = u(x0, y0).

(jf,v)->(jr 0 <yo)

Condition (13) actually contains conditions (11) and (12), since the existence of the
quantity on each side of the equation there is implicitly understood to exist. For
example, if u(x, y) = x3/(x2 + y2) when (JC, y) ¥- (0, 0) and if w(0, 0) = 0, then we
have already seen that u(x, y) —> 0 as (x, y) —> (0, 0) so that conditions (11), (12),
and (13) are satisfied. Hence u(x, y) is continuous at (0, 0).
 2.4 Limits and Continuity 57

Let/(z) be a complex function of the complex variable z that is defined for all
values of z in some neighborhood of z0- We say that f is continuous at z0 if the
following three conditions are satisfied:
(14) limf(z) exists.
Z->ZQ

(15) /(¾)) exists.

(16) lim/(2)=/(¾).

A complex function/is continuous if and only if its real and imaginary parts
u and v are continuous, and the proof of this fact is an immediate consequence of
Theorem 2.1. Continuity of complex functions is formally the same as in the case
of real functions, and the sum, difference, and product of continuous functions are
continuous; their quotient is continuous at points where the denominator is not zero.
These results are summarized by the following theorems, and the proofs are left as
exercises.

Theorem 2.3 Letf(z) = u(x, v) + iv(x, y) be defined in some neighborhood

ofzo. Then/is continuous at zo — XQ + iyo if and only ifu and v are continuous
at (*o, y0).

Theorem 2.4 Suppose that f and g are continuous at the point zo. Then the
following functions are continuous at Zo'.
(17) Their sum f(z) + g(z).
(18) Their difference f(z) - g(z).
(19) Their product f(z)g(z).
f(z)
(20) Their quotient provided that g(zo) 9^ 0.
g(z)
(21) Their composition f(g(z)) provided that f{z) is continuous in a neighborhood
of the point g(zo).

EXAMPLE 2.13 Show that the polynomial function given by

w = P(z) = a0 + a\z + a2z2 + • • • + anzn
is continuous at each point z0 in the complex plane.

Solution Observe that if a0 is the constant function, then lim ^ a0 = a0\

and if a\ # 0, then we can use definition (5) with f(z) = axz and the choice
8 = e/|ai | to prove that lim,.^ axz = aiZo- Then using property (9) and mathe-
matical induction, we obtain
(22) lim akzk = akz^ for k = 0, 1, 2, . . . , n.
58 Chapter 2 Complex Functions

Property (8) can be extended to a finite sum of terms, and we can use the result of
equation (22) to obtain

(23) lim P(z) = lim ( J ) akzk) = j ? akz* = P(zo).

Z->ZQ Z->ZQ \k=Q / k=0

Since conditions (14), (15), and (16) are satisfied, we can conclude that P is con-
tinuous at zo-

One technique for computing limits is the use of statement (20). Let P and Q
be polynomials. If Q(zo) ^ 0, then

r P(z) P(ZQ)
lim = .
z-*Zo Q(Z) (2(¾)
Another technique involves factoring polynomials. If both P(zo) = 0 and
Q(zo) = 0, then P and Q can be factored as P(z) = (z - Zo)P\(z) and
Q(z) = (z - Zo)Qi(z). If Gi(zo) ^ 0, then the limit is given by

r P(z) y (z - Zo)Pj(z) Pdzo)

lim = lim = .
z-*zo Q(Z) Z-+* (Z ~ Zo)Ql(z) (2,(¾))

EXAMPLE 2.14 Show that lim , Z" " 2i = 1 - / .

z->\+n2 - 2z + 2

Solution Here P and Q can be factored in the form

P(z) = (z - 1 - i)(z + 1 + /) and Q(z) = (z - 1 - i)(z - 1 + /)

so that the limit is obtained by the calculation

r z2 - 2/ v (z - 1 - /)(z + 1 + 0
lim — = lim
z-i + ; z2 - 2Z + 2 z - i + i (z - I - 0(z ~ 1 + 0
= hm = 1 - 1 .
z->\+i Z — 1 + I

EXERCISES FOR SECTION 2.4

1. Find lim (z2 ~ 4z + 2 + 5/). 2. Find lim Z + 4z + 2
.
^2 +/ c^/ Z+ 1
z4 - 1 z2 + z - 2 + /
3. Find lim r • 4. Find lim — 2 .
z-+i Z ~ I z^\+i Z ~ 2z + 1
z2 + z - 1 - 3i. X'
5. Find lim — — by factoring. 6. Show that lim — = 0.
 2.4 Limits and Continuity 59

7. State why lim (e'cos y -f ix2y) ~ eX(,cos y0 + ixly().

8. State why lim [ln(*2 + y2) + /y] = lnC*2, + y(2,) + /vo provided that | zo | ^ 0.

Iz P
9. Show that lim = 0.
-*<> z
10. Let/(z) = -T-T; 7—^ .
I z |- x~ 4- y-
(a) Find lim/(z) as z —> 0 along the line v = x.
.--0
(b) Find lim/(z) as z —» 0 along the line v = 2x.
.-•• » 0 '

(c) Find lim/(z) as z --> 0 along the parabola y = x2.

•»o'

(d) What can you conclude about the limit of/(z) as z —> 0?
11. Let/(z) = zVz. Show that/(z) does not have a limit as z —» 0.
12. Does u(x, y) = U3 - 3.ry2)/(x2 4- y2) have a limit as (x, y) -^ (0, 0)?
13. Let/(z) - z1/2 = rl/2[cos(6/2) + / sin(0/2)], where r > 0 and -TC < 0 < TC. Use the
polar form of z and show that
(a) /(z) -> i as z —> — 1 along the upper semicircle r = 1, 0 < 0 < TC.
(b) f(z) —> — / as z —» — 1 along the lower semicircle r = 1, — K < 0 < 0.
14. Does lim Arg z exist? Why? Hint: Use polar coordinates and let z approach - 4 from
the upper and lower half planes.
15. Determine where the following functions are continuous.
z2 + 6z + 5
4 2
(a) z - 9z + iz ~ 2 (b) - y i - y (c) z2 + 3z + 2
4 X• + iy
z + 1 x + i\
(d) , , . , 0 (e) =- (f)
z- + 2z + 2 .v - 1 1 zj - 1
16. Let/(z) = [z Re(z)]/1 z | when z / 0 , and let /(0) = 0. Show that/(z) is continuous for
all values of z.
17. Let/(z) = xey 4- (v2^ \ Show that /(z) is continuous for all values of z.
18. Let/(z) = U2 4- /y2)/ | z | 2 when z ^ 0 , and let/(0) = 1. Show that/(z) is not continuous
at zo = 0.
19. Let/(z) = Re(z)/|z| when z ¥" 0, and let/(0) = 1. Is/(z) continuous at the origin?
20. Let/(z) = [Re(z)] 2 /|z| when z 7^ 0, and let/(0) = 1. Is/(z) continuous at the origin?
21. Let/(z) - z l/2 = r1/2[cos(6/2) 4- i sin(0/2)], where r > 0 and -TC < 0 < TC. Show that
/(z) is discontinuous at each point along the negative x axis.
22. Let/(z) = In I z I + / Arg z, where —TC < Arg z ^ TC. Show that /(z) is discontinuous at
Zo = 0 and at each point along the negative x axis.
23. Let A and B be complex constants. Use Theorem 2.1 to prove that
lim (Az + B) = Azt) + #•
24. Let Az = z - z<>. Show that lim f<z) = vv() if and only if lim /(z () + Az) = vv0.
25. Let |g(z)| < M and lim/(z) = 0. Show that lim/(z)#(z) = 0.

26. Establish identity (8). 27. Establish identity (9). 28. Establish identity (10).
29. Let/(z) be continuous for all values of z.
(a) Show that g{z) = fiz) is continuous for all z.
(b) Show that h(z) = fiz) is continuous for all z.
30. Establish the results of (17) and (18). 31. Establish the result (19).
32. Establish the result (20). 33. Establish the result (21).
60 Chapter 2 Complex Functions

2.5 Branches of Functions

In Section 2.3 we defined the principal square root function and investigated some
of its properties. We left some unanswered questions concerning the choices of
square roots. We now look into this problem because it is similar to situations
involving other elementary functions.
In our definition of a function in Section 2.1 we specified that each value of
the independent variable in the domain is mapped onto one and only one value of
the dependent variable. As a result, one often talks about a single-valued function,
which emphasizes the only one part of the definition and allows us to distinguish
such functions from multiple-valued functions, which we now introduce.
Let w = f(z) denote a function whose domain is the set D and whose range is
the set R. If w is a value in the range, then there is an associated inverse function
z = g(w) that assigns to each value w the value (or values) of z in D for which the
equation/(z) = w holds true. But unless/takes on the value w at most once in Z),
then the inverse function g is necessarily many valued, and we say that g is a
multivalued function. For example, the inverse of the function w = f(z) = z2 is the
square root function z = g(w) = wU2. We see that for each value z other than
z = 0, the two points z and — z are mapped onto the same point w = f(z)\ hence g
is in general a two-valued function.
The study of limits, continuity, and derivatives loses all meaning if an arbitrary
or ambiguous assignment of function values is made. For this reason we did not
allow multivalued functions to be considered when we defined these concepts. When
working with inverse functions, it is necessary to carefully specify one of the many
possible inverse values when constructing an inverse function. The idea is the same
as determining implicit functions in calculus. If the values of a function f are de-
termined by an equation that they satisfy rather than by an explicit formula, then
we say that the function is defined implicitly or t h a t / i s an implicit function. In the
theory of complex variables we study a similar concept.
Let w = f(z) be a multiple-valued function. A branch o f / i s any single-valued
function / 0 that is continuous in some domain and, at each point z in the domain,
assigns one of the values o f / ( z ) .

E X A M P L E 2 - 1 5 Let us consider some branches of the two-valued square root

function/(z) = z l / 2 . We define the principal square root function as
ft ft
(1) f{z) = r1/2cos — + /r1/2sin — = r1/2e'e/2,

where we require that r > 0 and — n < 6 < n. The function/ is a branch of/ We
can find other branches of the square root function. For example, let

(2) f2(z) = r l/2 cos 9 + 2n

+ /r1/2sin d
+ 2n = , 1 / 2 ^ + 2 ^

where r > 0 and —7t < 0 < 71.

 2.5 Branches of Functions 61

e + 27i e e + in e
If we use the identities cos — = —cos — and sin = —sin —
2 2 2
then we see that

fi(z) — — r I/2 cos e e = _rl/Ve/2 = _/l(z)>

2 2
s o / i a n d / 2 can be thought of as " p l u s " and " m i n u s " square root functions.

The negative real axis is called a branch cut for the functions f\ and/ 2 - It is char-
acterized by the fact that each point on the branch cut is a point of discontinuity for
both functions/1 and/ 2 -

E X A M P L E 2 . 1 6 To show that the function f\ is discontinuous along the neg-

ative real axis, let zo = r{)e±iK denote a negative real number. Now we compute the
limit of/i(z) as z approaches zo through the upper half plane {z: Im(z) > 0} and the
limit off\(z) as z approaches z0 through the lower half plane {z: Im(z) < 0 } . In polar
coordinates these limits are given by
ft ft 1/2
lim /i(re' H ) = lim r l/2 l cos \- i sin — ir and
(r,e )-»(/•(). JC) (r,H)-»(r 0 ,re) \ 2 2

0
lim /i(re' e ) = lim cos h i sin — _,vi/-
(/•,e)->(r0.-7i) (r,H)-»(r0,-7t)

Since the two limits are distinct, the function f\ is discontinuous at zo- Likewise, f2
is discontinuous at ZQ. The mappings w = f\{z) and w = f2(z) and the branch cut are
illustrated in Figure 2.15.

w=f,(z) *•">•*%*

[Link]' ^ X

m
I

i
w =/ 2 (z) -.^^4'
t;.;.;:;.l:,.l.M,.l;;Ml.!;M;.!Y,Y!!, » ^

FIGURE 2.15 The branches/, and/ 2 of/U)

62 Chapter 2 Complex Functions

Other branches of the square root function can be constructed by specifying

that an argument of z given by 6 = arg z is to lie in the interval a < 0 < a + 271.
Then the branch fa is given by
A ft
1/2
(3) fa(z) = r cos — + /r l/2 sin — , where r > 0 and a < G < a + 2TC.

The branch cut f o r / a is the ray r > 0, 6 = a, which includes the origin. The point
z = o, common to all branch cuts for the multivalued function, is called a branch
point. The mapping vv = fa(z) and its branch cut are illustrated in Figure 2.16.

V
i1
'A.
I
OL
'&&?/<••••
p
•}:Mr.v
™=fjz) •.
b V
J* 1_^ u

FIGURE 2.16 The branch/, of/(z) = z]n.

The Riemann Surface for w = z1/2

A method for visualizing a multivalued function is provided by using a Riemann
surface. These representations were introduced by G. F. B Riemann (1826-1866) in
1851. The idea is ingenious, a geometric construction that permits surfaces to be
the domain or range of a multivalued function.
Consider vv = f(z) = z 1/2 , which has two values for any given z (except, of
course, for z = 0). Each function fx(z) and/ 2 (z), given in Example 2.15 is single-
valued on the domain formed by cutting the z plane along the negative x axis. Let
Dx and D2 be the domain of fY(z) and f2(z), respectively. The range set /i(z) is
the set Hx consisting of right half plane Hx plus the positive v axis, and the range
set f2(z) is the set H2 consisting of left half plane Hi plus the negative v axis. The
sets Hi and H2 are "glued together" along the poitive v axis and the negative
v axis to form the vv plane with the origin deleted.
Stack D\ to D2 directly above each other. The edge of Dx in the upper half
plane is joined to the edge of D2 in the lower half plane, and the edge of D\ in the
lower half plane is joined to the edge of D2 in the upper half plane. When these
domains are " g l u e d ' ' together in this manner they form R, which is a Riemann
surface domain for the mapping w = f(z) - zU2. The portion of Du D 2 , and R that
satisfy \z\ < 1 are shown in Figure 2.17.
 2.5 Branches of Functions 63

(a) A portion of D\ and its image under w = z]/2.

(b) A portion of D2 and its image under w = z1

(c) A portion of R and its image under w = z l/2 .

FIGURE 2.17 Formation of the Riemann surface for w - z]/2.

EXERCISES FOR SECTION 2.5

1. Let/i(z) and/ 2 (z) be the two branches of the square root function given by equations
(1) and (2), respectively. Use the polar coordinate formulas in Section 2.3 to
(a) Find the image of quadrant II, x < 0 and v > 0, under the mapping w = f\(z).
(b) Find the image of quadrant II, x < 0 and y > 0, under the mapping w = f2(z).
64 Chapter 2 Complex Functions

(c) Find the image of the right half plane Re(z) > 0 under the mapping w = f\(z).
(d) Find the image of the right half plane Re(z) > 0 under the mapping w = fi(z).
2. Let a = 0 in equation (3), and find the range of the function w = f(z).
3. Let a = 2n in equation (3), and find the range of the function w = f(z).
4. Find a branch of the square root function that is continuous along the negative x axis.
5. Let/Hz) = r l/3 cos(8/3) + ir ,/3 sin(6/3), where r > 0 and -TC < 6 < % denote the principal
cube root function.
(a) Show that/j is a branch of the multivalued cube root function/(z) = z1/3.
(b) What is the range of/i?
(c) Where is/i continuous?
6. Let/ 2 (z) = r l/3 cos[(8 + 2n)/3] + *>1/3sin[(8 + 2rc)/3], where r > 0 and - 7 i < 6 < n.
(a) Show that/ 2 is a branch of the multivalued cube root function/(z) = z1/3.
(b) What is the range of/ 2 ?
(c) Where is/ 2 continuous?
(d) What is the branch point associated with/?
7. Find a branch of the multivalued cube root function that is different from those in
Exercises 5 and 6. State the domain and range of the branch you find.
8. Let/(z) = z]/n denote the multivalued nth root function, where n is a positive integer.
(a) Show t h a t / i s in general an n-valued function.
(b) Write down the principal nth root function.
(c) Write down a branch of the multivalued nth root function that is different from the
one in part (b).
9. Describe a Riemann surface for the domain of definition of the multivalued function
w = f(z) = z ,/3 .
10. Describe a Riemann surface for the domain of definition of the multivalued function
w = / ( z ) = zm.
11. Discuss how Riemann surfaces should be used for both the domain of definition and the
range to help describe the behavior of the multivalued function w = /(z) = z2/3.
12. Show that the principal branch of the argument Arg z is discontinuous at 0 and all points
along the negative real axis.

2.6 The Reciprocal Transformation w= Mz

(Prerequisite for Section 9.2)
The mapping w = Mz is called the reciprocal transformation and maps the z plane
one-to-one and onto the w plane except for the point z = 0, which has no image,
and the point w — 0, which has no preimage or inverse image. Since zz = \ z | 2 , we
can express the reciprocal transformation as a composition:
_ z
(1) w —Z and Z = -—r-.
\z\2
The transformation Z = zl \ z | 2 is called the inversion mapping with respect to the
unit circle \z\ = 1. It has the property that a nonzero point z is mapped onto the
point Z such that

(2) |Z| |z| = 1 and arg Z = arg z.

Hence it maps points inside the circle \z\ = 1 onto points outside the circle
| Z | = 1, and conversely. Any point of unit modulus is mapped onto itself. The
inversion mapping is illustrated in Figure 2.18.
 2.6 The Reciprocal Transformation w= Mz 65

FIGURE 2.18 The inversion mapping.

The geometric description of the reciprocal transformation is now evident

from the composition given in expression (1). It is an inversion followed by a re-
flection through the x axis. If we use the polar coordinate form

(3) w = pe1^ = —e~iQ, where z = re'8,

r
then we see that the ray r > 0, 0 = a is mapped one-to-one and onto the ray
p > 0, <>| = - a . Also, points that lie inside the circle | z | = 1 are mapped onto
points that lie outside the circle \w\ — 1, and vice versa. The situation is illustrated
in Figure 2.19.

y v

FIGURE 2.19 The reciprocal transformation w = 1/z.

It is convenient to extend the system of complex numbers by joining to it an

4
'ideal" point denoted by <» and called the point at infinity. This new set is called
the extended complex plane. The point °o has the property that
(4) lim zn = °° if and only if lim \zn\ = °°.
66 Chapter 2 Complex Functions

An £-neighborhood of the point at infinity is the set {z: \z\ > 1/e}- The usual way
to visualize the point at infinity is accomplished by using the stereographic projec-
tion and is attributed to Riemann. Let 17 be a sphere of diameter 1 that is centered
at the (0, 0, y) in the three-dimensional space where coordinates are denoted by the
triple of real numbers (JC, y, £). Here the complex number z = x + iy will be asso-
ciated with the point Qc, y, 0).
The point M = (0, 0, 1) on fl is called the north pole of O. Let z be a complex
number, and consider the line segment L in three-dimensional space that joins z to
the north pole K. Then L intersects Q in exactly one point i£. The correspondence
z <-» ££ is called the stereographic projection of the complex z plane onto the
Riemann sphere £1. A point z = x + iy of unit modulus will correspond to
z nas mocUims
-S6 = ("T > "T > ~T )• ^ grater than 1, then !£ will lie in the upper
hemisphere where £ > y . If z has modulus less than 1, then X will lie in the lower
hemisphere where £ < y . The complex number z = 0 corresponds to the south pole
J = (0, 0, 0). It is easy to visualize that z ~^ °° if and only if i£ —> K. Hence M
corresponds to the "ideal" point at infinity. The situation is shown in Figure 2.20.

FIGURE 2.20 The Riemann sphere.

Let us reconsider the mapping w = 1/z. Let us assign the images w = <*> and
w = 0 to the points z = 0 and z = °°, respectively. The reciprocal transformation
can now be written as

( l/z
0
co
when z ^ 0, z ^ °°
when z = °°
when z = 0.
It is easy to see that the transformation w = f(z) is a one-to-one mapping of the
extended complex z plane onto the extended complex w plane. Using property (4)
of the point at infinity, it is easy to show t h a t / i s a continuous mapping from the
extended z plane onto the extended w plane. The details are left for the reader.
 2.6 The Reciprocal Transformation w= Vz 67

E X A M P L E 2 . 1 7 Show that the image of the right half plane Re(z) > y , under
the mapping w = 1/z, is the disk | w - 1 | < 1.

Solution The inverse mapping z = 1/w can be written as

1 u — iv
(6) x + iy z = U2 + V2 '

Equating the real and imaginary parts in equation (6), we obtain the equations
u . —v
(7) and y
u2 + v2 u2 + v2
The requirement that x > y forces the image values to satisfy the inequality
u Jt_
(8)
u2 + v2 2 '
It is easy to manipulate inequality (8) to obtain
(9) u2 - 2u + 1 + v2 < 1,
which is an inequality that determines the set of points in the w plane that lie inside
the circle with center H>0 = 1 and radius 1. Since the reciprocal transformation is
one-to-one, preimages of the points in the disk | w — 1 | < 1 will lie in the right
half plane Re(z) > y . The mapping is shown in Figure 2.21.

FIGURE 2.21 The image of Re(z) > 1/2 under the mapping w = 1/z.

EXAMPLE 2,18 Find the image of the portion of the right half plane
Re(z) > y that lies inside the circle | z ~ y | < 1 under the transformation w = 1/z.
68 Chapter 2 Complex Functions

Solution Using the result of Example 2.17, we need only find the image
of the disk I z — y I < 1 and intersect it with the disk \w — 1 I < 1. To start with,
we can express the disk | z < 1 by the inequality
(10) x2 + y 2 - j c < | .
We can use the identities (7) to show that the image values of points satisfying
inequality (10) must satisfy the inequality
1 u 3_
(ID
u2 + v2 u2 + v2 < 4 "
Inequality (11) can now be manipulated to yield
(T)2 < (a + I) 2 + v\
which is an inequality that determines the set of points in the w plane that lie exterior
to the circle | w + y | = Therefore, the image is the crescent-shaped region
illustrated in Figure 2.22.

FIGURE 2.22 The mapping w = \lz that is discussed in Example 2.18.

To study images of "generalized circles," let us consider the equation

(12) A(x2 + y2) + Bx + Cy + D = 0
where A, B, C, and D are real numbers. Then equation (12) represents either a circle
or a line, depending on whether A ^ 0 or A = 0, respectively. If we use polar
coordinates, then equation (12) has the form
(13) Ar2 + r{B cos 6 + C sin 6) + D = 0.
Using the polar coordinate form of the reciprocal transformation given in equation
(3), we find that the image of the curve in equation (13) can be expressed by the
equation
(14) A + p(B cos $ - C sin ¢) + Dp2 = 0,
 2.6 The Reciprocal Transformation iv= Mz 69

which represents either a circle or a line, depending on whether D ^ 0 or D = 0,

respectively. Therefore, we have shown that the reciprocal transformation w = l/z
carries the class of lines and circles onto itself.

E X A M P L E 2 . 1 9 Find the images of the vertical lines x = a and the horizontal

lines y = b under the mapping w = l/z.

Solution The image of the line x = 0 is the line u = 0; that is, the y axis
is mapped onto the v axis. Similarly, the x axis is mapped onto the u axis.
If a ¥" 0, then using equations (7), we see that the vertical line x = a is mapped
onto the circle

(15)
u2 4- v2
It is easy to manipulate equation (15) to obtain

ul
1
u H
1 , ,
+ vl = I u
±Y + ,= - (± v
a Aa2 2a/ \2aj
which is the equation of a circle in the w plane with center w0 = 1 / ( 2 a ) and radius
I 1/(2«) | .
Similarly, the horizontal line y = bis mapped onto the circle

u2 + v2 + —v + — 2 = u2 + ( v + ~ 2
b Ab \ 2b, 2b)'
which has center w0 = -i/(2b) and radius | 1/(2Z?)|. The images of several lines are
shown in Figure 2.23.

I !

•b = \ j
w = -
z
b = Vi

- • x

b = -½
£ = -1

FIGURE 2.23 The images of horizontal and vertical lines under the reciprocal
transformation.
70 Chapter 2 Complex Functions

EXERCISES FOR SECTION 2.6

For Exercises 1-8, find the image of the given circle or line under the reciprocal
transformation w = IIz.
1. The horizontal line Im(z) - y . 2. The circle | z + //2 | = y •
3. The vertical line Re(z) = - 3 . 4. The circle \z + 2 | = 1.
5. The line 2x + 2y = 1. 6. The circle | z - U2 | = 1.
7. The circle \z - y | = 1. 8. The circle \z + 1 - i| = 2.
9. (a) Show that lim (1/z) = 0. (b) Show that lim (\lz) = ~.
10. Show that the reciprocal transformation w = \lz maps the vertical strip 0 < x < y onto
the region in the right half plane Re(w) > 0 that lies outside the circle | w — 1 | = 1.
11. Find the image of the disk | z + 2//3 | < y under the reciprocal transformation.
12. Show that the reciprocal transformation maps the disk \z — 1 | < 2 onto the region that
lies exterior to the circle | w + y | = y .
13. Find the image of the half plane y > y — x under the mapping w = 1/z.
14. Show that the half plane y < x - y is mapped onto the disk | w - 1 - i | < Jl by the
reciprocal transformation.
15. Find the image of the quadrant x > 1, y > 1 under the mapping w = 1/z.
16. Show that the transformation w — 2/z maps the disk | z — i \ < 1 onto the lower half
plane Im(w) < — 1.
17. Show that the transformation w — (2 - z)lz = - 1 + 2 / ^ maps the disk | z - 11 < 1
onto the right half plane Re(w) > 0.
18. Show that the parabola 2x = 1 — y2 is mapped onto the cardioid p = 1 + cos $ by the
reciprocal transformation.
19. Limits involving °°. The function/(z) is said to have the limit L as z approaches oo, and
we write
lim/(z) = L

if for every e > 0 there exists an R > 0 so that

\f(z) - L\ < £, whenever \z\ > R.

Use this definition to prove that

20. Show that the complex number z = x + iy is mapped onto the point

( x y x2 + y2 \

x2 + y2 + 1 ' x2 + y2 + 1 ' JC2 + y2 + 1/

on the Riemann sphere.
21. Explain how are the quantities +°°, — «>, and °o different? How are they similar?
22. Write a report on Mbbius transformation. Include ideas and examples that are not men-
tioned in the text. Resources include bibliographical items 12, 23, 24, 30, 36, and 43.