# 8.

Mapping by Elementary Functions
 Introduction
The main purpose of this chapter is to build up a basic library of conformal mappings which is crucial in the solution of 2 - D
Laplace equations with complicated boundary conditions.
A collection of some mappings of interest can be found in App 2 of Churchill.
The program F(z).exe can be used if more details are desired.
The complete mapping of a complex function cannot be shown in a single graph since it involves a space of 4 dimensions. The
approach taken here is to list the mappings of a few representative domains. Since in applications, one is required to find a
mapping which reduces a complicated boundary to a simple one, some familarity with a collection of such mappings is needed
to get one started. Fortunately, simple composites of elementary functions already provide a respectable arsenal adequate for
most situations of interest.
Some domains of common interest are
1.
Fixed points: z = f HzL
2.
3.
4.
5.
6.

Lines. Of particular interest are the horizontal & vertical ones.
Circles.
Rectangles, half-planes, quadrants, strips.
Disks, annular region.
Sectors of disks or annular region.

 w=az+b

Alias: Linear Transformations
b
Fixed points at z =
Ha ¹ 1L
1 - a
 w=az

Let
a =  a¤ ei Φ0 ,
z =  z¤ ei Φ

w =  a¤  z¤ ei H Φ+Φ0 L
Interpretation: expand (  a¤ > 1 ), contract (  a¤ < 1 ), & rotate.
 w=z+b

Interpretation: translation
 w = 1z

Fixed points at z2 = 1

or

z = ±1

 1st Interpretation

w* =

1

=

z*

z

z¤2
1

arg w* = arg z
z¤
Interpretation: move points inside the unit circle to the outside radially & vice versa
w* =



w=

1
z

=

z*

z¤2

w¤ =

1

arg w = arg z* = - arg z
z¤
Interpretation:
1st:
move points inside the unit circle to the outside radially & vice versa


r2 = 0 A line by: y=mx+b Therefore.bL2 = r2 or x2 + y2 .2 a x . Using  x = u I x2 + y2 M y = v I x2 + y2 M I u2 + v2 M I x2 + y2 M = 1 The eq A I x2 + y2 M + B x + C y + D = 0 becomes  I x2 + y2 M H A + B u + C v L + D = 0 A + B u + C v + D I x2 + y2 M = 0 a2 + b2 . with: B D m= b= HC¹0L C C mx+b=Bx+D=0 HC=0L In the case of a circle._MappingByElementaryFunctions. . the eq A I x2 + y2 M + B x + C y + D = 0 describes a circle. with B C a=b=2A 2A D 1 r2 = a2 + b2 .2 08. the distance from its center Ha.  w=u+iv= 1 = z  z= u= 1 = w  z*  z¤2 x 2 x +y u-iv = 2 x-i y = x2 + y2 x  z¤ y v=- 2 2 x +y 2 =- y  z¤2 u2 + v2 x= u u2 + v2 Interpretation: = u v v  w¤2 u2 + v2  w¤2 transforms lines into circles & vice versa. when A = 0.aL2 + H y . bL to the origin is r0 = D Its radius r is r2 = r0 2 .arg z  z¤ Interpretation: 1st: move points inside the unit circle to the outside radially & vice versa then: reflects about the real axis. yL plane is described by: Hx . when A ¹ 0. The inside of unit disk  y=- =- region outside.= I B2 + C2 .nb 1 w= = z z*  z¤2  w¤ = 1 arg w = arg z* = . A \ The circle passes through the origin when D = 0.4 A D M > 0 A 4 A2 or a line.  2nd Interpretation A circle in the Hx.2 b y + a2 + b2 ..

lines & circles are mapped into lines & circles. 0 & passes through the origin.  Summary 1. A line passing through the origin: Bx+C y=0 is mapped into a line passing through the origin: Bu+Cv=0  Example 1 Vertical line x=c The quickest way is to apply the formulae developed above.  Example 2 ._MappingByElementaryFunctions. centered at 2c . 2. 2c Note also that 1. A circle not passing through the origin: A I x2 + y2 M + B x + C y + D = 0 is mapped into a circle not passing through the origin: A + B u + C v + D I x2 + y2 M = 0 2.   or z=c+i y c . A circle passing through the origin: A I x2 + y2 M + B x + C y = 0 is mapped into a line not passing through the origin: A+Bu+Cv=0 A line not passing through the origin: Bx+C y+D=0 is mapped into a circle passing through the origin: B u + C v + D I x2 + y2 M = 0 4.iy 1 w= = =u+iv z c2 + y2 y c u= v=c2 + y2 c2 + y2 1 u u2 + v2 = = c c2 + y2 u - 1 2 1 + v2 = 2c 2 2c 1 which is a circle of radius 1 . The size of the circle is inversely proportional to the distance of the line from the y -axis.08.nb Using  3 x = u I x2 + y2 M y = v I x2 + y2 M I u2 + v2 M I x2 + y2 M = 1 The eq A I x2 + y2 M + B x + C y + D = 0 I x2 + y2 M H A + B u + C v L + D = 0 becomes A + B u + C v + D I x2 + y2 M = 0  Hence. direct derivation for each special case is more appealing to those who detest memorization. The circle & line are on the same side of the y -axis. However. 3.

_MappingByElementaryFunctions. centered at .nb Example 2 Horizontal line y = c z=x+ic 1 x . The circle & line are on the opposite side of the x -axis.4 08. centered at 0.ic  w= = =u+iv z x2 + c2 x c u= v=x2 + c2 x2 + c2 1 v  u2 + v2 = =2 2 c x + c 2 2 1 1 or u2 + v + = 2c 2c 1 1 which a circle of radius .  Example 3 x³c Half plane   z=x+i y x .d ± Hd . Note also that 1.aL z .aL2 . 2c . 2. 2c 2c & passes through the origin.iy 1 w= = =u+iv z x2 + y2 y x u= v=x2 + y2 x2 + y2 1 u v u2 + v2 = = =2 2 x y x + y x=  Hc>0L u ³c>0 u2 + v2 u u2 +v2 - £0 c or u - 2 1 + v2 £ 2c 1 2 2c 1 which denotes the region inside a circle of radius 2c  w= az + b cz + d Had -bc¹0L Fixed points at c z2 + Hd . 0 & passes through the origin.b = 0 1  z= : a . The size of the circle is inversely proportional to the distance of the line from the x -axis.4 c b > 2c  1 .

1 provided d 1. D = -b. we have: a b w= z+ d d d b z= wa a The transform is 1 . 5 .nb  Mobius Transformation  czw -az +d w. assuming c ¹ 0: az + b a w= = + cz + d c Let Z=cz + d  w= a c + b- ad c W= b- × 1 cz + d 1 = Z cz + d 1 ad W c Provided ad-bc¹0. d For d = 0: w= a + c b 1 × is a combination of translation & inversion with c ¹ 0.1 provided a. B = -a. d ¹ 0.08. c a 2. Since A = c.a ¹ 0 or w ¹ . c z + d ¹ 0 or z ¹ .b=0 º Azw+Bz +Cw+D The last form is called the Mobius or bilinear transformation. C = d._MappingByElementaryFunctions. c For c = 0. \ lines & circles  lines & circles The inverse of the transform az + b w= cz + d is: -dw + b z= cw . c w . it is a composition of linear transformations & inversion..b c = -B C + D A ¹ 0  Properties in the finite plane For c = 0: w= a z+ d b is a linear transform with d ¹ 0.  a d . the transform is 1 .a ad-bc¹0 For c ¹ 0. c z In general.

i i .i i + z . we need to consider the transforms for the extended point ¥._MappingByElementaryFunctions.6 08. for c ¹ 0. Thus. T -1 = ¥ c In addition. the above pair of transform is everywhere 1 .2 b = 0 or c= a a + b a  i= = or b = -i a b2 b + b a \  d = -i a c = -a z .1. d 1. To be more precise.z w= = -z . i Let the transformation be wHzL = az + b cz + d The mapping of the 3 points give: -a + b b a + b -i = 1= i= -c + d d c + d  b=d -a + b a + b -i = i= -c + b c + b a . a TH¥L = c d T -1 H¥L = c For c = 0.nb  Extended Plane For the extended complex plane which includes the point ¥ . 0. 1 into -i.b a + b  = -c + b c + b b2 2 2 a c . T =¥ c a 2. 1.a Points at which these transforms fail to exist in the finite complex plane now becomes well defined. let az + b THzL = cz + d -dw + b T -1 HwL = cw . the undefined points of the finite plane transform are simply ¥ so that TH¥L = ¥ Hd=0L -1 T H¥L = ¥ Ha=0L  Example 1 Find transformation which maps -1.

z1 L = Hz3 . However.z1 L Hw3 . 1 into 1.z1 L Hw3 .w3 L H z .1 w= = = H1 . zi L " i = 1.w2 z .z2 L H z2 .w1 w .z1 z3 .w2 z .w2 z2 .z2 = w3 .z3 L = H w . One can prove the above assertion by demonstration ( see Churchill sec 71 ).iL a z + i Hz + iLH1 + iL H1 + iLz + i .b = -i Ha + bL or b= 1+i 1-i a= ia c = H1 .w1 L (3) we see that it reduces to the bilinear form after expanding the 1st two factors on each side & collecting terms.z2 L H z2 .z1 L H w2 .  proof It's well-known that a linear transformation w= Az + B can be written as w . We now try to obtain a similar form for the bilinear ( Mobius ) transformation: Azw+Bz +Cw+D=0 The presence of the H z w L term then suggests the form w .w2 L 7 . both sides = 0. Writing (2) as H w . we have H w3 .z1 = (1) w2 .w1 L H z .z1 L H z2 .w1 z .w1 L H z .z1 z . this leaves the form itself a product of pure inspiration. The following derivation aims to rectify this shortcoming. we can write H w .z2 L H z2 . 0. For i = 1.08. 2. For i = 3.z2 × = × (2) w2 .w1 z2 .w2 L = H z .z1 L H w .nb  Example 2 Find transformation which maps -1. 2._MappingByElementaryFunctions.w1 L H w2 .w1 w3 . ¥.iLz 2z 2z  Implicit Form Since 3 points defines a linear fractional transformation ( see eg 1 & 2 ). w .w2 z3 .z1 L which is called the implicit form of the linear fractional transformation.z1 showing that 2 points determine the transformation uniquely.w2 L Hz3 .z2 The use of the reciprocal is to create ( z w ) terms.w1 L Hz2 .z2 where we've multiply (1) with the reciprocal of another linear transform.w2 L H w2 .z2 L H w2 . i Let the transformation be wHzL = az + b cz + d The mapping of the 3 points give: -a + b b a + b 1= ¥= i= -c + d d c + d  d=0 -a + b a + b 1= i= -c c  \  a . What remains is to show that (3) is satisfied by points H wi . 3.w1 L Incorporating this condition into (3) then gives H w .w1 L H z .

w1 L (3) we see that it reduces to the bilinear form after expanding the 1st two factors on each side & collecting terms.z1 L H z2 .w1 L H z .z1 L H z2 .z1 L = Hz3 .w1 L H z .w1 L H w2 .z1 L H w2 . To bring it into complete agreement with eq(1) in sec71 of Churchill.1LH1L w+ i z + 1  i= w.i H z + 1 L w= iHz .z2 L H z3 .z2 L H w3 .w2 L H z2 .w2 L H w3 .w2 L H z . 1 into -i.w1 L z3 .1 Hz .w1 L H w2 .nb Writing (2) as H w .w2 L Hz3 .w2 L Hz3 . zi L " i = 1.z1 H w .z2 L H z2 .w3 L H w2 .z3 L H z2 .1L .w1 L H z .w3 L H w2 .H1 + i L = Hi .w3 L H w2 .1 -z + i = z + i =  Example 2 Find transformation which maps -1.w2 L = H z . For i = 1.z1 L = H z .z3 L H w .z3 L H w . The implicit form is particularly suited for cases where the transorm of 3 points are known.  Example 1 Find transformation which maps -1. 0.8 08. What remains is to show that (3) is satisfied by points H wi .w1 L H z .z1 L we have H w . 3.w1 L H w2 .z1 L H w . 2.z1 L H w .1Lz .i z.w1 L H z .w3 L H z .z3 L = H w .z3 L H z2 .w3 L H z .z2 L H z2 .w1 L Hz2 . 0.iLH1 + iL Hz. 3 & 3. as will be illustrated in the following examples.z1 L Hw3 . 1 in the numerator & denominator.z1 L we have H w + i L H 1 .w1 L H w3 . 3. 1 into 1. 2. i Using H w .i L z .Hz + 1L H 1 . For i = 3. we have H w3 . ¥.1 L H ¥ L H z + 1 L H -1 L = = . 2.z1 L H z2 . 1. both sides = 0.w3 L H z . resp.z1 L H z3 .z2 L H w2 ._MappingByElementaryFunctions.z1 L The last form is easier to memorize since the subscripts appear in the order 1.1 = iz .H1 + iL -i z . we make the interchange of index 2 « 3. giving H w .1L .w1 L Incorporating this condition into (3) then gives H w .z2 L  = H w .z1 L Hw3 .z3 L H z2 .w1 L H z . 2.z1 L which the correct transform satisfying wi = wHzi L for i = 1. i Using H w . 2.i L H z + 1 L H -1 L = Hw .

b. Thus: 0£Ζ £Π 0£W £1 b = B ei Β d = D ei ∆ Consider 1st points on the real axis: Im z = 0.i z._MappingByElementaryFunctions.w1 L H w2 .w1 L H z .w3 L H w2 .z3 L H z2 .08.nb Using H w .1 w= z .1 i Hz + 1 L + z .1 L H ¥ L H z + 1 L H -1 L = Hw .z3 L H w . It also emphasizes the fact that all 'points' with z = ¥ is considered as a single point ¥. we'll adopt the convention f iΖ z=Ze w = W ei W a = A ei Α c = C ei Γ º F for all complex number f . d.1LH1L w -1 z + 1  =w.z0 Im z0 > 0 z . z = ± Z: ± Za + b 1= ± Zc + d For the origin z = 0: B 1=  D=B¹0 D For Z ® ¥ : A 1=  C= A¹0 C For the rest of the upper plane: Im z > 0 az + b a w=  c z + d z®¥ c A \ W  =1 z®¥ C This is to be expected from the real axis case. C = A ¹ 0 we can write a z + w= × c z + b a d c . Using the fact that D = B ¹ 0.1 = 2z =  Mappings of Upper Half Plane w = ei Φ × z .z1 L H z2 .z1 L we have H w . For convenience.iLH¥L Hz. c.1 + z + 1 H1 + iLz + i .w3 L 9 H z .z0 * is the mapping which maps the upper half plane into the unit circle: w <1 Im z > 0 >  : Im z = 0 w =1  proof Assuming the transform to be bilinear: az + b w= Had-bc¹0L cz + d The task is to determine a.

z0 * with 3 free parameters Φ. To ensure the upper plane is mapped inside the circle. See example 2 for the case Im z0 = 0.i z + i z .nb For the rest of the upper plane: Im z > 0 az + b a w=  c z + d z®¥ c A \ W  =1 z®¥ C This is to be expected from the real axis case. It also emphasizes the fact that all 'points' with z = ¥ is considered as a single point ¥. can be beautified with the following notations: Α-Γ=Φ B A e i H Β-Α L = -z0 = -Z0 ei Ζ0 so that w = ei Φ × z .z0 z .z0 1> z .sinΖ0 < sinΖ0  Π > Ζ0 > 0 ie.ΓL  w = ei H Α . we need z .i = ei Π × z + i z0 = i Im z0 = 1 > 0 .2 A B cosH Α .Α L = cosH ∆ . & Ζ0 .Α = -H∆ .Α = ∆ .z i + z w= -  z ._MappingByElementaryFunctions.10 08. Z0 .Β L . The 2nd option.  cosH Β .i H Β-Α L The 1st choice is clearly too restrictive.Ζ0 L for 0<Ζ <Π -sinΖ sinΖ0 < sinΖ sinΖ0 . Im z0 > 0 Note: This condition can be obtained by inspection from fig 79 since z0 * is always further away from z than z0 does whenever z is in the upper plane.Γ L w=e iHΑ-ΓL z + × z + B A B A e i H Β-Α L e. C = A ¹ 0 we can write b a a z + w= × c z + d c iHΑ-ΓL = e z + B A z + B A × e i H Β-Α L e i H ∆-Γ L For points on the real axis: ±Z + 1= ±Z + Using a-b 2 B A e i H Β-Α L B i H ∆-Γ L e A 2 2 = A + B .Γ L We therefore have either Β . which we shall adopt.z0 *  cosH Ζ + Ζ0 L < cosH Ζ . Using the fact that D = B ¹ 0.  Example 1 w= i .Γ  or Β .

 Example 2 w= z . The upper half z . This mapping is many to 1. let x = -1 + Ε wHxL =  wHxL = H-1 + Ε L2 . 11 .  w= = = \ x .plane with wH¥L = -1. Moving upward along the line ( y increases ) results in moving counterclockwise on the circle ( Φ increases )._MappingByElementaryFunctions.  w = ez Let z=x+i y  w = ez = e x ei y = Ρ ei Φ Ρ = ex Φ= y+2nΠ The mapping is in general many to 1.1 + i yLHx + 1 .1 + y2 H x + 1 L + y2 2 Im w > 0 +i 2y H x + 1 L2 + y2 for y > 0.plane.08. The horizontal line y = c. x2 . Any y interval of length 2 Π will cover the whole circle once.1 + iy x + 1 + iy Hx .i = ei Π × z + i  z0 = i Im z0 = 1 > 0 \ w is of the form that maps the upper plane into a circle.i z + i z . The real axis y = 0 is mapped onto the real axis of the w .1 = 1- Ε2 -2 + Ε = Ε 2 ±¥ Ε Ε® 0¡ which shows the entire real w axis is indeed covered. The vertical line x = c. or z = x + i c is mapped into w = e x ei c with w = ex which is a ray ( line extending from the origin to infinity ) in the direction Φ = c.1 z + 1 z0 = 1  Im z0 = 0 The range of the mapping is no longer confined as witnessed by wH-1L = ¥ Using z = x + i y .1 H x + 1 L2 For points near -1.plane is mapped onto the upper half w .i yL H x + 1 L2 + y 2 x2 .nb w= - z . or z = c + i y is mapped into w = ec ei y w = ec with c which is a circle of radius e . y > 0.

y = 0.12 08.e .plane. Let z = r ei Θ Α£Θ£Α+2Π  w = ln z = ln r + i Θ which is simply the inverse of the exponential mapping.plane into the infinite strip 0 < Im w < Π. A rectangular region a£x£b c£ y£d is mapped into w = Ρ ei Φ with limits ea £ Ρ £ eb c£Φ£d An infinite strip a £ y £ b is mapped into w = Ρ ei Φ with a £ Φ £ b which is a fan shape region of infinite extend.1 only within a single branch of ln z._MappingByElementaryFunctions.plane into the infinite strip. & Log maps the upper half Z .nb The horizontal line y = c.yL cosh y = H e y + e . or  Ranges From the definitions 1 1 sinh y = H e y .yL = cosh y -¥ £ sinh y £ ¥ 1 £ cosh y £ ¥ . or z = x + i c is mapped into w = e x ei c w = ex with which is a ray ( line extending from the origin to infinity ) in the direction Φ = c.  w = ln z The mapping is 1 . Fixed point at x = sin x cosh y y = cos x sinh y  z=0 x = 0.  Example w = Log Let  Z= z -1 z + 1 z -1 z + 1 w = Log Z Since Z maps the upper half z .plane into the upper half Z .yL 2 2 One sees that for -¥ £ y £ ¥ sinh H-yL = -sinh y cosh H. the composite maps the upper half z .  w = sin z Let z=x+i y  w=u+iv w = sin z = sin x cosh y + i cos x sinh y u = sin x cosh y v = cos x sinh y The mapping is many to 1 with period 2 Π in the x direction.

we have u or v = 0 for n = even or odd.sinh2 y = 1 Using u2  £ sinh b cos c v2 =1 cos2 c sin2 c which is a hyperbola with foci points at w=± - v sin2 c + cos2 c = ± 1 The signs of sin c & cos c determine into which portion of the hyperbola is mapped. we have cosh c £ cosh y £ cosh d where c=: min 8 a . b are of the : same sign.  Vertical Line The vertical line x = c or z = c + i y is mapped into w = sin c cosh y + i cos c sinh y u = sin c cosh y v = cos c sinh y The bounds for u & v must be carefully calculated according to the rules given in sec Ranges._MappingByElementaryFunctions. resp. opposite The situation for sin x & cos x is even more complicated owing to their periodicity. c¹n 2 we have u 1£ £ cosh b 0£ sin c cosh2 y .yL = cosh y -¥ £ sinh y £ ¥ 1 £ cosh y £ ¥ 13 a£ y£b For a finite range sinh a £ sinh y £ sinh b.nb -¥ £ y £ ¥ One sees that for sinh H-yL = -sinh y cosh H. Further more. 0 d = max 8 a .08. which corresponds to the imaginary or real axis. b < b < if a. For the case Π 0 £ y £ b. Π If c = n . We'll dispense with an enumeration of the results for every possible combinations in the perhaps unjustifiable hope that an average student should be able to work it out for his/her-self. . the 2 n = odd case uses only portions of real axis which satisfies 1 £ u sin In 2 M Π £¥.

-1D v = 0 For the horizontal boundaries: Π Π 1. cosh bD v= 0 Π 2. x= 0£ y£b 2 is mapped into a segment of the real axis with u = cosh y Î @1.£x£ y=b 2 2 is mapped into a segment of the upper ellipse with u2 v2 + =1v>0 cosh2 b sinh2 b Π Π 2._MappingByElementaryFunctions.plane.  Horizontal Line The horizontal line y = c or z = x + i c is mapped into w = sin x cosh c + i cos x sinh c u = sin x cosh c v = cos x sinh c u v  -1 £ £1 -1 £ £1 cosh c sinh c Using cos2 y + sin2 y = 1  u2 v2 =1 cosh2 c sinh2 c which is an ellipse with foci points at w=± + cosh2 c .14 08.sinh2 c = ± 1  Rectangular region Π Region - Π £x£ 2  0£ y£b 2 -1 £ sin x £ 1 1 £ cosh y £ cosh b u = sin x cosh y  -cosh b £ u £ cosh b 0 £ cos x £ 1 0 £ sinh y £ sinh b v = cos x sinh y 0 £ v £ sinh b For the vertical boundaries: Π 1.nb  Vertical Strip For the semi-infinite vertical strip Π Π . .cosh y Î @.£x£ y³0 2 2 we have -1 £ sin x £ 1 0 £ cos x £ 1 0 £ sinh y £ ¥ 1 £ cosh y £ ¥  -¥ £ u £ ¥ 0£v£¥ which is the upper half w . 1D v= 0 .cosh b.£x£ y=0 2 2 is mapped into a segment of the real axis with u = sin x Î @-1.0£ y£b 2 is mapped into a segment of the real axis with u = . . x=.

.nb For the horizontal boundaries: Π Π 1. 15 .£x£ y=0 2 2 is mapped into a segment of the real axis with u = sin x Î @-1. the same result can be obtained as follows: 1 w=e2 = ln z 1 = e2 r e i Θ 2 8 ln r + iH Θ + 2 Π n L < +iΠn A fan shaped region 0£r£R a£Θ£b is mapped into another fan shaped region a b 0£ Ρ£ R £Φ£ 2 2 where w = Ρ ei Φ.  w = sinh z w = sinh z = i sin Hi zL = i sin Z where Z = i z The mapping is therefore a composition of a rotation._MappingByElementaryFunctions. . Alternatively. 1D v= 0  w = cos z Π w = cos z = sin z + = sin Z 2 Π where Z = z + . a sin transform. & then a rotation. 2 The mapping is therefore a composition of a translation & sin transformation. 1  w = zn Generalization of the above treatment to this case is obvious.£x£ y=b 2 2 is mapped into a segment of the upper ellipse with u2 v2 + =1v>0 cosh2 b sinh2 b Π Π 2. 1 -Π < Θ < Π.  w= z Let z = r ei Θ = r ei H Θ + 2 Π n L  w= z = r e i Θ i HΘ+2ΠnL 2 +iΠn = r e 2 n = 0 is the principal branch when n = 0.08.

It's also legitimate to use 2 lines each extending from a branch point to ¥ in an arbitrary direction. ie. we have: z-1 = r. where Θ = 1 . ( roots of even order are not branch points ) Branch cuts are obtained by joining pairs of branch points.remains in a specific branch. With the line between z = ± 1 as the branch cut.  w= Pn HzL Here Pn is a polynomial of order n.L + H n+ + n. Since ¥ is a single point. Caution: z = ¥ is not a branch point since w z®¥ z2 = z is single valued.L Π 2 2 with a range of 2Π. which is a point just above the branch cut.= Π & Φ+ = 2 Π.1 L H z + 1L w= 1 8 logH z-1L + logHz+1L < = e2 Branch points are at z = ± 1. the number of choices of cuts is infinite & one just pick the most convenient ones.L Π 2 If Φ+ & Φ.Π -Π < Φ.e i i Φ2 Φ+ + i n.L Π < Θ < + H n+ + n.16 08. we see that Π 3Π . The examples in Churchill were already discussed in Chap 3 & 6 so we won't repeat here.nb  Example w= sin z = Z = sin z Z Π with Z >0 0 < Arg Z < 2 This composite mapping is most easily illustrated graphically. ( see fig 91 ).= -Π & Φ+ = 0. these 2 lines are actually 1. Pn HzL = â ak zk = an ä Iz . The minimum of Θ occurs at Φ.ei Θ w= H Φ+ + Φ.  r + r . an ä Iz . The maximum of Θ occurs at Φ. The branch cut is therefore the line segment joining points z = ± 1.z j M 2 n  w= 1 j=1 We have therefore n + 1 branch points at z = z j & ¥ if all roots are distinct. Since ¥ represents the entire region outside the finite plane.  Example w = f HzL = z2 . which is a point just beneath the branch cut.< Π +in Π + z + 1 = r+ e 2 0 < Φ+ < 2 Π where any 2 consecutive n 's are adequate to denote the 2 branches._MappingByElementaryFunctions. This confirms the correctness of the branch cut assignment. In general.+ H n+ + n.1 H z . The situation is best visualized in the Reimann sphere.z j M n n k=0 j=1 where z j are its roots. there are m + 1 branch points where m is the number of distinct roots of odd order.

= Π & Φ+ = 2 Π.+ H n+ + n.Π -Π < Φ._MappingByElementaryFunctions.L Π 2 2 with a range of 2Π.e i i Φ2 Φ+ + i n.< Π + i n+ Π z + 1 = r+ e 2 0 < Φ+ < 2 Π where any 2 consecutive n 's are adequate to denote the 2 branches.  r + r . where Θ = 1 17 . we see that Π 3Π .L Π 2 If Φ+ & Φ.ei Θ w= H Φ+ + Φ. which is a point just beneath the branch cut. This confirms the correctness of the branch cut assignment.nb With the line between z = ± 1 as the branch cut. The minimum of Θ occurs at Φ.L + H n+ + n. we have: z-1 = r. which is a point just above the branch cut.remains in a specific branch.= -Π & Φ+ = 0.L Π < Θ < + H n+ + n.08. The maximum of Θ occurs at Φ.