Analytic Geometry With Calculus PDF

311AIVNV 39 13W
al
MUM J
-
ANALYTIC
GEOMETRY
with
CALCULUS
ROBERT C. YATES
University of South Florida
Englewood Cliffs, N. J. PRENTICE-HALL, INC. 1961

Q 1961, by Prentice-Hall, Inc., Englewood
Cliffs, New Jersey. All rights reserved. No
part of this book may be reproduced in any
form, by mimeograph, or any other means,
without permission in writing from the
publishers.
Library of Congress Catalog Card No. 61-9224
Printed in the United States of America
08488-C
Preface
In recent years analytic geometry and the calculus have been combined into
one course for the first or second year of college mathematics, and several
excellent texts have been published for this purpose. However, these texts
give primary emphasis to the calculus with a correspondingly reduced
content in analytic geometry. As a consequence, many students are not
acquiring a sound knowledge of analytic facts and techniques. For this
reason, this book emphasizes a full-bodied treatment of analytic geometry
in which the fundamental principles of calculus are introduced and used in
a supporting role. The result is a text, to follow algebra and trigonometry,
in which the student is more adequately prepared for the subject matter
of calculus.
Certain features of the book are listed here as a guide to the reader
before he begins a detailed study of the text. In the geometry of the plane,
rectangular and polar coordinates are introduced at the same time and used
interchangeably throughout. Lines are characterized by direction numbers
for wider application and to facilitate transition to three-space geometry.
The concepts of derivative and its inverse are presented early so that their
applications to direction of a curve, motion of points, plane areas, tangents
and normals to surfaces and curves in three-space, and volumes bounded
by surfaces considerably enhance the subject matter. Discussion of the
conies begins with the fundamental consideration of plane sections of a
right circular cone, thus establishing their proper designation as conic sec-
tions. The customary `higher plane' curves make their natural appearance
as important loci connected with various mechanical devices such as cams,
gears, and linkages. There are treatments of diameters of the conics and
diametral planes of the quadrics which provide a firm foundation for further
study of geometry. The usual treatment of transformations is amplified
and extended to plane mappings, some given in matrix form. Ruled surfaces
are studied and attention is focused upon their important application in
the construction of space gears.
This book contains ample material for a full semester course of four
hours a week or for a trimester of five, lours a week. For a shorter course of
vi Preface
three semester hours, certain sections (indeed, whole chapters) marked

with stars may be omitted without discontinuity. Some of these sections,
however, contain material that may well stimulate interest and should not
be omitted without considering the student's future.
The book has been written for the student. It is supposed that this is
his first introduction to analytic geometry, and to calculus. However, it is
assumed that he has some knowledge of algebra (with determinants) and
of trigonometry. The concepts of limit and derivative are presented in a
manner to give the student clear comprehension and understanding. The
formalized 'epsilon-delta' language is left to a later period when the student
will be more mature and capable of better appreciation. However, half-
truths and twilight meanings have been avoided.
The subject is to be enjoyed. It is often in this material that students
first realize the fascination and compelling absorption of mathematics.
The author gratefully acknowledges the invaluable help of Mrs. Al
Ferguson and Miss Carolyn Washer in preparing the typed manuscript,
the cooperation of students at the College of William and Mary for their
constructive and uninhibited criticism of several versions of trial texts,
and the meticulous preparation of line drawings by his son, Daniel S. Yates.
ROBERT C. YATES
Tampa, Florida
Contents
I. THE PLANE
1. The Real Number System-Graphs . . . . . . . . . . 3

1.1 The Real Number System and the Continuum Axiom . . . . 3
1.2. Coordinate Systems in the Plane . . . . . . . . . . . . 4
1.3. Graphs: Mathematical Statements . . . . . . . . . . . . 5
1.4. Graphs, Rectangular (General Remarks) . . . . . . . . . 8
1.5. Guide Lines (Rectangular Coordinates) . . . . . . . . . . 13
* 1.6. Techniques (Rectangular Coordinates) . . . . . . . . . . 15
1.7. Graphs, Polar . . . . . . . . . . . . . . . . . . . . . 20
2. Measures . . . . . . . . . . . . . . . . . . . . . . . 25
2.1. Distance. Midpoint of a Segment . . . . . . . . . . . . 25
2.2. Direction Cosines. Direction Numbers . . . . . . . . . . 26
2.3. Angle Between Two Line Segments . . . . . . . . . . . 28
2.4. Parallel and Perpendicular Line Segments . . . . . . . . . 29
2.5. Slope . . . . . . . . . . . . . . . . . . . . . . . . 30
2.6. Collinearity . . . . . . . . . . . . . . . . . . . . . . 31
2.7. Angle of Triangle Formed by P1P2P3 . . . . . . . . . . 31
3. Variables, Functions, Limits, Continuity . . . . . . . 33

3.1. Variable . . . . . . . . . . . . . . . . . . . . . . . 33
3.2. Function . . . . . . . . . . . . . . . . . . . . . . . 33
3.3. Notation . . . . . . . . . . . . . . . . . . . . . . . 34
3.4. Limits. The Symbol . . . . . . . . . . . . . . . . . . 34
3.5. Geometric Progressions . . . . . . . . . . . . . . . . . 36
3.6. A Special Limit . . . . . . . . . . . . . . . . . . . . 37
3.7. Limit Theorems . . . . . . . . . . . . . . . . . . . . 39
3.8. Continuity . . . . . . . . . . . . . . . . . . . . . . 39
viii Contents
4. The Derivative . . . . . . . . . . . . . . . . . . . . . 44
4.1. The Derivative in Rectangular Coordinates . . . . . . . . 44

4.2. Slope of a Curve y = f (x) . . . . . . . . . . . . . . . . 46
4.3. Maximum-minimum in Rectangular Coordinates . . . . . . 47
4.4. The Derivative in Polar Coordinates . . . . . . . . . . . 49
4.5. Derivatives of uv, u/v, u" . . . . . . . . . . . . . . . . 52
4.6. Differentiation of Implicit Functions . . . . . . . . . . . 54
5. Functions Whose Derivatives Are Known . . . . . . . . 56
5.1. Determination of the Function . . . . . . . . . . . . . 56

5.2. Area. Rectangular Coordinates . . . . . . . . . . . . . 59
5.3. Area. Polar Coordinates . . . . . . . . . . . . . . . . 61
k 5.4. Area. Parametric Equations . . . . . . . . . . . . . . . 63
6. The Straight Line . . . . . . . . . . . . . . . . . . . 66
6.1. Concept . . . . . . . . . . . . . . . . . . . . . . . 66
6.2. Slope . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.3. Families . . . . . . . . . . . . . . . . . . . . . . . 70
6.4. Concurrency . . . . . . . . . . . . . . . . . . . . . 74
6.5. Distance From Point to Line . . . . . . . . . . . . . . 77
6.6. Angle Bisectors . . . . . . . . . . . . . . . . . . . . 77
7. The Circle . . .
...
. . . . . . . . . . . . . . . . . 80
7.1. General Equations . . . . . . . . . . . . . . . . . . . 8o

7.2. Families of Circles . . . . . . . . . . . . . . . . . . . 83
k 7.3. Power of a Point . . . . . . . . . . . . . . . . . . . 89
k 7.4. Radical Axis . . . . . . . . . . . . . . . . . . . . . 91
8. Conics . . . . . . . . . . . . . . . . . . . . . . . . 95
8.1. General. Discussion . . . . . . . . . . . . . . . . . . . 95

8.2. Special Properties . . . . . . . . . . . . . . . . . . . 98
8.3. Equations of the Conics . . . . . . . . . . . . . . . . 102
8.4. More General Equations . . . . . . . . . . . . . . . . 105
8.5. Constructions . . . . . . . . . . . . . . . . . . . . . I09
Contents ix
9. The Conies: Their Properties and Applications . . . . . 112

9.1. The Reflective Property . . . . . . . . . . . . . . . . 113
9.2. Tangents . . . . . . . . . . . . . . . . . . . . . . . 116
* 9.3. Areas . . . . . . . . . . . . . . . . . . . . . . . . 117
9.4. The Parabolic Cable . . . . . . . . . . . . . . . . . . 119
9.5. Plane Motion of a Point . . . . . . . . . . . . . . . . 120
k 9.6. LORAN . . . . . . . . . . . . . . . . . . . . . . . 124
10. Some Mechanical Motions and Loci . . . . . . . . . . 127

10.1. Instantaneous Center . . . . . . . . . . . . . . . . . 127
10.2. The Trammel (Ladder) . . . . . . . . . . . . . . . . 128
k 10.3. The Conchoid . . . . . . . . . . . . . . . . . . . . 129
10.4. The Cycloids . . . . . . . . . . . . . . . . . . . . . 130
k 10.5. The Limacons . . . . . . . . . . . . . . . . . . . . 135
k 10.6. Linkage Motion . . . . . . . . . . . . . . . . . . . 135
10.7. An Involute of a Circle . . . . . . . . . . . . . . . . 139
k 10.8. Cams . . . . . . . . . . . . . . . . . . . . . . . . 139
k 10.9. Line Motion Linkages . . . . . . . . . . . . . . . . . 141
k 10.10. The Lemniscate . . . . . . . . . . . . . . . . . . . 142
11. Transformations . . . . . . . . . . . . . . . . . . . 145

11.1. Translation . . . . . . . . . . . . . . . . . . . . . 145
11.2. Rotation . . . . . . . . . . . . . . . . . . . . 145
11.3. Illustrations . . . . . . . . . . . . . . . . . . . . . 146
11.4. Removal of the xy-Term . . . . . . . . . . . . . . . . 148
11.5. Identification of Conics . . . . . . . . . . . . . . . . 150
k 11.6. Diameters of Conics . . . . . . . . . . . . . . . . . . 153
k 11.7. Principal Diameters . . . . . . . . . . . . . . . . . . 155
12. * Mapping . . . . . . . . . . . . . . . . . . . . . . 159

...
12.1. Translations . . . . . . . . . . . . . . . . . . . 159
12.2. Rotations . . . . . . . . . . . . . . . . . . . . . . 160
12.3. The Affine Linear Transformation . . . . . . . . . . . 163
12.4. Inversion . . . . . . . . . . . . . . . . . . . . . . 169
12.5. The Joukowski Airfoil . . . . . . . . . . . . . . . . 173
x Contents
II. THREE-SPACE
13. Coordinate Systems-Measures . . . . . . . . . . . . 177
13.1. Coordinate Systems . . . . . . . . . . . . . . . . . . 177

13.2. Distance . . . . . . . . . . . . . . . . . . . . . . . 180
13.3. Angle Between Two Line Segments . . . . . . . . . . . 181
13.4. Parallelism . . . . . . . . . . . . . . . . . . . . . . 182
14. Planes and Lines . . . . . . . . . . . . . . . . . . . 184
14.1. Definition . . . . . . . . . . . . . . . . . . . . . . 184

14.2. Definition . . . . . . . . . . . . . . . . . . . . . . 186
14.3. Families of Planes Through a Line. Projection Planes of a
Line . . . . . . . . . . . . . . . . . . . . . . . . 188
14.4. Normal to a Plane . . . . . . . . . . . . . . . . . . 189
14.5. Distance from Point to Plane . . . . . . . . . . . . . 190
14.6. Determination of Equations of Planes . . . . . . . . . . 192
14.7. Direction of a Line . . . . . . . . . . . . . . . . . . 195
14.& Determination of Equations of Lines . . . . . . . . . . 195
15. Surfaces and Curves . . . . . . . . . . . . . . . . . . 200
15.1. Definition . . . . . . . . . . . . . . . . . . . . . . 200

15.2. Definition . . . . . . . . . . . . . . . . . . . . . . 201
15.3. Graphical Representation of a Surface . . . . . . . . . . 202
15.4. Families of Surfaces and Projection Cylinders . . . . . . 204
16. Tangents and Normals . . . . . . . . . . . . . . . . 207
16.1. Tangent Line and Normal Plane to a Curve . . . . . . . 207

16.2. Tangent Plane and Normal Line to a Surface . . . . . . 210
17. Special Surfaces . . . . . . . . . . . . . . . . . . . . 214

17.1. Surfaces of Revolution . . . . . . . . . . . . . . . . . 214
17.2. Ruled Surfaces . . . . . . . . . . . . . . . . . . . . 218
17.3. The Cone and Cylinder . . . . . . . . . . . . . . . . 222
k 17.4. Gears . . . . . . . . . . . . . . . . . . . . . . . . 226
Contents xi
18. * Volumes . . . . . . . . . . . . . . . . . . . . . . 228

18.1. General Discussion . . . . . . . . . . . . . . . . . . 228
18.2. Illustrations . . . . . . . . . . . . . . . . . . . . . 229
19. Quadric Surfaces . . . . . . . . . . . . . . . . . . . 236

19.1. Plane Sections . . . . . . . . . . . . . . . . . . . . 236
* 19.2. Parallel Chords . . . . . . . . . . . . . . . . . . . . 237
k 19.3. Diametral Planes . . . . . . . . . . . . . . . . . . . 237
k 19.4. Principal Planes . . . . . . . . . . . . . . . . . . . 239
k 19.5. Classification of Quadric Surfaces . . . . . . . . . . . . 241
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
I.
THE
PLANE
1.
The Real Number

System-Graphs
1.1. The Real Number System and the Continuum Axiom

The counting numbers 1, 2, 3, are called natural numbers. A basic
assumption here is that for every such number n there is a successor n + 1.
There is, accordingly, no last natural number. Together with 0 these
numbers and their negatives comprise the set of integers.
A rational number is one that can be expressed as the quotient of two
integers, viz., 1, 6, 1, - s, 1.25, 0.454545 .
An irrational number is a real number that is not rational; i.e., one
that cannot be expressed as the quotient of two integers, viz., V2, 7r, ' - 7.
The set of all rational numbers (which includes the integers) and the
irrational numbers form the set of real numbers. This set may be displayed
as points on a line in the following fashion:
-3 -2 -1 0 1 2 3
V
3
Fig. 1.1
From any selected zero point 0 we mark regularly spaced points to the
right and left and label them by the integers in order of magnitude. The
3
4 THE REAL NUMBER SYSTEM-GRAPHS Sec. 1.1
rational numbers may then be assigned to intermediate points and the

irrational numbers to still other points.
The continuum axiom is the assumption that all real numbers may be
paired uniquely with points on a line and conversely; i.e., each point of a
line has a representative number. With points so labeled their order of
magnitude has geometric meaning. Thus the pair of statements
3<5 and -7<-2
means the point 5 lies to the right of 3, and -7lies to the left of -2.
A further interpretation is possible and necessary. We wish to think of a
real number not only as representing a point on a line but also as the signed
length of the line segment to that point from the origin (the zero point).
Segments are thus directed.
1.2. Coordinate Systems in the Plane

(A) Rectangular Coordinate System. We shall assume a plane upon
which we select two perpendicular lines. Upon these lines, called x- and
y-axes, directed segments measured from their intersection 0 (the origin)
are taken to represent pairs of real numbers. We choose equal uniform
scales on both axes. Such pairs (a, b) of ordered numbers locate points P,
and conversely. That is, to every point P belongs a unique pair of num-
bers and to every pair of numbers there corresponds a point P. This is a
one-to-one correspondence. These number pairs are called coordinates of
P: the x-coordinate is the abscissa, the y-coordinate is the ordinate. It is
to be understood that the first number in a pair represents a segment
length in the x-direction, the second in the y-direction.
The selected lines taken as coordinate axes separate the plane into four
regions called quadrants, labeled as shown. Signs of number pairs in these
quadrants are, respectively (+, +) (-, +) (-, -) (+, -).
y
aP:(a,b)
b
-+ ±+ 1 9- x
Fig. 1.2 Fig. 1.3

(B) Polar Coordinate System. A second serviceable system in our
assumed plane has for reference an initial ray, or `half' line, and its initial
Sec. 1.2 THE REAL NUMBER SYSTEM-GRAPHS 5
end point 0, called the pole. An ordered pair of numbers (r, B) then locates
a point P whose directed distance from 0 is r, the radius vector of P; and
this vector makes the angle 0 with the initial ray. The angle of the pair is
positive when measured from the initial ray in the counterclockwise direction,
negative when clockwise. A negative distance r is to be interpreted as the
extension of the radius vector "backward" through the pole 0. For ex-
ample, (-2, 30°) is plotted by draw-
(1,) (2 30°) ing a radius vector at +30° from the
initial ray, then extending this line
(-1,0) 0 Initial ray backward from 0 a length two units.
(t''r)
(1, 330°)
It should be noted that although a
(-1,150°) single point is determined by a given
(-2,30_ )
pair of polar coordinates, the converse
Fig. 1.4 is not true. A selected point has an
unlimited number of coordinate pairs.
For example, the pairs (2, 30°), (-2, 210°), (-2, -150°), and (2, -330°)
all designate the same point.
If the angle is given in radian measure, there will be no symbol attached,
thus (2, a/6), (2, 1.5).
(C) Exchange of Systems. The rec-
tangular and polar coordinate sys- y
tems may be exchanged one for the p
other by making the pole and the
origin coincident, and the x-axis co- r y
incident with the initial ray as shown.
Thus a point P may have coordinates 8 x x
(x, y) in the rectangular system and
(r, 0) in the polar system. Relation-
ships between these two sets of co- Fig. 1.5
ordinates are apparently
x = rcos0 r= x2+ y2
and
y = r sin B 0 = are tan (y/x).
These relationships permit the transfer of coordinates in one system to the
other. For example, (J3, 1) in rectangular coordinates can be written as
(2, 30°) in polar coordinates; whereas (-1, 5ir/6) polar, is (d3/2, - z )
rectangular.
1.3. Graphs: Mathematical Statements

Equations or inequalities which relate x and y or r and 0 define loci in
the reference plane. That is, the set of those points (and only those) whose
roordinates satisfy such a statement forms the graphical representation or locus

of the mathematical statement. *
For example, 3x + 2y = 12 represents a straight line: the set of all
points whose coordinates (x, y) satisfy the equation. All such pairs of
numbers taken from the domain of all real numbers represent points on the
line shown. Note that the line intercepts lengths of four and six units on
the x- and y-axes, respectively.
Fig. I.6 Fig. 1.7
As another example, r = 2 sin 0 represents a circle. A table of coordi-

nates satisfying the equation is given, and the set of all such points forms
the circle shown. This circle passes through the pole (since for B = 0,
r = 0) and has its center at (1, ir/2). Such matters will be discussed fully
in later sections.
Let us determine the rectangular equation of this circle. It is con-
venient (and legitimate) to multiply first by r
r2 = 2r sin 0
before replacing these polar coordinates by their rectangular equivalents.
Since r2 = x2 + y2 and r sin 0 = y, we have at once
x2+y2=2y.
We will have occasion to determine points of the plane restricted by
conditions of the sort
{y=5l, y5x2).
The set A of points satisfying the first restriction, y S 1, is the shaded
region shown in the left-hand picture. For all points in this region x may
be any real number, but y may not be greater than + 1. The set B for
which y > x2 consists of those points on the parabola y = x2 together with
* Numbers which satisfy a mathematical statement are those which make the state-
ment true.
Fig. 1.8
all points àbove' the curve. The set of points common to the two sets
(points whose coordinates satisfy both conditions) is called their intersec-
tion and denoted by the symbol A (1 B. This intersection set is shown as
the cross-hatched region of the right-hand picture.
In working exercises pertaining to graphs, the student should list a table
of number pairs for each equation. Chosen values of one variable should be
listed in order of magnitude (from the domain of permissible values) and
corresponding points then plotted. If it is not prohibited in some way, a
smooth curve is then drawn through these points in order of their tabular
arrangement. In doing this, it should be realized that an assumption has
been made; i.e., points on the graph actually exist whose coordinates, if
found, would satisfy the equation, and, of course, there are points on the
curve only for real number pairs.
EXERCISES
1. Locate the points whose rectangular coordinates are
(2, 3), (-2, 3), (-4, -2), (0, 0), (0, 3), (3, 0), (2, 1),
(0, -2), (-2, 0)
2. Show that it radians = 180°
3. Convert to radian measure

0°, 90°, 180°, 270°, 30°, 45°, 60°, 120°, 135°, 150-
210', 225°, 240°, 315°, 330°, - 30°, -225-
4. Convert to degree measure

a/6, it/4, u/3, v/2, 2ir/3, 31r/4, ir, 5rr/4, - 77r/6, - 5-/4,
- 71r/8, 2, 3.5, -1, -.75
5. Sketch the region of the plane such that

(a) {x> 2,y> 1}
(b) 11<x<3,1<y<2}
(c) {-1<x<0,-1<y<0}
(d) {x22<4,y2<9}
(e) {x<2,y< 1}
(f) {I x1: 1,1y1<_2}
(a) {05657r/6,0<r54}
(b) {058<a,i<r<2}
(c) I
(d) {7r/4 < 8 < 3ir/4, r sin B < 31

7. (a) Change the following polar coordinates to rectangular coordinates and
plot the points
(a, 0), (0, a), (1, it/2), (-1, 7r/2), (-1, -,r/2), (-2, 120°),
(-1, 7r)
(b) Determine for each of the following points in rectangular coordinates
four sets of polar coordinates with 0 < 0 5 21r
(0, 1), (1, 0), (-1, 0), (0, -1), (3, 1), (1, -1)
8. (a) Change the following statements to polar coordinates and sketch
x + y = 3,2x- 3y < l,y= l,x= -2,x2+ y2= y
(b) Change the following statements to rectanguhr coordinates and sketch
r sin 0 = 1, r cos 0 = -2, 2r sin 8 + 3r cos 8 = 5,
r=4,0=it/4,r=sin0,r=2cos0,0<r-cos0<1
1.4. Graphs, Rectangular (General Remarks)
In drawing graphs in rectangular coordinates the student will be guided
by the following remarks (additional details will be supplied in later
chapters).
(1) The graph of
y=ax+b
is a straight line, and the location of y

two of its points determines its posi-
tion. Note particularly the points (0, b)
(0, b) and for a 54 0, (-b/a, 0)
where the line crosses the axes. These
lengths b and - b/a are intercepts. The
graph of x = k is a line parallel to the
y-axis; y = k is a line parallel to the
x-axis. Fig. 1.9
(2) The graph of
y=ax2+bx+c, aX 0
Fig. 1.10
is a parabola of the types shown. The domain of x is unrestricted, and for

sufficiently large values of x, either positive or negative, the term ax2 is
numerically dominant. Thus if a > 0, the curve extends upward; if a < 0, it
extends downward. The y-intercept is c; the x-intercepts (if any) are the
real roots of ax2 + bx + c = 0.
If we solve
ax2+bx+c - y = 0
for x:
b -4ac ++4ay
x
2a 2a
it is evident that lines y = constant

(those parallel to the x-axis) intersect
the curve in either two points or in
no point depending upon the sign of
the quantity (b2 - 4ac + 4ay). If this
quantity is 0, however,
b2-4ac+4ay =0; Fig.1.11
4ac-b2
y =
4a
Then
and there is but one point of intersection. This y-coordinate is the maxi-
mum (y,u) if a < 0; the minimum (y.m) if a > 0.
We shall re-examine this important feature from a somewhat different
point of view. Let a line be drawn through P,(x,, y) and P2(x2, y), two
points on the curve with the same y-coordinate. Since they are points of
the curve, their coordinates satisfy the equation; thus,
y =ax;+bx,+c
y = axe + bx2 + c.
Subtracting these,
0 = a(x2, - xz) + b(xi - x2)
or
0 = (XI - x2) [a(xl + x2) + b].

Now, if P, and P2 are distinct points, then x, x2, and thus necessarily
a(x, + x2) + b = 0.
We now move the line P,P2 always parallel to the x-axis so that the
points approach each other. Their x-coordinates then approach a common
value X, which we indicate by writing
x,
-> X, (xi Fd x2).
x2
The last equation then approaches

2aX+b=0.
This X, called the critical value of x, is the value obtained previously and
will give the maximum or minimum value of y.
Note that the x-coordinate of the midpoint of any of the segments
P, P2 is
(x, + x2) /2.

THE REAL NUMBER SYSTEM-GRAPHS 11
Sec. 1.4
Since
a(xi+x2) + b =0,
then
xi+x2 b
2 = 2d
and this is the line of symmetry of the curve.
Fig. 1.12
(3) The graph of

y = ax= + bx2 + ex + d, a 0
is "S"-shaped as shown. The y-inter-
cept is d, the x-intercepts are the real
roots of axa + bx2 + cx + d = 0.
Since every polynomial equation of
odd degree with real coefficients has
either one or three real roots, the
curve intersects the x-axis once or
three times.
We now determine critical values
of x using the method described under
the parabola. Two points, Pi(xi, y)
and P2(x2, y) having the same y-
Fig. 1.13
coordinate, are taken on the curve.
Then
y=axi+bxi+ex,+d
y = axe + M, + cx2 + d.
Their difference gives
0 = a(x; - x2) + b(x, - x2) + c(xi - x2)

or
0 = (xl - x2) [a(xi + x1x2 + x2) + b(x, + x2) + c].

For distinct points Pi, P2, we must have x, x2 and thus
a(x2+x,x2+x2) +b(x,+x2) +c =0.
Let the segment P,P2 move parallel to the. x-axis so that P, -+ P2. Then,
x, and x2 approach a common value Y. The last equation approaches
3a12+2bX+c=0.
The real roots (if any) of this quadratic are the critical values of x and
give, on substituting in the equation of the curve, the maximum and
minimum values of y. Note that the curve will have maximum and mini-
mum points only if this quadratic has real roots; that is, only if
b2-3ac>0.
The curve has a point of symmetry, called a point of inflexion. To show
this we take three points P,, P2, and P3, on the curve with the same y-
coordinate and ask that P2 be the midpoint of P,P3; that is,
Si xa
x2 = or 2x2 = x, + xa or 3x2 = x, + x2 + xa.
2
Now, the sum (x' + x2 + xa) of the roots of
axa+bx2+cx+d = 0
is (- b/a) . Accordingly, the midpoint P2 of P3Pa is such that
b
3x2 = -ba or x2
3a
We now shall move the curve

without distortion so that the mid-
point is at the origin of coordinates.
%,y) Then x2 = 0, and thus the coefficient
x of x2 is 0. Moreover, the constant
term in the new equation is 0 since
(0, 0) is now a point of the curve.
We thus have y = Axa + Bx as repre-
Fig. 1.14
senting the cubic curve with point
of inflexion at the origin. Since
(-y) = A(-x)a + B(-x)
yields
y = Ax' + Bx,
then for each point (x, y) on the curve there is also a point (-x, -y)
on the curve. Such points are symmetric to the origin and the point of
inflexion is thus the point of symmetry.
EXERCISES
(a) {x+ y5 1,x 0, y>_ 01
(b) {:C+y:5 1,y 1,x>=-2{
(c) {x2+y54,y01
(d) {x2+y4,y> x}
(e) {y= I x-1 5 x5 1)
(f) {y>x2,y<x)
10. Sketch each of the following, giving x- and y- intercepts. Find coordinates
of maximum and minimum points if any.
(a) y = -3x (b) 4y = x
(c) y = 2x - 3 (d) 2y = 3x + 6
(e) y = x2 - 1 (f) y = .0+1
(g) y = x2 - 5x + 6 (h) y x2- x
(i)y=-x2+5x-6 G)y=x-x2
(k) y = x2 + x + 1 (I) y x2- 2x + 3
(m)y=xa (n) y xa- 1
(o) y x8+z+2 (p)y=x'-5x+4
(q) y = x(x - 1)(x - 2) (r) y = x(x - 1)2
(s) y = x2(1 - x) (t) y = (x - 1)2(2 - x)
1.5. Guide Lines (Rectangular Coordinates)

Certain lines (called guide lines or asymptotes), associated with a curve,
are important and useful. We discuss this feature with two examples.
x2
(A) y ` (x - 1) (x + 1).
For values of x near +1 or -1, the denominator of the fraction is near 0

and values of y are comparatively large. Thus
1.111
if x = 1.01, y = X55;
(0.01) (2.01)
if x = -0.99, y = -49.
If x is assigned values nearer and nearer to +1 or -1, values of y increase
(or decrease, i.e., grow negatively large). The lines x - 1 = 0 and
x + 1 = 0 thus serve as guides in sketching the curve.
Fig. 1.15 Fig. 1.16
Furthermore, if we write the equation as
y=1+xs-1'
we note that, as assigned values of x grow positively or negatively large,
y becomes nearer to + 1. The line y = 1 thus also is a guide line as shown.
(B)
xs-3x+4 1 1
2(x - 1) 2 x-1
It is evident that as x nears +1, y increases without bound.* The line
x = 1 is thus a guide line. Furthermore, as x grows positively or negatively
large, the fraction 1/(x - 1) nears 0, and the value of y grows nearer the
* y is said to increase without bound if it represents numbers arbitrarily large. This is
expressed symbolically by "y-+-."
value -x - 1. Accordingly, y = jx - 1 is a guide line as shown. Note that

the y-intercept is - 2, but there are no x-intercepts since for y = 0,
x2-3x+4=0
yields no real roots. The curve is called a hyperbola. There are two branches.
EXERCISES
11. Determine equations of guide lines and sketch
(a) y= x+
X- 1 =
1+ 2
x- 1
1
3x -- 4 4
(h) y _ = 3 -F
X x
x2- 5x
(c) y=
x- 1 =x-4--X- 4
x2- 1 3
(d) y= x+2=x-2-1-x+2
x - 5x + 6
(e) y =
x+3
3x2 + 2
(f) y = x-2
(g) y (x + 1):
x2
(h) y = x+4
-
*1.6. Techniques (Rectangular Coordinates)

It is advantageous in some instances to sketch relations simpler than
the given one; then combine these to form the required graph. The follow-
ing examples will make the procedure clear.
(A) Addition of Ordinates. Consider
1
yx 2+-x
THE REAL NUMBER SYSTEM-GRAPHS Sec. 1.6
16
y We draw the graphs y1 = x2, 112 =

1/x and from these, by addition of
ordinates, form the graph of y = y1 +
y2. The graph of y1 = x2 is a para-
bola through the origin with y-axis
for line of symmetry. The graph of
y2 = 11x is a hyperbola with the lines
x = 0, y = 0, (the coordinate axes)
as guide lines. For any x, y is the sum
of the directed distances y1, y2
We may locate the minimum point
by the method described in Par. 1.4
Fig. 1.17 (2) and (3) by forming
y =X + 1
XI
=x2+1x2
so that
xi-x2+--=0 1
XI x2
1
or
(x1 - x2)[(x1 + X2) - 1.

x1x2
] = 0.
Since x1 54 X2, then
(x1 + x2) - 1
x1x2
= 0.
X'
Letting -> x, we have
x2
Thus
x = 2-111 0.79
is the critical value and the corresponding
y = 3(2-211) - 1.89, the minimum y.
As a second illustration consider
y=x± /4-x2,
and let y, = X, Y2 = 4 - x2 so
that y = y, f y2. Note first that x
is restricted to the interval -2 <
x5 +2. The graph of,yi=xisa
x line through the origin. The graph of
y2 = 1/4 - x2 is the semicircle with
center at the origin and radius 2. For
each x, there are two values y:
y=y1+Y2, y=Y1-y2.
The resulting curveis called an ellipse.
It is symmetric to the origin.
Fig. 1.18
EXERCISES
12. Use addition of ordinates to sketch
(a) y= x2- 1
x
(b) y = x2 + 3x + 2
(c) y = x + (d) y = x2 + I
x3-1
(e) y =
x2
(g) y = x2 f Vx (h) y=x - I± 1+x

Y (B) Reciprocal Ordinates. Con-
(x, Y1) sider the Witch
1
Yi
Y
Y1
and let y, = x2 + 1, so that y = I/
y,. The graph of y, = x2 + 1 is a para-
bola with y,-intercept 1. For each x,
Fig. 1.19 the ordinate to the parabola is the
reciprocal of the ordinate to the Witch. As x increases positively or nega-

tively, yl increases positively, and y grows nearer zero. The line y = 0 is
thus a guide line. All values y are positive and since the parabola is sym-
metric to the y-axis, so is the Witch.
Consider
y- (x-1)(x-2)' 1
and let yl = (x - 1) (x - 2) so that y = 1/yl. The graph of yj _

(x - 1) (x - 2) is a parabola with yi-intercept 2 and x-intercepts 1 and 2.
The line x = 2 is a line of symmetry of the parabola and the required curve.
Fig. I.20
Note that, as x nears 1 or 2, yi nears 0 and y increases numerically with-

out bound. Thus x = 1 and x = 2 are guide lines.
Consider the Serpentine
x
y
xt + 1'
and let
s
yi
z 1 = x -I-, so that y = yl.
The graph of yi = x + (1 /x) is formed by combining as in (A) the

ordinates of the line and hyperbola shown. Its guide lines are x = 0 and
y = X.
As x increases to the right (and left), y grows nearer 0 and the Serpen-
tine thus has y = 0 as a guide line. It is tangent to the line y = x at the
origin.
Following the method of Par. 1.4 (2) and (3), we locate the maximum
and minimum points. Writing
x, X2
y= x2 +x2+1'
-
we have
xl x2
xi+x2+ = 0
or
xix2 + x1 - x21x2 - x2
- = U.
(xi + 1) (x2 + 1)
The numerator here is the zero factor, thus
(x, - x2) (1 - x,x2) = 0,
and, for xi x2,
1 - xix2 = 0.
xi
As --+ x, this approaches 1 - x2 = 0, from which x 1, -1, the critical
x2
values. Maximum and minimum values of y are

yat = #, Y. = - I.
EXERCISES
13. Use reciprocal ordinates to sketch each of the following.
Determine maximum and minimum points where necessary.
1 _ 1
(a) y = x2 + 4 (b)
x2 - 4
2 2x
(e) v = (d) v =
(x + 1)(x - 2) x2 + 9
x
(e) y = 9 (f) y = (x + 1)2
x2
1.7. Graphs, Polar

In sketching curves with equations given in polar coordinates, we list
values for the angle 0 in order of magnitude with corresponding values r.
The smooth curve through these points is, in general, the locus.
It should be recalled here that the trigonometric functions are periodic;
that is,
sin (B + 2ir) = sin B, cos (B + 21r) = cos B, tan (B + ir) = tan 8.
Moreover,
sin (ir - 0) = sin B, sin (-B) _ - sin B,
and
cos (-,r - 8) = - cos 0, cos (-B) = cos B.
Thus, an equation involving only the cosine function represents a curve
symmetric to the polar axis; only the sine function represents a curve
symmetric to the line perpendicular to the polar axis at the pole.
The following examples illustrate these and other features.
(A) Consider
r=4cos(0-:30°)
We locate the point D: (4, 30°), and let P: (r, 9) be representative of all
points of the locus. Then, from Fig. 1.22, . DOP = 9 - 30°. Since the
given equation requires that
r
cos (0 - 30°) = 4,
it is necessary that r (= OP) be a leg of a right triangle having the seg-

ment OD (=4) as hypotenuse. Accordingly, P is the right-angle vertex of
triangles with the fixed hypotenuse OD. The locus is thus a circle through
the pole with radius 2 and center (2, 30°).
Fig. 1.22 Fig. 1.23

Consider the form
r = a cos 0 + b sin 0
and let P(r, 0) be representative of all points of the locus.

If we multiply and divide the right member by 1/a2 + b2,
a
r= a2 }
__ )2 cos 9 + - - az +22 sin B
;
b2 [-\/-aT
we may set
a b
= COS a, a2 + b2 = sin a
a2+ )2
as shown.
Thus, r = -\/a2 + b2 (cos a cos 0 + sin a sin 0)
or
r=1/a2+b2cos(0-a).
This is the form just discussed. The equation thus represents the circle
through the pole with center ( -\/a2 + b2, a). Note that intercepts on the
Fig. 1.24
0° and 90° rays are a and b, respectively. Note further that it is quite
evident that the maximum value of
acos0+bsin0
is
1/a2 + b2
THE REAL NUMBER SYSTEM-GRAPHS Sec. 1.7
22
Fig. 1.25
(B) Consider the Limaeon
r=2+cos0
and let rl = 2, r2 = cos 0 so that r = rl + r2. We draw the circle r2 = cos 0
as shown and, for each point, (r2, 0), add to the radius vector r2 the vector
r, = 2 units. Note particularly that the addition is algebraic; that when
r2 is positive, r, + r2 is an extension. However, when r2 is negative,
(a/2 < B < 37r/2), the vectors rl and r2 are oppositely directed. In the
figure, for example, if 0 = 4 X OP = 120°, r2 = OQ = -12 and r = r, +
r2 = 2 - I = z = OQ + QP. The curve is oval-shaped.
(C) Consider the Parabola
I
r =
1 + cos 0
We employ an idea similar to that of reciprocal ordinates discussed in Par.

1.6(B). We first draw the Cardioid
r=1+cos0
after the fashion it was drawn in Example (B) just preceding. Then for
each 0 construct r so that
r=-r,.
1
900
Fig. 1.26 Fig. 1.27
It is advisable to draw the unit circle with center 0 to help determine

reciprocals of the numbers r1.
(D) Consider the Rose
r = sin 36
As 0 varies from 0 to 30° to 60°, r varies continuously from 0 to 1 to 0,

and corresponding points form the "leaf" symmetric to the line B =30°. For
60° < B < 120°, r is negative and the leaf symmetric to the line 0 = 90°
is formed. Repetition occurs for 0 > 180°.
EXERCISES
In the following, after checking for symmetry, etc., make use of the tech-
niques of addition of radii vectors and of reciprocal radii. Where necessary, tabulate
values of (r, 0). Use polar coordinate graph paper if available.
14. Draw the circles
(a) r = 4 sin B (b) r = 4 cos B (c) r = -4 sin B
(d) r=-4 cos 0 (e) r = 3 sin B+ 4 cos 0
(f) r2 = r sin 0 + 2 (g) r = 3 sin 0 - 4 cos 0
(h) r2 = r cos 0 - 2 (i) r + sin 0 - cos 0 = 0
15. Sketch the Limafons
(a) r = 1 + sin 0 (b) r = 1 - cos B (c) r = 1 - 2 sin 0
(d) r = 1+2cos0 (e) r = 2+sin0 (f) r = 2 - cos9
16. Sketch the Conics (P: parabola; E: ellipse; H: hyperbola)

(a) P: r(1 - cos 0) = 1 (b) P: r(1 - sin B) = 1
(c) P: r(1 + sin 0) = 1 (d) P:2r(1+cos0)=3
(e) E: r(2 + sin B) = I (f) E: r(2 - sin B) = 2
(g) E: r(2 + cos B) = 5 (h) E: r(3 - cos B) = 6
(i) H: r(1 + 2 sin 0) = 1 (j)H:r(1-2sin8)=2
(k) H: r(1 + 2 cos B) = 5 (1) H: r(1 - 3 cos B) = 6
0
(m) r = sect 0 (n) r = csc2
2
17. Sketch the Roses

(a) r = sin 8 (b) r = sin 20 (4 leaves)
(c) r = cos 30 (3 leaves) (d) r = cos 50 (5 leaves)
18. Using radian measure and the interval -2ir 5 0 5 2w, sketch the Spirals
(a) r = 20 (Archimedean) (b) r8 = 1 (reciprocal)
(c) r = 02 (d) r2 = 0 (parabolic)
(e) r20 = I (Lituus) (f) r = 1 + 0 (Archimedean)
(g) r = 20 (equiangular) (h) r = 92-e (equiangular)
2.
Measures
In the first chapter we have discussed, to some extent, the formation of

graphs or loci of given mathematical relations. The converse problem-
that of determining the equation or relation of a specified locus-is equally
important. To deal effectively with this problem, we must introduce such
concepts as distance, direction numbers, slope, and area, then determine
mathematical expressions for these concepts. This is the subject of this
chapter.
2.1. Distance. Midpoint of a Segment

Distance. Consider two points P,: (xi, yi) and P2: (x2, y2). The pro-
jections of the directed segment PP2 on the coordinate axes are the di-
rected lengths x2 - xl and y2 - yi as shown. By the right-triangle law (the
theorem of Pythagoras)
d = I P1P2 I = '\/ x2-XI 2 + (y2 - y 2
is the length of P,P2.

Midpoint. Let M: (x, y) be the midpoint of the segment P1P2. Then
25
26 MEASURES Sec. 2.1
Fig. 2.1
triangles HP1M and KMP2 are congruent, and thus
x2 - x = x - x1 or x = (x1 + x2)/2
y2-y=y-y1 or y = (y1 + y2)/2
2.2. Direction Cosines. Direction Numbers

Let a, p be angles formed by the coordinate axes and the segment P1P2
(or its extension) as shown. To standardize procedures and relations, we
shall measure a positively from the x-axis to the segment so that
Fig. 2.2
05 a<I800
Its cosine may then be positive or negative. The angle,6 will be acute and be
Sec. 2.2 MEASURES 27
measured positively or negatively from the y-axis to the line segment. That is,
-900 < 0 < +90°

I
and its cosine is thus always positive.

These direction cosines may be obtained from the coordinates of the
end points of the segment and its length
Cos a =
x2 - xi
cosh = y2-y'
d ' d
with y2 - yi, chosen positive.

If we form the sum of their squares
(x2 - x1)2 + (y2 - y1)2
cost a + Cos2 _ d2
we find the important relation
cost a + cost $3 = 1
Let two numbers A, µ, called direction numbers, be proportional to the

direction cosines of PiP2; i.e.,
A = k cos a, kcos$3, k P6 0.
For various values of k, the factor of proportionality, there are many such
pairs of these numbers. For all, however, there corresponds but one set of
direction cosines. This is evident if we sum their squares:
A2 + µ2 = k2(cos2a + Cos2N) = k2,
from which
A µ
COS a =
A2 + µ2' COS =
µ2
where the sign of the radical (the proportionality factor k) is chosen as

that of k to conform to the stipulation that (3 be acute. Then, since direc-
tion cosines may be obtained from direction numbers, the latter properly
establish the sense and direction of a line segment. Thus, for example, if
X = 3, µ=-4, then k2=(3)2+(-4)2=25
and
COS a = - 5i COS S = +6,
the sum of whose squares is unity. Any other pair, proportional to [3, -4]
such as [z, -2] or [6, -8] or [1, - ] leads to the same set of direction
cosines. It is important to note that the differences x2 - x1 and y2 - yl of
coordinates of any two points on a segment are direction numbers of the
segment.
EXERCISES
1. Sketch each segment whose end points are given and find its (1) length,
(2) midpoint, (3) direction numbers and cosines.
(a) (1, 2), (3, 6) (b) (-1, 2), (3, - 6)
(c) (0, 5), (8, 5) (d) (0, 0), (a, b)
(e) (-a, - b), (a, b) (f) (1, 5), (1, 1)
(g) (-1, - 3), (-1, 7) (h) (4, - 5), (0, 5)
2.3. Angle Between Two Line Segments

Consider two lines with direction angles al, #1 and a2, #2 respectively;
and let 0 be an angle they form with
each other. We wish to determine this
angle in terms of the directions of the
II
lines.
In order to minimize algebraic
manipulation, we draw two lines
through the origin parallel to the
Q(x2,y21
lnes consl; Aere d , th us 41- rming the
11
same angle 0. Then select a point on

each line, P(xi, y1) and Q(x2, y2), and
Fig. 2.3 let r1, r2 represent the distances OP
and OQ. Then by the law of cosines,
(PQ) 2 = (OP)2 + (OQ) 2 - 2(OP) (OQ) cos 8
or
(xl - x2)2 + (y1 - Y2)2 = (X12 + yl) -I- (x2 + y,) - 2r1r2 cos B.
Reduction gives
=x1.22 + y1 y2
cos 9
r1 r2 r1 r2
Now, the terms in the right member here are direction cosines of the lines
OP and OQ and also of the given lines; i.e.
X1 y1 X2 Y2
- = cos a1, = cos,61 and - = cos a2, = cos 02
ri r1 r2 r2
Accordingly,
cos 0 = Cos a1 COS a2 + cos 01 cos 132
The particular product-sum form of the right member will be referred to

as the scalar product of the pairs [ cos a,, cos,6,] and [ cos a2, cos 02].
If [A,, µ,] and [A2, 42] are direction numbers of given lines, then A, _
k cos a,, etc., so that
A1X2 + 111/12
cos 6 = %,/Ki
µi 112
2.4. Parallel and Perpendicular Line Segments

It is at once evident that if two line segments are parallel their direction
angles as well as their direction cosines are equal and their direction num-
bers are proportional; and conversely so.
If two lines are perpendicular, 0 = 90°, cos 0 = 0 and
Cos a, Cos a2 + cos 16, cos 02 = 0
That is, the scalar product of their direction cosines is 0. Conversely, if this
scalar product is 0, then cos 0 = 0 and 0 is 90° or -90°; thus the lines are
perpendicular.
Moreover, since
A1X2 + µ1F12
cos0 = ,
-I/T2 +112
then, if 0 = 90°,
A1A2 + µ1µ2 = ,
and, if this is 0, 0 = 90° or -90°. Thus, two lines are perpendicular if the
scalar product of their direction numbers is 0, and conversely.
MEASURES Sec. 2.5
30
2.5. Slope
The angle a is designated as the inclination of P1P2 and tan a as the
slope of the segment. We have, if x1 X x2;
Y2-y1 CosB_µ
tan a
X2 - x1 Cos a
For two lines with slopes
Cos 01 cos 02
tan a, _ tan a2 =
COS al COS a2
to be perpendicular, then from the preceding paragraph,
COS a, COS a2 + COS 01 COS 02 = 0

or
cos Cos a2
cos al cos fit'
1
tall a,
tan a2
Thus if two lines are perpendicular, their slopes are negative reciprocals.*
(The student should now state and prove the converse.)
EXERCISES
2. Find the cosine of the angle between the following pairs of segments of
Problem 1
(a), (b) (c), (d) (e), (f) (g), (f)
3. Prove that if tan a1 = -1/tan a2, then the lines with inclinations al and
a2 are perpendicular.
4. Calculate the slope of a line segment perpendicular to each of the segments
of Problem 1.
5. By finding direction numbers of each of the following pairs of segments,
show they are perpendicular
(a) (1, 1), (4, 3) and (2, 3), (4, 0)
(b) (7, 5), (-5, -3) and (-2, 9), (-4, 12)
6. Find the slope of a line segment through (-1, 2) which is (a) parallel,
(b) perpendicular to the segment joining (1, 6) and (-2, -3).
' That is, if they are not parallel to either coordinate axis.
2.6. Collinearity
We have now a means of testing collinearity of three (or more) points;
i.e., of determining whether or not all three lie on the same line.
If P1: (x1, y1), P2: (x2, y2), P3: (x3, ys) are collinear, the slopes of P,P2,
P1P3, and -2P3 are equal, and conversely. That is (x1 96 x2 /- X3);
Y2-Y1 Y3-y1 y3-Yz

X2 - x1 x3 - x1 x3 - X2
Note that any one of the three equations is a linear combination of the
other two; i.e., collinearity is established if any two of the slopes of the three
directed segments P;P; are equal.
2.7. Area of Triangle Formed by P1P2P3

The area A of triangle P1P2P3 may be measured as follows. Draw
parallel to the coordinate axes the sides of the rectangle circumscribing
the triangle. The area of this rectangle less the sum of its triangular areas
exterior to P1P2P3 is A. Thus
A = (x3 - x1) (y3 - y2) - z E (x2 - x1) (y1 - y2) + (xs - x1) (y3 - y1)
+ (x2 - x2) (y3 - y2) ]
a(- xsyz + x321 - x221 + x223 - x123 + x122)
Ix1 y1 11
I
W
X2 y2 1
Y3-Y1
Ix3 y3 11
(The algebraic sign of this numerical Y3-Y2

measure is unimportant. It will be Y1-Y2
positive if Pi, P2, P3 are selected in
counterclockwise order.) Note that if x3-x2
PiP2Ps are collinear, this determinant I

is 0, and conversely. Thus here is a
second device for testing collinearity Fig. 2.4
of points.
EXERCISES
7. Draw and determine the area of each triangle whose vertices are the triples:
(a) (0, 0), (0, 1), (6, 0) (b) (0, 0), (0, 1), (6, 2)
(c) (8, 9), (- 6, 1), (0, - 5) (d) (-8, 9), (6, 1), (5, 0)
8. Show that the triangle given in 7(c) is isosceles.

9. Show that (8, -1), (- 6, 1), (2, -7) forma right triangle. Do this in two
ways.
10. For the right triangle of Problem 9 show that the length of the segment
from the midpoint of the hypotenuse to the right angle vertex is half the
hypotenuse.
11. Given the quadrilateral (-2a, 0), (0, 2b), (2c, 0), (2d, -2e), show that the
midpoints of the sides form a parallelogram. (Build a four-bar linkage
with elastic thread through the midpoints of the sides for demonstration.)
12. Given the triangle (- a, 0), (0, b), (c, 0), show that the segment joining
the midpoints of two sides is parallel to and half the third side.
13. Find the area of each of the following quadrilaterals.
(a) (1, 2), (-4, 1), (-2, -2), (6, 0)
(b) (-1, 2), (0, - 6), (4, - 2), (3, 2)
14. Using the determinant form for area, show that each of the following
triples of points are collinear, then check by calculating slopes.
(a) (0, 4), (6, 0), (12, -4) (b) (a, b + a), (-b, 0), (-a, b - a)
15. Using alphabet letters for coordinates of vertices, prove that the diagonals
of any rhombus are perpendicular.
3.
Variables, Functions,
Limits, Continuity
3.1. Variable
Given a set of numbers, if x represents any number of the set, x is a
variable.
3.2. Function
Consider two sets of real numbers. (These sets may contain some or all
the real numbers and the two sets may have numbers in common.) Let x
and y represent, respectively, any numbers in the two sets. If to each num-
ber x there is paired one, and only one, y, then y is the value of a function, f,
of x. This we write as y = f (x).
We have already had examples of functions in the equations of preceding
exercises (e.g., 2y + 3x = 6; r = sin B). However, functions may exist
without equations or relationships explicitly given. Thus the table
x -1 3 5
y 0 I 7 I 11
defines y as values of a function of x or x as values of a function of y.

33
34 VARIABLES, FUNCTIONS, LIMITS, CONTINUITY Sec. 3.3
3.3. Notation
Familiarity with the symbol f ( ) which is used to represent a function
is essential in all that follows. For example, a function f ( ) may be given
by the mathematical form
f()=()2+(1)-2.
The unfilled parentheses here are to contain any of the numbers in the set x
and the indicated operations then performed. The results yield the function
values. Thus, if the set x contains all real numbers except 0, we have for the
example :
f(-2) = (-2)2 + 1/(-2) - 2 = 4
f(5) = (5)2 + 1/(5) - 2 = 23.2
f(1/z) = (I/Z)2 + z - 2
f(1+k) =k2+2k+1/(1+k) -1, k -1
This is a precise functional relationship which yields function values for all
values in the set x.
The exclusion of the value 0 for x is demanded since 1/0 is meaningless. To
clarify this, consider 12/3 = N, a statement which has meaning if, and only if,
3N = 12. N thus is 4. With the form 12/0 = N there is meaning if, and only
if, there is a number N, such that 12. There is no number N whose
product by 0 is 12. Thus
a/0 is no number, a 5-6 0.
The inverse relation 0/12 = N or 12N = 0 provides the unique result N = 0.
But the peculiar form 0/0 = N or 0 N = 0 may yield any number. The form
does not represent a defined unique number; thus
0/0 is undefined.
3.4. Limits. The Symbol ->

We wish at times to let x represent any number in a set near a specified
number a such that the difference* I x - a I can be made as near 0 as we
please for every x in the set. This is indicated by the symbol
x ->a.
Now as x -- a (read "x approaches a") it may happen for a given f ( )
* The symbol I x - a I is read as "the absolute value of (x - a)." It signifies the
signless number. Thus 3 - 5 1 = 2.
Sec. 3.4 VARIABLES, FUNCTIONS, LIMITS, CONTINUITY 35
that the function value f (x) approaches a constant L. That is,

as x ---, a, then f (x) -> L,
so that the value of the difference
If(x) - LI
becomes and remains arbitrarily near 0. In this event, we say that L is
the limit of f (x) as x approaches a. This concept is contained in the symbol
Limit f (x) = L.
x -+a
To illustrate, consider
f(x)=
x2-x-6
x-3
and let x -> 3. We list tables of values to give numerical evidence that
f(x)->5.
x 0 1 2 2.5 2.9 2.99 2.999 ...->3
f(x) 2 3 4 4.5 4.9 4.99 4.999
x 6 5 1 4 3.5 3.1 3.01 3.001
f(x) 11 8 1 7 1 6 1 5.5 1 5.1 j 5.01 1 5.001 J ... -* 5

[The student should check these values for self-assurance and realize that
no numerical array such as this can prove that f (x) -> 5].
Let us take a different approach to the problem for mathematical
certainty by setting
x=3+Ax
where the symbol Ax (read "delta x") is a quantity that will be made
arbitrarily near 0; thus, as Ox - 0, then x -i 3.
Setting 3 + Ox in the function
( )2- ( ) -6
f( )
( ) -3
we have
_ (3 +x)2 - (3 + Ax) - 6
f(3+Ax) (3+Ax) -3
9+6(Ox) + (Ax)2- 3- (Ax) - 6
3+Ax-3
5(Ax) + (Ax)2
Thus
f(3+Ax) = 5 + (Ax), Ax 0.
With this last form, it becomes conclusively evident that
as AX -*0, then and f(x) ->5.
In symbolic form, we write
Limit f (x) = 5.
X-3
Note particularly, however, that

f (3) = 0/0 is undefined.
3.5. Geometric Progressions

The familiar geometric progression,
Sn = a + ar + ar2 + + arn'',
where Sn indicates the sum of the first n terms, may shed further light on
the concept of limit. We multiply the expression through by the ratio r,
r S = ar+arz+ar3+ +arn,
and subtract. Then
(1 - r) Sn = a - arn
or, if r 34 1,
a a
Sn = rn
I - r 1 - r
Now, if r is numerically less than unity, i.e., I r I < 1, then as the number of
terms increases; i.e., as n grows arbitrarily large:
a
rn -* 0 and S. -, 1-r
Formally,
a
Limit S _ Ir l < I
n- co 1-r
where n -+ ao is to be read: "as n increases without bound." To illustrate,

consider the unending repeating decimal 0.454545 . For the terminating
decimal of n groups 45, we write

Sn = 45(.01) + 45(.01)2 + ... + 45(.01)
Multiplying by the ratio (.01),
(.01) Sn = 45(.01)2 + (.01) 3 + ... + 45(.01)"+1,
and subtracting,
(.99)Sn = 45(.01) - 45(.01)°+'.
Thus
S. = fl. - Jj(0.1)n.
Asn - oo, (.01)n- *0 and S.-+ 4 5 .
Let the student now reverse the process by dividing 45 by 99.
Consider another important geometric progression which we treat in
the same manner.
Sn = 1 +x+x2 + x3 + ... +xn-1
X Sn = x + x2 + a3 + x4 + + xn
(1-x)Sn=1-xn,
and, if I x I < 1,
Sn -> as n -> oo
1 -1 x
We shall have occasion to refer to this result in later sections. Let the stu-
dent now divide 1 by (1 - x) through "long" division justifiable only if
I x I <1.
3.6. A Special Limit

One of the most important and fundamental limits is
sin B
Limit =ly measured in radians
0 -- 0 B '
To establish this, let 0 be a central

angle AOP in a circle of radius
OA = 1, shown in Fig. 3.1. Draw
PMQ perpendicular to OA, and RT
tangent to the circle at A. It should
be recalled that if 0 is measured in
radians, then the area of a circular
sector of radius r and central angle
0 is Zr29. The area of the sector OAP
is thus B. For this area and those of
triangles OQP and ORT it is ap-
parent for 0 < 0 < 7r/2 that
OQMP< OQAP < ORAT. Now
since OM = cos 0, QP = 2 sin 0,
and RT = 2 tan 0, this inequality
Fig. 3.1
may be written as
sin 0 cos 0 < 0 < tan B.
Dividing by sin 0,
0
cos0 < -<
sin 0 cos 0
Inverting,
1 sin 9
> > cos 0.
cos B 0
These values are arranged on a line about the unit point as shown.
sin 9 e
e sin e
cos 9
Cos e
Now, as 0 -> 0, cos B --> 1 and (1/cos 0) -* 1. Since (sin 0)/0 and
(0/sin B) are always intermediate in value between cos 0 and (1/ cos 0), then
9 sin 0
-->1 and
sin 8 0
A serviceable corollary is
Limit 1-cos9=0
0
0 0
This becomes evident on replacing 1 - cos 0 by 2 sins (0/2) ; thus
1 - cos 0 = 2 sin (8/2) -sin (0/2) = sin_(0/2).sin

(0/2).
0 8 8/2
As 0 ---> 0, the factors in the last expression approach 1 and 0, respectively,

and the limit is established.
3.7. Limit Theorems

A covering theorem concerning limits is stated here without proof.
This is necessary for the operation called "differentiation"-the subject
of the next chapter.
If for two functions f (x) and g(x),
Limit f (x) = A and Limit g(x) = B.
z- a xa
Then
Limit [ f (x) + g(x) ] = A + B;

x -.a
Limit
x-a
ff(x) A
Limit B 5-6 0
X -, a g (x) J B
To illustrate in the order given:

(1) Limit (x2 + 3x) = (Limit x2) + (Limit 3x) = 4 + 6 = 10
x--.2 x-*2 x-+2
(2) Limit (Limit x2) (Limit/) = 16.2 = 32
x -+ 4 x --+ 4 x --' 4
x2 + 3x Limit (x2 + 3x) 4

(3) Limit 3
x ....1 x+2 Limit (x + 2) 3
3.8. Continuity
A concept of fundamental importance concerns the behavior of a func-
tion for values of the variable near a number a. Consider f (x) and the num-
* This includes the form Limit k g (x) - k Limit g(x).
ber x = a. For this number, there are the several possibilities:
f (a) may or may not exist (i.e., be a number),
and/or
Limit f (x) may or may not exist.
z-'a
Definition. If f(a) and Limit f(x) both exist and
x ---, a
Limit f (x) = f (a)

x a
the function is continuous at x = a.

The graph of the function contains the point with coordinates x = a,
y = f (a), and points arbitrarily near this point. Graphs of functions
discontinuous at some x = a are broken.
Intuitive notions of continuity and discontinuity were presented
graphically in Par. 1.5 where the matter of guide lines was discussed. The
student should review those illustrations now in the light of this formality.
Several items of interest here must be noted:
(1) A polynomial function f(x),
.f (x) = anx^ + .....1.. aix + ao
wherein the coefficients a, are real numbers (not all 0) and n is a positive
integer, is continuous for all real numbers x. Graphs of all such functions
are unbroken, and for each x there is one, and only one, y.
(2) The rational fractional function,
f(x) = P(x)
Q(x)
is continuous for all real numbers x except those for which Q(x) = 0.
* IP(x) and Q(x) are polynomials]
For example,
x2 - 2x x2 - 4
Y= y
x 1 x - 2
Fig. S.2 Fig. 3.3

has no point for which has no point for which
x=1 x=2
We list here without proof a set of theorems concerning continuity of
functions f (x) and g(x) that are necessary for work to follow.
If f(x) and g(x) are functions continuous in an interval, then in that
interval
(1) f(x) + g(x) is continuous
(2)
f(x) is continuous except where g(x) = 0

g (Z)
EXERCISES
1. (a) Given f(x) = x2 + 4x - 2.
Find f(-2), f(-1), f(0), f(1/x), fl(x - 1)/xJ.
Find x if f (x) _ - 2; if f (x) = 1.
(b) Given f(x) = x2 - 5x.
Find f(- J), f(0), f(5), f(x + 2), f[(1 - x)/x].
Find x if f (x) = 4; if f (x) = 9/4.
nZ - 5n + 6
2. (a) Given f(n) _
n- 2
Find f(1), f(2), f(3), f(4), Limit f(n).
n-+ 2
n2 + n - 2
(b) Given f(n)
n- 1
Find f(-3), f(-2), f(0), f(1), Limit f(n).
n --- 1
3. Calculate the sum of n terms and the limit as n --> co of the geometric
series:
+;+}+... ,
(a)
(1)) 1 _ 2+ 4 s+ .. .
4. Represent as the quotient of two integers the repeating decimal:
(a) 1.010101 .
(b) 2.123123123
5. Show that
sin B - B 1 - cos B
(a) Limit - =0 (b) Limit =
0--+ 0 0 0-0 02
6. Using a table of sines and cosines, draw in rectangular coordinates the

graphs of
(a) y = 2 sin x (b) y = sin 2x (c) y = sin (x/2) (d) y = 1- cos x.
7. Draw and discuss the graph of
x2 - 4
(a) y= x- 2 (= x + 2, x 0 2).
x2-5x+6 (= x - 2, z 0 3).
(b) y
x- 3
8. Draw the region defined by
(a) {r = sin B, r S cos 01 (b) {r < sin B, r sin 9
9. Draw(( the graph of
1
(a) 1r sin6+rcos06 4,050= 34 }
341
(b) rr= 1,-6595
10. Draw the graph of

sin x 1- cos x
(b)
Zr Zr
11. Prove that Limit (1 - cos n0)/B = 0 (Hint: Multiply numerator and de-
0-->0
nominator by n.)
12. Prove that Limit (sin k0)/0 = k. (Hint: Multiply numerator and denom-
0 0
inator by k.)
13. Given Sn = [1 - (1/22) + (1/23) - (1/24) + f (1/24)] 0. Show that
as n --> co, S. --> - 0. Then devise a bisection process to approximate the
trisection of angle 0.
4.
The Derivative
We introduce here the derivative concept, one of the most important

and certainly most useful ideas that ever occurred to man. First, the method
of finding a derivative is presented; then some interpretations and applica-
tions are given. We begin with an example.
4.1. The Derivative in Rectangular Coordinates

Consider
y=x2-x-6,
where the domain of x is the set of all real numbers. We shall take x to-
gether with its paired value y as particular but unspecified numbers. We
add to x an arbitrary number Ax. This changes the value of the function by
a determinate number Ay. Thus
y+Ay = (x+ax)2- (x+Ax) -6
or
y + Ay = x2 + 2x(Ox) + (0x)2 - x - (Ax) - 6
Since y = x2 - x - 6, we may eliminate these terms and so have
Ay = (2x - 1) (Ox) + (,IX) 2.
Dividing by Ax,
-4x= 2x - 1 + (Ax).
44
Sec. 4.1 THE DERIVATIVE 45
Now, as Ax --> 0, but remains different from 0, then Ay --' 0, and
-Ox-, 2x - 1
or
AY
Limit = 2x - 1
AX-0 Ox
The expression (2x - 1) is the derivative of the function x' - x - 6,

and the process involved in finding this derivative is called differentiation.
Definition. Formally, if y = f (x), the derivative is defined by either of
the equivalent expressions
Ay + ox) - f (x)
Limit or Limit f (x
ox-> 0 Ox AX -- o AX
The derivative is a new function derived from y, and we shall represent

it by the symbol Day, or f'(x). At times we may omit the subscript if there
is no need to call particular attention to the independent variable.
Note that the defining forms just given contain explicit directions for
finding the derivative. The procedure, called the "delta process," consists of
the following four steps:
(1) Replace x, y by x + Ax, y + Ay, respectively.
(2) Subtract the original equation, and simplify.
(3) Divide by Ox.
(4) Write the limit approached by each term.
For a second example, consider
xy-y-4=0.
Here the variable y is the value of a function of x given in implicit form.
Explicitly,
4
x 0 1,
y =
x- 1'
which for all real numbers, except x = 1, gives numbers y. Thus, assured
that y is indeed a function of x, x 0 1, we apply the delta process to the
THE DERIVATIVE Sec. 4.1
46
original equation. Thus

(1) (x+Ax)(y+Ay) - (y+Dy) -4=0,
or
xy + x(Ay) + y(Ax) + (Ax) (Ay) - y - (Ay) - 4 = 0.
(2) x(Ay) + y(Ax) + (Ox) (Ay) - (Ay) = 0

oy Ay
(3) x-+y+
O:r
(oy) - Ax
=0
1 1 1 1
(4)
Thus
x0 1.
x
EXERCISES
Using the delta process, find Dry:
1. xy + x2 = 1 2. xy- x2= 1
3. x2y - 4 = 0 4.xy2-4=0
5. y = Vx- 6. y
I
7. y = 8. Y
x-3 x+4
9. y2+x2+x+2y= 0 10.y2-x2+x+2y=0
4.2. Slope of a Curve y =f (x)
We display here the geometry involved in the process of finding Dxy
and give interpretations of utmost importance. Consider
y = x2
the parabola shown. Let P(x, y) be some particular but unspecified point
on the curve. Then (x + Ax, y + Ay) is a neighboring point Q. Since Q
also lies on the curve,
y + Ay = (x + Ax)2 = x2 + 2x(Ax) + (Ax)2
or, with y = x',
AY = 2x(Ax) + (Ax)'.
THE DERIVATIVE 47
Sec. 4.2
Then
y = 2x + (Ox).
Ox
Now if Ax - 0, then Ay - 0 and

Q follows the curve toward P.
Throughout this phase, Ay/Ax is
the slope of the variable line through
the fixed point P and the positions
Qi, Q2, Qa, . - - of Q. But as Ax -' 0,
Ay Fig. 4.1
- 2x(=Dsy).
Ox
Thus, as Q approaches the position of P, the slope of the line PQ approaches
the value of the derivative Dxy at P. It seems reasonable, then, to make the
following definition
Dzy is the slope of the tangent line at (x, y)
or, the line through (x, y) with slope Dsy is tangent to the curve
y = f(x)
4.3. Maximum-minimum in Rectangular Coordinates*
A curve f (x, y) = 0 may have points at which Dxy = 0 where the tan-
gent lines are parallel to the x-axis. t These are called critical points.
Y
Dxy=O
DxY,.
%Y=+
OxY=+
DxYz0
.x
xZ
Fig. 4.2
The student should review here the material given previously in Par. 1.4.
t Or at which Dsy fails to exist.
THE DERIVATIVE Sec. 4.3
48
A critical point is a "high"

point if the tangent turns through
it from positive to negative slope; a
"low" pointif from negative topositive.
Such critical points locate rela-
tive maxima and minima values of
y and offer valuable information in
curve sketching. For example,
y= x3-3x2+2
(2,-2)
Fig. 4.3 with
Dzy = 3x2 - Gx
has critical points where Dxy = 3x(x - 2) = 0; i.e., at x = 0, 2. These
points are (0, 2) and (2, -2). The accompanying table helps disclose the
nature of these critical points.
x 0 i 1 2 4
2 2
Y
DTy 0 - I 0
11
I max I min
I Note that any value of x < 0, any

value between x = 0 and x = 2, and
any x > 2 would serve to determine
the necessary signs of Day.
If the tangent turns through a
point at which D2y = 0 from posi-
tive to positive (or negative to
negative) slope, the point is in
general a flex point. In such regions,
Fig. 4.4 the curve is "S"-shaped as shown.
EXERCISES
Find the slope at (x, y) of each of the following curves and coordinates of
their maximum, and minimum points, if any. Draw the graph of each
equation.
11.y=x3-4x 12.y=4-6x-x2
13. y = x3- 8 14. ,y= x'-- 8x
15.y=- 1
1 16 .y=x2+1
17. x2y + 4y - 8 = 0 18. (1 + x)y2 = x2(3 - x)
19. x2 + 9y2 - 4x = 0 20. 9x2 + y2 - 4y = 0
4.4. The Derivative in Polar Coordinates

Let us apply the delta process to the relation r = f (8) and then seek a
geometric interpretation of the derivative Der. Consider the Cardioid
r=1+sin0
We have
(1) r+Ar=l+sin(B+A8)
(2) Ar = sin (8 + AB) - sin 8
= cos 0 sin (AB) + sin 0 cos (AB) - sin 8
cos 9 sin (A9) - sin 0 [1 - cos (A0) ]
Ar sin (AB) 1 - cos (A8)
(3)
AB
= cos 0
AB
- sin 0
AB
(4) Der = (cos 0) (1) - (sin 0) (0). (See 3.6).
Thus,
I Der = cos 9
Consider now the general situation, given r = f(0). Let P be the point
(r, 0) and Q the point (r + Ar, 0 + AO); let 4 represent the angle made by
the radius vector with the tangent line to the curve at P. Draw PA per-
pendicular to OQ. Then
PA = r sin (A0),
and
AQ = (r + Ar) - r cos (A0)

and
PA = r sin (AB)
tan (4Z AQP) =
AQ Ar + r[1 - cos (A9)]
50 THE DERIVATIVE Sec. 4.4
Fig. 4.5 Fig. 4.6

or, dividing numerator and denominator by (O9),
sin (M9)
O9
tan (4 AQP)
Ar 1 - cos (09)
O9{ AO
Now, as 09 ---> 0, AQP -> :/i, and thus,
r
tan ¢ =
Der*
If we draw a perpendicular to r through the pole, it meets the perpendicular

to the tangent at P in the point N and ONP Then the length ON
is Der.
Note this important difference between the derivatives in rectangular
and polar coordinates. While Der has an interpretation of length., Dxy is a
ratio of lengths.
Returning to the previous illus-
tration r = 1 + sin 0 with Der =
cos 0, we make use of this informa-
tion in the construction of the tan-
gent to the Cardioid at any point P.
Draw the circle r = sin 0 and its
chord ON perpendicular to OP. (N
lies on the diameter through H.) It
is Pvidpnt. thA.t OM = eoc A = Dor.
'1'nus the tangent at t- is the per-
pendicular to NP. Let the student Fig. 4.7
now show that P = 45° + 0/2 using the fundamental form tan 4, = r/Der.
The Derivative of cos 0. With r = cos 0,

(1) r+Ar=cos(0+AB)
= cos 0 cos (A8) - sin 0 sin (A8)
(2) Ar = -sin 0 sin (A8) + cos 8 cos (A8) - cos 0
Ar _ sin (A8) 1 - cos (A8)
(3) - - -sin 8 A8
- cos 8 -- AB
(4) Der = - (sin 0) (1) - (cos 8) (0)

or
Der = - sin 0
1
(The student should make special effort to remember the minus sign.)
The Derivative of sin ke. Consider
r = sin ko.
Then
r + Ar = sin k(6 + A8),
and
Ar = sin k(8 + A8) - sin k8
= sin k8 [cos I] + cos k8 sin
Thus
(k (k A8)
kB
A8 A8 A8
We supply the factor k in the fractions of the right member. Thus

Ar cos (k A8) - 1 sin (k AB)
ze
- kO kB
k.AB
and
1 1 1
Der = 0 + k cos k8
EXERCISES
Using the delta process, find Der and tan ¢. Sketch the ones marked *.
*21. r = sin 0 + cos 0 22. r2 = sin 0
23. is = cos 8 *24. r + r cos 8 = 1
*25. r - rsin8= 1 *26. r = I+cos6
*27. r = 3 + 2 cos 0 *28. r = sin 20
*29. r = cos 29 30. r = I + sin2 0

31. r = 1 - cost 0 *32. r(1 + 3 cos 0) = 3
*33. r(1 + 3 cos 0) = 2 34. r cos 0 = 2 sin2 0
35. r = tan 0 36. r = cot 0
*37. r2 = sin 20 *38. r-2 = cos 20
39. r coslt9 40. r = sin KO
4.5. Derivatives of uv, a/v, u"

In order to expedite the work of finding derivatives, we investigate
products, quotients, and powers of functions. These functions of x are repre-
sented here by the letters u and v and it is assumed that they themselves
have derivatives. Thus, given
y =uv; y=u/u, v-0; y=u",
we wish to find the form for Dy in each instance. Before doing so, note that
if x is changed by an arbitrary amount Ax, then u and v change by some
determinate amounts Au and AV, respectively. These changes cause y to
change by some amount Ay. Note also, since u, v, and y are continuous
functions of x, that as Ax -* 0, then Au --j 0, Av --+ 0, Ay -* 0, and
Au AV
Ay --> Dxy, - Dxu; - Dxv.
Ax Ax Ax
(A) y = uv
(1) Y + Ay = (u + Au) (v + AV) = uv + u(AV) + v(Au) + (Au) (AV)

(2) Ay = u(Av) + v(Au) + (Au) (AV)
AY _ AV Au AV
+ (Au) Ax
Ax - uAx
(3)
vAx
1 I 1 1
(4) Dxy = uDxv + vDxu + 0
Thus
Dxy = uDxv + vDxu
(B) y = u/v or vy = u, v 5,6 0.

From the foregoing result,
vDxy + yDv = Dxu.
Replacing y by u/v and solving for Day, we have
vDsu - uDsv
D=y = - -
(C) y = u" where n is a positive integer.

(1) y + Ay = (u + Au)" and, by the Binomial Theorem:
= u^ + nu"-'(Au) + n(n2! 1)u"-2(Au)2 + ... + (Au)
(2) Ay = nu"-'(Au) + (Au)2[ ]

Ay = nuri- Au Au
(3) (Au) Ax[ I
Ax Ax +
(4) D=y = nu"`'D=u
Note the important special case for u = x
y = x" D=y = nx"-'
The result for (C) holds for any constant n, rational or irrational, positive
or negative. We shall make free use of this fact but leave its proof to a course
in the Calculus.
Illustrations.
(A) y = (x2 + 3x - 1) (x4 + 7x + 5). [Form: y = uv].

Here let u = x2 + 3x - 1 with Dzu = 2x + 3
v =x4+7z+5with Dv =4x1+7.
Then
D.y = (x2+3x- 1)(4x1+7) + (x'+7x+5)(2x+3).
The student should experience the extra trouble in multiplying the paren-
thetical expressions of the given function and then differentiating.
2 u
(B) y = x + 1' x X -1. [Form: y = v
DZy _ (x + 1) (2x) - x2 _ x(x + 2)

(x + 1)2 (x + 1)2
(C) y = (x' + 1)1/2. [Form: y = u°]
3x2
D=y = '(x' + 1)-12(3x2) _ 2(x' + 1)1/2 '
(D) r = sin2O - cos 0.
Here sine 0 is in the form u2, so that
Der = 2( sin 9) cos B + sin B.
4.6. Differentiation of Implicit Functions

The differentiation formulas just developed are especially useful in the
differentiation of functions given implicitly. For example, consider
x2+xy+y2-y- 1 =0.
Here y represents values of two functions-those obtained by solving the
quadratic for y in terms of x. We shall regard y as representing the value of
either of the two functions.
Returning to the given equation, the term xy will be regarded as the
product of the values of two functions of x; the term y2 in the same spirit in
which we treated u^. The derivative then is
2x + xDy + y + 2yD=y - D=y = 0,
and
Dy - 2x + y
x+2y- 1
As another example, consider (the parabolic arc, Fig. 4.8) :
x1/2 + yl2 = 1.
Differentiating,
Jx-1/2 + 4y-u2DEy =0
or
Day = -,j
Consider the polar coordinate equa-
tion
r2 = sin 50 + 3 cos 0.
Then
2rDer = 5 cos 56 - 3 sin 0
or
Fig. 4.8
Der = 1( cos 50 - a sin 0).
EXERCISES
Using derivative forms for uv, u/v, u", find Day.
41. V-x - =1 42. x2i$ + y213 = at
43. x3 + y3 - 3xy = 0 44.x3+y3=x-y

45. x2y2 = x2 + y2 46. y2(2 - x) = x3
47.x2(2-y)=y3
x- y
48.y=x+y
49. y =
x+y
x2
50. x - 2xy+y=1
Using derivative forms for uv, u./v, u°, find Der.
51. r2= sin 0 + cos 0 52. Y1= cos 20
53. r(1 - cos B) = 1 54. r = I + sine 0
55. r = I + cost 0 56. r8 = 1
sin 6
57. r20 = 1 58. r = tan 0 =
cos 0
5.
Functions
Whose Derivatives
Are Known
The determination of the function whose derivative is known leads to

matters of fundamental importance. First we discuss details and start with
some examples.
5.1. Determination of the Function

(A) Given
D zy = 3x2 - 7x + 2,
then
y=xe-gx2+2x+C
a result justifiable by differentiation. The letter C represents any constant.
(B) Given
Der = cos2B,
which we rewrite as
Der = J+ 4 cos 2B,
56
Sec. 5.1 FUNCTIONS WHOSE DERIVATIVES ARE KNOWN 57
then
20 + 1 sin 20 -f- C.
r=1 2 2
We pause here to discuss uniqueness-to answer the question: "Given

the derivative, is there but one function with that derivative?" To consider
the matter, let f, ) and f2 ) be two functions whose derivatives are
equal; i.e.,
Df,(x) = Df2(x).
Then
DCJi (x) - f2(x)I = 0.
Now, the only function with zero derivative is a constant. Thus
.f, (x) - f2 (x) = C;

i.e., two functions with equal derivatives differ only by a constant. Unique-
ness, except for a possible additive constant, is thus established.
We turn now to the problem of finding the equation of a curve given
certain of its properties involving derivatives. For example,
(A) Find the equation of the curve with slope (1 - 2x) at any point
and which passes through (2, 1).
Since Dzy represents slope, we have
Dzy = 1 - 2x
and thus
y=x-x2+C.
Now (2, 1) lies on the curve and its coordinates must satisfy the equation.
Accordingly,
1 =2-4+C,
therefore, C = 3 and
y=x-x2+3,
the parabola shown.
Questions.
(1) Find the maximum y.
(2) Where is the slope equal to 3?
(3) Find the slopes at the three
points where the curve crosses the %
axes.
(4) Find the slopes at y = 2.
(5) What is the line of symmetry of
the curve? Fig. 5.1
58 FUNCTIONS WHOSE DERIVATIVES ARE KNOWN Sec. 5.1
(B) Find the equation of the curve through (ir, 1) such that 24, + 0 =
2ir at each of its points.
We have
r
Der = tan ,y = tan (a - 2 J
B\ - - tan
0 1 - cos 8
2 sin B
or
Der(1 - cos 0) + r sin 0 = 0.
On careful inspection this appears to be
De[r(1 - cos0)] = 0.
Thus
r(1 - cos0) = C.
But the coordinates (7r, 1) must satisfy this equation; i.e.,
(1) (1 - cos 7r) = C = 2.
Accordingly,
r(1 - cos 8) = 2
a parabola with focus at the pole.

Note that the initial line is the axis of symmetry of the curve
[ cos 0 = cos (- 0) ]. Let G be the intersection of the tangent with the axis
of symmetry. Then, sinceXFP =
9ZFGP + -CGPF, and 4GPF =
8/2, we have FGP = 0/2. If we
now draw a line through P parallel
to the axis of symmetry, this line
also makes 0/2 with the tangent.
Thus, light rays emanating from the
focus F are reflected by the curve
in lines parallel to the axis. This is
the reflective property of the para-
bola, a principle utilized in search-
Fig. 5.2
lights, sound reflectors, microscope
and telescope lenses, radar screens,
etc.*
* An interesting "triple play" on words is built upon the story of the three brothers
who inherited a cattle ranch. They named the ranch THE FOCUS because that was
where The Sons (Sun's) Raise (Rays) Meat (Meet).
EXERCISES
For each of the following, find the equation of the curve satisfying the given
conditions. Draw a sketch of each curve.
1. Dry= 4x;y= 9forx= 2
2. Dyy= 2x- l;y= 3forx= 0
3. Der = sin 8 - cos 8; r = -1 for 8 = 0
4. rDer = cos 28; r = I for 8 = 7r/2
5. Dry = x2/y; y = O for x = 0
6. Dry = x/y2; y= 0 for x= 0
7. Dzy= x/y; y = 2 for x = 1
8. Dry= -y/x;y= lforx= 1
9. Dry = (1 + x)/(1 + y); y = 4 for x = 2
10. Dry= (1- x)/(l+ y); y 3 for x = 2
11. Der = 4 cos 20; r = 2 for 0 = a/4
12. Der= -4sin26;r= 2,8= 0
13. Der=0;r=Ofor0=0
14. Der= 2;r= 2for0= 0
5.2. Area. Rectangular Coordinates
Consider the curve y = f (x) in the interval from x = a to x = b,
and let y > 0 in this interval be
single valued. Let the number A
represent the variable area bounded
top and bottom by the curve and
the x-axis respectively, and left and
right by the fixed ordinate at R,
and the variable ordinate at P (x, y).
Now let Ax be an arbitrary
change in x. Then A is changed by
some amount A. This area ùnder'
the curve is greater than that of the
inner rectangle with dimensions
Ax and y but less than the outer
Fig. 5.3 rectangle with dimensions Ax and
y + Ay. That is,
y(Ax) < 4A < (y + Ay) (Ax).
From this,
AA
< < y + Ay.
y Ax
Now, A is apparently a function of x, since its value is determined by

x. Thus as Ax --> 0, AA -+ 0 and AA/Ax --f D.A. The left member, y, of
the inequality is unaffected by Ax; the right member approaches y as
Ax - 0. Thus, as Ax --* 0,
AA
Ax -*y
Accordingly,
DzA = y
If the .function f ( ) were decreasing in the arbitrary interval from

x to x + Ax, the development would be the same except that Ay would be
essentially negative. The inequality signs would be reversed but the result,
D.A = y, would be the same. The student should write this up in complete
form with the book closed.
To illustrate, let us find the area bounded by the parabola
y=2+x-x2
and the x-axis. We have at once
DxA = y = 2 + x - x2 and thus A = 2x -E

2
- a
-f- C.
The value of C here can be deter-

mined by referring to the picture.
It is clear that for x = -1, the
curve crosses the x-axis and for
this value of x, A = 0. Thus, from
the last equation,
0=-2+z+}+C, or C=g.
Z Then,
Fig. 5.4 X2
A = 2x +
3 +L
an expression which gives the measure of the variable area under the arch
from the point (-1, 0) to any point P(x, y), -1 S x 5 2. To calculate
the area to (2, 0), insert x = 2 to find

22 29 +
A = 2(2) -}- 2 - 3 W _ $.
We shall refer to the expression for A as the area measure function.

Problem.
(1) Find the area under the arch from x = 0 to x = a.
(2) Calculate the area of the rectangle circumscribing the arch and
note that it is 50 per cent larger than that under the arch. This fact was
known to Archimedes.
EXERCISES
Sketch and find the area bounded by the curve, the x-axis, and the line.
15. y= x-x2,z=O,x= 1
16. y= x+xa,x= 0,x= -1
17. Y= x- x,x= O,x= 1
18.y= x+x3,x= 0,z= -1
19. y = x2 + x + 1, x = - 1, x = 2
20. y= x2+3x+4,x= -2,x= 1
21. y= sinx,x=0,x=7r
22. y= cosx,z= 0,x=7r/2
23. x2y= I,x= 1,x= n
24. xy2= l,x= 1,x= n
Sketch and find the area bounded by
25.y=9-z2,y=z+3
26. y= 16- x2,y= x+4
27. y = x(x - 3), y = 2z
28. y = x(x - 2), y = 3z
5.3. Area. Polar Coordinates

We determine plane area in polar coordinates in much the same way as
described in Par. 5.2. Here, however, we base the fundamental inequality
r=f(0)
Fig. 5.5
on areas of circular sectors. Consider the curve r = f (0), where there is

but one value r for each 0.
Let A represent the variable area between the curve, the polar axis,
and the ray r = OP drawn at any angle 0. A then is a function of 0. Now
let 00 be a change in 0, AA = area OPQ, the corresponding change in A.
With center 0, draw circular arcs through P and Q. Then
Area OPR < AA < Area OSQ
or
jr2(00) < AA < J(r + ir)2(A6).

From this
jr2 < Q < 1(r + Or)2.
Now, as A0 --> 0, AA -+ 0, and DA/A0 -> D.,A and, since the right member
of the inequality approaches the left member, AA /AO approaches that value
also. That is, AA/00 -> 1r2, or
D=A = jr2
The student should now recreate the argument, assuming that r decreases
(Or essentially negative) from 0 to 0 + MB and discover the same result.
Illustration. For the Cardioid
r=1+cos0, 050<2a,
we have
DzA = jr2 = l}(1 + cos 0)2 = I(1 + 2 cos 6 + cos; 0),
or
D=A +cos0+ +}cos20.
Then,
A =;8+sin0+ sin20+C.
Now, for 8 = 0, A = 0 and thus
(L)
C, = 0. The measure function for
area is thus
A= 40+sin0+sin20.
For 0 =7r, half the curve is generated

with area
Ae-r = 4a,
and the entire area enclosed by the Fig. 5.6
curve is I IT square units.
EXERCISES
Sketch and find the area bounded by the curve, and the given lines.
29. r = 4 cos 29, 0 = - a/4, 0 = 7r/4
30. r = 4 sin 28, 6 0, 8 = a/4
31. r = 2 cos 0, 8 = -7r/2, 8 = a/2
32. r = 2 sin 0, 6 = 0, 0 = 7r
33. r = 1- cosB,9= 0,0= 2a
34. r = 1 - sin 0, 0 = 0, 0 = 2a
35. r = 0, 0 = 0,8= 2a
36. rO = 1, 8 = a/4, 0 = IT
Sketch and find the area bounded by
37. r = cos 0, r = sin 0
38. r= 1 - cos8,r= 1+cos0
*5.4. Area. Parametric Equations
If x, y are the values of given differentiable functions of a parameter t,
the derivation of the area-measure function is as follows. Referring to
Fig. 5.3,
y(ax) < AA < (y + Ay) (,&x) (or the reverse).
Dividing by At,
Ax AA Ax
Y < (y + 4y) At
At < Al
Then as At -> 0, Ax -* 0, Ay -- 0,
and,
AX
-+D,x, Ay-4Ay, AA - DMA.
At At At
Accordingly,
D,A =
To illustrate, consider the area bounded by the parabola

{x=t2-1, y=l)
and the lines x = 0, y = 0, x = 3. We have
D,A = t(21),
so that,
A =;ts+C.
Fig. 5.7 Fig. 5.8

But for x = 0, t = 1 and A = 0. Thus, C = - 1. The area-measure
function is then A = 3(t3 - 1). For x = 3, t = 2, and
A =e
Consider the area bounded by the circle
{x = a sin t, y = a cost)
and the coordinate axes x = 0, y = 0.
Here
a21 + cos 2t
DMA = (a cos t) (a cos t) =
2
Then
A= a2t+sin2t+C.
C2 4 /
But for x = 0, t = 0 and thus C = 0. The measure function is accordingly
t sin 2t
A=a22+ 4
For x = a, t = Tr/2 and Ate, _ (ir/4) a2.
EXERCISES
39. Sketch and find the area bounded by each of the following.
(a) {x= t+2, y= 961- 1612},y= 0
(b) {x = tan t, y = cost t}, y = 0 (The Witch)
(c) {x = t2, y = t3}, x = 4, y = 0
(d) {x= asint,y= bcost}, y= 0
(e) fx= 2(1+ cost), y= 1 + sin t}
(f) {x = t - sin t, y = 1 - cos 1}, y = 0, 0 < t < 2x. (The Cycloid)
60
The Straight Line
Although the straight line, like the point, is an undefined element in

geometry, it may be variously described. Initially it is sometimes thought
of as the shortest distance between two points, or as the locus of a point
moving always in the same direction or as the set of all points whose co-
ordinates satisfy an equation of first degree. To each of these there are
objections. Outstanding is the fact that we need base lines from which to
measure distances, directions, and slopes.
In lieu, then, of any attempted definition, we choose the following as our
fundamental working concept.
6.1. Concept.
The straight line is the set of points, P, whose plane rectangular coordinates
x, y are expressible as values of functions
x=a+Xt, y=b+$Lt (1)
of the first degree in a variable parameter, t. The constants X, µ are not both 0
and the domain of t is the set of all real numbers.
If X = 0, and u 5x5 0, then x = a and the points are all those with con-
stant x, and thus lie parallel to the y-axis.
66
Sec. 6.1 THE STRAIGHT LINE 67
If X ;-4- 0, and it = 0, then y = b and the points are all those with con-
stant y, and thus lie parallel to the x-axis.
If X 0, and g 0, we may write (1) in the form
x-a y-b
and thus
x - a y - b
(2)
x µ
This is called the symmetric form

of the equation of the line and X and
µ are its direction numbers. If X and
are direction cosines, then t is the
directed distance from one point to
another. This fact plays an impor-
tant role at times hereafter, par-
ticularly in Par. 11.6. Note es-
pecially that (a, b) is a point of the
line since its coordinates satisfy
the equation.
Equation (2) may be written as
Fig. 6.1
a,x - Xy - (µa-Xb) = 0,
which has the form
Ax + By + C = 0 (3)
an equation of first degree in x and y.

Conversely, equation (3) may be re-expressed in the symmetric form
(2) and, accordingly, all linear equations in x, y represent lines. Thus, if
(x,, y,) is a point on the line (3), then
Ax,+By,+C=0,
and the difference of this and (3) is
A(x - x,) + B(y - y,) = 0. (4)
This is a scalar product relation between direction numbers [x - x,,
y - y,], of the line and the numbers [A, B]. Since this scalar product is 0,
68 THE STRAIGHT LINE Sec. 6.1
all lines with the fixed direction numbers [A, B] are perpendicular to
Ax + By + C = 0. In other words, [A, B] are direction numbers of paral-
lel normals to the line Ax + By + C = 0.
We may write (4) in the symmetric form (if A 0, B 54 0)
x- x1_y-Yi
B -A
and identify [B, - A] as direction numbers of the line itself. For example,
(A) The line 2x - 3y = 12 has direction numbers [-3, -2].
(B) The lines 4x - 6y = 7 and 3x + 2y = 1, with direction numbers
[-6, -4] and [2, -3], are perpendicular since the scalar product
-6(2) + (-4) (-3) isO.
(C) All lines perpendicular to
2x - by = 1 are given by
5x + 2y = k.
(Note the interchange of coefficients
of x, y and the change of sign). The
line through (2, 3), for example, is
that for which
5(2) + 2(3) = k.
Thus
5x+2y= 16.
(D) The line through (4, 5) and
(-1, 2) has direction numbers 5
and 3. For all points P(x, y) of this
Fig. 6.2 line, the direction numbers
[x+1,y-2],
must be proportional to [5, 3]. Thus
x + 1 _ y - 2
5 3
is the equation of the line. We might just as well have written [using the
point (4, 5)]
x-4 y-5
5 3
(E) The line through (a, 0) and (0, b) has direction numbers [a, -b]
and equation
x - a y
a -b
or
called the intercept form of the line.

This is an expression that frequently Fig. 6.3
will prove useful hereafter.
6.2. Slope
The slope of the line
Ax+By+C=O, B;-, 0
is the derivative Day, found from
A + B(D=y) = 0.
The slope, thus, is the constant
_ A
D:y B
Conversely, if a graph has the constant slope

A
Dsy
f3
then
A
y = -Bx+k,
and the graph is a line.
EXERCISES
Find direction cosines and angles measured from the axes of the following
lines. Draw a sketch of each.
x-1=y-2 2.x+I-y+2
5 12 3 4
3.1-x=y-2 4.2-x=3-y
3 4 5 12
5. y = 3x 6. x+y= 0
7. 2x - y = 1 8.x-2y=4
Write the equation of the line determined by the given conditions. Find
direction numbers of each and sketch.
9. Dxy = 4; x = 0, y = 0 10.D=y=-3;x=0,y=0
11.D=y=-3;x=4,y=0 12.D=y=3;x=O,y=-2
13. Through (3, 2) and (-2, 5) 14. Through (-2, 3) and (5, -2)
15. Through (0, -3) and (-2, 0) 16. Through (0, 3) and (-2, 0)
17. Through (xi, yi) and (X2, y2) 18. Through (3, 4) and (3, - 6)
19. Through (0, 0) bisecting the area formed by the triangle (0, 0), (2, 0),
(5, 6).
Write the equation of the line through the given point perpendicular to
the given line.
20. (-1, 2), 2x - y = 6 21. (2, -3), 3x - 4y = 6

22. (0, 0), 4x - 5y + 1 = 0 23. (-1, 0), 7x - 4y = 1
x-1 =
24. (1,1),2 y-3 25.(0,3),x+2=
y-1
5 3
6.3. Families
We expand ideas of Par. 6.1 and 6.2 by discussing roles that constants,
in the equation of a line, play in the families,
(A) y=3x+C (B) y-1 =m(x-2)

(C) 2x+y=k, x-2y=k (D) kfV4 F
(E) 2x + y - 7 + k(3x - y - 3) =0,
in which the letters C, m, k, represent arbitrary real constants. The word
"family" is appropriately chosen since it will be noticed that each equation
forms a set of lines having some common property.
IY
Y
'-g
Fig. 6.4 Fig. 6.5
(A) For each assigned value of (B) Since (2, 1) satisfies this
C, there corresponds a line of slope equation for each assigned value
3. Thus the equation represents a of m (the slope), the equation
family of parallel lines. represents all lines (except x = 2)
through the point (2, 1).
(C) 2x + y = k has direction

numbers [l, -2]; x - 2y = k has
direction numbers [-2, -1]. Thus
each is a family of parallel lines per-
pendicular to the other family.
(Note the interchange of x and y
coefficients with a change in one
sign, viz.: ax + by = h, and bx -
ay = k, produces perpendicular
Fig. 6.6 lines.)
(D)
k
t ---4 y -1
-- 72
represents for each k a line with

intercepts k and f 4 --k2, the
sum of whose squares is 4. That is,
the length of the line between the
coordinate axes is 2. If the sign of
the radical is negative, the line cuts
the negative y-axis. (This is the
trammel of Archimedes to be dis-
cussed later. All these lines are Fig. 6.7
tangent to the four-cusped hypocycloid called the Astroid, also to be

discussed later.)
(E) The family of lines 2x + y - 7 + k(3x - y - 3) = 0 is of

special importance. Consider the pair of lines
2x+y-7=0
3x-y-3=0
and let (a, b) represent their point
of intersection. Then x = a, and
y = b satisfy each; i.e., 2a + b -
7 = 0 and 3a - b - 3 = 0.
Now, the family equation
(2x+y-7) +k(3x- y- 3) = 0
is also satisfied by x = a, and
y = b for each value k since the in-
sertion of a and b makes each
X parenthetical expression 0. Thus all
members of the family pass through
the point (a, b). The fact that the
lines actually meet at (2 , 3) is of
Fig. 6 8
minor importance.
To illustrate the effective use of the idea, we shall find the equation of
the line through the point of intersection of
x-4y-1=0
and
2x +y+5 = 0
which also passes through (-1, 2).
Then
x-4y-1+k(2x+y+5) =0
represents (with one exception) all
lines through the intersection of the
two given lines. The one line of this
family which passes through
(-1, 2) must have an equation
satisfied by x = -1, and y = 2.
Thus
-1 -4(2) -1+k(--2+2+5) =0 Fig. 6.9
or k = 2.
This particular member is then

5x - 2y + 9 = 0.
Note that there was no need or interest in discovering the point of inter-
section of the given lines.
Project. Discuss the members of the family for which k = 4,, k
with reference to the "simultaneous" solution of the two given equations.
EXERCISES
Sketch several members of each of the following families and describe the
character of each family.
26. 2x - 3y = k 27.3.x+4y=k
28. y = 3x + C 29. y + 3x = C
30. y - I = m(x + 2) 31. y + 2 = mx
32. y = mx 33. x + my = 0
x
34. x+b= 1 35.+y=1
x y
36.a-+ =1 37. 1
9 y a 2+b- b=
Project. Given the family (2x/a) - (y/at) = 1, determine the plane
region where no lines of this family lie. (Hint: Solve for a in terms of x
and y. Then select those values of x and y which make a real. A clear picture
should be constructed.)
EXERCISES
Using the concept of a family of lines,. find the equation of the line through
the intersection of the given lines, satisfying the given condition.
38. x+y-1=0
12x -y+3= 0
(a) through (1, 2)

(b) parallel to 2x + 3y = 1
(c) perpendicular to 2x + 3y = 1
39. (2x- y-1=0

1 x + 3y - 2 = 0
(a) through (0, 0)
(b) parallel to x - 2y - 1 = 0
(c) perpendicular to x - 2y - 1 = 0
40. x+ y-1=0
2x + 3y - 6 = 0
forming with the positive coordinate axes a triangle of area z$ .
41. x+ y-2=0
with x-intercept 5
4x+3y-6=0
6.4. Concurrency
The necessary and sufficient condition that three lines (no two of which
are parallel),*
a,x+b,y+c,=0
a2x + b2y + C2 = 0
asx + b3y + ca = 0,
meet in a point (i.e., be concurrent) is that
Dm = 0.
(Necessity). If the three lines meet in a point, the common solution of

the first two equations must satisfy the third; i.e., the coordinates of the
point of intersection of
* If a,x + biy + c, - 0 and az - bsy + c2 - 0 are not parallel, then their slopes
are unequal; i.e.,
- at -
b,
as
b,
or albs - asb, =
a,
as
b,
b2
# 0.
Conversely, if this latter expression is different from 0, then -a,/b, 54 -aj/b,.

a,x+b,y+c, = 0
a2x + b2y + C2 = 0
c, b, a, C,
x-- c2
a,
b2
b,
y=- a2
a,
C2
b,
a2 b2 a2 b2
must satisfy a3x + bay + c3 = 0. Thus,

a, c,
a2 r2
- b3
a, b,
+C3=0,
a2 b2
or,
Cl b, a, C, a, b,
- as, - b3- +C3. = 0.
C2 b2 I I a2 c2 I I a2 b2
This is the expansion of the determinant D, and we have thus established

the necessary condition for concurrency.
(Sufficiency). To prove that this condition is sufficient we assume D = 0
and then show that the three lines contain a common point. Accordingly,
if D = 0, then
Cl b, a, c, b, a,
as* +b3 +C3 = 0,
c2 b2 a2 C2 b2 a2
or, since
a, b,
0,
a2 b2
c, b, a, c,
C2 b2 a2 C2
as
a, b,
+b3
a, b,
-Ca=0.
a2 b2 a2 b2
But this is evidently just the condition that the coordinates of the point of
intersection of the first two lines satisfy the equation of the third. Accord-
ingly, the three lines are concurrent.
EXERCISES
Find the number represented by
42. 43. 2 1 0
3 2 -1
0 2 5 5 4 -3
Show that the following triples of lines are concurrent. Sketch each set.
44. x+ y=3 45. 2x-y= 1

3x - 4y+ 5 = 0 3x+y=9
2x+ y = 4 y-x=1
46. x+ y= 2 47. x+ y = 1
3x + 4y = 10 3x+4y= 0
5x+3y=2 5x- y=23
Find the value of k which makes each of the following sets of lines concur-
rent.
48. x + 2y = 4 49. 2x+ y= 4

kx-4y=2 x+ky= 6
-x+5y=k x-2y=7
Find direction cosines and slopes of the lines
50. 51.
=0 =0
6.5. Distance From Point to Line
Let d represent the unsigned dis- ICY

tance from P: (x,, y,) to the line
Ax + By + C = 0,
A34 0, B 541 0.
The area of the triangle formed by
P and the intercepts - C/A and
- C/B of the line can be expressed
in determinant form-and also as sx
x
half the product of the altitude d
and the base length X Y as shown. Fig. 6.10
Thus
X1 y' i
1 C2 1
2d A2+ f2-2 -C/A

CZ
0 1
0 - C/B I
or, simplifying:
d=
Ax,+By,+C
A2 + B
a form that will be useful from time to time.
6.6. Angle Bisectors
The bisector of an angle formed by two lines is the locus of points equi-
distant from the lines. For example, the locus of P: (x, y) bisects an angle
formed by 3x + 4y - 12 = 0 and 5x + 12y - 20 = 0 if distances from P
to the lines are equal; that is, if
3x + 4y - 12 5x + 12y - 20
'
±-,/32_+_41 ±1/52 + 122
where we regard the radicals as presenting a choice in sign (the +, + and

-, - combinations yield the same result, and the +, - and -, + com-
binations, the same). For these two

choices
7x-4y-28=0,
and
4x + 7y - 16 = 0.
These are two perpendicular lines
bisecting the two angles formed by
Fig. 6.11 the two given lines.
Project.
(A) Show that the bisectors of the angles formed by any two lines, alx +
b,y + cl = 0 and a2x + b2y + c2 = 0, are perpendicular.
(B) Discuss the bisectors of the angles of the triangle formed by three
given lines. Show, using L, = 0, L2 = 0, and L3 = 0 as forms repre-
senting the lines, that there are four centers of circles touching the
three lines: an "incenter" and three "excenters." It will be noted that
these four points form an "orthocentric" set; i.e., any three of them
have the fourth as their orthocenter.
EXERCISES
Find the distance from the given point to the given line.
52. (5, 4), x + 2y = 1 53. ( 3, 2), 5x + 4y = 20
54. (-2, 1), x - 2y = 4 55. ( 1, -2), 2y - x = 4
56. (5, 3), 4x - 2y = 14 57. (4, 2), 3x - 2y = 8
Find the distance between the lines
58. 2x + 3y = 4 59. J3x - 4y = 10
2x+3y=7 3x-4y=-8
-1 y- 3 61. y= 3x+4
2 5
y= 3x- 5
Find the equations of the bisectors of the angles formed by each of the
following pairs of lines. Sketch each.
62. 3x - 4y = 6 63. 3x + 4y = 6
5x - 12y = 4 15x - 12y= 4
64. x+y=3 65. Ix - y=3
2x+y=5 x-2y+6 =0
66. Find the equations of the six bisectors of the angles of the triangle formed
by
3x + 4y - 10 = 0
7x - 24y + 75 = 0
4x - 3y - 40 = 0
Then find the coordinates of the incenter and the three excenters; the
inradius.
67. Find the locus of the centers of all circles tangent to the x-axis and the line
3x - 4y = 15.
68. Draw a network of points P with integer coordinates. Then
(a) find a set of such points (and their coordinates) on the line 3x - 5y +
1 = 0;
(b) show that no such points lie on the line y = V2x.
7*
The Circle
For general utility the circle is one of the most important of all plane
curves. Fortunately, it may be represented either in rectangular or polar
coordinates by relatively simple equations.
7.1. General Equations

(A) The circle is the set of points whose coordinates, x, y, are values of the
functions
x=h+Rcost, y=k+Rsint (1)
where h, k and R are constants and

t a variable parameter. R is the
radius, (h, k) the center, and t the
angle between a fixed line and any
radius.
Particularly useful forms of this
equation follow. Writing (1) as
x-h=Rcost, y - k=Rsint,
we square and add to find
(x-h)2+(y-k)2=R2 (2)
80
Sec. 7.1 THE CIRCLE 81
an expression indicating that all points P(x, y) are distant R units from
(h, k). Conversely, all points P whose coordinates satisfy (2) lie at the
constant distance R from (h, k).
The following is an important alternation of (2). We may write (2) as
x2+y2-2hx-2ky- (h2+k2- R2) =0
or
x2 + y2 + Ax + By + C =0 (3)
where we have set

A = -2h, B = -2k, C, = h2 + k2 - M.
It is thus apparent that we may identify a rectangular equation as repre-
senting a circle if it is of second degree, contains no x y-term, and has equal
coefficients for x2 and y2. Furthermore, the center (h, k) may be told off at
once as
h = -A/2, k = -B/2.
However, given form (2), it is best to put it into form (1) by completing
the squares on the terms in x and the terms in y to inspect for center and
radius. For example, in
x2+y2+4x-6y-3=0
we complete the two squares
(x2 + 4x + 22) + (y2 - 6y + 32)

3 + 4 + 9 = 16
to give
(x+2)2+(y-3)2=16
which, on comparison with form
(1), has center (-2, 3) and R = 4.
(B) Polar Coordinates. The polar
form of the general equation (3) is
(since x2 + y2 = r2, x = r cos B, Fig. 7.2
y = r sin 0)
74 + Ar cos 0 + Br sin0+ C = 0
82 THE CIRCLE Sec. 7.1
The student should note the important special cases listed below,
entering the rectangular and polar equations for each:
Rectangular Equation Polar Equation
1. C=0
2. A=0,C=0
3. B=0,C=0
EXERCISES
1. Find the center and radius of each of the following and sketch. In (a),
(b), (c), (d), (g), (h), (i), (j) find the equation of the tangent at the origin.
(a) x2 + y2 - 2x + 3y = 0
(b) 2(x2 + y2) + 5x - 6y = 0
(c) x2 + y2 = x
(d) x2 + y2 = y
(e) 3x2+3y2+7x+9y- 6=0
(f) 3x2 + 3y2 - 6x + 8y - 12 = 0
(g) r = cos 0
(h) r = sin 0
(i) r= 2sin0- 4cos0
(j) r= 4cos0- 2sin0
(k) r2 + 2r cos 0 - 3 = 0
(1) r2 - 6r sin 0 - 16 = 0
2. Show that there are no points whose coordinates satisfy
(a) x2+ y2+ x+1=0
(b) x2+ y2+ 6x+ 4y+ 15 = 0
(c) r2- 2rcos0+ 2 = 0
(d) r2 - 2r sin 0 + I = 0
EXTRA
All circles through the origin are given by
x2+y2+Ax+By=0.
Show that the tangent at the origin is
Ax + By = 0.
7.2. Families of Circles

The general equation
x2+y2+Ax+By+C =0
contains three essential constants whose values may be assigned or deter-
mined by certain specifications. Evidently, a circle* may be specified if
subjected to three conditions. It may, for example,
(a) pass through three given points (not collinear)
(b) pass through two given points and be tangent to a given line
(c) be tangent to three given lines (not concurrent or all parallel)
Illustrations.
(A) Find the equation of the circle through (-1, 0), (0, 2), (3, 0). The
coordinates of each point must satisfy (Fig. 7.3).
x2 ,+ y2 + Ax + By + C = 0
Thus
(-1)2 -A +C = 0 A -C = 1
22 +2B +C = 0 or 2B + C _ -4.
32 +3A +C = 0 3A +C = -9
From these,
and thus
A = -2, B =-, C = -3,
x2+y2-2x-ly-3=0
(What is the center; the radius?)
* Or a finite number of circles.
(0,2)
(-1,0) (3,0) g
Fig. 7.3 Fig. 7.4
(B) Find the equation of the circle through (0, 6) which is tangent to
y = x2 at (2, 4) (Fig. 7.4).
The coordinates (2, 4) and (0, 6) must satisfy
x2+y2+Ax+By+C=O.
Thus
2A + 413 + C = -20
6B + C = -36.
From these,
C= -36-613, A =8+13.
The equation of the circle may then be written:
x2+y2+ (8 + B)x + By - 6(6 + B) =0,
an equation that represents the family of all circles through the two
given points. We seek the member of this family whose slope at
(2, 4) is the same as that of the parabola at (2, 4). From the family
equation,
8 + B + 2x
2x + 2yDzy + (8 + B) + BD.y = 0 or D=y = - 2y+B
and from the parabola,
D=y = 2x.
For x = 2 and y = 4, these values of Dy must be equal. Thus
4 =4 or 13 = _54 .
8 8+ 13
Accordingly, the required circle is
44 84
x2+y2- 4x - 5y+-5 =0
The student should now rework the problem with the following
strategy. Using
(x - h)2 + (y - k) 2 = R2,
this equation is satisfied by (0, 6) and (2, 4) and, moreover, the line
joining (h, k) to (2, 4) is perpendicular to the tangent y = 4x - 4
to the parabola.
(C) Find the equation of the circle through P: (-2, 5) tangent to (Fig.
7.5)
3x-4y-19=0 and 4x+3y-17=0.
The desired circle is a member of the family (x - h) 2 + (y - k)2 =
R2 whose centers (h, k) lie on the angle bisector
3x - 4y - 19 4x + 3y - 17
5 = 5
or x+ y+2=0;
that is,
h+7k+2=0.
Furthermore, 1? is the distance from (h, k) to either given line. Thus,
R3h - 54k-19I-I5(k+1)I
Fig. 7.5
Accordingly,
[x- (-7k- 2)]2+ (y - k)2 = 25(k + 1)2
represents a family of circles tangent to the two given lines. If a circle
of this family is to pass through (-2, 5), then
[-2-(-7k-2)]222+(5 -k)2=25(k+1)2
or k = 0, b2. There are thus two circles:
(x + 2)2 + y2 = 25 and (x+ 64)2+(y- 52)2=289.
To construct, draw any circle, with center D, tangent to the two
lines. Let the line QP cut this circle at M1, M2. Lines through P
parallel to D1671, DM2 meet the angle bisector in centers of the re-
quired circles.
(1)) Find the equation of the family of circles through the origin whose
centers lie on the line 2y - x - 1 = 0 (Fig. 7.6).
Fig. 7.6
Any circle through the origin is given by

x2+y2+Ax+By = 0,
since (0, 0) satisfies the equation. Centers are (- A/2, - B/2) and
these coordinates satisfy 2y - x - 1 = 0. That is,
2/-\
2\
2/-1 =0
or
B= `2- 1.
Thus x2 + y2 + Ax + (4 _ 1)/ = 0, is the family.

Problem. Find those circles tangent to the line x - y + 1 = 0.
(E) Find the equations of the circles tangent to the lines
4x+3y-21 =0, 3x-4y-22=0, x+6=0.
Centers lie on the bisectors of the angles
4x+3,y-21
f5 - 3x-4y-22
f5 - x+6
ih
I
Using all combinations of signs, there are six bisectors in pairs at the
vertices A, B, C of the triangle:
A: x + 7y + I = 0, (I)
7x- y-43=0,(E)
x + 2y + 26 = 0, (E)
B:
2x- y+ 2=0,(I)
x - 3y + 51 = 0, (E)
C:
3 x + y + 3 =0, (I)
marked (I) and (E) for "internal" and "external" bisectors.
I
Fig. 7.7
Be THE CIRCLE Sec. 7.2
The three (I)'s meet at (-1, 0), the center of the in-circle.
There are three ex-circles with centers El: (-36, 5) ; E2: (9, 20) ;
and Es: (4, -15) where the external bisectors meet in pairs as shown
in the figure.
Radii are conveniently found at distances from centers to
x+6=0.
I:R=5; E,:R=30; E2:R=15; Ea:R=10.
The equations of the four circles are thus
1: (x+1)2+y2=25
E,: (x+36)2+(:y- 5)2=900
E.2:(x-9)2+(y-20)2=225
E3:(x-4)2+(y+15)2= 100
Problem. Show that I, E,, E2, Ea form an orthoeentric set and
that any point is the orthocenter of the other three.
EXERCISES
3. Write the rectangular and polar equations of the circle(s) determined by
the given conditions and sketch.
(a) Center (-3, 4), R = 5
(b) Center (4, -3), R = 5
(c) Center (-3, 4) and through (2, 0)
(d) Center (4, -3) and through (0, 2)
(e) Tangent to the axes with center on 3x - 4y = 12
(f) Tangent to the axes with center on 4x + :3y = 12
(g) Through (0, 0), (1, 2), (-2, 4)
(h) Through (-1, 1), (1, -1), (4, 5)
(i) Through (1, 3), (0, 2), center on x - y = 0
(j) Through (-2, 1), (1, -2), center on x + y = 0
4. Write the equation of the family of circles determined by each of the
following conditions. Sketch.
(a) Through (0, 0) and (0, 5)
(b) Through (0, 0) and (1, 2)
(c) Tangent to x = 0 and x = y
(d) Tangent to y = 0 and y = x
(e) Through the intersections of x2 + y2 - 4x = 0 and y = 2x
(f) Through the intersections of x2 + y2 - fnj = 0 and y = 3x
(g) Tangent to 7x + 12y = 23 at (5, -1)
(h) Tangent to 3x - 4y = 10 at (2, -1)
5. Find the equation of the circumcircle of the triangle formed by

(a) x+y=0 (b) 2x + y = 0
X- y= 0 x= 4
4y-3x=7 y=x-1
6. Find the equations of the four circles tangent to the lines
(a) 3x+4y- 15= 0,4x- 3y+25= 0,x+5= 0
(b) x -+- y- 1= 0, x- y+ 1= 0, x+ 2 y -+- 2= 0
7. Given
L1 x+y-1=0
L2 x-y-4=0
L3=3x+y+6=0
Form L1L2 + hL1L3 + kL2L3 = 0. Then determine h and k to make the
equation represent a circle (equal coefficients for x2, y2; no xy-tern).
Sketch and discuss the relationship of this circle to the triangle L1L2L3.
8. Find the equations of the circles tangent to the given circle at the given
point and with radius twice that of the given circle.
(a) x2 + y2 - 2x + 4y + 1 = 0 (h) x2 + y2 + 4x - 611 + 3 = 0
at (1, 0) at (1, 2)
9. Which members of the given family are tangent to the given circle?
(a) y = 3x + k (b) y - I = m(x - 2)
x2 + y2 - 2x = 0 xr2 + y2 - 2x = 0
(c) x2 + y2 - 4x = k (d) x2+112- 2x+4y= k
x2+y2-2y=0 x2 + y2 = 25
(e) %2 + y2 - kx = 9 (f) x2 + y2 - 2k-(x + y) + 7k2 = 0
x2 + y2 - 2y = 0 x2+y2= 25
k7.3. Pouter of a Poirtt
Consider the circle (Fig. 7.8) of radius R, and the point 0. Draw arbi-
Fig. 7.8
trary secants from 0 cutting the circle at A, B and D, E. Since opposite

angles of an inscribed quadrilateral are supplementary, -C OAD and C OEB
are equal and triangles OAD, OEB are similar. Thus
OA _OE (OA) (OB) = (OD) (OE)
or
OD ' OB
Thus for any point 0 in the plane of a circle, the product of collinear secant
lengths to the circle is constant. We call this constant the power of the point 0
with respect to the circle and designate it by the letter p. To evaluate p,
draw the special secant through the center C. Its segment lengths are
OC - R and OC + R and their product is
p = (OC)2-R2
It should be noticed that if the point 0 is either inside, on, or outside
the circle, then p <, =, > 0 respectively; and that if tangents may be
drawn from 0, the power is the square of the tangent length: viz.,
p = (OC)2- (CT)2 = P.
For an analytic discussion in rectangular coordinates, let the circle be
(x-h)2+(y-k)2=R2.
Then
p= (xo-h)2-+- (yo-k)2- R2,
Fig. 7.9
where xo and yo are coordinates of any point in the plane. By equation (3),
Par. 7.1, we may write
p=xo+yo+Axo+Byo+C.
For example, the square of the tangent length from the origin (xo = 0,
yo = 0) to
x2+y2+ Ax + By + C = 0
is C. Note thus that if C < 0, the circle encloses the origin.
*7.4. Radical Axis

Consider the two non-concentric circles:
r, °x2+y2+Aix+Bly+C, =0
r2=x2-{-y2+A2x+B2y+C2=0.
The powers of any point P (x, y) with respect to these circles are pi = x2 +
r,= + A,r B,y + C, and p2 = x2 + y2 + A2x + B2y + C2. We ask now
for the set of all points P(x,y) which have equal powers with respect to the
two circles. That is for which
x2+y2+A,x+B,y+Ci =x2+y2+A2x+B2y+C2
or
(A,-A2)x+(B,-B2)y+(C1-C2) =0.
This locus is a line called the Radical Axis of the two circles. We shall iden-
tify it further. Consider the family of circles
r,+kr2=0
and suppose the two circles intersect at (a, b) and (c, d).* Then
r,(a, b) = 0, r2(a, b) =0, r,(c, d) = 0, r2(c, d) = 0.
Accordingly, for all values of k, the
equation r, + kr2 = 0 is satisfied by
(a, b) and (c, d). Thus all members of
the family pass through these points
of intersection. The degenerate
member of the family is the one for
which k = -1; i.e., the radical axis.
This too contains the two points of
intersection; thus if two circles
intersect, their radical axis is co-
incident with their common chord.
For three non-concentric circles,
Fig. 7.10
r, = 0, r2 = 0, r3 = 0 there are
* Show that these are in fact circles. Then show that their centers lie on the per-
pendicular bisector of (a, b) and (c, d).
three radical axes, concurrent at a point called the Radical Center.* From
this point, if exterior to the circles, tangent lengths to the circles are
equal. It is thus the center of the circle orthogonal (meeting at right angles)
to all three.
Construction. The last discussion leads to a simple construction for the
radical axis of two given circles r, = 0, r2 = 0. Draw an arbitrary circle
intersecting r,, r2. The common chords locate P, the radical center of all
three. This point lies on the required radical axis.
Fig. 7.11
The construction of a circle through two points A, B, and tangent

to a line L as shown has several points of interest. Produce AB to
A
Fig. 7.12
Using the idea of equal tangent lengths, prove in this case that the three radical
axes are concurrent. Also make use of the determinant of Par. 6.4.
meet the line at 0 and let OA = a, OB = b. Imagine the circle

tangent to the line at P and set OP = m. Then by the power
theorem
m2 = ab,
and m is thus the "mean proportional" between a and b. Thus,

to construct the circle, mark 13- on AO extended, draw the circle
on A. as diameter, and OP1 perpendicular to AB. Then OP1 = m.
If we think of AB as a movie screen then the best seat in the
house on aisle L is at P, for P is the point of the line at which the
greatest angle is subtended by AB. At any other point, such as Q
of the line, the angle subtended by AB is measured by half the
difference of the intercepted arcs of the circle. (There is another
circle with P on the opposite side of the screen.)
EXERCISES
10. Calculate the power of each of the given points with respect to the circle
x2+y2- 3x+4y- 12= 0.
(-1, 2), (0, 0), (5, -6), (3, 2), (x1, yi), (z, -2)
11. Write the equation of the radical axis of
(a) x2+F12- x= 0, 2(x2+y2)-3y= 0
(h) 3 (.x2 + y2) - 4x -fry = 0, x2 + y2 - 25 = 0
(c) x2 + y2 - 6x = 0, x2 + y2 + 8x = 0
(d)x2+y2-2x+3y- 1=0, x2+y2+2x-3y=3
12. Find coordinates of the radical center of
(a) x2+y2-x=0, x2+y2-y=0, x2+y2-2x+4y-3=0
(b) x2 + y2 - x - y = 0, x2 + y2 - 2x + 4y - 1 = 0,
x'+y2-4x-5=0
13. Write the equation of the circle orthogonal to the three circles in 12 (a),
and in 12 (b).
14. Find the locus of points P such that the inclination of the line from (a, 0)
to P is twice that of the line from the origin to P
15. Find the locus of points P such that the distance from P to the origin is
k times the distance from P to (1, 0)
16. Find the locus of points P such that lines joining P to the origin and to
(a, 0) are perpendicular
17. Find the locus of points P such that lines joining P to the origin and to
(1, 0) form the angle a/4
18. Find the locus of midpoints P of line segments 8 units in length whose end
points lie on the x- and y-axes
19. Find the locus of midpoints P of lines joining (0, 0) to points on the circle
x2+y2-4x+3=0
20. Find the locus of points P such that the ratio of distances from P to two
fixed points is constant
21. Given the triangle A: (2, 8), B: (-4, 4), C: (2a , s) Find:
(a) the centroid (G)
(b) the orthocenter (H)
(c) the in-center (I)
(d) the circumcenter (0)
(e) the line HOG (the Euler line)
(f) the circumcircle
(g) the in-circle
(h) the three ex-centers El, E2, F,3
8.
Conics
The family of curves formed by plane sections of a right circular cone

are of three types: the Ellipse (including the circle), the Parabola, and the
Hyperbola. Of all curves, these are by far the most widely encountered and
applied. Applications are to be found in engineering structures, microscope
and telescope lenses, sound receivers and amplifiers, searchlight reflectors,
radar, Loran (Long Range Navigation), cams, and gears. An intimate
study of their properties is rewarding.
8.1. General Discussion
Consider the right circular cone, vertex 0, `base' angle 0, as shown in

Fig. 8.1. A sphere is dropped into the cone and lies tangent to the cone
along the circle of plane (i). The plane (c) is passed through the cone,
tangent to the sphere at F, and meets the other plane in the line DD'. Let a
be the angle between the planes (c) and (i). The curve of intersection of
this plane (c) and the cone is a conic section-in short, a conic.
Let P be representative of all points on the curve. Draw the element
OP which meets the tangent circle at A. Let C be the foot of the perpendi-
cular from P to plane (i). Draw CA. Then PCA is a right triangle and
95
THE CONICS Sec. 8.1
96
Fig. 8.1
PAC = R. From this

sin = PA.
Now pass the plane through PC perpendicular to DD', intersecting (i)
in CD. Then PDC is a right triangle and 4 PDC, = a. From this
PC
sina = PD'
Thus
sin a _ PA
sin 0 PD,
Sec. 8.1 THE CONICS 97
Now, lengths of tangents from an external point P to a sphere are equal.

The segments PA and PF are such tangents and are thus equal. Accord-
ingly, the last equation may be written
sin a PF
sin g PD
For a given plane and cone, this ratio is a constant which we have desig-
nated as `;e." This number is called the "eccentricity" of the conic. Its magni-
tude determines the type of section. The point F is called the focus and the
line DD' the directrix.
If a < 0, their e < 1, and PF < PD. The plane (c) then cuts all ele-
ments of one nappe and the curve is closed. This is called an ellipse, shown
in Fig. 8.3.
If a > 0, then e > 1, and PF > PD. The plane (c) then cuts elements
of both nappes and the curve has two parts or branches which are congruent
to each other, as shown in Fig. 8.4. Additional properties will be discussed
and displayed in Chapter 9.
If a = ,Q, then e = 1, and PF = PD. The plane (c) is then parallel to
an element of the cone. The curve is not closed and, since only one nappe
is cut, has but one branch. See Fig. 8.5.
We may summarize these remarks by the
following statements which are, in fact,
definitions.
D
(1) A conic section is the set of points common
to a right circular cone and an intersecting
plane, or
(2) A conic* is the plane locus of points P such,
that the distance from P to a fixed point F
(the focus), divided by the distance from P
to a fixed line (the directrix), is a constant
D'
( the eccentricity).
Fig. 8.2
Note an important fact (see Fig. 8.1). Since
sin a PF =
sin 0 PD e'
we may produce the three types of cones by assuming the cone given (i.e.,
that ,B is fixed), and varying the angle a of the cutting plane; or we may as-
* Other than a circle.
98 THE CONICS Sec. 8.1
sume the plane given in position (a is fixed), then vary the shape of the
cone by changing 0. t For example, we might fix a < 90°, then let # - 90°.
The cone becomes a cylinder and the section is an ellipse. This is the idea
of the following "project."
Project. Draw a right circular cylinder. Inscribe two equal spheres
tangent to a cutting plane. Show the section is an ellipse, as in Figure 8.2;
that is, that PF, + PF2 = PA + PB is a constant.
8.2. Special Properties

We give now some special definitive properties of the three general
types of conic sections for which e < 1 or e > 1 or e = 1.
(1) The Ellipse: e < 1, (a < 0) (See Fig. 8.3). Let two spheres be
PF,=PA
PF2=PB
PF,+PF2=PA+PB
=AB,const for all P
Foci : F1, F2
Directrices: D D', E E'
PF, E_ PFt _ PFD+PF2 PF1+PF2 20 a

PD PE PD+PE DE DE
DE=2 a ,thedistance between directrices.
Fig. 8.3
t The early Greeks did just this.
inscribed in one nappe of the cone and let the cutting plane be passed
tangent to both at F, and F2. Select a point P, representative of all points of
the curve of intersection, and draw the element OA PB of the cone, where
points A and B lie on circles in parallel planes. Then, since tangent lengths
from P to a sphere are all equal,
PF1 = PA and PF2 = PB.
Adding,
PF1 + PF2 = PA + PB = AB.
It is clear that for all points P of the locus,
AB = constant = 2a,
which is the slant height of the frustrum of the cone. Thus, with 2a repre-
senting this constant,
PF,+PF2=2a
This property gives rise to the alternate definition: The ellipse is the plane
locus of points P such that the sum of distances from P to two fixed points,
F, and F2, in the plane is constant.
Let F1F2 = 2c, and let the points where the line of foci meets the curve
be V, and T'2, called vertices. Then V1V2 = 2a, for, since V, and V2 are
points of the curve,
V1F1 + V1F2 = 2a and V2F2 + V2F1 = 2a,
or
V1F1 + V1F1 + 2c = 2a and V2F2 + V2F2 + 2c = 2a.
Thus
V,F,=a-c=V2F2j
and
T'1V2 = V1F1 + F1F2 + F2V2 = (a - c) + 2c + (a - c) = 2a.
The segment V1V2 is called the major diameter of the ellipse.
(2) The Hyperbola e > 1, (a > ,e) (See Fig. 8.4). Here the plane (r)
cuts both nappes and is tangent at F1 and F2 to spheres in each nappe.
Then, for P, a representative point of either branch of the curve,
PF1 = PA and PF2 = PB,
where PA and PB are measured along the element PBOA. Then
IPF,-PF2=JPA-PB1 =AB=2a,
PF1=PA
PFz=PB
PF,-PFZ=PA-PB
=AB,const
for all P
Fig. 8.4
a constant for all points of the curve. Thus the alternate definition: The
hyperbola is the plane locus of points P such that the unsigned difference of the
distances from P to two fixed points, F, and F2, in the plane is constant.
Again we set this constant equal to 2a and note that it is represented by
the segment V, V2, points where the line of foci meets the curve. This seg-
ment, V1V2, is called the transverse diameter.
(3) The Parabola e = 1, (a = 0) (See Fig. 8.5). The cutting plane (c)
touches the inscribed sphere at F and is parallel to an element HG of the
cone. The line of intersection of (c) and the plane of the tangent circle
Fig. 8.5
(through G and A) is the directrix DD'. From a representative point P

on the parabola, PD is drawn perpendicular to DD', and PA O is an ele-
ment of the cone.
Then, since PF and PA are tangents to the sphere,
PF = PA = HG.
But HG and PD are equal lengths since they are parallel segments
contained between two parallel planes. Accordingly,
PF=PD
Thus the parabola is the plane locus of points P equidistant from a line (the
directrix) and a distinct point (the focus) in the plane.
8.3. Equations of Lite Conics

IJ (A) Polar Equations. Let a
a
directrix be k units distant from the
corresponding focus. If we take the
focus as the pole, and the perpen-
dicular OX to the directrix as the
initial line, then P: (r, 0) will be
representative of all points of the
conic. From the fundamental
definition.
PO
= E.
Fig. 8.6 PD
r Ek
That is =E or r = E X 0.
k+rcos8 1 - Ecos0'
For relative positions of directrix, focus, and initial line as shown
(Fig. 8.7), equations would be as indicated. The student should check
each one for self-assurance.
_ Ek ek _ Ek
r r 1-E-sin 8
In each case, the pole 0 is a focus
Fig. 8.7
(B) Rectangular Equations. It is
convenient to take the y-axis coin-
cident with the directrix, with the
x-axis passing through the focus.
Then,
% (N/x-k)2+y2=EIx1,
or
(1 - E2) x2 + y2 - 2kx -f. k2 = 0

Fig. 8.8 represents any conic.
Note that:
(1) the equation is of the second degree
(2) e < 1, ellipse, coefficients of x2, y2 have the same sign
(3) e = 1, parabola, coefficient of x2 is 0
(4) e > 1, hyperbola, coefficients of x2, y2 have different signs
For central conies (ellipse, hyperbola) we choose foci at (-c, 0), (c, 0) and
set V1V2 = 2a.
Fig. 8.9 Fig. 8.10
PF2 + PF1 = 2a, a>c PF2 - PF1 = 2a

and
PFi - PF2 = 2a, a<c

or or
PF2 = 2a - PF1 PF2 = ±2a + PF1
V1F1 = e (V1D1)
a - c =e1E-a) c - a = e(a -
C = at
We shall derive the equation for the hyperbola, leaving the analogous
work for the ellipse to the student as a closed-book exercise. We have
1'(x+c)2+y2 = f 2a+V(x- c)2+y2.

Squaring and dividing by the factor 4a,

C
( X x -c +y.
a
Squaring again and setting c = aE,

(ex-a)2 =x2-2ax+a2E2+y2
or
(E2 - 1)x2 - y2 = a2(E2 - 1).
Now, since e > 1, we set b2 = a2(e2 - 1) = c2 - a2, writing finally,
x2 y2
-- 1 c2 = a2 .+ b2
a2 b2
For the ellipse, since e < 1, take b2 = a2(1 - E2) = a2 - c2.

Note:
ELLIPSE HYPERBOLA
OW = b, the semi-minor axis OV1 = a, the semi-transverse axis
OV, = a, the semi-major axis V1B = b, the semi-conjugate axis
F2W=F1W =a OB=c=OF1
Foci lie on the major axis. Foci lie on the transverse axis.
Guide Lines (Asymptotes). The

Hyperbola possesses guide lines--a
feature not shared by the Ellipse
or the Parabola. Let P1) be the
distance from the representative
point P on a curve f (x, y) = 0 to
the line y = mx + k. This is a guide
-11 line if, as OP -> co, then
Fig. 8.11
PD-+0 and Dxy->m
where D=y is the slope of the curve.
For the hyperbola
z
b2x2 - a2y2 = a2b2 with D2y =
a =fyy
b
a. Vi x2
the line bx - ay = 0 is a guide line. For, in the (+, +) quadrant,

bxl - ay, bxl - ay, bxl + ay, b2x1 - a2yi
PD = '/a2 + b2 (bxi + ayi)
a2 + b2 a2 + 62 bxl + ay,
Now, since P(xi, yi) lies on the curve, then
b2x -a2yi = a2b2.

j
Thus
a2b2
PD =
/a2 + b2 (bxl + ay,)
In the first quadrant, as OP -- oo, then xl - oo, yl -> oo, and PD --> 0
Furthermore, the slope of the curve Dzy -* b/a, the slope of the guide line.
Since the curve is symmetrical to both axes, the reflection
bx + ay = 0
of the first line is also a guide line. Accordingly, the factorable equation
b2x2 - a2y2 = 0
yields both guide lines. Note that these are diagonals of the rectangle with
their center at the center of the curve, and sides 2a, and 2b parallel to the
axes of symmetry of the curve. In sketching the hyperbola, this rectangle
and its diagonals should invariably be drawn first. The curve is then freely
sketched with little trouble.
8.4. More General Equations

If coordinate axes were chosen parallel to the axes of the central conics
with centers at (h, k), equations would be
x -a2h) 2 + (y k) 2
=1;
b2
- (y-k)2=I.
(x- h)2
a2 b2
g
(x- )2+ (y-k)2=1
(L2 b2
Fig. 8.12 Fig. 8.13
The Parabola has no center. If

we choose coordinate axes with the
x-axis parallel to the line of sym-
metry of the curve having its vertex
at (h, k), directrix x = h - p/2, we
have
v(h,k) V(x - h - p/2)1 + (.g - k)2
=ix-(h- 2)
Fig. 8.14 Squaring,

(y-k)2=x2-2hx+px+(h-2)n-x2+2hx+px+(h+2)2
or
(y-k)2=2p(x-h)
Alternate forms.
(1) If p is negative, the parabola "opens" toward the negative x-axis.
(2) If the y-axis is chosen parallel to the line of symmetry, the equation of
the parabola is
(x-h)2=2p(y-k)
"upward" if p > 0, "downward" if p < 0.

The two preceding forms are essentially
x=Aye+By+C
y = Axe+Bx+C
which may be returned to those forms by completing the square.
Of special importance are the simple forms
Y
..g
Fig. 8.15 Fig. 8.16
yz = 2px
x2 = 2py
for parabolas with vertices at the origin.
EXERCISES
1. The equation of a directrix and the coordinates of a focus of a conic are
given. Derive the rectangular equation in terms of e. Then obtain the
special equation for each given value of e. Determine the other focus and
equation of other directrix for the central conics.
(a) x+2= 0, (2,0);E= 1, 2, z
(b) x - 2= 0, (-2,0);e= 1, 2, i
(c) y+3 0, (0,3);e= 1,4,1
(d) y-3 0, (0 ,-3); e= 1, 4,
(e) x + 2 = 0, (3 , 4); e = 1, 2,
(f) x-20 ,( - 3 ,4 ) ; 1, 2,
(g) x+y= 1, (0,0);e= 1,2,i

(h) x + y + 1 = 0, (0,0);e= 1,2,
(i)x + y = V2, (-1,-1);= 1,2,
(j) x+y+ =0,(1,1);e= 1,2,'
2. For each of the following conics, determine, where applicable, the center,
foci, directrices, eccentricity, and guide lines. Sketch each curve.
(a) x2 + 4y2 - 8x - 16 = 0 (e) r(2 - cos 8) = 2
(b) 4x2 + y2 - 8y - 16 = 0 (f) r(2 + cos 8) = 2
(c) x' - 4y2 - 8x - 16 = 0 (g) r(2 + 3 sin 8) = 6
(d) 4x2 - y2 - 8y - 16 = 0 (h) r(2 - 3 sin 0) = 6
3. Find two numbers whose sum is 12 such that their product is greatest.
Sketch the product function.
4. Of all positive numbers, which provides the least sum of four times the
number added to its reciprocal?
5. Find the equation of the locus of points P, the sum of whose distances
from the given points is the given constant. Sketch each.
(a) (- 1, 0), (1, 0), 4 (b) (0, -1), (0, 1), 4
(c) (2, 3), (-2, 3), 8 (d) (2, -3), (2, 3), 8
(e) (1, 1), (- 1, - 1), 4 (f) (1, 1), (2, 2), 4
6. Find the equation of the locus of points P, the difference of whose distances
from the given points is the given constant. Sketch each.
(a) (-1, 0), (1, 0), 1 (b) (0, -1), (0, 1), 1
(c) (2, 3), (- 2, 3), 2 (d) (2, - 3), (2, 3), 4
(e) (1, 1), (- 1, - 1), 2 (f) (1, 1), (2, 2), 1
7. Find the equation of the Parabola with the given properties. Sketch each.
(a) Vertex (-2, 0) (b) Vertex (0, -3)
Focus (0, 0) Focus (0, 0)
(c) Vertex (-4, 4) (d) Vertex (-4, 4)
Focus (-4, 0) Focus (0, 4)
(e) Dzy = 1 - 4x through (0, 0) (f) D=y = 2x - 5 through (0, 0)
8. Guide lines of a hyperbola are
x - 2y - 4 = 0, x + 2y - 3 = 0.
Find the equation of the curve if it passes through (-2, 1).
9. Find the equation of the locus of points P(x, y) if an associated point
Q(x, 3y) is at a constant distance 5 from (0, 0). Sketch.
10. Sketch several members of the family of conics (parameter k).
(a) y2= k(x+k) (b) k+ky24 = 1.

?
It. Determine the equation of the boundary of the region through which no
line of the following family passes. (Hint: Solve for m and determine
relative values of x and y that yield non-real values of m). Sketch the
region.
(a) y = 2mx + m2
(b) y = mx -l-
4m
(c) (y - mx) = f 9 + 18 t2 (a family of twins)
(d) (y - mx) = f N/16 nm2 - 9 (a family of twins)
(e) 1/ -}- r11x = f N/2 -in
x y
(f) + =1
m S- m
8.5. Constructions
The Ellipse. Since PF2 + PF1 = 2a and F2F1 = 2c, then the total
distance PF2 + F--F, + PF1 = 2a + 2c, (a > c) is constant. Thus if an
inextensible string (ends tied together) be looped over thumb tacks at
F1F2, the track of a pencil point at P, moving with the string taut, is an
ellipse.
Fig. 8.17 Fig. 8.18
Draw two concentric circles, radii a and b (a > b). Let OBA be a ray
making angle 0 with an x-axis OX. Draw BP parallel to OX, AP parallel
to OY. Then P is on an ellipse. For, if coordinates of P are x, y and OA =
a, OB = b, then x/a = cos 0, y/b = sin 0, or
2
x2
a2 + b2 = 1.
The Parabola. Draw lines parallel to the directrix OD, meeting the axis
FX of the parabola in points M. With F as center and 0111 as radius, draw
arcs locating points P on the set of parallel lines. It is evident that FP =
OM = the distance P to the directrix.
Fig. 8.19
The Hyperbola. Draw two concentric circles, radii a, b. Let OCA be a

ray making angle 0 with an x-axis. Draw AM tangent to the circle at A,
and TC tangent to the other circle at T. Draw MP perpendicular to OX,
C P parallel to OX. Then P is on a hyperbola. For, if coordinates of P are
x, y and OA = a, and OB = b, then
x/a = sec 0, y/b = tan 0,
and thus
x2 y2
= 1.
a2 b2
An inextensible string F1PA of

length k + 2a has one end fixed at
F1, the other to a point A on a
straightedge. Another point of the
straightedge, distant k from A, is
fixed at F2. A pencil keeps the string
Fig. 8.21 taut against the straightedge as
shown and traces the hyperbola
with F1, F2 as foci, 2a as transverse axis. For,
AP+PF1- (AP+ PFZ) =k+2a-k

or
PF1 - PH2 = 2a.
THE CONICS 111
Sec. 8.5
EXERCISES
12. Use string to construct the following ellipses.
(a) Major axis: 2a = 8 (b) Major axis: 2a = 26
Focal distance: 2c = 6 Minor axis: 2a = 14
13. Construct (a) the ellipse and (b) the hyperbola by the method of con-
centric circles of radii four and five inches.
14. Make and use the mechanism described for the hyperbola in 8.5 (R).
15. For the given values K, draw and discuss members of the families
(a) y= x2- 2x+K,K= 0,1,-1,-2
(b) y= x2+Kx+ 1, K= 0,2,-2,3
(c) y = Kx2 - 2x + 1, K = 0, 1, -1, -2
16. Sketch and discuss the extent of
(a) {x = sin t, y = csc t}
(b) (x = see t, y = cos t}
(c) {x= sin t, y= csc t} U {x= sec t, y = cost}
17. Solve Axe + 2Bxy + Cy2 + 2Dx + 2F,y + F = 0, C 5 0 for y in terms
of x. Then determine the condition
A B D
D C E =0
DEF
which makes the square root expression a perfect square. This is a condi-
tion sufficient for the conic to degenerate into two lines.
18. Show that the orthocenter of three arbitrary points (a, (1/a)], [b, (1/b)],
[c, (1/c)] on the hyperbola xy = 1 also lies on the curve.
19. Show that (ay --- bx) (by + ax) + cx + dy + e = 0 is the most general
equation of the rectangular hyperbola. Give equations of its guide lines.
9.
The Conies:
Their Properties
And Applications
The Conies stand first in the parade of all curves on two counts: the
variety of their properties and the extent of their applications. This chapter
presents some (by no means all) of these-properties that underlie the
construction of lenses in microscopes and telescopes, the description of
paths of projected bodies (either locally or in far distant spaces), the forms
of cables in suspension bridges, the arches of viaduct supports, and the long-
range charts of air and sea navigation. It is said that Manaechmus in-
vented (or discovered) the conics in trying to solve the ancient problems of
Trisecting the Angle, Squaring the Circle, and Duplicating the Cube. More
certainly it is known that when Kepler sought to explain the interaction of
heavenly bodies, he found, all developed and waiting, a great body of
knowledge of these curves, which was pointedly pertinent. The application
of the hyperbola became important in locating positions of enemy guns
during World War I and again in World War II in positioning planes and
ships by means of newly developed electronic methods. The elliptical
paths of satellites and the parabolic shapes of radar receiving screens are
matters of interest now to everyone. We discuss details here.
112
Sec. 9.1 THE CONICS: THEIR PROPERTIES AND APPLICATIONS 113
9.1. The Reflective Property

The three conies have an important common property that has varied
and interesting applications. This is the reflective property (already dis-
cussed for the Parabola in Par. 5.1, Fig. 5.2.).
The tangent to a conic at any point P

forms equal angles with the focal lines
PF1 and PF2.
We shall establish this property by first considering again (see Chapter

4) the formation of a tangent. Any conic (whose axes are parallel to the
coordinate axes) may be represented by an equation in rectangular coordi-
nates of second degree, i.e., by
Axe+By2+Cx+Dy+F=0,
and any line by
y = mx + k.*
Their points of intersection are at most two, x-coordinates of which are
given by the quadratic
(A+Bm2)x2+(C+2Dm.k)x+F+k.2=0, A+Bm2*0.f
The tangent line is formed when
these roots, x, are not distinct, but Link lengths:
equal. Thus the tangent to a conic F, G2 = G, F2 =2a
F, F2 = aG,Gc2 =2c
has only one point in common with
the curve, and, if a line (not parallel
to the axis of a parabola or parallel iG
to an asymptote of a hyperbola)
has but one point in common with
a conic, it is tangent.
i'
The Ellipse. With the foregoing Q/
principle in mind, we build the
plane crossed parallelogram linkage
shown whose joints are F1, F2, Fig. 9.1
G1, G2. Let P be the variable inter-

section of the longer links. It is evident that F2P = PG2, and thus
F1P + F2P = F1P + PG2 = 2a.
* Except x = constant.
t For A + Bm2 = 0, A and B have different signs, and the conic is a hyperbola with
guide-line slopes ± -%/ - (A/B). Lines with these slopes each meet the curve in only one
point (the quadratic in x reduces to a first degree equation) and thus could not be tangent.
114 THE CONICS: THEIR PROPERTIES AND APPLICATIONS Sec. 9.1
Accordingly, P lies on the ellipse with foci F,, F2, (or G,, G2) and major
axis 2a. As the linkage is deformed with F,, F2 fixed, the point P traces the
ellipse.
Draw the line QPR bisecting F,PG,. Then G1, G2 are reflections of
F,, F2 in this line. We wish to show that P is the only point of this line in
common with the ellipse and that the line is thus tangent. To this end,
consider any other point Q of the line. Then, since for any triangle the sum
of any two sides is greater than the third:
QF, + QF2 = QG, + QF2 > G,F2 = 2a.
Link lengths: Thus all points Q of the line, other
=2a than P, lie "outside" the ellipse
F,F2=G1G2=2c and the line is therefore tangent.
c>a
Since QPR bisects G,PF,, then
F2PQ = F,PR
The Hyperbola. We make use

of the same mechanism here, except
that the shorter links are extended
(as in the Parallel Ruler) beyond
the joints F,, F2, G1, G2
Let P be the intersection of the
Fig. 9.2 extended shorter links. It is evident
that PF2 = PG2, and thus, for the
branch pictured,
PF1 - PF2 = PF1 - PG2 = 2a.

Accordingly, P lies on the hyperbola with foci F,, F2 (or G,, G2) and trans-
verse axis 2a.
Draw the line PQ bisecting c F2PF1. Then G1, G2 are reflections of
F,, F2 in this line. Now P is the only point of this line in common with the
hyperbola. For, any other point Q of the line is such that
QF1 - QF2 = QFi - Q02 > F,G2 = 2a,
and thus Q cannot be a point of the curve. The line PQ is accordingly tan-
gent and
F2PQ = F, PQ.
The Parabola. Our general statement of the reflective property needs
interpretation in the case of the parabola since this conic has only one focus.
Referring to the section of the cone

shown in Fig. 8.1, let the plane (i)
be fixed and let a - 0 with a < 0. P
, F2
The plane (c) rotates about the
directrix DD' and approaches paral-
lelism with an element of the cone.
The focus F, remains on the fixed
sphere while F2 recedes indefinitely
along the axis of the conic; i.e.,
F, F2 - ao and the conic approaches
the form of the parabola. Thus for
the parabola we interpret "the focal
Fig. 9.3
line PF2" as the line through P
parallel to the axis of the curve. This line and PF1 form equal angles
with the tangent, a fact already established in Chapter 5 where polar
coordinates were used.
Problem. Prove the reflective property for the parabola with rectangular
coordinates using y2 = 2px and its tangent at P(x,, y,). [flint: Show that
the tangent, the axis of the curve, and PF1 form an isosceles triangle. The
focus is F,(p, 0).]
EXTRA
Folding the Conies. The reflective property permits a pleasant and
attractive project in the creation of the conics by folding paper. Using good
quality wax paper, draw circles with centers F, and radius R as shown.
Fig. 9.4
Select an arbitrary point F2 inside the circle for the ellipse, outside for the
hyperbola. Fold and crease the paper so that any point Q of the circle falls
on F2. The crease is the perpendicular bisector of QF2 and is tangent to the
conic with foci F,, F2. * Let P be the point of the crease on QFI, or QFI
extended. Draw F2P = u, FIP = v. Then, since P is equidistant from
F2 and Q, PQ = u, and also, evidently, for the
ellipse: u + v = R; hyperbola: is - v = R;
a constant for all points Q of the circle. Moreover, angles formed by the
crease with the focal lines PF1, and PF2 are equal and, by the reflective
property, these creases are tangents.
Continue making creases (three dozen or more) by folding over other
points Q onto F2 and mount the finished model on a dark background.
The Parabola again presents a special case. Start with a line (in the
realm of fantasy, this is a circle with "infinite" radius) instead of a circle.
Fold this over onto a selected point F. The line is the directrix and the
creases are tangents to the parabola. The student should prepare an ex-
planation of this construction and formation.
9.2. Tangents
Consider the point P: (x1, yl) on the conic
Ax2+Ry2+2Cx+2Ey+F=0.
The slope of the curve at any point is given by
Ax+ByDzy -l- C+EDZy = 0;
that is,
Ax + C
By + E*
The line tangent at P is then
Axl+C
y-yl`Byl+ E(xx,)
* In creasing a piece of paper, two points R and S

are folded over and held upon each other. An arbitrary
point P on the crease is obtained by pulling against R,
S. Distances RP and SP will be equal. Other points 1'
are located in the same way. The crease will be the per.
pendicular bisector of the original plane segment RS.
Fig. 9.5
or
Axx,+Byy,+Cx+Ey = Ax,+By;+Cx,+Ey,.
Since P lies on the curve, its coordinates x,, y, satisfy the equation, and thus
the right member of the last expression can be exchanged for
- Cx, - Ey, - F.
The tangent line is then
Axx,+Byy,+C(x+xt)+E(y+y,)+F=0
a form that can be conjured up at once from the equation of the curve on
replacing
x2 by xx,, y2 by yyl,
x byY 2' y by
y 2
* 9.3. Areas
We have already (Chapter 5) determined areas of parabolic segments.
The determination of areas of segments of the hyperbola would lead us
into matters that properly belong in the study of Calculus and shall be
omitted here. The Ellipse, however, presents little difficulty. Consider the
semi-circle
y, = .V/a2 - x2
and the semi-ellipse
Y2 = aa2 - x2.
The area between the elliptical arch

and the x-axis is given by
b
D.A =y2 = a2-x2.
a
Fig. 9.6
The corresponding area for the circle is given by
D=B=y,= a2-x2
where, because both y's increase as x decreases, we measure areas from a
to X.
THE CONICS: THEIR PROPERTIES AND APPLICATIONS Sec. 9.3
11 g
Accordingly,
D=A =
and thus
A= ()3
a
+ C.
Now for x= a, y,= 0, y2= 0, and A = 0, B = 0. Thus C = 0. We have

then for any x,
A = \a) ,3.
Particularly, for x = 0, B = 'aa2/4 and the quadrant area of the ellipse is
A =-.
a 4
b 7ra2
The entire area of the ellipse is thus
A = 7rab
EXERCISES
1. Calculate the areas of the ellipses described in Problem 5, Chapter 8.
2. Calculate the areas of the ellipses
(a) x2 + 9y2 - 12x = 0
(b) 9x2 + y2 + 12y = 0
(c) 2x2 + y2 - 6x + 4y = 0
(d) x2+2y2+6x-4y=0
(e) x2+ 9y2+ 10x+ 16 = 0
(f) 4x2 + 9y2 + 24x + 36y - 72 = 0
3. Write the equation of the line tangent to the given conic at the given point.
(a) Axe + Bye + Cx + Dy = 0 at (0, 0)
(b) y = A x2 + Bx at (0, 0)
(c) x2 - y2 = x + y at (4, 3)
(d) 9x2 - 16y2 - 3x + 4y = 0 at (8, 8)
4. For each of the following conics, a focus is at the origin and the focal line
is along the x-axis. Find the equations of directrices and coordinates of
the other focus (if any).
(a) Through (-4, 3), = 1 (d) Through (7, 12), e =
(b) Through (3, 4), e = 1 (e) Through (-5, 12), e = 2
(c) Through (-3, 4), e = z (f) Through (4, -3), e = 2
5. Show that the tangent to b2x2 + aY = a2b2 at (xi, yj) and the tangent to
x2 + y2 = a2 at (x1, y2) have the same x-intercept.
6. Find the area bounded by y = x2 and y = x(6 - x).
7. The equation (x - 1)2 + y2 = 2x2 represents all conics with focus (1, 0),
directrix x = 0 and eccentricity .
(a) Find k for the lines y = kx to be tangent to members of the family of
tonics
(b) Find the locus of all points of tangency
(c) Identify the types of conies for which the inclination of the tangent is
<, =, >45°.
Draw a figure showing the three types of tonics and their tangents.
8. Draw an arbitrary circle with center at the focus F of a parabola meeting
the curve in P and its axis in N and T. Show that PT and PN are tangent
and normal to the curve at P.
*9.4. The Parabolic Cable

A suspension bridge is made by hanging a roadbed on a cable anchored
at two points of support. For our
theoretical discussion we assume the
cable and vertical stringers weight-
less, and take W as the constant
weight per linear foot of the road-
bed. We take x, y axes as shown with
the origin at the lowest point of
the cable.
Since the cable (and structure)
are in equilibrium, the tension at any
point is tangent to the cable. At the
origin, the tension is horizontal and Fig. 9.7
equal to a constant H. At P: (x, y),
let the variable tension be T and its inclination 0.
For the section of the structure from 0 to P, there are thus three forces:
H, T, Wx
which maintain equilibrium. The force Wx is the vertical weight of x feet

of the roadbed. For equilibrium, the sum of these three forces is 0. Thus,
resolving them into vertical and horizontal components,
T sin 0 = Wx
T cos B = H.
Dividing and setting Dzy = tan 8, we have
jj7
tan g=Dy=Hx.
From this, the curve of the cable is [since (0, 0) is on the curve]
y = W x2
2H
a Parabola.
9.5. Plane Motion of a Point

Velocity. If a point P(x, y) moves
to (x + Ax, y + Ay) in time At, its Y
average velocities in the x- and y-di-
rections during this time interval are
Ax Ay Vy
and
At At* P
If x and y are functions oft so that x

.x
the position of P may be calculated
if time is known, then the instan- Fig. 9.8
taneous velocities at time t are
Ax Ay
vZ = Limit and vy = Limit
At-+O At At-+0 At
These are derivatives with respect to t, and we denote them by the sym-
bols D,x and D,y. Thus
vz = D,x and v = Dy.
Acceleration. If, further, the velocities change from vZ, v,, to vZ + AV.,
v + w, in time At, the average accelerations in the x- and y-directions
during this period are
AvZ Ov
and
At At
Instantaneous accelerations are then
-Z and a,, = Limit

Av
aZ = Limit
At'o At At-+0 At
-
These are recognized as second derivatives of x and y with respect to I.

Accordingly,
yr = Dx, v,, = Dy
aZ = D 1 Z = D2x, av = D V, = D2y
provide information concerning the plane motion of a point.

Note that in general, the ratio of the average velocities is
Dy Ax AY
At At Ox'
and, as At -> 0, Ax --> 0, and Ay - 0, then the limit of this ratio is
- = -D(y
V,
vZ
=DZy
Dix
i.e., the direction of motion is along the path (the velocity vector v is tangent to
the path).
Two studies are now possible.
(1) If x and y are values of given functions of t, we may find velocities and
accelerations in the directions of the axes.
(2) If accelerations a.,, ay are values of given functions of t, together with
given particular values at some time t, (called "initial conditions"),
we may find velocities vZ, vy and the values of position functions x
and y.
1 22 THE CONICS: THEIR PROPERTIES AND APPLICATIONS Sec. 9.5
To illustrate the first study, let

t t
x = 2sin2, y = 3 cos
2
t
then Dx = v= = cos , Dy = v,, sin
2
2
and D2x = aZ = - z sin D2y = a = -; cos 2.
2
For specified times t = 0, t = 7r/2,
for example,
t -0:
x = 0, y = 3, vz = 1, v = 0,
a:=0, -i
ar
x=JZ,
vx= 2, vy 4
3J2
Fig. 9.9 a.
V2
all =- 3v2
8.
The path is the ellipse 1x2 + $y2 = 1 and the point P starting at (0, 3)
moves in a clockwise direction. The velocity vector and its components are
drawn to scale and in position at t = Za.
For the second study, we consider the motion of a point mass P fired at
t = 0 from the origin with x and y velocities:
vo cos a and vo sin a.
At some time t the mass m is at
P(x, y). In flight the only force act-
ing on the mass is (-mg), which is
due to the gravitational pull in a
negative y-direction. Newton's law
of motion is that force produces an
acceleration proportional to the
force; i.e.,
F = ma. Fig. 9.10
Thus for forces at P in the x and y directions,

0 = mD2x, -mg = mD2y
or
D2x = 0, D2y = - g (_ - 32 ft/sect). (1)
From these equations,

Dx=C1, Dy=-gt+C2.
But when t = 0,
Dx = vo cos a, Dy = va sin a
Thus
Dx = (vo cos a), Dy = -gt + (vo sin a) (2)
and consequently,
x = (vo cos a)t + K1, y = -2t2 + (vo sin a)t + ill.

However, when t = 0, x = 0 and y = 0. Thus
K1=0, K2=0.
Finally,
x = (vo cos a) t and y = - 2V+ (vo sin a) t (3)
Thus equations (1), (2), and (3) give complete information about the
projectile. Equation (3) gives its position at any time t; equation (2) its
velocity components; equation (1) its acceleration.
The rectangular equation of the path is found by eliminating tin (3).
We find
y a)zx2 + (tan a)x (4)

-2vo ( cos
9
and recognize it as a parabola.

Problem.
(1) The horizontal range R is the x-intercept of the path (4). Find R;
then calculate the angle a which gives maximum R.
(2) Find the time of flight [from equations (3) ].
(3) Find the maximum height (maximum y) for given a and vo.
(4) Find the coordinates of the focus of the parabola (4). Then show that
the foci of all parabolic paths produced by varying a, with vo constant
lie on a circle.
(5) Find the range up the straight hill y = kx.
EXERCISES
9. For each of the following, draw and identify the path of the point P (x, y);
obtain expressions for vr, v and a%, ay at any time t; then calculate the
velocities and accelerations at the specified time and draw these compo-
nents in position.
t2,y=t},t=1
t- 1, y t2},t= 2
Cost,y= Kill e},t= 7r/3
I - Cos1,y= Sill t},t= 703
2cost, y = 3sint},t= 37x/4
-3cost, y = 4sint}, t= 37x/4
sec t, y tant},t=7r/4
tan t, y sect}, t = 7r/4
t, y 110, t= 2
12,y 1/t},t= 2
t+]/t,y= 1 1/t},t=2
1 - sint,y= I- cost},t= 7/6
10. In each of the following, the accelerations of a point at any time t are
given together with velocities and position at a specified time. Find the
velocities at any time t and the equations of the path. Draw the path.
(a) a.= O,av= -32;v2= 3,vy= 4andx= O,y= Oatt= 0
(b) ax= 0, ay= -32;vx= -3, vv= 4andx= 0, y = Oatt= 0
(c) a.= 2, ay= 0; 1,vy= 2andx= 1,y= I att= .12
(d) ax= 2, ay= 4;v: 4,v,, -17ancix= 4,y= -42att= 2
(e) a,z= -3 sin t, ay= 3cost;v., = 0, v,, = 3andx= 3, y = 3att = 7r/2
(f) az = -16 cos 2t, ay = -12 sin 2t; v= _ -S, vy = 0 and x = 0,
y=3att=7r/4.
*9.6. LORAN
Electronic pulses travel with the speed of light; i.e., at the constant rate
of 186,000 miles per second. Reducing this to more workable time units,
scientists use the microsecond: one-millionth of a second. One microsecond
thus measures the distance 0.186 miles traveled by a pulse in 10-6 seconds.
If two stations send out pulses at the same time, a receiver P could cal-
culate the distance between the stations by noting the difference in time of
his receptions of these pulses.
The Loran system consists of a Master transmitting station M and a
Slave station S located at some known and charted time-distance 2c from
M. Station M emits a pulse signal that is received at S in time 2c and at P
in time t,. When the signal reaches S it activates a second signal which is
received by P at time tz.
Fig. 9.11
1 26 THE CONICS: THEIR PROPERTIES AND APPLICATIONS Sec. 9.6
If the navigator P steered a course that keeps the difference of times

tz - t, constant (equal to 2a), he would track a hyperbola with M and S as
foci and 2a as transverse axis.
Charts are prepared with such hyperbolas drawn about M and S. Each
curve of this family is that along which tz - t, is a constant, and each is
marked with this time difference for identification. The navigator tunes his
receiver to the established frequency of the station and locates on his chart
the hyperbolic branch marked with his constant time difference.
A second pair of stations M and S on another frequency establish a
second set of hyperbolas. The ship then locates itself on one of these curves
and her position is thus completely determined by the intersection of these
two branches of hyperbolas in the two families. The same master station (at
Bermuda, perhaps) may activate two slaves as shown in the diagram.
The navigator is never confused over choice of branch of a hyperbola.
The other branch wanders off in a far-away region-beyond the tolerance of
even a foolish error.
10.
Some Mechanical
Motions And Loci
We devote this chapter to plane mechanical motions that have im-

portant implications in geometry and fundamental applications in the
design of machinery. Linkages, cams, and gears have for their purpose the
motions that package cigarettes and cut turbine blades, bottle drinks, and
stamp radio grids, drive automobiles, and play juke boxes.
The elementary and principal motion is circular. With a plane upon
which to operate, an inextensible link with one fixed point is sufficient to
create circular motion with all the accuracy desired. From this basic motion
the designer fabricates a movable structure to create motion of desired
character. We begin with basic concepts.
10.1. Instantaneous Center*

Consider motion (not pure translation) of a rigid body (a lamina) in
a plane. We suppose that the velocities of two points A and B of the body
are known at an instant, as shown in the "snapshot" of the action. Draw
* See William F. Osgood, Mechanics. New York: The Macmillan Co., pp. 154-160.
127
128 SOME MECHANICAL MOTIONS AND LOCI Sec. 10.1
through A and B perpendiculars

AM11 and BN to their respective
velocity directions.
Now, any point of AM is mov-
ing-if at all-parallel to the direc-
tion of A since the body is rigid
(there could be no motion of any
N such point toward or away from
A). Likewise. any point of RAT
moves-if at all-at this instant
parallel to the direction of B.
Fig. 10.1 Let the point of intersection of
AM and BN be I. This point can-
not. nruvc parallel to two different directions and thus must have zero ve-
locity at the instant pictured. It is called the instantaneous center of motion.
Any other point P of the rigid body then moves in a direction perpendicu-
lar to the line joining P to I. We have thus a means of constructing tangents
to paths traced by points P of rigid bodies which move so that two of their
points A and B have prescribed directions.
10.2. The Trammel (Ladder)

Suppose a ladder of length a + b moves with its ends A and B on two
perpendicular lines, chosen as the x- and y-axes. Let P: (x, y) be a point
Fig. 10.2 Fig. 10.3
fixed on the moving ladder, a units from one end, b units from the other,
and let B be an angle of inclination of the ladder as shown. Then
x = a cos 0, y=bsin0
Sec. 10.2 SOME MECHANICAL MOTIONS AND LOCI 129
and the path of P is thus an ellipse. The instantaneous center of motion is

the intersection of perpendiculars at A and B to their directions along the
axes. Thus the tangent t to the ellipse at P is perpendicular to IP. Note
several items of interest.
(1) The locus of M, the midpoint of the ladder (P chosen so that a = b)

is a circle, center at 0. This `stepladder' arrangement (called the Scott-
Russell mechanism and used as the "walking beam" on some river
boats) permits disposal of one of the lines.
(2) The point I is a vertex of the rectangle having the ladder as a diagonal.
The locus of I (called the centrode) is thus the circle of radius a + b
with center 0.
(a) (b)
Fig. 10.4
(3) The ladder is in all positions tangent to the Astroid, Fig. 10.4(a)
(proof omitted).
(4) Members of the family of ellipses traced by all points of the ladder are
tangent to the Astroid, Fig. 10.4(b) (proof omitted).
*10.3. The Conchoid
Given a fixed point 0 distant a units from a given line L, let R be a

point fixed on a straight rod which passes always through 0. As R moves
along L, the locus of P, Q, two points distant b units from R on the rod is
the conchoid (shell-shape).
SOME MECHANICAL MOTIONS AND LOCI Sec. 10.3
130
Fig. 10.5
To derive the equation of this two-branched curve, let 0 be the pole

and OX, parallel to L, the polar axis. Since OR = a csc 0, then
OP = a csc 0 + b, OQ = a csc 0 - b.
Thus
r = a csc 0 f b
represents the locus of both P and Q.

Note :
(1) The curve is symmetrical to the line through 0 perpendicular to L.
(2) There is a loop if b > a, a cusp if b = a.
(3) The point of the rod at 0 has direction of motion along the rod, the
point R moves along L. The instantaneous center I is thus located
and tangents at P and Q are perpendicular to PI and QI, respectively.
Problem. From the polar form just given, produce the rectangular
equation.
10.4. The Cycloids

The family of curves, called cycloids, formed by rolling circles upon
circles, is of importance in the construction of gears and in the study of
planetary motion. In addition to characteristics which make the curves of
practical value, there are features which endear them to the geometer.
Definition. The Cycloid is the

path of P, a point of a circle rolling
without slipping upon a fixed circle
or line. The curves have cusps and
arches as shown and fall into three
classes:
(1) Epicycloids:* rolling circles
òutside' the fixed circle
(2) Hypocycloids: f rolling circles
ìnside' the fixed circle
(3) Ordinary cycloids: rolling cir- Fig. 10.6
cles on a line
(A) The Line Segment and Ellipse. Consider a circle, center C, radius
a, rolling inside a fixed circle, center 0, radius 2a. Let T be the point of
tangency, and P the tracing point. Draw OCT and OP extended to P1. If
we let TOP1 = 0, then 4 TCP = 20. It is evident that
are TP = a(20) and are TP1 = 2a(0)
and thus
are TP = are TP1.
Fig. 10.7 Fig. 10.8
Accordingly, if the smaller circle rolls without slipping, P will fall at P1

and is in all positions collinear with 0 and P1. Thus P travels along the fixed
diameter through P1.
* Epi = "on top" or "upon" as in epidermis.
t Hypo = "beneath" as in hypodermic
132
Let Q be diametrically opposite P. The diameter traced by Q is perpen-

dicular to OP, since triangle QOP is inscribed in a semicircle. Thus the
diameter ? , 2a in length, travels with its ends on two perpendicular lines
and acts precisely as the Ladder of Par. 10.2. Then any point F of PQ
(and thus any point rigidly attached to the rolling circle) traces an ellipse.
The fixed circle upon which the smaller circle rolls is recognized as the locus
of the instantaneous center I of the ladder.
(B) The Astroid. The hypocy-
eloid of four cusps is the Astroid
where the radius of the rolling circle
is one-fourth that of the fixed circle
(radius a). We have encountered it
as the envelope of the ladder and
thus know that its tangent length
measured between the axes of sym-
metry is constant.
We select the x-axis through a
cusp as shown. Since there is no
slippage, are TP = are TX. Thus if
TOX = 0, then TCP = 40. If
CRM is perpendicular and PR
parallel to OX, then PCM =
90° - 30. (To verify, add the angles
Fig. 10.9 OCM, PCT, and PCM). We have
x = OM - RP = 4 a cos 0 + 4 cos 30
y = MC - RC = ; a sin 0 - 4a sin30.
Using the identities,
cos 30 = 4 cos' 0 - 3 cos 0
sin 30 = 3 sin 0 - 4 sin' 0,
the forms are reduced to
I
x = a cos' 0, y = asin' 0 or x213 + y2!3. = a213
At the moment pictured, T is the instantaneous center of motion and the

perpendicular to TP is tangent to the Astroid at P. Show that this tangent
passes through S, the point where OT meets the rolling circle.
(C) The Ordinary Cycloid. This

curve is traced by a point P of the
circle, radius a, rolling upon a line. We
take the x-axis through the cusps, the
y-axis at one of them. Let 0 be the
angle through which the circle has
turned after passing the cusp 0.
Then since OM = are PM, ON = x, Fig. 10.10
NP=y,
x=OM-PR=aO-asin0, y=MC - RC = a-acos0
or
x = a(0 - sin 0), y = a(l - cos 0)
Note that:
(1) The span 00' of one arch is equal in length to the circumference, 21ra,
of the circle.
(2) The instantaneous center is M. The tangent to the cycloid at P ac-
cordingly passes through the topmost point of the circle.
(3) The length of one arch is 8a (Proof omitted).
(4) The area bounded by the line and one arch is 3(ira2) (Proof omitted).
Question. Find the slope of the curve in terms of 0.
(D) The Cardioid. The epicy-
cloid, formed as one circle rolls
upon another of the same radius a,
is called a Cardioid. Placing the
pole at the single cusp and the polar
axis through the fixed center, P will
have coordinates r and 0. Are
lengths IO and IP are equal and
subtend angles 0. The four points
D, 0, P, and C form an isosceles
trapezoid. If DO, OP, and PC be Fig. 10.11
projected onto I)C, we have
2a = acos0+r+acos0
or
r = 2a(l - cos 0)
134
Note that:
(1) Each of the four equations r = 2a (1 f cos 0), r = 2a (1 ± sin 0)
represents a Cardioid.
(2) I is the instantaneous center of motion, and thus the tangent to the
curve at P passes through P and T.
(3) The are length of the curve is 16a (Proof omitted).
(4) The area is 6ira2 (See Par. 5.3).
(5) Light rays emanating from a point source on a circle are reflected
from the circle along rays tangent to a Cardioid: the tangent TP forms
0/2 with DT. Draw the circle, center D, radius 3a, and mark its inter-
section S with DO. Draw ST.
Since ,COST and ODT sub-
tend the same are, COST =
0/2. Furthermore, since SDT is
isosceles, -OST = STD = 2
0/2. Thus a ray ST meets the I
reflecting circle at T and pro- S
ceeds along TP, tangent to the
Cardioid. Make the straight
line drawing by starting with
a circle and one of its points S.
Lay off equal arcs, numbered
1, 2, 3, , from S. Rays inci- Fig. 10.12
dent at n are reflected to 2n.
Draw only these reflected rays (about 40). Notice how this bright
curve (a caustic) occurs on the surface of coffee or milk in a cup, on
the bottom of a shiny cyclindrical cooking pan, etc.
Project.
(A) Draw lines tangent to a circle. Draw their perpendiculars from a
selected point on the circle. Sketch the locus of the intersections of
these tangents and their perpendiculars. Such a process forms the
pedal of a curve.
(B) Draw a circle. With centers on this circle, draw (about 40) circles pass-
ing through a selected point of the first circle.
(C) Derive the equation of the two-cusped epicycloid called the Nephroid.
This is the envelope of rays reflected by a circle where the point
source of light is "at infinity" and incident rays are parallel to each
other.
(D) Mark equal arcs as in the sketch of the Cardioid caustic and draw
reflected rays from n to 3n for the Nephroid.
*10.5. The Lima cons

Let P be rigidly attached, at a
distance b from the center, to the
rolling circle whose radius equals
that of the fixed circle. We take the
pole 0 on the fixed diameter through
the cusp of the Cardioid such that
DO = b. Then, projecting OP = r,
CP, and DO onto DC,
b cos 0 + r + b cos 0 = 2a
or
r = 2(a - b cos 0)
These are trochoidal curves forming.

a family of Lima cons. They are of
three types : Fig. 10.13
(1) b > a, interiorly looped

(2) b = a, cusped (the Cardioid)
(3) b < a, oval-shaped (some dimpled)
Note that :
(1) The four equations r = a f b cos 0, r = a f b sin B represent Lima-
gons.
(2) Inverses of these curves may be formed through the relation rp = 1
(See Par. 12.5). Thus the inverse of r = a - b cos B,
1
P =
a - b cos 0
is a conic. For a <, _, > b, these are respectively hyperbolas, para-

bolas, and ellipses.
Question. Sketch the Limason and its inverse for b = 2a, for a = 2b.
*10.6. Linkage Motion

The simple jointed parallelogram in its crossed form is endowed with
important properties. Linkage arrangements involving this crossed paral-
lelogram shed considerable light upon matters of concern to the mathema-
tician and engineer.
(A) Conic Gears. (Refer to Figs. 9.1, 9.2). Let F1F2G1G2 be joints of a
crossed parallelogram with F1F2 = G1G2 = 2c; F1G= = G1F2 = 2a, a > c.
The intersection P of the longer bars lies on two congruent tangent el-
lipses with foci at the joints and major axis equal to 2a. Since F1G2 = 2a,
these two points may he fixed as shown, and the elliptical plates will turn
about these as centers. (The bar G1F2 = 2a may be retained to keep con-
tact.) We thus have elliptical gears used in changing uniform circular
motion into variable circular motion. Such quick-return mechanisms are
found in machines designed for a slow working stroke followed by a fast
return to a new working position.
Fig. 10.14 Fig. 10.15
We take the same linkage with the shorter links extended to meet in a
variable point P which lies (Par. 9.1) on two congruent tangent hyperbolas
with foci F1, F2 and G1, G2. Hyper-
bolic plates may be attached to the
links F1F2 and G1G2 to form hyper-
bolic gears. If F1 and G2 are fixed to
the plane, uniform motion about F,
produces a desirable variable mo-
tion about G2 (the bar F1F2 is re-
tained to keep contact).
(B) A Linma4on Linkage. The
crossed parallelogram ABCQ with
AQ = BC = a; AB=QC=b,
a> b, has variable angles BAQ
and BCQ always equal (marked 0).
If a second crossed parallelogram
00'AB, with AB = 00'= b and
FA = R-0 = c, is joined to the
Fig. 10.16 larger one as shown, then . 0'A B
_ 0'OB. If we wish 0'OR = OCP, then the two parallelograms

must be similar with sides proportional, and conversely; i.e., if
a _ b
or b2 = ac,
b c
then
0'OB = OCP = 0.
With this stipulation we create the linkage shown in Fig. 10.16 where we
have fixed the points 0 and 0' to the plane determining the fixed line OX.
Now, BOO' = XOC = OCQ. As OC rotates about 0, CQ turns
about C through the same angle but in opposite direction. This is precisely
the action of two equal circles rolling one upon the other as sketched. P
traces the Cardioid, other points Q of CQ trace the general Limacons.
(C) The Lemniscate. The three-
bar "kite" has 0, 0' fixed with
AB=BP=OB=a, and O'A =
00' = aVV. ABP is a continuous
straight bar. We take 0 as pole,
0'0 as polar axis, and P with co-
ordinates r, 0. Since A, P, and 0
are equidistant from B, they lie on
Fig. 10.17 a circle with center B, and thus
AOP = 90°. From this,
-,CO'OA = -CO'AO = 90° - 0. Thus
OA = 2(aV) cos (90° - 0) = 2av2 sin 0.
Furthermore, since
(OA)2 + (OP)2 = (AP)2,
8a2 sine 0 + r2 = 4a2,
or
r2 = 4a2 (1 - 2 sine 0) = 4a2 cos 20
the "figure-eight" curve shown.

Project. Using rectangular coordinates, find the equation of the locus of
P(x, y) such that the product of distances PF1, PF2 is the constant a2,
where F1 is (-a, 0), F2 is (a, 0). Then revert to polar coordinates and
identify this as a Lemniscate. Finally, determine the area enclosed by the
curve from its polar equation. The instantaneous center of the bar AP
(Fig. 10.17) is the intersection of

0'A and OB. Put the linkage in
position where P is at 0. The in-
stantaneous center, then, is A. Show
that the slopes of the curve at 0
are + 1, - 1.
(D) The Cissoids. Six bars are
jointed together in the plane as
shown in Fig. 10.18 and have the
following lengths [a > b]:
QA=QB=a; OA=OB=AP=RP=b.
It is evident that Q, 0, and P are collinear through all deformations. If M
is the midpoint of the variable rhombus OAPB, then
(QM)2 + (MA)2 = a2 and (OM)2 + (MA)2 = b2.
By subtraction,
(QM)2 - (OM)2 = a2 - b2 = (QM + 0141) (QM - OM).
But since OM = MP, we may write this last relation as
(QP) (QO) = a2 - b2
Fasten the joint 0 to the plane and attach a radial bar 0'Q (=c) fixed at 0'
so that 00' = 0'Q = c. Then, Q moves on a circle that passes through 0.
Using 0 as pole, and 0'0 as the polar axis, coordinates of P are r, B. Noting
that . 0'OQ = 0'Q0 = 8, then QO = 2c cos 0 and
(2c cos 0) (2c cos 0 + r) = a2 - b2; or
a2 - b2
r=
0
-2 B
This is a family of Cissoids for values of the parameter c. There is a loop

if 2c > ct2 - b2; a cusp if 2c = a2 - b2; an isolated point if 2c < a2 - b2.
Project.
(1) Transfer the equation to rectangular coordinates, and find the equa-
tion of the guide line. Sketch the curve.
(2) Build the linkage for a = 5, b = 3, and c = 8 inches; discuss the one
for a = 5, b = 3, andc=6.
10.7. An Involute of a Circle
Let a rigid straightedge be rolled

without slipping on a circle of radius
a. The locus of a point P of the
straightedge is an Involute of the
circle. (We may choose, instead, to
think of a string PA kept taut and
unwound.) Let axes be selected as
shown in Fig. 10.19 with the
straightedge tangent to the circle
at A, and let 0 = XOA. Then if
P started at B, are AB=AP = a0
(radian measure). Draw AN per-
pendicular and PM parallel to OX. Fig. 10.19
Then
x = ON + MP = acos0+aOsin0
y = NA - MA =asin0-a0cos0
Note that:
(1) Any point of the straightedge traces an Involute of the circle (such
curves are called `parallel' to each other).
(2) The instantaneous center of motion is A, and the direction of motion
of P is perpendicular to AP (i.e., parallel to OA).
(3) It is a curve frequently used in the construction of gear teeth (see
Curves, p. 137).*
*10.8. Cams
A lamina plate rotates about a fixed point 0, pushing a pin P into a
straight slot or cylinder, as shown in Fig. 10.20. Such a plate is called a
cam. Motions imparted to a pin (or another cam plate) of this type are
essential elements in sewing machines, juke boxes, gasoline pumps, etc.
We consider a few of these cam motions.
(A) The Circle. Let a circle be pivoted at its center 0 (see Fig. 10.21)
and carry a point A which bears on a bar rigidly attached at right angles to
* It. C. Yates, Curves and Their Properties. Ann Arbor, Mich.: Edwards Bros., 1952.
Fig. 10.20 Fig. 10.21
a pin. The pin is slotted in line with 0 and held against the point A by some
spring device.t Let OA = k, .)AOP = 0, OP = r. Then
r = kcos0
Accordingly, the pin oscillates. If

the circular cam turns with constant
angular velocity, P moves with
harmonic motion. Rest positions
occur when A is collinear with 0
and P; the maximum velocity of P
is at 0, the center of oscillation.
Fig. 10.22 (B) The Limacon. Harmonic mo-
tion may be produced with a
Limayon cam (Fig. 10.22). Let the cam rotate about the pole 0 of
r=a+bcose
where 8 is the angle between OP and the axis of symmetry as shown. Here
a < b. Rewriting the equation as
(r - a) = b cos 0
this is recognized as similar in form and character to the relation in (A).

Oscillation takes place between those extreme positions of P where 0 = 0
and 7r; that is, for r = a + b, a - b. The center of motion is thus at r = a.
t Or, the point A is a knob riding in a slot in the vertical bar.
(C) The Archimedean Spiral. The Spiral r = aO, (0 --<_ 0 < rr) forms
the two symmetric halves of a cam
pivoted at the pole 0 (Fig. 10.23).
If the cam turns with constant
angular velocity, the variable dis-
tance r = OP changes at a constant
rate, a times the angular rate.* Thus
P has uniform linear motion. Certain
sewing machines use this cam in the
bobbin-winding attachment so the Fig. 10.23
thread guide moves uniformly back
and forth. Some gasoline pumps use the principle to register gallons and
cost.
*10.9. Line Motion Linkages

The mechanical problem of cre-
ating linear motion is neither foolish
nor elementary. The most natural
and most available motion, given a
working plane, is circular, and it is
from this basic motion that we wish
to produce linear motion.
To this end, we utilize the
crossed parallelogram, discussed in
several earlier paragraphs. We start
with four bars equal in pairs
OB=0'A=a; 00'=AB=b;
b < a;
and note that always
z O'OB = O'AB = 0.
We now add two bars OD and CD, attached, as shown, to the first
parallelogram with
DC=00'=b and OD=O'C=c, c<b,
as in Par. 10.6 (B). If we selected the bar length c at random, we could not
expect . DOO' to be equal to 0. If, however, it were equal to 0, the parallelo-
grams would be similar. If they are similar, their corresponding side lengths
would be proportional; that is, alb = b/c or b4 = ac.
* That is, since r = ae, then Dgr = a D,8.
Taking lengths in this manner, we fix 0 and 0' to the plane, extending
the bar OD to E so that OE = a.
Then, completing the rhombus
OEPB, we establish the point P
which is always collinear with 0 and
0'. The student should construct a
model of this with cardboard strips
and eyelets, taking, for example,
a = 6, b = 3, and c = z inches.
Question. Extend the bar PE
to Q so that EQ = a. What is the
path of Q? (Refer to the Ladder of
Par. 10.2.)
We consider here an arrange-
Fig. 10.25 ment of six bars (see Fig. 10.25), re-
sembling the one of Par. 10.6 (D),
with
AP=PB=BQ=QA=a, OA=OB=b, a>b.
Then (see the Power Theorem of Par. 7.3),
I OP I J OQ J = IOM J ON = (a - b) (a + b).
We attach a seventh bar 0'Q of arbitrary length c and fix 0' to the plane
so that 070 = 0'Q. The path of P is a line perpendicular to 00'. [Prove
this fact analytically, using polar coordinates with 0 as pole.]
*10.10 The Lemniscate

The crossed parallelogram may
be used to draw the Lemniscate
(compare Par. 10.6 Q. Let 0, P,
Q be midpoints of bars and fix A
and D to the plane. Using P as pole
and APD as axis, coordinates of Q
are r, 0. In triangle BAD, OP =
(BD)/2; in ABC, OQ = (AC)/2.
Thus
(OP) (OQ)
= (BD) (AC)/4 = (a2 - b2)/4
or
(RD) (AC) = a2 - V. Fig. 10.26

Let PM = z = MQ. Then

/a
BD = 2(BM) cos 0 = 212 - z cos 0
AC = 2(A31) cos 0 = 2( + z) cos 0.
Thus
(BD) (AC) = (a2 - 4z2) cos2 0 = az - b2.
Furthermore,
r = 2z cos 0.
Accordingly
a2 cos2 0 - r2 = a2 - b2, the locus of Q.
Now, if
a = bV'2
r2 = 2b2 cos2 0 - b2 = b2(2 cos2 0 - 1)

or
9.2 = b2 cos 20 the Lemniscate.
Note that the instantaneous center of motion is I, the intersection of

AB and CD produced. This point is on the hyperbola having A and D as
foci and b as the transverse axis length.
EXERCISES
1. Draw a picture of a drive wheel, connecting rod and cylinder of a railroad
engine. Find the instantaneous center of motion of some selected point of
the rod. Find the path of the midpoint of the rod if the rod has the same
length as the radius of the wheel.
2. The flange of a railroad car tire projects 3 inches beyond the circumference
of the tire. If the wheel has 30 inch diameter, find the equations of the path
of a point on the outer edge of the flange.
3. A "kite" linkage ABCD has AB = AD = 6, BC = CD = 4. If AD is
fixed, what is the locus of the instantaneous center of the bar BC?
4. Derive equations for the epicycloid generated by a circle of radius b rolling
on a circle of radius a.
5. Derive equations for the hypocycloid generated by a circle of radius b
rolling on a circle of radius a.
Fig. 10.27 Fig. 10.28
6. A 12-inch diameter wheel rolls along a straight line, 360 revolutions per
minute. Find components of the velocity of a point on the circumference
2 seconds after contact with the line.
7. A straight stick AB moves with its ends on a circle of radius 0.4 = R.
The point P of the stick, a units from one end, b units from the other, de-
scribes another circle of radius r concentric with the first. Show that the
difference in areas of these two circles is the area of an ellipse with semi-
axes a, b. (Hint: Draw the diameter through P and make use of the Power
Theorem, Par. 6.3). (This is a special case of a rather remarkable theorem:
see it. C. Yates, Curves and Their Properties, p. 111).
8. Given the circle of radius a/2 and the fixed point 0 distant a/VV units from
the center, find a polar equation of the locus of points P on a variable line
through 0 such that OP = AB (l.emniscate).
11.
Transformations
It is interesting and sometimes useful to exchange a given rectangular
coordinate system for another with a new origin and direction. We wish to
determine for a representative point P the relations between the coordinates
x, y and x,, yl in the two systems. These relations are analytic forms of
transformations.
11.1. Translation
Let a new set of axes be drawn through 01 at (h, k) parallel to the old
set and with the same sense. If P has coordinates x1, yl in the new set, then
x=x,+h., y=y1+k (1)
are the equations of translation.
11.2. Rotation
Let a new set of perpendicular axes be drawn without change of origin
at an angle a as shown. From the figure
OQ = x1, QP = yJ
and
x=OT - RQ, y =TQ+RP.
145
146 TRANSFORMATIONS Sec. 11.2
*Y P
P
od x,
LY
Ik
0
I
iQ
h
O-g 0 '-g
Fig. 11.1 Fig. 11.2
't'hus
x= x,cosa-y,sina, y= X1 Sill a+ylcosa ` (2)
are equations of rotation.
11.3. Illustrations
We demonstrate effects of such transformations by the following
examples.
(A) Translate the axes to the new origin (2, 3) and find the new
equation of the parabola,
y2-6y-4x+17 = 0
From (l),
x=x,+2, y=y,+3.
Then
y2-6y-4x+17=0
becomes
(y,+3)2- 6(y,+3) -4(x1+2) + 17 = 0
or
y;+6y,+9- 6y,- 18-4x1-8+ 17 =0.
Thus
and the point x = 2, y = 3 is recognized as the vertex of the parabola.

(B) Rotate the axes through 45° and find the new equation of the
Sec. 11.3 TRANSFORMATIONS 147
Fig. 11.3 Fig. 11.4

hyperbola
y2 - x2 = 1
From (2), setting a = 45°, sin a = cos a = V21z, x = (V2/2) (x1 - yj),
and y = (V/2) (x, + y,). Then y2 - x2 = 1 becomes
(x,+y,)2- z(x,-y,)2=1
or
2x,y, = 1
(C) Rotate the axes through 30° and find the new equation of the
ellipse
2x2+d3xy+y2=5
Since sin 30°'= z and cos 30°
= V3/2, the equations of rotation
are x = (d3 x, - y,), y =
z (x, + V3 zy,). The given equation
becomes
2(d3 x, - y,)2
+3(v x,-yl)(x,+VSy,)
+ (x, + d3 y,)2 = 20
from which
5x + yl = 10
Fig. 11.5
TRANSFORMATIONS Sec. 11.3
148
Note that there is no product term x1y1 and that the curve is symmetric to
the new axes. We discuss this important feature in the next paragraph.
EXERCISES
1. Translate the axes in each, of the following to the point indicated and
sketch.
(a) 4y = x + y2; (4, 2)

(b) 3x - 4y = 12; (-1, 2)
(c) 4x2 + 4x - y = 4; (-2, -4)
(d) x2+ y2- 6x+4y 1; (3,-2)
(e) (y - 1)2 = 4(x - 3); (3, 1)
2. Setting x = x, + h and y = yj - k in each of the following, find (h, k) so
that linear terms do not appear. Sketch each.
(a) x2 - 3y2 - 10x + 12y + 12 = 0
(b) 4x2 + 9y2 - Sx + 36y + 4 = 0
(c) 9x2 - 4y2 - 54x - 8y = 41
(d) 3x2 + 3y2 - 12x + 6y - 4 = 0
11.4. Removal of the xy-Term

Consider the general equation
Axe+Bxy+Cy2+Dx+ Ey+F =0, B00 (3)
of the second degree. We show that it is always possible by special choice of

angle a to rotate axes so that the transformed equation does not contain
the product term x,y,.
For this purpose we substitute in (3) the rotation relations
x = x, cos a - y, sin a, y = x, sin a + y1 cos a
and focus attention only on the collection of the coefficient of x1y1. This
collection is
-2A sin a cos a + B (cost a - sine a) + 2C sin a cos a
or
B cos 2a - (A - C) sin 2a.
This coefficient is 0 if, and only if, (A - C) sin 2a = B cos 2a or, A 5,4- C,
B
tail 2a = (4)
A - C
If A = C, cos 2a = 0 or 2a = 90°, and a = 45°.

Thus, since for all real numbers B/(A - C) there exist angles with these
numbers as tangents, then it is always possible to rotate axes to produce from
(3) a new equation with no product term. This new equation will be of the
form
ALxi + CLyi + D,x1 + E1y1 + F = 0 ,
an equation which has already been identified as representing a conic.

It is important to notice that the equations of rotation (2) are linear
in the coordinate variables, and accordingly their substitution in (3) does
not alter degree. We therefore may make the important statement
Each equation of second degree

in rectangular coordinates
represents a conic.
Example. Rotate axes to re-express
66x2 - 24xy + 59y2 + 108x + 94y + 76 = 0
as an equation with no product term; then identify the conic.

Since A = 66, B = - 24, and C = 59, then tan 2a = - ;'. Thus
cos 2a = - = 2 cos' a - 1, and cos a = , sin a = . Equations of
rotation are then
3x1 - 4y1 4x1 + 3y1
x= 5 -, y= 5
On substituting these in the given equation we have

Ya(3x1 - 4y1)2 -4(3x1 - 4y1) (4x1 + 3y1) + 5(4x1 + 3y1)2
+ 1(3x1 - 4y1) + 't(4x1 + 3y1) + 76 = 0
Y
r
-g
24
Fig. 11.6 Fig. 11.7
or, after reduction
(x1 + 8)2 (yl s)2 = 1

+
i T
recognized as an ellipse with center x, = - 7, y, = 5.
EXERCISES
3. Rotate axes for each of the following through the angle indicated and
sketch.
(a) x2 + 4xy + y2 = 2; 45°
(b) xy + x + y = 0; 45°
(c) 4x2 + 4y2 = 17xy; 450
(d) 5x2 + 2V xy + 7y2 = 16; 60° [compare (e)J
(e) 7x2 + 2V3 xy + 5y2 = 16; 30° [compare (d)]
4. Show that the degree of the algebraic equation f(x, y) = 0 is unaltered by
(a) the translation x = x3 + h, y = yl + k
(b) the rotation
{ x = xl cos a - yi sin a, y = xI sin a + yi cos a),
11.5. Identification of Conics

We have just established the fact that in rectangular coordinates each
equation of second degree represents a conic. These conic sections may, of
course, include lines and points as special cases. We turn now to the impor-
tant matter of identification and show that, using a simple device, we can
determine by inspection the type of conic (ellipse, parabola, hyperbola)
when we are given an equation of the second degree.
Thus consider the general equation
Ax2+Bxy+Cy2+Dx+Ey+F =0
This curve is cut by any line of the family

y = mx
in at most two points. The x-coordinates of such points are given, after
eliminating y, by
(A+Bm+Cm2)x2+ (D+Em)x+F =0.
Now, there may be no value x-
or one-or two. We focus attention IY
on only those lines y = mx which
cut but once. If this is the case,
A+Bm+Cm2=0.
Notice now the distinctions in this
respect among the three types:
For the ellipse, there is no line
(other than tangents which yield
two equal values) of the family Fig. 11.8
which cuts the curve just once. The
preceding quadratic in m thus has no real roots. Accordingly,
B2- 4AC<0.
For the hyperbola, there are two lines (those parallel to the guide lines
of the curve) which cut the curve just once. Thus
B2-4AC>0.
For the parabola, there is but one line which cuts the curve just once
(the line parallel to the axis of symmetry). Thus
B2-4AC=0.
We may consolidate this form, called the discriminant, and its results
into
B2 - 4AC<,=,>0
e, p, h
for convenient reference in applications.

Examples.
(a)
xy=1; A=0, B=1, C=O. Thus B2 -4AC,=1>0
Hyperbola.
(b)
x2 + 2y2 = x; A = 1, B = 0, C = 2; B2 - 4AC = -8 < 0
Ellipse.
(c)
x2 + 2xy + y2 = x ; A = 1, B = 2, C = 1; BZ - 4AC = 0
Parabola.
EXERCISES
5. Identify the following conics, rotate axes to remove the product term, and
sketch
(a) x2 + 6xy + 9y2 - 4x = 0
(b) (x + 2y)2 = x - y
(c) xy = x - y
(d) 7x2 + 2V. xy + 5y2 = 8
(e) 2x2+ 4xy+ 5y2+ 4x+ 16y+ 2 = 0
(f) 6x2 + 12xy + y2 = 1
6. Show that (ax + by) (cx + dy) = 1 represents a hyperbola.
7. Show that
(a) y = x f x- 5x is a hyperbola
(b) 2y = - x f x - I is a parabola
(c) y = x f 4 - x2 is an ellipse
and graph each.
*11.6. Diameters of Conics

Consider the conic
f(x,y) =Ax2+2Bxy+Cy2+2Dx+2Ey+F=0
where not all coefficients of the second degree terms are 0, and the family
of parallel lines
{x = xo+Xt, y = yo+.ut), X2+M2 =1 (2)
through points (x, y) of the conic

and arbitrary points (xo, yo). Here
[X, ,u] are direction cosines, and thus
t measures directed distances along (xa, Yo)
each line from (xo, yo) to points on
the curve.
We eliminate x, y between equa-
tions of line and curve to find
(x,y)
f (xo, yo) + 2[AXxo + BAxo
+ BXyo + Cµyo + DX + EA]t
+ [AX2 + 2BXA + C42]t2 = 0. (3)
If we wish the locus of those points Fig. 11.9
(xo, yo) which are midpoints of such
parallel chords, then the preceding quadratic in t must have roots ti t2
or ti + t2 = 0. The coefficient of t,
(A xo + Byo + D)X + (Bxo+Cyo+E)µ =0 (4)
is thus 0. Since this equation is of the first degree in no, yo, the locus is evi-
dently a line. It contains all midpoints of the given family of chords and
thus must contain the center of the conic-if there is a center. It is called a
diameter of the conic.
The form Ax2 + 2Bxy + Cy2 + 2Dx + 2Ey + F = 0 may be displayed as
x(Ax+By+D)+y(Bx+Cy+E)+(Dx+Ey+F)=O.
Or (for the student who has knowledge of matrices) in the matrix product form:
AB 1) x
(x, y, 1) - B C F, y = 0
UEF 1
Let us now assign various sets of values to [X, u] thus obtaining various
other diameters. All of these diameters must pass through the intersection
(if any) of
Axo+Byo+D=0 and Bxo+Cyo+E=0. (5)
1 ,'or, these are coefficients of X, , in (4), and with their common solution
(x,,, yo), arbitrary values of X and µ satisfy (4).
If the determinant
A B
a= =AC-B2
B C
is not 0, we may solve for xo, yo and these are coordinates of the center of
the conic. Thus, dropping subscripts, the center is
D B A D
El
x = - EA C , y = - IB a
If a = 0, the conic has no center.
To illustrate, consider the conic
x2+2xy- y2+4x-6y- 1 =0.
.g Identifying coefficients by compari-
son with (1),
A = B = 1, C = - 1,
D = 2, E _ - 3.
Thus
1 1
a= -2,
Fig. 11.10 1 ,1
and
1 2
X =
1 1 -3 5
y
2' -2 2
is the center. The student should now translate axes to the center (4, --)
as new origin and see that linear terms disappear. Conversely, if D = E _
0, there are no linear terms and (if A 0) the center is at the origin.
If A = AC - Ii2 = 0, there is no unique solution of Equations (5).
The conic is then called non-central.
EXERCISES
8. For each of the following conics find the diameter which bisects chords
with the given direction [X, µ] and its center, if any.
(a) y2 = x; X = {c = 1/V2
(b) x2 + y2 = x; X = I, µ = V3/2
(c) x2+ 2xy - y2= 1; X -1,u=.3/2
(d) x2+2xy+y2= x;X -1,µ=v/2
(e) x2+xy+y2- 6x+1= 0; A= -V3/2, R= 3
(f) x2 - 9y2 + 12x - 36y + 9 = 0; A = 1/v, µ
(g) x2 - 4y = 0; A = 2A
(h) b2x2 + ay = a2b2; A = mA
(i) b2x2 - a'y2 = a2b2; Et = mX
(j) xy = 6; chords: 3x + 4y = k
k11.7. Principal Diameters

A principal diameter of a conic is one that is perpendicular to the family
of parallel chords it bisects. We re-express the Equation (4) of a diameter
in the form
(AX+BA)xo+ (BA+Cp)yo+DA+Eµ = 0. (6)
For this to be a principal diameter, its normals must also have direction
[X, µ]. That is
AA+Bp=kA and BX+CA=kµ
or
(A-k)X+Bµ=0
(7)
BA+ (C-k)µ = 0
If there are numbers A, k (not both 0 since A2 + µ2 = 1) which satisfy
Equations (7) then the determinant of coefficients must be 0. That is
JA - k B
=0
B C-k
or
k2- (A + C) k + A C - B2 = 0
1 56 TRANSFORMATIONS Sec. 11.7
This quadratic in k is called the characteristic equation. Its roots, both

real, are
A+C=1= /(A - 2+4B2
k,, k2 = 2
For each k (inserted in Equations

7) there is a set of numbers X, µ to
determine the equations of the
principal diameters.
Certain facts are now evident.
(1) If the two roots are equal
(not 0), then (A - C)2 +
4B2 = 0, or A = C and B =
0. The conic is thus a circle.
(2) If one root is 0 then AC -
B2 = 0. The conic is thus a
parabola, and the second de-
gree terms form a perfect Fig. 11.11
square.
Illustration. Find the principal diameters and the center of the ellipse
x2+xy+y2- 10x-8y+1 =0
and sketch.
The locus of midpoints of chords
{x=xo+At, y=yo+µt}, X2 + µ2 =1
is the diameter
(2X +µ)x+ (X +21i)y- 10X-8µ =0.
For principal diameters,
2a + µ = kX and X + 2p = kµ
which give the characteristic equation
12-k 1
=0
1 2 -
with roots k = 1, 3.
For
k=1: X =-fit and x-y=2
k=3: X= µ.and x+y=6,
the principal diameters. The center is their intersection (4, 2).
We may shed further light on principal directions and the character of

conics through Equations (7) in the form
AX + Bµ = kX, BA + CIA =kg.
We multiply the first by µ, the second by X and obtain
(AA+Bµ)µ= (BX+Cµ)X
or
B(µ2-X2)+(A-C)aµ=0.
Let one principal diameter have inclination 0 so that X = cos 0 and µ =
sin 0. The preceding equation may then be written as
B(sin2 0 - cos' 0) + (A - C) sin a cos 0 = 0
or
(A - C) sin 20 = 2B cos 20,
which is
2B
tan 20 =
A - C
The angle 0 so determined (as shown in Par. 9.4) is that through which the
coordinate axes may be rotated to remove the product term xy. Since
2 tan 0
tan 20 =
1 - tan2 0'
the last expression may be rewritten as
A
tang 0+ B C tan 0- 1 = 0,
a quadratic in (tan 0) whose roots tan 01, tan 02 have their product -1.
Principal directions are thus orthogonal.
EXERCISES
9. Find equations of the principal diameters and center (if any) of each of
the following conics. Then determine an angle 0 through which axes may be
rotated to remove the product term xy. Identify and sketch each curve.
(a) xy = 4
(b)xy+x+y=0
(c) x2 + 6xy + 9y2 = x
1$$ TRANSFORMATIONS Sec. 11.7
(d) x2 + xy + y2 = I
(e) 5x2 + 6xy + 5y2 = 9
(f) x2+xy=x-y
(g)x+y=1
(h)x-y=1
(i) 3x2 + 4--y = 19z + 10y (Ans. 2x + y = 6, 2x - 4y = 11
0 2 -4 x
0 [Ans. x -]- 2y = 7,
G)(xyl) 2 3 -12 y
2x - y = 4]
)()
(k) (x + y)2 = y - x [Ans. x + y = 0]
(1) (3x ± 4y)2 = 75x [Ans. 3x + 4y = 41
10. Find the directions of maximum and minimum distances from the origin
to the conic
Axe + Bxy + Cy2 = 1, 0 0.
Then show these directions are orthogonal and are in fact principal direc-
tions. Then discuss the case 0 = 0.
12.
*Mapping
Important concepts arise when transformations are considered from a

point of view different from that taken in Chapter 11. Instead of thinking
of points P as fixed under a change in position of the reference axes, we
shall think of retaining the axes in position and changing the location of
points. This is a relative matter. With this point of view, a transformation is
regarded as a mapping of points. We shall consider translations and rota-
tions under this plan and then shall discuss a few important new trans-
formations.
12.1. Translations
The transformation
x,=x+h, y,=y+k
maps points P(x, y) into P, (x,, y,) by adding constants h and k to variable
x, y. This may be regarded as a shift of all points P through a distance
h2 -j k2 in the direction a = Are tan (k/h). For example, all points of
the region
(x+2)2+(y-1)251
are mapped without distortion of the region by the translation
x,=x+5, y,=y+1
159
160 MAPPING Sec. 12.1
Y
Pt{x,,Ytl x
Ik
x
'-g
g
Fig. 12.1 Fig. 12.2
into the region

(x1 - 3)2 + (yl - 2)2 5 1 as shown.
12.2. Rotations
We turn points of the plane
about the origin 0, through the
angle a in the clockwise direction as
shown, mapping P into P1. Let
XOP = 8. Then
x1 = (0P1) cos (0 + a)
= (0P1) (cos 0 cos a - sine sin a)
Y, = (OP3) sin (0 + a)
= (0P1) (sin 0 cos a + cos B sin a).
But
x = (OP) cos 0 = (0P1) cos 8 Fig. 12.3
y = (OP) sin 8 = (0P1) sin 0,
and thus
x cos a - y sin a = x1
x sin a + y cos a = yl
The rectangular display

C09a
R=
sin a
Sec. 12.2 MAPPING 161
called a matrix, characterizes the rotation. We say that the transformation R

carries P into P, and symbolize this as
R(P) --+ P,
or as
COS a - Sin a [ x \ Cxi
sin a cos a y y,
For example, the matrix
rotates points about the origin through the angle (+ 30°), whereas
J3
2 2
V3 t
2 2
rotates points through (- 60°).

Note particularly that this mapping (as in the case of translation) does
not alter lengths or angles; lines map into lines, circles into circles, and
conies into tonics (without change in type). The effect is that of a rigid
motion of the entire plane and any configuration in it. Lines forming a tri-
angle ABC map into lines each turned through the angle a. Lengths OA
Fig. 12.4 Fig. 12.5

and OC are unaltered as is 4AOC = 6. Thus the length of the side AC is

invariant and so is the shape and size of the triangle.
The skew parallelogram is an instrument for plane rotations. Two adja-
cent sides of a jointed parallelogram are replaced by plates that are sec-
tors of circles containing the same angle a. With 0 fixed, motion of the
mechanism rotates points P into Pl through the constant angle a. We omit
proof. If the student wishes to establish the principle, he must show that
the variable distances OP and OPl are always equal and that the angle
POP1 is constant and equal to a.
Translations and rotations to-
gether form a set of transformations
that characterize plane Euclidean
geometry. Before going further, note
that the two transformations are
not commutative. That is, if T repre-
sents a translation and R a rotation,
Z T followed by R is not, in general,
the same as R followed by T. For
example, if P(x, y) is (0, 1), T is a
translation to the right one unit,
and R is a clockwise rotation of
90°, then
T(P) -+ P1(1, 1), R(P1) -+ P2(1, -1)

whereas
R(P) -- Qi(1, 0), T(Q1) - Q2(2, 0).
EXERCISES
1. Map each of the following by the rotation R and translation T; first, R
followed by T, then T followed by R. Draw figures for each order of
transformation.
(a) {0<x<1, 0<y<41;R:a=45°,T:(1,4)

(b) The triangle A : (0, 0), B : (1, 1), C : (3, 0) and its interior; R : a =
-45°, T : (-1, -2).
(c) zy= 1;R:a=45°,T:(1,1)
(d) z2+4y2< 4;R : a = 90°, T : (1, 2)
(e) y2 = x3; R : 90°, T : (1, 2)
12.3. The Affine Linear Transformation

Consider the transformation
x1 = X IX + X 2Y + Xi
Y1 = µ1x + µ2y + µa
with
>11 X2
3 0.*
This is the affine linear transformation, so called because points with finite
coordinates (x, y) are mapped into points with finite coordinates (x,, y1).
The transformation is linear, and the degree of any transformed expression
is unaltered. Thus lines map into lines, conics into conics, and closed curves
into closed curves. Translations and rotations are special cases for certain
values of the X's and µ's. We discuss the following additional special cases.
(A) Similitude.
x, A 0 z
0 < X F& 1
y1 0 X y
or, in polar coordinates,
r1=Xr, 01=0
* If
= 0,
xl As
/11 xl
then
xl xs
µ1 µs
Thus
A1x X2y X1x + A2y
µ1x 121 µ1x + dill
Accordingly,
X1 - XI X1 X2
Y1 - Al µ1 µ2
a line of points (x1, y1). Thus all points (x, y) in the plane map into a single straight line.
This is too peculiar for our consideration.
B Fig. 12.7
-g
Points P(x, y) transform into points P1(xl, y1) collinear with 0 and P.
Regions are stretched if it > 1, contracted if it < 1. To consider shape, it is
evident that changes are proportional; i.e.,
AX, = X(Ax), AY, = it (AY)
and
AY, Ay
AX, Ax
Accordingly, slopes of lines and curves at corresponding points are equal.
Shape, therefore, is preserved, and the transformation is one of similitude.
Points are said to be projected from 0.
A jointed parallelogram with one side extended forms the pantograph,
the basic instrument of similitude. Select arbitrary lengths OA = BC =
a, OC = AB = b, and take Pi on AB extended such that AP = ab,
A > 1. Then mark the point P on BC collinear with 0 and P1. Since
triangles OCP and AP1O are always similar,
OP1 _ OP OP1 OP
or
AP1 OC Ab = b
Thus
OP1 = X (OP) or r1 = Ar
the transformation of similitude.

(B) .Stretch.
0 x
Xp>0,
y1 (10 p X);
The transformation maps points P (x, y) into P, (x,, yj) through a distance
(a - 1)2x2 + (µ 1)2jj2
in the direction I
Are tan [(µ - 1)/(X - 1)] (y/x).
This represents a two-way stretch
(or contraction) in the x- and y-
directions. Degree of a transformed
algebraic expression is invariant.
Conics go into conies and, of par-
ticular importance, the type of
conic is invariant. For, the conic Fig. 12.9
Axe -1- Bxy + Cy2 -F Dx + Ey + F = 0
transforms into
A2 + B-x,yl + u2y,
C2 D E
X2x, + x, + uy, + F = 0.
The discriminant of this transformed conic is
1,2 AC I .(B2-4AC),
lf2-4Azµz = X2,42
a number that has the same sign as B2 - 4AC.

Example. Map the circular region (x - 3)2 + y2 < 4 by
x, 2 0 (x).
y, 0 y
We have
(xi-3)\\2 +(2y')24
l2
or
(xi - 6) 2 A
16 + 1
an elliptical region formed by stretching the circle in the x-direction and

contracting it in the y-direction.
Example. The circle x2 + y2 = a2 is contracted by xi = x, yi = ky,
k < 1 into
y n
X1
Yi-
a2 + k2a2 - 1'
Fig. 12.10 Fig. 12.11
an ellipse with semiaxes a and ka. We note an interesting detail. Tangent

lines drawn to the circle and the ellipse at points having the same x-co-
ordinate have the same x-intercept. This is evident for several reasons.
The x-intercept is independent of the y-coordinate; and, if the circle and its
tangent are "turned" about the x-axis, the configuration may be viewed as
the ellipse. The line rotates about its intercept point and remains tangent.
This relation between tangents to the two curves affords a simple con-
struction for the tangent to the ellipse (see Problem 5, Chapter 9).
(C) The transformation
x1 Al x2 x X1 A2
y1 121 122 J Al 122
leaves the origin unchanged. Moreover, lines through the origin map into
lines through the origin. In particular, the x-axis (y = 0) maps into
xl = Xx, yl = 12x or
X1y1 - 121x1 = 0 f
the y-axis (x = 0) into
Fig. 12.12 X2y1 - 122x1 = 0
If we wish the x, y-axes to map into perpendicular lines, then X1y1 -

121x1 = 0 must be perpendicular to X2y1 - 122x1 = 0; i.e.,
X1X2 + 121122 = 0
a condition both necessary and sufficient. The transformation is then called

orthogonal.
If [AI, u,] and [A2, 92] are direction cosines of the x, y-axes measured
with respect to the x1, y1-axes, then
Ai + ui = 1 and A2 + µz = 1.
We may determine the inverse of the orthogonal transformation
x = Alxl + A2y1, y = uixl + 02yt
by solving for x1 and y1 in terms of x and y. To this end, multiply x by A1,

y by u1, and add. We have
x1 = A1x + µ1y.
Multiplying x by A2, y by 112, and adding,
y1 = A2x + u2y.
Both sets of equations of transformations-the given set and its inverse-

can be read conveniently from the table:
X1 Y1
x Al A2
y u1 u2
EXERCISES
2. Map each of the following by the transformation indicated. Draw a figure
for each and describe the mapping in words.
(a) (y>x2,y<x2+11;
(c)
(d) r-
(b) r = 1 + cos 0; r = r1i e = 01
{x5oiv<'J;
(e) 1<x2+y2<4;
2
0 -3
0
it
0
1
0
0
2
;2r= r1,9=01
0
(f) 0 < 4x2 -!- 9y2 < 36;
0 2
(g) 4x2 - 9y2 = 36;
3. Show that the transformation x = Ax1, y = tsyl is equivalent to

r = v2 + X tan 9 = µ tan 91
4. The transformation
x = 2x1 - 1 = 2(x1 - a),
y = 3y1 - 2 = 3(y1 -
may be regarded as a stretch S followed by a translation T, or an appro-
priate translation T followed by a stretch S. Draw figures for each of the
following to show the effect ST, then TS.
(a) x2 + y2 < 1
(b){-1<x<1,0<y<2}
(c) {yx2,y<2}
(d) {x= Isint1,y= cost
(e) {x= 1- cost,y= t- sin t),0<t<2,r
(f) {x= t - cost,y= 1 -sint},0<t<22r
4 -3
5. Use the transformation
3 4
to map each of the following. Figure required.

(a) {0<x<3,0<y<11
(b) ( y<sin x,0<x<ir)
(c) 9x2 + 16y2 < 144
(d) {xy<sinx,-2r<x<ir,x*0}
6. Express the transformation of Problem 5 in the polar coordinates (r, 0)
and (r1, 01).
7. Show that
-1 0
0 1
is a reflection of all points in the y-axis; that
is a reflection in the x-axis. Then apply the first transformation to y

4 - x2; the second to y2 = 4 - x.
8. Show that
-1 0
0 -1
is a rotation of 180° about the origin 0. Then show, for example, that the
Serpentine y = x/(x2 + 1) has symmetry with respect to 0.
9. Describe effects of the transformation
10. The transformation
gives a shearing effect parallel to the y-axis; while
is a shear parallel to the x-axis. Discuss the effect of these shears on the
region (0< x< 2,0<y< 11 fork= 2;fork= z.
12.4. Inversion
Of particular importance in some applications is the transformation
rr1 = k2, 0 = 81
called inversion. Here the point P1(r1j 01) maps into P(r, 0), a point col-
linear with P1 and the origin 0 and such that the product of distances from 0
is the constant P. That is, (OP) (OP1) = V. The points P and P1 are said
to be mutually inverse with respect to the circle, center at 0 and radius k. We
will take this as the unit circle, setting k = 1 (a scale choice). All points,
except 0, within the circle map into points outside and .vice versa. The unit
circle itself is the locus of fixed points.
Construction. For the construction of the point P inverse to P1 (here
taken outside), draw a tangent P1T to the circle and then the perpendicu-
lar TP to OP1. Since triangles 0 TP and OP1T are similar,
OP_ OT or (OP) (OP1) = 1.
OT OP1'
Fig. 12.13 Fig. 12.14
Inverse Curves. Given a relation in polar coordinates defining a set of

points Pi, the relation defining the set of inverse points P is obtained (since
0 = 01) by the simple expedient of replacing r1 by 1/r. In doing this, the
subscript may be dropped for convenience without danger of confusion
(the reference system remains fixed). Thus, for example, the Cardioid
r=1-cos@
inverts into the parabola
r(1 - cos 0) = 1
and the circle 4r = 3 cos 0 into the line 3r cos 0 = 4.
Unit Unit
circle
P,
Fig. 12.15a Fig. 12.15b
Of special interest are the inverses of circles

r2+ Arcos9+ Brsin0+ C = 0.
Their inverses are
Cr'+Arecos0+Br1sin0+ 1 = 0
which, for C 0 0, are themselves circles. However, if C = 0, the circles
(ignoring the factor r = 0),

r+Acos0+Bsin0=0,
all pass through the origin 0, and invert into the lines
Art cos 0 + Brt sin 0 + 1 = 0.
Some inverse curves are tabulated here.
rr1=1
Circle: Circle:
r2+Arcos0+ Brsin0+ C= 0 Cr12+Artcos0+Brtsin0+1=0
Circle: Line:
r2 + Ar cos 0 + Br sin 0 = 0 Art cos 0 + Br, sin 0 + 1 = 0
Circle: Circle:
r k r1= 1/k
Line: Line:
0=k 0= k
Hyperbola: Lemniscate:
r2 cos 20 = 1 ri = cos 20
The student should provide a sketch of each pair of curves given.

The Inversor. The linkage of Par. 8.6 (D), discovered independently
in 1864 by Peaucellier and Lipkin, permits mechanical inversion and, by the
addition of a radial bar, gives straight line motion. There are six bars
jointed together in a plane as shown such that
OA=OB=a; PA=AQ=QB=BP=b, a>b.
To establish the inversive prop-
erty, draw the circle, center A,
radius b, through P, Q. This meets
the line of the bar OA in X, Y.
By the Power Theorem (Par.
6.3), (OP) (OQ) _ (OX) (OY) =
(a - b) (a + b) = a2 - b2. We set
a2 - b2 = i (by choice) so that
(0p)(00 = 1
which is the fundamental inversive

relation among 0, P, and Q.
Fig. 12.17 Fig. 12.18
For line motion, fix 0 to the plane. Then attach a link CP of arbitrary
length to P and fasten its other end to the plane at C so that i7-O- = CP.
Now use the line CO as polar axis and show analytically that Q has line
motion. Make a model of this using carboard strips and eyelets for joints.
[Compare Par. 8.6 (D)].
The Crossed Parallellogram Inversor. Consider again the linkage
ABCD, BC = AD = a, AB = CD = b, a > b. Select points 0, P, Q, R
on the bars as shown. If these are taken on a line parallel to BD and AC,
they remain so as the parallellogram is deformed.
Now, PQ and AC have a common perpendicular bisector and thus lie on
a circle (that varies in size as the linkage deforms). Let H be the point
of intersection of this circle with the bar AB. By the Power Theorem
(Par. 6.3),
(BH) (BA) = (BP) (BC).
Since P is a point fixed on BC, the length BP is constant. Then, since
the factors (BA), (BP), and (BC) are constants, so also is BH. Thus H is a
point fixed in position on BA, and the variable circles through A, P, Q, C
all pass through H. Using the Power Theorem again,
(OH) (HA) _ (OP) (OQ)
in which the left member is, by the preceding argument, a constant. So
(OP) (OQ) = constant
This is the inversive property of the three collinear points 0, P, Q and

thus this four bar mechanism is entirely equivalent in this respect to the
Peaucellier inversor composed of six bars.
Take 0, P, Q as midpoints of bars. Since the product of the variable distances CP
and CQ is constant, we have the privilege of selecting a special position of the linkage to
evaluate the constant. For this purpose, let the linkage form a rectangle with P, Q
coincident. Then
(0p) (0Q) a' - b=
= (OP)' = 4
Project. Fasten 0 to the plane. Then attach a fifth bar to P and fix its
other end to the plane at a point whose distance from 0 equals the length
of this fifth bar. Q describes a straight line. Build this model selecting
0, P, Q as midpoints of bars and explain.
EXERCISES
11. Find the equation of the curve inverse to each of the following. Sketch
each curve and its inverse.
(a) r= 2 cos 0 (b) r = 3 sin 0
(c) r = sin 0 + cos 0 (d) r sin 0 = 1 + r cos 0
(e) r2 sin 20 = 1 (f) r2 cos 20 = 1
(g) r = 1 + cos 9 (h) r = 1 + sin 0
(i) r(1 + 2 cos 0) = 1 (j) r(2 + cos B) = 1
12.5. The Joukowski Airfoil

The design of an airfoil involves mapping by inversion. The flow of
fluid past a cylinder is studied in detail and then mapped into flow about
a wind surface whose shape gives desirable effects in the flow pattern. This
particular mapping of points P is as follows. The point P (r, B) is inverted to
Q (1/r, 0), then reflected in the polar axis to 0(1/r, - 0) . The mapped point
P, is then located by drawing PPl equal and parallel to OQ and in the same
sense as shown.
Fig. 12.19 Fig. 12.20
For an airfoil, points P on a circle through (-1, 0) and enclosing (1, 0)

are mapped into points Pi. The position of the circle through (-1, 0) deter-
mines the particular shape of the airfoil.
The student should construct an airfoil by this method on polar co-
ordinate paper. Draw a unit circle and choose another through (-1, 0).
Then map points of this circle sufficient in number to give shape to the air-
foil.
II.
THREE
SPACE
13.
Coordinate Systems
-Measures
The mathematical machinery of three-space geometry is for the most

part a natural extension of that in plane geometry. We shall use this
machinery-the formulas and relations-to study surfaces and space
curves and their properties. Here the development of spacial perception is
of prime importance; the student will discover that a mental picture of a
spacial configuration is a first requi-
site for an analytic treatment. Our
discussion starts with reference sys-
tems of coordinates by means of
which measurements can be made
and geometrical forms and princi-
ples may be formulated in algebraic
terms.
13.1. Coordinate Systems

(A) Rectangular Coordinates. We Fig. 13.1
choose three directed and mutually
perpendicular lines intersecting at 0 (the origin) for coordinate axes. The
three pairs of these lines determine three coordinate planes. On each axis
177
178 COORDINATE SYSTEMS-MEASURES Sec. 13.1
there are assigned + and - directions. Any point P then is uniquely lo-
cated by the ordered set (x, y, z) of numbers representing directed dis-
tances from the coordinate planes, and conversely.
The display of a three-dimensional geometry on a plane may perhaps
seem strange at first. To help himself, the student might use the three lines
of intersection of floor and walls in the corner of a room to visualize the
situation. Pictures drawn on a book page are ìn perspective'; here the
yz-plane is the plane of the page and the x-axis is perpendicular to the
page.
Space is separated by these coordinate planes into octants, in each of
which are located points whose
coordinates (x, y, z) have one of the
eight arrangements of plus and
minus signs.
(B) Cylindrical Coordinates. A
useful set of coordinates of points P
in three-space is called cylindrical.
This system is a hybrid composed of
polar coordinates in a reference plane
(e.g., the xy-plane) together with a
directed distance to the point from Fig. 13.2
that plane. As shown, we choose the
xy-plane as the polar coordinate plane; then any point P has coordinates
r, 0, z.
We may exchange the rectangular system for this cylindrical system
by means of the equations of transformation
x = rcos0, y = rsin0, z = z
Thus, for example, the equations

x2 + y2 + z2 = 1 and r2 + z2 = 1
Z are equivalent, the first in rec-
i
tangular, the second in cylindrical
coordinates.
F--- /P pi /
/, x-i_'r! I
I
(C) Spherical Coordinates. A
third convenient system, of spher-
ical coordinates, locates a point P
by its distance p from a fixed point
0 (the origin) and angles ', meas-
ured from the positive z-axis to
OP(=p), and 0, measured from the
Fig. 13.3 positive x-axis to the projection OQ
Sec. 13.1 COORDINATE SYSTEMS-MEASURES 179
of OP on the xy-plane. (Geographical terms for these angles are longitude

and co-latitude, if the xy-plane is that of the Equator.)
We shall agree to restrict the angles so that
-2ir5B<21r; 0 5cp <a.
The radial distance p may be either positive or negative.
An exchange is now possible from rectangular to polar coordinates and
vice versa. From the picture OQ = p sin (D and
x= p sin 4) cos B, y = p sin 4' sin 0, z = p cos 4'
Note that x2 + y2 = p2 sin2 4) and x2 + y2 + z2 = p2. Thus, for example,

the equations
x2+y2+z2 - 6x =0 and p = 6sin4cos0
are equivalent, the first in rectangular, the second in polar coordinates.
EXERCISES
1. Complete the following table.
Rectangular Cylindrical Spherical
x2+Y2+ z2= 2x - 4y
r = z(l + cos 9)
p = 8 cos -
z_
x2+J2
_
r= 1-sing
p sin2 4' = 6 cos 8
xy = z I
z sin 0 = 3r
p sin 4' = tan 6

180 COORDINATE SYSTEMS-MEASURES Sec. 13.2
13.2. Distance
Direction. Consider two points
1'1(x1, y1, Si) and P2(x2, y2, z2)
Through P1 draw lines parallel to
the x, y, z-axes, and project the
directed length P1P2 onto these
lines as shown. The projected
lengths are x2 - x1; y2 - y1; zs - z1
These are the three dimensions of
the rectangular `box' of which
PiP2 is a diagonal. Accordingly, the Fig. 13.4
distance P1 to P2 is
P1P2 _ (x2 - x1)2 + (y2 - y1)2 + (Si - z1)
Direction Cosines. Let angles made by Pi-P2 with the three lines drawn
parallel to the axes be a, 0, y, respectively, each numerically not greater
than r. Then from the right triangles pictured,
x2 - x1 y2 - y1. z2 - z1
cos Of = I cos =I cos y = I I.
p1 P2 P1P2 1; PiP2
These numbers are called direction cosines of the directed line segment
P1P2 (The student should read again Par. 2.2 where these matters are
discussed in connection with the line in plane geometry.)
If we form the sum of their squares,
(x2 - x1)2 + (y2 - y1)2 + (Si - Zl)2
cost a + cost rg + cost y = ,
I P1P2 I t
we find
Cost a + cost 0 + cost y = 1
a fundamental relation.
Direction Numbers. Let three numbers [X, µ, v] be proportional to the
direction cosines of 751P2; i.e.,
X = k cos a, µ = k cos fl, v = k cos y.
For random positive values of k (the factor of proportionality) there are

many such triples. For all, however, there corresponds but one set of
direction cosines. This is evident on summing the squares
X2+p2+v2 =kt (costa+cos'/3+costy) =k2,
from which
A
cos a
0
cos $ +v2,
v
cos y = /X2 + µ2 + v2
Then, since direction cosines may be obtained from direction numbers, the
latter triples properly establish the sense and direction of a line segment.
Thus, for example, if
X = 3, µ = - 4, v = 5,
then
X2 + µ2 + P2 = 50,
and
cos a =
3
cos
-4 cosy
5
the sum of whose squares is unity. Any other triple, proportional to [3,
-4, 5], such as
[ , `-2, 5]; [6, -8, 10]; [1, -s, $]; [9, -12, 15],
leads to the same set of direction cosines. If k is negative, the sense, not the
direction, is changed.
Direction numbers and cosines are a basic concept and will play an
important role in all discussions of three space.
13.3. Angle Between Two Line Segments

If a line is moved parallel to it-
self, its direction angles are not
changed. With this in mind, we de-
fine the angle between two lines in
space as the angle formed by their
parallels drawn through the origin.
Thus consider two lines with direc-
tion angles al, #I, yl, and a2, $2, 72
which form an angle 0. If

Pi (xl, Yi, Si) and P2 (X2, Y2, Z2) are
Fig. 13.5
points on their parallels (with the
COORDINATE SYSTEMS-MEASURES Sec. 13.3
1gq
same sense) through the origin and distant r1, r2 from 0 and distant d
from one another, then, by the law of cosines
da = ri + r2 - 2r1r2 cos 8
or
(X2 - x1)2 + (y2 - Y1)2 + (Z2 - Z1)2
=x2+y22+z2+xi+y2+zi-2r1raCOS8.
Reducing,
cos o =+-- +
x1x2
r,r2
ylya
r1r2
Z1Z2
r1r2
or OS 0 = COS al COS a2 + COS 01 COS 02 + COS 'Y1 COs y2
As we did in Par. 2.3, here also we shall name the right member the scalar
product of the two sets [cos a,, cos 01, cos y1] and [cos a2, cos, 2, COS 721-
13-4. Parallelism
It is at once evident that if two line segments are parallel, their direc-
tion angles and cosines are equal, and conversely. Furthermore, all sets of
direction numbers of the two parallel lines are proportional. Thus, for
example, if
[-1, 2, -3] and [-4, 8, -12]
are direction numbers of two lines, they are parallel.
Perpendicularity. If two lines are perpendicular, their angle 0 is 7/2
(or -vr/2) and cos 8 = 0. From the preceding form for cos 8,
COS a, COS a2 + COS 01 COS 02 + COS 71 COS 72 = 0.
Conversely, if this scalar product is 0, then cos 0 = 0, and 0 is an odd mul-

tiple of 7r/2.
If [X,, t11, v1] and [X2, µ2, v2] are direction numbers of two perpendicular
line segments, the preceding zero scalar product can be written, since
X, = k cos a,, etc., as
X1X2 1.1)22 1112 _
k2 k2 k2
-0
or X1X2 + M1M2 + V1P2 = 0
the necessary and sufficient condition for perpendicularity. Thus, for example,
the two lines with direction numbers
[-1, 2, -3] and [12, 9, 2]
are, since
(-1)(12) + (2)(9) + (-3)(2) = 0,
perpendicular.
EXERCISES
2. Find the distance between the following pairs of points:
(a) (2, 3, 1); (-4, 5, -1) (b) (-5, 2, -2); (5, -2, 2)
(c) (-6, 1, 1); (6, -2f -1) (d) (a, b, c); (b, c, a)
(e) (-3, 4, -2); (3, -1, 1) (f) (1, 1, 1); (8, -7, -6)
3. Find the midpoint of each line segment in Problem 2.
4. Find direction numbers and direction cosines of each line segment in
Problem 2.
5. Find the angles formed by lines from the origin to
(a) (0, -4, 3), (-4, 0, 3) (b) (-2, -1, 5), (2, 1, 4)
(c) (6, -2, 0), (4, -1, 3) (d) (3, -4, 1), (1, -5, 2)
6. Describe the set of all points for which
(a) y = 90° (b) a + 13 = 90° (c) a = 30° (d) a = y
7. Show that the following pairs of points are collinear with the origin:
(a) (-1, 3, 4), (2, -6, -8) (b) (1, 1, -5), (-4, -6, 30).
8. Show that the three lines whose direction numbers are [3, -2, 4]; [4, 8, 1];
[-34, 13, 32] are mutually perpendicular.
9. Give a set of direction numbers of a line perpendicular to one whose direc-
tion angles are a = 30°, # = 60°.
10. Describe the set of points for which
(a) x = 4 (b) x=-3 (c) x = y
(d) y2<1 (e) x254 (f) z>2
(g) r = 5 (h) 4, = a/4 (i) cos 0 = 0.5
(j) r cos 0 = 2.
11. For what value of k is the line segment with direction numbers [-k, 3, k]
perpendicular to a segment with direction numbers [3, 2, 5]?
12. Find the angles formed by the pair of lines with direction numbers
(a) [-1, 3, 2]; [2, 5, 01 (b) [1, -2, 31; [0, 1, -2].
13. For what values of x and y will the points (x, 1, 2), (3, -2, 4), (-2, y, 1)
be collinear?
14.
Planes And Lines
14.1. Definition
The plane is an undefined element in the structure of geometry. For our
purposes, it is the set of points P whose coordinates x, y, z are values of func-
tions
Ix = a,s + bit + ci, y = a2s + b2t + C2, z = a3s + b3t + C31
of the first degree in two independent variables s and t. * Thus, for example,
(A) {x=t+s, y= t-s, z= 2t+3)
(B) {x=2t-1, y=1-3s, z=4t+12sI
are planes.
We may eliminate s and t among the equations defining x, y, and z
to produce the rectangular equation of the plane. For the two examples,
(A) x + y = 2t =z - 3,
and thus
x+y-z+3=0.
x + 1 1 - y
3,
(B) t = 2 4 t + 12s = z,
" The variables s and t are independent if s Pd f (t) or t 0 g(s).

184
Sec. 14.1 PLANES AND LINES 185
and thus
2x - 4y - z + 6 = 0.
Generally, the elimination of the variables s and t yields a single equation
of the first degree in x, y, and z:
Ax + By + Cz + D = 0
which may be used as an alternate definition of a plane. That is, a plane

is the set of points whose coordinates
x, y, and z satisfy an equation of the
first degree.
Traces and Intercepts. Those
points in a plane Ax + By + Cz +
D = 0 for which z = 0 all lie in the
xy-coordinate plane. The line of
these points
Ax + By + D = 0
is called a trace of the plane. Traces
in the xz- and yz-planes are those Fig. 14.1
points for which y = 0 and x = 0,
respectively; i.e.,
Ax+Cz+D =0
and
By+Cz+D =0.
Points of the given plane that lie on the coordinate axes are (x, 0, 0),
(0, y, 0), (0, 0, z). These mark intercepts of the plane and, of course, are
intersections of traces. Intercepts and traces should be drawn to display
the plane.
EXERCISES
1. Sketch the following planes, giving traces and intercepts of each.
(a) {x=t,y=2s,z=s+t}
(b){x=-t,y=t+s,z=s}
(c) {x=t,y=t-s,z=s}
(d) {x= 1-i,y= 1-s,z=s+t}
(e) {x= l+s+t,y= 1- s+t,z= 1+s-t}
(f) {x= 2 +t,y= 3+8,z= o}
PLANES AND LINES Sec. 14.1
186
(g) 12x = t, 3y = s, i z = 4s + 2t - 121

(h) {3x = s- 1, 2y= 3s - t, 5z = 41+6}
(i) 2x - y + 3z = 1
(j) x+2y+3z= 6
(k) x+3y= 4
(1) y + 2z = 5
(m)x+4z= 6
(n)x+y+z=0
2. Locate all points
(a) equidistant from (0, 0, 0) and (0, 6, 0)
(b) equidistant from (0, 0, -2) and (0, 0, 7)
(c) where z = 4
(d) for which x + y - 1 = 0.
3. What angle is made by
(a) y = 2x with x = 0? (b) y = VZz with y = 0?
14.2. Definition.
A line is the set of points P whose coordinates x, y, and z are values of
functions
(x =alt+b1, y =all+b2, z =aal+ba}
of the first degree in one variable t. Thus, for example,
(A) {x=t+1, y=2t, z=1- :3t}
(B) Y =
t-1 z = 6t+5}
3 ,
are lines.
The variable t may be eliminated (in many ways) in such sets of equa-
tions to produce rectangular equations of the line. For the given examples,
we may find
(A)
(x -1= 2 = 2 -z or the independent pair
{x+y+2z-3=0, x-2y-2z+1 =0}.

(B) {2x = 3y + 1 =
z
6 5}, or the pair
{z=12x+5, 3y+1=2x}.
Since all equations here are of the first degree in x, y, and z, each repre-
sents a plane, and the line is the set of points common to those planes. The
line is thus the intersection of the planes.
Generally, the elimination of t yields a pair of independent equations
{Aix + Eiy + Ciz + Di = 0, A2x + B2y + C2z + D2 = 01 1
Accordingly, we have the alternate definition: a line is the set of points com-
mon to two planes; the line is their intersection.
A line may be pictured by find-
ing points where it pierces the co-
ordinate planes. For instance, the
line I X110,5,31
{x+y-z-2=0,
2x + y + 3z - 14 =0}
pierces the yz-plane where x = 0;
i.e., in the point of intersection of
y - z - 2 = Oandy + 3z - 14 = 0
which is at y = 5, z = 3. The line Fig. 14.2
pierces the xz-plane where y = 0;
i.e., at the point of intersection of x - z - 2 - 0 and 2x + 3z - 14 = 0
which is at x = 4, z = 2. The join of these points (0, 5, 3) and (4, 0, 2) is
shown.
EXERCISES
4. Draw the given line and find a set of its rectangular equations.
(a) {x=t- 1,y=2-f-3t,z= 1-4t}
(b) {2x= i+t,3y= 2 - 3t,z= 2t}
188 PLANES AND LINES Sec. 14.2
(c) {x = 1 + (t/2), y = 1 - (t/3), z = (t/6)}

(d) {x=2,y=t,z=4-t}
(d) {x+ 4 = 3t, y - 2 = 5t, z= 15t+4}
(f) {x + y = 2t, x - y = 3t, z - x = 4t}
5. Draw each of the following lines, finding points of the line where it pierces
the coordinate planes.
(a) {x+y-1=0,z=2}
(h) {x+y - 1=0,y=2}
(c) {x+ y- 1 = 0,x+y+z= 2}
(d) {2x- y+4z- 1 = 0,x+2y-3z- 2= 01
(e) {x+y+z-3=0,y+2z-4=0}
(f) {3x+2y-4z-6=0,x-y+z-3=0}
14.3. Families of Planes Through a Line. Projection Planes of a
Line
Consider the fancily of planes
A,x+B,y+Ciz+D,+k(A2x+L'2y+C2z+D2) =0
where k is any real number. All points P on the line
{AIx+B,y+CIz+D,=0, A2x+B2y+C2z+D2=01
have coordinates which satisfy the equation of the family. Thus each plane
of the family passes through the line. Among the planes in this family are
three, called the projection planes of the line, which stand perpendicular to
the coordinate planes. These are found by selecting values of k that elimi-
Fig. 14.3 Fig. 14.4

pate each of the variables in turn. For example, the line

{5x+2y+5z-26 =0, 5x-2y+3z- 14 =01
defines the family of planes
5x + 2y + 5z - 26 + k(5x - 2y + 3z - 14) =0.
For k = -q, 1, -1 we have
15x - 8y + 4 = 0; 5x + 4z - 20 = 0; 2y + z - 6 = 0};
the three projection planes of the line. The student will find these planes
useful as shown in the pictorial representation of the line.
EXERCISES
6. By considering a family of planes, find projection planes for the lines of
Problems 4 and 5. Draw these planes and the lines they determine.
7. Find by construction the projection in the [+ + +] octant of the region
Ix +2y-250,y-x+150,y-3x+350,2y-x-250,2=0}
onto the plane x + y + z = 4. (This represents an important part of a
typical problem in the field of "Linear Optima.")
14.4. Normal to a Plane

Let P,(xi, yi, zi) and P2 (X2, Y2, Z2)
be any two points of the plane
Ax + By + Cz + D = 0.
Then
Axi + By, + Czt + D = 0,
and
Axe+By2+Cz2+D = 0. Fig. 14.5
By subtraction,
A(xi-x2) +B(y,- Y2) +C(zt-z2) = 0,
the form of a scalar product. The triple [A, B, C] of constants is a set of
direction numbers of a fixed direction; the triple
[(x1 -X2), (. - y2), (xt - Z2) ]
is a set of direction numbers of a variable line segment lying wholly in the
plane. The lines with direction numbers [A, B, C] are thus perpendicular to
all lines in the given plane, and conversely. These lines are called normal
to the plane.
We may use the preceding principle to find, for example, the equation
of the plane containing the point (1, 2, 3) and having a normal with direc-
tion numbers [5, -4, 6].
Let P(x, y, z) be representative of all points in the plane. Then all
lines with direction numbers
[x- 1, y-2, z-3]
are perpendicular to the normal direction [5, -4, 6]. Accordingly,
5(x- 1) -4(y-2) +6(z-3) =0,
the equation of the plane containing those points, and only those, whose
coordinates satisfy the equation.
EXERCISES
8. Find a set of direction numbers for the normal direction of each plane of
Problem 1.
9. Use the concept of scalar product to find the equation of the plane through
the given point with the given direction numbers for its normal. Sketch
each plane.
(a) (0, 0, 0) [1, 1, 1] (h) (0, 0, 0) [-1, 0, 1]
(c) (1, 0, 0) [2, - 1, 0] (d) (0, 0, 1) [1, 2, 3]
(e) (2, -1, 1) [2, -1, 1] (f) (1, 1, 4) [-1, 1, 1
14.5. Distance from Point to Plane

Consider the plane
x(cos a) + y(cos 0) + z(cos y) _ p (1)
and let h (Fig. 14.6) represent the perpendicular distance from the origin to
the plane at (x, y, z). Then h, being normal to the plane, has direction
cosines [cos a, cos j3, cos y], and thus
x = h cos a, y = h cos,B, z = h cosy.
From the equation of the plane, then,
h(cos2a+COS2/3+cos2y) = p.
Thus h = p; that is, the given plane is p units distant from the origin.
Let Pl (x,, y,, z1) (Fig. 14.7) be a point d units distant (in the same direc-
tion as p) from the plane
xCosa+ycos/3+zcosy = p.
Fig. 14.6 Fig. 14.7
The plane parallel to this through P, is, since it is perpendicular to the

same normal direction,
(x- x1)Cosa+(y-y1)cos$+(z-z1)cosy=0
or
xcosa+ycos0+zcosy =x,cosa+y,cos0+z1cosy.
But since this second plane is p + d units distant from the origin, its
equation can be written as
xcosa+ycos+zcosy =p+d. (2)
Thus
p + d = x, cos a+y,cos0+z1cos-1
or
d = x, cos a + y1Cos13+z,cosy - p (3)
an expression for the directed distance d from (x,, y,, z1) to the plane (1).
If the plane is given in the form Ax + By + Cz + D = 0, this may
be rewritten in the form (1) by dividing throughout by the square root
of the sum of the squares of the coefficients A, B, and C to reduce them to
direction cosines. Thus
A B C
2+B2+ 2x+',/ +B2+C2
+VA2+B2+ 2z
D
= 0.
+ 2 + B -+-U2
PLANES AND LINES Sec. 14.5
192
The distance to the plane from the origin is
-D
P _ /A2 + B2+ 2
and, using (3), the distance to the plane from (x,, y,, z,),
d=
V4T1
EXERCISES
10. Write the equations of the three planes all with normal direction
[2, -3, -5] which are distant, respectively, 7, -4, and k units from the
origin.
11. Find the distance from (2, 3, 4) to each of the planes of Problem 1.
12. Find the distance from (2, 3, 4) to
(a) {.r.+y-1=0,z=2}
(b) {x+y-1=0,x+y+z-2=0)
14.6. Determination of Equations of Planes
Since the general equation of a plane
Ax + By + Cz + D = 0
contains three essential independent constants, a plane is determined by
three independent conditions; for example, by three points, or a line and a
point, or two intersecting lines, or its normal direction and a point. We
clarify these ideas with the following illustrations.
(1) Find the equation of the plane through the non-collinear points
(1, 2, 0), (2, -1, 7), (-1, 3, 1). Since each set of coordinates must satisfy
the equation Ax + By + Cz + D = 0 of the plane, we have
A+2B +D =0
2A - B+7C+D =O
-A+3B+ C+D=0.
Solving for A, B, C in terms of D,
A -- ID, B =-ID, C=-}D
the equation is then

-sDx - BDy - jDz + D = 0
or
2x + 3y + z - 8 = 0.
(The student invariably should check that the given coordinates actually
satisfy the equation presented as the solution of the problem).
Problem. Show that the equation of the plane is represented by the
determinant relation
x y z 1
1 2 0 1
= 0.
2 -1 7 1
-1 3 1 1
Note the following. For given points P1(x1, y1, z1), P2(x2, y2, z2), and
P3(xa, ya, z3) the equation
=0
X-3 ya z3 11
represents the plane through P1, P2, P3. For, it is of first degree in x, y, z
and, moreover, is satisfied if x, y, z is replaced by coordinates of P1, P2, Pa
in turn. Thus Pi, P2, P3 are among the set of points in the plane.
(2) Find the equation of the plane through (5, 2, 1) parallel to 3x -
4y - 6z - 7 = 0. All planes parallel to the given plane have equations
3x-4y-6z+D=0.
The one through (5, 2, 1) is such that
3(5) - 4(2) - 6(1) + D = 0
or D = -1. Thus the required equation is
3x - 4y - 6z - 1 = 0.
(3) Find the equation of the plane with trace x + 3z = 12 which forms
with the coordinate axes a tetrahedron of volume 96 units. Planes
x+By+3z = 12
all have the given trace. Their x-, y-, z-intercepts are 12, 12/B, 4. The
volume of the tetrahedron (one-third base area times height) formed with
the axes then is
3(12)( 12)(4)2 = 96.

Thus B = 1 and
x+y+3z = 12
is the required plane.
EXERCISES
13. For each of the following, find the equation of the plane containing the
given set of three points. Sketch each plane.
(a) (1, 0, 1), (2, -1, 0), (5, 4, 3)
(b) (0, 0, 1), (1, 0, - 2), (3, - 1, - 2)
(c) (5, 7, -1), (2, ', -2), (1, 1, 1)
(d) (2, 1, 0), (1, 0, 2), (0, 1, 2)
14. The points of each of the following sets are collinear. Show that the equa-
tion of the plane through them cannot be uniquely determined. Sketch.
(a) (1, 2, 3), (5, - 2, 7), (2, 1, 4)
(b) (-2, 1, 2), (0, 4, 6), (4, 10, 14).
15. Show by using direction numbers that the points of Exercise 14 are
collinear.
16. Show that the following pairs of planes are perpendicular. Sketch each
pair.
(a) 2x+3y+4z= 1 (b) x- y- z= 1
x+2y-2z=7 x+4y-3z=6
(c) x- jy- 3z= 4 (d) 10x- 8y+ 15z = 4
3x+4y+1z=5 x+5y+ 2z=3.
17. Find the value of k for which 3x + ky + 3z = 1 is perpendicular to
x - y+2z= 3.
18. Find the point of intersection of each of the three planes
(a) x- y+ z- 1 = 0 (b) 2x+5y- 10z= 30
x+2y+3z- 6= 0 x+ y+ z= 8
4x- y-2z+8=0 x-2y- 3z= 1
19. Find the point of intersection of each of the lines

X- 1
(a) { 2 =y-2=z+3
(b) {x= t, y= 1+2t, z 1-1}
with the plane 2x - y - 4z = 8.
14.7. Direction of a Line

Let the line
{A1x+Bly-f- C,z+D, =0, A2x+B2y+C2z+D, =0}
have direction numbers [a, b, c].
Since this line is perpendicular to
the normals to both planes, the [A,.B,,Ci]
scalar product of their direction
numbers is 0. Thus
aA1 + bB, + cC1 = 0
and Fig. 14.8
aA2 + bB2 + cC2 = 0.
From these equations it is generally possible to solve for the ratios of
a, b, c. We illustrate.
Find direction numbers [a, b, c] of the line
12x-y-z+3 =0, 2x+y-2z+2=0).
We have
2a - b - c = 0
2a + b - 2c = 0
from which
a=4c, b=jc
so that
a b c
4 1 1.
Direction numbers of the given line are thus C4, , 1] or [3, 2, 4]; etc.
14.8. Determination of Equations of Lines

Since the general equations of a line
{Aix+ B,y+ C1z+D, = 0, A2x+P,2y+ C,2z+D2 = 01
together contain two essential in-

dependent constants, a line is de-
termined by two conditions. Illus-
trations follow.
(1) Find equations of the line
with direction numbers [3, 4, 5]
which passes through (2, 1, 6).
Let P(x, y, z) be representative of
all points on the required line. Then
[x - 2, y - 1, z - 6] are direction
numbers of all segments of the line.
Fig. 14.9 These must be proportional to
3, 4, 5. That is
{x-2=3t, y-1 =41, z-6=5t}

where t is an arbitrary constant of proportionality. Upon its elimination,
{--
x- 2
3- =
y
4
1_ z - 6
5
or, for instance,

{4x- 3y-5 =0, 5y-4z+19 =0},
two projection planes that define the line.
(2) Find equations of the line through (-2, 3, -5) and (2, 6, 3).
Direction numbers of the line segment joining the two points are found by
subtracting coordinates in order: [4, 3, 8]. Then, if P(x, y, z) is represen-
tative of all points on the line, direction numbers of all segments
[x-2, y-6, z-3]
are proportional to 4, 3, 8. Thus
{x-2=4t, y-6=3t, z-3=8t}
represent the line. We might just as well have used the point (-2, 3, -5) ;
thus
{x+2 = 4s, y - 3 = 3s, z+5 = 8s}.
On eliminating t, the first set yields
_4_=y- 36 =g},
Jx 2 z- 3
and the second set, on eliminating s, yields

x+2 _ y - 3 _z+5
{
4 3 8
The student can easily show that both results may be re-expressed as
{3x-4y+18=0, 8y-3z--39=0}.
(3) Find equations of the line through (1, 2, 3) parallel to the line
{x-y+2z-4 =0, 2x+3y+3z-8 =0}.
Direction numbers [a, b, c] of the given line are such that
a- b+2c =0 and 2a+3b+3c =0
and thus a: b: c = -9: 1: 5. The required line is parallel to the given line
and thus has direction numbers [-9, 1, 5]. Accordingly, the equations of
the line are
{x - 1 = -9t, y - 2 = t, z - 3 = 5t}
or
jx-91 =y-2=z 53
(4) Find equations of the line (called the common normal) perpendicu-
lar to the two lines
(I): {x-z+3 =0, y+z - 1 =0)
and
(II): {2x-y+1 =0, 3y+2z-17=0).
Our strategy here is as follows. First
we find the equation of the plane
(A) which contains the line (II)
and which is parallel to line (I).
Then we obtain the plane through
(II) perpendicular to (A). This to-
gether with the plane through (I)
perpendicular to (A) determines the
common normal. The family
Fig. 14.10
equation
2x-y+1 +k(3y+2z-17) =0
or
2x- (1 - 3k)y+2kz-l-1 - 17k =0
represents planes containing line (II). The plane of this family parallel
to (I) has its normal direction [2, 3k - 1, 2k] perpendicular to (1):
[1, -1, 1]. Thus

2+I-3k+2k=0 or k=3
and (A) is
x + 4y + 3z - 25 = 0.
The plane through (II) perpendicular to (A) has its normal direction
[2, 3k - 1, 2k] perpendicular to the normal direction of (A) : [1, 4, 3].
Thus
2+4(3k- 1) +6k =0 or k=$.
This, then, is
9x - 3y + z - 4 = 0
The family
x-z+3+k(y+z- 1) =0
or
x + ky + (k - 1)z + 3 - k = 0
with direction numbers of its normal [1, k, k - 1] contains line (I). The
plane of this family perpendicular to (A) has its normal perpendicular to
the normal direction [1, 4, 3] of (A). Thus
1+4k+3(k-1) =0 or
This, then, is
7x+2y-5z+19=0.
The pair of planes
{7x+2y-5z+19=0, 9x-3y+z-4=0}
is the required common normal. The student will find that a mastery of
the geometrical thought and the analytical processes involved in the solu-
tion of this problem will make him fully aware of three-space concepts
concerning lines and planes.
An interesting configuration may be observed from considering suc-
cessive common normals. For two given lines, there is a common normal.
For these three lines taken in pairs, there are two additional common nor-
mals. Their common normals result in other lines. Although it might seem,
if this process is continued, that there would be an unlimited set of lines
produced, the process brings forth repetitions after a total of ten lines.
EXERCISES
20. Find the equation of the plane through the given point perpendicular to
the given line.
(a) (1, -2, 3); {x= y,x+z= 4}
(b) (2,-3,3); {x=t+l,y=3t-2,z=4- 2t}
21. Write the equations of the line through the given point with the given
direction numbers. Sketch each.
(a) (0, 0, 0) ; [2, 1, 51 (c) (-2, 1, 4); [2, 3, -4]
(b) (0, 0, 0); [- , 3, -s] (d) (1, *, 11 );
22. Write x, y, z as some particular linear functions of s and t, thus defining a
plane. From these obtain the rectangular equation and the direction num-
bers of the normal. Sketch.
23. Write x, y, z as some particular linear functions of t, thus defining a line.
Give coordinates of two points on your line and sketch.
24. Find the equation of the locus of points equidistant from the pair of planes
(a) x+ 2y+2z- 1 = 0 (b) x- y+z - 2= 0
x-1/5y+2z-3=0 x=y
25. Find the equations of the locus of points equidistant from the three points
(0, 0, 4), (0, 3, 0), (5, 0, 0).
26. Find where the line
{x = 2t + 1, y = t - 3, z = 2 - 3t)
pierces the plane
{x= 21+s, y = 1- s, z= 1 - t).
27. Find the point of intersection of the lines
(x+y+z- 9= 0,2x- y+3z- 13= 0)
and
(5x+3y-5z+1=0, 5x+y+5z-33=0}.
15.
Surfaces And Curves
15.1. Definition
A surface is the set of points P whose coordinates x, y, z are values of
functions
{x = x(s, t), y = y(s, t), z = z(s, t) }
of two variables s and t.'* Thus, for example,

(A) {x = -s + t, y = s + t, z = 2t + 1}
(B) z = 1s+ t, y = -s1 } 3t, z=

11
t
are surfaces.
If it is possible to eliminate the variables s and t (the parameters) among
the equations defining x, y, z we may produce the rectangular equation of
the surface. For example (A) : x + y = 2t and thus
x+y=z-1
is the rectangular equation of the surface, recognizable as a plane.
* The variables s and t are independent parameters-not functionally related.
200
Sec. 15.1 SURFACES AND CURVES 201
For the example (B) : x + y = 4t, and thus

4
Z =
x + y
Generally, the elimination of the parameters s and t produces a single

equation
f (x, y, z) = 0
which may be used as an alternate definition of a surface.
15.2. Definition
A curve is the set of points P whose coordinates x, y, z are values of func-
tions
{x = x(t), y = y(t), z = z(t) }
of one variable t. Thus, for example,

(C) {x=t, y=I - t, z =3t+2}
(D) {x=t, y= 1-t1, z=sint}
are curves.
Elimination of the variable t in example (C) produces the pair of
equations
{y=1-x, z=3x+2}
for the rectangular equations of the curve. They are recognizable here as
two planes defining a line as the set of their common points.
For example (D), elimination of t gives the pair of equations
{y = 1 - x$, s = sin x}
representing two surfaces whose intersection is the curve, called a Helix.
In general, such elimination produces rectangular equations of two
surfaces
{ f(x, y, z) = 0, g(x, y, z) = 0)
which define the curve as the set of points common to the surfaces.
202 SURFACES AND CURVES Sec. 15.2
EXERCISES
1. Obtain the rectangular equation of each of the following surfaces.
(a) {x= t2,y= s2,z= s+t}
(b) {x= s + t, y= s-- t,z= st}
(c) {2x=s2+ 1,3y= t2- 1,4z= s+t}
(d) {x+ y = t, x - y= s, z= st- 11
(e) fx=sint+coss,y=sint-coss,z=cos2s}
2. Obtain rectangular equations for each of the following curves.
(a) {x= t2,= t+1,z=-t+l}
(b) (x- y= t2,2x+3y= t- 2,z= t+ 1}
(c) {x= 2cost,y= 3sint,z= t}
(d) {x = t, y = sec t, z = tan t}
15.3. Graphical Representation of a Surface

If a surface be given, its sections with planes parallel to the coordinate
planes are a means of representing the surface pictorially. For example,
sections of the surface
x2+y2=z
by planes z = k > 0 (planes parallel to the xy-plane) are the circles {x2 +
y2 = z, z = k 1. As k increases, the radius of the circular section increases.
The trace formed by the surface with x = 0 (the yz-plane) is the
parabola y2 = z; whereas that with y = 0 (the xz-plane) is the parabola
x2 = z. The surface is bowl-shaped and called a paraboloid of revolution.
For another example, consider
x2+4y2=(z-1)2
Sections of the surface made by planes z = k are ellipses
{x2 + 4y2 = (k - 1)2, z = k}
with major and minor diameters of lengths 2 1 k - 1 I and I k - 11. For
k = 1, the section is the point (0, 0, 1). Sections made by y = 0 are the
lines
{X2 = (z - 1)2, y = 0};
i.e.,
{x = z - 1, y = 0} and {x = 1 - z, y = 0};
zzk
Fig. 15.1 Fig. 15.2
and by x = 0, the lines

(2y=z-1,x=0J and 12y=1-z,x=O).
Planes through the z-axis, y = hx, cut the surface in the lines
{x2 + 4h2x2 = (z - 1)2, y = hx
i.e., in
{x 1+412z-1,y=hxJ
and
{x1/1+4h2=1- z,y=hxJ.
The surface is an elliptical cone.
As a third illustration, consider the surface
(y-2)2=3-z
I
This surface is the set of points

whose coordinates y, z satisfy the 1z
given equation with x-coordinates
completely arbitrary. Thus, for
example, points (x, 3, 2) for all
x's lie on the surface. These par-
ticular points lie on the line parallel
to the x-axis which pierces the yz-
plane at y = 3, z = 2. This is
further substantiated by the sec-
tions of the surface made by planes
y = k. These sections are the lines Fig. 15.3
{z=3- (k-2)2,y=kJ,
lying parallel to the x-axis. The surface is a parabolic cylinder, and each of
the parallel lines lies wholly on the surface and is called an element of the.
cylinder. Note further that the curves
{(y-2)2=3-z,x=k}
are all congruent parabolas.
EXERCISES
3. For each of the following surfaces, find sections made by given plane(s)
and sketch each surface.
(a) {x2 + 4y2 = z; z = k}
(b) {9x2 + y2 = z; z = k}
(c) { x2 + y2 = z2; z = k}
(d) { x2 + y2 + z2 = l ; z = k}
(e) { (y - 3x)2 = z2; z = k}
(f) {(2y - x)2 = z2; z = k}
(g) {x2 + y2 - 2x = 0; x = k}
(h){y2+z2-4=0;y=k}
(i) {x2+y2+z= 4;z= k}
(j) {x2+y2-z=4;z=k}
(k) {x2+y2- z2= 1;z= k,x= k}
(I) {9x2 + 4y2 - z2 = 36; x = k, y = k, z = k}
(m){x2- y2- z2= 4;x= k,y= k,z= k}
(n) {x2-4y2-3z2= 12;x=k,y=k,z=k}
15.4. Families of Surfaces and Projection Cylinders
Consider the pair of surfaces
{f(x, y, z) = 0, g(x, y, z) = 01
defining a curve as the set of those points whose coordinates satisfy both
equations. Then each surface of the family,
f(x, y, z) + y, z) = 0
contains those points. We have here an unlimited number of surfaces pass-

ing through the given curve.
Fig. 15.4 Fig. 15.5
Among this family are cylindrical surfaces of the type described in the
third illustration of the preceding paragraph. Each such surface, called a
projection cylinder, has a set of line elements perpendicular to one of the
coordinate planes. For instance,
x2+y2-2y=0
is the set of points whose xy-coordinates satisfy the given equation, and
whose z-coordinate is completely arbitrary. Such points are all located on
lines perpendicular to the xy-plane drawn through points on the circle
{x2-+2-2y=0, z=0).
These lines are elements of the right circular cylinder shown.
We utilize these ideas in the (+ + +) octant representation of the
curve
{x2+2y2+z2-2y-4=0, 2x2-+2-z2-4y+4=0{.
Each member of the family
x2+2y2+z2-2y-4+k(2x2+y2-z2-4y+4) =0
contains the curve. By taking k = 1, then k = - Z, we select two projec-
tion cylinders whose intersection is the given curve; thus
{x2+y2-2y=0, y2+z2-4=0).
These two surfaces are right circular cylinders with elements parallel to the
z- and x-axes, respectively. Points on their curve of intersection are found
as follows.
Noting that the variable y is the only one that appears in the equations
of both surfaces, we set y = k and find points of each cylinder lying in this
plane, shown as A BCP. These plane sections of the cylinders are the lines
{x= 2k-k2, y=k}
and
{z= 4-k2, y=k},
each of which (AP and PC) is an element. These line elements meet at
P, a point on both surfaces, and thus on their curve of intersection. The
process is repeated by taking other values of k, 0 S k 5 2, to locate as
many points P as desired.
EXERCISES
4. For each of the following curves, write the rectangular equation of a family
of surfaces containing the curve. Then from this family select projection
cylinders, and use these to sketch the curve as illustrated in Par. 13.4.
(a) {:0 + z2= 16,x2+y2+z2= 25}
(b) {x2 + z2 = 4, x2 + y2 = z}
(c) { 9x2 + 4y2 = z2, y2 + z2 = 41
(d) (.0+ ?12 + z2 = 40, y2 + z2 = 41

(e) 4.r, x2 + y2 = 4x}
{y2 + z2 =
(f) {x2+y2= (z- 1)2,x2+y2= }
(g) {y2 + 3z2 = 4(2x + 3z), 2y2 + z2 = 4(z - x)}

(1i) 1.r t'-- 1, y= t-f- 1, z =
(i) {X = cos t, y = sin t, z = 2t}
(j) {x= I - cost, y= 1- sin t,z+x= 1}
(k) {r = 5, z = 30) ((ylindrical coordinates)
(1) ( x2 + y2 + z2 = 16, 4.r' -j- (z - 2)2 = 4)
(m) {x2 + y2 + z2 = 25, 4x2 + z2 = y2}
1)2 I
(o) {3x2+y2- 8x- z+4= 0,5x2+y2- 8x+s= 41

(p) {5y2- 2z2+3x+y+2z= 0,y2+222+x- 2y- 2z 01
(q) {x = 4(1 - sin t), y = 2 cos t, z = 3 sin t}

16.
Tangents
And Normals
In this chapter we shall consider tangents and normals to three-space
curves and surfaces. It will be found that curves generally possess at a point
a tangent line and a normal plane; whereas surfaces have tangent planes
and normal lines. Such concepts have far-reaching importance in the fur-
ther study of geometry, in problems of maxima and minima, and in physi-
cal applications concerning problems of flow. We deal first with curves.
16.1. Tangent Line and Normal Plane to a Curve

Let the curve
{x = x(t), y = y(t), z = z(t) }
be defined by functions each single-
valued and differentiable for the
domain of real values t.* Let
P(x, y, z) for some unspecified t be
representative of all points of the
curve. Further, let Fig. 16.1
Q (X + Ax, Y + Dy, z + Az)
be another point of the curve for the value t + At.
* Functions having such desirable properties are called "well-behaved."
207
208 TANGENTS AND NORMALS Sec. 16.1
The line PQ then has direction numbers

Ax Ay As
I At At' At
We now let At -> 0. Then Ax - 0, Ay -' 0, Az --> 0, Ax/At -* Dtx,

Ay/At ---> D,y, Az/At -* D,z, Q -* P and the line PQ approaches a position-
direction that we feel should be tangent to the curve at P. In this spirit, we
define the limits of the direction numbers Ax/At, Ay/At, Az/At as direction
numbers of the tangent line to the curve at P.
(A) For example, the curve
{x=t2, y=13, z=t-1}

has
D,x = 2t, Dty = 3t2, Dz = 1
for direction numbers of its tangent line at any point. If we choose a particu-
lar point by setting t = 3, for instance, the point has coordinates (9, 27, 2),
and the tangent line has direction numbers [6, 27, 1]. Its equations are then
x- 9 y 27 _z-2
{ 6 27 1
The normal plane is that plane through P perpendicular to the tangent

line. Its equation is given at once by the scalar product relation:
6(x - 9) +27(y-27) + (z - 2) = 0.
(B) Consider a second example. The curve
{x-y2-z2=0, x + 3z - 1 = 01
is not given in the primary form

with x, y, z expressed as values of
functions of t. However, we may
regard each of these rectangular
coordinate variables as values of
functions of t and can obtain direc-
tion numbers of the tangent line by
implicit differentiation. Thus
D,x-2yD,y-2zD,z=0
Fig. 16.2 D,x + 3Dz = 0.
Sec. 16.1 TANGENTS AND NORMALS 209
From these equations we solve for Dtx and D:y in terms of Dez:
Dx = -3D,z, Ay = - 2z2y+ 3,Diz, Dtz = Do.
These are a set of direction numbers of the line tangent to the curve at
(x, y, z). Since any set proportional to these defines the same direction, we
have more simply
[-6y, - (2z + 3), 2y].
For a particular point on the curve, for instance, (z, 17/6, 11), direction
numbers are [3N/-17, 10, - 17] and the tangent line is
x-l y- 17/6 z-g
3N/17 +10 - 17 '
the normal plane

3 17(x - J) + 10(y - /6) - 17(z - e) = 0.
EXERCISES
1. Find equations of tangent line and normal plane to each of the following
curves at the designated point. Sketch each curve by means of projection
cylinders.
(a) {x= acos0,y= asin0,z= k9},9=ir/4
(b) {x= acos0,y= b sin 0, z = kd},9= it/6
(c) {x= t,y= t2+ 1,z= 1 - t},t= I
(d) {x= 1- t,y= 1+t,z= I-t2},t= #
(e) {x= ta,y= t- 1,z= 2t2+ 1},t=
(f) {x= t- 1,y= t+ z= t'},t= 1
(g) {x= tcost,y= tsint,z= at},t=ir/3
(h) {x2+y2= z2,y+z= 8},(4,3,5)
(i) {x2 + y2 + z2 = 3x, 3x - 2y + 4z = 51, (1, 1, 1)
(j) {x2+y2+2z2= 4x,x+3y-4z=0},(1,1,1)
(k) {x2+3y2= 4,y+2z= 71,(1,1,3)
(1) {y2 - z2 = 9, x = 2}, (2, 5, 4)
(m) {x2 + 4y2 = 8z, x + 2y + 3z = 61, (2, 1, 1)
2. Find points on each of the following curves where the tangent line is
parallel to the xy-plane.
(a) {3x2 + 2y2 + z2 = 4, z22 = 9(x22 + y2)}
(b) {x = t2, y = 1 - t, z = to - 12t}

3. Show that the line tangent to the Helix
{x = sin 0, y = cos 0, z = aOl
makes a constant angle with the plane z = 0.
16.2. Tangent Plane and Normal Line to a Surface

Let the surface be represented by the functions
{x = x(s, t), y = y(s, t), z = z(s, t) }
which we shall suppose are single-
valued and differentiable for neces-
sary domains of s and t.
Points P(x, y, z) on the surface
are determined by values of s and t.
If we fix upon a constant to for t
and let s vary the point P travels on
the surface along that curve for which
t equals to. The direction numbers of
the tangent line to this curve at P
Fig. 16.3 are, according to Par. 16.1, the
s-derivatives of x, y, z:
[D,x, D,y, D.z].
If we fix s, setting s = so, and allow t to vary, P travels along another
curve on the surface. Direction numbers of the line tangent to this curve at
P are then the t-derivatives
[Dex, Day, Dez].
These two lines meeting at P determine the plane tangent to the sur-
face at P. (It is shown in the Calculus that all curves on the surface through
P have tangent lines which lie in this plane.)
The line through P perpendicular to the tangent plane is called normal
to the surface. A set of direction numbers [X, µ, v] may be found from the
direction numbers of the tangent lines by using the scalar product relation.
We illustrate.
(A) Find equations of Tangent Plane and Normal Line to the surface
{x = 3s + 41, y = 4s - 3t, z = I + s2 + t2j
at the point for which s = 2, t = 1;

i.e., at (10, 5, 6).
For t constant,
D,x=3, Day=4, D,z=2s.
For s constant,
D,x = 4, D, y = - 3, D jz = 2t.
These are direction numbers of
a pair of lines tangent to the surface
at any point designated by selected
s and f values. For the point in
question, s = 2, t = 1, the two
Fig. 16.4
directions are
[3, 4, 4] and [4, -3, 2].
The normal line with direction numbers [a, µ, v] is perpendicular to
both tangent lines. Accordingly,
3A+4µ+4v = 0
4X-3µ+2v=0.
From these we obtain the ratios
_ µ v
20 I0 - 25'
and choose
X=4, µ=2, v=-5.
The equation of the tangent plane at (10, 5, 6) is thus
4(x - 10) + 2(y - 5) - 5(z - 6) = 0, (TP)
and those of the normal line
Jx- 10 _ y 5 z 6
(NL)
4 2 -5 }
The surface is a paraboloid of revolution.

We discuss the same problem now with the surface given by the equiva-
lent form
` x2+y2-252+25=0
and select the same point (10, 5, 6).
We regard x, y, and z as function values of s and t and employ implicit
differentiation. Then, fixing t as a constant, the implicit` s-derivative is
2x I D,x + I
2y I D,y I
- 25 , D,z = 0,
while the t-derivative is
2x D:x + 2y Djy - 25 Dlz = 0.
Now it should he understood that [D,x, D,y, D,z] and [D:x, Day, Dz] are
direction numbers of lines through P, tangent to the surface and that the
normal line is their common perpendicular. Upon inspection of the two
derivative equations, it is evident that the coefficients
2x 2y - 25
of the derivatives are in fact these direction numbers of the normal.

For the given point (10, 5, 6), these direction numbers are [20, 10,
-25] or the equivalent set [4, 2, -5]. Equations of tangent plane and
normal line are at once determined.
Note should be taken that either derivative equation yields direction
numbers of the normal line.
EXERCISES
4. Find equations of tangent plane and normal line to each of the following
surfaces at the designated point. Sketch each surface.
(a) x2 + J2 + z2 = 14, (1, 2, 3)
(b) x2 + y2 + z2 = a2, (xo, yo, Zo)
(c) x2 + y2 + z2 = 2y + 10, (1, 2, 3)
(d) 4x2 + 9y2 + 25z2 = 206, (5, 3, 1)
(e) axe + bye + ae = 1, (xo, yo, Zo)
(f) (x + y)z = 2y, (2, -4,4)

(g) x2 + y2 = 1 + 2z, (3, 4, 12)
(h) yz+xz+xy= 8, (1,2,2)
(i) {x= 3s+ 1, y= 41+ 2,z= s2+t2},s= 1,t= -2
(3) 1x= I/$, y= 1 / 1 , z = s+ t}, s= 2, t 4
(k) {x= cos113t,y= sin" t,z= s},s= 3,=ir/4

(1) {x= s- t,y= s+t,z= s2- t21,s= 2,t= 1
(m){x= t(s+t),y=s(s+t),z=s+t},s= I,t= 2
5. Find points on each of the following surfaces where the tangent plane is
parallel to the xy-plane.
(a) z = xy + (27/y) + (27/x)
(b) z = sin x + sin y
(c) z = xy(3x + 2y + 1)
(d) z = x2 - 3xy + 3y2 - 7x + 12y + 10
6. Find the maximum value of x° - y4 + 4xy - 2y2.
17.
Special Surfaces
Surfaces having special characteristics are always of interest. Those

formed by revolving a curve about an axis and those upon which lie whole
families of lines are particularly interesting and fundamentally important
for their applications. The latter ones, called ruled surfaces, are, for in-
stance, the basic forms of space gears that transfer motion from one axis
to another.
17.1. Surfaces of Revolution

If a surface is such that sections made by all planes perpendicular to a
fixed line are circles with centers on that line, the surface is one of revolution.
The line of centers of the circles is the axis of revolution. We restrict our
discussion to those surfaces having circular sections formed by planes
x = k or y = k or z = k. Accordingly,
(x - a)2 + (y - b)2 = f(z)

(y - a)2 + (z - b)2 = g(x)
(z a)2+(x-b)2=h(y)
are surfaces of revolution, the first having circular sections with planes
z = k, the second with x = k, and the third with y = k. We assume, of
course, that f (k) > 0, g(k) > 0, h(k) > 0 for some domain of k.
214
Sec. 17.1 SPECIAL SURFACES 215
Familiar surfaces of revolution encountered already herein are:
x2 + y2 = 4 (cylinder) x2 + y2 = z2 (cone)
y2
x2 + y2 = z (paraboloid)
x2
- - - = 1 (hyperboloid)
We turn to the generation of surfaces of revolution and the determina-

tion of their respresentative equations. Note the selected arrangement of
reference axes and let the plane curve y = f (x) be revolved about the x-
axis. Let the point Po(xo, yo, 0) on the given plane curve trace out the
circular section upon which P(x, y, z) is a representative point. Then the
radius of the circular section through PO and P is yo. Accordingly,
y2 + z2 = y0 - U (X-) 12,
and, since x = xo,
y2 + z2 = C f(x) ]2
the equation of the surface.

For example, the surface generated by revolving
{y2=x, z=0)
about the x-axis is
y2+z2=x
a paraboloid of revolution.
-4kxl
x=xc
Fig. 17.1 Fig. 17.2

216 SPECIAL SURFACES Sec. 17.1
For another example, suppose the line {y = 3, 5x - 3z = 0} be

revolved about the z-axis. Let P(x, y, z) be representative of all points of
the path of Po(xo, 3, zo), a point of the given line. Then
5xo - 3zo = 0 with zo = z.
The radius of the circular section at P is
OQ = OQo = 1x2 + y2 = 9 + xo.
Then
x2+y2 = 9+xo = 9+z = 9+z2 T-K
or
x2 y2 z2
9 +9 25-1
a hyperboloid of revolution. (Note that sections made by planes x = k and

y = k are sets of hyperbolas.)
Problem.
(1) Show that the preceding hyperboloid is also generated by revolving
the line
{y=3, 5x+3z=01
about the z-axis.
VD
Fig. 17.3
(2) Show that the preceding hyperboloid is also generated by revolving the
hyperbola
{ 25y2 - 9z2 = 225, z - 01
about the z-axis.
(3) Make two circular discs with equally spaced holes in their circumfer-
ences. Lace elastic thread through the holes. By holding the discs
apart in parallel planes and twisting them, have the model show the
formation of these surfaces which include the cone and cylinder.
EXERCISES
1. Show that each of the following are surfaces of revolution; find a set of
planes which cut the surface in circles, and name the axis of revolution.
Sketch each surface.
(a) x2 + 9y2 + 9z2 - 6x = 0
(b) 4x2 - 5y2 - 5z2 + 4x = 0
(c) 4x2 + 4y2 = 5 + 922
(d) x2 + y2 = 4z2
(e) x2-y2+ z2= 1-2y
(f) z'2+z2= 4y
X2 y2 22
(g)
a'
+ a'+b2= 1, a > b (Oblate spheroid)
x2 y2 22
(h) as+ + b2
= 1, a > b (Prolate spheroid)
x2 y2 22
(i) a2+a2- -= I
x2 y2 z2
G)
a'
+a2- = -1
x2 y2 22
(k) a2+ a2- b_ 0. [Show how this surface is related to (i), (j).]
(1) x2 + z2 = ay2 + by + C
(m) ( x2 -+y2 - b)2 + z2 = a2 (Torus)
2. Derive the equation of the surface generated by revolving the given curve
about the given axis. Sketch,each surface.
(a) {x + 2y = 4, z = 0} about x-axis; y-axis
(b) 1 y2 = 4z, x = 0} about z-axis; y-axis
(c) {x2 - 9y2 = 1, z = 0} about x-axis; y-axis

(d) { x2 + 9y2 = 1, z = 01 about x-axis; y-axis
(e) {x - 4z, y - 01 about z-axis; x-axi9
(f) {z2 = 4y, x = 01 about y-axis; x-axis
(g) {2x + z = 2, z + 2y = 2} about z-axis, y-axis
(h) {2z + z - 2, z + 2y = 21 about x-axis
(i) 12x + z = 2, z + 2y = 2} about y-axis
(j) {y(1 + x2) = 1, z = 01 about y-axis
(k) {y2 + (z - 3)2 = 1, x = 0) about y-axis (Torus)
(1) { x2 + (y - 2a)2 = a2, z - 01 about x-axis (Torus)
17.2. Ruled Surfaces

If each member of a family of lines lies wholly on a surface, the surface is
ruled; the lines are its rulings. The hyperboloid of revolution, considered
in Par. 17.1, is an example of such a surface-an unusual one, in fact,
since it contains two sets of rulings. Let us consider the matter.
The equation of the surface
x2 y2 Z2
25=1
9 +9
may be written in the factored form
\3 5/\3 + 5/ = \1 3/\1 + 3/'

which prompts consideration of the family of lines
= k\1 + 3/' 1-3

{3 -5 k\3 + 5J =
It is evident that all points, whose coordinates x, y, z satisfy the equations

of the family of lines, lie also on the hyperboloid. This set of lines, for all
values k, thus lies wholly on the surface.
The family
obtained through choice of other factors in the equation of the surface, is

different from the preceding set of lines and presents other rulings as was
pointed out in Par. 17.1.
It should be noted that the tangent plane to the surface at any point is
the plane formed by the two rulings through that point. Conversely, a
plane tangent to the surface cuts out two rulings on the surface.
Consider a second example, the cone
(x - 1)' + (y - 2)' _ (z - 3)'
Rewriting the equation in factored form,

(x- I)(x - 1)= [(z-3) - (y-2)][(z-3) + (y-2)]
=(z-y-1)(z+y-5),
we set up equations
{x- 1 =k(z - y - 1), k(x- 1) =z+y-51
of a family of lines. All points of the family of lines are also points of the
cone, and the lines are thus rulings.
We note that every line of the family contains the point (1, 2, 3) and
thus identify the point as the vertex of the cone.
General Remarks. Our treatment of ruled surfaces has been rather special
in that equations of rulings were determined only for surfaces whose
equations were readily factorable. A general investigation must proceed as
follows.
For a surface
f(x,y,z) -0
to have rulings, it must be possible to determine a set [X, µ, r] of direction

numbers of a line through any point (xo, yo, zo) of the surface so that all
points (x, y, z) of the line lie also on the surface. Thus, since any point on
the line through (xo, yo, zo) with direction [x, p, r] is
l xo + at, yo + f+t, zo + V11,
then between the equations

f (xe, yo, zo) = 0 and f (xo + At, yo + pt, zo + pi) - 0,
the second of which is identically zero

for all values t, we must be able to
solve for the ratios of X, µ, P. We
illustrate.
It is at once evident that the
surface
z=xy+1
has the two sets of rulings:
Fig. 17.4 {y=k, z=kx+1j
and
{x=k, z=ky+1).
However, we treat the problem by the general method outlined above.
If Po(xo, yo, zo) is a representative point of the surface, then
zo = xoyo + 1.
For points on a line through Po(xo, yo, zo) with direction [X, µ, v] to lie on
the surface:
ZO + Pt = (xo + At) (Yo + Fit) + 1,
or
Pt = (xo o + yoX)t + Aµt2,
an identity in t. Thus
v = xoµ + yoa, Aµ = 0
We have here a choice:

(1) A=0, µO0 and v=xoo.
Direction numbers of rulings for this choice are accordingly:
[0, o, xo µ] or more simply [0, 1, xo].
Equations of the set of these rulings are then
z-zpl
y - Yo xo
or
fx=xo, xoy - z=xoyo - zo= -1}.
Here xo plays the role of the parameter k in the factor treatment.
(2) The alternative A 0, 0, v = yol

yields direction numbers
[x, 0, yoX] or [1, 0,110
and a second set of rulings
x-xo = z-z0 , y =yo

I yo
or
(xyo - z =xoyo - zo= - 1, y =ya}
where yo plays the role of k in the factor treatment. The surface is
hyperboloid.
EXERCISES
3. Each of the following is a ruled surface. Use the method of factoring to
find equations of the rulings and sketch the surface. State which surfaces
are doubly ruled.
(a) xy = 1
(b) xz- y= 0
(C) x2 + y2 = 25
(d) 9x2 + z2 = 36
(e) (x - 1)2 + y2 = z2
(f) (x + 2)2 - (y + 3)2 + (z - 1)2 = 0
x2 y2 Z2
(g) 9 + 16 25 =1
2
(h) 25 6 -+- 4 = 1
(i) 9x2 - 4y2 = 6z

(j) z2 - 16y2 = 9x
(k) z3 - 3zx + 8y = 0
(1) (x - 1) (y - 2) _ (z - 3)
4. Use the general method discussed in Par. 17.2 to find rulings for
(a) z = x(y - 3) (b) x2 = (y - 1)2(z - 2)2
5. Find the equation of the surface whose rulings are the family of lines
with parameter k.
(a) 12x - 3z = k(1 - y), k(2x + 3z) = 1 - y)
(b) 12x + 3y = k, k(2x - 3y) = z{
(c) { y - z = kx, k(y + z) = x}
((I) {2x - y - z = kz, 2x + y - 1 + k(2x + y) - 01
(e) {x=ky,kx-z)
17.3. The Cone and Cylinder
Because of their frequent oc- z
currence we pay special attention to
cones and cylinders.
Definition. A cone is a ruled sur-
face whose rulings (also called
elements) are lines through a fixed
point , ca lle d th e vert ex, and a ll po in t s 11 Y
of a curve called the directrix. (If
the direetrix is a plane curve, the
vertex is not in the plane of the Fig. 17.5
curve.)
To transcribe the geometric definition into analytical form, let the ver-
tex V and directrix D be
V: (a, b, c), D:{f(x,y,z) = 0, g(x,y,z) =01.
if P0(xo, yo, zo) be representative of points on the directrix, the points
P(x, y, z) on the ruling through Po and V are such that
xo -ayou -bzo c k
x- a y- b z- c
or
xo=a+k(x-a), yo=b+k(y-b), zo=c+k(z-c)

where the ratio is set equal to a parameter k for temporary convenience.
Now, since f(xo, yo, zo) = 0, g(xo, yo, zo) = 0, then
{f[a+k(x-a), b+k(y-b), z+k(z-c)]=0,

g[a+k(x-a), b+k(y-b), z+k(z-c)] =O
Elimination of k produces the equation of the cone.
To illustrate, consider the cone with
directrix: {x2 + y2 = 2z, z = 4) and vertex: (1, 2, 6)
We have
xo-1 yo2-zo-6=k
x- 1 y- 2 z- 6
or
{xo=1+k(x-1), yo=2+k(y-2), zo=6+k(z-6)).
And, since
{xo + y' = 2z(., zo = 4)
then
{[1+k(x- 1)]2+[2+k(y-2)]2=2[6+k(z-6)],
6 + k(z - 6) =4).
Eliminating k, the equation of the cone is
(x - 1)2-1- (y-2)2-{-2(x- 1)(z-6) +4(y-2)(z-6)

=3(z - 6)2
Note that the replacement of the

quantities
xo - a by k(x - a)
yo - b by k(y - b)
zo - c by k(z-c)
has the analytical effect of estab-
lishing homogeneity in the quan-
tities (x - a), (y - b), (z - c).
Accordingly, any equation homoge-
neous in time quantities is a cone. Fig. 17.6
The fact is useful in recognition. For
example,
2x2y - xz2 + 2y' - 2yz2 = 0
is homogeneous in x, y, z and thus represents a cone with vertex at (0, 0, 0).

For another example,
x(y + 2) + (z - 1) (y + 2) = x(z - 1)
is homogeneous in x, (y + 2), and (z - 1) and thus is a cone with vertex

at (0, -2, 1).
Definition. A cylinder is a ruled surface whose rulings (also called
elements) are lines all parallel to a fixed direction and which pass through
points of a curve called the directrix. Thus for a given direction [A, µ, v]
and given directrix {f(x, y, z) = 0, g(x, y, z) = 0) then points P(x, y,
z), on the rulings that pass through Po(xo, yo, zo) on the directrix, are
such that
x - xo y- yo z- zo =k
A µ v
and
f(xo, yo, zo) = 0, g(xo, yo, zo) = 0.
These equations usually are sufficient equipment to permit the elimination
of xo, yo, zo and thus obtain the equation of the cylinder in terms of x, y, z,
A, p, P. To establish techniques, consider for directrix the curve
{x2+4y2=36, y=z)
and for the fixed direction of rulings the numbers
[A=1, o=2, v=1].
Rulings through (xo, yo, zo) are then
jx - xo=y 2 yo=z-zo=k},
and, since Po(xo, yo, zo) is a representative point of the given directrix
{xo + 4y0 ' = 36, yo = zo),
we have
(x-k)2+4(y-2k)2=36, y-2k=z-k.
Or, eliminating k,
(x-y+z)2±4(2z-y)2 =36 ,
for the equation of the cylinder.

We now turn the problem Z

around and seek equations of the
rulings. Writing the equation of the
surface in factored form
(x-y+z)(x-y+z)
= [6 - 2(2z - y) ][6 + 2(2z - y)]
= (6+2y - 4z) (6 - 2y+4z),
we establish the family Fig. 17.7
{x-y+z=k(6+2y-4z), k(x-y+z) =6-2y+4z}

of lines. These are
{x- (1+2k)y+ (1+4k)z =6k, kx+(2-k)y+(k-4)z=6},
and all lines of this family have direction numbers proportional to [l, 2, 1];
(the student should verify this fact).
EXERCISES
6. Identify each of the following as cones; locate the vertex, and give equa-
tions of rulings.
(a) x2+y2= (z- 1)2
(b) z2 - z' = y2
(c) (x - 1)' + (y + 2)2 = z2
(d) 4(x- 1)'+9(y+2)'= z2
(e) x2+y'-z2-2z+4y+5=0
(f) x2 + 4y2 - z2 - 8y + 4 = 0
(g) xa + y8 + ze - 3x2y + 3xy2 = -1
(h) x2y - xyz = za
7. For each of the following, derive the equation of the cone with the given
vertex (V) and directrix (D).
(a) V:(0,0,1) D:{y2= x,z= 0}
(b) V: (0, 0, 1) D: { (x - 1)2 + y2 = 1, z = 0}
(c) V: (0, 0, 1) D: {b2x2 + aY = a2b2, z = 0}
(d) V: (0, 0, 1) D: {b2x2 - ay = a2b2, z = 0}
(e) V: (0, 0, 0) D: {x2 + y2 - 2x - 4y = 0, z2 = 4y}
(f) V: (0, 0, 0) D: {x2+y2= 2z, z = 11
8. Obtain equations of rulings of each of the following surfaces. Show these

rulings all parallel, and thus identify the surface as a cylinder.
(a) x2 + y2 = x (b) y2 - z2 = y
(c) y(I + x2) = I (d) y = sin z
(e) x+y2+z=4 (f) x2+ y2- 2xy+x+z= 6

(g) (x + z)2 + y2 = 1 (h) 2(x - y)2 + z2 = 4(x - y)
9. Write equations of cylinders having the following curves as directrices
and having rulings with the given direction.
(a) {y2= x+ 1,z = 0}
(b) {x2+y2= 2x, z = 01
(c) {x2 + y2 = 25, z = 0}
(d) {x2 - z2 = 1, y = 0)
(e) { x2 + y2 = 4, y + z = 21
(f) {x2 + y2 = 1, z = 0)
* 17.4. Gears
The problem of transferring circular motion from one axis (shaft) to
another in three space is of permanent importance in the design of machin-
ery. Cylinders, cones, and ruled hyperboloids of revolution are the most usual
devices-each of which permits thrust over a comparatively long tooth.
The rulings of a pair of these surfaces, cut into the surface as teeth, come
into complete mutual contact with each other during motion of the surfaces
about fixed axes.
Fundamentally, all such gears must be surfaces of revolution since the
axes of rotation must be in fixed positions. For contact, therefore, tooth
lines must be rulings-or curves upon the surface which are congruent and
in position to mesh.
The general problem in the design of gears falls into three categories,
depending upon the transfer desired. Motion about a fixed axis is to be
changed into motion about another axis which is (1) parallel to the first,
(2) non-parallel to but intersecting the first, and (3) non-parallel to and
not intersecting the first. These are described below.
In designing teeth, properly shaped to minimize friction (and noise,
an audible loss of energy) and to insure continuous thrust without un-
desirable velocity changes, a study is made of the motion of a circle rolling
upon another or upon a line (the last for the design of rack-and-pinion
gears). It is found that tooth surfaces must be cut in the shape of the
cycloid or an involute of the circle (see Chapter 10). The manufacture of
gears is an industry which demands the rare combination of mathematician
and expert mechanical draftsman.
Fig. 17.8 Fig. 17.9
Two circular cylinders (not necessarily congruent) form contact along

rulings while rolling about fixed parallel axes. The rotating speed of one is
transferred to the other which rotates in the opposite direction at reduced
(or increased) speed.
For some purposes, it is desirable to cut teeth upon the surfaces of
cylinders along Helices, and gears based upon this design are known as
spiral gears. Such a design permits thrust along parts of several tooth
edges at the same moment, thus dis-
tributing the work load over several
teeth. '
^:
Right circular cones form bevel Ì
gears with shaft lines intersecting
at a common vertex of both cones.
Contact is made along rulings and
m o tion is t ra ns ferre d f rom one ax is
to another at a desired angle. Fig. 17.10
Hyperboloids of revolution (not
necessarily congruent) form space gears as shown. They commonly are called
hypoid gears. The shafts about which they rotate are non-parallel and non-
intersecting. The driving gears in the differential box of present-day
automobiles are of the hypoid variety. (A certain unwanted wobble at the
edges necessitates high-pressure lubrication.)
Fig. 17.11
18.
*Volumes
We return to the process of differentiation and its inverse in order to

determine volumes enclosed by some surfaces. Our discussion is restricted
to surfaces whose sections, made by planes perpendicular to an axis, are
of determinable area. These include surfaces of revolution.
18.1. General Discussion

Consider a surface whose sec-
tions by planes perpendicular to
the x-axis have areas A (x), as values I
Q
of known functions of x. Let V (x)
represent the volume bounded by Ip 7 4-A(x+px)
the surface, a fixed plane x = a,
and a variable plane parallel to the
fixed one at P (x, y, z), a representa-
tive point on the surface. Let an-
other parallel section be taken at
x + Ax. Its area is then A (x + Ax).
We shall let AV represent the
volume bounded by the surface and d4-- % -*A- Ax
the two parallel planes, one at x,
the other at x + Ox. Then A V is Fig. 18.1
intermediate between the volumes
of the two cylinders shown, one ìnside,' the other òutside' of the surface.
228
Sec. 18.1 VOLUMES 229
That is,
A(x) Ax < AV 5 A(x + Ax) Ax.
Dividing by Ax,
QV
A(x) < A(x + Ax)
then as Ax - 0, Q -> P, iV/ix --> D=V, and the limit of AV/Ax must be
the limit of A (x + Ox) ; i.e.,
D=V = A (x)
The surface may be such that the sense of the inequality is reversed.
The limit, however, would be the same. The student should construct a
picture of this and present an argument analogous to the one given here-
with closed book.
18.2. Illustrations
The following illustrations clarify the matter in various circumstances.
(A) Find the volume bounded by the ellipsoid
x2 y2 z2
a2 -1- b2 -{- C2 = 1.
The surface
Y2 z2 x2
b2 -FL = 1 - a2
C2
has elliptical sections cut by planes

x = constant. For a particular x,
the semi-major and minor diameters
of the section are (setting z = 0, Fig. 18.2
then y = 0)
b 1 -x2s
a
and c- 1- x2
a2
The area A(x) of such an elliptical section is then (see Par. 9.3).
A (x) = irbc(1 -
x
az!
230 VOLUMES Sec. 18.2
and thus
/ x
D=i" = irbc{ l -
a2
From this
V = -,rbc1 x - 3a2 + C.
For x = 0, V = 0 and thus C.' = 0. Then for x = a,
Since this is the volume enclosed by only half the surface, the entire volume
is
l' = 4rabc
Note several points of interest. If the ellipsoid is one of revolution, e.g.,

x2 y2 z2
1,
2
its volume is "rab2.

For the sphere, as another special case,
x2+y 2 + z2 = a2;
the volume is
4
-rat'
.3
(B) A surface has its base

d
formed by the parabola {y2 = x,
z = 01 and has semi-circles as plane
sections perpendicular to the x-axis.
Find the volume bounded by this
surface and x = k.
Let a plane section perpendicular
to the x-axis be taken at P(x, y) on
y2 = x; i.e., at P(x, 1/x). The
Fig. 18.3 radius of the circular section is then
-0 and
A (x) = xx.
Accordingly,
Dz V = Tx,
and
But for x = 0, V =0andthus C =0.Forx =k,
irk2
Vk = (units)'
2
(C) Find the volume common to the two cylinders
x2 + z2 = a2 and y2 + z2 = a2
Since the surface bounding the

volume is symmetric to all co-
ordinate planes, we determine one
octant part of the volume, then
multiply the result by 8.
At a representative point
P(x, y, z) on both cylinders,
x= a2 - z2, y= a2 - z2.
A plane section at P perpendicular
to the z-axis thus is a square with
area Fig. 18.4
A(z) =a'-z2.
Accordingly,
D,V = a2 - z=.
Thus
a
V =a2z- 3+C.
Now for z = 0, V = 0 and thus C = 0. For z = a, the volume is V. =
232 VOLUMES Sec. 18.2
2as/3. The entire volume is thus
V= 3a3 (units)3
Note that
x2 + z2 - a2 - (y2 + z2 - a2) = 0
or
x-y=0, x+y=0
are planes through the curves of
intersection of the two cylinders.
The curves are thus plane sections
of right circular cylinders and these
are known (by Chapter 9) to be
ellipses. (A useful model may be
turned out on a lathe.)
(D) A sur face has, f or i ts b ase, a
circle of radius a. Sections made by
{ y- - - p lanes perp endicular to a certain
xZ
diameter of the circle are isosceles
Fig. 18.5 triangles of height h. Find the
volume.
We take the base circle as
{x2+y2=a2, z=0}
and the particular diameter as the x-axis.
The triangular section taken through (x, y) on the circle has altitude h
and base length 2-\/a2 - x2. Its area is then
A(x) =h \/a2 -x2=DxV.
The evaluation of the radical in the right member of this expression has
already been considered in Par. 9.3 in connection with the area of the circle
of radius a. There, for D.A = a2 - x2, we found in the range of x from
-a to +a that A = ira2/2.
Accordingly, for the volume standing on the entire circular base
V = 2a2h (units)3
This volume may be pictured if one imagines a rubber sheet spanning

the circle. If a diameter beneath the sheet be lifted as a ridge pole of a tent
the rubber sheet forms the lateral surface.
(E) The area bounded by
{y2=x, x=0, y=4}
in the plane z = 0, is revolved about the y-axis. Find the volume generated.
A plane section of the surface at P(x, y) on y2 = x is a circle of radius x.
Its area is
A (y) = 7rx2 = iry' =
Thus
V=75 +C and C=O.

For y = 4,
1024
V= 7r (units)
5
Fig. 18.6 Fig. 18.7
(F) Find the volume generated by revolving about the y-axis the
area bounded by
y =2x-:0 and y = 0, in the plane z = 0.
For a selected y there are two points P,, P2 on the curve with x-values given
by
x2- 2x + y = 0;
PI:x=1-
and
P2:x=1+ Y.
VOLUMES Sec. 18.2
234
The plane section of the surface at P is bounded by two circles (a hollow

disc) of radii 1 - N/1 - y and 1 + N/1-- V. Thus
A (y) = 7r[1 + - /1 - y]2 - 1r[1 - N/1 - T

= 4a(1 - y)
Accordingly,
47r(1 - y)"2
From this,
(1 - y)3/2 + C.
V = -47r-
2
But for y = 0, l' = 0 and thus C = 81r/3. For y = 1 (the greatest or-
dinate),
8
87 (units),
EXERCISES
1. The base of a solid is bounded in the plane z = 0 by y2 = x and x = 1.
Sections made by planes x = k are squares. Sketch and find the volume.
2. The base of a solid is bounded in the plane z = 0 by y2 = x and y = x2.
Sections made by planes y = k are equilateral triangles. Sketch and find the
volume.
3. Sketch and find the volume of the solid described in Problem 2 if the sec-
tions are semicircles.
4. The base of a solid is bounded in the plane z = 0 by x2 + y2 = 2x. Sections
made by planes x = k are equilateral triangles. Sketch and find the volume.
5. Sketch and find the volume of the solid described in Problem 4 if the sec-
tions are semicircles.
6. A twisted horn (Fig. 18.8) is such that every plane section perpendicular
to a fixed (vertical) direction is a circle with radius proportional to its
distance from the vertex. Find the volume for horn height h.
7. A solid has its base bounded by Ix + 2z = 4, y = 01. Sections made by
planes x = k are quadrants of circles. Sketch and find the volume.
8. Find the volume of the wedge bounded by x2 + y2 = a2, z = 0, and z = by.
9. The reflecting backboard for a bandstand has the semicircular base
{ x2 + y2 = 2y, z = 0). Sections made by planes y = k are sectors of para-
bolas with foci on the y-axis. Fiad the volume from y = 0 to y = 1.
Fig. 18.8
10. Sketch and find the volume bounded by the cone x2 + y2 = z2 and the
plane z = a.
11. Sketch and find the volume bounded by the cylinder x2 + y2 = 1 and the
planes z = 0, z = 1 - y.
12. Sketch and find the volume bounded by the cone b2x2 + a) = z2 and the
plane z = k.
13. Find the volume of the surface generated by revolving
{+b=
x
a 1,z=0
for x >_ 0, y Z 0 about the x-axis; the y-axis.
14. Find the volume of a sphere of radius a.
15. Find the volume generated by revolving about the indicated axis the area
bounded in the plane z = 0 by the following curves.
(a) {y = x2, x = 1, y = 0} about the x-axis.
(b) {y = (x - 1)(x - 2), y = 01 about the x-axis.
(c) {2y = Vx-, x = 6, y = 0) about the x-axis.
1
(d) {y = x -4- -,
X x = 1, x = 9, y = 01 about the x-axis.
(e) { y = 1/x, y = x} about the x-axis; the y-axis.

(f) {y = 7 - x2, xy = 6) about the x-axis; the y-axis.
(g) b2x2 + a2y2 = a2b2 about the x-axis; the y-axis.
(11) 1 aye = z I, y = a, x = 0 } about the line y = a.
19.
Quadric Surfaces
The most general equation in rectangular coordinates of the second de-
gree with real coefficients is
f(x, y, z) = Axe + By2 + Cz2 + 2(Dyz + Exz + Fxy)

(1)
+2(Gx+Hy+ Jz) +K =0
in which not all coefficients of the second degree terms are 0. Its graphical
representation is called a quadric surface. Equation (1) may be displayed
to advantage as the matrix product
A F E G x
F B D H y
0
(xyzl)
L+' D C J z
G H J K 1
19.1. Plane Sections

Outstandingly important is the
fact that all plane sections of a
quadric surface are conics. This is
intuitively evident if the configura-
tion of plane and quadric be moved
rigidly so that the intersecting plane
is coincident with the xy-plane X
and had equation z = 0. The inter- Fig. 19.1
236
Sec. 19.1 QUADRIC SURFACES 237
section is then (in new variables)

(Ax2+By2+2Fxy+2Gx+2Hy+K=O, z=0},
recognizable as a conic.
* 19.2. Parallel Chords

Consider the family of parallel lines*
(x = xo + At, y = yo + At, z = zo + vtl, (2)
all with direction cosines EX, µ, v], through points (x, y, z) of the surface
and arbitrary points (xo, yo, zo) (Fig. 19.2).
Since [A, µ, v] are direction cosines, t measures distances along each line
from (xo, yo, zo) to points on the surface. We eliminate x, y, z between equa-
tions of line and surface to find
f(x(,, yo, zo) + 2[(AAxo + Bµyo + Cvzo)
+ D(µzo + vyo) + E(vxo + Azo) + F(Ayo + µxo)
+ GA + Hµ + Jv]t
+ [AA2 + Bµ2 + CO + 2Dµv + 2EAv + 2FAµ]t2 = 0, (3)
a quadratic in t. We call its roots ti, t2.t For given (xo, yo, zo), the numbers
t1, t2 are directed distances from the given point to the surface.
* 19.3. Diametral Planes

We seek now the set of those
points (xo, ye, zo) which are mid-
points of such parallel chords; i.e.,
for which ti = -t2 or t, + t2 = 0.
The quadratic then must have its
coefficient of t equal to 0; i.e.,
(Axe + Fyo + Ezo + G)A
+ (Fxo + Byo + Dzo + H)µ
+ (Exo+Dyo+Czo+J)v
Fig. 19.2
= 0. (4)
* The student here should review Paragraphs 11.6 and 11.7.
f If the coefficients in this quadratic are all 0, the line lies entirely on the surface.
Why?
238 QUADRIC SURFACES Sec. 19.3
This is of the first degree in xo, yo, zo, and thus the locus of the midpoints of
parallel chords is a plane. Since it contains all midpoints, it must contain
the center of the surface-if there is a center.
Let us now consider various families of parallel lines (2) by assigning
various sets of values to [X, jc, v]. Each family gives a diametral plane (4).
And all these planes pass through the intersection of
A xo + Fyo + Ezo + (: = 0 1
Fxo + Byo + Dzo + H = (5)
Exo+Dyo+Czo+J=0J
For these are coefficients of [X, µ, v] in (4), and with their common solution
(xo, yo, zo), arbitrary values of [X, µ, v] satisfy (4).
If the determinant of coefficients in (5),
AFE
A° F B D, (6)
ED C.
is not 0, we may solve for xo, yo, zo, and these are coordinates of the center
of the quadric surface. If this determinant is 0, there is no center.
For example, find the center of the quadric surface
yz + xz + xy - x + 2y - z = 9.
Identifying coefficients by comparison with the form (1), we find
A=B=C=0;D=E=F=1;(:=JH=1;K=-9.
Equations (5) are then (dropping subscripts)
y+z=1, x+z=-2, x+y=1,
and these planes meet at (- 1, 2, - 1), the center of the surface.
EXERCISES
1. Find the centers of the following quadric surfaces.
(a) x2 - xz - yz - z = 0 [Ans: (0, -1, 0)]
(b) x2 - 2y2 + 6z2 + 12xz = 11 [Ans: (0, 0, 0)]
(c) 36x2 - 9y2 - 4z2 - 72x + 36y + 24z = 72 [Ans: (1, 2, 3)]
(d) 4x2 + y2 + 8yz - 4xy + 24x - 20y - 48z = -7 [Ans: (0, 6, 1)1
The intersection of Equations (5) may not be a single point but a line of
points-or even a plane of points. If such a line does not lie on the surface,
it is the central line. This is the case in Exercise 2. The case of a central
plane is involved in Exercise 3.
EXERCISES
2. Find the equations of the central line of the quadric surface
(a) x2 + 4y2 - z2 + 4yz + 2xz + 4xy + 2x + 4y - 2z - 0
(b) x2 + 4y2 + 12yz + 6xz + 4xy - 8x - 16y - 6z = 0
3. Show that the plane x - y + z = # is central to the surface
x2+y2+z2-2yz+2xz-2xy-x+y-z=0
4. Show that x2 + 4y2 + z2 - 4yz - 2xz + 4xy + lOx + 5y - 7z + 15 = 0
has no center.
fc 19.4. Principal Planes

A principal plane of a quadric is one that is perpendicular to the family
of parallel chords that it bisects. We re-express equation (4) of a diametral
plane in the form
(A),+ Fµ+ Ev)xo+ (FX + Bu+Dv)yo+ (EX +Dµ+ Cv)zo
+(GX+Hit +Jv) =0. (7)
't'his bisects the chords with direction [X, µ, v] and is a principal plane if its
normal also has direction [A, u, v]; i.e., if
AX + F;& + Ev = kX, Fa + Bµ + Dv = kµ, EX + Dµ + Cv = kv,
or
(A - k) X + Fµ + Es, = 0
Fa + (B - k)µ + Dv = 0 (8)
Ea + D1 +(C-k)v=0
If there are numbers [X, p, v] (not all 0 since A2 + µ2 + v2 = 1) which
satisfy these equations, then the determinant of coefficients of equations
(8) must be 0; i.e.,
A-k F E
F B-k D =0.
E D C-k
This determinant equation is a cubic in k, called the characteristic of

the surface, and it can be shown that its roots k1, k2, k3 are all real but not
all 0. For each number k; there is a set of direction cosines [A;, µ;, v,]
obtained as the common solution of Equations (8). The three sets of num-
bers A;, u;, v; determine coefficients in Equation (7), thus yielding the
equations of the principal planes.
Illustration. Find principal planes of the quadric
3x2+2y2+z2-4yz-4xy+2=0.
By comparison with (1), coefficients are
A = 3, B = 2, C = 1,
D = -2, E = 0, F = -2,
G=H=J=0, K =2,
and
3 -2 0
0 = -2 2 -2 = -10
0 -2 1
Equations (5) are

3x-2y=0, -2x+2y-2z=0, -2y+z=0,
and their common solution gives the center (0, 0, 0).
The characteristic is
3-k -2 0
-2 2-k -2 =0
0 -2 1 -k
or
k3-6k2+3k+10=0.
Its roots are
k = 2, 5, -1
For each of these numbers k, there corresponds a principal plane whose

normal has direction [A, µ, v] obtained from Equations (8). Equations
(8) are,
(1) fork=2:A-2µ=0, -2A-2v=0, -2µ-v=0, from which,
since A2 + µ2 + v2 = 1,
A=3,
(2) for k=5:-2X-2µ=0, -2X-3p-2v=0, -2p-4v=0,

from which
µ=-1, v=1;
a=1,
(3) for k=-1:4X-2u=0, -2X+3z-2v=0, -2p+2v=0,
from which
X=J, JA v=1.
Principal planes are, accordingly,
2x+y-2z=0, 2x-2y+z=0, x+2y+2z=0.
The student should verify that these are mutually perpendicular. The
surface is a hyperboloid of two sheets.
EXERCISES
5. Find centers and principal planes of
(a) z = xy + 1 (See Par. 17.2)
(b) x2+ y2+4z2- 2x- 4y- 8z+8= 0
(c) x2 - 3y2 + 3z2 + 8yz + 6x + 29 = 0
(d) x2 - 2z2 + 4xz - 4y + 12z + 6 = 0
* 19.5. Classification of Quadric Surfaces

If principal planes of a given quadric are chosen as coordinate planes
and the transformation made, all equations of second degree may be re-
duced to the standard form
Lx2+My2+Nz2+Px+Qy+Rz+ S =0 (9)
in which product terms yz, xz, xy do not appear. The characteristic equation
is unchanged by this transformation. For (9), the characteristic is
IL-k 0 0 I
0 M-k 0 =0
0 0 N-k
or
(k-L)(k-M)(k-N) =0.
It is thus apparent that the roots ki, k2, k2 of the characteristic are the coeffi-
cients of the terms of second degree in Equation (9). From this the following
facts concerning these roots ki and the nature of the quadric are now
evident.
(1) If two roots are equal but different from 0, the quadric is a surface
of revolution.
(2) If two roots are 0, quadratic terms in the equation of the quadric
surface form a perfect square.
(3) If one root is 0 and A 0, the quadric has no center.
(4) If all three roots are equal (0), the quadric is a sphere.
We are now prepared to classify all quadrics into the following cate-
gories. Forms and figures are listed here for occasional reference. Note that
this classification (as in the classification of the conics) is made entirely on
the coefficients of second degree terms. Linear terms give information
concerning the center (or midpoints of chords).
1. Central Quuadrics: L 5 0, M 0, N 5.6 0.
Equation (9) may be written in some one of the following forms, each of
which is now identifiable. Centers are at (u, v, w).
(x - u)2 (y - v)2 (z - w)2

a2
+ b2 + c2
= 1, Ellipsoid
Fig. 19.3
(x - u)2 (y - v)2 (z - w)2 _ Hyperboloid

a2 + b2 mm c2 - 1' (one sheet)
Fig. 19.4
(x-u)2 (y-v)2 (z-v,)2-

b2 C2
Hyperboloid
(two sheets)
Fig. 19.5
(x - u)2 + (y -_v)2
b2 + (z - w)2 No locus
a2 c=
(x 21) + (y y) (z cZw)2 = 0,
a2 b2
+ A point
(x i`)2 v)2 (z - w) z = 0,
a2 + (y. c, ' Cone
b2
II. 1 Non-central Quadr~ics: L Fl- 0, A'! x 0, N = 0.

I
Equation (9) may be written in some one of the forms
(x - u)2 (y - v)2
a2 + b2
=cz,
Elliptic Paraboloid
Fig. 19.7
- -- (y-v)2
(x-u)2
a2 b2
- =cz,
Hyperbolic Paraboloid
("saddle")
Fig. 19.8
or the special cases of cylinders or planes.
III. Non-central Quadrics: L : 0, M = N = 0.

Equation (9) may be transformed into one of the forms
x2 = ay or x2 = bz, Cylinders
Fig. 19.9
x2 = a, Planes
Index
(Numbers refer to pages)
Acceleration 121 Continuity 39

Addition of ordinates 15 Continuum 3
Affine transformation 163 Coordinate systems 4ff, 177ff
Airfoil 173 Critical points 47
Angle between lines 28, 181 values 10
bisectors 77 Crossed parallelogram 113, 136, 141, 142,
Area of triangle 31 172
measure function 60ff Cubics 11
of ellipse 117 Curves 200ff
parametric 63 Cycloid 65, 130ff, 226
polar coordinates 61 Cylinder 222, 244
rectangular coordinates 59 projection 204
Astroid 72, 129, 132 Cylindrical coordinates 178
Asymptotes 13, 104
Derivative 44, 49, 52
Bevel gears 227 Diameters of conies 153
Bisectors 77, 87 principal 155
Diametral planes 237
Cable, parabolic 119 Differentiation, implicit 54, 208
Cams 139 Direction cosines 26, 67, 180
Cardioid 22, 49, 50, 62, 133, 136, 170 Direction numbers 26, 67, 180
Caustic 134 Direction of line 195
Center, instantaneous 127 Directrix 97, 222
Central quadrics 242 Discriminant 151
Centrode 129 Distanee 25, 180
Characteristic equation 156, 240 point to line 75
Circle 80ff point to plane 190
Cissoid 138
Collinearity 31
Conchoid 129
Eccentricity 97
Concurrency 74 Ellipse 95ff
Cone 95, 219, 222ff, 243 Ellipsoid 229, 242
elliptical 203 Elliptic gears 136
Conics 95ff Elliptic paraboloid 243
folding 115 Epicycloid 131, 143
properties 112ff Excenter 78
Construction of conies 109ff Excircle 88
245
246 INDEX
Families of circles 83 Lipkin 171

of lines 70, 109, 218 Loran 124
of planes 188
of surfaces 20.1 Mapping 159ff
Flex points 48 Matrix 161
Folding the conies 115 Maxima, minima 47
Function 33 Mechanical motions 127
differentiation of implicit 54 Midpoint 25
Motion of a point 120
Gears (elliptic, hyperbolic, etc.) 136, 226
Geometric progression 36 Nephroid 134
Graphs (definition) 5 Non-central quadrics 243
of surfaces 202 -Normal line 210
polar 20 plane 207ff
Guide lines 13, 104 Normal to plane 189
Number, irrational 3
Harmonic motion 140 rational 3
Helix 201, 210, 227 Number system 3
Hyperbola 95ff
Hyperbolic gears 136
Hyperbolic paraboloid 243 Oblate spheroid 217
Hyperboloid 216, 218, 221, 242 Ordinates, addition of 15
Hypocycloid 131, 143 reciprocal 18
Hypoid gears 227 Orthocentric set 78, 88, 111
Identification of conics 150 Pantograph 164

Incenter 78 Paper folding 115
Incircle 88 Parabola 9, 58, 95ff
Inflexion 12, 48 Parbolic cable 119
Instantaneous center 127 cylinder 204
Intercept form of line 69 Paraboloid 202, 211, 215
Inverse 135, 173 Parallelism 182
Inversion 169ff Parallelogram, crossed 113
Inversor 171 Peaucellier 171
Involute 139, 226 Pedal 134
Irrational number 3 Perpendicularity 29, 182
Plane motion 120
Joukowski airfoil 173 Planes 184ff, 244
diametral 237
Kite 137 principal 239
Point of inflexion 12, 48
Ladder 128 Polar coordinate system 4
Lemniscate 137, 142, 171 Power of point (Power theorem) 89,171
Line motion 141 Principal diameters 155
Lines 9, 66ff, 184ff planes 239
concurrent 74 Progressions, geometric 36
Linkage motion 135ff, 141, 172 Projectile path 122
Limapon 22, 135, 136, 140 Projection cylinders 204
Limit 34 planes 188
Limit (sin 9)/0 37 Prolate spheroid 217
Limit theorems 39 Properties of conics 112
INDEX 247
Quadrics, central 241 Spirals 24

non-central 243 Archimedean 141
Quadric surfaces 236ff Straight line 66
Stretch 164
Radical axis 91 Surfaces 200ff, 214ff
center 92 quadric 236ff
Range of projectile 123 ruled 218
Rational number 3 Symmetric form 67
Real number system 3
Reciprocal ordinates 18 Tangent length 90
Reflective property 58, 113 Tangent to curve 47, 116, 166, 207ff
Removal of xy term 148 plane 210
Rose 23 Torus 217
Rotation 145, 160 Trace 185
Ruled surfaces 218 Trammel 71, 128
Rulings 218ff Transformations 145ff
affine 163
Saddle 243 orthogonal 167
Scalar product 29, 182 Translation 145, 159
Serpentine 18 Trisection 43
Shear 169
Similitude 163 Variable 33
Skew parallelogram 161 Velocity 120
"lope 30, 46, 69 Volume 228ff
Spherical coordinates 178
Spiral gear 227 Witch 18

Analytic Geometry With Calculus PDF

Uploaded by

Copyright:

Available Formats

Analytic Geometry With Calculus PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Analytic Geometry With Calculus PDF

Uploaded by

Copyright:

Available Formats

311AIVNV 39 13W

Englewood Cliffs, N. J. PRENTICE-HALL, INC. 1961

Library of Congress Catalog Card No. 61-9224

Printed in the United States of America

three semester hours, certain sections (indeed, whole chapters) marked

1. The Real Number System-Graphs . . . . . . . . . . 3

3. Variables, Functions, Limits, Continuity . . . . . . . 33

4.1. The Derivative in Rectangular Coordinates . . . . . . . . 44

5. Functions Whose Derivatives Are Known . . . . . . . . 56

5.1. Determination of the Function . . . . . . . . . . . . . 56

6. The Straight Line . . . . . . . . . . . . . . . . . . . 66

7.1. General Equations . . . . . . . . . . . . . . . . . . . 8o

8.1. General. Discussion . . . . . . . . . . . . . . . . . . . 95

9. The Conies: Their Properties and Applications . . . . . 112

10. Some Mechanical Motions and Loci . . . . . . . . . . 127

11. Transformations . . . . . . . . . . . . . . . . . . . 145

12. * Mapping . . . . . . . . . . . . . . . . . . . . . . 159

13. Coordinate Systems-Measures . . . . . . . . . . . . 177

13.1. Coordinate Systems . . . . . . . . . . . . . . . . . . 177

14. Planes and Lines . . . . . . . . . . . . . . . . . . . 184

14.1. Definition . . . . . . . . . . . . . . . . . . . . . . 184

15. Surfaces and Curves . . . . . . . . . . . . . . . . . . 200

15.1. Definition . . . . . . . . . . . . . . . . . . . . . . 200

16. Tangents and Normals . . . . . . . . . . . . . . . . 207

16.1. Tangent Line and Normal Plane to a Curve . . . . . . . 207

17. Special Surfaces . . . . . . . . . . . . . . . . . . . . 214

18. * Volumes . . . . . . . . . . . . . . . . . . . . . . 228

19. Quadric Surfaces . . . . . . . . . . . . . . . . . . . 236

The Real Number

1.1. The Real Number System and the Continuum Axiom

rational numbers may then be assigned to intermediate points and the

1.2. Coordinate Systems in the Plane

Fig. 1.2 Fig. 1.3

1.3. Graphs: Mathematical Statements

roordinates satisfy such a statement forms the graphical representation or locus

Fig. I.6 Fig. 1.7

As another example, r = 2 sin 0 represents a circle. A table of coordi-

3. Convert to radian measure

210', 225°, 240°, 315°, 330°, - 30°, -225-

4. Convert to degree measure

5. Sketch the region of the plane such that

(d) {7r/4 < 8 < 3ir/4, r sin B < 31

is a straight line, and the location of y

is a parabola of the types shown. The domain of x is unrestricted, and for

it is evident that lines y = constant

0 = (XI - x2) [a(xl + x2) + b].

The last equation then approaches

(x, + x2) /2.

(3) The graph of

0 = a(x; - x2) + b(x, - x2) + c(xi - x2)

0 = (xl - x2) [a(xi + x1x2 + x2) + b(x, + x2) + c].

We now shall move the curve

(s) y = x2(1 - x) (t) y = (x - 1)2(2 - x)

1.5. Guide Lines (Rectangular Coordinates)

For values of x near +1 or -1, the denominator of the fraction is near 0

Fig. 1.15 Fig. 1.16

Furthermore, if we write the equation as

value -x - 1. Accordingly, y = jx - 1 is a guide line as shown. Note that

*1.6. Techniques (Rectangular Coordinates)